Interventional Neuroradiology

Interventional Neuroradiology

Interventional Neuroradiology Edited by

Robert W. Hurst

Hospital of the University of Pennsylvania Philadelphia, Pennsylvania, USA

Robert H. Rosenwasser

Thomas Jefferson University Philadelphia, Pennsylvania, USA

Informa Healthcare USA, Inc. 52 Vanderbilt Avenue New York, NY 10017 # 2008 by Informa Healthcare USA, Inc. Informa Healthcare is an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-8493-9562-3 (Hardcover) International Standard Book Number-13: 978-0-8493-9562-8 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequence of their use. No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Interventional neuroradiology / edited by Robert W. Hurst, Robert H. Rosenwasser. p. ; cm. Includes bibliographical references. ISBN-13: 978-0-8493-9562-8 (hardcover: alk. paper) ISBN-10: 0-8493-9562-3 (hardcover: alk. paper) 1. Nervous system—Interventional radiology. I. Hurst, Robert W. II. Rosenwasser, Robert H. [DNLM: 1. Cerebrovascular Disorders—radiotherapy. 2. Cardiovascular System—anatomy & histology. 3. Central Nervous System—blood supply. 4. Radiology, Interventional—methods. WL 355 I6074 2008] RD594.15.I62 2008 616.80 04757—dc22 Visit the Informa Web site at www.informa.com and the Informa Healthcare Web site at www.informahealthcare.com

2007023346

To the mentors, students, and patients who have shown me how much there is to learn and all too often, how little time in which to accomplish it. To my wonderful wife Marilyn, and my children, Jonathan and Katherine, who make family the greatest happiness of my life. I must thank them for the commitment in time that has made this endeavor possible. —Robert W. Hurst

I would like to dedicate this book to my wife Deborah August, M.D., who has been my partner and pillar of strength and supported me without hesitation in all my endeavors. In addition, I wish to express my gratitude to William A. Buchheit, M.D., my Neurosurgical Mentor and Friend, and the individual who supported the early concept of Endovascular Therapy for disorders of the Nervous System . . . a man way ahead of his time. —Robert H. Rosenwasser

Preface

This book is intended to provide the clinical practitioner with background information and specific descriptions of the anatomy, techniques, disorders, procedures, and decisions most commonly encountered in interventional neuroradiology. Throughout the past decade, interventional neuroradiological techniques have revolutionized therapy for vascular disorders of the head, neck, and central nervous system. These procedures now provide noninvasive treatment for many of the most common neurological disorders and make possible treatment of numerous patients for whom there were no reasonable therapeutic options before. With progress, however, comes the requirement for increased knowledge and technical skill to deliver these treatments safely and effectively. Areas of fundamental knowledge in interventional neuroradiology cross the boundaries of classically delineated medical and surgical specialties, including neurosurgery, neuroradiology, and neurology. Required knowledge includes familiarity with neuroradiological imaging of vascular disease, knowledge of vascular anatomy, and thorough understanding of cerebrovascular disorders and their endovascular treatments. Most importantly, skill in basic interventional techniques must be coupled with good clinical judgment in patient management and decision making. Recent rapid advances in neuroimaging mean that practitioners of interventional neuroradiology must have excellent diagnostic skills with noninvasive neuroimaging modalities to identify the presence of cerebrovascular disease, evaluate its effects, identify potential candidates for neurointerventional procedures, and document the effects of the treatment. Separate chapters on CT, MR, and ultrasonographic evaluation of cerebrovascular disease emphasize the current noninvasive evaluation of disorders that are of interest to neurointerventionalists. In addition, the authors have made every effort throughout the text to illustrate the integration of current neuroimaging into the performance and decision making associated with interventional neuroradiological procedures. As in all radiological- or surgical-based specialties, thorough understanding of pertinent anatomy is essential. For the neurointerventionalist, cerebrovascular anatomy is the workplace. Anatomic knowledge underlies the understanding of many, if not all, cerebrovascular disorders, provides routes of endovascular access, and defines the scope of treatment options. Chapters covering pertinent vascular anatomy of special importance to neurointerventional procedures have been included. These chapters are directed at key anatomic concepts as well as specific anatomic features of the head, neck, brain, and spine vasculature. It is through basic neurointerventional techniques that treatment is delivered to the individual patient. No amount of theoretical understanding can overcome poor technique in an environment as unforgiving as the cerebrovascular system. Discussion of basic techniques with appropriate illustrations should prove useful for readers at all levels of experience, from students entering the field to experienced practitioners who may benefit from review or additional technical options. Coupled with anatomic and technical knowledge is the requirement for understanding the epidemiology, pathophysiology, and clinical features of the increasing numbers of cerebrovascular disorders that are now amenable to endovascular treatment. Recognized experts in the field have authored clinically oriented discussions of the most common conditions of interest to interventional neuroradiologists. Treatment discussions are illustrated with current images to emphasize pertinent technical and anatomic details. Extensive and current references are included to serve as a basis for further research.

vi

Preface

Perhaps most essential to successful neurointerventional practice is the requirement for correlating the appropriate application of knowledge and technical skills to the care of patients. This book is designed to illustrate and emphasize the importance of integrating clinical information, knowledge of disease processes, and technical skill through the use of good clinical judgment to formulate and perform effective neurointerventional procedures. Robert W. Hurst Robert H. Rosenwasser

Contents

Preface . . . . v Contributors . . . . ix 1. Vascular Anatomy of the Head, Neck, and Skull Base . . . . . . . . . . . . . . . . . . . . . 1 Michele H. Johnson, Hjalti M. Thorisson, and Michael L. DiLuna 2. Applied Neurovascular Anatomy of the Brain and Skull . . . . . . . . . . . . . . . . . 23 Randy S. Bell, Alexander H. Vo, and Rocco A. Armonda 3. Vascular Anatomy of the Spine and Spinal Cord . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Armin K. Thron 4. Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 David S. Liebeskind 5. CT Imaging and Physiologic Techniques in Interventional Neuroradiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 Ronald L. Wolf 6. MR Angiography: Principles and Applications in Interventional Neuroradiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Neerav R. Mehta and Elias R. Melhem 7. Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Jaroslaw Krejza 8. Techniques and Devices in Interventional Neuroradiology . . . . . . . . . . . . . . 161 Jeffrey M. Katz, Y. Pierre Gobin, and Howard A. Riina 9. Balloon Occlusion, Wada, and Pharmacological Testing . . . . . . . . . . . . . . . . . 183 Linda J. Bagley 10. Endovascular Management of Tumors and Vascular Malformations of the Head and Neck . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Johnny C. Pryor, Joshua A. Hirsch, and Robert W. Hurst 11. Dissections of the Carotid and Vertebral Arteries . . . . . . . . . . . . . . . . . . . . . . . . 213 Qaisar A. Shah, Scott E. Kasner, and Robert W. Hurst 12. Direct Carotid Cavernous Fistula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 Uday S. Kanamalla, Charles A. Jungreis, and Jeffrey P. Kochan 13. Endovascular Management of Intracranial Aneurysms . . . . . . . . . . . . . . . . . . . 239 Darren Orbach, Tibor Becske, and Peter Kim Nelson 14. Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Jonathan L. Brisman, David W. Newell, and Joseph M. Eskridge 15. Endovascular Management of Brain Arteriovenous Malformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 John B. Weigele, Riyadh N. Al-Okaili, and Robert W. Hurst

viii

Contents

16. Endovascular Treatment of Acute Ischemic Stroke . . . . . . . . . . . . . . . . . . . . . . . 305 Mayur A. Paralkar, Alexandros L. Georgiadis, Adnan I. Qureshi, and Qaisar A. Shah 17. Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Eric Sauvageau, Robert D. Ecker, Junichi Yamamoto, Ramachandra P. Tummala, Elad I. Levy, and L. Nelson Hopkins 18. Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 Nabil M. Akkawi and Ajay K. Wakhloo 19. Endovascular Management of Dural Arteriovenous Fistulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335 J. Marc C. van Dijk, Robert A. Willinsky 20. Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 Nicholas J. Patronas and Donald L. Miller 21. Endovascular Treatment of Spinal Vascular Malformations . . . . . . . . . . . . . 363 Mayumi Oka and Kieran Murphy 22. Percutaneous Vertebroplasty Mary E. Jensen Index . . . . 411

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

Contributors

Nabil M. Akkawi Division of Neuroimaging and Intervention, University of Massachusetts Medical School, Worcester, Massachusetts, U.S.A. Riyadh N. Al-Okaili Department of Radiology, King Abdulaziz Medical City, Riyadh, Saudi Arabia. Rocco A. Armonda Departments of Neurosurgery and Radiology, National Naval Medical Center, and Comprehensive Neurosciences Program, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A. Linda J. Bagley Departments of Radiology and Neurosurgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A. Tibor Becske Departments of Neurology, Neurosurgery, and Radiology, New York University Medical Center, New York, New York, U.S.A. Randy S. Bell Departments of Neurosurgery and Radiology, National Naval Medical Center, and Comprehensive Neurosciences Program, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A. Jonathan L. Brisman Department of Cerebrovascular and Endovascular Neurosurgery, Winthrop University Hospital, Mineola, Long Island, New York, U.S.A. Michael L. DiLuna Department of Neurosurgery, Yale University School of Medicine, New Haven, Connecticut, U.S.A. Robert D. Ecker Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A., and Department of Neurological Surgery, U.S. Naval Hospital, Okinawa, Japan. Joseph M. Eskridge Department of Interventional Neuroradiology, Seattle Neuroscience Institute, Seattle, Washington, U.S.A. Alexandros L. Georgiadis Department of Neurology, Zeenat Qureshi Stroke Research Center, University of Minnesota, Minneapolis, Minnesota, U.S.A. Y. Pierre Gobin Departments of Radiology and Neurosurgery, New York Presbyterian Hospital, Weill Medical College of Cornell University, New York, New York, U.S.A. Joshua A. Hirsch Department of Interventional Neuroradiology and Endovascular Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A. L. Nelson Hopkins Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A. Robert W. Hurst Departments of Radiology, Neurology, and Neurosurgery, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A. Mary E. Jensen Departments of Radiology, and Neurosurgery, University of Virginia Health Systems, Charlottesville, Virginia, U.S.A.

x

Contributors

Michele H. Johnson Interventional Neuroradiology, Departments of Diagnostic Radiology and Surgical Otolaryngology, Yale University School of Medicine, New Haven, Connecticut, U.S.A. Charles A. Jungreis Temple University Hospital, Temple University School of Medicine, Philadelphia, Pennsylvania, U.S.A. Uday S. Kanamalla Temple University Hospital, Temple University School of Medicine, Philadelphia, Pennsylvania, U.S.A. Scott E. Kasner Department of Neurology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A. Jeffrey M. Katz Department of Radiology, New York Presbyterian Hospital, Weill Medical College of Cornell University, New York, New York, U.S.A. Jeffrey P. Kochan Temple University Hospital, Temple University School of Medicine, Philadelphia, Pennsylvania, U.S.A. Jaroslaw Krejza Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.; Department of Nuclear Medicine, Medical University of Gdansk, Poland. Elad I. Levy Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A. David S. Liebeskind UCLA Stroke Center, University of California, Los Angeles, California, U.S.A. Neerav R. Mehta University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A. Elias R. Melhem University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A. Donald L. Miller Department of Radiology, National Naval Medical Center and Department of Radiology, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A. Kieran Murphy Department of Radiology, Division of Interventional Neuroradiology, Johns Hopkins University, Baltimore, Maryland, U.S.A. Peter Kim Nelson Departments of Neurology, Neurosurgery, and Radiology, New York University Medical Center, New York, New York, U.S.A. David W. Newell Department of Neurosurgery, Seattle Neuroscience Institute, Seattle, Washington, U.S.A. Mayumi Oka Department of Radiology, Division of Interventional Neuroradiology, Johns Hopkins University, Baltimore, Maryland, U.S.A. Darren Orbach Departments of Neurology, Neurosurgery, and Radiology, New York University Medical Center, New York, New York, U.S.A. Mayur A. Paralkar Department of Medicine, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, U.S.A. Nicholas J. Patronas Department of Radiology, National Institutes of Health Clinical Center, Bethesda, Maryland, U.S.A. Johnny C. Pryor Department of Interventional Neuroradiology and Endovascular Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A.

Contributors

Adnan I. Qureshi Department of Neurology, Zeenat Qureshi Stroke Research Center, University of Minnesota, Minneapolis, Minnesota, U.S.A. Howard A. Riina Departments of Radiology and Neurosurgery, New York Presbyterian Hospital, Weill Medical College of Cornell University, New York, New York, U.S.A. Eric Sauvageau Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A., and Department of Neurological Surgery, University of South Florida College of Medicine, Tampa, Florida, U.S.A. Qaisar A. Shah Department of Neurology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A., and Department of Neurology, University of Minnesota, Minneapolis, Minnesota, U.S.A. Hjalti M. Thorisson Department of Diagnostic Radiology, Yale University School of Medicine, New Haven, Connecticut, U.S.A. Armin K. Thron Department of Neuroradiology, University Hospital, RWTH Aachen University, Aachen, Germany. Ramachandra P. Tummala Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A. J. Marc C. van Dijk Department of Neurosurgery, University Medical Center, Groningen, Groningen, The Netherlands. Alexander H. Vo Departments of Neurosurgery and Radiology, National Naval Medical Center, and Comprehensive Neurosciences Program, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A. Ajay K. Wakhloo Division of Neuroimaging and Intervention, University of Massachusetts Medical School, Worcester, Massachusetts, U.S.A. John B. Weigele Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A. Robert A. Willinsky Department of Medical Imaging, Toronto Western Hospital, Toronto, Ontario, Canada. Ronald L. Wolf Department of Radiology, Neuroradiology Section, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A. Junichi Yamamoto Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A.

xi

1 Vascular Anatomy of the Head, Neck, and Skull Base Michele H. Johnson, Hjalti M. Thorisson, and Michael L. DiLuna Interventional Neuroradiology, Departments of Diagnostic Radiology and Surgical Otolaryngology;, Department of Diagnostic Radiology; and , Department of Neurosurgery, Yale University School of Medicine, New Haven, Connecticut, U.S.A.

INTRODUCTION The emphasis of this chapter is on the anatomy and anatomic variations of the vasculature of the head and neck beginning in the thorax at the level of the aortic arch and extending superiorly to the level of the skull base (vascular entrance through the dura). Selective catheterization is predicated on familiarity with these anatomic features. Cross-sectional (vascular) imaging, including CTA and MRA, has supplanted catheter studies for the purposes of pure diagnosis. Identification of the common and uncommon variations and their adjacent soft tissue relationships is important to the neurointerventionalist when assessing the crosssectional imaging prior to therapeutic intervention. The anatomy of this region will be explored using a combination of CTA, MRA, and conventional angiographic images and case examples to demonstrate features important to the neurointerventionalist (1).

EMBRYOLOGY The embryology of the aortic arch development is complex and beyond the scope of this chapter; however, a few relevant embryologic considerations provide a basis for understanding important normal variants that may have an impact on catheterization and image interpretation (2–6). The convexity of the aortic arch forms from the left fourth primitive aortic arch. The innominate or brachiocephalic artery (BCA), the left common carotid artery (LCCA), and the left subclavian artery (LSUB) arise sequentially from the aortic arch (from proximal to distal) (Fig. 1A). In the majority of cases, the LCCA arises distinctly separate from the BCA; however, in approximately 20% of patients, the LCCA may arise in conjunction with the BCA in a bovine configuration (Fig. 1B) (7,8). In a small percentage of patients, the left vertebral artery may arise as a branch of the aortic arch (Fig. 1C). Even more rarely, the right vertebral artery may arise directly from the aortic arch (Fig. 1D, E) (9).

In rare cases, the arch is derived from the right primitive arch and the brachiocephalic vessels arise as a mirror image of the normal arrangement (Fig. 1F). More commonly, an aberrant right subclavian artery (RSUB) is present that is characterized by the right common carotid as the first branch from the aortic arch, followed by the LCCA, the LSUB, and finally the RSUB, which arises distally and proceeds toward the right behind the esophagus to give rise to the right vertebral artery and remaining subclavian artery branches (Fig. 1G). A focal dilatation of the aorta adjacent to the origin of the aberrant right subclavian is referred to as Kommerell’s diverticulum and may occasionally become aneurysmal and require surgical repair (Fig. 1H) (10,11).

AORTIC ARCH AND BRANCHES The aorta arises from the heart and emerges from the pericardium in the superior mediastinum, where it forms the ascending aortic arch (AOA) anterior to the trachea at the level of the sternal manubrium. From this ascending arch arise three major branches: the BCA, the LCCA, and the LSUB (Fig. 1A). The BCA crosses obliquely cephalad into the right anterior to the trachea before bifurcating into the right common carotid artery (RCCA) and RSUB behind the sternoclavicular joint. Fluoroscopic recognition of the head of the clavicle as the location of the bifurcation of the BCA can be a useful adjunct to selective catheterization of the RCCA and RSUB (Fig. 2). The anterior location of the RCCA in relationship to the RSUB can be exploited in the selective catheterization of the subclavian artery by turning the patient’s head toward the left and extending the arm to accentuate the separation between these two vessels. The right vertebral artery arises from the RSUB just opposite the origin of the internal mammary (INM) artery. The left vertebral artery can arise directly from the aorta in 5% of cases (9). Additional subclavian branches include the ascending cervical artery, the thyrocervical trunk,

2

Johnson et al.

Figure 1 (A) Normal LAO arch configuration. Note the typical configuration of the great vessels and the marked vertebral artery asymmetry (right > left). (B) Bovine arch. Note the common origin of the BCA and the LCCA. The left vertebral artery is larger than the right. (C) LAO arch injection demonstrates the origin of the left vertebral artery from the aortic arch between the origins of the LCCA and the left SUB. Note the absence of vertebral originating from the left SUB. (D) Right vertebral artery arising from the arch demonstrated on posterior view of 3D CTA. (E) Spontaneous aortic dissection in a patient with aberrant right subclavian and a bovine arch configuration. The patient presented with chest and right arm pain. Note the false lumen (FL) and the dissection flap (arrows). (F) Ehlers-Danlos with aberrant right subclavian, bovine origin, and multiple aneurysms (arrows). (G) Right aortic arch with aberrant left subclavian and tracheal ring. Note the diverticulum of Komerell (arrows). (H) Massive oral bleeding. Aortic arch arteriogram demonstrates a normal arch confirguration; however, there is an increased distance between the BCA and RCCA and the LCCA (arrows) secondary to mediastinal hematoma. (I) Massive oral bleeding is associated with extravasation of contrast from this left common carotid blow-out. (J) Massive oral bleeding is associated with extravasation of contrast from this left common carotid blow-out. Abbreviations: BCA, brachiocephalic artery; LCCA, left common carotid artery; SUB, subclavian artery; RCCA, right common carotid artery; LAO, left anterior oblique.

Figure 2 (A,B) 3D CTA demonstrates the normal relationships of the BCA as it bifurcates into the subclavian and carotid arteries on the right bifurcation. The BCA road-map image demonstrates the clavicle as a landmark for the bifurcation in the AP view. Abbreviation: BCA, brachiocephalic artery.

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

3

Figure 3 Proximal subclavian branches (A–C) SUB injection demonstrates proximal branches supplying T2 vertebral tumor. (D) The ascending cervical artery is a potential collateral source to the vertebral artery. Abbreviation: SUB, subclavian artery.

Figure 4 Vertebral artery cervical branches. AP view (A) and lateral view (B) of the cervical vertebral artery demonstrate small muscular and vertebral body branches (arrows).

and the costocervical trunk (Fig. 3A–C). These branches are important to identify in the analysis of pathologic processes of the lower neck as well as vascular malformations and other pathologic lesions involving the cervical and/or upper thoracic vertebral bodies and spinal cord.

VERTEBRAL ARTERIES The vertebral arteries ascend posterior to the common carotid between the longus colli and scalenus anterior muscles, entering the transverse foramen at C6. They traverse the transverse foramen of the cervical vertebral body between C6 and C2. After exiting the transverse foramen at C2, the vertebral artery proceeds posterolaterally through the transverse foramen of C2 and posteromedially between C1 and the occiput, before entering the foramen magnum (1,4,5). The cervical vertebral artery provides small branches to supply the vertebral bodies and the adjacent cervical musculature (Fig. 4A, B). The cervical course is usually straight, although tortuosity

may limit distal microcatheterization and/or may lead to confusion when the transverse foramen is enlarged (Fig. 5). It is also important to recognize the potential for luminal narrowing and/or flow alteration within the vertebral artery as a consequence of normal head turning. This normal phenomenon may be accentuated by the presence of osteophytes encroaching on the artery within the transverse foramen (12). Provocative maneuvers during angiography or, alternatively, during noninvasive vascular imaging may demonstrate these findings, which may correlate with clinical hypoperfusion symptoms such as lightheadedness or vertigo (Fig. 6A, B) (13,14). The left vertebral artery is dominant (larger and responsible for the majority of the posterior fossa flow) almost half of the time, while the right vertebral artery is dominant 25% of the time (12,15–17). No size or flow dominance is present in the remaining cases (12,15–17). Anastomoses exist at multiple levels with the external carotid artery (ECA), the thyrocervical trunk, and the costocervical trunk.

4

Johnson et al.

Figure 5 Vertebral artery tortuosity versus dissection on CTA (A). Note the tortuosity without dissection flap on the AP angiogram (B).

Figure 6 Syncope on head turning. (A) Left vertebral artery: neutral position. Note the compression of the cervical vertebral artery by uncovertebral joint degenerative osteophytes accentuated on moderate (B) and maximal (C) head turning.

The vertebral arteries proceed through the dura at the level of the foramen magnum and join to form a common basilar artery. The posterior inferior cerebellar artery (PICA) is the largest, though frequently variable, branch of the vertebral artery and usually arises proximal to origin of the basilar artery. It can arise as a single trunk or in duplicate, and occasionally the vertebral artery can terminate as the PICA (18,19). There is a balance between distal branches of the PICA and hemispheric branches of the anterior inferior cerebellar artery (AICA) such that an AICAPICA variant may be an absent PICA, with the PICA territory supplied by distal branches of the AICA, or vice versa (18–20). (Fig. 7A, B) The posterior spinal artery often arises from the vertebral artery at the level of the medulla oblongata

or may arise from the PICA, coursing posteriorly and dividing into anterior and posterior branches to anastomose with small perforators from the vertebral artery. The ascending cervical artery, posterior intercostal arteries, and lumbar arteries may each contribute collateral supply to the posterior spinal arteries at their respective levels. The anterior spinal artery arises from the distal end of the vertebral artery and descends anterior to the medulla oblongata, joining with its contralateral branch to descend as a single vessel, forming multiple anastomoses with similar segmental perforators (as the posterior spinal artery), to supply the anterior spinal cord to the filum terminale. The posterior meningeal artery arises from the cervical vertebral artery to supply the bone and dura of the posterior fossa (12). Multiple small spinal

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

5

Figure 7 Distal vertebral artery variations. (A) AP and (B) lateral vertebral ends in PICA. Vertebral artery fenestrations (C–E); T1-weighted sagittal MRI (F) and AP (G) and lateral (H) vertebral angiograms demonstrate an AVM fed by the ASA; AP (I) and lateral (J) views demonstrate the origin of PICA below the foramen magnum. Abbreviations: AVM, arteriovenous malformation; ASA, anterior spinal artery; PICA, posterior inferior cerebellar artery.

branches enter the vertebral canal through the intervertebral foramina to supply the spinal cord. Muscular branches at the level of the lateral mass of C1 supply the deep cervical musculature.

medially to enter the carotid canal at the skull base (Fig. 8C, D).

COMMON CAROTID ARTERIES

The ECA arises at the bifurcation of the common carotid artery in the neck and supplies the face, scalp, and dura primarily, with potential collateral contributions to the brain parenchyma and orbital contents (23). The ECA branches have many variations. (Fig. 9). However, true ECA anomalies are rare, the most common being a so-called nonbifurcated common carotid artery, where the ECA branches arise separately from the common carotid trunk (24,25). Anomalous origin of the ECA from the aortic arch is also rarely encountered (26). The ECA courses anterolaterally from its initial position along the lateral pharyngeal wall as it passes beneath the posterior belly of the digastric and stylohyoid muscles and pierces the parotid fascia. The deep lobe of the parotid gland separates the ECA from the ICA (1,4, 5). Two schemes for categorizing the ECA branches according to cranial caudal or anterior and posterior locations have been proposed to predict the vascular source of neovascularity or bleeding on the basis of cross-sectional imaging prior to intervention. In one scheme, the ECA branches are

The common carotid arteries proceed cephalad within the fibrous carotid sheath along with the internal jugular vein, the vagus nerve, and the ansa cervicalis. The common carotid arteries have no normal branches before the carotid bifurcation, although rare variations may occur (Fig. 8A, B) (20). The terminal common carotid artery dilates to form the carotid bulb and bifurcates into the ICA and ECA. The bifurcation is typically located between the level of thyroid cartilage and the greater horn of the hyoid bone, although carotid bifurcations may lie either above or below this level (reported at the C1–C2 to the C6–C7 levels) (21,22). The anatomic level of the carotid bifurcation is more important when surgical rather than endovascular correction of carotid atherosclerotic disease is planned. The bifurcation is located between C3 and C5 in approximately 80% of patients, with the next common location at the C5–C6 level (13%) (22,23). The internal carotid artery (ICA) courses posterolateral to the ECA and then proceeds

External Carotid Artery

6

Johnson et al.

Figure 8 Cervical carotid variations. (A) Normal bifurcation, (B) cervical loop, and (C) ascending pharyngeal artery arise from the ICA. (D) Hypoglossal artery with ICA occlusion. (E) Hypoglossal artery CT. (F) Hypoglossal artery angiogram. (G) Fibromuscular dysplasia (FMD). Abbreviations: ICA, internal carotid artery.

Figure 9 (A) Lateral CCA injection reveals a large facial artery (FAC) with dominant nasal supply compared with the smaller IMA contribution. Note the prominent nasal blush in this patient with epistaxis (arrows). (B) Lateral CCA injection reveals a large IMA and smaller FAC in another patient. Rapid visualization of collateral circulation (C) from IMA to facial territory following FAC embolization and (D) from FAC to IMA territory demonstrated in two different patients during embolization therapy for epistaxis. Abbreviations: CCA, common carotid artery; IMA, internal maxillary artery.

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

Figure 10 Terminal ECA branching. Lateral view demonstrates the terminal branches of the ECA, the IMA, and the STA. Note the normal origins of the MMA and AMA from the IMA. Abbreviations: ECA, external carotid artery; IMA, internal maxillary artery; STA, superficial temporal artery; MMA, middle meningeal artery; AMA, accessory meningeal artery; IMA, internal maxillary artery.

conceptually divided into three segments: (1) the lower cervical segment, (2) the middle segment (at the mandibular angle), and (3) the upper segment (in the area of the parotid gland). An alternative organizational method is to consider the ECA branches as anterior and posterior branches. The anterior branches, listed in proximal to distal order, are the superior thyroid, lingual, and facial arteries. The posterior branches in proximal to distal order are the ascending pharyngeal artery (APA), occipital, and posterior auricular arteries. The branch order corresponds to the associated soft tissue structures, after which the vessels are named. The ECA terminates by bifurcating into the internal maxillary and superficial temporal arteries (Fig. 10A, B) (23). Superior Thyroid Artery

The superior thyroid artery is usually the most proximal and anterior ECA branch and can be readily identified by the prominent thyroid blush after contrast injection. This artery may also arise from the carotid bifurcation or, occasionally, directly from the cervical common carotid artery (Fig. 11) (23,27). The superior thyroid artery arises from the anterior surface of the ECA and courses directly inferiorly alongside the gland to supply the superior pole of the thyroid gland and larynx. There is extensive collateralization with the contralateral superior

7

Figure 11 Superior thyroid artery (SUT). The normal SUT (arrow) is the first branch of the ECA and provides a dense arterial blush (*) to the richly vascular thyroid gland. Note the presence of multiple branches and the incidental anterior communicating artery aneurysm. Abbreviation: ECA, external carotid artery.

thyroid artery and the inferior thyroid artery, which originates from the thyrocervical trunk. Rarely, injury to the artery may occur at the time of tracheostomy or laryngeal surgery, resulting in bleeding and/or pseudoaneurysm formation (Fig. 12A, B). Ascending Pharyngeal Artery

The APA is the first posterior ECA branch (23). Anteriorly directed APA branches supply the pharynx and eustachian tube. Posteriorly directed branches supply the tympanic cavity and prevertebral muscles (Fig. 13A, B). The main trunk of the APA parallels the course of the ICA and can be occasionally mistaken for the ICA on ultrasound in the setting of internal carotid occlusion (a source of false-negative ultrasound screening examinations) (Fig. 14A, B) (28). The APA, in its location adjacent to the pharyngeal mucosal space, can be eroded by tumor and become the source of intractable bleeding (Fig. 15A, B). A small but clinically important branch vessel is the neuromeningeal branch, which supplies both the dura and lower cranial nerves. There are extensive anastomoses between the APA and the intracranial vasculature. These include rami anastomosing with the middle meningeal and accessory meningeal arteries of the external carotid circulation (23,29). There are also anastomoses with the internal carotid system via the inferior tympanic artery, which

8

Johnson et al.

Figure 12 Superior thyroid artery (SUT) pseudoaneurysm: peritracheal bleeding nine days after radical neck surgery, layngectomy, and tracheostomy. Oblique RCCA injection demonstrates faint blush from the distal SUT (A) better seen on microcatheter injection (B). It was successfully embolized with acrylic (C). Abbreviation: RCCA, right common carotid artery. From Endovascular today and neurosurgical clinics.

Figure 13 Ascending pharyngeal artery. AP (A) and lateral (B) angiograms of the normal APA that divides into an anterior (pharyngeal) and posterior division. Collaterals exist between the posterior division and the vertebral and between the anterior division and the internal carotid artery. (C) Note the extensive neovascular supply (arrows) from the anterior division of the APA to this JNA. Abbreviations: APA, ascending pharyngeal artery; JNA, juvenile nasal angiofibroma.

anastomoses with the caroticotympanic artery of the petrous internal carotid. Other variable anastomoses may also exist between the APA and the vidian artery and inferolateral trunk. The APA may also anastomose with cervical branches of the vertebral artery via

the artery of the odontoid arch (Fig. 16). These potential pathways of collateralization are extremely important to keep in mind during therapeutic embolization of the APA territory (23,29). Careful angiographic evaluation of the APA branching pattern

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

9

Figure 14 Ultrasound pitfalls. (A–C) High-grade carotid stenosis misdiagnosed as ICA occlusion by ultrasound. An 80-year-old female had known bilateral carotid stenoses, which had previously been estimated at 90% on the right by ultrasound 5 months ago. Pulsed Doppler of right ICA (A) shows increased peak systolic velocity and prominent diastolic flow. The more distal ICA could not be visualized, and the study reported RICA occlusion. By comparison, pulsed Doppler of the left ICA (B) shows a symmetric appearance of the diastolic flow on the left as compared with the right. A true occlusion of the RICA would demonstrate no diastolic flow in the ICA proximal to the occlusion. (C) A CTA 3D volume-rendered image with curved reformation confirms that the right ICA is severely narrowed, but patent. (D, E) APA mistaken for the ICA by ultrasound. Pulsed Doppler ultrasound (D) image of a patient with congenital absence of the left ICA (same patient as shown in Fig. 32C–E) demonstrates an artery in the expected location of the ICA, which has low resistance flow. This vessel was mistaken for the ICA on initial ultrasound interpretation. A CTA 3D volume-rendered images shows the absent ICA (E). The APA lies parallel to the carotid sheath, and the presence of APA to ICA anastomoses lead to ‘‘internalization’’ of the waveform pattern, thus potentially causing confusion on ultrasound. Abbreviations: ICA, internal carotid artery; APA: ascending pharyngeal artery; RICA, right internal carotid artery.

Figure 15 APA extravasation (EXTRAV). (A) Lateral CCA injection demonstrates faint extravascular contrast in the region of the APA. (B) Microcatheter injection reveals frank extravasation. (C) Acrylic injection (CAST) resulted in cessation of bleeding. Abbreviations: APA, ascending pharyngeal artery; CCA, common carotid artery. From Endovascular today.

10

Johnson et al.

Figure 16 Artery of the odontoid arch. (A) Selective APA injection before embolization of skull base giant cell tumor (arrows) demonstrates a midline vessel filling the vertebral artery from the APA. (B) Note the vessel filling on the selective vertebral artery injection. Safe embolization requires the catheter to be positioned distal to the collateral branch. Abbreviation: APA, ascending pharyngeal artery.

Figure 17 Lingual artery normal. (A) AP and (B, C) lateral early- and late-phase images.

is imperative before therapeutic embolization is performed. Lingual Artery

The lingual artery arises from the anterior surface of the ECA, loops upward, and proceeds anteriorly along the hyoid and deep into the hypoglossal muscle, to supply the ipsilateral tongue, sublingual gland, pharynx, and hyoid musculature (1,23). It may occasionally arise from a common trunk with the facial artery (Fig. 17). The lingual artery has a characteristic U-shape on AP and lateral views. Lingual artery injury, erosion, or laceration may result in pseudoaneurysm formation and massive bleeding (Fig. 18A, B). Facial Artery

The facial artery is the third anteriorly oriented ECA branch. It ascends along the superior constrictor

muscle, passes deep into the stylohyoid and digastric muscles, and loops over the submandibular gland. It crosses the anterior aspect of the mandible and branches into the submental artery inferiorly, to supply the floor of mouth and submandibular gland. The facial artery and its superior branches course in an oblique fashion from the inferolateral aspect of the face, supplying the lips, face, palate, pharynx, and floor of the nasal cavity before terminating as the angular artery near the medial canthus of the eye (Fig. 19) (1,23). Occipital Artery

The occipital artery is the second posteriorly oriented ECA branch, arising opposite the facial artery. It passes beneath the posterior belly of the digastric and sternocleidomastoid muscles providing muscular penetrating branches. It courses within the subcutaneous tissues of the posterior scalp and supplies the posterior

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

11

Figure 18 (A) Massive oral bleeding. Axial CTA demonstrates radiation seeds in the tongue/floor of mouth on the left. A rounded collection of contrast is identified in the tongue consistent with a lingual pseudoaneurysm (arrow). (B) CTA coronal reconstruction demonstrates the pseudoaneurysm (arrow) and correlates with the (C) AP external carotid arteriogram, where the pseudoaneurysm is identified as arising from the LIN. (D) Lateral ECA arteriogram demonstrates the markedly irregular lingual artery and the contrast extending into the pseudoaneurysm (arrow) arising from the irregular segment. Abbreviations: ECA, external carotid artery; LIN, lingual artery; FAC, facial artery.

Figure 19 Facial artery nasal supply. (A) AP and (B) lateral views of a FAC injection demonstrates marked vascular blush to the nasal arcade and a focal PSA in this patient with epistaxis. Abbreviations: FAC, facial artery; PSA, pseudoaneurysm.

skin, muscle, and meninges of the posterior fossa (1,24,30). Prominent muscular branches provide anastomoses between the occipital and vertebral arteries, particularly in the setting of proximal stenosis or occlusion (Fig. 20A, B). It is important to recognize that meningeal branches pass intracranially through the hypoglossal and mastoid canals as well as through the jugular foramen. These branches can become enlarged in the setting of dural arteriovenous malformation (Fig. 21).

Posterior Auricular Artery

The posterior auricular artery arises from the posterior aspect of the ECA just above the level of the occipital artery (23). It may occasionally arise from or as a combined trunk with the occipital artery (30). The stylomastoid branch of the posterior auricular artery enters the stylomastoid foramen and sends branches to the chorda tympani within the tympanic cavity, the mastoid, and the semicircular canals. The

12

Johnson et al.

Figure 20 Occipital artery. (A) Lateral selective occipital artery injection demonstrates scalp branches and distal meningeal branches supplying a hypervascular meningioma (arrows). (B) Lateral common carotid injection demonstrates prompt filling of the vertebral artery from muscular collaterals of the occipital artery. The anterior circulation fills via the posterior communicating artery in this patient with occlusive disease of the internal carotid artery. (C–E) OCC to vertebral muscular collaterals are demonstrated in this patient with left subclavian origin occlusion. (C) Illustrates reconstitution of the intracranial vertebral artery, while later phase lateral (D) and AP (E) views demonstrate reconstitution of the cervical vertebral artery and distal subclavian. Abbreviation: OCC, occipital.

mastoid artery anastomoses with petrosal branches from the middle meningeal artery. Superficial Temporal Artery

The ECA terminates within the parotid gland in the superficial temporal artery (STA) and the internal maxillary artery. From its origin within the parotid gland, the STA proceeds cephalad over the arch of the zygoma and divides into frontal and parietal branches. The STA is primarily a cutaneous artery supplying the anterior two-thirds of the scalp, the underlying cranium and musculature, and portions of the parotid gland, ear, and temporomandibular joint (1,23). Small local branches anastomose with the maxillary and facial artery branches of the upper portion of the face. The STA has a characteristic ‘‘hairpin’’ turn on angiography as it courses over the zygoma (Fig. 23). The superficial course of the STA renders it vulnerable to direct injury with resultant pseudoaneurysm formation. The pseudoaneurysms commonly present as pulsatile ‘‘lumps’’ on the forehead or scalp following remote trauma (Fig. 24). Figure 21 Enlarged dural branches with dural arteriovenous malformation. Lateral ECA arteriogram demonstrates an enlarged middle meningeal artery with shunting into the transverse sinus and middle meningeal vein. In addition, there are enlarged dural branches of the occipital artery shunting into the abnormal, distally occluded transverse sinus. Abbreviation: ECA, external carotid artery.

auricular branch supplies the scalp, the pinna, and the external auditory canal. A prominent but normal vascular blush is noted in the pinna after injection of the posterior auricular artery (Fig. 22). The stylo-

Internal Maxillary Artery

The internal maxillary artery courses deep to the neck of the mandible and enters the infratemporal fossa. It commonly passes horizontally between the heads of the medial and lateral pterygoid muscles and through the pterygomaxillary fissure into the pterygopalatine fossa (1,4,5,23). Three segmental divisions of the internal maxillary artery are defined by the position of the artery relative to the pterygoid muscle. The first segment gives rise to the inferior alveolar artery, which extends inferiorly along with the mandibular nerve to the mandibular foramen (Fig. 25). The middle and accessory meningeal arteries pass through the

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

13

Figure 22 PAA supplying AVM. (A) MRA 3D TOF axial source image demonstrates enlargement of the left pinna and increased signal intensity consistent with hypervascularity. Digital AP (B) and early- and late-phase lateral (C, D) views of a selective OCC artery injection demonstrates the PAA arising from the OCC (a normal variant) and a prominent blush with early venous drainage into the external jugular system secondary to a high-flow AVM of the pinna. Abbreviations: AVM, arteriovenous malformation; OCC, occipital; PAA, posterior auricular artery; TOF, time of flight.

Figure 23 Scalp AVM. (A, B) AP and lateral selective STA angiograms demonstrate the enlarged feeders from the anterior division of the STA to an AVM of the scalp. Note the normal size of the STA posterior division and the early draining vein. Abbreviations: STA, superficial temporal artery. Abbreviation: AVM, arteriovenous malformation.

14

Johnson et al.

Figure 24 STA pseudoaneurysm. (A, B) Two patients with typical STA aneurysms (arrows) following direct trauma. Abbreviations: STA, superficial temporal artery; MMA, middle meningeal artery.

Figure 25 Internal maxillary artery. (A, B) AP and lateral selective IMA injections demonstrate the three segments and the important branches. Abbreviations: IMA, internal maxillary artery; MMA, middle meningeal artery; ACM, accessory meningeal artery; STA, superficial temporal artery.

foramen spinosum and ovale, respectively. The middle meningeal artery has a characteristic curve as it exits the foramen spinosum that parallels the floor of the sella on lateral angiogram. The meningeal branches can be differentiated from the scalp branches by their straight rather than tortuous course. Remembering that ‘‘you can wrinkle your forehead, but you cannot wrinkle your dura’’ is a helpful key to differentiating these branches (Fig. 26). The middle meningeal artery may be variable in size and may occasionally give rise to, or arise from, the ophthalmic artery (31). The deep auricular artery that supplies the external auditory canal and the anterior tympanic artery that supplies the tympanic membrane both arise from the first segment of the internal maxillary artery. The pterygoid segment (middle) is located in the high, deep masticator space and gives rise to masseteric, buccal, and deep temporal arteries. These supply the pterygoid and temporalis muscles and the lingual and buccal nerves. The third or sphenopalatine segment of the internal maxillary artery lies within the pterygopalatine fossa and sends branches along with each nerve to the pterygopalatine ganglion

(Fig. 27). It terminates in multiple branches to the nasal cavity supplying both nasal wall and septum. The posterior superior alveolar artery supplies the palate and posterior wall of the maxilla. The infraorbital artery passes through the infraorbital fissure along the orbital floor. External Carotid Anastomotic Network

The importance of external carotid to internal carotid collaterals and potential anastomotic pathways cannot be overemphasized in the setting of disease and neurointervention (23,29,30,32). These interconnections are dynamic and may change in appearance and flow rate during the interventional procedure, becoming most dangerous near the end of the procedure. The IMA has numerous extensive anastomoses with other ECA branches in the face. It is clinically relevant to appreciate the extensive collateral network between the lingual, facial, and internal maxillary artery branches. A complex hemodynamic balance exists between these pedicles. If a hypoplastic facial artery is present, large buccal and masseteric branches

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

15

Figure 26 Middle meningeal artery variations. The ophthalmic artery arises from the MMA in this patient. The reverse can also occur, posing potential problems for embolization. Abbreviations: MMA, middle meningeal artery.

Figure 27 IMA nasal arcade (A) AP DSA injection into the distal ECA demonstrates the branches of the internal maxillary artery and the nasal arcade (arrows). Note the STA and MMA arteries. (B) Magnified superselective AP view better demonstrates the nasal arcade and prominent mucosal blush in this patient with epistaxis. Abbreviations: IMA, internal maxillary artery; DSA, digital subtraction angiography; ECA, external carotid artery; STA, superficial temporal artery; MMA, middle meningeal artery.

will be present from the internal maxillary artery, and vice versa. During embolization therapy for epistaxis, it is not uncommon to appreciate anastomotic branches restoring proximal flow to an embolized territory (Fig. 28). External carotid to internal carotid anastomoses exist, and flow may proceed in either direction depending on the location and nature of the diseased vasculature. The distal ethmoidal branches of the IMA anastomose with distal ethmoidal branches of the ophthalmic artery. Thus the IMA, via these ethmoidal collaterals, may provide a supply route to the supraclinoid ICA via reversal of flow through the ophthalmic artery. The vidian artery anastomoses with the petrous ICA. The artery of the foramen rotundum and the inferolateral trunk anastomose with the cavernous ICA. These ECA-ICA anastomoses vary to a significant degree among patients and offer a clinically significant

collateral pathway between the ECA and the ICA systems, seen most prominently in the setting of occlusive vascular disease (Fig. 29A, B) (23,29). With occlusion of the ECA, ICA branches may collaterally restore external carotid flow (Fig. 30) (23,29).

Internal Carotid Artery The ICA enters the skull base through the carotid canal ascending anterior to the jugular bulb and posterior to the eustachian tube (1,4,5,33). The ICA petrous segment courses anteromedially to the tympanic cavity, giving rise to the caroticotympanic artery (to the tympanic cavity), the vidian artery, and small periosteal branches (34). The ICA courses superiorly, extending above the foramen lacerum to pierce the dura and enter the posterior aspect of the cavernous sinus (Fig. 31). The ICA is occasionally congenitally

16

Johnson et al.

Figure 28 Epistaxis: the importance of ophthalmic collaterals. (A) ICA injection at the time of initial epistaxis embolization demonstrates normal terminal ophthalmic artery branches. (B–D) One month later, the patient presents with recurrent epistaxis, and sequential ICA images demonstrate reconstitution of the nasal arcade by ophthalmic collaterals. Abbreviation: ICA, internal carotid artery.

Figure 29 Extensive collaterals to the petrous, cavernous, and supraclinoid ICA from the branches of the internal maxillary artery. (A) IMA to OPH to ICA ethmoidal collaterals. (B) Vidian artery and inferolateral trunk to the petrous ICA. Note the occipital to vertebral artery muscular collaterals. Abbreviations: ICA, internal carotid artery; AMA, acessory meningeal; MMA, middle meningeal artery; OPH, ophthalmic artery.

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

17

Figure 30 Restoration of ECA flow. (A) Lateral CCA arteriogram demonstrates extravasation of contrast from the proximal ECA at the origin from the carotid bulb. Note the radiation seeds and the small occipital artery identified before embolization. (B) Following partial embolization of the ECA and occlusion of the right CCA. Control arteriogram demonstrates filling of the ipsilateral vertebral artery with filling of a large muscular collateral with reconstitution of the occipital artery and retrograde filling of the ECA with continued extravasation into the pharynx. Control of bleeding required particulate embolization for occlusion and disconnection of the muscular collateral to the occipital artery. Abbreviations: ECA, external carotid artery; CCA, common carotid artery.

Figure 31 The dural ring. The ICA enters the cavernous sinus dura, traverses the sinus and exits at the dural ring. This patient presented with SAH and demonstrates and ICA posterior wall aneurysm. The arrow marks the location of the dural ring on the conventional angiogram (A) and on 3D CTA (B). The trigeminal artery is a primative communication between the cavernous carotid segment and the distal one-third of the basilar artery identified on conventional angiogram (C) and on 3D CTA (D). Abbreviation: ICA, internal carotid artery; SAH, subarachnoid hemmorhage.

18

Johnson et al.

Figure 32 ICA anomalies. (A) CT and (B) AP angiogram demonstrate the aberrant ICA with the characteristic lateral position of the ICA (arrow) projecting into the middle ear cavity behind the tympanic membrane. The carotid canal is incomplete and the carotid is usually narrowed just distal to the middle ear segment. (C) CT and (D, E) CTA in agenesis of the ICA demonstrate absence (*) of the carotid canal in addition to the absence of the ICA. Abbreviation: ICA, internal carotid artery.

absent and can be differentiated from acquired occlusion by the absence of the carotid canal at the skull base (Fig. 32A, B) (35). Nomenclature varies, but the four-part division of the internal carotid, designated as C1–C4 and described in the radiology and surgical literature, is useful. The cervical segment (C1) begins proximally at the origin of the ICA with the CCA and extends cephalad to the external orifice of the carotid canal. The petrous segment (C2) traverses the carotid canal and enters the cavernous sinus (dura), where the cavernous segment (C3) begins. The cavernous segment ends where the ICA pierces the dural roof of the cavernous sinus. The supraclinoid segment (C4) begins where the ICA exits the dural ring and enters the subarachnoid space, and it ends at the internal carotid bifurcation into anterior and middle cerebral artery branches (34,36). The supraclinoid segment passes medially to the anterior clinoid and below the optic nerve. Together, the C3 and C4 segments form the characteristic ‘‘S’’ shape seen on lateral and oblique angiographic views of the skull base. C1 does not normally provide any branches. C2 gives rise to three potential branches: the caroticotympanic branch supplying the middle and inner ear; the vidian artery, or the artery of the pterygoid canal,

which goes through the foramen lacerum; and the artery of the foramen rotundum (35,36). C3 gives rise to three trunks. The posterior trunk, or the meningohypophyseal trunk, branches into the tentorial artery (of Bernasconi and Casinari) supplying the tentorium, the inferior hypophyseal artery supplying the posterior pituitary capsule, and the dorsal meningeal artery supplying the abducens nerve and the clivus (35,36). The lateral trunk, or inferior cavernous sinus artery, supplies the inferolateral cavernous sinus wall and region of the foramen ovale and spinosum. The medial trunk, or McConnel’s capsular artery, supplies the anterior and inferior pituitary capsules and is present in only 28% of the population (33,34,36). A pituitary blush is commonly identified on lateral internal carotid arteriograms. These small branches become important in the analysis of skull base tumors and provide potential anastomoses with external carotid branches in the setting of disease (Figs. 32 and 33) (33,34,36).

VEINS OF THE HEAD, NECK, AND SKULL BASE The venous drainage of the face is predominantly superficial and empties into the external jugular drainage pathways (1,4,5). The supraorbital and supratrochlear

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base

19

Figure 33 Cervical and facial veins. (A) The proximal internal and external jugular veins are demonstrated as approached from the femoral route. (B, C) Facial veins drain into the EJV. (D, E) Nasal and facial structures may drain superiorly into the superior and or inferior ophthalmic veins. Abbreviations: EJV, external jugular vein; SOV, superior ophthalmic veins.

veins of the face join to become the angular vein and proceeds as the facial vein over the angle of the mandible (1,4,5). The pterygopalatine venous plexus is located around and within the lateral pterygoid muscle. It may be recognized on CT as a focal area of irregular enhancement adjacent to the muscle. It is often identified as a variation in the cerebral venous drainage pattern on cerebral angiography, receiving flow from the greater middle cerebral (sylvian) vein (Fig. 34A, C). The pterygopalatine venous plexus drains into a pair of maxillary veins, which lie deep in the neck of the mandible and join with the temporalis vein draining the temporal region of the face and scalp to form the retromandibular vein. The inferior ophthalmic vein travels with the infraorbital artery and drains into the cavernous sinus intracranially and the pterygopalatine venous plexus extracranially.

Occasionally, the facial veins will drain superiorly into the ophthalmic veins and into the cavernous sinus as a normal variation in the absence of shunting (37). The retromandibular vein passes through the parotid gland and divides into anterior and posterior branches that drain into the internal and external jugular veins, respectively. The deep facial vein represents the anastomosis between the pterygopalatine venous plexus and the facial vein. The anterior jugular veins lie in the submental region extending inferiorly to the suprasternal notch, where they communicate with the external jugular vein deep to the sternocleidomastoid muscle. The external jugular also receives the posterior auricular vein. The external jugular vein empties into the subclavian vein near the midpoint of the clavicle. The internal jugular vein originates from the jugular bulb receiving blood from

20

Johnson et al.

Figure 34 Prominent PVP (A) Axial CTA image demonstrates asymmetry of the PVP, prominent on the left (arrows)—possible normal variant versus AVM. Digital subtraction images of the venous phase after ICA injection in AP (B) and lateral (C) projections demonstrate a prominent sylvian (greater middle cerebral) vein (arrows) draining into an unusually large PVP (arrow), which subsequently drains into the external jugular vein. This arrangement is a normal anatomic variant. Abbreviation: PVP, pterygopalatine venous plexus; AVM, arteriovenous malformation; ICA, internal carotid artery.

Figure 35 Veins at the skull base. (A) Late venous-phase DSA image in AP projection after arterial injection shows the normal course of the skull base venous sinuses. (B, C) DSA images in AP and lateral projections demonstrate the course of the IPS, which is oriented medially and anteriorly. (D) DSA image in AP projection from another patient after right IJV injection demonstrates venous communication with contrast filling the right IPS and refluxing into the left IPS. Note the cavernous sinus filling. Abbreviations: DSA, digital subtraction angiography; IPS, inferior petrosal sinus; SIG, sigmoid sinus; IJV, internal jugular vein.

the sigmoid sinus and its first extracranial tributary, the inferior petrosal sinus (Fig. 35A–C) (37,38). It descends behind the ICA directly adjacent to the arch of C1, where it joins the subclavian vein to become the brachiocephalic vein. The left brachiocephalic vein joins at the right of the second costal cartilage to become the superior vena cava.

SUMMARY Knowledge of the normal and variant anatomy of the head, neck, and skull base is critical to the under-

standing of its vascular pathology and to the safe performance of diagnostic and therapeutic angiographies. Correlation with cross-sectional imaging is useful in the anticipation of vascular supply and dangerous anastomoses.

REFERENCES 1. Johnson MH. Head and neck vascular anatomy. Neuroimaging Clin N Am 1998; 8:119–141. 2. Langman J. Medical Embryology. 3rd ed. Baltimore, MD: William & Wilkins, 1975.

Chapter 1: Vascular Anatomy of the Head, Neck, and Skull Base 3. Haughton VM, Rosenbaum AE. The normal and anomalous aortic arch and brachiocephalic vessels. In: Newton TH, Potts GN, eds. Radiology of the Skull and Brain: Angiography, vol 2. Great Neck, NY: Mosby, 1974:1145– 1163. 4. Warwick R, Williams PL, eds. Gray’s Anatomy. 36th ed. Philadelphia: WB Saunders, 1981. 5. McMinn RMH. Last Anatomy, Regional and Applied. 8th ed. Edinburgh, UK: Churchill Livingstone, 1990. 6. Faggioli GL, Ferri M, Freyrie A, et al. Aortic arch anomalies are associated with increased risk of neurologic events in carotid stent procedures. Eur J Vasc Endovasc Surg 2007; [Epub ahead of print]. 7. De Garis C, Black I, Riemenschneider E. Patterns of the aortic arch in American white and Negro stocks, with comparative notes on certain other mammals. J Anat 1933; 67:599–618. 8. Layton KF, Kallmes DF, Cloft HJ, et al. Bovine aortic arch variant in humans: clarification of a common misnomer. AJNR Am J Neuroradiol 2006; 27(7):1541–1542. 9. Lemke AJ, Benndorf G, Liebig T, et al. Anomalous origin of the right vertebral artery: review of the literature and case report of right vertebral artery origin distal to the left subclavian artery. AJNR Am J Neuroradiol 1999; 20 (7):1318–1321. 10. Caus T, Gaubert JY, Monties JR, et al. Right-sided aortic arch: surgical treatment of an aneurysm arising from a Kommerell’s diverticulum and extending to the descending thoracic aorta with an aberrant left subclavian artery. Cardiovasc Surg 1994; 2(1):110–113. 11. Cina CS, Arena GO, Bruin G, et al. Kommerell’s diverticulum and aneurysmal right-sided aortic arch: a case report and review of the literature. J Vasc Surg 2000; 32 (6):1208–1214. 12. Newton TH, Mani RL. The vertebral artery. In: Newton TH, Potts GN, eds. Radiology of the Skull and Brain: Angiography, vol 2. Great Neck, NY: Mosby, 1974:1659– 1709. 13. Paksoy Y, Vatansev H, Seker M, et al. Congenital morphological abnormalities of the distal vertebral arteries (CMADVA) and their relationship with vertigo and dizziness. Med Sci Monit 2004; 10(7):CR316–CR323. 14. Kashimada A, Machida K, Honda N, et al. Measurement of cerebral blood flow with two-dimensional cine phase-contrast MR imaging: evaluation of normal subjects and patients with vertigo. Radiat Med 1995; 13 (2):95–102. 15. Mitchell J. Differences between left and right suboccipital and intracranial vertebral artery dimensions: an influence on blood flow to the hindbrain?. Physiother Res Int 2004; 9 (2):85–95. 16. Seidel E, Eicke BM, Tettenborn B, et al. Reference values for vertebral artery flow volume by duplex sonography in young and elderly adults. Stroke 1999; 30(12):2692–2696. 17. Smith AS, Bellon JR. Parallel and spiral flow patterns of vertebral artery contributions to the basilar artery. AJNR Am J Neuroradiol 1995; 16(8):1587–1591. 18. Johnson MH, Christman CW. Posterior circulation infarction: anatomy, pathophysiology, and clinical correlation. Semin Ultrasound CT MR 1995; 16:237–252. 19. Lister JR, Rhoton AL Jr., Matsushima T, et al. Microsurgical anatomy of the posterior inferior cerebellar artery. Neurosurgery 1982; 10(2):170–99. 20. Matsumoto M, Okuda H, Ishidoh E, et al. An anomalous case of the common carotid artery giving off several

21.

22.

23.

24.

25. 26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

21

branches and high division of the internal carotid artery. Okajimas Folia Anat Jpn 1986; 63(1):37–43. Lo A, Oehley M, Bartlett A, et al. Anatomical variations of the common carotid artery bifurcation. ANZ J Surg 2006; 76(11):970–972. Thomas JB, Antiga L, Che SL, et al. Variation in the carotid bifurcation geometry of young versus older adults: implications for geometric risk of atherosclerosis. Stroke 2005; 36:2450–2456. Djindian R, Merland JJ. Normal superselective arteriography of the external carotid artery. In: Djindian R, Merland JJ, eds. Superselective Arteriography of the External Carotid Artery. New York: Springer-Verlag, 1978:1–46. Morimoto T, Nitta K, Kazekawa K, et al. The anomaly of a non-bifurcating cervical carotid artery. Case report. J Neurosurg 1990; 72(1):130–132. Ooigawa H, Nawashiro H, Fukui S, et al. Non-bifurcating cervical carotid artery. J Clin Neurosci 2006; 13(9):944–947. Cakirer S, Karaarslan E. Aortic arch origin of the left external carotid artery. AJNR Am J Neuroradiol 2003; 24 (7):1492; author reply 1492. Toni R, Della Casa C, Castorina S, et al. A meta-analysis of superior thyroid artery variations in different human groups and their clinical implications. Ann Anat 2004; 186(3):255–262. Wei CJ, Chang FC, Chiou SY, et al. Aberrant ascending pharyngeal artery mimicking a partially occluded internal carotid artery. J Neuroimaging 2004; 14(1):67–70. Mishkin MM, Schreiber MN. Collateral circulation in Newton TH and Potts DG, eds. Angiography. Radiology of the Skull and Brain, vol 2, book 4. St Louis, MO: Mosby Yearbook, 1977:2344–2374. Alvernia JE, Frazier K, Lanzino G. The occipital artery: an anatomical study. Surgical anatomy and technique. Neurosurgery 2006; 58(1 suppl 1):ONS-114–ONS-122. Kawai K, Yoshinaga K, Koizumi M, et al. A middle meningeal artery which arises from the internal carotid artery in which the first branchial artery participates. Ann Anat 2006; 188(1):33–38. Johnson MH, Chiang VL, Ross DA. Interventional neuroradiology adjuncts and alternatives in patients with head and neck vascular lesions. Neurosurg Clin N Am 2005; 16:547–560. Lasjuanias P, Berenstein A. The internal carotid artery (ICA). In: Lasjuanias P, Berenstein A, eds. Surgical Neuroimaging: Functional Vascular Anatomy of Brain, Spinal Cord and Spine. 3rd ed. Berlin: Springer-Verlag, 1990. Tubbs RS, Hansasuta A, Loukas M, et al. Branches of the petrous and cavernous segments of the internal carotid artery. Clin Anat 2007; 20 (online). Worthington C, Olivier A, Melensen D. Internal carotid artery agenesis: correlation by conventional and digital subtraction angiography and by computed tomography. Surg Neurol 1984; 22(3):295–300. Harris FS, Rhoton AL. Anatomy of the cavernous sinus. A microsurgical study. J Neurosurg 1976; 45(2):169–180. Hacker H. Normal supratentorial veins and dural sinuses. In: Newton TH, Potts GN, eds. Radiology of the Skull and Brain: Angiography, vol 3. Great Neck, NY: Mosby, 1974:1851–1877. Huang YP, Wolf BS. Veins of the posterior fossa. In: Newton TH, Potts GN, eds. Radiology of the Skull and Brain: Angiography, vol 2. Great Neck, NY: Mosby, 1974:2155–2219.

2 Applied Neurovascular Anatomy of the Brain and Skull Randy S. Bell, Alexander H. Vo, and Rocco A. Armonda Departments of Neurosurgery and Radiology, National Naval Medical Center, and Comprehensive Neurosciences Program, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A.

INTRODUCTION Advances in imaging and materials technology have expanded the array of pathologies treatable through less invasive endovascular approaches. The resultant benefit to the patient manifested by increased interventional success rates and reduced morbidity and mortality cannot be overstated. However, the utility of even the most advanced biplanar machine with 3D rotational capabilities is limited without a thorough understanding of the craniocerebral angiographic anatomy. This understanding must, of necessity, include the significant arterial anastomoses and collateral circulatory patterns that should be considered during any intervention. Collateral circulation may prevent significant neurologic deficit should parent artery occlusion (PAO) be required. That said, known circulatory anastomoses could result in infarct distal to the area of embolization or PAO. The purpose of this chapter is to provide an in-depth review of the normal cerebrovascular angiographic anatomy as well as the significant internal, external, and vertebrobasilar anastomoses. Additionally, high-quality gross anatomic specimens will be shown with the basic angiogram to emphasize the importance of surrounding neurologic structures. The importance of the contribution of individual anatomy in the formulation of any treatment plan will also be emphasized. Because thorough reviews of anatomic variants have been provided elsewhere (1–5), only brief descriptions will be highlighted where considered relevant.

INTERNAL CAROTID ARTERY The internal carotid artery (ICA) originates from the common carotid artery in the neck at the approximate level of the fourth cervical vertebrae. Though several segmental naming schemes exist, this chapter will refer to that provided by Rhoton (5). The cervical segment (C1) ascends to the base of the skull without producing any branches (Fig. 1B). It enters the skull through the carotid canal to become the horizontal petrous portion (C2). This segment is seen as the first medial turn on a standard anterior-posterior (AP) projection and as the

first anterior turn on a lateral projection (Fig. 1A). The vidian and caroticotympanic branches originate from this segment. The artery then takes a 908 superior turn at the foramen lacerum and becomes the cavernous portion (C3). The segment is manifested as a double arterial density on a standard AP projection and as an anteriorly projecting hairpin turn on a lateral projection. Branches within this segment include the meningohypophyseal trunk, the inferolateral trunk, and McConnell’s capsular arteries. The ophthalmic artery may occasionally originate from this segment. The meningohypophyseal trunk gives rise to the tentorial artery of Bernasconi and Cassinari [important during embolization of tentorial meningiomas or tentorial arteriovenous malformation (AVM) (Figs. 2 and 3)], the dorsal meningeal artery, and the inferior hypophyseal artery. The artery then progresses caudally and laterally as it exits the cavernous sinus and enters the subarachnoid space through an inner and outer dural ring to become the supraclinoid segment (C4). The ophthalmic, superior hypophyseal, posterior communicating, and anterior choroidal arteries arise from this segment (Fig. 4). The internal carotid then bifurcates into the anterior and middle cerebral arteries (MCAs).

The Ophthalmic Artery The ophthalmic artery arises from the anterior wall of the ICA as its first intradural branch (Fig. 4). In 8% of cases, the artery may arise from within the cavernous sinus (5). It then travels in an anterior direction and enters the orbit through the optic canal along with the optic nerve. A recurrent meningeal branch may intermittently arise from the orbital portion of the ophthalmic artery, traveling back through the superior orbital fissure to supply the meninges in that area. It continues forward and gives rise to the anterior and posterior ethmoidal arteries. The remaining terminal branches of the ophthalmic artery are the central retinal, lacrimal, long and short ciliary, supraorbital, medial palpebral, infratrochlear, supratrochlear, and dorsal nasal arteries (5). Significant collateral circulation exists between the ophthalmic artery and the internal maxillary (long sphenopalatine communication via the ethmoidal

24

Bell et al.

Figure 1 AP (A) and lateral (B) angiogram of the ICA. The segmental scheme of Rhoton is provided. C1 extends from the carotid bifurcation to its entrance into the carotid canal. C2 extends from this point, through the petrous bone, to the foramen lacerum. C3 constitutes the intracavernous segment. The supraclinoid segment (C4) continues from the cavernous sinus to the ICA bifurcation. Abbreviation: ICA, internal carotid artery.

Figure 2 T1-weighted axial MRI through the rostral midbrain. A tentorial AVM is shown. The black arrow (Figure 3) indicates the location of the tentorial artery of Bernasconi and Cassinari within the paramesencephalic cistern.

Figure 3 Lateral angiographic projection of the right ICA. The black arrow again indicates the location of the tentorial artery. Note the venous outflow to the superior sagittal sinus. Abbreviation: ICA, internal carotid artery.

arteries), the middle meningeal (via the ethmoidal arteries), and the superficial temporal artery (via the lacrimal and zygomatic-orbital arteries) (1,2).

circulation to the PCA territory (1,4,5). Anatomically, it courses below the edge of the tentorium just superior to the third cranial nerve (Fig. 5). There are multiple small perforating arteries that arise from the posterior communicating artery. The largest of these arteries is called the premamillary artery (1,5). The perforating arteries are divided into anterior and posterior perforating arteries. The anterior perforators supply neurologic tissue within the posterior limb of the internal capsule, the anterior thalamus, the posterior hypothalamus, and the anterior one-third of the optic tract, while the posterior perforators penetrate the rostral midbrain and supply the subthalamic nucleus (1).

The Posterior Communicating Artery The posterior communicating artery arises from the posteromedial aspect of the ICA (Figs. 4, 5, and 13). It terminates at the posterior cerebral artery (PCA) and is the boundary between the P1 and P2 segments of that artery. In approximately 22% of cases, the posterior communicating artery is larger than the PCA or fails to fuse with the PCA and becomes the dominant

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

25

Figure 4 Lateral angiogram of the C4 segment of the ICA. Note the anterior course of the ophthalmic artery as it passes through the optic canal with the optic nerve. The black arrow identifies the course of the artery as it passes superior to the optic nerve. Significant terminal branches are noted. The posterior communicating and anterior choroidal arteries are also easily seen in this view. Abbreviations: ICA, internal carotid artery; AchA, anterior choroidal artery; pCom, posterior communicating artery; LA, lacrimal artery; CrA, central retinal artery; PeA, posterior ethmoidal artery; AeA, anterior ethmoidal artery.

Figure 6 Lateral angiogram, arterial phase, of the ICA. The branches and segments of the ACA are delineated. The A1 segment extends from the carotid bifurcation to the anterior communicating artery. The A2–A5 segments are then named based on their location with respect to the corpus callosum. The anterior and posterior internal parietal arteries are not well visualized on this injection. Abbreviations: ACA, anterior cerebral artery; ICA, internal carotid artery; A1, precommisural segment; A2, infracallosal segment; A3, precallosal segment; A4, supracallosal segment; A5, posterocallosal segment; FpA, frontopolar artery; AntIFA, anterior internal frontal artery; CmA, callosomarginal artery; MidIFA, middle internal frontal artery; PostIFA, posterior internal frontal artery; PeriA, pericallosal artery; PcA, precentral artery.

Figure 5 Gross anatomic specimen from Rhoton. The branches of the ICA are viewed in an axial plane from below. Note the location of the CN III with respect to the PCoA. and the AChA. The cisternal segment of the AChA can be seen traveling toward the temporal horn of the lateral ventricle and the choroid plexus within. The M1 segment of the MCA is also shown with numerous lenticulostriate arteries traveling superiorly through the anterior perforated substance to supply portions of the basal ganglia. Abbreviations: ICA, internal carotid artery; AChA, anterior choroidal artery; PCoA, posterior communicating artery; CN III, third cranial nerve; LentStrA, lenticulostriate arteries; MCA, middle cerebral artery. Source: From Ref. 5.

The Anterior Choroidal Artery The anterior choroidal artery is the last named branch arising from the ICA prior to its bifurcation (Figs. 4 and 5). It travels in a posterolateral direction toward the choroidal point and the choroid plexus of the temporal horn of the lateral ventricle. The artery is

broken into the cisternal (within the subarachnoid cisterns) and plexal (within the lateral ventricles and choroids plexus) segments. Small perforating vessels arise from the cisternal segment and are not visualized on a lateral angiogram. These arteries supply the optic tract, the cerebral peduncle, the mesial temporal lobe, and the lateral geniculate body.

The Anterior Cerebral Artery The anterior cerebral artery (ACA) can be broken down into five anatomic segments on the basis of its location with respect to the underlying corpus callosum (Figs. 6 and 7). The A1 segment, or the precommunicating segment, extends from the ICA bifurcation to the anterior communicating artery. Usually, small perforating branches feed the optic chiasm, hypothalamus, and anterior corpus callosum, though they are not typically visible on a normal fourvessel cerebral angiogram. The recurrent artery of Heubner may occasionally arise from this segment, though it primarily originates from A2 (Fig. 13). This artery is not commonly visualized on a basic

26

Bell et al.

Figure 7 Gross anatomic specimen, sagittal plane through the falx cerebri. The segments of the ACA are shown with surrounding neurologic structures (compare with Fig. 6). The ACA feeds the paracentral lobule, which is responsible for motor control of the contralateral leg. Abbreviation: ACA, anterior cerebral artery; CallMargA, callosomarginal artery. Source: From Ref. 5.

Figure 8 AP (A) and lateral (B) injection of the ICA. The MCA segmental anatomy is shown. M1 starts at the ICA bifurcation and ends at the LI. M2 continues to the sharp (>908 on AP) turn at the CS. M3 travels over the frontal, temporal, and parietal Op to terminate as the distal, cortical M4 branches. Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery; LI, limen insula; CS, circular sulcus; Op, operculum.

diagnostic angiogram. The remaining postcommunicating segments include the infracallosal (A2), the precallosal (A3), the supracallosal (A4), and the posterocallosal (A5). The A2 segment typically starts at the anterior communicating artery and extends to the bifurcation of the pericallosal and callosomarginal arteries. The frontopolar and orbitofrontal arteries arise from this segment. Subserved neurologic tissue includes the hypothalamus, septum pellucidum, anterior commisure, columns of the fornix, and portions of the basal ganglia. The A3 segment includes the callosomarginal and pericallosal arteries. The A4 and A5 segments involve the terminal branches of the ACA, including the arteries that provide collateral flow to certain areas within the MCA and PCA distributions (Fig. 8).

The Middle Cerebral Artery The MCA originates from the internal carotid and travels in a course parallel to the floor of the middle cranial fossa. The artery partitions into four anatomic

segments, and like the other cerebral arteries organized in this fashion, the segments are based on surrounding cerebral anatomy rather than arterial branch points (Fig. 8A, B). The M1 segment extends from the ICA to the 908 turn that the artery takes at the limen insula (Figs. 8 and 9). The MCA bifurcation may occur prior to or after this point. The M1 segment is characterized by multiple small perforating arteries that feed the lentiform nuclei and the anterior limb of the internal capsule. These lenticulostriate arteries are divided into medial, intermediate, and lateral groups and originate from the superior wall of the M1 segment and travel through the anterior perforated substance to the deep hemispheric nuclei (Figs. 12 and 13). The M2 segment extends from the limen insula to the second turn of the artery at the circular sulcus. Although the M2 branches are distinguishable on a lateral angiogram, they appear as a group of double densities on an AP view. The gross anatomic lateral view of the insula clearly displays the complex arterial anatomy in this region (Fig. 11). The M3 segment specifically refers to the course of the vessels over the frontal, parietal, and

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

27

Figure 9 Gross anatomic specimen, coronal view of the brain. The MCAs are shown with surrounding anatomy (compare with Fig. 8A). The single white area indicates the location of the limen insula. The double white arrow highlights the sharp turn of the MCA at the circular sulcus. Three white arrows approximate the location of the operculum. Abbreviations: PCA., posterior cerebral artery; LentNucl, lentiform nuclei; LentStrA, lenticulostriate arteries; MCA, middle cerebral artery; ACA, anterior cerebral artery. Source: From Ref. 5.

temporal opercula. The M4 segment refers to the terminal cortical branches (Fig. 10). It is important to remember that some of the distal MCA vessels feed non-eloquent cortex (i.e., the temporopolar artery); however, vessels feeding the central area bilaterally (primary motor cortex) and the angular area on the left could result in significant neurologic deficit should sacrifice occur. Specifically, sacrifice of the superior trunk of the MCA on the left could result in a Gerstmann’s syndrome (right–left dissociation, acalculia, agraphia without alexia, finger agnosia), while sacrifice of the same artery on the right might result in asomatagnosia.

Anatomic Considerations Aneurysms and other vascular malformations of the distal intracranial circulation present difficult treatment scenarios. As a general rule, aneurysms distal to the circle of Willis tend to rupture regardless of size (6,7) and therefore require treatment at the time of diagnosis. The possible exception may include the distal aneurysms that result from the aberrant highflow state associated with AVMs (8,9). Although open surgery was advocated prior to the advent of endovascular treatment strategies, a combined approach is now often necessary. The following will examine the unique anatomic circumstances that must be considered prior to treatment of distal intracranial circulation pathology. The Anterior Cerebral Artery

The ACA and its branches supply the cortex within the interhemispheric fissure. Since this network includes the cingulate cortex and the paracentral lobule, clinical consequences from pathology in this area may manifest as lower extremity paresis or memory impairment (Fig. 7) (8,10,11). Surgical approaches to vascular

Figure 10 Late arterial phase injection, lateral view, of the ICA. The distal MCA vessels are shown with approximate named locations. The dense tangle of vessels approximates the location of the insula. Abbreviations: MCA, middle cerebral artery; Ofr, orbitofrontal artery; PreFr, prefrontal artery; PreCen, precentral artery; Cen, central artery; Ang, angular artery; OccTemp, temporooccipital artery; PostTemp, posterior temporal artery; MidTemp, middle temporal artery; AntTemp, anterior temporal artery; Tp, temporopolar artery.

28

Bell et al.

Figure 11 Gross anatomic specimen of the arterial anatomy of the insula (compare with Fig. 8B). The opercular cortices have been retracted to show the insula and vessels. Abbreviations: SupTr, superior trunk; InfTr, inferior trunk. Source: From Ref. 5.

Figure 12 AP angiogram of the ICA, magnified view of the ICA bifurcation. The ML, IL, and LL arteries can be seen. Note the origin from the superior aspect of the MCA with a course through the anterior perforated substance to supply the lentiform nuclei. Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery; ML, middle lenticulostriate; IL, intermediate lenticulostriate; LL, lateral lenticulostriate.

pathology in this area are technically difficult and require extensive preoperative planning (6,8). Considerations include the possibility of disruption of venous drainage to the superior sagittal sinus from surgical exposure and retraction, the neurologic consequences to frontal lobe retraction, and variable anatomy. Endovascular treatments of distal ACA pathology are technically challenging because of vessel tortuosity and reduced distal vessel caliber. However, the possible neuropsychiatric consequences of open surgery can be avoided. Treatment with platinum coils of narrow-

Figure 13 Gross anatomic specimen, anterosuperior view of the ICA bifurcation (compare with Fig. 12). The small lenticulostriate arteries are indicated by white arrows. The recurrent artery of Heubner can also be seen arising from the A2 segment of the ACA. Abbreviations: ICA, internal carotid artery; ACA, anterior cerebral artery; PCoA, posterior communicating artery; RecA, recurrent artery of heubner; MCABr, middle cerebral artery branches; FrontBr, frontal branch. Source: From Ref. 5.

necked, distal ACA saccular aneurysms alone is a reasonable approach because anterograde flow through the preserved parent artery may be possible. However, it may be necessary at times to sacrifice the parent artery when balloon remodeling or stent-assisted coiling of a wide-necked or fusiform aneurysm is not feasible (Fig. 14A, B). In this case, sufficient collateral circulation from the PCA circulation may be seen (Fig. 23). Parent artery occlusion (PAO) proximal to the pericallosalcallosomarginal artery bifurcation (A2–A3) may result in some or all of the neurologic consequences previously outlined. Occlusion distal to this vessel segment may be clinically silent secondary to the extensive collateral

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

29

Figure 14 Magnified lateral (A) and plane lateral injection of the ICA. An irregular dilation of the A3 segment of the ACA (A, black arrow) is noted in a patient with traversing penetrating head trauma and subarachnoid hemorrhage. The treatment strategy included PAO using detachable platinum coils (B, black arrow). Abbreviation: ACA, anterior cerebral artery; PAO, parent artery occlusion.

circulation between the anterior and posterior cerebral circulation through the posterior pericallosal and splenial arteries. Kim et al. describe an elegant combined open surgical and endovascular approach to an A2 aneurysm, where the distal artery arose from the aneurysm dome. In this case, a side-to-side pericallosal-pericallosal anastomosis was performed prior to unilateral endovascular PAO of the A2 segment. This anastomosis resulted in preservation of distal flow to both hemispheres and obliteration of the complex aneurysm (12). The Distal Middle Cerebral Artery

The MCA supplies hemispheric structures, including the lentiform nuclei, the lateral aspect of the frontal, parietal, and temporal cortices, and the insular cortex. Clinical sequelae from arterial occlusion are largely based on which segment of which MCA (right or left) is occluded, with symptoms ranging from contralateral hemiparesis or hemianesthesia to aphasia and calculation difficulties. MCA bifurcation aneurysms are difficult to treat endovascularly because of the tendency toward widenecked morphologies or because of bifurcation arteries distal to the aneurysm arising from the aneurysm dome. Fusiform aneurysms or giant MCA bifurcation aneurysms may require PAO in combination with open surgical clipping and bypass (13,14). Weill et al. describe two cases of giant MCA trifurcation aneurysms that were successfully treated with EC-IC bypass and subsequent PAO. In these cases, the M1 segment was coil occluded in a patient with an intact circle of Willis, while the supraclinoid internal carotid was coiled in the patient with an absent anterior communicating artery (14).

Mycotic aneurysms that form in the distal MCA circulation may rupture and cause subarachnoid hemorrhage, or they may resolve on antibiotics (15). Recurrent hemorrhage or interval angiographic enlargement may push the surgeon to intervene. Several studies have demonstrated successful obliteration of the distal aneurysm through endovascular occlusion (16–20). Sodium amytal injection has been used to determine whether parent artery occlusion (PAO) will be tolerated. However, this procedure’s low sensitivity may not accurately reflect the lack of postocclusion neurologic deficit with a negative result. In cases of distal aneurysm formation in eloquent arterial territories, it may be necessary to accept the possibility of neurologic deficit to reduce the risk of potentially fatal subarachnoid hemorrhage (Fig. 15 A, B). For M4 occlusions, collateral flow from the ACA may reduce the neurologic deficit that would occur from a central arterial occlusion (motor strip).

PERSISTENT CAROTICOBASILAR ANASTOMOSES Persistent fetal circulatory patterns refer to anastomoses between the carotid and basilar arterial systems, including the persistent trigeminal, otic, hypoglossal, and proatlantal arteries (Fig. 16). These structures are present embryologically, normally recede through vessel atresia, and are associated with other vascular malformations (21–26). The trigeminal artery is the most common, occurring in 0.1% to 0.2% of the general population. This artery typically arises from the precavernous carotid and anastomoses with the distal basilar artery (Fig. 17). The persistent otic artery arises from the petrous portion of the carotid artery and terminates at

30

Bell et al.

Figure 15 Lateral angiogram (A) and magnified lateral angiogram of the ICA. The black arrows indicate the focal dilation of the distal MCA consistent with the presence of a mycotic aneurysm. Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery.

Figure 16 Lateral angiographic cartoon representation of the types of persistent caroticobasilar anastomoses. A fetal PCA (a) is shown, though this type is common enough to be considered a normal variant. The persistent trigeminal artery (b) originates from the cavernous segment of the ICA and terminates at the basilar artery. The persistent otic artery (c) originates from the petrous portion of the ICA and terminates at the proximal basilar artery. Both the persistent hypoglossal (d) and persistent proatlantal (e) arteries arise from the extracranial ICA and terminate at the vertebral artery. C1 and C2 indicate the location of the first and second cervical vertebrae. Abbreviation: PCA, posterior cerebral artery; ICA, internal carotid artery.

the anterior inferior cerebellar artery (AICA) or basilar artery (1). There are less than five reported cases of this type in the literature. The persistent hypoglossal artery arises from the extracranial ICA, passes through the anterior condyloid foramen, and terminates at the distal basilar artery. The persistent proatlantal artery is a

Figure 17 Diagnostic angiogram of the ICA, lateral projection. A persistent trigeminal artery is seen leaving the internal carotid in the cavernous (C3) segment and entering the basilar artery Note the filling of both the PCA and SCA circulation. Abbreviations: ICA, internal carotid artery; PCA, posterior cerebral artery; SCA, superior cerebellar artery.

primitive anastomosis between the ICA or ECA and the cervical vertebral artery.

Anatomic Considerations The presence of persistent caroticobasilar anastomoses should be ruled in or out before certain procedures. Wada testing is employed to localize language and memory dominance prior to partial or complete amygdalohippocampectomy. During this test, sodium amobarbital is injected into the ICA on one side to effectively anesthetize the ipsilateral hemisphere. Should a persistent caroticobasilar anastomosis exist, the target of the amobarbital would be the brain stem,

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

31

Figure 18 Lateral common carotid artery injection (A) and selective, late arterial phase injection of the ECA (B). The branches of the ECA are identified. Note the tortuous course of the distal vessels (B) characteristic of external carotid vessels. The black arrow indicates the sharp turn the middle meningeal artery makes just after passing through the foramen spinosum. Abbreviations: ST, superior thyroid; L, lingual artery; F, facial artery; ECA, external carotid artery; ICA, internal carotid artery; AscP, ascending pharyngeal artery; IMax, internal maxillary artery; MMA, middle meningeal artery; Occ, occipital artery; STA, superficial temporal artery; d, distal.

potentially resulting in respiratory arrest, stroke, or death (3). In this case, it would be necessary to select the artery distal to the anastomosis to avoid this complication. The same principle applies to balloon test occlusion studies (27). Among the list of complications for carotid endarterectomy is the possibility of brain stem infarct from a fractured embolus that occurs during this procedure should a persistent hypoglossal artery be present (28,29). Though this complication has not yet been reported in the setting of carotid artery stenting, it should be considered if a persistent hypoglossal artery is present. In this case, the choice of stent length and position would be important. A Y-stent might be the optimal solution preserving flow routes to both eloquent vascular distributions.

EXTERNAL CAROTID ARTERY The external carotid artery (ECA) originates from the common carotid artery in the neck. The named branches in order of origin are the superior thyroid, lingual, facial, ascending pharyngeal, occipital, posterior auricular, superficial temporal, and internal maxillary arteries (Fig. 18A). The ascending pharyngeal artery (APA) further bifurcates into pharyngeal and neuromeningeal trunks. The internal maxillary artery terminates in the middle meningeal, accessory meningeal, and sphenopalatine arteries (Fig. 18B). Further terminal branch description will be given later in this text where relevant anastomoses apply.

Anatomic Considerations: The Ascending Pharyngeal Artery The APA is unique because it provides anastomotic channels to the internal carotid, the vertebral artery, and other branches within the external carotid circulation (Fig. 19A) (1–4,30). It typically arises from the ECA, but it can occasionally arise from the proximal ICA or an aberrant posterior inferior cerebellar artery (PICA) (30–34). The APA starts as a common trunk and then bifurcates into pharyngeal and neuromeningeal trunks. The pharyngeal trunk terminates as the superior, middle, and inferior pharyngeal branch, providing rich anastomotic connections to the internal maxillary artery (middle pharyngeal via the descending palatine artery and pterygovaginal artery via the accessory meningeal artery) and the ICA (superior pharyngeal via the inferolateral trunk and the recurrent artery of the foramen lacerum). As the name implies, this portion of the artery supplies the tissue of the oropharynx. Although the artery is often difficult to see on an external carotid angiogram, its clinical importance exceeds its size in certain circumstances. Specifically, the anastomotic channels previously described can cloud the results of test balloon occlusion of the ICA by providing collateral circulatory routes. These channels also become important during the embolization of glomus jugulare, vagale, or tympanicum tumors supplied predominantly by this artery. The neuromeningeal trunk courses in a posterosuperior direction toward the foramen magnum. Its

32

Bell et al.

Figure 19 (A) Cartoon angiographic representation of the ascending pharyngeal artery and its anastomoses. (a) Middle pharyngeal artery to internal maxillary artery via the descending palatine artery. (b) Pterygovaginal artery (terminal branch of the superior pharyngeal artery) to internal maxillary artery via the accessory meningeal artery. (c) Superior pharyngeal artery to the ICA via the recurrent artery of the foramen rotundum and the inferolateral trunk. (d) Clival branches (terminal branches of the neuromeningeal trunk) to the ICA via the meningohypophyseal trunk. (e) inferior tympanic artery to ICA via the caroticotympanic branch. (f) Hypoglossal artery to the vertebral artery via the odontoid arch system. (g) Neuromeningeal trunk to the vertebral artery via the odontoid arch system. (h) Neuromeningeal trunk to the odontoid arch system. The odontoid arch then connects, at times, to the occipital artery. (B) Selective vertebral artery injection. A direct anastomosis from the vertebral artery to the ECA via the APA is shown. Abbreviations: APA, ascending pharyngeal artery; NmT, neuromeningeal trunk; PhB, pharyngeal branch; CCA, common carotid artery; ECA, external carotid artery; Vert, vertebral artery; IMax, internal maxillary artery; dICA, distal internal carotid artery.

branches include the inferior tympanic, musculospinal, hypoglossal, and jugular arteries, with additional terminal branches to the internal auditory canal, the clivus, and the odontoid arch. Clinically relevant anastomoses occur between the hypoglossal and musculospinal arteries to the vertebral artery, the inferior tympanic branch to the ICA through the caroticotympanic artery, lateral clival branches directly to the ICA, and ECA to ECA connections from the odontoid arch system to the occipital artery (1,4,30). An example of a direct anastomosis between the ascending pharyngeal and vertebral artery is given in Figure 19B.

THE VERTEBROBASILAR SYSTEM The vertebral arteries typically arise from the subclavians bilaterally (V1). They proceed superiorly and dorsally to enter the foramen transversarium at the level of C6. The arteries subsequently travel to the arch of C1, giving off a variable number of small spinal muscular and segmental arteries (V2). Two characteristic right-angle turns are noted on both AP and lateral angiographic projections at C1 and C2 (V2) (Figs. 20A and 22), which have been described as a box on the AP projection. The artery then processes dorsally to the atlanto-occipital joint and travels in an

anterosuperior direction to enter the dura (V4). Prior to the vertebrobasilar junction, the artery gives off the anterior spinal artery (ASA) and posterior inferior cerebellar artery (PICA). The ASA supplies the anterior spinal cord, and the PICA supplies the lower brain stem, cerebellar tonsils, and the inferior aspect of the cerebellar hemispheres. The basilar artery then travels anterior to the brain stem, giving off the anterior inferior cerebellar artery (AICA), multiple small pontomesencephalic perforators, and the superior cerebellar artery (SCA) (Figs. 20 and 21). There are multiple areas of collateral circulation between the SCA, AICA, and the PICA, and distal parent artery sacrifice in this area may be clinically silent (Fig. 23, white arrows). The AICA and PICA may, at times, arise from a common trunk (Fig. 22). This particular anatomic variant may alter the treatment plan in certain circumstances. The basilar artery subsequently bifurcates within the crural cistern into the PCA.

The Posterior Cerebral Artery The PCAs, like the MCA and ACA, are segmentally organized on the basis of the relevant surrounding anatomy (Figs. 23–25). The P1 segment starts at the basilar bifurcation and extends to the insertion of the

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

33

Figure 20 Selective vertebral artery injection, transmaxillary projection (A) and gross anatomic specimen for comparison (B). Note the course of the vertebral and basilar arteries with respect to the brain stem and cranial nerves. Abbreviations: Vert, vertebral artery; PICA, posterior inferior cerebellar artery; AICA, anterior inferior cerebellar artery; B, basilar artery; SCA, superior cerebellar artery; PCA, posterior cerebral artery; CN, cranial nerve; C1, first cervical vertebrae; C2, second cervical vertebrae. Source: From Ref. 5.

Figure 21 Lateral angiogram of the vertebral artery. Abbreviations: Vert, vertebral artery; SmB, spinomuscular branch; PICA, posterior inferior cerebellar artery; B, basilar artery; AICA, anterior inferior cerebellar artery; SCA, superior cerebellar artery; PCA, posterior cerebral artery; C1, first cervical vertebrae; C2, second cervical vertebrae.

Figure 22 Transmaxillary projection of the vertebral artery. Note the bilateral absence of a true PICA, and a prominent bifurcation of the AICA characteristic of an AICA-PICA complex. Abbreviations: PICA, posterior inferior cerebellar artery; AICA, anterior inferior cerebellar artery.

34

Bell et al.

Figure 23 Lateral projection of the basilar artery from a vertebral artery injection. Collateral circulatory pathways are shown with the segmental anatomy of the PCA. Note the filling of the internal parietal arteries from this injection via the splenial artery. The splenium of the corpus callosum is outlined by the splenial artery and the lateral and medial posterior choroidal arteries. The black arrow indicates the faint contrast blush within the posterior thalamoperforating arteries from the basilar apex and P1. The white arrows depict the collateral circulation of the distal cerebellar vessels. Abbreviations: PCA, posterior cerebral artery; B, basilar artery; MpcA, medial posterior choroidal artery; LpcA, lateral posterior choroidal artery; S, splenial artery; CalcA, calcarine artery; PoA, parietooccipital artery; IpA, internal parietal artery; CC, corpus callosum.

Figure 24 Projection through the foramen magnum (Townes projection), vertebral artery injection. The segmental anatomy of the PCA is shown. P1 starts at the basilar artery and ends at the insertion of the posterior communicating artery. P2 is divided into a P2A within the crural cistern and P2P within the ambient wing cistern. P3 begins at the quadrigeminal plate cistern and ends at the origin of the calcarine and parietooccipital artery. The P4 segment includes the terminal cortical branches. The digitally subtracted shadow of a 7.62-mm bullet is seen (FB). Abbreviations: PCA, posterior cerebral artery; AtA, anterior temporal artery; P2A, anterior segment; P2P, posterior segment; PtA, posterior temporal artery; PoA, parietooccipital artery; FB, foreign body.

Figure 25 Gross anatomic specimen of the medial undersurface of cerebrum. The course and segmentation of the PCA is shown. Note the course of the calcarine artery within primary visual cortex (Cuneus and Lingula). The brain stem can also be seen medial to the P2 segment. Perforating arteries from this segment supply the lateral brain stem and optic tract. Abbreviations: PCA, posterior cerebral artery; ACA, anterior cerebral artery; ICA, internal carotid artery; CalcSulcus, calcarine sulcus; ParOccip, Sulcusparietooccipital sulcus. Source: From Ref. 5.

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

35

posterior communicating artery. The P2 segment is divided into the anterior (P2A) and posterior (P2P) on the basis of its cisternal location. The P2A starts at the posterior communicating artery and travels around the anterolateral aspect of the mesencephalon in the crural cistern. The P2P continues posteriorly within the paramesencephalic and ambient wing cisterns and ends at the quadrigeminal plate cistern. Small perforating arteries, not well visualized on an angiogram, arise from this segment and supply the cerebral peduncles, brain stem, optic tracts, thalamus, choroids plexus, and hippocampus. The posterior temporal artery also emerges in this region. The P3 segment starts at the quadrigeminal cistern (tectal plate) and continues to the origin of the parieto-occipital and calcarine arteries (P4).

Anatomic Considerations: The Distal Posterior Cerebral Artery Distal vascular anomalies in the posterior cerebral circulation are difficult entities to treat. The surgical approaches to the crural, paramesencephalic, ambient wing, and quadrigeminal plate cisterns are certainly elegant, but the associated morbidity may lend itself to a treatment strategy that incorporates a less invasive approach (7, 35–38). Patients with pathology in this region may present with hemiparesis (brain stem perforators or compression), homonymous hemianopsia (compression of optic tract, infarction of calcarine cortex), and occasionally fourth nerve compression (Fig. 26) (7,39). The elegant nature of the neural tissue fed by the PCA system complicates the open surgical or endovascular repair of vascular abnormalities in this region. Surgical bypass followed by PAO is one successful approach to complicated PCA pathology. Endovascular treatment without open surgery can reduce morbidity, but the often necessary PAO can result in additional deficit. There are diverging published opinions concerning where along the PCA circulation PAO is safe (36,40). Ciceri et al. treated 21 aneurysms in 20 patients with endovascular coil occlusion (36). The parent artery was preserved in 14 patients, and PAO was performed without preoperative test balloon occlusion. Though relatively successful with proximal PAO, the general recommendation posited by the authors was that occlusion distal to P2 could be tolerated secondary to adequate distal collateral circulation provided by the posterior temporal, lateral posterior choroidal, medial posterior choroidal, and splenial arteries. Conversely, proximal PCA occlusion could be associated with brain stem infarct. However, Hallacq et al. occluded the parent artery in 10 patients with P2 segment aneurysms without postocclusion deficit, concluding that P2 occlusion was in fact safe.

THE CEREBRAL VEINS It is convenient to think of the cerebral venous anatomy as a construct of channels (sinuses) and veins organized into a deep and superficial system. The

Figure 26 Artists rendition of the course of the distal PCA under the surface of the tentorium cerebelli. An aneurysm in this location could conceivably cause a fourth cranial nerve palsy secondary to its proximity to that nerve. Abbreviation: PCA, posterior cerebral artery. Source: From Ref. 39.

sinuses receive the bulk of venous outflow from the brain and terminate in the internal jugular veins. The following will review the deep and superficial venous anatomy seen on normal angiographic studies.

The Superficial Venous System The superior sagittal sinus is the large midline vein easily visualized on both AP and lateral angiographic projections (Fig. 27). It receives direct inflow from the hemispheres via the superficial frontal, parietal, and occipital cortical veins as well as extra-axial inflow from the diploic and meningeal veins. The superior anastomotic vein of Trolard is the largest of the cortical venous inflow tracts and is typically located in the middle third of the sinus. The superior sagittal sinus terminates at the torcular herophili at the confluence of the sinuses. Here the straight sinus and the superior sagittal sinus become the paired transverse sinuses. The straight sinus receives venous inflow from the inferior sagittal sinus, the vein of Galen, and meningeal veins from the tentorium. The transverse sinuses make a 908 turn under the asterion of the skull to become the sigmoid sinuses. On a lateral angiogram, the inferior anastomotic vein of Labbe and the superior petrosal sinuses can be seen draining into this area.

Figure 27 Lateral (A) and AP (B) projection, venous phase. Normal venous anatomy is shown. Abbreviations: IJ, internal jugular vein; SigS, sigmoid sinus; TS, transverse sinus; SPS, superior petrosal sinus; SS, straight sinus; SSS, superior sagittal sinus; VoL, inferior anastomotic vein of Labbe; BVoR, basal vein of Rosenthal; TsV, thalamostriate vein; VoG, vein of Galen; IcV, internal cerebral vein; AfV, anterior frontal vein; MfV, middle frontal vein; PfV, posterior frontal vein; PrecV, precentral vein; VoT, superior anastomotic vein of Trolard; ApV, anterior parietal vein; PpV, posterior parietal vein.

Figure 28 Representation of the main cerebral venous and sinus anatomy. Abbreviations: Ant, anterior; Post, posterior; Mid, middle; Cent, central; Med, medial; Precent, precentral; Pericall, pericallosal; Car, Carotid; Cav, cavernous; SphenPar, sphenoparietal; Sup, superficial; Sag, Sagittal; Inf, inferior; Paracent, paracentral; Cer, cerebral; Front, frontal; Bas, basal; Men, meningeal; Temp, temporal; Tent, tentorial; Occip, occipital; Post, posterior; Calc, calcarine; Lat, lateral; Str, straight. Source: From Ref. 5.

Chapter 2: Applied Neurovascular Anatomy of the Brain and Skull

Figure 29 Lateral injection of the ICA. A direct CC fistula is shown. The cavernous sinus is best visualized during pathologic conditions. Note the cerebellar cortical venous reflux, making this particular CC fistula prone to causing subarachnoid hemorrhage. Abbreviations: CC, carotid cavernous; Pp, pterygoid venous plexus; IpS, inferior petrosal sinus; SpS, superior petrosal sinus; CavS, cavernous sinus; RV, retinal vein.

The base of the brain contains multiple venous sinuses that are clinically relevant. The cavernous sinus is a paired structure composed of multiple venous sinusoidal channels that anatomically encircle the sella turcica (Fig. 30). This structure is not overtly obvious on a normal angiogram, but becomes prominent in the setting of direct carotid cavernous fistulas (Fig. 29). Just posterior and inferior to this structure is the midline basal sinus. This sinus receives inflow from the superficial sylvian vein via the sphenoparietal sinus, and outflow from the superior petrosal sinus toward the sigmoid sinus.

The Deep Venous System The deep venous structures are also easily visible on a lateral angiogram. The basal veins of Rosenthal become visible within the crural cisterns (Figs. 27A and 28). They receive flow from the deep middle cerebral veins and the anterior cerebral veins (Fig. 28). They subsequently course within the paramesencephalic and ambient wing cisterns with the PCA to finally end at the vein of Galen. The vein of Galen also receives venous drainage from the internal cerebral veins. This paired venous outflow receives predominant feeders from the thalamostriate veins (Fig. 27A), the anterior caudate veins, and the anterior septal veins lining the

37

Figure 30 AP projection of an ICA injection. A direct CC fistula is shown. The cavernous sinus clearly outlines the sella turcica in this view. Abbreviations: CC, carotid cavernous; ST, sella turcica; CavS, cavernous sinus; Pp, pterygoid plexus.

lateral ventricles. The internal cerebral veins course within the velum interpositum with the medial posterior choroidal arteries and terminate at the vein of Galen. The vein of Galen then empties into the straight sinus.

REFERENCES 1. Armonda R, Rosenwasser R. Vascular anatomy of the central nervous system. In: Awad I, Rosenwasser R, eds. Vascular Malformations of the Central Nervous System. Philadelphia: Lippincott Williams & Wilkins, 1999:19–45. 2. Lasjaunias P, Berenstein A, ter Brugge K. Surgical Neuroangiography: Clinical Vascular Anatomy and Variations. 2nd ed. New York: Springer, 2001. 3. Morris P. Practical Neuroangiography. 1st ed. Baltimore, MD: Lippincott Williams & Wilkins, 1997. 4. Osborn A. Cerebral Angiography. 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 1999. 5. Rhoton AL. Cranial Anatomy and Surgical Approaches. Schaumburg, IL: Lippincott Williams & Wilkins, 2003. 6. Ohno K, Monma S, Suzuki R, et al. Saccular aneurysms of the distal anterior cerebral artery. Neurosurgery 1990; 27(6):907–912; discussion 912–913. 7. Hamada J, Morioka M, Yano S, et al. Clinical features of aneurysms of the posterior cerebral artery: a 15-year experience with 21 cases. Neurosurgery 2005; 56(4):662–670. 8. Wisoff JH, Flamm ES. Aneurysms of the distal anterior cerebral artery and associated vascular anomalies. Neurosurgery 1987; 20(5):735–741. 9. Kim EJ, Halim AX, Dowd CF, et al. The relationship of coexisting extranidal aneurysms to intracranial hemorrhage

38

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21. 22.

23.

24.

Bell et al. in patients harboring brain arteriovenous malformations. Neurosurgery 2004; 54(6):1349–1357. Endo H, Shimizu H, Tominaga T. Paraparesis associated with ruptured anterior cerebral artery territory aneurysms. Surg Neurol 2005; 64(2):135–139. Ezura M, Takahashi A, Jokura H, et al. Endovascular treatment of aneurysms associated with cerebral arteriovenous malformations: experiences after the introduction of Guglielmi detachable coils. J Clin Neurosci. 2000; 7(suppl 1):14–18. Kim LJ, Albuquerque FC, McDougall C, et al. Combined surgical and endovascular treatment of a recurrent A3-A3 junction aneurysm unsuitable for stand-alone clip ligation or coil occlusion. Technical note. Neurosurg Focus 2005; 18(2):E6. Greene KA, Anson JA, Spetzler RF. Giant serpentine middle cerebral artery aneurysm treated by extracranialintracranial bypass. Case report. J Neurosurg 1993; 78(6): 974–978. Weill A, Cognard C, Levy D, et al. Giant aneurysms of the middle cerebral artery trifurcation treated with extracranialintracranial arterial bypass and endovascular occlusion. Report of two cases. J Neurosurg 1998; 89(3):474–478. Corr P, Wright M, Handler LC. Endocarditis-related cerebral aneurysms: radiologic changes with treatment. AJNR Am J Neuroradiol 1995; 16(4):745–748. Asai T, Usui A, Miyachi S, et al. Endovascular treatment for intracranial mycotic aneurysms prior to cardiac surgery. Eur J Cardiothorac Surg 2002; 21(5):948–950. Chapot R, Houdart E, Saint-Maurice JP, et al. Endovascular treatment of cerebral mycotic aneurysms. Radiology 2002; 222(2):389–396. Erdogan HB, Erentug V, Bozbuga N, et al. Endovascular treatment of intracerebral mycotic aneurysm before surgical treatment of infective endocarditis. Tex Heart Inst J 2004; 31(2):165–167. Khayata MH, Aymard A, Casasco A, et al. Selective endovascular techniques in the treatment of cerebral mycotic aneurysms. Report of three cases. J Neurosurg 1993; 78(4):661–665. Watanabe A, Hirano K, Ishii R. Cerebral mycotic aneurysm treated with endovascular occlusion–case report. Neurol Med Chir 1998; 38(10):657–660. Bohmfalk GL, Story JL. Aneurysms of the persistent hypoglossal artery. Neurosurgery 1977; 1(3):291–296. Garza-Mercado R, Cavazos E, Urrutia G. Persistent hypoglossal artery in combination with multifocal arteriovenous malformations of the brain: case report. Neurosurgery 1990; 26(5):871–876. Kobayashi H, Munemoto S, Hayashi M, et al. Association of persistent hypoglossal artery, multiple intracranial aneurysms, and polycystic disease. Surg Neurol 1984; 21 (3):258–260. Nishida C, Ashikaga R, Araki Y, et al. Persistent hypoglossal artery associated with arteriovenous malformation: a case report. Eur J Radiol 2000; 33(1):59–62.

25. Patel AB, Gandhi CD, Bederson JB. Angiographic documentation of a persistent otic artery. AJNR Am J Neuroradiol 2003; 24(1):124–126. 26. Brick JF, Roberts T. Cerebral arteriovenous malformation coexistent with intracranial aneurysm and persistent trigeminal artery. South Med J 1987; 80(3):398–400. 27. Allen JW, Alastra AJ, Nelson PK. Proximal intracranial internal carotid artery branches: prevalence and importance for balloon occlusion test. J Neurosurg 2005; 102 (1):45–52. 28. Cartier R, Cartier P, Hudan G, et al. Combined endarterectomy of the internal carotid artery and persistent hypoglossal artery: an unusual case of carotid revascularization. Can J Surg 1996; 39(2):159–162. 29. Megyesi JF, Findlay JM, Sherlock RA. Carotid endarterectomy in the presence of a persistent hypoglossal artery: case report. Neurosurgery 1997; 41(3):669–672. 30. Hacein-Bey L, Daniels DL, Ulmer JL, et al. The ascending pharyngeal artery: branches, anastomoses, and clinical significance. AJNR Am J Neuroradiol 2002; 23(7): 1246–1256. 31. Lasjaunias P, Guibert-Tranier F, Braun JP. The pharyngocerebellar artery or ascending pharyngeal artery origin of the posterior inferior cerebellar artery. J Neuroradiol 1981; 8(4):317–325. 32. Lasjaunias P, Moret J. The ascending pharyngeal artery: normal and pathological radioanatomy. Neuroradiology 1976; 11(2):77–82. 33. Quisling RG, Seeger JF. Ascending pharyngeal artery collateral circulation simulating internal carotid artery hypoplasia. Neuroradiology 1979; 18(5):277–280. 34. Wei CJ, Chang FC, Chiou SY, et al. Aberrant ascending pharyngeal artery mimicking a partially occluded internal carotid artery. J Neuroimaging 2004; 14(1):67–70. 35. Chang HS, Fukushima T, Takakura K, et al. Aneurysms of the posterior cerebral artery: report of ten cases. Neurosurgery 1986; 19(6):1006–1011. 36. Ciceri EF, Klucznik RP, Grossman RG, et al. Aneurysms of the posterior cerebral artery: classification and endovascular treatment. AJNR Am J Neuroradiol 2001; 22(1): 27–34. 37. Gerber CJ, Neil-Dwyer G. A review of the management of 15 cases of aneurysms of the posterior cerebral artery. Br J Neurosurg 1992; 6(6):521–527. 38. Kitazawa K, Tanaka Y, Muraoka S, et al. Specific characteristics and management strategies of cerebral artery aneurysms: report of eleven cases. J Clin Neurosci 2001; 8(1):23–26. 39. Hall JK, Jacobs DA, Movsas T, et al. Fourth nerve palsy, homonymous hemianopia, and hemisensory deficit caused by a proximal posterior cerebral artery aneurysm. J Neuroophthalmol 2002; 22(2):95–98. 40. Hallacq P, Piotin M, Moret J. Endovascular occlusion of the posterior cerebral artery for the treatment of p2 segment aneurysms: retrospective review of a 10-year series. AJNR Am J Neuroradiol 2002; 23(7):1128–1136.

3 Vascular Anatomy of the Spine and Spinal Cord Armin K. Thron Department of Neuroradiology, University Hospital, RWTH Aachen University, Aachen, Germany

INTRODUCTION Because of progress in microneurosurgery and interventional neuroradiology, intramedullary spinal vascular lesions have become more and more accessible and treatable. Unfortunately, a lack of knowledge about spinal vascular anatomy is evident in many conferences with neurologists and sometimes even with neurosurgeons and neuroradiologists. This lack of knowledge might be a reason for unsatisfactory clinical results in the treatment of spinal vascular diseases by invasive therapeutic techniques. Furthermore, magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) of blood vessels in and around the spinal cord have substantially improved. To provide a correct anatomical interpretation of the demonstrated blood vessels, knowledge of the anatomy of spinal cord blood vessels is the first prerequisite. At the end of the 19th century, Kadyi (1) gave the most precise and detailed anatomical description of these blood vessels. His work was published in 1889, seven years after the first extensive and comprehensive study performed by Adamkiewicz (2). This chapter deals with the essentials of spine and spinal cord blood vessel anatomy (in parts adapted from 3), outlines the possibilities of identifying these vessels on tomographic images, and illustrates the main problems and pitfalls in the anatomical evaluation of spinal vascular malformations.

ARTERIAL BLOOD SUPPLY Sources of Arterial Blood Supply to the Spine and Spinal Cord The blood supply to the vertebral body, the paraspinal muscles, the dura, the nerve root, and the spinal cord is derived from segmental arteries (Fig. 1). These vessels persist as intercostal and lumbar arteries in the majority of the thoracolumbar region. Several segments in the upper thoracic region have a common feeder, which is the supreme intercostal artery. Following intrauterine vascular rearrangements, longitudinal arteries are established in the cervical region. On each side, three vessels are potential

sources of spinal blood supply, namely, the vertebral artery, the deep cervical artery, and the ascending cervical artery. In the sacral and lower lumbar region, sacral arteries and the iliolumbar artery (supplying the L5 level) derived from the internal iliac arteries are the most important supply to the caudal spine. Generally, the segmental arteries supply all the tissues on one side of a given metamere, with the exception of the medulla. A spinal branch of the posterior intercostal artery enters the vertebral canal through the intervertebral foramen and regularly divides into three branches: an anterior and a posterior artery of the vertebral canal, which supply the spinal column, and a radicular artery, which supplies the dura and nerve root at every segmental level. The hemivertebral blush, resulting from injection of a segmental artery, may help in identifying the artery. At the thoracic level, the artery is named according to the number of the rib under which it courses. The segmental arteries are connected across the midline and between levels above and below, through highly effective anastomoses (Figs. 2 and 7). At certain segmental levels, this radicular artery has persisted as a radiculomedullary artery, which means that it follows the anterior and/or posterior nerve roots to form and supply the superficial spinal cord arteries (Fig. 1). The number of these radiculomedullary arteries is reduced during an embryonic transformation process. From 2 to 14 (on average 6) anterior radiculomedullary arteries persist as a result of this ontogenic reduction of feeding vessels. The posterior radiculomedullary arteries are reduced less drastically from 11 to 16 vessels. Figure 3 demonstrates schematically the typical potential sources of arterial supply to the anterior axis of the spinal cord.

Extra- and Intraspinal Extradural Anastomoses 1.

An extraspinal system connects the neighboring segmental arteries longitudinally. The vessels course on the lateral aspect of the vertebra or transverse process (Figs. 1, 2, and 7). This system is highly developed in the cervical region where the vertebral artery and the deep cervical and

40

Thron

Figure 1 Blood supply of the spinal column and spinal cord.

2.

ascending cervical arteries form the most effective longitudinal anastomoses. The intraspinal extradural system is mainly a transverse anastomosis, but it also has longitudinal interconnections. The retrocorporeal and arteries are the relevant vessels for the supply of bone and dura (Figs. 1, 2, and 7). These anastomoses provide an excellent collateral circulation and therefore numerous segmental arteries can be visualized by injection of one segmental artery (Fig. 2).

The extra- and intraspinal anastomoses protect the spinal cord against ischemia when pathologies, such as arteriosclerotic disease of the aorta, cause focal vessel occlusion.

Radicular Supply and Superficial Spinal Cord Arteries Several nomenclatures and classifications have been used to describe spinal cord arteries. This variation is an ongoing cause of misunderstanding. A recent classification proposed by Lasjaunias et al. (4) differentiates three types of spinal radicular arteries: radicular, radiculopial, and radiculomedullary. The first type of spinal radicular artery is a small branch present at every segmental level, which is restricted to the supply of the nerve root. The second type supplies the nerve root and superficial pial plexus (e.g., posterior radicular artery). The third type supplies the nerve root, pial plexus, and medulla. This classification may offer some advantages to the interventional neuroradiologist when compared

with classical differentiations because it stresses the importance of the anterior supply for the gray matter of the spinal cord parenchyma. From an anatomical and linguistic point of view, however, it is not a clear-cut differentiation because anterior and posterior radicular arteries share in the blood supply of the medulla and the posterior radicular arteries do contribute to the supply of the central gray matter, especially of the posterior horn. We therefore suggest only a slight modification of the older anatomical classification with the following differentiation of spinal radicular arteries: radicular arteries supplying only the nerve root and the dura mater, but not the spinal cord; anterior radiculomedullary arteries in which the persistent medullary branch runs with the anterior nerve root to join the longitudinal trunk, which has been called the anterior spinal artery (Figs. 1 and 3); and posterior radiculomedullary arteries in which the persistent medullary branch accompanies the posterior nerve root and joins the longitudinal systems of posterolateral and/or posterior spinal arteries. The first lies laterally and the second medially to the posterior root entry zone. These longitudinally oriented vessels are not continuous and may replace each other (Figs. 1, 6, 7C, D, and 8C). As has already been mentioned, the anterior radiculomedullary supply is reduced to an average of 6 radiculomedullary arteries (Fig. 3), whereas from 11 to 16 posterior radiculomedullary arteries persist after embryonic life.

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

Figure 2 Extra- and intraspinal extradural anastomoses. Selective injection in the first lumbar artery on the left opacifies homolateral arteries as well as contralateral vessels. The typical hexagonal configurations of the retrocorperal intraspinal anastomosis (small arrows) as well as the extraspinal pretransverse and anterolateral anastomoses (large arrows) are demonstrated. The injected artery gives rise to an anterior radiculomedullary artery (arrowheads), probably the Adamkiewicz artery. Note the hemivertebral blush corresponding to the injected artery.

The thoracolumbar enlargement is the region where the dominant anterior radiculomedullary artery (arteria radicularis magna, or Adamkiewicz artery) arises. However, in this region several posterior radiculomedullary arteries may also be large-sized vessels, which furnish blood supply to this area. They are connected to the anterior spinal artery through two anastomotic semicircles, called the arcade of the cone (Figs. 3 and 6A). The superficial distribution of blood to the spinal cord is achieved by the above-mentioned anterior and posterior longitudinal vessels, which have been named anterior and posterior/posterolateral spinal arteries. Both systems supply a superficial network of smaller pial arteries that covers the spinal cord, termed the vasocorona(Figs. 8 and 9). The anterior

41

Figure 3 Sources of supply to the anterior spinal artery.

spinal artery may be small or absent as a continuous tract in the upper thoracic and upper cervical regions of the spinal cord (Figs. 3 and 4). The main source of arterial supply to the cord is the anterior spinal artery (ventral axis), with a multisegmental distribution of blood and a distinct territory of supply. It gives rise to the hemodynamically important central (centrifugal) system, which supplies the major part of the gray matter. Additionally, there are branches to the pial system on the anterior and lateral surface, supplying the ventral two-thirds of the vasocorona (Figs. 8 and 9A). The posterior/posterolateral spinal arteries distribute blood to the dorsal one-third of the

42

Thron

Figure 5 Demonstration of an anterior radiculomedullary feeder to the cervical spinal cord by injecting the costocervical trunk on the left. The ascending and descending branches (arrowheads) are forming the anterior spinal artery at this level.

Figure 4 Anterior spinal artery in the cervical spinal cord. The pattern of supplying vessels varies considerably, especially in the upper spinal cord. (A) Photograph of an injected specimen. Plexiform pattern of arterial supply in the upper cervical levels, without formation of a midline anterior spinal artery. (B) X-ray film in AP view. The anterior spinal artery is formed by a large unilateral descending branch from the left vertebral artery (arrow) and is reinforced by a large anterior radiculomedullary artery at the C5 level on the right. The small descending branch coming from the right vertebral artery (arrowhead) ends in this network of small tortuous superficial arteries. Source: From Ref. 13.

vasocorona, and in this way they share with central artery branches in the supply of the posterior horn and marginal parts of the central gray matter (Fig. 9A). The posterior/posterolateral arteries do not have such a distinct territory of supply as the anterior spinal artery, which means that they predominantly reinforce the rope ladder–like network of posterior pial arteries (Fig. 8C).

Differences in Arterial Supply of the Spinal Cord Depending on Regions Cervical Region

One of the ventral radicular feeders between C5 and C8 is often distinctly larger (400–600 mm) than the others and was termed the artery of the cervical enlargement by Lazorthes (5,6). It is more often derived from the deep and ascending cervical arteries than from the vertebral artery (Figs. 3 and 5). Therefore, these vessels that originate from the thyrocervical and costocervical trunks, respectively, must be demonstrated on angiography for diagnostic and interventional procedures. The average number of anterior radicular feeders to the cervical medulla is 2 to 3. The ventral feeders to the upper cervical cord, originating from the intracranial section of the vertebral artery, may be very small. Their demonstration on angiography is often impossible. If there are two descending branches from both vertebral arteries, the smaller or rudimentary vessel does not join the main

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

43

the posterior system of the upper cervical region. The vessel may have a large caliber and originates in a lateral position (lateral cervical artery) (7). The number of central arteries in the cervical enlargement is about 6/cm. They take a horizontal course.

Thoracic Region

Figure 6 X-ray film (AP view) of a contrast-injected human spinal cord specimen with arterial filling. (A) Lumbar and (B) thoracic regions. Note the different calibers of the anterior (large arrowheads) and posterior radiculomedullary feeders (small arrowheads) and of the anterior spinal artery (small arrows) at different spinal cord levels. The important supply coming from the ‘‘artery of the lumbar enlargement’’ is obvious as well as its connection with the posterior arteries around the cone (arcade of the cone) (black arrow). The system of posterior/posterolateral arteries is discontinuous; the largest posterior radiculomedullary feeder enters below and contralateral to the artery of Adamkiewicz in this specimen (oblique arrowhead). Source: From Ref. 13.

midline trunk but ends separately as a large central artery (Figs. 3 and 4). Duplication of the anterior spinal artery over some distance is frequent in this region; a pseudoisland formation and even a net-like plexiform pattern of arteries may be observed. Continuity of an anterior spinal artery may not exist. All these variations have to be regarded as a state more closely related to the embryonic (or ontogenetic) condition (Fig. 4). A descending branch from the vertebral artery or posterior inferior cerebellar artery (PICA) constitutes

Occasionally, one segmental artery branches and supplies two intercostal spaces. In this case, the dorsal and spinal branch of one of the two segments may not be seen. Therefore, a small posterior intercostal artery, from which the spinal branch of this metamere arises, must be looked for. This procedure may be crucial, for example when searching for the site of a dural arteriovenous fistula. The upper and midthoracic regions are mostly supplied by small radicular arteries (200–400 mm), making angiographical demonstration difficult (Fig. 3). In addition, the ventral anastomotic tract (anterior spinal artery) may be discontinuous throughout these regions. The pial system plays an important role in this spinal cord region, which has relatively less gray matter and more white matter tracts (Figs. 9A and 10A). On the posterior surface of the cord, the longitudinal tracts may run in posterior/posterolateral positions, thus indicating the functional identity of these vessels (Figs. 6 and 8C). The number of central arteries is only 2 to 3/cm for this region. This fact explains the prevalence of steeply ascending and descending central artery branches (Fig. 10A). As pointed out earlier, the impression of an intrinsic longitudinal anastomosis on sagittal images was not confirmed on coronal images of our microangiographic studies.

Thoracolumbar Region and Cauda Equina

One of the ventral feeders between T9 and L1 (exceptionally at L2 or L3) is always dominant (80–100 mm) and is therefore called the artery of the lumbar enlargement (Lazorthes) or the great radicular artery (Adamkiewicz) (Figs. 3 and 6). Below its level of entrance, additional significant ventral feeders are unusual. Supply to the posterior system in this region often includes two equally large dorsal feeders (400– 500 mm) that enter the spine above or below the great radicular artery (Fig. 6). The ventral and dorsal systems are connected to each other around the conus (arcade of the cone, rami anastomotici arcuati) (1). This pattern may constitute a significant anastomosis, comparable to the circle of Willis. The densest concentration of central arteries is found in the thoracolumbar enlargement, where 6 to 8 vessels/cm can be counted on microangiograms (Fig. 8B).

44

Thron

Figure 7 Angiographical demonstration and identification of spinal cord arteries. (A, B) Selective injection of the 11th intercostal artery with normal findings. (A) Frontal view. Typical hairpin configuration and midline position of the anterior radiculomedullary and anterior spinal arteries [small ascending and larger descending branch at this level (arrowheads)]. Note the extra- and intraspinal longitudinal and transverse anatomoses (arrows) and the hemivertebral blush. (B) Lateral view. Anterior location of the artery demonstrated in (A) (arrowhead). The lateral projection is very helpful to differentiate the anterior or posterior position of the artery with certainty, which may be difficult in cases of scoliosis, and especially AVMs. (C, D) Injection of the ninth intercostal artery in case of an intraspinal tumor. (C) Frontal view. The segmental artery gives rise to an anterior (arrowhead) and posterior (arrow) spinal cord supplying artery. Note the different positions of the ‘‘hairpin curve’’ and of the descending branches. (Displacement of the anterior and posterior spinal arteries below the level of the injection and the equal size of both vessels are due to an intraspinal neurinoma.) (D) Lateral view. The anterior (arrowhead) and posterior positions (arrow) of the anterior and posterior spinal artery can be distinguished. Abbreviation: AVM, arteriovenous malformation.

Intrinsic Spinal Cord Arteries The arteries directly supplying the spinal cord are central (or sulcal) arteries originating from the anterior spinal artery and perforating branches arising from the pial network which covers the spinal cord. The first type of perforating arteries constitutes a centrifugal system. Each central artery (inner vessel diameter, 100–250 mm) penetrates the parenchyma to the depth of the anterior fissure, courses to one side of the cord, and branches mainly within the gray matter. The second type of perforating arteries arises from the pial covering of the cord (vasocorona) and penetrates the white matter tracts from the periphery (centripetal

system). These vessels are numerous, with a diameter of up to 50 mm. Both types of intrinsic artery and their region of supply can be appreciated on the axial section of the microangiogram demonstrated in Figures 8, 9A, and 10A.

Superficial and Intrinsic Arterio-Arterial Anastomoses Arterio-arterial anastomotic interconnections are frequent in the spinal cord. The anterior spinal artery may be regarded as the largest and most constant longitudinal anastomosis (Figs. 3 and 6). The posterior systems are not constantly developed or continuous. They include longitudinal and

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

45

Figure 8 Anterior spinal artery, intrinsic arteries, and pial plexus demonstrated on microangiograms of the lumbar spinal cord. The spinal cord has been cut into three coronal sections, each 2 to 3 mm thick; section length, 2.5 cm. (A) Anterior coronal section. Numerous transverse and oblique branches from the anterior spinal artery (arrow) supply the anterior part of the superficial pial plexus called ‘‘vasocorona.’’ (B) Middle coronal section. The central arteries, derived from the anterior spinal artery, course to one side of the cord and branch mainly within the gray matter. From the surface of the spinal cord, perforating branches of the vasocorona penetrate and supply the outer rim of fiber tracts and parts of the posterior horn. (C) Posterior coronal section. The larger posterolateral (arrows) and smaller posterior spinal arteries (arrowheads) form a rope ladder–like network, supplying the posterior part of the vasocorona. The posterolateral arteries are running laterally, the posterior arteries medially of the posterior roof entrance zone. Source: From Ref. 13.

transverse components and serve as anastomotic pathways and distribution channels, at least over some segments (Figs. 6 and 8C). The arcade of the cone has an anastomotic function, comparable to that of the circle of Willis. Superficial interconnections between two or several central arteries exist predominantly in the thoracic region. They run immediately deep and parallel to the anterior spinal artery at the entrance of the anterior fissure within the pial system (Fig. 10A). Additionally, there are horizontal anastomoses between central artery branches and the superficial systems, especially in a centroanterolateral or centroposterolateral direction. However, they do not seem to

play an important role in the plasticity of blood supply to the spinal cord. The most important observation to note is that we could not demonstrate an intrinsic longitudinal anastomoses between the ascending and descending branches of the central arteries within the spinal cord parenchyma (13), as was assumed by Adamkiewicz (2) and Fazio and Agnoli (8).

VENOUS DRAINAGE The pattern of venous drainage deviates substantially from that of the arteries. Their arrangement will be

46

Thron

Figure 9 Comparison between intrinsic spinal cord arteries and veins demonstrated on microangiograms of axial sections of 2-mm thickness. (A) Arteries at different levels of the spinal cord (anterior spinal artery, arrow; posterior/posterolateral spinal arteries at both sides of the posterior root entry zone, small arrowheads). The central arteries (large arrowhead) are the predominant intrinsic feeders at the level of the cervical and lumbar enlargements. They run within the anterior fissure and continue either to the right or left side of the hemicord as a centrifugal system. The perforating branches are the predominant feeders of the thoracic spinal cord. They originate from the superficial vasocorona as a centripetal system, and their territory of supply can very well be differentiated from the system of central arteries. (B) Veins at different levels of the spinal cord. Radial and central veins are of almost equal size and drain to the pial covering of the spinal cord (anterior and posterior median spinal veins, arrows). Source: From Ref. 13.

described in the direction of venous drainage from the spinal cord parenchyma to the epidural plexus.

Intrinsic Veins Radial veins drain the blood of the spinal cord parenchyma. They show a horizontal, radial, and symmetrical course in most parts of the spinal cord (Fig. 9B). Only in the lower thoracic cord, from the lower lumbar enlargement to the conus medullaris, are the

sulcal veins (100–250 mm) larger than the numerous radial veins.

Superficial Veins At the level of the spinal pia mater, blood is accumulated in essentially two longitudinal collectors: the anterior and posterior median spinal veins (Figs. 10–14). The anterior midline vein is located under the anterior spinal artery (Fig. 11C). It has its largest caliber lumbosacrally. In about 80% of cases, it runs together with the

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

47

Figure 10 Comparison between intrinsic spinal cord arteries and veins demonstrated on microangiograms of sagittal sections through the midline of the spinal cord (section length, 2.5 cm; thickness, 2 mm). (A) Arteries at a lower thoracic level. Anterior spinal artery (large arrow) with loss of contrast filling in small sections. The ascending course of the central arteries with more vertical than horizontal arborization within the gray matter is demonstrated. The impression of an intrinsic longitudinal anastomosis is not supported by coronal images. Compare the small perforating arteries of the posterior columns, originating from the pial network of the vasocorona. (B) Veins at a lower thoracic level. Anterior median vein (arrow) and posterior median vein (double arrow) with loss of contrast filling in sections. The sulcal veins are less numerous but larger than the posterior veins. Several of them join to form a common stem (arrowhead). The different pattern of intrinsic arterial supply and venous drainage at approximately the same spinal cord level is well demonstrated in this comparison. Source: From Ref. 13.

filum terminale as a sometimes very large terminal vein to the end of the dural sac. The venous longitudinal system on the anterior and posterior surfaces of the cord is more variable in course, size, and localization than the anterior spinal artery (Fig. 11). The longitudinal midline veins are not always continuous tracts and may be replaced by secondary systems of smaller caliber. The posterior median spinal vein takes a course independent from the posterolaterally located arteries and is especially large above the thoracolumbar enlargement. Varicose convolutions are frequent (Figs. 11B and 14B). The posterior veins of the thoracolumbar enlargement are undoubtedly the medullary vessels of largest caliber (up to 1.5 mm in diameter) and are rarely matched by superficial cervical veins. These are the vessels most likely to be seen on MR images (Figs. 13A and 14A). The vessels are part of a pial

vascular network, which has been called the venous plexus of the pia mater (9), the coronal pial plexus (10,11), or the venous pial plexus (12).

Intraparenchymal Venous Anastomoses These anastomoses are quite common. However, they are not distributed uniformly over the length of a spinal cord. They are of two types. Anastomoses of the first type are complex and connect central and peripheral branches (sulcal and radial veins 100–200 mm in diameter). They are very frequent and drain to smaller veins of the superficial pial plexus. More important are anastomoses of the second type, which are transmedullary midline anastomoses from 300 to 700 mm in diameter, connecting the median veins on both sides of the cord. They do not

48

Thron

Figure 11 Superficial spinal cord veins. (A) Photograph of the dorsal aspect of a spinal cord specimen following ink injection into the veins at cervical and thoracic levels. The posterior median vein has a variable size at different levels. Three large radiculomedullary veins (arrows) and some smaller ones can be seen. (B) X-ray film (AP view) of a contrast-injected thoracolumbar spinal cord specimen. There is much tortuosity mainly of the posterior spinal cord veins (posterior venous plexus). Three large radiculomedullary veins accompany lumbar or sacral nerve roots to reach the epidural space (arrows). (C) Photograph of an injected spinal cord specimen with filling of the ventral veins at the thoracic level. Note the hairpin configuration of the AMV, where it continues as RV. This configuration is very similar to the arterial one. The nonfilled anterior spinal artery is running beside or over the vein (arrow). Abbreviations: AMV, anterior median vein; RV, radicular vein.

receive tributaries from the intrinsic vessels. Because of their size, they are not only seen on microangiograms (Fig. 13B) but may also be seen on angiography or MRI (Fig. 13A). Through these large anastomoses, blood can easily be directed from one side of the cord to the other (13).

Differences in the Venous Drainage Depending on Spinal Cord Region Cervical Region

Radial symmetry of intrinsic veins is very pronounced in the cervical spinal cord. Transmedullary midline anastomoses are also very frequent, but they are of

smaller caliber in the upper cervical region than those in the lower cervical and upper thoracic regions. The anterior median vein was frequently larger than the posterior median vein in our material. Both veins connect to the brain stem veins and basal sinuses around the foramen magnum. Additionally and predominantly radicular outflow to the epidural plexus occurs at many levels (Fig. 11A). Thoracic Region

The greatest concentration of large transmedullary anastomic veins is found in the cervicothoracic region (1–2/cm) followed by the mid- and lower thoracic levels (Fig. 14B), where they are more widely

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

49

Figure 12 DSA of an artery of Adamkiewicz with arterial and venous phase. (A) Typical hairpin course between the anterior radiculomedullary artery and the descending branch of the anterior spinal artery. (B) Venous phase showing the radiculomedullary vein coincidentally at the same level. The configuration between midline and radiculomedullary veins is the same. This configuration is important to know for the interpretation of angiograms in case of an AVM with early venous filling or of spinal MRAs with insufficient time resolution. Abbreviations: DSA, digital subtraction angiography; AVM, arteriovenous malformation. Source: Courtesy of Prof. G. Schroth, Berne, Switzerland.

Figure 13 Venous midline anastomoses. (A) T1weighted sagittal MRI following injection of the contrast medium. Demonstration of a large intramedullary midline anastomosis between the anterior and posterior midline veins in a normal subject (arrow). Source: Courtesy of Prof. D. Petersen, Lu¨beck, Germany. (B) Transparenchymal anastomosis near the medullary cone with a caliber of 0.7 mm (arrows). Microangiogram of a midsagittal cut with venous filling (same specimen as in Fig. 10B). Note the larger caliber of the anterior vein at the level of the cone compared with the posterior vein. Source: From Ref. 13.

50

Thron

Figure 14 Superficial spinal cord veins in the lumbar spinal cord and cauda equina. (A) MRI of T1-subtraction image. The anterior median vein is running together with the filum terminale to the end of the caudal sac (arrowheads), which can be an abnormal arterialized vein or a normal variant as is shown in (B, C). (B) X-ray film of an injected specimen in lateral projection. Tortuous posterior venous plexus at the level of the lumbar enlargement (small arrows). The anterior median vein (arrowhead) continues as a terminal vein (vein of the filum terminale). Several transmedullary anastomoses can be assumed in this projection radiography (small arrowheads). (C) Microangiogram of the cone with a large terminal vein (arrowhead). Source: (B) and (C) from Ref. 13.

separated. Anterior and posterior median veins are mostly of equal size. Lumbar Region

In this region, sulcal veins may be considerably larger than radial veins (Fig. 9B). The posterior median spinal vein is particularly large above the thoracolumbar enlargement (Fig. 14B), frequently forming varicose convolutions (the so-called posterior venous plexus). The anterior median spinal vein reaches its maximum caliber in this region, and it is important to note that the vein of the filum terminale is the continuation of this anterior median vein (Fig. 14). Alternatively, the anterior vein can follow a sacral nerve root to reach the sacral epidural space (Fig. 11B). The midline veins

of the thoracolumbar enlargement are the largest blood vessels of the spinal cord (Figs. 13 and 14). When demonstrated on contrast-enhanced MRI studies, or CT myelography, they should not be mistaken for spinal cord arteries.

Radiculomedullary Veins and the Transdural Course The superficial venous blood collectors drain into the epidural venous plexus through radicular veins (Fig. 11). The transition of the midline vessel to the radicular vein forms a hairpin course, similar to the arterial configuration (Fig. 11). Therefore, on angiographic images, the vein might be mistaken for an

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord

51

Figure 15 AVM at the level of the cone, supplied by two posterolateral feeders and the anterior spinal artery. (A) MRA showing the malformation as a whole and the main drainage into a considerably enlarged terminal vein (arrow). (B) Unsubtracted angiogram. Typical midline position of the anterior spinal artery. Supply from this vessel is mostly running through the arcade of the cone (arrowheads) to the posterior surface. (C–E) DSA. Note the somewhat different hairpin configurations of anterior and posterolateral feeders in AP view (arrowheads). The largest part of the nidus is lying posteriorly (arrows). (F) Horizontal interconnections (black arrows) between the posterolateral tracts (arrowheads) are visualized with increasing peripheral resistance during embolization.

artery (Fig. 12), particularly when an arteriovenous malformation (AVM) with early venous filling is present. For the same reason, it may be impossible to distinguish an anterior spinal artery from an anterior spinal vein on magnetic resonance images as long

as no sufficient time-resolved MRA is available. From an anatomical point of view, the number of venous outlets is high. In some studies, on average 25 radicular veins were counted on the anterior and posterior surfaces of the cord (14,15). However, purely

52

Thron

Extradural Venous Spaces and the Extraspinal Venous System The extradural plexus, well demonstrated by spinal contrast venography, extends as a continuous system from the sacrum to the skull base. It drains the spinal cord and surrounding structures. Drainage of blood from the spine (including spinal cord) occurs through the internal and external venous vertebral plexus and also extends as a continuous system from the sacrum to the base of the skull. They are identifiable as anterior and posterior systems; the anterior internal vertebral plexus is larger than the posterior internal system. The external systems run anterior to the body of the vertebrae (anterior external plexus), while the posterior external plexus lies posterolateral to the vertebral bodies. This valveless system is connected with the azygos and hemiazygos venous systems by intercostal or segmental veins and in the cervical region with the vertebral and deep cervical veins. The segmental veins in the lumbar region are connected by the ascending lumbar vein, joining the azygos (right side) and hemiazygos veins (left side) (3).

Figure 16 SDAVF T5 level. The AV shunt is at the level of the dura mater (arrow) and is directed at two veins coming from upward and downward. The mass of dilated veins resembles an AVM, but no spinal cord artery is involved. Abbreviations: DSA, digital subtraction angiography; SDAVF, spinal dural arteriovenous fistula; AVM, arteriovenous malformation.

ANATOMICAL EVALUATION OF AVMs Some important problems and pitfalls in the clinical application of blood vessel anatomy concerning spinal AVMs should be mentioned and explained. They are illustrated in Figures 15–19. 1.

radicular veins might have been included in this number. If smaller veins (<250 mm in diameter) are excluded, the number of radiculomedullary veins draining the spinal cord is from 6 to 11 for the anterior and from 5 to 10 for the posterior systems (1,16). These latter studies are in agreement with a study performed by Jacobs (17). In addition, Moes and Maillot (18) described fibrotic radicular veins at thoracic levels. These veins may contribute to the vulnerability of the spinal venous system, such as in the chronic impairment of venous drainage in Foix and Alajouanine disease (19) as a late complication of dural arteriovenous fistula (AVF) (20). The transdural course of radicular veins exhibits special features (3). The presence of venous valves described by Oswald (21) could not be confirmed in later studies. Instead, an oblique and zigzag course with considerable narrowing of the lumen was first described by Tadie et al. (22). They concluded that this configuration might act as an anti-backflow system, protecting the spinal cord against high pressure in the extraspinal veins. A study performed by Otto (23) is in agreement with their findings, although this arrangement did not always prevent reflux from the epidural plexus to the superficial spinal cord veins when a contrast medium was injected in the postmortem specimen (23).

2.

3.

4.

Prior to a therapeutic intervention, it is essential to identify the feeders of a spinal AVM from both the anterior and posterior circulation. Figure 15 illustrates the different configuration of the hairpin curve in anterior and posterior radiculomedullary arteries. Nevertheless, it is essential to have a lateral projection for a definite identification. Spinal dural arteriovenous fistulas (SDAVFs) may look very similar to AVMs (Fig. 16). But as long as no typical radiculomedullary artery is involved, the arteriovenous (AV) shunt is much more likely situated at the level of the dura mater. Discrimination between a perimedullary fistula (fistulous type of an AVM) and a SDAVF may be difficult if the arterialized vein looks like an anterior radiculomedullary feeder with a hairpin curve (Fig. 17). This event is not rare in SDAVF at lower lumbar levels when the arterialized vein is one of the large lumbar veins shown in the postmortem specimen of Figure 11B. To avoid misinterpretation, careful analysis of the vessel anatomy in the region of the intervertebral foramen is important as well as a look at the further course of the vessel on the spinal cord surface on later images (Fig. 17C). If you are unable to identify an AV shunt (fistula) on or within the spinal cord, ask yourself whether the intradural blood vessels as a whole could not be veins (Fig. 17). Spinal cord supplying arteries and a SDAVF may be observed at the same level and same side as

Figure 17 DSA of a SDAVF at the L3 level (arrowhead). (A) The blood vessel that is opacified first resembles a radiculomedullary artery. It runs upward to the cone and lumbar enlargement and exhibits a narrow curve. (B) Anterior position of the blood vessel in the lateral view. (C) Filling of typical veins on the later image in the AP view confirms that the whole of intradural vessels are veins. The configuration of a radiculomedullary vein can be very similar to that of an artery (compare Figs. 11 and 12). Abbreviations: DSA, digital subtraction angiography; SDAVF, spinal dural arteriovenous fistula.

Figure 18 DSA of a SDAVF with an anterior radiculomedullary artery entering at the same foramen. (A) AP view. The artery (arrowheads) is partially superimposed by the enlarged veins (arrow). (B) The lateral view demonstrates the anterior position of the artery (arrowheads) and the posterior position of the mass of veins (arrow). Abbreviations: DSA, digital subtraction angiography; SDAVF, spinal dural arteriovenous fistula.

54

Thron

Figure 19 AVM of the filum terminale (DSA, AP views). (A) The anterior spinal artery (black arrows) is not unusually enlarged, but it continues without a change in caliber below the level of a cone in normal position. (B) At the L4 level, a second blood vessel is opacified that is a little bit larger than the artery and runs in upward direction (artery, black arrow; arterialized terminal vein, arrowheads). The clinical significance of this small AV shunt (white arrow) was considerable. Abbreviations: AVM, arteriovenous malformation; DSA, digital subtraction angiography.

5.

demonstrated in Figure 18. They are better discriminated on lateral views. If an anterior or posterior spinal artery seems to be too large for the region to be supplied or if it extends below the level of the cone, it should be followed caudally. This is the only way not to miss the small AVMs of the filum terminale (Fig. 19).

ACKNOWLEDGEMENT Special thanks to Walter Korr, RWTH Aachen University, for technical assistance in the computer graphic design of Figs. 1 and 3.

REFERENCES ¨ ber die Blutgefa¨ße des menschlichen Ru¨cken1. Kadyi H. U markes. Lemberg: Grubnowicz u Schmidt, 1889. 2. Adamkiewicz A. Die Blutgefa¨ße des menschlichen Ru¨ckenmarkes. II.Teil: Die Gefa¨ße der Ru¨ckenmarkoberfla¨che. Sitzungsberichte der Akademie der Wissenschaften in Wien, Mathematisch-Naturwissenschaftliche Klasse 1882; 85:101–130 (abstr 3). 3. Thron A. Vascular anatomy of the spine. In: Byrne James, ed. Interventional Neuroradiology. Oxford: Oxford University Press, 2002. 4. Lasjaunias P, Berenstein A, ter Brugge K. Surgical Neuroangiography. Volume1: Clinical Vascular Anatomy and Variations. 2nd ed. New York: Springer, 2002.

5. Lazorthes G, Poulhes J, Bastide G, et al. La vascularisation arte´rielle de la moelle. Recherches anatomiques et applications a` la pathologie medullaire et a` la pathologie aortique. Neuro-Chirurgie 1958; 4:3–19. 6. Lazorthes G, Gonaze A, Djindjian R. Vascularisation et circulation de la moelle epinie`re. Paris: Masson, 1973. 7. Lasjaunias P, Vallee B, Person H, et al. The lateral spinal artery of the upper cervical spinal cord. J Neurosurg 1985; 63:235–241. 8. Fazio C, Agnoli A. The vascularization of the spinal cord. Anatomical and pathophysiological aspects. Vasc Surg 1970; 4:245–257. 9. Tveten L. Spinal cord vascularity. The venous drainage of the spinal cord in the rat. Acta Radiol Diagn 1976; 17:653–662. 10. Gillilan LA. Veins of the spinal cord. Neurology 1970; 20:860–868. 11. Turnbull JM, Breig A, Hassler O. Blood supply of cervical spinal cord in man; a microangiographic cadaver study. J Neurosurg 1966; 24:951–965. 12. Crock HV, Yoshizawa H. The blood supply of the vertebral column and spinal cord in man. Wien, New York: Springer, 1977. 13. Thron AK. Vascular anatomy of the spinal cord. Wien, New York: Springer, 1988. 14. Jellinger K. Zur Orthologie und Pathologie der Ru¨ckenmarkdurchblutung. Wien, New York: Springer, 1966. 15. von Quast H. Die Venen der Ru¨ckenmarkoberfla¨che. Gegenbaurs Morphologisches Jahrbuch 1961; 102:33–64. 16. Suh TH, Alexander L. Vascular system of the human spinal cord. Arch Neurol Psychiat 1939; 41:659–677. 17. Jacobs T. Venae radiculares. Anatomische Untersuchungen zur veno¨sen Drainage des menschlichen Ru¨ckenmarkes.

Chapter 3: Vascular Anatomy of the Spine and Spinal Cord Thesis. Medizinische Fakulta¨t der Rheinisch-Westfa¨lischen Technischen Hochschule Aachen, 1996. 18. Moes P, Maillot C. Les veines superficielles de la moelle epinie`re chez l’homme. Essai de systematisation. Archives d’Anatomie, d’Histologie et d’Embryologie Normales et Expe´rimentales. Extrait du tome 64. Paris, Colmar: E´ditions Alsatia, 1981:5–110. 19. Foix Ch, Alajouanine TH. La mye`lite ne´crotique subaigue. Rev Neurol 1926; 33:1–42. 20. Thron A, Koenig E, Pfeiffer P, et al. Dural vascular anomalies of the spine—an important cause of progressive myelopathy. In: Cervos-Navarro J, Ferszt R, eds.Stroke and Microcirculation. New York: Raven Press, 1987.

55

21. Oswald K. Untersuchungen u¨ber das Vorkommen von Sperrmechanismen in den Venae radiculares des Menschen. Thesis. Berlin, 1961. 22. Tadie M, Hemet J, Aaron C, et al. Le dispositif protecteur anti-reflux des veines de la moelle. Neuro-Chir 1979; 25:28–30. 23. Otto J. Morphologie des Sperrmechanismus am Duradurchtritt der Venae radiculares des Menschen. Neuroradiologische und histologische Befunde. Thesis. Medizinische Fakulta¨t der Rheinisch-Westfa¨lischen Technischen Hochschule Aachen, 1990.

4 Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology David S. Liebeskind UCLA Stroke Center, University of California, Los Angeles, California, U.S.A.

INTRODUCTION Collateral circulation in the brain compensates for obstruction to arterial inflow or venous drainage (1). Descriptions of collateral vessels date back to the founding of neurology. Centuries after Sir Thomas Willis described arterial collaterals and their potential significance in disease, angiography illustrated the influential role of these routes. Hemodynamic studies later emphasized the critical impact of collaterals, yet subsequent imaging advances diverted attention away from angiography, seeking neuroprotection and targeting tissue ischemia. In the routine clinical practice of interventional neuroradiology, arterial and venous intracranial collaterals are influential factors in the diagnosis, treatment, and prognosis of various cerebrovascular disorders. Knowledge of collateral anatomy and pathophysiology may expand our understanding of numerous disorders. Correlative studies of imaging features and angiography may facilitate diagnosis and broaden perspectives on novel treatment strategies. This chapter reviews current knowledge of arterial and venous collaterals, emphasizing the specific implications of collaterals in various disorders.

ANATOMY The anatomy of intracranial collaterals greatly influences the capacity of these channels to provide alternative blood flow routes across different regions, with collateral capacity primarily determined by luminal caliber. A description of arterial collateral anatomy may be subdivided between common routes, including Willisian collaterals at the circle of Willis or leptomeningeal anastomoses, and atypical circuits that may develop in response to particular lesions. Similarly, venous collateral anatomy may be described through the typical anastomotic routes and the atypical, a category in which the diversity and complexity of routes is enormous. Some connections such as the posterior communicating artery (PCoA) represent embryonic remnants, whereas other routes form only in response to disease. There

are also numerous collateral extracranial-intracranial (EC-IC) routes, not discussed herein, for both arterial and venous flow diversions. Much of the knowledge regarding the anatomy of intracranial collaterals stems from historical descriptions over the last few hundred years. Only recently, with the advent of angiography and modern imaging techniques, have the functional correlates of these blood flow routes been established. Prior reports have classified collateral routes as primary or secondary functional routes on the basis of anatomical location, yet this classification may be oversimplified, as great variability exists. The knowledge of intracranial collateral anatomy in humans is particularly important in understanding ischemic stroke and other clinical cerebrovascular disorders, as there are considerable differences in anatomy that may preclude successful translation of therapeutic approaches studied in animals (2). Species differences in the configuration of collaterals may also be compounded by differences in collateral anatomy among various individuals or populations. The circle of Willis provides numerous potential routes for blood flow diversion (Fig. 1). All of the Willisian segments, including the anterior communicating artery (ACoA), the proximal anterior cerebral artery (ACA), the PCoA, and the proximal posterior cerebral artery (PCA), may facilitate flow diversion in either direction depending on intraluminal pressure gradients. All of these segments may also be atretic or hypoplastic, yet they retain the ability to develop significant blood flow capacity and luminal expansion. These arterial segments are relatively closely matched in size and vessel wall characteristics with respect to their parent arteries. This configuration allows for interhemispheric collateral flow or compensation for gradients that may develop between the anterior and the posterior circulations. Much emphasis has been placed on the anatomy of the PCoA (3,4). Various terms, including persistent or fetal PCoA anatomy, have been used to differentiate the status of this segment on the basis of diameter measurements at autopsy or on imaging studies, such as magnetic resonance angiography (MRA), where the status of this vessel or dominance is described in

58

Liebeskind

Figure 1 Schematic illustration of the circle of Willis and potential Willisian collaterals, including ACoA (a), proximal ACA (b), PCoA (c), and proximal PCA (d). Abbreviations: ACoA, anterior communicating artery; ACA, anterior cerebral artery; PCoA, posterior communicating artery; PCA, posterior cerebral artery.

relation to the proximal PCA. Descriptive terms of the opposite situation, in which there has been involution of the PCoA’s embryonic origin from the internal carotid artery (ICA) resulting in a hypoplastic PCoA, have questionable validity, as even small-diameter remnants may once again provide blood flow if the need arises. Arterial patterns at the circle of Willis have been categorized by citing the prevalence of certain configurations, but such descriptive data are also questionable, as anatomy may change with disease and age or vary among populations. The leptomeningeal anastomoses bridging distal reaches of the major cerebral arteries are small (*50–400 mm) arteriolar connections that allow for retrograde perfusion of adjacent territories (Fig. 2) (5,6). Such connections display variable configurations, including end-to-end anastomoses, end-to-side connections, and azygous variants (6). These arteriolar anastomoses adjoin the middle cerebral artery (MCA) with both the ACA and the PCA. Anastomoses from the ACA potentially feed the superior or anterior divisions of the MCA, with most of the posterior or inferior division MCA collateral flow arising from the PCA. Such connections are relatively sparse between the ACA and the PCA. The seminal work of Vander Eecken and Adams on 20 human cadavers delineated the principal characteristics of leptomeningeal anastomoses, illustrating considerable variability in the size, number, and location of these collaterals (6). Such

Figure 2 Schematic illustration of principal supratentorial leptomeningeal anastomoses in the brain, including ACA-MCA (a) and PCA-MCA (b) routes. Abbreviations: ACA, anterior cerebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

59

great variability likely influences the results of any particular anatomical study and accounts for much controversy in correlative studies of collateral function with age. Anastomoses also converge over the cerebellar convexities, where the distal branches of the posterior inferior cerebellar arteries (PICAs), anterior inferior cerebellar arteries, and superior cerebellar arteries (SCAs) meet (Fig. 3). Because of the symmetric anatomy of posterior fossa structures, such anastomoses may allow for collateral flow between cerebellar hemispheres and from proximal to distal aspects of the basilar distribution. In cases in which flow demands and pressure gradients exceed the capacity of primary arterial routes and Willisian or leptomeningeal collaterals, atypical routes of collateral flow may develop. Some collateral routes may utilize the paths of normal variants, such as azygous connections between the ACAs. The anterior and posterior choroidal arteries may distribute blood flow in either direction between the anterior and posterior circulations. In cases of moyamoya, this choroidal network is commonly recruited. Other moyamoya arterioles pervade subcortical structures, meandering around occluded MCAs. Anastomoses may shunt flow between the PCA and the SCA at the tentorial edge. More unusual arterial collateral routes may also arise, commonly in association with prominent EC-IC collaterals. Atypical collaterals can be demonstrated in almost any configuration, limited to an extent solely by physical barriers such as the falx or tentorium.

Venous collateral anatomy is best understood in light of typical venous flow patterns (Fig. 4) (7,8). Venous drainage is balanced by superficial and deep systems, with the transcerebral veins allowing for potential shunting in either direction. The superficial system, including the cortical veins and superior sagittal sinus, typically empties the majority of outflow toward the right transverse and sigmoid sinuses and into the jugular. The anastomotic veins of Trolard and Labbe´ shunt flow across the cerebral hemisphere to drainage pathways with lower pressures. Similarly, cortical veins share connections, allowing for diversion of flow. The deep system includes the choroid plexuses and draining veins of the thalami, striatum, periventricular white matter, limbic regions, and rostral brain stem. Larger emissaries of this system include the basal veins, vein of Galen, and straight sinus. The deep system may drain via the straight sinus and into the left transverse system or, alternatively, send flow anteriorly toward the basal veins. Numerous anastomoses abound toward the inferior surface of the brain, allowing for drainage of the deep system. The deep middle cerebral vein, inferior and superior petrosal sinuses, and the basilar plexus may shuttle flow across these regions. Because of the variability in venous outflow patterns and potential anastomoses to relieve focal venous hypertension, minimal attention has been placed on systematic characterization of venous collateral anatomy.

Figure 3 Schematic illustration of cerebellar anastomoses, demonstrating potential collateral flow between SCA (a), AICA (b), and PICA (c). Abbreviations: SCA, superior cerebellar artery; AICA, anterior inferior cerebellar artery; PICA, posterior inferior cerebellar artery.

Figure 4 Schematic illustration of intracranial venous anatomy and typical collateral routes, including vein of Trolard (a), vein of Labbe´ (b), deep middle cerebral vein (c), superior petrosal sinus (d), pterygoid plexus (e), inferior petrosal sinus, and basilar plexus (f).

60

Liebeskind

Figure 5 Right ICA injection on angiography, demonstrating Willisian and leptomeningeal collateral flow in acute stroke due to left ICA occlusion. Abbreviation: ICA, internal carotid artery.

EPIDEMIOLOGY The epidemiology of intracranial collaterals has occasionally been broached in the literature, citing variations in Willisian anatomy or unfounded theories related to collateral development in different cohorts, such as the elderly. Most of these studies have used anatomical data based on autopsy series. Unfortunately, this approach of using anatomical postmortem data to describe potential collateral function does not make sense when one considers the dynamic changes in collateral flow that take place during life. Studies of Willisian configuration in normal individuals are also limited, as collaterals are irrelevant in the absence of disease. Functional assays such as angiographic demonstration of collateral flow during particular clinical scenarios, such as acute ischemic stroke, provide further information (Fig. 5), yet serial changes or a reflection of collateral development may still go unresolved. Other reports in the literature have extrapolated findings on coronary or peripheral arterial collaterals to the cerebral circulation, without validation. Much speculation has addressed the influence of age on collateral flow, yet considerable variability with intracranial collaterals likely occurs with increasing age (9). There is scant epidemiological data on arterial collateral flow, via Willisian or leptomeningeal routes, even within a specific disease state, and the epidemiology of venous collaterals is unknown.

PATHOPHYSIOLOGY The pathophysiology of collateral circulation in the brain has largely been unexplored. Much of the knowledge regarding arterial collaterals has been extrapolated from studies of collateral circulation in other vascular beds or in animal models in which vast differences exist with respect to collateral anatomy. Venous collateral pathophysiology remains virtually completely unknown. Most of the very few studies of

collateral flow in humans relate to anatomical patterns and resultant blood flow, yet very little is known about collateral recruitment. The process of arteriogenesis, or the recruitment and development of preexisting arterioles to accommodate significant flow changes, must be distinguished from angiogenesis, the de novo growth of vessels (10). Features of both may be simultaneously involved with various cerebrovascular disorders, yet the role of these processes is quite distinct. Furthermore, although arterial collaterals may be emphasized in acute ischemic stroke, there are likely changes that take place in the venous system as well. Similarly, failure of venous collateralization in cerebral venous thrombosis (CVT) may ultimately affect arterial inflow, leading to ischemia. As a result, the arterial and venous components must be considered in concert. Venogenesis, the venous counterpart of arteriogenesis, is assumed to be similar to the pathophysiological events that accompany the arterial process. Time is also a critical variable, as the capacity of collaterals to adapt to blood flow derangements changes with time. The particular role or influence of collaterals in specific disorders is considered in subsequent sections of this chapter. With all of these entities, however, it remains important to distinguish the presence of collaterals at a specific time point versus the development or collateralization process itself. The presence or extent of collaterals defined on angiography or imaging reflects the result of an adaptive process responding to significant blood flow alterations. The process of collateralization may be best studied in cases in which collaterals are suboptimal or in cases with progressive ischemia or congestion, allowing for investigation over a prolonged time course. In cases of acute stroke with exuberant collaterals, the process may be obviated or fully realized. The pathophysiology of arteriogenesis has been established in the peripheral and coronary circulations (11,12). Arteriogenesis is principally mediated by

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

61

Figure 6 Table summary of the critical differences between arteriogenesis and angiogenesis and implications in acute cerebral ischemia.

increased fluid shear stress due to mechanical forces that accompany pressure gradients across anastomotic vessel segments. Inflammation plays a key role, incited by cytokine upregulation and macrophage infiltration due to mechanical events at the anastomoses (13). Vascular remodeling allows for potential expansion of the anastomotic vessel radius, thereby increasing flow and alleviating fluid shear stress. This process has considerable differences with respect to angiogenesis (Fig. 6). Arteriogenesis may rapidly culminate in dramatic increases in blood flow, whereas angiogenesis is a local phenomenon that increases permeability and relatively fragile capillary growth without the capacity for significant increases in blood flow. Angiogenesis in the brain occurs in perilesional areas around arteriovenous malformations, tumors, and stroke (14). Recently, the potential for angiogenesis and concomitant neurogenesis has been the focus of investigation in studies of stroke recovery or restorative neurology. The potentially beneficial role of inflammation in cerebral arteriogenesis has yet to be established. In other arterial beds, inflammation simultaneously promotes atherosclerosis and corresponding arteriogenesis. Very recently, genetic upregulation of the actin-binding Rho-activating protein triggered by mechanical factors at anastomotic sites has been discovered (15). Although the basic vascular pathophysiology of arteriogenesis and collateralization is likely to be similar, the anatomy of intracranial collaterals and resultant pathophysiology may be quite distinct (16). Willisian collaterals allow for prompt flow diversion across relatively small distances between arterial territories. Pressure differentials allow for potential circuits to open, causing flow to course toward the ischemic vessel or territory. The diameter of these connections may be quite variable across individuals, likely reflecting developmental variation and subsequent evolving changes during life. Willisian collateralization and the appearance of the circle of Willis is therefore a dynamic process (17). Leptomeningeal anastomoses may also evolve in response to environmental stressors, yet the nature of leptomeningeal collateral perfusion is quite complex. The elongated pathways bridging arterial territories provide blood flow via a limited number of distal anastomoses that perfuse the ischemic territory in a

retrograde fashion. Such reverse arterial flow via selective daughter branches is extremely unusual (Fig. 7), unlike other blood flow routes in the systemic circulation. This pattern of blood flow violates the major hemodynamic principal of Murray’s law, where flow is configured in a manner that is energy efficient (18). It remains unknown whether the distal arterial tree adapts to conform to this ideal mode of blood flow by constricting adjacent daughter arteries. The resulting slow flow is largely diverted toward the parent occluded arterial segment. Intravascular deoxygenation likely occurs because of slow flow past ischemic endothelium and neighboring ischemic brain parenchyma (19).

Figure 7 Diagram of retrograde leptomeningeal flow in the setting of MCA occlusion (a), illustrating anastomotic inflow via isolated distal segments (b, c) and predominant flow toward the trunk of the occluded parent artery (d). Abbreviation: MCA, middle cerebral artery.

62

Liebeskind

In response to ischemia, the microcirculation adapts through loss of flow heterogeneity to accommodate maximal oxygen extraction (20). Low-perfusion hyperemia, the expansion of cerebral blood volume (CBV) despite diminished blood flow due to arterial occlusion, relies heavily on the venous system (9). The mechanisms underlying venous engorgement remain unclear, but progressive expansion of the venous bed downstream from the ischemic arterial territory has been well documented (9). A critical and potentially influential question addresses what leads to the demise of this compensatory mechanism. Cerebral venous steal, reduction of the critical pressure gradient to maintain collateral arterial inflow, and the venoarterial reflex have been postulated as potential factors (21). These factors may also be important in the process of arterialization of the venous system that accompanies other cerebrovascular disorders. Paradoxically, much of the vascular pathophysiology relating to cerebral hemodynamics and intracranial collateral flow was uncovered more than 25 years ago. Although angiography was pivotal in these investigations, it was subsequently replaced by more noninvasive imaging modalities. The unrealized hopes of neuroprotection and isolated focus on the ischemic cascade without consideration of blood flow diverted attention away from hemodynamics and vascular pathophysiology that interventional neuroradiologists often observe in the angiography suite.

CLINICAL CORRELATES The clinical features associated with collateral circulation are often manifested as a dramatic minimization of symptoms despite severe obstruction to normal blood flow. Examples of this phenomenon include asymptomatic acute occlusion of the MCA or clinically silent occlusion of the ICA. Similar events may occur even more frequently with venous collateralization. For instance, CVT involving the principal dural sinuses may go undetected (Fig. 8). Such examples of collateral ability to ameliorate or minimize clinical symptoms are often recognized only when dynamic changes cause transient loss of this ability. In such situations, wide fluctuations in symptoms or neurological deficits may be apparent. These fluctuations are most commonly observed during the very early stages of acute ischemia, during the first minutes and hours after presentation. In cases of MCA ischemia triaged in the prehospital setting as soon as 15 minutes after symptom onset, deficits are often quite minimal, followed by considerable changes and often devastating consequences at later time points. Certain clinical features may also be described with specific disorders. Collateral failure may occur during subacute stroke despite previously sustained perfusion and no apparent blood pressure or hemodynamic changes. In a similar fashion, the limb-shaking transient ischemic attacks (TIAs) of moyamoya may represent only transient collateral failure. Referred auditory phenomena or bruits may indicate venous collateralization. Many of these clinical features are often suspected to be

Figure 8 CTV demonstrating transverse and sigmoid sinus thromboses (arrows) with isolated headache. Abbreviation: CTV, computed tomographic venography.

mediated by collaterals, yet imaging or angiography is often required to substantiate these claims.

IMAGING Unlike the principal arterial and venous routes in the brain, imaging of collaterals evades most current techniques (22), partly because when disease alters the normal pathways for blood flow, collaterals will develop via numerous trajectories. Furthermore, collateral anastomoses tend to be diminutive, as they are recruited only as they are needed. As a result, the goal of imaging collaterals often follows an indirect path where much is inferred on the basis of vascular distributions and the oxymoronic objective of attempting to see what cannot be seen. There is no ideal imaging modality for demonstration of collaterals. Although conventional angiography has been extremely influential in characterization of collaterals and angiographic correlation is often used to substantiate noninvasive markers of collateral flow, there remain qualitative aspects of collateral perfusion that evade angiography. As a result, imaging characterization of intracranial collaterals is founded on integration of findings from various studies. Each modality brings a specific advantage or limitation. For instance, MRA may fail to demonstrate flow in a functional ACoA if a specific threshold is not met. In contrast, computed tomographic angiography (CTA) may demonstrate fairly extensive leptomeningeal collaterals, yet the flow in these segments may be quite minimal

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

Figure 9 CTA source images depicting contrast opacification of leptomeningeal vessels (arrows) in the setting of acute left MCA occlusion. Abbreviations: CTV, computed tomographic venography; MCA, middle cerebral artery.

(Fig. 9). Differences inherent to each modality may accentuate flow or anatomical patency to varying degrees. For most of the clinical disorders encountered in interventional neuroradiology that are described in this chapter, angiography remains paramount for definitive characterization of collateral flow. Whereas Willisian routes are more easily depicted with various imaging modalities, leptomeningeal collaterals are more difficult to delineate. Many of the noninvasive imaging correlates beyond definition of collaterals on conventional angiography have been described and are based on findings in acute ischemic stroke. Extrapolation from acute ischemia to other variants, such as near occlusion or recurrent ischemia bordering on critical perfusion thresholds, has provided insight into other clinical scenarios where arterial collaterals are pivotal (22). Paradoxically, the acute ischemic stroke imaging findings of collateral perfusion may even have valuable information related to venous collateral system as well. For instance, imaging of congested venous drainage in low-perfusion hyperemia may be similar to the findings noted in CVT. Imaging of collaterals is best described by distributions, direct visualization of the anatomical structures themselves, and functional aspects including perfusion. The advent and increasingly routine clinical application of multimodal CT and MRI, incorporating parenchymal images, some extent of angiographic depiction of proximal lesions and corresponding collateral circulation, as well as perfusion may be gleaned.

63

The vascular distributions of arterial or venous collaterals mirror normal patterns of arterial supply or venous drainage. For instance, the borderzones of the MCA territory are based on the normal pattern for the periphery of blood flow in this artery. Unfortunately, these boundaries shift on the basis of variations in normal anatomy and with disease. In general, regions deep within the expected primary vascular distribution are collateral poor, whereas those at the periphery are collateral rich. The extreme variability of venous collateral anatomy makes it quite difficult to infer such distributions. CT or MRI parenchymal sequences may demonstrate patterns suggestive of collateral recruitment. Insular vulnerability in MCA occlusion suggests collateral salvage of more peripheral cortical regions (Fig. 10). Similarly, borderzone infarcts may suggest collateral hemodynamic insufficiency. Direct visualization or imaging of Willisian routes may be feasible with most diagnostic modalities. The short segmental collaterals at the circle of Willis may be demonstrable with transcranial colorcoded Doppler ultrasonography, CTA, MRA, and conventional angiography. In the setting of acute ischemic stroke, Willisian flow patterns reflect changes that took place shortly after arterial occlusion.

Figure 10 CT in acute right MCA stroke with isolated hypodensity of the insular region (arrows). Abbreviations: CT, computed tomography; MCA, middle cerebral artery.

64

Liebeskind

Once flow is restored with proximal recanalization, such diversion of flow and the pattern of Willisian collaterals may change abruptly. Changes in Willisian flow with apparent arterial diameter expansion may also be evident in serial imaging of cases with chronic hypoperfusion or ischemia (17). Leptomeningeal collaterals may be evident on conventional angiography and CTA, and only in rare circumstances with MRA. The slow flow in leptomeningeal collateral routes precludes adequate visualization of these segments with MRA. CTA source images may provide an indication of the extent of leptomeningeal collaterals when viewed in axial format. The ability to depict venous collaterals is analogous to demonstration of leptomeningeal arterial collaterals: conventional angiography and CTA may illustrate these channels, yet MRA or magnetic resonance venography (MRV) is limited. On review of parenchymal sequences, venous collaterals may be seen as engorged or dilated structures with prominent flow voids. Such an appearance may indicate the presence of a peripherally situated arteriovenous malformation (Fig. 11). Conventional angiography may easily demonstrate the presence of arterial or venous collateral routes, with some information regarding functional capacity evident by the temporal appearance of delayed opacification or washout. Such images provide a link between the

anatomical information of vessel appearance and functional aspects of resultant perfusion. Aside from demonstrating the presence of collateral routes, imaging may also provide some insight into the functional aspects or capacity of collaterals. Various modalities may characterize features of collateral blood flow and nutrient or oxygen exchange. The amount of flow in various Willisian collaterals may be estimated from transcranial Doppler (TCD); however, velocity measures alone may be deceiving, as diameter changes may accompany collateral recruitment. In contrast to the previous discussion regarding direct visualization of collaterals, MRA or MRV may have an advantage over CTA or computed tomographic venography (CTV): MRA or MRV accentuates flow characteristics rather than anatomy. Therefore, standard time-of-flight (TOF) MRA may provide very useful information regarding capacity of specific collateral routes. Conventional MRI sequences may provide some subtle, yet very useful, findings related to collateral flow. Fluid-attenuated inversion recovery (FLAIR) MRI vascular hyperintensity (FVH) may be evident in distal aspects of an occluded artery because of slow, retrograde leptomeningeal collateral filling of the artery (Fig. 12) (23,24). Deoxygenation in such distal arterial segments may be evident with gradient-recalled echo (GRE) sequences (19). Such signal loss on GRE associated with deoxygenation may also be observed in draining veins from the ischemic territory in stroke or in engorged venous

Figure 11 MRI evidence of flow voids (arrows) associated with a previously undiagnosed CAVM. Abbreviation: CAVM, cerebral arteriovenous malformation.

Figure 12 Slow, retrograde leptomeningeal collateral filling of the left MCA demonstrating FVH (arrows). Abbreviations: MCA, middle cerebral artery; FVH, FLAIR MRI vascular hyperintensity.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

65

regarding collateral flow. Both modalities demonstrate delay and dispersion of contrast passage that are characteristic of collateral flow (Fig. 14). CBV is often elevated, and microcirculatory changes may be evident if one analyzes the tissue concentration curves in detail. When considering perfusion imaging techniques, one must remember that specific patterns may change rapidly with time and that certain perfusion findings may have different implications in acute versus chronic settings. During chronic phases, specific perfusion abnormalities may be better tolerated.

DISORDERS

Figure 13 GRE prominence of the draining basal vein (arrow) suggesting deoxygenation in the setting of acute stroke. Abbreviation: GRE, gradient-recalled echo.

collaterals because of thrombosis (Fig. 13). Recent developments in MRI have capitalized on the ability to encode spatial or directional information with phase-contrast (PC) MRA techniques, or selective labeling of specific arterial inflow routes with selective arterial spin-labeled (SASL) perfusion (25). Arterial spin-labeled perfusion MRI may reveal delayed arterial transit effects because of slow, leptomeningeal flow supplying the periphery of an ischemic lesion. Commonly used contrast-bolus perfusion techniques with CT or MRI also provide important information

Arterial and venous disorders affecting the brain invariably involve some element of collateral circulation. Collaterals may serve a compensatory role to sustain oxygen and nutrient delivery scaled to metabolic demand, or these alternative blood flow routes may maintain homeostasis through relief of venous congestion. These beneficial roles are complemented by potentially detrimental aspects. For instance, collateral arterial feeders and venous routes may hinder treatment of arteriovenous malformations as these channels proliferate because of humoral and mechanical influences. Although the extent of collaterals may only marginally influence current clinical decision making, the goals of revascularization procedures or treatments are often synonymous with collateralization. Similarities exist in the anatomy of collateral routes and related pathophysiology, yet the role of collaterals is best understood within the context and following discussion of specific cerebrovascular disorders.

Ischemic Stroke Collaterals play a crucial role in acute ischemic stroke (1,22,26). Although not all strokes are associated with thromboembolic occlusion of an intracranial artery or arteriole, ischemia in an arterial territory or bed is universal. Progressive stenosis of a proximal artery may also incite ischemia and elicit collateral recruitment. The degree or extent of collateral compensation varies, as distal cortical branch occlusions or lacunar strokes have limited collateral routes to balance

Figure 14 Schematic illustration of the normal tissue concentration curve (A) and the delay and dispersion associated with collateral flow (B) in the setting of acute stroke.

66

Liebeskind

diminished antegrade flow. The vast knowledge regarding intracranial arterial collateral pathophysiology has been garnered from clinical observations and imaging correlates during acute or subacute cerebral ischemia. During these dynamic early stages of collateral adaptation to ischemia, patients often undergo various imaging studies, including angiography. At later stages, a more stable balance between residual antegrade flow and collaterals develops. As a result, some of the observations regarding collaterals in acute ischemic stroke may be relatively unique, precluding translation of these observations to other clinical settings. The critical role of collaterals is accentuated by the impressive impact of collateral perfusion on recanalization and the fallacy of neuroprotection without blood flow to the penumbra beyond the occluded vessel segment (2). Great emphasis has duly been placed on proximal recanalization; however, such approaches are often futile, and sustenance of the penumbra via collaterals may be the only viable therapeutic option. To capitalize on potential collateral therapeutic interventions, attention must be focused on integration of the wealth of clinical, imaging, and angiographic data, which are often collected during early stages after symptom onset (Fig. 15). Collateral pathophysiology in acute stroke may be ideally described in the setting of MCA occlusion. As soon as distal intraluminal arterial pressure beyond the clot plummets because of failure of antegrade flow, collaterals are recruited. Ischemia associated with a large pressure gradient, and not hypoxia, is the principal driving force that encourages blood flow to traverse the leptomeningeal anastomoses between the distal reaches of the ACA and the PCA and into the MCA field. Augmented flow in these small anastomoses causes a dramatic rise of fluid

shear stress and resultant vascular remodeling because of arteriogenesis. Upregulation of various cytokines and macrophage invasion leads to permeability derangements in these areas at the far periphery of the ischemic field. Eventually, this process leads to an increase in the radii of these small collateral routes. Release of angiotensin II and neuropeptide Y may cause systemic hypertension, yet ironically the relatively intact vasoconstrictive capacity of these distal arterioles may offset attempted hypertensionmediated increases in flow. Retrograde MCA flow is highly energy inefficient, and even slight reductions in the driving pressure gradient may cause collateral failure. CBV elevations, principally due to venous engorgement and loss of flow heterogeneity in the microcirculation, allow for optimal oxygen and nutrient extraction. Eventually, however, a series of detrimental events may ensue, where CBV drops and collateral failure is manifest. The triggers for failure of such beneficial early stages of CBV elevation that has been termed low-perfusion hyperemia remain unclear. Unless correlative imaging or angiographic studies are acquired, the dynamic clinical fluctuations due to collateral flow during acute ischemic stroke may go unfounded. Rapid changes in head positioning and dramatic increases in volume due to fluid boluses may produce profound changes and even normalization of the neurological examination, despite persistent arterial occlusion. Unfortunately, such changes may be transient, as sudden deterioration due to collateral failure may also occur. This paradigm is most worrisome when early hemodynamic improvement deters the clinician from intravenous thrombolysis within three hours and subsequent deterioration occurs well beyond this limited therapeutic window.

Figure 15 Diffusion-weighted imaging (A), time-to-peak PWI map (B), and angiogram (C) in acute left MCA occlusion. Abbreviation: MCA, middle cerebral artery.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

67

Figure 16 Acute left MCA occlusion on MRA (A) with unrevealing diffusion-weighted imaging (B) despite extensive time-to-peak abnormalities on perfusionweighted imaging (C). Abbreviations: MCA, middle cerebral artery; MRA, magnetic resonance angiography.

Case 1

A 92-year-old woman presented with acute onset of right hemiparesis and aphasia. Emergent MRI was acquired, and it revealed occlusion of the left MCA without diffusion-weighted imaging evidence of tissue injury (Fig. 16). FVH illustrated slow, retrograde collateral filling of the left MCA (Fig. 17). After 20 minutes in supine position during the MRI, her neurological deficits completely resolved. On return to the ER, she sat upright and her prior deficits of aphasia and hemiparesis recrudesced. Robust leptomeningeal collaterals were evident on angiography (Fig. 18), and after complete recanalization with mechanical thrombectomy her exam normalized again. Her transient collateral failure associated with changes in head positioning prompted the decision to proceed with thrombectomy. This case demonstrates that collaterals may avert tissue injury despite abrupt cessation of arterial flow and that vigorous collaterals may be evidenced even with advanced age. Almost every imaging modality provides some information regarding collateral flow in acute ischemic stroke. TCD ultrasonography may exhibit flow diversion at the circle of Willis during acute MCA occlusion; increased velocities in other arterial segments may

signify collateral flow. Transcranial color-coded ultrasonography may also provide direct visualization of such Willisian correlates. Often, the most demonstrable indirect evidence of collateral flow is loss of the insular ribbon on noncontrast CT. This finding suggests collateral preservation of the remainder of the MCA field. Infarct growth in the setting of persistent occlusion is also partially a reflection of collateral failure. MRI offers several further facets of collateral flow in acute stroke. FVH in distal segments of the MCA or occluded vessel is due to slow, retrograde leptomeningeal collateral flow (23,24). As the days from symptom onset lapse, this finding subsides because of stabilization or equilibration of collateral flow with infarct growth. Correlation with conventional angiography proves that FVH is not due to thrombosis itself. GRE MRI sequences may depict deoxygenation in distal leptomeningeal collaterals, in draining veins, and in the ischemic tissue as well. Permeability derangements at the borderzones associated with collateral recruitment may also be depicted as subarachnoid hyperintensity on FLAIR (Fig. 19) or with dedicated permeability imaging techniques. Collateral perfusion is most readily identified on perfusion CT or MRI techniques. The footprints of collateral perfusion are evident as prolongation in time-to-peak contrast bolus, elevated

68

Liebeskind

Figure 17 FVH in the distal left MCA (arrow) reflecting predominantly PCA to MCA collateral flow. Abbreviations: FVH, FLAIR MRI vascular hyperintensity; MCA, middle cerebral artery; PCA, posterior cerebral artery; MCA, middle cerebral artery.

mean transit times, augmented CBV, and microcirculatory measures demonstrating loss of flow heterogeneity. These individual parameter maps may be generated with either CT or MRI perfusion techniques.

Other imaging techniques, such as single-photon emission computed tomography (SPECT) or positron emission tomography (PET), may provide additional hemodynamic or even metabolic information related to collateral perfusion, yet such approaches are often cumbersome or impractical in the setting of acute ischemic stroke. The utility of such perfusion imaging studies to depict regions dependent on collateral flow gave rise to the development of mismatch as an imaging surrogate of salvageable penumbra. Various definitions or iterations of mismatch have been developed to ideally select candidates for therapeutic intervention while minimizing risk. Although in the literature much emphasis has been placed on imaging identification of mismatch, incredibly few have substantiated the basis of this approach addressing the actual source of collateral perfusion. Furthermore, it is often forgotten that such imaging techniques provide only a snapshot in time of an extremely dynamic process that may radically differ within minutes. Others have attempted to utilize noninvasive angiographic depictions of collateral flow. CTA source images may provide some indication for the extent of collateral perfusion, yet the prolonged imaging acquisition obliterates temporal information related to flow in order to achieve more anatomical images. MRA may fail to demonstrate leptomeningeal collaterals, yet ipsilateral changes in the PCA may be indicative of PCA to MCA collateral flow in acute stroke (Fig. 20). Such changes may include prolongation or extension of the apparent PCA course on MRA reconstructions, or increases in the apparent PCA diameter (27). Ultimately, definitive proof of collateral supply depends on conventional angiography (28). However, correlation of angiographic findings with the often subtle noninvasive imaging findings noted above provides important information in other cases when angiography is not available or for ongoing imaging research related to collateral circulation. Angiography may reveal flow

Figure 18 Retrograde leptomeningeal collateral filling of the left MCA territory demonstrated with angiography on a left common carotid artery injection. Abbreviation: MCA, middle cerebral artery.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

Figure 19 Subarachnoid hyperintensity on FLAIR due to increased permeability and contrast leakage at the leptomeningeal borderzones. Abbreviation: FLAIR, Fluid-attenuated inversion recovery.

Figure 20 Ipsilateral prominence of the PCA (arrows) on MRA in the setting of acute right MCA occlusion. Abbreviations: PCA, posterior cerebral artery; MRA, magnetic resonance angiography; MCA, middle cerebral artery.

69

diversion via Willisian routes and leptomeningeal sources of perfusion during the arterial phase. Adjacent arteries such as the ACA or the PCA are initially visualized, followed by a momentary delay during transit through anastomoses beyond the resolution of conventional angiography, culminating with retrograde filling of the MCA. Similarly, PICA to SCA anastomoses over the cerebellar convexities may bypass severe stenoses or occlusions of the basilar. The extent, but also the temporal features, of such filling patterns are important for adequate characterization of collateral flow. Several scales have been developed to capture such information, incorporating the delay of collateral perfusion that may be prolonged well beyond the normal capillary filling and into the late venous phases (29,30). Such prolongation of venous perfusion may also provide important information regarding the venous congestion associated with elevated CBV and the low-perfusion hyperemia of acute stroke. As most of the limited number of angiographic scales that capture information on collateral flow emphasize arterial filling, angiographic correlation with perfusion mismatch may be somewhat inaccurate. Following effective reperfusion due to recanalization and cessation of collateral dependence, all of these imaging or angiographic markers of collateral flow disappear. In fact, persistence of such markers of collateral flow may be indicative of incomplete reperfusion. Many of these imaging markers of collateral flow may be seen with other cerebrovascular disorders, but multimodal correlation is often best with the contemporaneous imaging approach unique to acute ischemic stroke. The reliance on angiography for validation of collateral supply largely limits observations on collateral flow in acute stroke to cases in which endovascular therapy is entertained or to the decreasing number of cases in which diagnostic conventional angiography is pursued. Collateral flow has been demonstrated as a strong predictor of favorable clinical outcome in intra-arterial thrombolysis and mechanical thrombectomy (31,32). Collateral flow does not appear to influence the success of proximal recanalization, yet ischemic injury may be lessened in tissue supplied by collaterals beyond the occlusion, or such regions may be sustained until partial restoration of antegrade flow is established. Collateral flow may also thereby decrease the risk of hemorrhagic transformation. The pattern of collateral filling, such as Willisian diversion and configuration of potential ACA collateral flow in ICA occlusion, may have a substantial effect on outcome. The unusual filling pattern of retrograde arterial flow in the ischemic field may also determine the quality or effects of collateral perfusion (Fig. 21). Willisian collaterals have recently been used for delivery of endovascular therapy (33). The first endovascular device utilizing collaterals, NeuroFloTM, is also currently being studied in clinical trials. The device employs augmentation of cerebral blood flow that accompanies titration of concomitant supra- and infrarenal artery aortic balloon inflation during acute stroke (Fig. 22). However, the mechanism of this approach remains to be elucidated. Once proximal recanalization or antegrade

70

Liebeskind

Ongoing investigations of collateral circulation in acute cerebral ischemia may elucidate important clinical features, imaging correlates, and undisclosed pathophysiology of collateral perfusion. Such studies may also provide relevant information for translation to the management of other cerebrovascular disorders. These findings may cease the unshakable failure of neuroprotection related to ongoing disregard for collateral perfusion and facilitate the development of collateral therapeutics (2,26). Endovascular therapy for proximal recanalization may be refined, allowing for collateral augmentation after failed recanalization and for prolonged windows of opportunity. The calculations of time is brain assuming a linear function may also be clarified through consideration of collaterals and the ability to maintain tissue for prolonged periods of time. Revision of this concept may recognize that time is brain because collaterals may fail with time.

Intracranial Atherosclerosis

Figure 21 Frontal projection of a left ICA injection on angiography demonstrating retrograde filling of the MCA. Abbreviation: ICA, internal carotid occlusion.

flow is restored, angiographic collaterals dissipate. In clinical practice, the appearance of robust collaterals on angiography may be deceiving in decision making. One may be compelled to forgo relatively risky interventions to establish antegrade flow when collaterals are excellent. Unfortunately, when left untreated, many of these cases may be prone to collateral failure. Alternatively, the degree of collaterals may lessen stroke severity or clinical outcome even after failed recanalization. Despite these ostensibly critical implications of collateral flow in acute stroke, collaterals are often regarded as only a curious finding on angiography in acute stroke. Most multicenter trials of endovascular therapy to date have considered collaterals only in post hoc analyses.

Although the intracranial arterial collateral circulation has been well described in acute ischemic stroke and in chronic extracranial occlusive disease, knowledge of collaterals in chronic intracranial occlusive disorders is largely limited to moyamoya. In chronic intracranial atherosclerotic disease, arterial stenosis may be isolated to a specific arterial segment, invoking a particular pattern of collateral development. Furthermore, antegrade flow in that territory may not be viable via shorter segmental bypasses provided by the lenticulostriate collaterals of moyamoya. In contrast to acute ischemic stroke, where complete or subtotal occlusion is common, a wide range in the degree of stenosis may be present with intracranial atherosclerosis. The influence of time or temporal features may be quite distinct, as the pace of intracranial atherosclerosis may allow for more considerable collateral compensation (Fig. 23). Collateral flow should theoretically be inconsequential or nonexistent if the stenosis is not hemodynamically significant, exceeding luminal stenoses beyond 60% to 70%. Nevertheless, anecdotal descriptions relate collateral findings with even mild to moderate stenoses. The question remains whether such stenoses are actually hemodynamically significant because of factors beyond luminal stenosis. Collaterals with intracranial occlusive disease may be far more complex than in extracranial disease, as both leptomeningeal and Willisian routes are commonly utilized. If one segregates focal intracranial lesions by potential collateral routes, a different balance may exist between leptomeningeal and Willisian collateral influences. For instance, leptomeningeal collaterals may be pivotal in MCA stenosis, whereas Willisian routes may provide retrograde flow distal to a basilar stenosis. These differences underscore the unique aspects of intracranial collaterals in atherosclerotic disease. Despite these potentially important aspects of collateral flow with intracranial atherosclerosis, the

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

71

Figure 22 Aortagram during placement and titration of balloons on the NeuroFlo device for potential collateral augmentation in acute stroke.

Figure 23 Frontal projection of an angiogram showing retrograde collateral flow in severe atherosclerotic stenosis of the left MCA. Abbreviation: MCA, middle cerebral artery.

subject remains unexplored except for sporadic case series or isolated reports that skirt the topic. Several reasons for this lapse may exist. Intracranial atherosclerosis has only recently been studied in a systematic fashion in the Warfarin Aspirin Symptomatic Intracranial Disease (WASID) trial (34). The study was stopped prematurely on the basis of the futility of detecting a significant difference in treatment between warfarin and aspirin. A parallel investigation of noninvasive imaging correlates, the Stroke Outcomes and Neuroimaging of Intracranial Atherosclerosis (SONIA) study, demonstrated the relatively marginal performance of MRA and TCD for detection of angiographic stenoses in a multicenter setting (35). Willisian collaterals may be readily detected with such noninvasive techniques, yet leptomeningeal collaterals may require conventional angiography (Fig. 24). As a result, many clinicians have deliberated the role of imaging versus angiography and potential treatments for intracranial atherosclerosis. Only very recently has the potential impact of intracranial angioplasty and stenting revived the consideration of conventional angiography and concomitant characterization of collaterals. Future studies will likely need to heed the impact of collaterals on stroke risk and stenting for a given stenosis. Such analyses of collaterals may reveal differences in the role of intracranial collaterals at various stages of disease. Specific collateral patterns, such as distal flow reversal in the basilar

72

Liebeskind

Figure 24 Frontal projection of a left ICA injection on angiography of concomitant ACA and distal MCA stenoses (arrows), where (A) represents early and (B) later phases of angiogram. Abbreviations: ICA, internal carotid occlusion; ACA, anterior cerebral artery; MCA, middle cerebral artery.

collateral flow may also be used in the future to decide when stenting is not indicated despite severe stenoses.

Moyamoya

Figure 25 Angiography demonstrating retrograde leptomeningeal filling of the MCA beyond a proximal stenosis. Abbreviation: MCA, middle cerebral artery.

or leptomeningeal recruitment with MCA stenosis (Fig. 25), may be predicted on the basis of luminal stenosis or provide critical clinical information related to stroke risk. Similarly, the presence of beneficial

Moyamoya is the quintessential model of collateral circulation in the brain. The term has been used to describe a severe multifocal steno-occlusive intracranial arterial disease that most frequently affects young women of Asian descent. Moyamoya syndrome refers to a similar pattern of predominantly proximal anterior circulation occlusive lesions and exuberant collateral formation that occurs in other cohorts or settings (Fig. 26) (36,37). Although much debate has focused on distinguishing this syndrome from the disease, the late-stage pathophysiology relating to collateral flow is same (37). The demographic and clinical features of moyamoya cases in the United States may be strikingly different than classic Asian descriptions (38). As an example, a moyamoya pattern may be seen in older patients with severe atherosclerotic disease because of numerous vascular risk factors. Imaging definitions have been used to describe a moyamoya pattern. Specific MRI criteria have arisen from conventional angiographic stages, delineating patterns that correlate with disease progression. Unfortunately, many aspects continue to fuel debate. When unilateral or subtle findings are noted, many question the diagnosis of moyamoya. Others resist usage of the term when the pathognomonic fine network of lenticulostriate collaterals is inapparent. Irrespective of the diverse range of conditions that has been reported in association with moyamoya, particular features are universal, including initial diversion of flow through

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

73

Figure 26 TOF MRA illustrating multifocal anterior circulation occlusions in moyamoya syndrome. Abbreviation: TOF MRA, time-of-flight magnetic resonance angiography.

Willisian collateral routes and crucial recruitment of leptomeningeal collaterals to supply the vascular territory distal to the steno-occlusive lesions. Abnormal hemodynamics or particular flow patterns may predispose to the development of stenotic lesions, and at later stages, further flow disturbances may lead to aneurysm formation. Moyamoya patterns have been described with various concomitant neurovascular lesions, including atypical aneurysms, vascular anomalies, and arteriovenous malformations (39,40). The clinical features of moyamoya syndrome have remained obscure, as these patients often present with diverse demographic backgrounds and various comorbidities and often have minimal clinical symptoms due to well-developed leptomeningeal collaterals. Patients may present with migrainous headaches due to leptomeningeal dilatation, seizures, or TIAs. Sensory TIAs may be ascribed to migrainous events, yet these brief ischemic episodes may result from transient failure of parietal collaterals. After recovering from such brief symptoms, there is often little impetus to pursue further diagnostic studies. However, devastating strokes, including hemorrhages, may occur. Imaging features, such as the ivy sign (Fig. 27), may be subtle, and vascular disease may go unsuspected unless a dedicated angiographic (noninvasive or conventional) study is acquired (41). Because of such poor recognition of this disorder and the reliance on conventional angiography, angiographers such as interventional neuroradiologists often encounter these patients. Although angiographic descriptions have often focused on the steno-occlusive lesions, angiography of collateral patterns is often dramatic and may be helpful in characterization of

Figure 27 FLAIR depiction of the ivy sign in moyamoya, demonstrating subtle hyperintensities of the subarachnoid space (arrows). Abbreviation: FLAIR, Fluid-attenuated inversion recovery.

the disorder. Aside from the fine, lenticulostriate collaterals that bypass segmental occlusions of the MCA or the ACA, the PCA is often markedly enlarged or prominent, with vigorous leptomeningeal collaterals that supply the cerebral convexities. Progressive enlargement of the PCoA after proximal PCA stenosis follows obliteration of normal antegrade blood flow routes in the anterior circulation (17). Deep transcerebral collaterals may be evident as medullary streaks on MRI (42,43). At later stages of the disorder, enlargement of collaterals between the anterior and posterior choroidal arteries may herald intracerebral hemorrhage (44). The lack of prospective studies of moyamoya, especially within the United States, has lead to a clinical quagmire where little knowledge has been garnered regarding treatment of patients with moyamoya. In general, once an imaging study or conventional angiography confirms the diagnosis, most patients are referred to select vascular neurosurgeons for potential bypass or synangiosis (45). Medical treatment of moyamoya remains uncharted. The specific extent of collateral formation or perfusion derangements on noninvasive studies is rarely used to select candidates for intervention (46). Delineation of an exhausted oxygen extraction fraction on PET may be useful in guiding future standardized approaches (47). Intracranial angioplasty and stent placement has only rarely been described, perhaps because of the fear of

74

Liebeskind

dissection or perforation of the stenotic artery with presumed inflammatory infiltrates. The decision to proceed with EC-IC bypass or synangiosis may be influenced by angiographic features. Following revascularization of such cases, these patients may have limited clinical follow-up by neurologists, but neuroradiologists may serially monitor them with multimodal CT, MRI, or conventional angiography. Following revascularization, clinical symptoms of this progressive disorder may abate because of adequate collateral augmentation (48,49). Interestingly, focal revascularization also appears to improve global perfusion because of easing of demand on various collateral channels (50). Future studies may focus on moyamoya to model collateral flow in acute stroke or to further characterize the pathophysiology of collateral failure.

Extracranial Arterial Stenosis or Occlusion Prominent pressure differentials exerted at the circle of Willis and resultant shifts in blood flow may occur with stenosis or occlusion of the extracranial carotid or vertebral arteries. Although alternative EC-IC routes of blood flow diversion are frequently noted, these changes are accompanied by shifts in blood flow in various Willisian segments. Unilateral carotid occlusion or even vertebral occlusion with a contralateral hypoplastic vertebral artery may elicit such changes.

Willisian segments are able to rapidly shunt flow to the potentially ischemic region or hemisphere. Stenoses must exceed 60% to 70% before hemodynamic implications are evident, yet severe stenoses or occlusions are necessary to cause flow redistribution at the circle of Willis. Moderate stenoses of the extracranial ICA, for instance, may not be hemodynamically significant, but embolic risk may be high. As Willisian collaterals respond only during considerable intraluminal pressure shifts, even severe, ulcerated carotid plaques may not elicit Willisian changes unless hemodynamically significant. Rapid downstream pressure changes due to plaque rupture and sudden carotid occlusion may not be adequately predicted on the basis of Willisian flow patterns unless the culprit lesion is hemodynamically significant at the baseline. More subtle changes may be evident with progressive stenoses, allowing Willisian segments such as the PCoA to grow with time (Fig. 28). The end-diastolic velocity of the CCA on duplex ultrasonography of carotid stenoses may be able to determine the hemodynamic significance of such lesions as correlated with Willisian collateral patterns (51). Once flow is restored, these changes may be readily reversed. For instance, rapid changes in collateral flow and cerebral blood flow distribution may occur after endovascular or surgical revascularization of extracranial stenoses (52,53). Carotid revascularization of stenosis contralateral to an occluded carotid may also improve ACoA flow to the contralateral hemisphere (54).

Figure 28 Willisian collateralization (A, B) of the PCoA chronicled with TOF MRA. Abbreviations: PCoA, posterior communicating artery; TOF MRA, time-of-flight magnetic resonance angiography.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

A multitude of reports have described extensive extracranial occlusive disease with good clinical outcomes. Alternatively, in cases with rapid ICA occlusion due to thromboembolic disease, failure of Willisian segments to compensate for reduced blood flow may lead to devastating strokes. Time appears to be a critical factor—if stenoses or occlusions develop over a long period of time, almost any degree of occlusive disease may be tolerated (55). Even bilateral common carotid occlusion may be sustained with a good clinical course (40). The configuration of Willisian segments and metabolic demand of downstream territories may determine the size, severity, and pattern of cerebral infarction (Fig. 29) (56,57). Presence of ophthalmic flow reversal and leptomeningeal recruitment may signify relative insufficiency of Willisian segments (58). The specific Willisian segments may also differentially affect the pattern of cerebral ischemia. ACoA flow may determine the size and occurrence of borderzone infarction, whereas PCoA flow may be inconsequential (57,59). Almost every diagnostic modality employed in prior reports has demonstrated that collateral compensation and downstream blood flow requirements may play a critical role in delineating asymptomatic and symptomatic carotid occlusions (60). Prediction of recurrent stroke risk with symptomatic carotid occlusion has yielded conflicting results. Some have reported high-residual flow rates in other arterial segments and suggested that prominent collateralization via PCoA flow may identify patients at high risk for recurrent ischemia (61).

75

Improved oxygen extraction has been associated with increased collateral flow after carotid occlusion (62). However, after symptomatic carotid occlusion, recurrent stroke may not be offset by improved collateral flow alone (63). The size of the baseline lesion and subsequent demand likely influences the need for collateral blood flow via the circle of Willis. Differences in technique and patient characteristics have likely influenced the results of numerous studies attempting to conclusively delineate the nature of this relationship (54,62,64). Angiographic definition of collateral flow patterns, including Willisian diversion, pial supply, and delayed venous opacification, may provide important information regarding ischemic risk after symptomatic carotid occlusion. Brief angiographic evaluation of Willisian segments alone may not accurately predict misery perfusion on PET (64). As much controversy persists regarding the role of EC-IC bypass surgery, detailed evaluation of angiographic, hemodynamic, and metabolic status with PET (Fig. 30) is currently being used to identify candidates for revascularization in the Carotid Occlusion Surgery Study (COSS) (65,66). It has also been suggested that the etiology of proximal ICA occlusion may influence outcome. ICA occlusion due to dissection may produce larger infarcts compared with progressive atherosclerotic disease due to the relative insufficiency of collaterals with rapid occlusion following dissection (67). The extent of Willisian collaterals after an occlusion due to dissection may also influence the likelihood for spontaneous recanalization, as robust collaterals may hinder reestablishment of patency in the proximal dissected segment. Various imaging techniques and provocative maneuvers have been used to assess not just stroke risk, but the need for shunting or other periprocedural interventions for carotid revascularization (25,68,69). The absence of ACoA or PCoA flow on angiography has been used to predict the need for shunting during carotid revascularization (70). Phase-contrast MRA, because of its ability to reflect not just the presence of flow but also direction, may be useful to predict changes that may occur with temporary carotid occlusion (71). Prediction of ischemia and the need for shunting may ideally be defined on the basis of noninvasive studies prior to revascularization.

Cerebral Venous Thrombosis

Figure 29 FLAIR demonstration of a relatively small infarct in left ICA occlusion due to dissection and adequate collateral capacity. Abbreviations: FLAIR, Fluid-attenuated inversion recovery; ICA, internal carotid artery.

CVT is relatively uncommon, yet it is often considered the prototypical venous disorder. The cerebral venules and draining sinuses account for more than 60% to 80% of CBV; however, much of the complex physiology in the cerebral venous system remains unexplored. The diverse nature of CVT-associated predisposing conditions or prothrombotic states has attracted much attention. In fact, most of the literature on CVT focuses on the thrombotic aspects, without considering venous flow patterns (Fig. 31). Several neurovascular lesions such as arteriovenous malformations or fistulas may have complex angioarchitecture that promote venous thrombosis, but venous collaterals are otherwise rarely

76

Liebeskind

Figure 30 Oxygen-15 PET data showing increased oxygen extraction fraction in the right hemisphere of a patient with carotid occlusion. Abbreviation: PET, positron emission tomography.

Figure 31 MRV illustration of prominent collateralization in extensive CVT. Abbreviations: MRV, magnetic resonance venography; CVT, cerebral venous thrombosis.

considered. The remarkable distensibility and ability to compensate for pressure differentials within the cerebral venous system have implications for every aspect of CVT from diagnosis to treatment. Thrombosis of a venous sinus or draining vein is offset by diversion of flow into neighboring channels. Unless considerable stasis ensues, the thrombus will remain isolated to the occluded segment until endogenous thrombolytic mechanisms allow for recanalization. Venous pressure may rise in adjacent areas, but this rise is generally well tolerated. Areas of the brain with relative venous insufficiency may be prone to venous hypertension, with subsequent vasogenic edema, hemorrhage, and ultimately ischemia. Venous hypertensive hemorrhage is more common in areas with relatively poor venous collaterals even with small amounts of clot, whereas extensive thrombosis of several major dural sinuses may be inconsequential. Because of extreme variability in venous collateral networks, venous hemorrhage may be difficult to recognize on the basis of location alone, as the principal venous territories are often vague (8,72). Hemorrhage confined to the deep territory of the vein of Labbe´ (Fig. 32) may be one of the few exceptions. The relative dominance of right- versus left-sided drainage of the superficial and deep venous territories

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

Figure 32 Intracerebral hemorrhage due to occlusion of the left vein of Labbe´.

influences venous hypertension and lesion location (8,72,73). The medullary or transcerebral veins may also divert flow in either direction between the superficial and deep systems. The clinical presentation and subsequent course of CVT is completely determined by collaterals (74). In fact, many CVT cases have been estimated to go undiagnosed likely because of considerable compensation by venous collaterals. Even though isolated cortical vein thrombosis may cause neurological deficits in some individuals, the pursuit of this diagnosis is often tempered because it is generally considered a benign disorder due to collateral outflow. When patients present with CVT, headache, seizures, and focal neurological deficits may be noted. Sensory complaints, transient in many cases, may occur because of venous congestion of parietal regions with transverse or sigmoid sinus involvement. Some patients may describe ear fullness, bruits, or other auditory complaints associated with shunting of venous flow (Fig. 33). Dependent head positioning may elicit dramatic increases in symptoms or jugular venous distention. On occasion, a patient may present with an intracerebral hemorrhage of unclear etiology until venous thrombosis or prominent venous collateralization is noted. This broad spectrum of clinical manifestations and imaging presentation with hemorrhage has caused much confusion. Imaging correlates are extremely variable and best defined with MRI. Any angiographic technique (CTV, MRV, or conventional

77

Figure 33 Prominent venous collaterals on CTV causing auditory phenomena in CVT. Abbreviations: CTV, computed tomographic venography; CVT, cerebral venous thrombosis.

angiography) can illustrate thrombotic occlusion and some degree of venous collateralization. MRI offers particular advantages, including demonstration of isolated cortical vein thromboses, prominence or distention of medullary veins, and silent edema or dramatic parenchymal lesions including hemorrhage that may easily resolve over time (56,75,76). MRI may also show mastoid fluid collection due to venous congestion and attempted outflow via collaterals (Fig. 34). Angiography has assumed a minimal role in diagnosis of CVT and is increasingly reserved for rescue treatment when patients deteriorate. Angiography may depict extensive venous collaterals in cases of dural sinus thrombosis. Following thrombolysis or thrombectomy, such venous collaterals may resolve, but the time course may be protracted if thrombus is retained or stasis continues. Such residual venous collaterals may persist indefinitely, causing other clinical symptoms. Residual symptoms such as tinnitus or nystagmus may be partially due to collaterals. These seemingly detrimental manifestations of distended venous collaterals offset the potentially high mortality rate of an otherwise relatively benign disorder.

Dural Arteriovenous Shunts The development of dural arteriovenous shunts or fistulas (DAVFs) has not been elucidated, although

78

Liebeskind

involving the sinuses may promote thrombosis or engender cognitive deficits, including a rapidly progressive dementia (77). Abnormal flow in DAVFs may also be associated with the development of concomitant aneurysms or cerebral arteriovenous malformations (CAVMs). Because of the substantial complexity and variable drainage patterns of these lesions, DAVF classification standards define cases on the basis of specific venous outflow patterns. These specific drainage patterns are also used to guide embolization or surgical resection. Venous drainage of DAVFs through collateral channels may present a far more difficult therapeutic challenge than management of venous congestion in CVT, as the presence and development of a shunt increases the complexity of hemodynamic factors (78).

Cerebral Arteriovenous Malformations

Figure 34 Mastoid fluid collection on MRI in the setting of CVT. Abbreviation: CVT, cerebral venous thrombosis.

angiogenic factors are thought to promote vascular conduits between superficial arteries and veins. Such shunts may be expected to produce arterial steal syndromes, but symptoms typically result from venous outflow disturbances (Fig. 35). Cortical venous reflux and diversion of flow via venous collaterals may produce focal neurological symptoms, tinnitus, or bruits. More diffuse venous drainage patterns

The complex arterial and venous angioarchitecture of CAVMs is akin to the diverse anatomy and blood flow derangements that accompany DAVFs. Unlike DAVFs, however, the contribution of pial collaterals and the influence on more proximal intracranial arterial patterns are greater with CAVMs. CAVMs represent a subset of vascular malformations in the brain. These lesions incorporate arterial and venous segments, typically centered about a nidus, where blood flow changes may induce angiogenesis (Fig. 36). Concomitant arteriogenesis, or development of preexistent arterioles, may also be accompanied by venous recruitment, or venogenesis. Such angioectatic elements are important correlates of collateral circulation that continuously adapt to the evolving hemodynamic disturbances within and around a CAVM. Collateralization is one component of a reactive process within CAVMs in response to hemodynamic disturbances that may diminish tissue perfusion or exacerbate venous congestion. The presence of coexistent angiogenesis within the nidus and more peripherally situated arteriogenesis and venogenesis of

Figure 35 Lateral projection of a left common carotid injection demonstrating a left transverse sinus DAVF with numerous extracranial carotid artery feeders and prominent venous cortical reflux. Abbreviation: DAVF, dural arteriovenous fistula.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

Figure 36 Angiographic demonstration of a complex CAVM with multiple feeders and venous drainage pathways. Abbreviation: CAVM, cerebral arteriovenous malformation.

79

culminate in a moyamoya pattern or vasculopathy (39). A combination of arteriogenic and angiogenic factors likely leads to the proliferation of finer collateral vessels in such cases. Complete occlusion of feeding arteries to CAVMs has also been reported, with all such patients developing exuberant pial collateral supply (81). The venous outflow of CAVMs may be exceedingly complex, with variations of normal drainage patterns in up to one-third of cases (82). Detailed angiographic evaluation of the venous phase may delineate or distinguish abnormal outflow tracts with respect to normal venous drainage. The venous collaterals associated with a CAVM may continuously evolve in response to local changes in the CAVM and remote or diffuse vascular events with age. Clinical manifestations may also depend greatly on the nature of the contributing vessels and resultant perfusion patterns around a CAVM. Considerable leptomeningeal supply may be associated with headaches or seizures. Focal neurological deficits typically result from venous congestion. Cortical venous reflux may result in congestion and compromised perfusion, leading to symptoms. CAVMs situated in closer proximity to the draining venous sinus are less likely to be symptomatic. Although arterial and venous collaterals are an important component of CAVM pathophysiology, numerous other mechanisms are also influential.

Case 2

respective feeders and drainage routes offers an ideal model for the study of collateralization in the brain. Unfortunately, the complexity of these related, but distinct, processes and the diverse anatomy of each particular lesion limit standardized assessment of these important pathophysiological events. As a result, only basic accounts regarding collateral circulation can be described with respect to CAVMs. High-flow states with rapid shunting and diminished tissue perfusion adjacent to a CAVM may result in capillary proliferation around the nidus because of angiogenesis (79). Hypoxia triggers angiogenesis and the formation of new capillaries. Mechanical influences, such as shear stress, drive arteriogenesis or venogenesis in the larger vessels that supply and drain the CAVM. Marked hemodynamic changes due to diversion of arterial flow may result in shifts or transfers of the normal watershed or borderzone regions. Leptomeningeal anastomoses from adjacent arterial territories may contribute to such dramatic shifts in perfusion. When a CAVM resides predominantly within a specific arterial territory, such changes may be less apparent, with only slight variations in leptomeningeal circulation noted distal to such lesions. As with other cerebrovascular lesions, CAVMs may be associated with persistent embryonic variants (e.g., trigeminal, hypoglossal arteries) (80). Variations in the configuration of the circle of Willis are also frequently noted with CAVMs, particularly when these lesions are situated near borderzone regions. Arterial stenoses that develop proximal to a CAVM may rarely

A 49-year-old man underwent embolization of an extensive CAVM of the posterior fossa. Shortly after the procedure, he began experiencing severe retroorbital headaches, which exacerbated when he placed his head in a dependent position. He also noted that when he would rest his head with his hand on the right side of his neck, these symptoms would become quite severe. Serial angiography revealed marked reduction of the nidus (Fig. 37), yet MRV showed enlarged venous collaterals abutting the right aspect of the tentorium (Fig. 38). His referred trigeminal pain syndrome due to engorgement of venous collaterals was likely exacerbated by compression of the right internal jugular vein and abated within weeks because of initiation of gabapentin. Following an acute change of the hemodynamic milieu within a CAVM, collaterals may rapidly adapt in response to pressure changes (83,84). Such changes may occur following rupture, embolization, surgery, or radiation of the CAVM (85). Predicting such changes in flow patterns may be quite difficult (86). Therefore, care must be individualized to the specific case on the basis of the anatomy, flow physiology, clinical manifestations, and technical factors associated with any planned multidisciplinary intervention. Intraoperative angiography may be used to guide surgical management (87). Various approaches, including combinations or staged procedures, are utilized in clinical practice (88). After embolization of a CAVM, serial angiography over a period of months or one to two years may be necessary to

80

Liebeskind

Figure 37 Extensive posterior fossa CAVM on angiography before (A) and after (B) embolization with a small residual nidus. Abbreviation: CAVM, cerebral arteriovenous malformation.

Figure 38 Gadolinium-enhanced MRV demonstrating prominent venous collateral drainage along the right tentorial surface. Abbreviation: MRV, magnetic resonance venography.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

demonstrate adequate obliteration of the nidus. Incomplete obstruction of the nidal-venous junction may allow for angiogenesis and persistence or regrowth of the vascular lesion. In general, the potential collateral supply to a given region influences the likely success of therapeutic embolization of a CAVM (89). Intranidal deposition of embolic material may be necessary to avoid collateral recruitment and regrowth of the CAVM.

Aneurysms The relatively proximal location of most intracranial aneurysms influences the role of collaterals with such lesions. Aneurysms are some of the few abnormalities that directly involve several of the potential collateral segments at the circle of Willis. ACoA aneurysms constitute 30% to 35% of all intracranial aneurysms, whereas disease of this segment is otherwise uncommon. PCoA aneurysms often involve only the origin of this segment, yet other pathology of this vessel is unusual (90). The embryonic development of these anastomotic segments and blood flow changes that may occur at these sites may predispose to aneurysm formation. Abrupt hemodynamic changes may cause rapid shifts in these communicating arterial segments because of pressure differentials, but progressive ischemia may also impose significant flow demands at these junctures. Flow redistribution following occlusion of a proximal vessel may also impart complex hemodynamic changes, leading to aneurysm formation that extends beyond these short diversion segments at the circle of Willis (91). Size and location are important variables in aneurysm management, but also as they relate to collateral circulation. Most small intracranial aneurysms do not invoke or affect collateral flow; however, collaterals play an important role in the context of giant cerebral aneurysms. More distal aneurysms may spare proximal Willisian routes for

81

blood flow diversion, yet these distal lesions may also be more difficult to treat because of variability in leptomeningeal collateral capacity (92,93). As most current approaches for aneurysm treatment involve surgical clipping or endovascular coil embolization with attempted parent vessel preservation, collateral flow may not be relevant. However, when endovascular or surgical parent vessel sacrifice is entertained, testing of collateral supply is mandatory and EC-IC bypass surgery may even be indicated prior to definitive aneurysm treatment. Evaluation of collateral circulation with functional studies is essential to properly gauge the risk of parent vessel sacrifice with giant aneurysms (Fig. 39). Sources of actual and potential collateral flow, including extracranial routes, leptomeningeal anastomoses, and Willisian segments, must be carefully documented. Various protocols have been utilized in the past, including clinical, imaging, and specific angiographic measures. Provocative maneuvers, including induced hypotension, have also been employed. Temporary balloon test occlusion of the parent artery is rapidly stopped if the patient becomes symptomatic. Other aneurysms, when present, should be treated before tolerance testing for parent artery sacrifice. Angiographic measures of adequate collateral circulation may involve stump pressure measurements and preserved perfusion throughout all phases of the injection of the contralateral carotid and/or vertebral arteries during balloon inflation (94,95). To ensure adequate tolerance testing, balloon placement may need to be moved distally in cases in which angiography demonstrates potential collateral channels at the skull base (96). In general, tolerance to parent artery occlusion is greater in the pediatric population, whereas variability in collateral circulation with increasing age makes tolerance testing imperative in adults. Unruptured aneurysms may grow with subsequent thrombosis and occlusion of the parent artery. Mass effect from these lesions may also precipitate

Figure 39 After initial diagnosis of a large left cavernous carotid aneurysm (A) with adequate PCoA (B) and ACoA (C) collateral supply, parent vessel sacrifice was performed (D). Abbreviations: PCoA, posterior communicating artery; ACoA, anterior communicating artery.

82

Liebeskind

flow diversion via collateral routes. These events may culminate in strokes, yet adequate collateral routes may compensate for distal hemodynamic insufficiency. In the setting of occluded or thrombosed giant cerebral aneurysms, collateral circulation typically involves pericallosal anastomoses between the ACA and the PCA, or lateral geniculate anastomoses between the anterior choroidal artery and the PCA. At more distal sites, the aneurysmal vessel and associated collateral blood flow routes may greatly influence the risk of stroke. For instance, giant or fusiform thrombosed aneurysms of the PCA may be offset by efficient collateral anastomoses through the geniculate network or via retrograde leptomeningeal supply (94,97). Interestingly, giant or fusiform thrombosed aneurysms of the MCA rarely recruit adequate leptomeningeal collaterals. Following rupture of an intracranial aneurysm, vasospasm may ensue. Aneurysmal subarachnoid hemorrhage may seem radically dissimilar with respect to ischemic stroke, yet the role of leptomeningeal collaterals in vasospasm may be closely related to such influential blood flow routes in the setting of acute cerebral ischemia. Although complete occlusion of a proximal artery is commonly encountered in acute ischemic stroke, vasospasm only partially diminishes antegrade flow. Despite these differences in the degree of patency of the proximal arteries, retrograde leptomeningeal collaterals are influential in both disorders. Recent studies of vasospasm suggest that various medical therapies and even investigational devices, such as NeuroFlo, may be used to augment collateral flow and improve clinical outcome (98–101). The disproportionately greater opportunity for studying the angiographic aspects of leptomeningeal collateral circulation in vasospasm may provide further insight into the influence and therapeutic manipulation of leptomeningeal collaterals in ischemic stroke.

and collateral recruitment in acute stroke or chronic ischemia. Intracranial tumors may obliterate primary blood flow routes causing diversion of flow through collaterals, or such lesions may also utilize collaterals to sustain ongoing tumor growth. This latter mechanism may be used to treat tumors with intra-arterial delivery of chemotherapeutic agents or for embolization of nutrient vessels. As tumors generally do not respect vascular distributions, the anatomy of collateral vessels may be quite complex or even unique in a particular case. Angiography is therefore critical for diagnostic and therapeutic purposes (103). Occlusion of arterial inflow due to invasion or compression of a proximal vessel and diversion of blood through collateral channels is most common with rapidly growing lesions or vascular tumors such as meningiomas (Fig. 40). Unlike the abrupt arterial occlusion that commonly occurs in acute ischemic stroke, tumor encroachment on a proximal artery typically follows a prolonged course that allows adequate collaterals to develop. Carotid occlusion due to compression from a meningioma may therefore be accompanied by adequate diversion of flow through the circle of Willis. Mass effect and tumor compression may be silent or more obscure on the venous side of the circulation. Because of marked venous distensibility and often complex patterns of venous collateralization, tumor compression may be clinically inapparent. In isolated cases, diversion of blood flow through venous collaterals may elicit headaches, bruits, or tinnitus. During surgical resection of intracranial or even extracranial tumors, knowledge of collateral drainage patterns may be important to

Tumors The inherent vascular correlates in cancer or neoplasia provide insight into the role of collaterals and arteriogenesis in vascular disease. Much of the current knowledge regarding arteriogenesis has emerged from ongoing investigations of tumor angiogenesis, the growth of new vessels feeding cancerous lesions (102). In oncology, the therapeutic goal is antiangiogenesis, or cessation of new vessel growth. This mechanism may be important in the treatment of arteriovenous malformations, yet the principal objective in most cerebrovascular disorders is to grow or recruit additional compensatory blood flow routes. Most molecular studies capitalizing on insight from vascular correlates in tumor pathophysiology remain at very early stages in preclinical development, yet imaging of brain tumors has already fostered translation of dedicated imaging techniques to the vascular realm. Perfusion CT or MRI techniques focusing on CBV and permeability of the blood-brain barrier in tumors provide the ability to investigate arteriogenesis

Figure 40 Progressive encroachment and compression of the torcula and proximal transverse sinuses due to a large meningioma seen on MRI.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology

avert postoperative complications (74,104,105). Most recently, MRI techniques have been developed to delineate cortical venous drainage patterns prior to resection (106). Collateral routes may be utilized for intra-arterial chemotherapy or embolization. Effective treatment may depend on such collateral feeders that may allow continued tumor growth. Prior to therapeutic embolization, assessment of the collateral circulation is mandatory in order to avoid resultant ischemia. Test occlusions with demonstration of collateral compensation, including angiographic flow diversion, stump pressure measurements, and various imaging techniques, may be used for this purpose (95).

Other Considerations A description of collateral anatomy and pathophysiology in the most common disorders encountered within interventional neuroradiology may be outlined on the basis of the primary neurovascular lesion, although systemic aspects remain quite complex. Clinical encounters typically center on restoration or obliteration of blood flow to a focal region, but such lesions may engender manifestations or reflect diffuse pathophysiology within the systemic circulation. Cerebral ischemia due to atherosclerotic disease or moyamoya may cause systemic upregulation of various inflammatory cytokines associated with arteriogenesis. Similarly, cerebral aneurysms or arteriovenous malformations may be associated with remote vascular lesions due to common underlying pathophysiology. Such examples emphasize the important role of endogenous homeostatic mechanisms and common vascular pathophysiology despite our focus on the cerebral circulation. Even within the cerebrovasculature, concomitant lesions may be intertwined, such as aneurysms in the setting of moyamoya or the association of venous thrombosis with arteriovenous malformations. Our current knowledge of cerebrovascular anatomy and flow physiology requires further investigation with respect to the molecular and genetic determinants associated with collateral recruitment.

CONCLUSIONS The essential role of collateral circulation in cerebrovascular disorders has been recognized for centuries, yet detailed characterization of these blood flow routes remains an elusive goal. Collaterals develop in concert with underlying vascular lesions or disorders, compensating for potential blood flow derangements but also serving as a marker of disease with manifestations that may be detrimental. Angiography persists as the principal modality for defining the anatomy and associated blood flow routes of collaterals in the cerebral arterial and venous systems. Advanced noninvasive imaging modalities provide additional information regarding collateral pathophysiology, yet correlation with conventional angiography is often needed. The unique features of collateral circulation in humans limit translation of

83

animal research to the clinical realm, thereby reinforcing the need to learn from correlative studies in our patients. Diagnostic, therapeutic, and prognostic implications of intracranial collaterals underscore the importance of these blood flow routes in interventional neuroradiology. Ongoing refinement of current therapeutic approaches for cerebrovascular disorders will undoubtedly depend on further knowledge of collateral perfusion.

REFERENCES 1. Liebeskind DS. Collateral circulation. Stroke 2003; 34:2279–2284. 2. Liebeskind DS. Neuroprotection from the collateral perspective. IDrugs 2005; 8:222–228. 3. Alastruey J, Parker KH, Peiro J, et al. Modelling the circle of Willis to assess the effects of anatomical variations and occlusions on cerebral flows. J Biomech 2007; 40(8): 1794–1805. 4. Jongen JC, Franke CL, Ramos LM, et al. Direction of flow in posterior communicating artery on magnetic resonance angiography in patients with occipital lobe infarcts. Stroke 2004; 35:104–108. 5. Brozici M, van der Zwan A, Hillen B. Anatomy and functionality of leptomeningeal anastomoses: a review. Stroke 2003; 34:2750–2762. 6. Vander Eecken HM. [Morphological significance of leptomeningeal anastomoses confined to the territory of cerebral arteries.] Acta Neurol Psychiatr Belg 1954; 54:525–532. 7. Andeweg J. The anatomy of collateral venous flow from the brain and its value in aetiological interpretation of intracranial pathology. Neuroradiology 1996; 38:621–628. 8. Andeweg J. Consequences of the anatomy of deep venous outflow from the brain. Neuroradiology 1999; 41:233–241. 9. Tomita M, Gotoh F, Amano T, et al. ‘‘Low perfusion hyperemia’’ following middle cerebral arterial occlusion in cats of different age groups. Stroke 1980; 11:629–636. 10. Buschmann I, Schaper W. Arteriogenesis versus angiogenesis: two mechanisms of vessel growth. News Physiol Sci 1999; 14:121–125. 11. Buschmann I, Schaper W. The pathophysiology of the collateral circulation (arteriogenesis). J Pathol 2000; 190:338–342. 12. Scholz D, Cai WJ, Schaper W. Arteriogenesis, a new concept of vascular adaptation in occlusive disease. Angiogenesis 2001; 4:247–257. 13. Buschmann I, Heil M, Jost M, et al. Influence of inflammatory cytokines on arteriogenesis. Microcirculation 2003; 10:371–379. 14. Wei L, Erinjeri JP, Rovainen CM, Woolsey TA. Collateral growth and angiogenesis around cortical stroke. Stroke 2001; 32:2179–2184. 15. Heil M, Schaper W. Influence of mechanical, cellular, and molecular factors on collateral artery growth (arteriogenesis). Circ Res 2004; 95:449–458. 16. Liebeskind DS. Anatomic considerations in therapeutic arteriogenesis for cerebral ischemia. Circulation 2004; 109:e4; author reply e4. 17. Liebeskind DS, Sansing LH. Willisian collateralization. Neurology 2004; 63:344. 18. Kassab GS. Scaling laws of vascular trees: of form and function. Am J Physiol Heart Circ Physiol 2006; 290:H894–H903. 19. Liebeskind DS, Ances BM, Weigele JB, et al. Intravascular deoxygenation of leptomeningeal collaterals detected with gradient-echo MRI. Stroke 2004; 35:266.

84

Liebeskind

20. Tsai AG, Johnson PC, Intaglietta M. Oxygen gradients in the microcirculation. Physiol Rev 2003; 83:933–963. 21. Pranevicius M, Pranevicius O. Cerebral venous steal: blood flow diversion with increased tissue pressure. Neurosurgery 2002; 51:1267–1273; discussion 1273–1264. 22. Liebeskind DS. Collaterals in acute stroke: beyond the clot. Neuroimaging Clin N Am 2005; 15:553–573, x. 23. Liebeskind DS, Bemporad JA, Melhem ER. FLAIR vascular hyperintensity as a marker of leptomeningeal collaterals in subacute stroke. Proceedings of the ASNR 41st Annual Meeting, Washington, DC, 2003. 24. Liebeskind DS, Cucchiara BL, Kasner SE, et al. FLAIR MRI vascular hyperintensity reflects perfusion status in cerebral ischemia. Presented at the 53rd Annual Meeting of the American Academy of Neurology, Philadelphia, PA, 2001. 25. Hendrikse J, van Osch MJ, Rutgers DR, et al. Internal carotid artery occlusion assessed at pulsed arterial spinlabeling perfusion MR imaging at multiple delay times. Radiology 2004; 233:899–904. 26. Liebeskind DS. Collateral therapeutics for cerebral ischemia. Expert Rev Neurother 2004; 4:255–265. 27. Liebeskind DS, Weigele JB, Hurst RW. Collateralization of the posterior cerebral artery. Stroke 2004; 35:266. 28. Liebeskind DS. Location, location, location: angiography discerns early MR imaging vessel signs due to proximal arterial occlusion and distal collateral flow. AJNR Am J Neuroradiol 2005; 26:2432–2433; author reply 2433–2434. 29. Liebeskind DS, Sayre JW, Weigele JB, et al. Angiographic collateral scales for intra-arterial thrombolysis. Proceedings of the ASNR 42nd Annual Meeting, Seattle, 2004. 30. Higashida RT, Furlan AJ, Roberts H, et al. Trial design and reporting standards for intra-arterial cerebral thrombolysis for acute ischemic stroke. Stroke 2003; 34:e109–e137. 31. Kucinski T, Koch C, Eckert B, et al. Collateral circulation is an independent radiological predictor of outcome after thrombolysis in acute ischaemic stroke. Neuroradiology 2003; 45:11–18. 32. Liebeskind DS, Hurst RW for the MERCITM. Investigators. Angiographic collaterals and outcome in mechanical thrombolysis. Stroke 2005: 36:449. 33. Kole MK, Pelz DM, Lee DH, et al. Intra-arterial thrombolysis of embolic middle cerebral artery using collateral pathways. Can J Neurol Sci 2005; 32:257–260. 34. Chimowitz MI, Lynn MJ, Howlett-Smith H, et al. Comparison of warfarin and aspirin for symptomatic intracranial arterial stenosis. N Engl J Med 2005; 352: 1305–1316. 35. Stroke Outcome and Neuroimaging of Intracranial Atherosclerosis (SONIA) Trial Investigators. Stroke outcome and neuroimaging of intracranial atherosclerosis (SONIA): design of a prospective, multicenter trial of diagnostic tests. Neuroepidemiology 2004; 23:23–32. 36. Zipfel GJ, Fox DJ Jr., Rivet DJ. Moyamoya disease in adults: the role of cerebral revascularization. Skull Base 2005; 15:27–41. 37. Peerless SJ. Risk factors of moyamoya disease in Canada and the USA. Clin Neurol Neurosurg 1997; 99(suppl 2): S45–S48. 38. Numaguchi Y, Gonzalez CF, Davis PC, et al. Moyamoya disease in the United States. Clin Neurol Neurosurg 1997; 99(suppl 2):S26–S30. 39. Nawawi O, Sinnasamy M, Ramli N. Unilateral moyamoya disease with co-existing arteriovenous malformation. Br J Radiol 2006; 79:e12–e15. 40. Liu HM, Lai DM, Tu YK, et al. Aneurysms in twig-like middle cerebral artery. Cerebrovasc Dis 2005; 20:1–5.

41. Maeda M, Tsuchida C. ‘‘Ivy sign’’ On fluid-attenuated inversion-recovery images in childhood moyamoya disease. AJNR Am J Neuroradiol 1999; 20:1836–1838. 42. Takanashi J, Suzuki H, Barkovich AJ, et al. Medullary streaks: dilated medullary vessels in chronic ischemia in children. Neurology 2003; 61:583–584. 43. Kimura H, Oka K, Ikeda K, et al. The clinical significance of cerebral veins in moyamoya disease. Clin Neurol Neurosurg 1997; 99(suppl 2):S90–S95. 44. Morioka M, Hamada J, Kawano T, et al. Angiographic dilatation and branch extension of the anterior choroidal and posterior communicating arteries are predictors of hemorrhage in adult moyamoya patients. Stroke 2003; 34:90–95. 45. Smith ER, Scott RM. Surgical management of moyamoya syndrome. Skull Base 2005; 15:15–26. 46. Togao O, Mihara F, Yoshiura T, et al. Cerebral hemodynamics in moyamoya disease: correlation between perfusion-weighted MR imaging and cerebral angiography. AJNR Am J Neuroradiol 2006; 27:391–397. 47. Piao R, Oku N, Kitagawa K, et al. Cerebral hemodynamics and metabolism in adult moyamoya disease: comparison of angiographic collateral circulation. Ann Nucl Med 2004; 18:115–121. 48. Scott RM, Smith JL, Robertson RL, et al. Long-term outcome in children with moyamoya syndrome after cranial revascularization by pial synangiosis. J Neurosurg 2004; 100:142–149. 49. Irikura K, Miyasaka Y, Kurata A, et al. The effect of encephalo-myo-synangiosis on abnormal collateral vessels in childhood moyamoya disease. Neurol Res 2000; 22:341–346. 50. Jefferson AL, Glosser G, Detre JA, et al. Neuropsychological and perfusion MR imaging correlates of revascularization in a case of moyamoya syndrome. AJNR Am J Neuroradiol 2006; 27:98–100. 51. Kamouchi M, Kishikawa K, Okada Y, et al. Reappraisal of flow velocity ratio in common carotid artery to predict hemodynamic change in carotid stenosis. AJNR Am J Neuroradiol 2005; 26:957–962. 52. Niesen WD, Weiller C, Sliwka U. Unstable cerebral hemodynamics in carotid artery occlusion and large hemispheric stroke: A cerebral blood flow volume study. J Neuroimaging 2004; 14:246–250. 53. Vriens EM, Wieneke GH, Hillen B, et al. Flow redistribution in the major cerebral arteries after carotid endarterectomy: a study with transcranial Doppler scan. J Vasc Surg 2001; 33:139–147. 54. Rutgers DR, Klijn CJ, Kappelle LJ, et al. Sustained bilateral hemodynamic benefit of contralateral carotid endarterectomy in patients with symptomatic internal carotid artery occlusion. Stroke 2001; 32:728–734. 55. Demaria RG, Albat B, Frapier JM, et al. Vertebral artery surgery with cardiopulmonary bypass and deep hypothermia. J Cardiovasc Surg (Torino) 2000; 41:299–302. 56. Yamauchi H, Kudoh T, Sugimoto K, et al. Pattern of collaterals, type of infarcts, and haemodynamic impairment in carotid artery occlusion. J Neurol Neurosurg Psychiatry 2004; 75:1697–1701. 57. Hendrikse J, Hartkamp MJ, Hillen B, et al. Collateral ability of the circle of Willis in patients with unilateral internal carotid artery occlusion: border zone infarcts and clinical symptoms. Stroke 2001; 32:2768–2773. 58. Hofmeijer J, Klijn CJ, Kappelle LJ, et al. Collateral circulation via the ophthalmic artery or leptomeningeal vessels is associated with impaired cerebral vasoreactivity in patients with symptomatic carotid artery occlusion. Cerebrovasc Dis 2002; 14:22–26.

Chapter 4: Intracranial Collateral Routes and Anastomoses in Interventional Neuroradiology 59. Bisschops RH, Klijn CJ, Kappelle LJ, et al. Collateral flow and ischemic brain lesions in patients with unilateral carotid artery occlusion. Neurology 2003; 60:1435–1441. 60. Vernieri F, Pasqualetti P, Matteis M, et al. Effect of collateral blood flow and cerebral vasomotor reactivity on the outcome of carotid artery occlusion. Stroke 2001; 32: 1552–1558. 61. Rutgers DR, Klijn CJ, Kappelle LJ, et al. Recurrent stroke in patients with symptomatic carotid artery occlusion is associated with high-volume flow to the brain and increased collateral circulation. Stroke 2004; 35:1345– 1349. 62. Derdeyn CP, Videen TO, Fritsch SM, et al. Compensatory mechanisms for chronic cerebral hypoperfusion in patients with carotid occlusion. Stroke 1999; 30:1019–1024. 63. Rutgers DR, Donders RC, Vriens EM, et al. A comparison of cerebral hemodynamic parameters between transient monocular blindness patients, transient ischemic attack patients and control subjects. Cerebrovasc Dis 2000; 10:307–314. 64. Derdeyn CP, Shaibani A, Moran CJ, et al. Lack of correlation between pattern of collateralization and misery perfusion in patients with carotid occlusion. Stroke 1999; 30:1025–1032. 65. Horn P, Vajkoczy P, Schmiedek P, et al. Evaluation of extracranial-intracranial arterial bypass function with magnetic resonance angiography. Neuroradiology 2004; 46:723–729. 66. Adams HP Jr., Powers WJ, Grubb RL Jr., et al. Preview of a new trial of extracranial-to-intracranial arterial anastomosis: the carotid occlusion surgery study. Neurosurg Clin N Am 2001; 12:613–624, ix–x. 67. Milhaud D, de Freitas GR, van Melle G, et al. Occlusion due to carotid artery dissection: a more severe disease than previously suggested. Arch Neurol 2002; 59:557–561. 68. Costin M, Rampersad A, Solomon RA, et al. Cerebral injury predicted by transcranial doppler ultrasonography but not electroencephalography during carotid endarterectomy. J Neurosurg Anesthesiol 2002; 14:287–292. 69. Ajisaka R, Masuoka T, Fujita T, et al. Effect of nifedipine on left ventricular function during exercise in patients with stable effort angina. Relation of its efficacy to the severity of coronary artery disease. Jpn Heart J 1989; 30:13–25. 70. Kim GE, Cho YP, Lim SM. The anatomy of the circle of Willis as a predictive factor for intra-operative cerebral ischemia (shunt need) during carotid endarterectomy. Neurol Res 2002; 24:237–240. 71. Bagan P, Azorin J, Salama J, et al. The value of phasecontrast magnetic resonance angiography of the circle of Willis in predicting cerebral ischemia-hypoxia (shunt need) during carotid endarterectomy. Surg Radiol Anat 2005; 27:544–547. 72. Bergui M, Bradac GB. Progressive stroke, lacunae, and systemic blood pressure. Stroke 2002; 33:2735–2736. 73. Kuker W, Schmidt F, Friese S, et al. Unilateral thalamic edema in internal cerebral venous thrombosis: Is it mostly left?. Cerebrovasc Dis 2001; 12:341–345. 74. Mendelowitsch A, Sekhar LN, Clemente R, et al. EC-IC bypass improves chronic ischemia in a patient with moyamoya disease secondary to sickle cell disease: an in vivo microdialysis study. Neurol Res 1997; 19:66–70. 75. Rottger C, Trittmacher S, Gerriets T, et al. Reversible MR imaging abnormalities following cerebral venous thrombosis. AJNR Am J Neuroradiol 2005; 26:607–613. 76. Kawaguchi T, Kawano T, Kaneko Y, et al. Classification of venous ischaemia with MRI. J Clin Neurosci 2001; 8(suppl 1):82–88.

85

77. Hurst RW, Bagley LJ, Galetta S, et al. Dementia resulting from dural arteriovenous fistulas: the pathologic findings of venous hypertensive encephalopathy. AJNR Am J Neuroradiol 1998; 19:1267–1273. 78. Flueler U, Taylor D, Hing S, et al. Hemifacial spasm in infancy. Arch Ophthalmol 1990; 108:812–815. 79. Pile-Spellman JM, Baker KF, Liszczak TM, et al. High-flow angiopathy: cerebral blood vessel changes in experimental chronic arteriovenous fistula. AJNR Am J Neuroradiol 1986; 7:811–815. 80. Raybaud CA, Livet MO, Jiddane M, et al. Radiology of ischemic strokes in children. Neuroradiology 1985; 27:567–578. 81. Enam SA, Malik GM. Association of cerebral arteriovenous malformations and spontaneous occlusion of major feeding arteries: clinical and therapeutic implications. Neurosurgery 1999; 45:1105–1111; discussion 1111–1102. 82. Vinuela F, Nombela L, Roach MR, et al. Stenotic and occlusive disease of the venous drainage system of deep brain AVM’s. J Neurosurg 1985; 63:180–184. 83. Kinouchi H, Mizoi K, Takahashi A, et al. Combined embolization and microsurgery for cerebral arteriovenous malformation. Neurol Med Chir (Tokyo) 2002; 42:372–378; discussion 379. 84. Jungreis CA, Horton JA, Hecht ST. Blood pressure changes in feeders to cerebral arteriovenous malformations during therapeutic embolization. AJNR Am J Neuroradiol 1989; 10:575–577. 85. Stein BM, Wolpert SM. Surgical and embolic treatment of cerebral arteriovenous malformations. Surg Neurol 1977; 7:359–369. 86. Oxelbark S, Mannting F, Morgan MG, et al. Revascularization of infarcted vs. viable myocardium. Effects on symptoms, physical performance and global/regional left ventricular function. Scand J Thorac Cardiovasc Surg 1991; 25:81–87. 87. Vitaz TW, Gaskill-Shipley M, Tomsick T, et al. Utility, safety, and accuracy of intraoperative angiography in the surgical treatment of aneurysms and arteriovenous malformations. AJNR Am J Neuroradiol 1999; 20:1457–1461. 88. Groden C, Grzyska U, Freitag HJ, et al. Two-step presurgical endovascular treatment of five arteriovenous malformations partially fed by single vessels en passage. Surg Neurol 1999; 52:160–165; discussion 165–166. 89. Fournier D, Terbrugge K, Rodesch G, et al. Revascularization of brain arteriovenous malformations after embolization with bucrylate. Neuroradiology 1990; 32:497–501. 90. Kaspera W, Majchrzak H, Kopera M, et al. ‘‘True’’ aneurysm of the posterior communicating artery as a possible effect of collateral circulation in a patient with occlusion of the internal carotid artery. a case study and literature review. Minim Invasive Neurosurg 2002; 45:240–244. 91. Wolf R, Lichtlen PR. [Relation between plasma fibrinogen and the function of collateral coronary vessels]. Z Kardiol 1995; 84:348–359. 92. Biondi A, Jean B, Vivas E, et al. Giant and large peripheral cerebral aneurysms: etiopathologic considerations, endovascular treatment, and long-term follow-up. AJNR Am J Neuroradiol 2006; 27:1685–1692. 93. Macchi PJ, Grossman RI, Gomori JM, et al. High field MR imaging of cerebral venous thrombosis. J Comput Assist Tomogr 1986; 10:10–15. 94. van Rooij WJ, Sluzewski M, Slob MJ, et al. Predictive value of angiographic testing for tolerance to therapeutic occlusion of the carotid artery. AJNR Am J Neuroradiol 2005; 26:175–178. 95. Morishima H, Kurata A, Miyasaka Y, et al. Efficacy of the stump pressure ratio as a guide to the safety of permanent

86

96.

97.

98. 99.

100.

101.

Liebeskind occlusion of the internal carotid artery. Neurol Res 1998; 20:732–736. Stolz E, Allendorfer J, Jauss M, et al. Sonographic harmonic grey scale imaging of brain perfusion: scope of a new method demonstrated in selected cases. Ultraschall Med 2002; 23:320–324. Hallacq P, Piotin M, Moret J. Endovascular occlusion of the posterior cerebral artery for the treatment of p2 segment aneurysms: retrospective review of a 10-year series. AJNR Am J Neuroradiol 2002; 23:1128–1136. Macdonald RL, Curry DJ, Aihara Y, et al. Magnesium and experimental vasospasm. J Neurosurg 2004; 100:106–110. Toyoda K, Chu Y, Heistad DD. Gene therapy for cerebral vascular disease: update 2003. Br J Pharmacol 2003; 139:1–9. Lynch JR, Wang H, McGirt MJ, et al. Simvastatin reduces vasospasm after aneurysmal subarachnoid hemorrhage: results of a pilot randomized clinical trial. Stroke 2005; 36:2024–2026. Kornblihtt LI, Cocorullo S, Miranda C, et al. Moyamoya syndrome in an adolescent with essential thrombocythe-

102.

103.

104.

105.

106.

mia: successful intracranial carotid stent placement. Stroke 2005; 36:E71–E73. Harrigan MR, Ennis SR, Sullivan SE, et al. Effects of intraventricular infusion of vascular endothelial growth factor on cerebral blood flow, edema, and infarct volume. Acta Neurochir (Wien) 2003; 145:49–53. Nelson PK, Levy D, Masters LT, et al. Neuroendovascular management of intracranial aneurysms. Neuroimaging Clin N Am 1997; 7:739–762. Wustenberg EG, Offergeld C, Zahnert T, et al. Extension of intracranial thrombosis after unilateral dissection of the internal jugular vein. Arch Otolaryngol Head Neck Surg 2005; 131:430–433. DiMeco F, Li KW, Casali C, et al. Meningiomas invading the superior sagittal sinus: surgical experience in 108 cases. Neurosurgery 2004; 55:1263–1272; discussion 1272–1264. Lin SK, Ryu SJ, Chang YJ, et al. Clinical relevance of musical murmurs in color-coded carotid and transcranial duplex sonographies. AJNR Am J Neuroradiol 2006; 27:1493–1497.

5 CT Imaging and Physiologic Techniques in Interventional Neuroradiology Ronald L. Wolf Department of Radiology, Neuroradiology Section, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A.

BACKGROUND

CT TECHNIQUES

The first clinical CT scans of the brain were obtained in 1972 on a prototype CT scanner developed by Hounsfield. The first clinical scanner, EMI Mark I, was introduced in 1973. The Nobel Prize in medicine was awarded to Sir Godfrey Hounsfield and Alan Cormack in 1979 for the development of computerassisted tomography, underscoring the impact of this achievement on clinical medicine. Improvements in design led to slip ring technology and thus helical or spiral scanning, which was first introduced in 1975 at Varian and then reintroduced with a more practical design in 1985 and 1987 by Toshiba and Siemens, respectively. Whereas the EMI Mark I scanner required approximately five minutes for the acquisition of one imaging section, spiral scanners could cover several centimeters in less than 60 seconds, obtaining nearly isotropic resolution over a small field of view but at the expense of tube heating (1). Multidetector CT (MDCT) was first implemented on the first-generation EMI Mark I scanner, which acquired two sections at a time. Elscint introduced the first helical scanner with dual detectors in 1992 and followed it by detector configurations of four channels or more starting around 1998. Current configurations have up to 64 channels, so that 30 to 40 cm can be covered in less than 30 seconds, obtaining nearly isotropic resolution (1). With MDCT configurations, thin and thick sections are effectively acquired simultaneously while covering a large distance along the z axis. Data for thin sections can thus be acquired and combined to reconstruct thicker sections for reading, while retaining advantages of thin-section scanning such as minimization of partial volume artifact and resultant streaking. As long as raw data are retained, additional thin-section reconstructions can be obtained retrospectively for multiplanar and 3D reformatting (2).

Conventional CT and CT Myelography/ Cisternography CT of the brain can be performed using sequential single slice, helical multislice, or multidetector multislice techniques. The American College of Radiology guidelines suggest section thicknesses in the supratentorial compartment of 10 mm or less in adults (5 mm or less in children under age 10), and 5 mm or less in the posterior fossa in adults or children. For the skull base, sections of 3 mm or less are preferred. If multiplanar reformats are required (e.g., facial and skull base fractures in the setting of trauma) or if 3D rendering is to be performed, sections of 2 mm or less should be obtained. With multidetector scanners, multiple data sets (e.g., for standard CT of the brain, face, skull base, and/or temporal bones) can be acquired prospectively and simultaneously by selecting detector spacing on the order of 1 mm or less, combining thin sections for evaluation of brain and soft tissues and reconstructing thin sections with overlap for reformatting or rendering retrospectively as needed for the face, skull base, and temporal bones (Fig. 1). The same is true for CT of the spine after myelogram or CT of the head and the face after cisternogram. Coronal sections can be directly acquired by angling the gantry and positioning the patient with the neck extended. Highquality reformatted images from axial data are now obtainable with newer MDCT systems, but direct coronal acquisitions are still needed if reformats do not provide sufficient detail for clinical decision-making. Also, if the patient moves during the axial acquisition, the reformats will of course be degraded. CT is the primary imaging modality for emergent indications such as trauma and acute changes in neurologic status, including ischemia and intracranial hemorrhage. For most applications concerning structural imaging of the brain, skull base, cranial vault,

88

Wolf

Figure 1 Simultaneous acquisition of standard NECT and high-resolution CT of skull base with MDCT. (A) NECT (3-mm section) shows pneumocephalus and temporal bone fracture (arrowhead). (B) 1-mm bone reconstructions from the same raw data also reveal fracture line through the right carotid canal (arrow), nearly invisible on standard NECT. Abbreviation: NECT, nonenhanced CT.

and spine, nonenhanced CT (NECT) is most often adequate. The primary indications for the use of IV contrast include infection and neoplasm, but in practice this use is relatively uncommon because CNS infection and neoplasm will almost always prompt an MRI, obviating the need for enhanced CT. IV contrast is, however, preferred for routine CT of the soft tissues of the neck and required for CT angiography (CTA). In the setting of penetrating trauma, NECT of the neck can be useful for assessing the trajectory of injury and proximity to vascular structures, as well as evaluating for foreign materials such as bullets and fragments. Intrathecal contrast is of course required for myelography or cisternography.

CT Angiography CTA has become an attractive alternative to digital subtraction angiography (DSA) for rapid evaluation of the cervical and cerebral vasculature. A complete evaluation of the brain with NECT, CTA of the head and/or neck, and perhaps CT perfusion and contrastenhanced CT (CECT) of the head can be obtained in less than 10 to 15 minutes. CTA is well tolerated and in many cases preferred by patients compared with MR imaging and MR angiography (3). Source images are available immediately and provide most of the diagnostic information necessary for decision making. Data for the head and the neck can be obtained during the same imaging run (and bolus), and smaller field-ofview reconstructions at more closely spaced intervals (i.e., on the order of 0.5 mm) reconstructed for the circle

of Willis using the initial data set. Creation of volumerendered, multiplanar reformatted, and/or maximum intensity projection (MIP) images can be performed relatively quickly at the scan console or on a separate workstation. Typically, consecutive thin-axial sections (*1– 2 mm) are obtained during IV contrast administration of 75 to 100 cc iodinated contrast at 3 to 4 cc/sec, followed by a saline chasing bolus. Reconstructing with an overlap of about 50% improves the appearance of volume-rendered and reformatted images. Timing the bolus on the basis of contrast opacification in the aortic arch or left ventricle allows optimal arterial opacification and minimization of venous interference. Timing strategies include automatic triggering using specialized software or using a fraction of the bolus administered during a cine acquisition to obtain an enhancement profile. Alternatively, at least for the circle of Willis, a standard delay of about 25 seconds nearly always gives good arterial opacification, with adjustments made for patients with poor cardiac output. A CT venogram (CTV) can be obtained by simply adding several seconds (about 6–8 seconds) to the delay. Protocols vary for different vendors and for different detector configurations and should be optimized for each site. There are many options for postprocessing (4), with the most commonly used techniques including volume rendering (VR), MIP, and oblique/orthogonal or curved reformatting. MIP, a ray-tracing algorithm, where the brightest pixel along a ray passed through the volume is displayed in a projection image, is

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

89

Figure 2 Stent evaluation. (A) Source and (B) curved reformatted CTA images show restenosis in a SMART1 stent (nitinol; Cordis, New Jersey, U.S.), poorly demonstrated on (C) contrast-enhanced MRA (open arrow). (D) CTA and (E) time-of-flight MRA show patent WALLSTENT1 (stainless steel; Boston Scientific, Massachusetts, U.S.). Ends of stent are better shown with CTA (arrows). (F) Neuroform1 stent (nitinol, Boston Scientific, Massachusetts, U.S.) on CTA (arrowhead).

probably the most useful. VR is generated by assigning colors and opacities to ranges of attenuation so that vessels appear distinct from bone and soft tissues. It is most helpful for intracranial applications in which 3D visualization is needed, especially aneurysms. VR is less helpful in the neck, where overlapping structures such as veins and the spine make postprocessing more difficult, but it can be helpful in certain situations such as visualizing the relationship of a high carotid bifurcation relative to the mandible prior to endarterectomy. Shaded surface display (SSD) methods are of limited utility (5), often underestimating degree of stenosis and now superseded by other rendering techniques. Other useful postprocessing techniques include automated vessel analysis techniques for calculation of stenosis severity and for vessel extraction, software for separating arteries and veins, and ‘‘fly-through’’ techniques. Subtraction or masking algorithms for bone and metal exist for CTA just as they do for DSA (6,7), but are not yet widely used. Overlap between attenuation of contrast in a vessel and adjacent bone or calcium is often present, limiting threshold-based segmentation approaches. Associated artifacts such as beam hardening and streaking can limit diagnostic accuracy with or without subtraction, an obvious problem with routinely used aneurysm clips, metallic coils in aneurysms or vessels, and stents. In general, CTA is preferable for the evaluation

of stents (Fig. 2), while coils are better evaluated using MRI techniques (Fig. 3). CTA will often be more successful than MRA in proximity to clips (Fig. 4) (8).

CT Perfusion There are essentially two imaging approaches for the measurement of cerebral blood flow (CBF) in clinical practice: (1) intravascular or nondiffusible tracer (bolus contrast) techniques and (2) diffusible tracer (the tracer can diffuse out of the vessels and into surrounding tissue) techniques. Most routine clinical CT and MR perfusion-weighted imaging studies use an intravascular contrast agent, rapidly injecting a bolus and analyzing the first pass. Diffusible tracer methods include stable xenon CT (XeCT) perfusion, H215O PET, and arterial spin-labeled perfusion MRI. Excellent discussions of different perfusion methodologies are available in articles by Wintermark et al. (9) and Latchaw et al. (10). Bolus Contrast CT Perfusion

Bolus contrast CT perfusion (CTP) is based on the linear relationship of attenuation to concentration of iodine in the brain. It is performed by scanning in cine mode at between one and four imaging locations, repeatedly imaging these locations over 40 to 50 seconds at a rate

90

Wolf

Figure 3 Intracranial embolization coils. Scout topogram and axial image from NECT (A, B) demonstrate large coil mass in basilar tip aneurysm with extensive artifact (B) limiting CTA. Conventional (C, E) and contrast-enhanced MRA (D, F) show aneurysm remnant (arrowheads) on source (C, D) and MIP (E, F) images with minimal artifact from coils (arrows). Thrombus was seen in coiled aneurysm inferiorly (E, block arrow). Abbreviations: NECT, nonenhanced CT; MIP, maximum intensity projection.

Figure 4 Aneurysm clips. (A, C) Axial source and (B, D) MIP images show that CTA (top) shows the A1 segment and ACoA complex near the clip (open arrow, B). On MRA (top), susceptibility artifact leads to extensive signal loss (arrows, C and D). Abbreviations: MIP, maximum intensity projection; ACoA, anterior communicating artery.

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

of about one image set every 1 to 2 seconds before, during, and after a bolus of iodinated contrast. A volume of 40 to 50 cc is infected at 5 to 8 cc/sec through a large-bore IV. Analysis of the time series of CT images results in a time-attenuation curve (TAC). Motion can significantly degrade the perfusion analysis, and although software is available that allows realignment of motion-degraded data to an extent, care should be taken to prepare the patient appropriately to minimize motion, including sedation if necessary. Hemodynamic parameters that are typically generated include measures of the time-to-peak (TTP, time from arrival of bolus in intracerebral arteries to peak concentration in tissue, units of seconds), cerebral blood volume (CBV, integral under the TAC normalized to intravascular attenuation in a large vessel such as sagittal sinus, units of cc/100 g), mean transit time (MTT, average time for contrast to pass from arterial to venous side, units of seconds), and cerebral blood flow (CBF, blood flow in volume of tissue, units of cc/100 g/min). For parameters TTP and CBV, calculations are straightforward and relatively easy to obtain. For MTT or CBF, calculations require more sophisticated analysis, which includes measuring and incorporating the arterial input

91

function (deconvolution techniques) or evaluating the shape of the TAC (nondeconvolution techniques). Deconvolution techniques correct for the imperfect bolus (in theory it should be an instantaneous bolus, but in reality it is spread out), deconvolving or removing the effect of the imperfect arterial input function from the TAC to obtain the residue function, from which the CBF and MTT can be generated (Fig. 5). The central volume principle describes the relationship between parameters as CBF = CBV/MTT (11,12). Nondeconvolution methods use the slope of the tissue TAC to measure the change in concentration of iodine over time, which is proportional to CBF and to the difference between iodine concentration in artery and vein (12,13). A high injection rate of 6 to 8 cc/sec is required, while deconvolution methods tolerate lower rates of injection on the order of 5 cc/sec. Absolute CBF quantitation is possible with deconvolution methods, but there are difficulties in assuring accuracy; for example, large vessels in the analyzed volume can lead to overestimation of CBF. Nondeconvolution methods tend to underestimate CBF. In practice, relative values for perfusion parameters are often used for interpretation, using normal-appearing and/or contralateral brain as an internal reference.

Figure 5 Bolus contrast CTP. NECT (A) shows SAH in left basal cisterns and minimal hydrocephalus. Regions of interest (ROI) for artery (arrowhead) and vein (arrow) are chosen (B) to generate time attenuation curves (C), from which parametric maps are generated such as (D) CBF, (E) CBV, and (F) MTT. Abbreviations: CTP, CT perfusion; NECT, nonenhanced CT; SAH, subarachnoid hemorrhage; CBF, cerebral blood flow; CBV, cerebral blood volume; MTT, mean transit time.

92

Wolf

Another approach, which does not rely on deconvolution or analysis of the bolus itself, generates maps of percent perfused blood volume (PBV) on the basis of a subtraction of registered unenhanced baseline CT images from CTA source images (14), with the change in parenchymal attenuation linearly proportional to the tissue concentration of iodine.

Other causes include reperfusion injury, amyloid angiopathy, coagulopathy, drug abuse, and intracranial neoplasms. Less common are entities such as venous hypertension or occlusion, eclampsia, vasculitis, and infection (17). This section focuses on common clinical entities that are most relevant to interventional neuroradiology: aneurysm, AVM, and venous hypertension/occlusion.

Stable Xenon Perfusion CT

Stable xenon can be used to measure absolute CBF. Xenon is lipid soluble and thus diffusible and leads to changes in attenuation, which can be measured on CT images. XeCT perfusion has been applied in several clinical settings such as cerebrovascular disorders, traumatic brain injury, balloon test occlusion, and subarachnoid hemorrhage (SAH) and vasospasm. However, it is not currently FDA approved, primarily because of reported adverse events such as apneic episodes or increased intracranial pressure (15). These events tend to be transient and the technique well tolerated, particularly with the lower concentration of inspired xenon (28%) that is now used (previously it was 33%). The study is performed by obtaining two baseline scans at two to eight imaging locations without xenon, followed by six additional scans at these locations during xenon inhalation. To determine the change in attenuation from xenon, the baseline scans are averaged for each location and subtracted from that location for each of the subsequent time points. Xenon is delivered mixed with oxygen at a concentration of 28%. End-tidal xenon concentration is measured, and end-tidal carbon dioxide is monitored, as well as any apneic episodes. Calculation of CBF is based on the Fick principle; that is, the amount of an indicator in a sample of tissue is proportional to the difference between the amount supplied in arterial blood and the amount carried away in venous blood. Modified Kety-Schmidt equations are used to describe the relationship of xenon concentration in the brain and in the arteries with the blood-brain partition coefficient and CBF (16). The xenon concentration in the brain is obtained from the CT measurements at baseline and during inhalation of xenon, and the timedependent arterial concentration is obtained by measuring the end-tidal xenon concentration, which is proportional to the time-dependent arterial concentration in patients without severe lung disease leading to significant dead space. The total time of acquisition is on the order of 5 to 6 minutes, and studies take about 20 minutes from start to finish, including data processing and creation of CBF maps. Repeat studies can be obtained 20 minutes after the end of a previous scan, allowing for washout of xenon.

CLINICAL APPLICATIONS Nontraumatic Hemorrhage Common causes of nontraumatic intracranial hemorrhage include ruptured aneurysm, arteriovenous malformation (AVM), hypertension, and prematurity.

Aneurysm

An aneurysm is essentially a circumscribed dilatation of an artery. There are different types and/or causes of aneurysms, but the most common is the berry, or saccular, aneurysm. Other types of aneurysms include mycotic, fusiform, dissecting, traumatic, and pseudoaneurysms. Aneurysms can be associated with abnormal vasculature in neoplasms. Venous aneurysms also occur, and both arterial and venous aneurysms can be seen with AVMs. The most common locations of berry aneurysms (ruptured or unruptured) are proximal in the circle of Willis. About 85% to 95% involve the anterior circulation, and 5% to 15% involve the posterior circulation. The most likely locations are anterior communicating artery (ACoA, 30%), internal carotid artery (ICA) [including the periophthalmic and posterior communicating artery (PCoA), 25%], middle cerebral artery (MCA, 20%), basilar artery (BA, 10%), and posterior inferior cerebellar artery (PICA, 5%). There are multiple aneurysms in about 20% of aneurysm patients (18). CT techniques play a prominent role in the evaluation of unruptured or ruptured aneurysms. For unruptured aneurysms, MRA is more commonly used as a screening modality, whereas CTA is more often used to verify suspicion of an aneurysm (e.g., one suggested but not certain on MRA), to better characterize an aneurysm detected on MRA or DSA (e.g., giant or cavernous aneurysms), and when patients cannot undergo MRI or MRA. Multiple studies have compared CTA with MRA, DSA, and/or rotational angiography. Its sensitivity for detection of aneurysms is at least as good as that of MRA but, as with MRA, drops off below 3 mm (19). CTA is also used to follow patients with known aneurysms. Although aneurysm clips currently placed are largely MR compatible, they create substantial artifact, which often renders MRA useless. Clips can also limit CTA, but diagnostic information is still often obtainable, even for previously clipped aneurysms. Previously coiled aneurysms are better evaluated with MRA. Subarachnoid hemorrhage and saccular aneurysmal rupture. The most common cause of nontraumatic SAH is ruptured aneurysm (75–85% of cases) (20), with significant associated morbidity and mortality. Other less common causes include perimesencephalic hemorrhage, AVM or arteriovenous fistula (AVF, in brain or spine, and possibly with associated aneurysm), intracranial dissection, drugs such as amphetamines or cocaine, coagulation disorders, vasculopathies such as sickle cell disease and moyamoya, and others (21,22). NECT is almost always the first imaging study obtained, in which blood appears

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

93

Figure 6 Aneurysm rupture and CTA. (A) NECT at presentation shows SAH in the sylvian fissure (white arrowhead) and a parenchymal hematoma (black arrowheads). This pattern is suspicious for MCA aneurysm rupture, confirmed on (B) MIP images from CTA (white arrow). A second unruptured MCA aneurysm is present (open arrow). Abbreviations: NECT, nonenhanced CT; SAH, subarachnoid hemorrhage; MIP, maximum intensity projection.

hyperdense. NECT allows rapid evaluation for the presence of SAH as well as immediate complications such as hydrocephalus, is available around the clock, and provides easy access to unstable patients. Acute blood appears dense on CT, depending on hematocrit and hemoglobin values (56 HU, with hematocrit of 45% compared with gray matter attenuation of just under 40 HU or CSF at around 0–5 HU) (23). Coagulopathies can lead to difficulty in visualizing acute blood; for example, a low hemoglobin value of less than 10 g/dL can be invisible. The sensitivity for NECT in detecting SAH is approximately 95% in the first one to two days, but decreases over time to 50% after one week and to almost 0% after three weeks. Negative NECT should be followed by lumbar puncture to increase sensitivity for SAH detection, assessing for blood and/or xanthochromia (depending on time after initial bleed). The pattern of SAH on NECT may suggest the most likely location of the ruptured aneurysm, at least for ACoA and MCA aneurysms (24). The presence of parenchymal hematoma with SAH increases accuracy (24,25). When multiple aneurysms are present, the pattern of hemorrhage and especially the

location of parenchymal hematoma may provide useful information in deciding which aneurysm is likely to have bled, complementary to other indicators such as aneurysm size and morphology (Fig. 6). However, the amount and distribution of blood are very often not predictive of the site of aneurysm rupture (25,26). Patterns associated with rupture of an ACoA aneurysm include symmetric SAH, blood in the anterior interhemispheric fissure, anterior pericallosal cistern and/or cisterna lamina terminalis, anterior interhemispheric clot, or inferior frontal parenchymal hematoma (Fig. 7). Intraventricular hemorrhage may also be associated secondary to rupture through the lamina terminalis. Other ICA aneurysms, including PCoA aneurysms, are more difficult to localize, often without lateralizing signs on CT. MCA aneurysms may demonstrate asymmetric density in the sylvian fissure. More specific localizing presentations include parenchymal hematoma or expansile clot in the sylvian fissure (Fig. 8). PICA aneurysms may show disproportionate blood in the posterior fossa and fourth ventricle, and basilar tip aneurysms might show SAH primarily in the interpeduncular and

Figure 7 Ruptured ACoA aneurysm pattern. (A) Axial NECT image shows nearly symmetric SAH and a small interhemispheric or midline parenchymal hematoma (white arrow). (B) DSA confirms ACoA aneurysm (black arrow), suspected for rupture. Small MCA and ICA aneurysms were also detected (black arrowheads). Abbreviations: ACoA, anterior communicating artery; NECT, nonenhanced CT; SAH, subarachnoid hemorrhage; DSA, digital subtraction angiography; ICA, internal carotid artery.

94

Wolf

Figure 8 Ruptured MCA aneurysm. (A, B) Axial NECT images at two locations show asymmetric SAH and focal hematoma, expanding left sylvian fissure (see also Fig. 9). Abbreviations: NECT, nonenhanced CT; SAH, subarachnoid hemorrhage.

Figure 9 Ruptured posterior circulation aneurysms. (A) NECT shows asymmetric blood in right cerebellopontine angle, suggesting right PICA aneurysm. (B) NECT from a different patient with ruptured basilar tip aneurysm shows focal blood near basilar tip. Abbreviations: NECT, nonenhanced CT; PICA, posterior inferior cerebellar artery.

prepontine cisterns (Fig. 9). Van der Jagt et al. (25) reported that validity of SAH distribution on CT was ‘‘inconsistent or low’’ for ruptured aneurysm arising from MCA, ICA, or posterior circulation aneurysms, unless a parenchymal hematoma was in proximity. Blood distribution was a better predictor for anterior cerebral artery (ACA) and ACoA aneurysms. An atypical presentation of a ruptured aneurysm is subdural hematoma (SDH), often with some SAH but rarely without any evidence of SAH. It has been described in ICA and ACoA aneurysms (27), also with pericallosal aneurysms (Fig. 10). Pericallosal aneurysms might also show parenchymal hematoma or large focal SAH above the corpus callosum. While the most common causes of SAH are trauma followed by intracranial aneurysm, this distinction cannot always be made clinically. Examples include unwitnessed falls, patients ‘‘found down’’ without overt evidence of trauma, or motor vehicle collision or a fall where a ruptured aneurysm may

have preceded the traumatic event. Findings on NECT favoring trauma are associated calvarial or skull base fractures, SDH, contusions, and a relative lack of blood in basal cisterns. On follow-up imaging, confidence is increased when evolving contusions or foci of diffuse axonal injury are clearly demonstrated. Findings favoring aneurysm include SAH in basal cisterns, excessive amount of SAH, and lack of obvious traumatic findings. Occasionally, the aneurysm itself can be clearly visualized on NECT (Fig. 11). In some cases, it is impossible on cross-sectional imaging to accurately assess whether the source of SAH is aneurysmal or traumatic, and conventional angiography may be necessary. Conventional angiography or DSA is still considered the gold standard for detection of aneurysms. However, CTA may be preferred in some instances, e.g., catastrophic SAH (28) (Fig. 6). In a systematic review of noninvasive imaging studies for aneurysm, White et al. (29) found that CTA was not as sensitive as

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

95

Figure 10 Atypical pattern of hemorrhage. (A–C) Axial NECT images at three locations show subdural hemorrhage (arrows, A and C), and relatively little SAH (B, arrowhead ). (D) Volume-rendered image from CTA demonstrates a pericallosal artery aneurysm. Abbreviations: NECT, nonenhanced CT; SAH, subarachnoid hemorrhage.

Figure 11 Demonstration of aneurysm on NECT. Axial NECT images from patients with (A) ruptured and (B) unruptured ICA aneurysms. In each case, the aneurysm is clearly seen without IV contrast (arrows). Hydrocephalus is also noted in (A). Abbreviations: NECT, nonenhanced CT; ICA, internal carotid artery.

96

Wolf

Figure 12 Giant aneurysm. (A) Sagittal T1-weighted MR image shows flow void/jet (arrowhead ) and thrombus (arrow ) in partially thrombosed giant aneurysm. (B) MIP image from MRA fails to distinguish thrombus from patent aneurysm. (C) CTA MIP image distinguishes patent (black asterix) from thrombosed (white plus sign) aneurysm. Abbreviation: MIP, maximum intensity projection.

DSA for small aneurysms, with an accuracy of 96% for aneurysms larger than 3 mm but only 61% for aneurysms 3 mm or less (overall accuracy of 89%). However, studies only up to 1998 were reviewed, and higherquality multidetector scanners have become more widely available from around 1998. More recently, Chappell et al. (30) looked at 21 studies in a metaanalysis comparing CTA and DSA (the most recent in 2002). The overall sensitivity and specificity weighted for number of patients per study were 92.7% and 77.2%, respectively. The specificity will likely be inaccurate, since the rate of true-negative cases is difficult to assess in most of these studies, most of which focus on cases in which an aneurysm was suspected clinically or radiographically (30). There is increasing evidence that missing a symptomatic aneurysm on CTA would be quite rare (19,31–33). Indeed, there are reported cases in which aneurysms are detected on CTA and not DSA, and at minimum, CTA is a valuable adjunct study (34,35). For example, the 3D anatomy of complicated or giant aneurysms, including those with significant intraluminal thrombus, aneurysm relationship to bony structures (e.g., paraclinoid aneurysms), and calcifications that might interfere with clipping, may be demonstrated more clearly (Fig. 12). Although there are still questions regarding sensitivity for detection of very small aneurysms (<3 mm) and false positives are also of concern (19), some centers currently use CTA for only routine aneurysmal SAH workup and have found it safe and reliable (36,37). SAH from aneurysmal rupture in the subacute and chronic setting. While MRI is more sensitive and specific for evaluation of ischemia, it is not always a viable option in a sick ICU patient. CT techniques are preferred for following SAH patients after presentation and treatment, evaluating for complications such as rebleeding, hydrocephalus, vasospasm, and ischemia or infarct. CTA with or without perfusion may be helpful in vasospasm cases (Fig. 13) (38). The

addition of CTP (9,39) or XeCT perfusion studies (9,16) may also be helpful. Effectiveness of treatment can be tested by obtaining perfusion data before and after changes in therapy. Other methodologies have also been applied to this problem, including PET, SPECT, and transcranial Doppler (TCD) (39), but none have been established as a definitive test (10). Imaging follow-up for the original and any additional aneurysms depends on prior intervention. Immediate (or intraoperative) conventional angiography is often performed to verify clip placement, but CTA can be used for follow-up in the subacute or chronic setting (40,41). There are some limitations related to artifacts from the clips (42), but at least some artifacts can be minimized with technique (Fig. 14). Coils most often cannot be evaluated with CTA (Fig. 3). Nonaneurysmal subarachnoid hemorrhage. Perimesencephalic hemorrhage is the cause of SAH in about 10% of cases and accounts for about 70% of SAH cases that have normal DSA (20). The typical pattern is blood localized in cisterns around the midbrain, without extension into sylvian fissures, interhemispheric fissure, or parenchyma. Intraventricular hemorrhage is also not typically present. Since aneurysms may be missed initially because of spasm, compression by hematoma, or perhaps suboptimal number or choice of views, DSA-negative SAH should be followed in about one week with another exam. CTA may provide the means to forgo follow-up (or initial) DSA. Other causes of nonaneurysmal SAH (about 5% of cases) include intracranial dissection, AVM/AVF, coagulopathy, drug use (e.g., amphetamines or cocaine), and vasculopathies such as sickle cell disease and moyamoya (21,22). AVM

CTA has been applied to diagnostic evaluation of AVM, but cannot replace DSA at this time (43,44).

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

97

Figure 13 CTA and vasospasm. (A) Axial source image from CTA acquisition demonstrates irregular narrowing of left M1 segment of MCA, not present on baseline DSA and suspicious for vasospasm. (B) DSA confirms M1 vasospasm, and also shows bilateral A1 and A2 segment as well as ICA bifurcation vasospasm. Abbreviations: DSA, digital subtraction angiography; ICA, internal carotid artery.

Figure 14 Aneurysm follow-up after clipping. (A) Source image from CTA demonstrates aneurysm remnant (arrowhead ) despite artifact from clip (block arrows). Slab MIP (B) and VR images (C) demonstrate topography of remnant (arrowheads). Abbreviation: MIP, maximum intensity projection.

CTA can be used as a complementary examination, primarily for depiction of 3D morphology and for stereotactic planning, but rotational angiography can also provide this information. AVMs and AVFs can be detected using CTA, and it can be used for following lesions after treatment; however, MRI and MRA probably have the edge in this regard because of superior evaluation of parenchyma and improving temporal resolution of MR DSA methods. In the setting of acute intracranial hemorrhage, nearly all patients will initially undergo NECT, and in some cases a presumptive diagnosis can be made on NECT even without hemorrhage (Fig. 15). Cavernomas and venous angio-

mas are commonly encountered and must be recognized to avoid unnecessary testing. CTA provides a fast evaluation of intracranial vasculature for emergent intervention (Fig. 16), but the temporal resolution of CTA is not adequate for completely evaluating AV shunting or delineating arterial or venous aneurysms, arterial feeders, and draining veins. Venous Occlusive Disorders

Venous occlusive disease may be suspected on initial NECT by demonstration of a high-attenuation clot in one or more venous sinuses or cerebral veins. As

98

Wolf

Figure 15 Vascular malformations on NECT. (A) Cavernous malformation (open arrow). (B) Developmental venous anomaly (white arrow). (C) AVM (arrowheads). Abbreviations: NECT, nonenhanced CT; AVM, arteriovenous malformation.

Figure 16 AVM with catastrophic presentation. (A) NECT showed fourth ventricular hemorrhage (block arrow), parenchymal hemorrhage (open arrow), SAH, and hydrocephalus. CTA obtained en route to surgery (B) showed left cerebellar AVM with focal aneurysms (MIP, black arrow, arrowhead ). On postoperative DSA, early (C) and delayed (D) phases from selective superior cerebellar artery injection confirmed findings. Abbreviations: AVM, arteriovenous malformation; NECT, nonenhanced CT; SAH, subarachnoid hemorrhage; MIP, maximum intensity projection; DSA, digital subtraction angiography.

opposed to arterial ischemia, venous ischemia or infarction more often presents with hemorrhage or with patterns of edema atypical for arterial ischemia. Examples include bilateral thalamic hypodensities with deep venous occlusive disease (although this event

could be mimicked by top of the basilar syndrome) and posterior temporal lobe hemorrhage suggestive of vein of Labbe´ or transverse sinus thrombosis. Highresolution imaging of the cerebral venous system can be obtained with CT venography (Fig. 17) (45).

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

99

Figure 17 Transverse sinus thrombosis. (A, B) NECT shows increased density in right transverse sinus indicating thrombus. (C) Posterior oblique MIP projection confirmed absent contrast opacification of right transverse sinus (arrowhead) on CTV. Abbreviations: NECT, nonenhanced CT; MIP, maximum intensity projection; CTV, CT venogram.

Other Causes of Intracranial Hemorrhage

There are multiple other causes of nontraumatic intracranial hemorrhage, including hypertensive hemorrhage, amyloid angiopathy, coagulopathy, drug abuse, and intracranial neoplasms. Additional less common causes are reperfusion injury, eclampsia, vasculitis, and infection. Some of these will be discussed in the next section, since they may also present as strokes or ‘‘stroke-like’’ syndromes, but a complete discussion is beyond the scope of this chapter.

Ischemia Stroke represents the third-largest cause of mortality and a leading cause of morbidity in the United States. The vast majority of ‘‘strokes’’ are ischemic in origin (80–85%), and the minority are hemorrhagic events. Most ischemic infarcts are thromboembolic. Prognosis, risk of recurrence, and management options are influenced by infarct subtype (Trial of ORG 10172 in Acute Stroke Treatment or TOAST criteria) (46). Subtypes include cardioembolic, large vessel, small vessel, other (determined) causes, and cryptogenic. Hyperacute and Acute Setting

NECT Acute imaging addresses the following questions: (1) Is there hemorrhage or other explanation for symptoms? (2) What is the etiology of the infarct and status of the vessel involved, if any? (3) What is the location and extent, and is there tissue still at risk? (47). The first study performed is usually NECT. In the acute setting (first 3 to 6 hours for anterior circulation and longer for posterior circulation), this study is often the only imaging test necessary for the stroke neurologist to decide to treat with tissue plasminogen activator (tPA). MRI with diffusion-weighted imaging (DWI) is more sensitive and specific for detection of acute ischemia. MRI is also sensitive to hemorrhage (47–49), though most practitioners still prefer CT.

CT may show findings of ischemia within the first 6 hours, often subtle initially but becoming obvious within 12 to 24 hours. Early signs of infarct include the dense artery or dot sign, loss of the insular ribbon, blurring of basal ganglia, sulcal effacement, and loss of differentiation between gray and white matter at the cortical margin (Fig. 18). The pattern may provide clues to etiology. Infarcts corresponding to one or more arterial territories with gray matter involvement are more likely embolic, while infarcts falling between vascular territories (borderzone or watershed distributions) tend to reflect a more proximal lesion such as large vessel stenosis or occlusion. Overlap in pathophysiologies and patterns occurs; for example, a high-grade carotid stenosis or occlusion may appear identical to an ICA terminus embolus, and large-vessel stenosis might present with associated in situ thrombosis. Small-vessel ischemic events will most often be invisible early, appearing in subacute and chronic stages as lacunar infarcts and subcortical white matter lesions. Global injuries such as those seen with cardiopulmonary arrests may be ischemic, hypoxic, or anoxic. Patterns include deep gray matter injury, cortical laminar necrosis, diffuse white matter injury, or a combination. Cardiopulmonary arrest may precipitate borderzone injuries when a preexisting large-vessel stenosis or occlusion is present. Hypoxic-ischemic injury or perinatal asphyxic injury in the newborn and premature injuries such as PVL are also examples, but beyond the scope of this chapter (50). CTA Noninvasive imaging methods such as CTA play an important role in the workup of cerebral ischemia, in acute as well as subacute or chronic settings. Multiple studies have shown good agreement with conventional DSA (ranging from 86% to 100%) and other imaging modalities (Fig. 18) (51–55). Followup studies (DSA, MRA, and brain imaging) confirm CTA findings in approximately 80% of cases (56). Potential for the greatest benefit from thrombolysis has been demonstrated in a subgroup of patients

100

Wolf

Figure 18 Acute ischemic infarct. (A, B) NECT images show dense MCA sign (A, arrow) and blurred gray-white matter borders (A and B, arrowheads). (C, D) CTA shows relative decrease in vascularity on source image (C, arrowheads) and occlusion versus high grade stenosis (D, arrow) with distal filling MCA branches. (E, F) Follow-up NECT more clearly shows infarct, as well as hemorrhagic transformation (F, block arrow). Abbreviation: NECT, nonenhanced CT.

with moderate or severe persisting deficit for less than three to six hours, MCA occlusion (as opposed to other sites such as ICA terminus), lack of extended infarct signs, and efficient collateral circulation (55). Also important is the ability to detect autolyzed thrombi and spontaneous recanalization. Perfusion imaging and comprehensive stroke evaluation. Perfusion changes are immediate in the setting of acute ischemia (i.e., evident prior to parenchymal abnormalities, including those identified on DWI), and information is also provided regarding ‘‘tissue at risk.’’ Some advocate a comprehensive exam consisting of unenhanced CT, CTA, and CTP (57,58) (Figs. 18 and 19), attractive because of more widespread availability of CT, better access to sick patients, patient tolerance, and speed. Others advocate a comprehensive stroke MR protocol. In practice, combinations are often used; for example, NECT of the brain and CTA of the head and the neck can be performed, followed by MRI with DWI, perfusion-weighted imaging (PWI), and gradient echo (susceptibility- weighted) imaging sequences, especially for subacute and chronic ischemia workup (47). Most of the diagnostic imaging information can thus be rapidly obtained in the hyperacute or acute

setting, with DWI improving detection and delineation of the extent of infarcted tissue and susceptibilityweighted imaging providing a sensitive evaluation for subtle hemorrhage. The goal for intervention in the acute setting is to reestablish blood flow without causing harm. Since use of tPA currently relies on clinical history and early presentation of the patient (i.e., within 3 hours for intravenous tPA and within 6 hours for intra-arterial tPA for anterior circulation, longer for posterior circulation), it is hoped that techniques such as perfusion imaging will help establish a new starting point for this time frame. If extent of existing infarct and perhaps age can be established and perfusion imaging can help assess for tissue at risk, the window of opportunity for treatment can be extended. Normal brain perfusion suggests that thrombolysis or other methods for augmenting CBF are not immediately necessary. Patients with ‘‘penumbral’’ tissue may benefit from thrombolysis, but reperfusion can lead to life-threatening hemorrhage in severely ischemic or infarcted tissue. CTP and (58–61) XeCT perfusion (62,63) have been applied in this setting. The Council on Cardiovascular Radiology of the American Heart

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

101

Figure 19 Acute ischemic infarct and CTP (same patient in Fig. 18). One time point (A) from singlesection CTP is shown. Very low CBF (B) and CBV (C) with prolonged MTT (D) confirm right temporal lobe infarct. Less severe decreased CBF, symmetric CBV, and prolonged MTT in right occipital lobe (arrows) indicate minimal penumbra. Abbreviations: CTP, CT perfusion; CBF, cerebral blood flow; CBV, cerebral blood volume; MTT, mean transit time.

Association recently published guidelines for perfusion imaging in cerebral ischemia (10). There is evidence of potential benefit and predictive value of perfusion methodologies such as XeCT, CTP, and slow-infusion PBV, but larger prospective studies are required to fully establish their role. Subacute and Chronic Setting

When thrombolysis is no longer an option, the same questions remain. Workup of subacute and chronic cerebrovascular disease usually involves multiple diagnostic imaging modalities, including combinations of NECT, CTA, CTP, DUS, MRI/MRA, and PWI, DSA, and/or XeCT perfusion. CT techniques are typically better first-line imaging strategies, while MRI/MRA and DUS are better in the subacute and chronic settings. MRI is the most sensitive and specific technique for evaluation of the brain parenchyma. After initial workup of ischemic stroke, NECT can be used for routine follow-up in the subacute setting, assessing for infarct evolution, hemorrhagic transformation, hydrocephalus, cerebral edema, or mass effect. Combinations of DUS, MRA, and/or CTA are typically used for cerebrovascular imaging. In the neck, DUS is often the first or screening exam. Accuracy (e.g., agreement with DSA) of

one noninvasive test can be improved by confirming findings with a second noninvasive test, and by adding a third if the first two are discordant (3,64,65). Cerebrovascular CTA Atherosclerosis and stenotic-occlusive disease CTA has been shown to be useful in evaluating carotid stenosis (64–67). In general, CT tends to underestimate the degree of stenosis compared with DSA, while MRA tends to overestimate. However, for severe carotid stenosis (70–99%, NASCET criteria), CTA has been shown to be fairly accurate. In a meta-analysis of CTA studies prior to 1998 (almost all single-detector CT acquisitions), Hollingworth et al. (68) reported a pooled sensitivity and specificity of 95% and 98%, respectively. CTA remained sensitive (95%) when stenoses greater than 30% were included, although specificity decreased to 92%. Another systematic review of CTA studies between 1990 and 2003 (69) (all single-detector CT scanners) reported pooled sensitivity and specificity for 70% to 99% stenoses of 85% and 93%, respectively, and for occlusions they reported sensitivity and specificity of 97% and 99%. Accuracy for carotid stenosis (64,66) and for occlusion (70) will likely improve with the increasing use of MDCT. Accuracy in detection of vascular wall pathology such as ulceration is unclear (Fig. 20), but neither

102

Wolf

Figure 20 Ulcerations. (A) DSA shows web-like focal stenosis in proximal ICA (arrow) with probable ulceration(s) more proximally (arrowhead ). (B) MIP image from CTA also shows the focal stenosis fairly accurately, but the ulceration is not as clearly depicted. However, ulceration can be seen on CTA source image (C) (arrowheads). Abbreviations: DSA, digital subtraction angiography; ICA, internal carotid artery; MIP, maximum intensity projection.

is DSA considered perfectly accurate in detection of ulcerated plaques. For intracranial occlusive disease, CTA performance is in general similar to that of MRA, except for demonstration of very slow flow and collaterals, where CTA is superior (Fig. 21) (55,71). In comparison with MRA and DSA in a retrospective study, Bash et al. (72) found a higher sensitivity of CTA for intracranial stenosis, higher positive predictive value for stenosis and occlusion, and higher interobserver reliability. They also found cases in which a false-positive occlusion on DSA may have been present that appeared stenotic but patent on CTA. Combining noninvasive imaging modalities likely increases confidence (73). False-positive occlusions in the cervical carotid also occur (66). It is often not possible to be certain of

this event, since time tends to separate the studies and interval occlusion or recanalization could occur between studies. DSA is considered the gold standard, but it is a projectional technique, and limited projections are obtained. CTA has essentially infinite projections in any orientation and thus is more likely to find the projection with the most narrowed lumen. Rotational angiography would be a better comparison test in this regard (66). Very slow flow in a smallcaliber artery may be easier to detect with CTA (72). In the case of hypoplastic or atretic carotid arteries, CTA is complementary and in some ways superior to DSA, with the detection of a hypoplastic petrous carotid canal establishing the diagnosis (Fig. 22). Problem areas for CTA in the head or the neck include overlapping venous structures and vessels in or around

Figure 21 Intracranial stenotic-occlusive disease and CTA. (A) Oblique axial MIP image from CTA shows high-grade left M1 stenosis (arrowhead ) with collateral filling in MCA branches (arrow). (B) Coronal MIP projection from MRA shows the abnormal M1 segment (arrowhead ) but no convincing collaterals because of slow flow. (C) Oblique DSA projection shows the M1 stenosis (black arrowhead) and delayed MCA filling. Abbreviations: MIP, maximum intensity projection; DSA, digital subtraction angiography.

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

103

Figure 22 ICA agenesis and atresia. CTA source images show hypoplastic right petrous carotid canal (A, black arrow), absent intracranial ICA (B, white arrow), and transsellar collateral (C, arrowhead) from left to right anterior circulation. Abbreviation: ICA, internal carotid artery.

bony structures and heavy vascular calcifications (64,66). CT can be used for direct evaluation of atherosclerotic plaque and vessel wall as well as lumen. The extent of calcification and other components relevant to vulnerable plaque such as ulceration and hemorrhage can be evaluated. MRI is likely superior in discriminating plaque architecture, however (74). Arterial dissection Dissections can be categorized as traumatic (discussed below) or atraumatic. Atraumatic dissections may be spontaneous or associated with a precipitating cause such as minor or trivial trauma (e.g., a movement or position not related to an external traumatic force). Genetic factors perhaps combined with environmental factors have been implicated in predisposition to spontaneous dissection (75). Inherited disorders predisposing to dissection include fibromuscular dysplasia, Marfan’s syndrome, EhlerDanlos, homocysteinuria, and others. Hypertension and smoking may also predispose to dissection.

Dissection of the extracranial ICA is the most common form of cerebrovascular dissection (75). Whereas atherosclerotic disease typically involves carotid bifurcation and bulb/ICA origin, spontaneous dissection often originates at least 1 cm beyond the bifurcation and involves more distal ICA. It typically does not extend into the petrous carotid artery, but does occur on occasion, and thrombus can also propagate distally (Fig. 23). Acute ischemic infarct in a younger patient should prompt a search for this entity, occurring more commonly in the age range of 35 to 50 years. Vertebral artery dissections more commonly involve the distal portions, where atherosclerotic disease is often found more proximally. Symptoms and findings of extracranial dissection include neck pain, headache, Horner’s syndrome, cranial nerve defects, pulsatile tinnitus and bruit, and of course transient or permanent ischemia. Compared with extracranial dissection, intracranial dissection is less common, is more

Figure 23 Distal cervical carotid dissection and CTA. (A) Narrowing of cervical ICA lumen (arrow) with dilatation of vessel overall (arrowheads). (B) Intimal flap (open arrow) with true and false lumen apparent. (C) Occlusion or high-grade stenosis ICA, possibly from propagating thrombus (black arrowhead), with distal petrous ICA filling. Abbreviation: ICA, internal carotid artery.

104

Wolf

Figure 24 Intracranial dissection presenting with SAH. (A) NECT shows SAH anterior to pontomesencephalic junction. (B) Oblique coronal MIP image from CTA shows fusiform dilatation of intradural left vertebral artery (arrowhead ), confirmed on AP (C) and lateral (D) DSA projections. The appearance suggests dissecting aneurysm. Abbreviations: NECT, nonenhanced CT; SAH, subarachnoid hemorrhage.

often traumatic, and has higher morbidity and mortality. Patients are even younger than those with extracranial dissection, and the ICA is involved more frequently than VA (76). Intracranial dissection more often presents with sudden early infarcts or SAH, which affects treatment decisions regarding anticoagulation (Fig. 24). Mass effect from pseudoaneurysms will be more problematic in the closed intracranial space. The supraclinoid ICA is the most commonly affected segment, followed by intradural and suboccipital VA (near intradural transition). Dissection represents a disruption of one or more layers of the arterial wall. Intracranial arteries lack a vasa vasorum, so an intimal tear is more likely in intracranial dissections. Dissection of blood between intima and media results in narrowing and potential occlusion of the lumen, whereas collection of blood between media and adventitia may result in expansion of artery diameter or pseudoaneurysm. Combinations of arterial dilation and luminal narrowing may occur. Imaging findings include narrowing of the lumen, especially smooth or slightly irregular tapered narrowing, intimal flaps with or without double lumens, eccentric or crescentic wall hematoma, and pseudoaneurysm formation (Figs. 23 and 24) (77). Intimal flaps are specific for dissection but are seen in a minority of cases. It is often difficult to distinguish intracranial dissection from other pathologies such as atherosclerosis or embolism causing partial or complete occlusion. Vasospasm can also mimic dissection.

Noninvasive evaluation of nontraumatic dissection has been described more extensively with MRI and MRA, but CTA has also been applied (78–81). Perfusion imaging and evaluating long-term risk of ischemic event. The significance of cervicocranial stenotic-occlusive disease is modified by collaterals, type of plaque (i.e., ‘‘vulnerable’’ plaque), autoregulation and cerebrovascular reserve, oxygen extraction fraction, cardiac status, etc. Cerebral perfusion imaging can be used to evaluate the hemodynamic effect of stenosis or occlusion, but interpretation is more complicated in the subacute and chronic setting than in hyperacute and acute settings. Symptoms may be intermittent with embolic and/or perfusional etiologies. Because of the modifying factors noted above, a single baseline perfusion test may not completely address the ‘‘significance’’ of a particular lesion (10,82). Baseline perfusion may be normal or show only prolonged TTP or MTT, but a delay is expected with high-grade stenosis and is difficult to translate to stroke risk. Qualitative analysis (i.e., comparing abnormal to contralateral side for relative measures) may be misleading when both sides are abnormal. Perfusion studies with a challenge can help identify patients who may benefit from revascularization. Hemorrhagic and Stroke-like Conditions

Hemorrhagic stroke or ‘‘stroke-like’’ entities include reperfusion or hyperperfusion syndromes, vasculitis/

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

vasculopathy, and posterior reversible encephalopathy syndromes (PRES). Other causes for hemorrhagic ‘‘stroke’’ include hypertensive hemorrhage, amyloid angiopathy, coagulopathy, drug abuse, and intracranial neoplasms. Venous hypertension and occlusion were discussed briefly above. Reperfusion and hyperperfusion. Attempts to establish reperfusion via medical or catheter-based thrombolysis are aimed at rescuing tissue at risk around an irreversibly injured core, but early reperfusion can lead to edema and hemorrhage, as well as neuronal injury in the penumbra (10,83). Potential indicators for hemorrhagic transformation include extent of parenchymal hypoattenuation on baseline CT, older age, and administration of aspirin prior to thrombolysis (84). Other indicators might include those based on perfusion studies. One retrospective study using XeCT perfusion reported that CBF values below 10 and perhaps even less than 15 cc/100 g/min in aggressively managed acute MCA infarct patients could be associated with increased risk of hemorrhage, edema, and herniation with or without reperfusion (85). Hyperperfusion syndrome occurs when brain tissue in a vascular territory experiencing low cerebral perfusion pressure (CPP) due to a flow-limiting vascular lesion is suddenly subjected to a normal CPP after revascularization (86). Autoregulation is impaired or overwhelmed, leading to development of symptoms such as headache, seizures, and hypertension minutes to hours after the procedure that in some cases appear

105

identical to those of infarct. Imaging may demonstrate breakdown of the blood-brain barrier with or without hemorrhage. CT, CTA, and perhaps perfusion studies could be of use in this setting (86). Vasculitis and vasculopathy. Vasculitides and vasculopathies encountered in the head and the neck include fibromuscular dysplasia, giant cell arteritis, Takayasu arteritis, granulomatous angiitis of the CNS, SLE, moyamoya disease, sickle cell disease, infectious diseases such as syphilis and herpes virus, and many others, including PRES such as eclampsia and hypertensive encephalopathy. Imaging findings include segmental narrowing or beaded appearance, multivessel or repeated dissections or pseudoaneurysms (nontraumatic or minor trauma), occlusions, and moyamoya pattern (Fig. 25). In some cases (e.g., PRES), the diagnosis can be made more effectively on parenchymal imaging of the brain, since the vascular findings on imaging may be nonspecific, subtle, or absent. In addition to the ability to image the brain, CT and MRI allow direct visualization of the vessel wall, which may be helpful in making the diagnosis of vasculitis (Fig. 26).

Trauma The mechanisms of extracranial and intracranial traumatic neurovascular injury are primarily penetrating and blunt, but all involve some form of disruption of

Figure 25 Vasculitis/vasculopathy. CTA MIP images demonstrating (A) neurosyphilis, with nonspecific segmental arterial narrowing; (B) fibromuscular dysplasia, with beaded appearance in both ICAs; and (C, D) AIDS vasculopathy, with fusiform arterial dilations. Abbreviations: MIP, maximum intensity projection; ICA, internal carotid artery.

106

Wolf

Figure 26 Cross-sectional imaging in giant cell arteritis. CTA source images show circumferential smooth thickening from great vessel origins through (arrows) common carotid arteries (A), not seen distal to carotid bifurcation in the ICA (B, arrowheads).

the vessel wall. This disruption can be subtle, as with slight separation of intima and media with small intramural hematoma. More severe injuries are intimal disruption with formation of flap and false lumen, pseudoaneurysm, occlusion or transection, and AVF. Morbidity and mortality increase with severity, primarily due to secondary CNS injury and typically due to thromboembolic disease, drop in CPP, and/or exsanguination. NECT of the head is often the first exam for evaluating intracranial injury. NECT of the face and the neck is also used in trauma, but usually for detection of craniofacial and spinal fractures as opposed to soft tissue evaluation. NECT is suboptimal for direct visualization of vascular injury, relying on indirect signs such as fractures predisposing to vascular injury, trajectory of penetrating injury,

proximity of penetrating injuries and/or bullet fragments, hematoma, or soft tissue swelling (Fig. 27). The gold standard for neurovascular injury is still conventional DSA. It is also considered the definitive study and can be combined with endovascular intervention. There are some advantages to using noninvasive imaging; for example, CTA is fast, provides information regarding nonvascular structures, and directly images vessel wall and lumen. CTA is a 3D technique, so unlimited projections are available as opposed to limited 2D projections obtained with conventional DSA. A disadvantage of CTA is low temporal resolution, making evaluation for AVF inadequate. Artifacts from bone, heavy calcifications, and metallic objects such as bullets can limit accuracy and render portions of the CTA nondiagnostic. Poor arterial

Figure 27 Indirect evidence of penetrating arterial injury on NECT. (A) Bullet trajectory passed through mandible and transverse process of C1 on right. Fracture includes transverse foramen (arrow), indicating potential vertebral artery injury. DSA confirmed occlusion. (B) Bullet and fragment trajectories (arrow) are concerning for MCA injury, in this case directly confirmed by demonstration of right MCA infarct (arrowheads). Abbreviations: NECT, nonenhanced CT; DSA, digital subtraction angiography.

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

107

Figure 28 Blunt carotid vertebral injury. CTA source images show abnormal contour of bilateral ICAs (A and B, arrowheads). (C) Left ICA caliber change on MIP image suggests dissection/pseudoaneurysm, improved on 12-month follow-up (D). Abbreviations: MIP, maximum intensity projection; ICA, internal carotid artery.

contrast opacification can lead to uncertainty, and venous opacification can limit evaluation of arterial structures. Small distal arteries such as external carotid branches are suboptimally evaluated. CTA interpretation begins with source images, evaluating for caliber change, nonanatomic cross sections, intimal flap, vessel wall abnormalities such as hematoma, and of course extravasation (Fig. 28). Normal arterial cross sections away from bifurcations, ‘‘kinked’’ vessels, and dramatic turns or loops are round or oval. Most diagnostic information is available from source data, but some pathology is seen best on rendered or reformatted images (Fig. 29).

Penetrating Injury

There is increasing evidence supporting the use of CTA in this setting; for example, a recent prospective study of 175 patients with suspected arterial injury from penetrating trauma using single-detector CTA (87) reported a sensitivity and specificity of 100% and 98.6% and positive and negative predictive values of 92.8% and 100%, respectively. Accuracy should improve with MDCT. A large study of CTA for intracranial penetrating trauma is not yet available, and DSA may still be required (88). Partial or complete occlusions are the most commonly identified carotid

Figure 29 Blunt neck trauma. (A) Filling defect in the common carotid artery proximal to bifurcation is noted on CTA source image, but better demonstrated on MIP (B). Abbreviation: MIP, maximum intensity projection.

108

Wolf

Figure 30 Penetrating vascular injury. (A) Source and (B) oblique MIP images from CTA show filling defects in left ICA lumen suggesting thrombi (arrows), poorly visualized on (C) DSA. Caliber change suggesting dissection is visible on both studies. Abbreviations: MIP, maximum intensity projection; ICA, internal carotid artery; DSA, digital subtraction angiography.

injury in penetrating or blunt trauma (Fig. 30). Pseudoaneurysms occur in about one-third of penetrating injuries to the ICA. Arterial extravasation or transection is less commonly seen with noninvasive imaging, since these patients likely require a more aggressive workup with DSA or surgery (Fig. 31). AVFs are less commonly seen acutely, but may become evident later on. Blunt Injury

Blunt carotid vertebral injury (BCVI) is less common than penetrating injury (Figs. 28,29). Aggressive

screening has been recommended on the basis of the impression that the rate of BCVI is higher than previously recognized, that patients are often asymptomatic for hours to days before an injury becomes evident, and that BCVI is treatable. Lesions for blunt carotid injury (BCI) are typically graded on a five-point scale (89): grade I, lumen irregularity/dissection (<25% narrowed); grade II, dissection/intimal flap or intramural/intralumenal thrombus (25% narrowed); grade III, pseudoaneurysm; grade IV, occlusion; and grade V, transection/extravasation (Fig. 32, see also Figs. 28–31). Early reports with single-detector CT

Figure 31 Arterial transection or extravasation from penetrating injury. CTA source images demonstrate distorted true lumen of ICA proximally (A, arrow) and faint filling of pseudoaneurysm or hematoma (B and C, arrowheads). Distal ICA was not visualized.

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology

109

Figure 32 Blunt trauma with carotid pseudoaneurysm formation. (A) Source image from CTA demonstrates intimal flap (arrowhead, A) with increased caliber overall. (B) Oblique MIP image better demonstrates pseudoaneurysm (arrow).

technology reported suboptimal accuracy of CTA, but accuracy will likely improve with MDCT technology. One recent study (90) reported an incidence of 0.60% for BCVI for all blunt trauma admissions, incidence of BCVI of 3.7% in screened high-risk patients, and sensitivity and specificity of CTA for BCVI of 100% and 94%, respectively.

Traumatic Intracranial Aneurysms These can be caused by blunt trauma, penetrating injuries, and/or fractures. Though the term ‘‘pseudoaneurysm’’ is often used to describe incomplete disruption of all layers of the vessel wall (including this chapter), a true pseudoaneurysm or false aneurysm is really complete disruption of all vessel wall layers, with blood contained by perivascular clot. It is difficult to distinguish different forms of traumatic aneurysm on imaging, and it is unclear whether noninvasive imaging is adequate for excluding traumatic intracranial aneurysms. These aneurysms can form in the cavernous ICA in the setting of skull base fractures or penetrating (or iatrogenic) injuries. Peripheral vessels can be involved as well, including MCA, ACA, and middle meningeal arteries and less commonly smaller branches. For example, pericallosal aneurysms can form from shearing or compressive injuries against the falx. They may have a delayed presentation.

REFERENCES 1. Mahesh M. Search for isotropic resolution in CT from conventional through multiple-row detector. Radiographics 2002; 22(4):949–962. 2. Flohr TG, Schaller S, Stierstorfer K, et al. Multi-detector row CT systems and image-reconstruction techniques. Radiology 2005; 235(3):756–773. 3. Patel SG, Collie DA, Wardlaw JM, et al. Outcome, observer reliability, and patient preferences if CTA, MRA, or Doppler ultrasound were used, individually or together, instead of digital subtraction angiography before carotid endarterectomy. J Neurol Neurosurg Psychiatry 2002; 73(1):21–28. 4. Tomandl BF, Kostner NC, Schempershofe M, et al. CT angiography of intracranial aneurysms: a focus on postprocessing. Radiographics 2004; 24(3):637–655.

5. Takahashi M, Ashtari M, Papp Z, et al. CT angiography of carotid bifurcation: artifacts and pitfalls in shaded-surface display. AJR Am J Roentgenol 1997; 168(3):813–817. 6. Lell M, Anders K, Klotz E, et al. Clinical evaluation of bonesubtraction CT angiography (BSCTA) in head and neck imaging. Eur Radiol 2006; 16(4):889–897; [Epub 2005 Nov 3]. 7. Venema HW, Hulsmans FJ, den Heeten GJ. CT angiography of the circle of Willis and intracranial internal carotid arteries: maximum intensity projection with matched mask bone elimination-feasibility study. Radiology 2001; 218(3): 893–898. 8. Masaryk AM, Frayne R, Unal O, et al. Utility of CT angiography and MR angiography for the follow-up of experimental aneurysms treated with stents or Guglielmi detachable coils. AJNR Am J Neuroradiol 2000; 21 (8):1523–1531. 9. Wintermark M, Sesay M, Barbier E, et al. Comparative overview of brain perfusion imaging techniques. Stroke 2005; 36(9):e83–e99. 10. Latchaw RE, Yonas H, Hunter GJ, et al. Guidelines and recommendations for perfusion imaging in cerebral ischemia: a scientific statement for healthcare professionals by the writing group on perfusion imaging, from the Council on Cardiovascular Radiology of the American Heart Association. Stroke 2003; 34(4):1084–1104. 11. Eastwood JD, Lev MH, Provenzale JM. Perfusion CT with iodinated contrast material. AJR Am J Roentgenol 2003; 180(1):3–12. 12. Wintermark M, Maeder P, Thiran JP, et al. Quantitative assessment of regional cerebral blood flows by perfusion CT studies at low injection rates: a critical review of the underlying theoretical models. Eur Radiol 2001; 11 (7):1220–1230. 13. Klotz E, Konig M. Perfusion measurements of the brain: using dynamic CT for the quantitative assessment of cerebral ischemia in acute stroke. Eur J Radiol 1999; 30 (3):170–184. 14. Hunter GJ, Hamberg LM, Ponzo JA, et al. Assessment of cerebral perfusion and arterial anatomy in hyperacute stroke with three-dimensional functional CT: early clinical results. AJNR Am J Neuroradiol 1998; 19(1):29–37. 15. Winkler S, Turski P. Potential hazards of xenon inhalation. AJNR Am J Neuroradiol 1985; 6(6):974–975. 16. Yonas H, Pindzola RP, Johnson DW. Xenon/computed tomography cerebral blood flow and its use in clinical management. Neurosurg Clin N Am 1996; 7(4):605–616. 17. Osborn AG. Intracranial hemorrhage. In: Osborn AG, ed. Diagnostic Neuroradiology. 1st ed. St. Louis: Mosby-Year Book, Inc., 1994:173.

110

Wolf

18. Greenberg MS. SAH and aneurysms. Handbook of Neurosurgery, vol. 2. 4th ed. Lakeland, FL: Greenberg Graphics, Inc., 1997:832. 19. Kangasniemi M, Makela T, Koskinen S, et al. Detection of intracranial aneurysms with two-dimensional and threedimensional multislice helical computed tomographic angiography. Neurosurgery 2004; 54(2): 336–340; discussion 340–331. 20. U-King-Im JM, Koo B, Trivedi RA, et al. Current diagnostic approaches to subarachnoid haemorrhage. Eur Radiol 2005; 15(6):1135–1147; [Epub 2005 Feb 12]. 21. Kirkpatrick PJ. Subarachnoid haemorrhage and intracranial aneurysms: what neurologists need to know. J Neurol Neurosurg Psychiatry 2002; 73(suppl 1):28–33. 22. van Gijn J, Rinkel GJ. Subarachnoid haemorrhage: diagnosis, causes and management. Brain 2001; 124(pt 2):249–278. 23. Latchaw RE, Silva P, Falcone SF. The role of CT following aneurysmal rupture. Neuroimaging Clin N Am 1997; 7 (4):693–708. 24. Karttunen AI, Jartti PH, Ukkola VA, et al. Value of the quantity and distribution of subarachnoid haemorrhage on CT in the localization of a ruptured cerebral aneurysm. Acta Neurochir (Wien) 2003; 145(8):655–661; discussion 661. 25. van der Jagt M, Hasan D, Bijvoet HW, et al. Validity of prediction of the site of ruptured intracranial aneurysms with CT. Neurology 1999; 52(1):34–39. 26. Nehls DG, Flom RA, Carter LP, et al. Multiple intracranial aneurysms: determining the site of rupture. J Neurosurg 1985; 63(3):342–348. 27. Krishnaney AA, Rasmussen PA, Masaryk T. Bilateral tentorial subdural hematoma without subarachnoid hemorrhage secondary to anterior communicating artery aneurysm rupture: a case report and review of the literature. AJNR Am J Neuroradiol 2004; 25(6):1006–1007. 28. Le Roux PD, Dailey AT, Newell DW, et al. Emergent aneurysm clipping without angiography in the moribund patient with intracerebral hemorrhage: the use of infusion computed tomography scans. Neurosurgery 1993; 33 (2):189–197; discussion 197. 29. White PM, Wardlaw JM, Easton V. Can noninvasive imaging accurately depict intracranial aneurysms? A systematic review. Radiology 2000; 217:361–370. 30. Chappell ET, Moure FC, Good MC. Comparison of computed tomographic angiography with digital subtraction angiography in the diagnosis of cerebral aneurysms: a meta-analysis. Neurosurgery 2003; 52(3):624–631; discussion 630–621. 31. Jayaraman MV, Mayo-Smith WW, Tung GA, et al. Detection of intracranial aneurysms: multi-detector row CT angiography compared with DSA. Radiology 2004; 230 (2):510–518; [Epub 2003 Dec 29]. 32. van Gelder JM. Computed tomographic angiography for detecting cerebral aneurysms: implications of aneurysm size distribution for the sensitivity, specificity, and likelihood ratios. Neurosurgery 2003; 53(3):597–605; discussion 605–596. 33. Wintermark M, Uske A, Chalaron M, et al. Multislice computerized tomography angiography in the evaluation of intracranial aneurysms: a comparison with intraarterial digital subtraction angiography. J Neurosurg 2003; 98 (4):828–836. 34. Villablanca JP, Jahan R, Hooshi P, et al. Detection and characterization of very small cerebral aneurysms by using 2D and 3D helical CT angiography. AJNR Am J Neuroradiol 2002; 23(7):1187–1198. 35. Hashimoto H, Iida J, Hironaka Y, et al. Use of spiral computerized tomography angiography in patients with subarachnoid hemorrhage in whom subtraction angiog-

36.

37.

38.

39.

40.

41.

42.

43. 44.

45. 46.

47. 48.

49.

50.

51.

52.

53.

raphy did not reveal cerebral aneurysms. J Neurosurg 2000; 92(2):278–283. Hoh BL, Cheung AC, Rabinov JD, et al. Results of a prospective protocol of computed tomographic angiography in place of catheter angiography as the only diagnostic and pretreatment planning study for cerebral aneurysms by a combined neurovascular team. Neurosurgery 2004; 54(6):1329–1340; discussion 1340–1342. Velthuis BK, Van Leeuwen MS, Witkamp TD, et al. Computerized tomography angiography in patients with subarachnoid hemorrhage: from aneurysm detection to treatment without conventional angiography. J Neurosurg 1999; 91(5):761–767. Anderson GB, Ashforth R, Steinke DE, et al. CT angiography for the detection of cerebral vasospasm in patients with acute subarachnoid hemorrhage. AJNR Am J Neuroradiol 2000; 21(6):1011–1015. Hoeffner EG, Case I, Jain R, et al. Cerebral perfusion CT: technique and clinical applications. Radiology 2004; 231 (3):632–644. Young N, Dorsch NW, Kingston RJ, et al. Intracranial aneurysms: evaluation in 200 patients with spiral CT angiography. Eur Radiol 2001; 11(1):123–130. van Loon JJ, Yousry TA, Fink U, et al. Postoperative spiral computed tomography and magnetic resonance angiography after aneurysm clipping with titanium clips. Neurosurgery 1997; 41(4):851–856; discussion 856–857. Sagara Y, Kiyosue H, Hori Y, et al. Limitations of threedimensional reconstructed computerized tomography angiography after clip placement for intracranial aneurysms. J Neurosurg 2005; 103(4):656–661. Anzalone N, Scomazzoni F, Strada L, et al. Intracranial vascular malformations. Eur Radiol 1998; 8(5):685–690. Tanaka H, Numaguchi Y, Konno S, et al. Initial experience with helical CT and 3D reconstruction in therapeutic planning of cerebral AVMs: comparison with 3D timeof-flight MRA and digital subtraction angiography. J Comput Assist Tomogr 1997; 21(5):811–817. Casey SO, Alberico RA, Patel M, et al. Cerebral CT venography. Radiology 1996; 198(1):163–170. Kolominsky-Rabas PL, Weber M, Gefeller O, et al. Epidemiology of ischemic stroke subtypes according to TOAST criteria: incidence, recurrence, and long-term survival in ischemic stroke subtypes: a population-based study. Stroke 2001; 32(12):2735–2740. Mullins ME. Modern emergent stroke imaging: pearls, protocols, and pitfalls. Radiol Clin N Am 2006; 44:41–62. Kidwell CS, Chalela JA, Saver JL, et al. Comparison of MRI and CT for detection of acute intracerebral hemorrhage. JAMA 2004; 292(15):1823–1830. Perl J II, Tkach JA, Porras-Jimenez M, et al. Hemorrhage detected using MR imaging in the setting of acute stroke: an in vivo model. AJNR Am J Neuroradiol 1999; 20 (10):1863–1870. Grant PE, Yu D. Acute injury to the immature brain with hypoxia with or without hypoperfusion. Radiol Clin N Am 2006; 44:63–77. Kilpatrick MM, Yonas H, Goldstein S, et al. CT-based assessment of acute stroke: CT, CT angiography, and xenon-enhanced CT cerebral blood flow. Stroke 2001; 32 (11):2543–2549. Knauth M, von Kummer R, Jansen O, et al. Potential of CT angiography in acute ischemic stroke. AJNR Am J Neuroradiol 1997; 18(6):1001–1010. Lev MH, Farkas J, Rodriguez VR, et al. CT angiography in the rapid triage of patients with hyperacute stroke to intraarterial thrombolysis: accuracy in the detection of large vessel thrombus. J Comput Assist Tomogr 2001; 25 (4):520–528.

Chapter 5: CT Imaging and Physiologic Techniques in Interventional Neuroradiology 54. Shrier DA, Tanaka H, Numaguchi Y, et al. CT angiography in the evaluation of acute stroke. AJNR Am J Neuroradiol 1997; 18(6):1011–1020. 55. Wildermuth S, Knauth M, Brandt T, et al. Role of CT angiography in patient selection for thrombolytic therapy in acute hemispheric stroke. Stroke 1998; 29(5): 935–938. 56. Verro P, Tanenbaum LN, Borden NM, et al. CT angiography in acute ischemic stroke: preliminary results. Stroke 2002; 33(1):276–278. 57. Tomandl BF, Klotz E, Handschu R, et al. Comprehensive imaging of ischemic stroke with multisection CT. Radiographics 2003; 23(3):565–592. 58. Lev MH, Segal AZ, Farkas J, et al. Utility of perfusionweighted CT imaging in acute middle cerebral artery stroke treated with intra-arterial thrombolysis: prediction of final infarct volume and clinical outcome. Stroke 2001; 32(9):2021–2028. 59. Eastwood JD, Lev MH, Azhari T, et al. CT perfusion scanning with deconvolution analysis: pilot study in patients with acute middle cerebral artery stroke. Radiology 2002; 222(1):227–236. 60. Koenig M, Klotz E, Luka B, et al. Perfusion CT of the brain: diagnostic approach for early detection of ischemic stroke. Radiology 1998; 209(1):85–93. 61. Wintermark M, Reichhart M, Thiran JP, et al. Prognostic accuracy of cerebral blood flow measurement by perfusion computed tomography, at the time of emergency room admission, in acute stroke patients. Ann Neurol 2002; 51(4):417–432. 62. Firlik AD, Kaufmann AM, Wechsler LR, et al. Quantitative cerebral blood flow determinations in acute ischemic stroke. Relationship to computed tomography and angiography. Stroke 1997; 28(11):2208–2213. 63. Hughes RL, Yonas H, Gur D, et al. Cerebral blood flow determination within the first 8 hours of cerebral infarction using stable xenon-enhanced computed tomography. Stroke 1989; 20(6):754–760. 64. Josephson SA, Bryant SO, Mak HK, et al. Evaluation of carotid stenosis using CT angiography in the initial evaluation of stroke and TIA. Neurology 2004; 63(3):457–460. 65. Lu X, Zhang W, Gui Q, et al. Evaluating non-invasive medical imaging for diagnosis of carotid artery stenosis with ischemic cerebrovascular disease. Chin Med J (Engl) 2003; 116(1):112–115. 66. Berg M, Zhang Z, Ikonen A, et al. Multi-detector row CT angiography in the assessment of carotid artery disease in symptomatic patients: comparison with rotational angiography and digital subtraction angiography. AJNR Am J Neuroradiol 2005; 26(5):1022–1034. 67. Randoux B, Marro B, Koskas F, et al. Carotid artery stenosis: prospective comparison of CT, three-dimensional gadolinium-enhanced MR, and conventional angiography. Radiology 2001; 220(1):179–185. 68. Hollingworth W, Nathans AB, Kanne JP, et al. The diagnostic accuracy of computed tomography angiography for traumatic or atherosclerotic lesions of the carotid and vertebral arteries: a systematic review. Eur J Radiol 2003; 48:88–102. 69. Koelemay MJ, Nederkoorn PJ, Reitsma JB, et al. Systematic review of computed tomographic angiography for assessment of carotid artery disease. Stroke 2004; 35 (10):2306–2312; [Epub 2004 Sep 02]. 70. Chen CJ, Lee TH, Hsu HL, et al. Multi-slice CT angiography in diagnosing total versus near occlusions of the internal carotid artery: comparison with catheter angiography. Stroke 2004; 35(1):83–85; [Epub 2003 Dec 18]. 71. Kinoshita T, Ogawa T, Kado H, et al. CT angiography in the evaluation of intracranial occlusive disease with col-

72.

73.

74.

75.

76. 77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

111

lateral circulation: comparison with MR angiography. Clin Imaging 2005; 29(5):303–306. Bash S, Villablanca JP, Jahan R, et al. Intracranial vascular stenosis and occlusive disease: evaluation with CT angiography, MR angiography, and digital subtraction angiography. AJNR Am J Neuroradiol 2005; 26(5):1012–1021. Hirai T, Korogi Y, Ono K, et al. Prospective evaluation of suspected stenoocclusive disease of the intracranial artery: combined MR angiography and CT angiography compared with digital subtraction angiography. AJNR Am J Neuroradiol 2002; 23(1):93–101. Fayad ZA, Sirol M, Nikolaou K, et al. Magnetic resonance imaging and computed tomography in assessment of atherosclerotic plaque. Curr Atheroscler Rep 2004; 6 (3):232–242. Thanvi B, Munshi SK, Dawson SL, et al. Carotid and vertebral artery dissection syndromes. Postgrad Med J 2005; 81(956):383–388. Pelkonen O, Tikkakoski T, Leinonen S, et al. Intracranial arterial dissection. Neuroradiology 1998; 40(7):442–447. Provenzale JM. Dissection of the internal carotid and vertebral arteries: imaging features. AJR Am J Roentgenol 1995; 165(5):1099–1104. Leclerc X, Godefroy O, Salhi A, et al. Helical CT for the diagnosis of extracranial internal carotid artery dissection. Stroke 1996; 27(3):461–466. Nakatsuka M, Mizuno S. Three-dimensional computed tomographic angiography in four patients with dissecting aneurysms of the vertebrobasilar system. Acta Neurochir 2000; 142(9):995–1001. Zuber M, Meary E, Meder JF, et al. Magnetic resonance imaging and dynamic CT scan in cervical artery dissections. Stroke 1994; 25(3):576–581. Chen CJ, Tseng YC, Lee TH, et al. Multisection CT angiography compared with catheter angiography in diagnosing vertebral artery dissection. AJNR Am J Neuroradiol 2004; 25(5):769–774. Derdeyn CP, Grubb RL, Jr., Powers WJ. Cerebral hemodynamic impairment: methods of measurement and association with stroke risk. Neurology 1999; 53(2):251–259. Schaller B, Graf R. Cerebral ischemia and reperfusion: the pathophysiologic concept as a basis for clinical therapy. J Cereb Blood Flow Metab 2004; 24(4):351–371. Larrue V, von Kummer RR, Muller A, et al. Risk factors for severe hemorrhagic transformation in ischemic stroke patients treated with recombinant tissue plasminogen activator: a secondary analysis of the European-Australasian Acute Stroke Study (ECASS II). Stroke 2001; 32(2):438–441. Firlik AD, Yonas H, Kaufmann AM, et al. Relationship between cerebral blood flow and the development of swelling and life-threatening herniation in acute ischemic stroke. J Neurosurg 1998; 89(2):243–249. Reigel MM, Hollier LH, Sundt TM Jr., et al. Cerebral hyperperfusion syndrome: a cause of neurologic dysfunction after carotid endarterectomy. J Vasc Surg 1987; 5 (4):628–634. Munera F, Soto JA, Palacio DM, et al. Penetrating neck injuries: helical CT angiography for initial evaluation. Radiology 2002; 224(2):366–372. Stallmeyer MJB, Morales RE, Flanders AE. Imaging of traumatic neurovascular injury. Radiol Clin N Am 2006; 44:13–39. Biffl WL, Moore EE, Offner PJ, et al. Blunt carotid arterial injuries: implications of a new grading scale. J Trauma 1999; 47(5):845–853. Berne JD, Norwood SH, McAuley CE, et al. Helical computed tomographic angiography: an excellent screening test for blunt cerebrovascular injury. J Trauma 2004; 57 (1):11–17; discussion 17–19.

6 MR Angiography: Principles and Applications in Interventional Neuroradiology Neerav R. Mehta and Elias R. Melhem University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A.

INTRODUCTION Magnetic resonance imaging has revolutionized noninvasive imaging in the late 20th century and the early 21st century. In particular, imaging of the vascular tree has become possible with MRI technology, with continual refinements and improvements driving it into the future. Magnetic resonance angiography has roots that date back to the early 1950s, when, in the Department of Physics at the Indian Institute of Sciences, G. Suryan, using a U tube and coils of wire, discovered inflow effects. Singer in the late 1950s applied Suryan’s discovery of inflow effects in vivo, using a mouse and a tourniquet. From these humble beginnings, MRI and MRA technology has progressed to its current level of sophistication, with both neurological and nonneurological applications (1,2).

TIME-OF-FLIGHT TECHNIQUE The vast majority of neurovascular MRA performed today is via time-of-flight (TOF) techniques. It is widely employed because of its ready availability as well as the ease of acquisition of diagnostic studies. Essentially, these techniques make use of blood inflow effects to produce high intravascular signal, while signal from background stationary tissues are minimized (3–6). Let us begin with a look at a TOF sequence. Initially, all spins are aligned along the bore of the magnet in the positive z axis. The spins are all precessing at the same Larmor frequency (Fig. 1). A slice select gradient is applied, and spins within a specific slice are tipped into the xy plane via a 908 radiofrequency (RF) pulse. If no other RF pulses are applied, then the spins will relax back toward the positive z axis. At time T1, approximately 63% of the magnetization has recovered to the positive z axis (Fig. 2). In TOF MRA, however, before the tissues can relax back to the positive z axis, additional RF pulses are applied. These pulses are applied at time TR, so that TR is less than T1 of the tissues. A closer look at this event reveals that the first RF pulse tips the spins in the xy plane. The xy component of the signal dephases quickly, and the z component begins to

grow, in accordance with the tissue’s T1 at the given magnetic strength. Before the z component has recovered, a second RF pulse is applied and the partially recovered z component is tipped into the xy plane. This action results in a smaller xy component, which subsequently dephases and results in a smaller recovering z component. A third, fourth, and fifth RF pulse are applied in a similar manner, each resulting in smaller and smaller z components and xy components. Eventually, this process reaches a steady state and the spins are saturated. The shorter the TR (the time between subsequent RF pulses), the greater the degree of saturation (Fig. 3). The more saturated the spins become, the less is the signal that can be measured from those spins (as the xy components get smaller in accordance with the RF pulse until a steady state is reached). Recall that both z components and xy components of the saturated spins are small. If fresh spins enter a slice with a full z-axis magnetization vector, then the moment they are tipped into the xy plane, they will produce a large amount of signal relative to the saturated spins. Hence, a blood vessel bringing in fresh spins will have high intravascular signal as the fresh spins traverse a slice of saturated stationary spins. This mechanism is the heart of TOF MRA (Fig. 4). If the flow of blood is slow and a volume of blood lingers in a slice for too long, then the repetitive RF pulses that saturate the stationary tissues also begin to saturate the blood. As mentioned previously, shorter TR leads to greater saturation, which also applies to blood, and a shorter TR will cause greater blood saturation once it has entered into the slice. When blood initially enters into the slice, it has its greatest magnetization in the z axis. This phenomenon is called entry slice phenomenon. As blood traverses a volume that is continuously receiving RF pulses, the blood itself gets saturated and the signal diminishes the further it travels into the volume of interest. In practice, a longer TR can be chosen in order to preserve intravascular signal over a large volume. However, this choice is made at the cost of decreased background tissue saturation. In TOF MRA, 908 RF pulses are not typically used; instead, pulses of varying flip angles are

114

Figure 1

Figure 2

Figure 3

Mehta and Melhem

employed. The greater the flip angle, the greater the background saturation of stationary tissues. This inference can be better understood by comparing the scenario of the 908 pulse with a scenario in which the flip angle is 18. The 18 flip angle would lead to negligible saturation of stationary tissues. A large z-axis component would exist even after multiple RF pulses are used. The signal obtained from inflowing blood would be poorly differentiated from stationary tissues. As the flip angle increases, the saturation of stationary tissues also increases. As in the case of shorter TR times, with larger flip angles and greater background stationary tissue saturation, there is also greater saturation of blood as it traverses a volume. Hence, the same caveat that applies to shorter TR times also applies to larger flip angles, and flip angles can be varied to preserve intravascular signal at the cost of decreased background tissue saturation. However, 908 flip angles are typically not used in

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

115

Figure 4

clinical practice. Blood flowing within the carotid and cerebral arteries usually has laminar flow, whereas blood along the center of the vessel moves at a faster velocity than blood along the periphery of the vessel. The blood along the periphery moves so slowly that the large, 908 flip angles quickly cause blood saturation, which is not the case for the faster moving blood along the central part of the lumen. The net result is the signal arising from the central portions of the vessel and lack of signal from the periphery, ultimately causing a perceived decrease in the caliber of the vessel. This observation should be noted on most MRA sequences, particularly when slow flow is involved in scans performed with larger flip angles. In Figure 5 (A through D), flip angles and TR were varied on the same subject to produce MRA images that are presented using collapsed maximum intensity projection (MIP) algorithm in the axial projection. Figure 5A serves as the reference 3D TOF MRA performed with a typical flip angle of 258 and TR of 42 milliseconds. Now, compare Figure 5A with Figure 5B, where the flip angle was decreased to 108 (TR remains 42 milliseconds). The signal within the large vessels is decreased; however, less blood saturation results in better visualization of the small peripheral arteries. Compare Figure 5B, where the flip angle is 108, with Figure 5C, where the flip angle is increased to 508 (TR remains 42 milliseconds). Note that there is increased signal within the large vessels; however, blood saturation has resulted in poor visualization of the small peripheral arteries. Compare Figure 5A with Figure 5D, where the flip angle is kept at 258, but the TR has been increased to 84 milliseconds. The increased TR leads to less blood saturation, and hence improved visualization of small peripheral arteries. However, the stationary tissues are also less saturated, resulting in less contrast between the arteries and surrounding tissues. Figure 5E demonstrates 3D TOF MRA performed on a 3.0-tesla magnet. The TR and the echo time (TE) are 24 and

3 milliseconds, respectively. The higher field magnet theoretically doubles the signal-to-noise ratio as well as increases vessel contrast with respect to surrounding tissues. Shorter TE times allow for less phase dispersion and hence higher intravascular signal within more peripheral vessels (see below for effect of TE times on TOF MRA). Compare Figure 5E with Figure 5A, both of which are performed on the same subject, to get a taste of what can be expected as routine MR imaging migrates from 1.5 to 3.0 tesla.

2D TOF 2D TOF involves a sequential acquisition, slice by slice. A thin slice is selected, and the spins within the slice are saturated. The blood flowing perpendicularly into the slice is bright, with high intravascular signal. The blood flowing into the slice at an oblique angle would have less intravascular signal, as it would have to traverse a greater distance than within the perpendicularly oriented vessel. As a volume of blood traverses a greater distance within a slice, it experiences a greater degree of blood saturation (Fig. 6). If MRA images were to be acquired at this point, the inflowing blood from both above and below the slice would provide intravascular signal. In the case of neck MRA, if a slice of the midneck were to be acquired, intravascular signal from both carotid and vertebral arteries and the jugular veins would be acquired. If a parallel saturation band is placed above the slice of interest, then the spins within the jugular veins would get saturated before they enter the slice of interest (Fig. 7). The resultant image would only have intravascular signal from the carotid and vertebral arteries. If a parallel saturation band is placed below the slice of interest, then the spins within the carotid and vertebral arteries would get saturated before they enter the slice of interest.

116

Mehta and Melhem

Figure 5 Compare the differences between the reference MRA data set performed at 1.5 tesla and those with varying flip angles and TR times. (A) The reference (TR 42, TE 3, flip angle 258). (B) Same TR and TE as the reference, but with a flip angle of 108. (C) Same TR and TE as the reference, but with a flip angle of 508. (D) Same TE and flip angle as the reference, but with a TR of 848. (E) MRA of the same subject performed at 3.0 tesla.

Figure 6

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

117

Figure 7

Figure 8

The resulting image would only have intravascular signal from the jugular veins (7). After multiple slices are acquired, the resultant data set is usually stacked and displayed using the MIP algorithm. Postprocessing of MRA data is required in order to display projection images reminiscent of traditional angiography. Simply stacking the data and then viewing the summation from the side does not work, as there is too much background tissue that overlaps and obscures the vasculature. MIPs transgress this limitation by displaying only the maximum pixel value for a given projection line. Multiple projection images can be calculated from the stacked data set to provide images at different angles (Fig. 8). By scrolling through a data set of MIPs, a 3D appreciation of vasculature can be obtained. At the Hospital of the University of Pennsylvania (HUP), projection images are calculated at 68 increments over 1808. Currently, 2D TOF is primarily used for imaging carotid and vertebral arteries in the neck. These vessels have an optimal orientation to acquire 2D slices as they enter each slice with an almost perpendicular orientation, lacking significant tortuosity. Currently, at HUP, 2D TOF of the neck is performed with a flip angle of 608, TR of 45 milliseconds, and TE of 6.9 milliseconds. The slice thickness of 1.5 mm is used with a

matrix size of 256
128 for a total of 128 slices. This procedure is then supplemented with a 3D gadolinium-enhanced sequence (see the next section).

3D TOF 3D TOF is the same concept as 2D TOF, except that a slab or volume (3–8 cm) instead of a thin slice (1.5 mm) is obtained. There is no slice selection; instead the z axis is partitioned into 32 to 64 slices with multiple phase-encoding steps. This partition results in very thin slices, usually 1 mm or less in thickness. However, since blood is flowing through a large 3D volume during the acquisition, in contrast to the thin slices of 2D acquisitions, the blood can be saturated as it courses through the slab. This technique does somewhat limit the evaluation of slower flow. Modifications of the technique include making the slabs thinner and performing multiple sequential acquisitions of these thinner slabs. This technique has been termed multiple overlapping thin slab acquisition (MOTSA). One of the major advantages of employing the MOTSA technique is the reduction of saturation effects. Flip angles are also adjusted accordingly, with 3D acquisition flip angles smaller than those of 2D acquisitions. With a smaller flip angle,

118

Mehta and Melhem

blood saturation can get minimized in the 3D TOF sequence. Flip angles for 3D TOF range from 158 to 358, whereas for 2D TOF the flip angle ranges from 408 to 908. Other methods to reduce saturation effects within the vessels and to enhance the visualization of the small peripheral intracranial arteries have been developed. These include tilted optimized nonsaturating excitation (TONE) or ramped RF and magnetization transfer imaging (MTI). The first (TONE or ramped RF) is based on varying the flip angles across a volumetric slab. A voxel of blood entering a slab experiences a smaller flip angle, and as it traverses the slab it experiences gradually increasing flip angles. This process can help minimize saturation of blood as it traverses the volumetric slab. The second (MTI) is based on additional saturation of brain tissue surrounding the small intracranial arteries. With magnetization transfer, the bound water within brain tissue is saturated with an RF pulse targeting the bound water’s proton precession rate (which is lower than that of free water). The bound water then interacts with local free water to exchange a saturated proton for an unsaturated proton. The net result is saturation, and hence suppression, of free water in the tissues. MTI can assist in background suppression without having to resort to increasing flip angles or decreasing TR (9–11). One of the major advantages of 3D acquisition over 2D acquisition is the characterization of flow in tortuous vessels, because 2D acquisitions are much more dependent on the angle of vessel entry into a given slice (see above). In 3D acquisitions, blood can flow in any direction and produce signal, as long as blood saturation does not occur. Given the tortuosity of the intracranial vessels, 3D TOF is much more widely employed for intracranial vasculature evaluation. At HUP, evaluation of intracranial circulation is primarily performed with 3D TOF acquisition, with a flip angle of 258, TR of 48 milliseconds, and TE of 5.6 milliseconds. A total of 60 slices are obtained with a 1 mm thickness and a 512
192 matrix.

Limitations The characteristics that limit TOF MRA include nonlaminar blood flow, slow flow, and tortuous vasculature. Nonlaminar flow of blood leads to mixing of blood of differing phases. If the blood of two different phases mixes within a voxel, the resulting voxel will have lower signal intensity. This condition takes on particular clinical importance in the assessment of vascular stenoses. The flow distal to a carotid stenosis is usually nonlaminar, which will subsequently result in phase dispersion. The image produced will then overestimate the degree and length of stenosis. Techniques to decrease the degree of phase dispersion include minimizing TE, acquiring thinner slices, and using flow compensation techniques (see the next section). Slow flow is yet another source of error in MRA sequences. Slow flow within a TOF acquisition results in blood saturation and hence signal loss. This loss can somewhat be compensated for by increasing TR or

decreasing flip angles; however, it is at the cost of background suppression and vessel signal. In general, slow flow is more of a problem for 3D techniques than for 2D techniques, the major reason for which is greater blood saturation within a thick slab during the 3D acquisition compared with a thinner slice on a 2D acquisition. If a smaller, slow-flow vessel needs to be imaged by increasing TR or decreasing flip angles, then MTI can be used as a tool to assist in background suppression. Tortuous vasculature is an intrinsic concern when evaluating the intracranial circulation, particularly at the level of the cavernous carotid artery. A tortuous vessel may lie parallel within an imaging slice and become subject to saturation effects. Tortuosity can also result in vessels entering the slice both from above and below the slice of interest, resulting in signal loss in patent vasculature secondary to parallel saturation bands.

Flow Compensation Flow compensation, also known as gradient moment nulling, is a necessity for high-quality MRA images. Flow compensation addresses the issue of phase dispersion as blood within a vessel moves at a constant velocity. First-order flow compensation accounts for velocity, second-order for acceleration, third-order for change in acceleration, and so on. During an MR acquisition, blood will flow through a given volume of interest. As the readout gradient is applied across a slice, the flowing blood experiences multiple different magnetic fields and changes phase accordingly. Assume that a voxel of blood traverses a volume that is experiencing a magnetic field gradient. As the voxel travels along, it experiences different magnetic field strengths and hence accumulates phase as it travels through everincreasing local magnetic fields. The signal from the flowing blood can either be in phase with the surrounding tissues or out of phase. It so happens, through a quadratic relationship of phase gain with time, that during odd echoes there is dephasing, while during even echoes there is rephasing. Hence, increased intravascular signal is seen during even echoes. Flow compensation techniques essentially change the shape of the magnetic field gradient in order to reproduce the even-echo rephasing effect during the very first echo, resulting in increased intravascular signal. This signal is used primarily in first-order flow compensation. The gradient shape can be changed to account for secondorder and third-order flow compensation; however, with each additional order of compensation, the time of application of the gradient increases. This increased time leads to increased TEs. In general, the shorter the TEs, the less the effects of signal loss from nonlaminar flow. Thus, a balance must be struck between how complex and long a shaped gradient can be applied and the TE times. The first-order flow compensation is optimal, whereas the second- and third-order flow compensation is not worth the cost of the increased TE times (12).

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

119

Figure 9 Comparison of out-of-phase (A) and inphase (B) MRA acquisitions. Note the prominence of the orbital fat in the in-phase sequence.

Echo Time The effect of TE on MRA is also crucial to the understanding and production of adequate MRA images, especially in the case of nonlaminar blood flow. Nonlaminar flow causes a loss of signal secondary to phase dispersion. Between the RF pulse and the time for readout of signal, i.e., TE, nonlaminar flow allows regions of blood within a vessel of differing phases to mix. This mixing leads to loss of intravascular signal. If the TE is decreased, there is less time for these regions of blood with differing phases to mix and hence less loss of intravascular signal. TE should also be chosen to suppress signal from adjacent fat. Recall that fat and water precess at slightly different frequencies. At 1.5 tesla, water precesses 220 Hz faster than fat. This difference in precessional frequency allows for spins of water and fat molecules to be either in phase or out of phase with each other. As both molecules exist within adipose tissue, fat can be suppressed by choosing a TE where the spins are out of phase and hence have decreased signal (Fig. 9; 9A is out of phase and 9B is in phase).

CONTRAST-ENHANCED MRA The administration of contrast at first appears to be a natural evolution of the MRA technique. The gadolinium molecule itself is paramagnetic and effectively serves to shorten the T1 of the blood around it. The shortened T1 has the potential to provide MRA images with high contrast-to-noise ratios and high signal-tonoise ratios, as well as potentially shortening acquisition times secondary to shorter TR and TE times. Gadolinium administration also reduces the saturation effects of slow-flowing blood. The major limitation to administering contrast is in the venous contamination. The bolus of gadolinium must be administered and timed accordingly to minimize both parenchymal and venous phases. In essence, speed of imaging is one of

the core issues with contrast- enhanced MRA. Unfortunately, for intracranial circulation, venous contamination in the cavernous sinuses and basal veins severely limits evaluation of the circle of Willis (Fig. 10). For this reason, contrast-enhanced MRA has yet to be widely implemented for studies targeted to the circle of Willis. This limitation is not an issue, however, in the neck, where gadolinium-enhanced MRA is now commonly used in the evaluation of carotid stenosis. In performing contrast-enhanced MRA, typically 20 cc of gadolinium is administered at a rate of 2 to 3 mL/sec using a power injector. Timing the bolus of contrast is critical and can be performed by a number of different methods. Currently, these include ‘‘best guess’’ methods, manual timing bolus, and automated timing bolus. Manual methods involve the administration of a small bolus of gadolinium, typically 2 cc, followed by a 20-cc saline flush. The volume of interest is imaged with fast 2D gradient-recalled echo sequence, usually at a rate less than 1 frame/sec. Once maximal contrast appears in the target vessel, the delay can be calculated accordingly. Automated timing bolus methods involve the beginning of the imaging once contrast is detected in the vessel of interest. Imaging is subsequently performed using 3D SPGR technique. Early or late bolus timing can lead to significant artifacts and venous contamination, respectively (Fig. 11). As a general rule, the acquisition of data to fill the center of k space has to occur at the time when the maximum concentration of contrast material is in the vessel of interest, as this region of k space is where contrast-to-noise ratio is maximized. The periphery of k space is used to increase the definition of edges and borders (Fig. 12). Contrast-enhanced MRA has become very valuable in neck MRA for the evaluation of carotid artery stenosis. The faster acquisition times can result in 3D slab acquisition times of less than 20 seconds, compared with 2D TOF methods that can last longer than 12 minutes for the same coverage. During respiration, there is motion of the neck vasculature, which subsequently causes artifacts on the 2D TOF sequence.

120

Mehta and Melhem

Figure 10 The top image is the arterial phase of gadolinium. The bottom image is delayed with both arterial and venous signal. The top image demonstrates a left posterior communicating artery aneurysm (confirmed on CT angiography). Venous opacification on delayed image obscures the aneurysm. The timing on contrast-enhanced MRA is critical, particularly in intracranial MRA, to avoid missing pathology.

Figure 11 The bolus was timed too early on the image on the left, whereas the bolus was too late on the image on the right.

Figure 12 Central k space provides contrast. Peripheral k space defines lines and borders.

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

121

Figure 13 Reduced FOV acquisition is ‘‘unfolded’’ to produce the final image. Abbreviation: FOV, field of view.

With the 3D gadolinium sequence, the entire acquisition can be performed in one breathhold. In addition, the TE times of contrast-enhanced 3D MRA are usually three to four times shorter than those of 2D TOF MRA. With shorter TE, the effects of intravoxel phase dispersion get minimized. This minimization is crucial for the accurate assessment of the degree of a carotid stenosis. Contrast-enhanced MRA has now widely supplanted 2D TOF sequences in the neck. 2D TOF data are commonly used to supplement interpretation and serve as a backup in case the bolus of gadolinium is poorly timed. Typical TR or TE times for 3D gadolinium-enhanced MRA are 4.4 msec/1.6 msec, with a 258 to 308 flip angle. Current research efforts are focused on faster acquisitions via parallel imaging and more efficient filling of k space (e.g., maximize the bolus of contrast with the acquisition of central k space to maximize contrast resolution, while acquiring spatial resolution data at the periphery of k space before or after the bolus). Parallel imaging with the SENSE or SMASH technique offers the ability to reduce scan times and improve temporal resolution. SENSE, which stands for sensitivity encoding, uses multiple coils in combination with a reduced field of view to reduce image acquisition times. If a reduced field of view were used with traditional 2D or 3D Fourier transform techniques, wraparound artifact would result. However, SENSE uses multiple coils, each with different sensitivity weightings determined by their orientation around the volume to be imaged. This configuration allows a reduced field-of-view image to be acquired. The reduced images are extrapolated to a full field of view, which appears ‘‘folded’’ by the wraparound artifact. The different sensitivity weightings from each coil are then used to ‘‘unfold’’ the image (Fig. 13). With the time savings of SENSE, contrast bolus can potentially be tracked into the arterial, capillary, and venous phase (13–18).

PHASE CONTRAST MRA Phase contrast angiography (PCA) is a third major technique that has been developed for MR. Unlike TOF angiography, which primarily uses magnitude data from the MR acquisition, PCA uses the phase data. Hence, while TOF images still have signal from surrounding anatomic structures, PCA images demonstrate signal only from flowing blood (Fig. 14). One of the major advantages of PCA over TOF imaging is in the assessment of flow direction and velocity. A second advantage of PCA is in the delineation of slow flow. TOF, however, is a faster technique that requires only one acquisition, whereas PCA requires four separate acquisitions to create one data set. Essentially, in PCA, contrast is achieved in blood vessels by tagging moving blood with phase changes. Moving objects develop a phase change as opposed to stationary objects during the acquisition. Moreover, the velocity and direction of the motion can be determined from the data set. The concept of phase contrast can be understood by taking the case of a blood vessel within a slice of interest and following the effects of the varying gradients on a voxel of blood traveling within that vessel. The particular gradient that makes phase contrast possible is the bipolar gradient (19–21). Take a slice of certain thickness with a blood vessel oriented within it so that the vessel is parallel to the x axis and perpendicular to the y axis. In practice, such a vessel could represent the M1 segment of the middle cerebral artery as a patient has been placed supine within a magnet bore, with the cranial direction corresponding to the positive z axis. Initially, prior to any RF pulse application, all spins are initially oriented so that their net magnetization points in the positive z direction, along the bore of the main magnetic field of the magnet. Spins of stationary tissue as well as spins within the blood vessel are all initially oriented in the same direction of the positive z axis.

122

Mehta and Melhem

Figure 14 The left image is a TOF image, and the right is a phase contrast angiogram. The TOF image demonstrates a saccular structure adjacent to the left posterior cerebral artery with similar signal characteristics of the surrounding vasculature, findings were suspicious for aneurysm. The phase contrast angiogram shows no signal in the suspected aneurysm and no flow. The subsequent digital subtraction angiogram demonstrates no aneursym. Abbreviation: TOF, time-of-flight.

Figure 15

Figure 16

A 908 RF pulse is applied. The spins that were pointing in the positive z axis are now flipped into the xy plane (Fig. 15). Both stationary tissue spins and spins within the blood vessel are flipped. A gradient is applied along the x dimension (Fig. 16). The gradient is such that spins on the left side of the slice experience a larger field than those on the right. The magnetic field on the left side is larger, tapering down to a smaller magnitude on the right side. Hence, the spins on the left side of the slice precess faster than those on the right. When the gradient is turned off, the spins throughout the slice once again precess at the same frequency, although with a change in phase. The spins on the left side of the slice have a different phase from the spins on the right side, they have gained phase. A second gradient is now applied along the x dimension. This gradient is in the opposite direction and of reversed polarity, but equal in magnitude to the first gradient (Fig. 17). It is such that the spins on the left side of the slice experience a smaller magnetic field than those on the right. The spins on the left,

Figure 17

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

therefore, precess more slowly than those on the right. When the second gradient is turned off, the spins once again precess at the same rate. During the gradient, the spins on the right gained phase, canceling the phase gain that the spins on the left had experienced during the first gradient. For tissues that were stationary during the time of application of the two gradients, no net change in phase was induced. Take the case of a blood vessel with a given volume of blood within it so that a volume of blood is initially located on the left side of the slice and eventually ends up located on the right. During the first gradient, the volume of blood experiences a larger magnetic field compared with spins on the right. This magnetic field induces a larger change in phase for that volume of blood. During the application of the first and second gradients, the volume of blood moves from the left of the slice (where the larger magnetic field was experienced and a large phase change occurred) to its right. During the second gradient, the volume of blood experiences a larger magnetic field, as it is now located on the right of the slice. So its phase never returns to its original phase. The volume of blood has gained phase relative to all the stationary spins. The blood that travels in the opposite direction would encounter a negative phase change. The distance that the volume of blood travels in the time the two gradients are applied determines the degree of phase change. If the blood has moved slowly and traveled a short distance, the phase change would be small. If the blood has moved fast, from the very edge of the slice on the left to the very edge on the right, then the phase change would be maximized. Hence, the change in phase is proportional to the velocity (Fig. 18).

The initial gradient is referred to as the ‘‘first lobe’’ of the bipolar gradient, while the second gradient is aptly referred to as the ‘‘second lobe’’ of the bipolar gradient. So a single application of a bipolar gradient (both lobes) results in a data set that demonstrates phase change in tissues with motion and no phase change in stationary tissues. In PCA, two acquisitions are performed with two different bipolar gradients. The first acquisition is as described above. The second is with a reversed bipolar gradient (Fig. 19). Once both sets of data are acquired, a subtracted data set is calculated. The method of subtraction is either via a complex difference technique or a phase difference technique. To further clarify the purpose of the second acquisition with reversed gradients and the subtraction technique, take the initial case of the first acquisition as described above. The initial bipolar gradient, performed during the first acquisition, results in zero phase change for stationary tissues. For moving blood within our hypothetical blood vessel, the blood that traverses the slice from the left to the right, assume that the first gradient induces a positive phase change. A second acquisition is subsequently performed, identical to the first acquisition but with the two lobes of the bipolar gradient reversed. This reversal again results in zero net change in phase for stationary tissues. However, for moving blood, the phase change is reversed compared with the initial acquisition: it is negative. If the phase data from the second acquisition are then subtracted from the phase data of the first, the phase of the stationary spins cancels out. The phase of the moving spins, on the other hand, being in opposite

Figure 18

Figure 19

123

124

Mehta and Melhem

Figure 20

directions, results in a phase value. As can be seen from Figure 20, the x components of the phase cancel, while the y components are additive. Notice that the result of the final calculated phase change demonstrates a y-component magnitude that is maximized only in certain conditions. If the phase shift of the moving protons after each bipolar gradient is 908, the shift has the greatest magnitude after subtraction, and hence the greatest signal intensity. This particular phase shift correlates with a given velocity—a parameter encoded by the MR technologist—termed ‘‘VENC.’’ Velocities less than this optimal encoded velocity, VENC, have phase shifts less than 908 (and greater than 08) for the first acquisition and phase shifts between 08 and 908 for the second acquisition. These lower velocities have correspondingly lower signal intensities. Velocities greater than VENC have phase shifts greater than 908 and less than 1808 and also have less signal intensity. Any phase shift less than, or greater than, 908 results in less signal, which is why the proper velocity encoding is critical in PCA. For spins moving in the opposite direction, the phase shift with the velocity corresponding to 908 has the greatest negative signal. The typical VENC for arterial flow is around 60 cm/sec, and for venous flow around 20 cm/sec. The PCA acquisition therefore requires that phase be utilized. In the vast majority of MR studies, magnitude images are primarily used. So how does one go about measuring phase? With quadrature coils, signal is measured in both the positive x axis and positive y axis. The signal in the former is the ‘‘real’’ (also known as the ‘‘in-phase’’ or ‘‘I’’) component, while the signal in the latter is the ‘‘imaginary’’ (also known as the ‘‘in-quadrature’’ or ‘‘Q’’) component. With both components, one can calculate either a magnitude image or a phase image. The magnitude data are commonly used in the majority of imaging sequences, including the TOF sequences. The phase data are used in PCA. With the acquisition of phase data encoded with a positive bipolar gradient along the x dimension subtracted from a second set of phase data encoded with a negative bipolar gradient along the x dimension,

stationary spins with the same phase get subtracted out, and moving spins with phase changes are displayed in the resulting image. When PCA was initially introduced, two sequences were performed along each axis for a total of six acquisitions. Positive and negative bipolar gradients were applied along each of the three dimensions (x, y, and z axes), which would result in a full phase contrast angiogram with phase contrast data in the x, y, and z dimensions, including data from the stationary tissue (Fig. 21). The number of acquisition has been decreased to 4 using the Hadamard multiplex flow-encoding approach, which yields the same data set. Given the additional time for the application of the bipolar gradients, TEs can be slightly longer than for TOF sequences. As in TOF, shorter TEs are desired to minimize artifacts from phase dispersion. Subtraction of two data sets allows for increased vessel conspicuity. Hence, even though blood saturation can occur with PCA, the increased conspicuity allows for a shorter TR (24 milliseconds at HUP, compared with 48 milliseconds for TOF at HUP). In general, maximizing TOF effects will also help maximize phase contrast, as phase contrast is dependent on both inflow effects and phase effects.

CLINICAL APPLICATIONS Intracranial Aneurysms An estimated 10 to 15 million people in the United States have intracranial aneurysms (22). Aneurysms that have come to the attention of physicians have primarily presented themselves in the form of subarachnoid hemorrhage (SAH). These ruptured aneurysms tend not to be imaged by MRA, as the presentation is acute and conventional angiography is performed expeditiously. At most institutions, patients have SAH detected by either unenhanced CT scan of the head or a positive lumbar puncture. Typically, patients continue on to CT angiography and finally conventional angiography. Assessment of the aneurysm at the time of conventional angiography

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

125

Figure 21 Phase contrast data at the level of the A1 and M1 segments (Fig. 20A) and the P1 segments (Fig. 20B). Note the bright signal in the vessels flowing toward the subject’s left and the dark signal in the vessels flowing toward the subject’s right, corresponding to the direction of flow.

most often determines whether the treatment will proceed with coiling or clipping. As can be seen, MRA has little role in these patients on presentation with SAH. It is, however, within the purview of MRA to play a role in the follow-up of patients who have had aneurysm coiling. Typically, these patients return for follow-up conventional digital subtraction angiography (DSA) to assess for coil compaction, residual neck, and parent vessel patency. Approximately 1% to 4% of patients who have had coiling may rebleed if no follow-up is performed, making follow-up imaging essential. 3D TOF MRA performed at 1.5 tesla has

been reported to have sensitivities of 75% to 92% in the detection of residual aneurysm neck (Figs. 22–24). The addition of gadolinium has been reported to increase sensitivity to 100%, with 96% specificity at 12 months postcoiling. These initial results demonstrate considerable promise in the future of MRA, particularly contrast-enhanced MRA, in the followup of coiled aneurysms (23–25). The second arena where MRA has a particularly important role to play is in the detection of unruptured aneurysms. Unruptured aneurysms (Figs. 25–27) can initially be brought to attention from screening MRA.

Figure 22 Coiled anterior communicating artery aneurysm with recanalization—discovered on MRA and confirmed on DSA. Abbreviation: DSA, digital subtraction angiography. Source: DSA courtesy of Dr. Mikolaj Pawlak, Department of Neuroradiology, University of Pennsylvania, Pennsylvania, U.S.A.

126

Mehta and Melhem

Figure 23 Coiled right supraclinoid ICA aneurysm with recanalization on 3D TOF MRA. Note the right carotid intracranial stent. Abbreviations: ICA, internal carotid artery; TOF, time-of-flight.

Figure 24 Recanalization in a coiled giant aneurysm.

There is controversy pertaining to the screening of asymptomatic individuals with a first-degree relative to a ruptured intracranial aneurysm (26,27). Screening is also considered for patients with preexisting conditions (e.g., adult polycystic kidney disease, fibromuscular dysplasia, collagen-vascular disease) that predispose to aneurysm formation, an area where noninvasive MRA can play an important role without the need for invasive conventional angiography. In the detection of aneurysms, studies have shown sensitivities for MRA ranging from 55% to 75% in comparison with conventional DSA. The sensitivity of the MR study can vary depending on reader experience, type of postprocessing algorithm applied, as well as the size of the aneurysm. Postprocessing algorithms typically

include MIPs; however, multiplanar reconstructions (MPRs) have been shown to be useful, particularly in the characterization of the aneurysm neck. MPRs are generated by reformatting the source data in any desirable plane, which then allows for the evaluation of the data set in coronal, sagittal, and oblique planes, in addition to the traditional axial acquisition. In addition to postprocessed data, source data must also be meticulously reviewed. Aneurysms less than 5 mm tend to be difficult to detect on MRA. One study found the sensitivity for detection to be 55% for aneurysms of 2 to 5 mm, whereas for aneurysms greater than 5 mm, sensitivity has been reported to be 88%. Systems of 3.0 tesla are also now coming into clinical practice with the FDA clearance in 1999, promising increased

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

Figure 25

signal-to-noise and contrast-to-noise ratios. Although the detection rate of aneurysms has not yet been shown to improve with 3.0 tesla systems compared with 1.5 tesla, increased image quality has been reported with improved aneurysm characterization (28–33).

Carotid Stenosis Carotid stenosis is a common indication for MRA of the neck vasculature. Traditionally, DSA has been the gold standard for depicting carotid stenosis. Benefit was shown in the NASCET trial for the treatment of symptomatic carotid stenosis greater than 70%.

127

Clearly, depicting the degree of stenosis is critical in the proper management of these patients. The 2D and 3D TOF techniques have been commonly used for imaging neck vasculature. In carotid stenosis, the blood that flows distal to the stenotic segment tends to be turbulent. Typically, the greater the degree of stenosis, the greater the turbulence, which in turn leads to signal loss on TOF imaging secondary to intravoxel dephasing. In fact, the presence of a flow void on 2D TOF MRA of the carotids has been demonstrated to have a positive predictive value of 84% for high-grade stenosis (greater than 70%). Unfortunately, the turbulence factor does lead to an overestimation of the degree of stenosis. This factor is rather critical in the case of patients who have intermediate stenosis but whose MRA overestimates it as greater than 70%. Hence, 2D TOF MRA has made its role in screening for stenosis rather than in becoming a substitute for DSA. The addition of gadolinium to the imaging algorithm has held additional promise in the characterization of stenosis. There are increased intravascular signal, shorter acquisition times, and TEs, as well as a high contrast-to-noise ratio compared with TOF MRA. With the increased intravascular signal from gadolinium, which can be substituted for a decrease in image voxel size, the effects of turbulence, and hence intravoxel dephasing, are minimized. The increased speed of scanning also helps with the minimization of motion artifacts. As in TOF MRA, gadolinium-enhanced MRA of the carotid vessels has a high sensitivity for the detection of high-grade stenosis (93–97%). However, this result is similar to the sensitivities of 3D TOF MRA without gadolinium. In intermediate stenoses, gadolinium has not been particularly helpful, with sensitivities ranging between 14% and 68% (Fig. 28) (33–40). MRI shows considerable promise in the characterization of arteriosclerotic plaque. In addition to assessing luminal stenosis with MRA, MRI has the ability to distinguish the various components of arteriosclerotic plaque and assess for high-risk lesions. The thickness of plaques is typically on the order of

Figure 26 Pseudoaneurysm secondary to dissection of PICA. Digital subtraction angiogram on the left, MRA on the right. Abbreviation: PICA, posterior inferior cerebellar artery.

128

Mehta and Melhem

Figure 27

millimeters; hence high-resolution MRI with phasedarray neck surface coils is necessary to achieve the spatial resolution required to separate intraplaque components. Important components that can be identified with MRI include the fibrous cap and the lipid core. Vulnerable plaques are thought to have thinned or ruptured fibrous caps and/or large lipid cores. The lipid core is typically of intermediate signal intensity on T1-weighted sequences and hypointense on T2-weighted sequences. However, the fibrous cap is

not readily identified on standard MRI sequences. Reports have been made of a hypointense band corresponding to the fibrous cap on TOF sequences. This band lies between the hyperintense arterial lumen and the isointense lipid core. The fibrous cap on hemorrhagic plaque can also be seen, as the hemorrhagic components can be T1 hyperintense (Fig. 29). One group has proposed the following in terms of fibrous cap characterization: thick and intact, thin and intact, ruptured. The thick and intact fibrous cap consists of a

Figure 28 Source: DSA courtesy of Dr. Mikolaj Pawlak, Department of Neuroradiology, University of Pennsylvania, Pennsylvania, U.S.A.

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

129

T1 and T2 sequences. Utilizing these criteria, the thin or ruptured fibrous cap was shown to have been associated with a recent transient ischemic attack and stroke (41–44).

Dissection

Figure 29 Hemorrhagic atherosclerotic plaque on TOF axial image. Between the hyperintense lumen of the internal carotid artery and hyperintense hemorrhage within the plaque, a thin dark fibrous cap can be seen. Source: Courtesy of Dr. Ronald Wolf, Department of Neuroradiology, University of Pennsylvania, Pennsylvania, U.S.A.

uniform hypointense band on TOF sequences. The thin and intact fibrous cap consists of nonvisualization of the hypointense band, but a smooth lumen surface of T1- and T2-weighted sequences. The ruptured fibrous cap consists of nonvisualization of the hypointense band and an irregular lumen surface on

Arterial dissections in the cervical carotid and vertebral arteries can either arise spontaneously or occur after a traumatic event. In both situations, MRA has supplanted conventional angiography in the initial diagnosis. The major findings to identify on MRA are a double lumen with an intimal flap or an intramural hematoma within the vessel wall. The hematoma is immediately isointense to slightly hyperintense to muscle during the first few days of a dissection (Fig. 30). Gradually, it becomes hyperintense and can remain so for months after the dissection. A fat-saturated T1-weighted sequence is also helpful for identifying the hyperintense signal of the intramural hematoma and subtract out the T1-hyperintense fat surrounding the artery. One can also see complications of dissection, such as pseudoaneurysm formation (Figs. 26 and 31) (45).

Intracranial Vascular Malformations Intracranial arteriovenous malformations (AVMs) are high-flow vascular lesions characterized by dysregulated angiogenesis. The lesions tend to form during development, with the vast majority being sporadic in occurrence and only 2% being part of a syndrome. Pathologically, these lesions demonstrate three major components: feeding arteries, a nidus, and draining veins. The feeding arteries are mature vessels that may or may not be enlarged. The arteries supply the AVM nidus, which is composed of numerous dysplastic,

Figure 30 Right carotid dissection. Fat-saturated T1-weighted image on the right, 3D TOF MRA in the middle, and MIP on the right. Note the T1 bright intramural hematoma in the fat-saturated sequence. There is high signal in the lumen and in the hematoma on the TOF sequence (and MIP). Abbreviations: TOF, time-of-flight; MIP, maximum intensity projection.

130

Mehta and Melhem

Figure 31

thin-walled vessels. These vessels are direct arteriovenous shunts, without intervening capillary network. Also note that no brain parenchyma is located within the nidus. The nidus subsequently empties into enlarged draining veins. Occasionally, there are associated prenidal, nidal, and postnidal aneurysms.

The imaging of these lesions has traditionally been the domain of conventional DSA. A conventional MRI demonstrates AVMs as multiple flow voids, with variable amounts of associated hemorrhage. MRA has been used to characterize the morphology of the AVM (Fig. 32). The traditional 3D TOF MRA can demonstrate high-flow feeding vessels; however, the technique is not very sensitive to slow-flowing draining veins. Using contrast-enhanced MRA with very short acquisition times (TR/TE 5/2 milliseconds) does improve the visualization of draining veins as well as feeding arteries (46). Unfortunately, current MRA techniques do not provide dynamic information like conventional DSA. Magnetic resonance digital subtraction angiography (MR DSA) is showing promise in providing this information. The technique essentially involves obtaining multiple fast T1-weighted acquisitions during the administration of gadolinium. The initial, precontrast image is used as a mask, which is then subtracted from the contrast-enhanced sequences. When performed with the 2D technique with a thick slice, on the order of 6 to 10 cm, an image is acquired approximately every 1.05 seconds to provide the necessary temporal resolution (47,48). As this is a 2D technique, spatial resolution is not as optimal as with the 3D technique. The recent use of keyhole imaging and SENSE has allowed for the 3D technique with MR DSA, acquiring an image every 1.7 seconds (49). Indeed, recent application of 3D MR DSA technique in the evaluation of residual AVM after radiosurgery demonstrated a sensitivity of 81% and a specificity

Figure 32 AVM on 3T TOF MRA. The image on the left is a MIP collapse, the central image is a source MRA slice, and the image on the right is FLAIR weighted showing flow voids. Abbreviations: AVM, arteriovenous malformation; MIP, maximum intensity projection; FLAIR, fluid-attenuated inversion-recovery.

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

of 100% in the identification of a nidus or draining vein, compared with DSA (50). As opposed to intracranial AVMs, dural arteriovenous fistulas (DAVFs) are thought to be acquired later in life, not during development. Although not completely understood, theories suggest that venous sinus thrombosis forms and triggers angiogenesis for recanalization. The angiogenesis then results in one of three types of DAVFs. The three classifications are based on venous drainage patterns, with type 1 draining anterograde into the venous sinus itself, type 2 draining reterograde into the venous sinus (2a) and retrograde into the subarachnoid/leptomeningeal veins (2b), and type 3 draining solely retrograde into the subarachnoid/leptomeningeal veins (51). Detection of DAVF by MRI is challenging, with one study finding that a majority of DAVFs were characterized by flow void clusters around a dural sinus (52). Type 2 and 3 lesions also tend to demonstrate dilated leptomeningeal and/or medullary veins. 3D TOF MRA has a 45% sensitivity for directly demonstrating the fistulas and a 91% sensitivity for detecting flow-related enhancement in draining veins. Dynamic MRA sequences also show promise in characterizing these lesions (Fig. 33).

Spinal Vascular Malformations Spinal vascular malformations are lesions that often require an exhaustive angiographic search with catheterization of multiple radicular arteries to help localize the malformations. There are four types of spinal vascular malformations, the most common (80%) being type 1 spinal DAVFs. These are direct arterial to venous connections located peripherally within the dura of a nerve root sleeve. The lesions receive supply from a dural branch of the radicular artery, with drainage into the cord pial veins. These type 1 lesions are best thought of as extramedullary, intradural, peripheral AV fistulas. Type 2 lesions are true arteriovenous malformations located within the

131

cord itself. These intramedullary lesions are supplied by branches of the anterior or posterior spinal artery and typically have a compact nidus, without intervening parenchyma. Type 3 lesions are more complex than type 2 lesions and have both intramedullary and extramedullary components. Type 4 lesions are extramedullary arteriovenous fistulas, like type 1 lesions; however, they are centrally located within the subarachnoid space. Their arterial supply arises from either the anterior or posterior spinal artery—and there is no nidus—with direct drainage into the spinal veins. When patients present with a suspected spinal vascular malformation (e.g., progressive myelopathy), they initially undergo an MRI evaluation of the spine. Conventional MR imaging shows T2 hyperintensity within the cord, with multiple flow voids from engorged venous structures. Unfortunately, the site of T2 hyperintensity in the cord, representing cord edema from venous hypertension and resultant cord ischemia, does not correspond to the level of a malformation. Venous flow voids may or may not be seen. As the majority of spinal vascular malformations are type 1, intramedullary flow voids are not reliable. Conventional angiography is then used to painstakingly evaluate multiple spinal levels in the hope of discovering the vascular lesion. Spinal MRA has reduced this painstaking search and assisted in the targeting of these lesions. A marked improvement has been reported in the true-positive detection rate of these lesions with 3D contrast-enhanced MRA, with the true positives improving from 15% with MRI data alone to 50% with combined MRI and MRA data (53). Recently, 3D contrast-enhanced MRA with an elliptic centric filling of k space was reported to have correctly localized spinal vascular malformations in eight out of nine patients (Fig. 34) (54). The elliptic centric algorithm fills the central portion of k space in the first one-tenth of the total scan time. The central k space determines contrast resolution, while the peripheral k space determines spatial resolution; hence, arterial contrast is maximized with elliptic centric filling.

Figure 33 DAVF on dynamic gadolinium-enhanced MRA. From left to right: precontrast, arterial phase, delayed phase. The filling of leptomeningeal veins and transverse sinus is immediately evident. Abbreviation: DAVF, dural arteriovenous fistula.

132

Mehta and Melhem

Figure 34 Spinal DAVF demonstrated at the T11 level by 3D contrast-enhanced MRA with an elliptic centric filling of k space. The level of the arterial feeder was subsequently confirmed on DSA and embolized. Abbreviations: DAVF, dural arteriovenous fistula; DSA, digital subtraction angiography.

Ischemic Stroke Ischemic stroke is one of the most common causes of morbidity and mortality. The five-year mortality rate for carotid territory infarction is 40% (55). Imaging has taken on an ever-increasing role in the diagnosis of ischemic stroke. Perhaps more enticing is the prospect of using imaging to help triage patients to proper therapy. Intense research efforts are attempting to better define ischemic penumbra and stratification for thrombolytic therapies. The pathophysiology of ischemic stroke essentially boils down to decreased cerebral blood flow (CBF). When CBF decreases to below 10–15 mL/100 g of brain parenchyma per minute, neuronal death ensues. Conventional MRI sequences initially demonstrate loss of flow void in the occluded vessel and T2-hyperintense signal in regions corresponding to ischemic tissue. These findings are analagous to the CT findings of hyperdense arterial attenuation from thrombus and associated hypodensity in the affected brain parenchyma. However, the true revolution in MRI of stroke came with the advent of diffusionweighted imaging (DWI) sequences. DWI sequences essentially measure the freedom of water. If a water molecule is within an environment where it can move freely in all three dimensions (e.g., outside of a cell), then it has quite a bit of freedom, and hence increased diffusivity. With the use of diffusion gradients, MRI can measure the diffusivity of water, creating maps of the apparent diffusion coefficient (ADC). Areas of the brain that have restricted diffusivity are those where water is primarily intracellular, not extracellular. Hypoxic neurons have failure of their sodium/potassium ATP ion pumps, leading to an influx of water into the cell, leading to restricted diffusion. DWI, in animal models, has the capability of imaging infarction as early as five minutes on the onset of ischemia. In humans, DWI abnormalities have been detected as soon as 39 minutes after the onset of stroke (56). Restricted diffusion then typically returns to normal by day 7.

While DWI has served to image metabolic changes occurring with a stroke, recent efforts have focused on perfusion. One of the observations made with DWI is that the area of restricted diffusion can enlarge over time following the onset of ischemia. The growth of this hypoxic region of tissue suggests an atrisk tissue not initially identified by DWI. Perfusionweighted imaging (PWI) sequences came into the picture to help identify this at-risk tissue, also known as the ischemic penumbra (57). PWI sequences are typically performed by administering a bolus of gadolinium, followed by rapid sequential images acquired as the bolus traverses the cerebral circulation from artery to vein. As gadolinium is paramagnetic, the T2* is shortened and is used to track the bolus and assess its associated signal changes in the vasculature and parenchyma. PWI sequences result in maps of CBF, cerebral blood volume (CBV), and mean transit time (MTT) (Fig. 35). These three variables are related by the central volume principle, according to which CBF ¼ CBV/MTT. Ischemic tissue has decreased CBF; decreased, normal, or elevated CBV depending on the degree of vascular reserve; and elevated MTT. Infarcted tissue has decreased CBF, decreased CBV, and elevated MTT. It is thought that elevated MTT may be the most sensitive indicator of brain tissue at risk for infarction. Another MR perfusion parameter, time to peak (TTP), which is delayed in ischemic tissue, can be used to identify the ischemic penumbra. On the assumption that the tissue with a restricted diffusion is infarcted, one can then subtract DWI data from the MTT map. The final image is that of the ischemic penumbra, the tissue that is presumably ischemic but not yet irreversibly infarcted. This is the tissue that is targeted for salvage with thrombolytics. There is still controversy as to whether all the tissue with decreased CBF will proceed toward infarction or not (58). There is also controversy as to whether the tissue identified on DWI represents infarcted tissue, as observations have shown that thrombolytic therapy can decrease the volume of DWI abnormality. Hence, current measurements of

Chapter 6: MR Angiography: Principles and Applications in Interventional Neuroradiology

133

Figure 35 The top right image is T2 weighted, top central and top left are DWI and ADC, and bottom row is perfusion data. Note how perfusion abnormalities encompass a large portion of the right MCA territory compared with the DWI images, which only show focal abnormalities in the right external capsule and right parietal lobe. The difference between perfusion and diffusion data is the ischemic penumbra. Abbreviations: DWI, diffusion-weighted imaging; ADC, apparent diffusion coefficient.

ischemic penumbra can be thought of as an overestimate of tissue that will proceed to nonsalvageable infarction without intervention. Also, the current measurements of DWI abnormality can be considered an overestimation of nonsalvageable infarcted tissue.

REFERENCES 1. Suryan G. Nuclear resonance in flowing liquids. Proc Indian Acad Sci 1951; A33:107–113. 2. Singer J. Blood flow rates by nuclear magnetic resonance measurements. Science 1959; 130:1652–1653. 3. Dumoulin C, Hart H. Magnetic resonance angiography. Radiology 1986; 161:717–720. 4. Keller P, Drayer B, Fram E, et al. MR angiography with two-dimensional acquisition and three-dimensional display. Radiology 1989; 173:527–532. 5. Suri JS, Liu K, Reden L, Laxminarayan S. A review on MR vascular image processing algorithms: acquisition and prefiltering: part I. IEEE Trans Inf Technol Biomed 2002; 6(4):324–337. 6. Ozsarlak O, Van Goethem JW, Maes M, et al. MR angiography of the intracranial vessels: technical aspects and clinical applications. Neuroradiology 2004; 46:955–972. 7. Felmlee J, Ehrman R. Spatial presaturation: a method for suppressing flow artifacts and improving depiction of vascular anatomy in MR imaging. Radiology 1987; 164:559–564. 8. Rossnick S, Laub G, Braeckle R. Three dimensional display of blood vessels in MRI. Proceedings of the IEEE computers in cardiology. New York: Institute of Electrical and Electronic Engineers, 1986:193–195. 9. Parker D, Yuan C, Blatter D. MR angiography by multiple thin slab 3D acquisition. Magn Reson Med 1991; 17: 434–451. 10. Purdy D, Cadena G, Laub F. The design of variable tip angle slab selection (TONE) pulses for improved 3D MR angiography. Berlin: The Society of Magnetic Resonance in Medicine, 1992:882. 11. Edelman R, Ahn S, Chien D, et al. Improved time-of-flight MR angiography of the brain with magnetization transfer contrast. Radiology 1992; 184:395–399.

12. Parker D, Goodrich C, Roberts J, et al. The need for phaseencoding flow compensation in high-resolution intracranial magnetic resonance angiography. J Magn Reson Imaging 2003; 18:121–127. 13. Isoda H, Takehara Y, Isogai S, et al. Technique for arterialphase contrast-enhanced three-dimensional MR angiography of the carotid and vertebral arteries. AJNR Am J Neuroradiol 1998; 19:1241–1244. 14. Weiger M, Preussmann K, Kassner A, et al. Contrastenhanced MRA using SENSE. J Magn Reson Imaging 2000; 12:671–677. 15. Leclerc X, Nicol L, Gauvrit J, et al. Contrast-enhanced MR angiography of supraaortic vessels: the effect of voxel size on image quality. AJNR Am J Neuroradiol 2000; 21:1021–1027. 16. Aoki S, Nakajima H, Kumagai H, et al. Dynamic contrastenhanced MR angiography and MR imaging of the carotid artery: high-resolution sequences in different acquisition planes. AJNR Am J Neuroradiol 2000; 21:381–385. 17. Yang C, Carr J, Futterer S, et al. Contrast-enhanced MR angiography of the carotid and vertebrobasilar circulations. AJNR Am J Neuroradiol 2005; 26:2095–2101. 18. Ozsarlak O, van Goethem J, Parizel P. 3D time-of-flight MR angiography of the intracranial vessels: optimization of the technique with water excitation, parallel acquisition, eight-channel phased-array head coil and low-dose contrast administration. Eur Radiol 2004; 14:2067–2071. 19. Dumoulin C, Souza S, Darrow R. Simultaneous acquisition of phase-contrast angiograms and stationary tissue with Hadamard encoding of flow induced phase shifts. J Magn Reson Imaging 1991; 1:399–404. 20. Dumoulin C, Souza S, Walker M, et al. Three-dimensional phase contrast angiography. Magn Reson Med 1989; 9:139–149. 21. von Schulthess G, Higgins C. Blood flow imaging with MR: spin-phase phenomena. Radiology 1985; 157:687–695. 22. Kim D, Van Ginhoven G, Milewicz D. Incidence of familial intracranial aneurysms in 200 patients: comparison among Caucasian, african-american, and Hispanic populations. Neurosurgery 2003; 53:302–308. 23. Gauvrit J, Leclerc X, Pernodet M, et al. Intracranial aneurysms treated with Guglielmi detachable coils: usefulness of 6-month imaging follow-up with contrast-enhanced MR angiography. AJNR Am J Neuroradiol 2005; 26:515–521.

134

Mehta and Melhem

24. Westerlann H, van der Vliet A, Hew J, et al. Time-of-flight magnetic resonance angiography in the follow-up of intracranial aneurysms treated with Guglielmi detachable coils. Neuroradiology 2005; 47:622–629. 25. Boulin A, Pierot L. Follow-up of intracranial aneurysms treated with detachable coils: comparison of gadoliniumenhanced 3D time-of-flight MR angiography and digital subtraction angiography. Radiology 2001; 219:108–113. 26. Rinkel G. Intracranial aneurysm screening: indications and advice for practice. Lancet Neurol 2005; 4:122–128. 27. Jacoby G, Sams R II, Biagioli F. Should people with a firstdegree relative who died from subarachnoid hemorrhage be screened for aneurysms?. J Fam Pract 2006; 55(1):59–60. 28. Atlas S, Sheppard L, Goldberg H, et al. Intracranial aneurysms: detection and characterization with MR angiography with use of an advanced postprocessing technique in a blinded-reader study. Radiology 1997; 203:807–814. 29. Adams W, Laitt R, Jackson A. The role of MR angiography in the pretreatment assessment of intracranial aneurysms: a comparative study. AJNR Am J Neuroradiol 2000; 21:1618–1628. 30. Chung T, Joo J, Lee S, et al. Evaluation of cerebral aneurysms with high-resolution MR angiography using a section-interpolation technique: correlation with digital subtraction angiography. AJNR Am J Neuroradiol 1999; 20:229–235. 31. Suzuki M, Matsui O, Ueda F, et al. Contrast-enhanced MR angiography (enhanced 3-D fast gradient echo) for diagnosis of cerebral aneurysms. Neuroradiology 2002; 44:17–20. 32. Gibbs G, Huston J III, Bernstein M, et al. 3.0 Tesla MR angiography of intracranial aneurysms: comparison of time-of-flight and contrast-enhanced techniques. J Magn Reson Imaging 2005; 21:97–102. 33. Gibbs G, Huston J III, Bernstein M, et al. Improved image quality of intracranial aneurysms: 3.0 T versus 1.5 T timeof-flight MR angiography. AJNR Am J Neuroradiol 2004; 25:84–87. 34. Nederkoorn P, van der Graaf Y, Eikelboom B, et al. Timeof-flight MR angiography of carotid artery stenosis: does a flow void represent severe stenosis?. AJNR Am J Neuroradiol 2002; 23:1779–1784. 35. Serfaty J, Chirossel P, Chevallier J, et al. Accuracy of threedimensional gadolinium-enhanced MR angiography in the assessment of extracranial carotid artery disease. AJR Am J Roentgenol 2000; 175:455–463. 36. Randoux B, Marro B, Koskas F, et al. Carotid artery stenosis: prospective comparison of CT, three-dimensional gadolinium-enhanced MR, and conventional angiography. Radiology 2001; 220:179–185. 37. Cosottini M, Pingitore A, Puglioli M, et al. Contrastenhanced three-dimensional magnetic resonance angiography of atherosclerotic internal carotid stenosis as the noninvasive imaging modality in revascularization decision making. Stroke 2003; 34:660–664. 38. Nederkoorn P, Elgersma O, van der Graaf Y, et al. Carotid artery stenosis: accuracy of contrast-enhanced MR angiography for diagnosis. Radiology 2003; 228:677–682. 39. Remonda L, Senn P, Barth A, et al. Contrast-enhanced 3D MR angiography of the carotid artery: comparison with conventional digital subtraction angiography. AJNR Am J Neuroradiol 2002; 23:213–219. 40. Borisch I, Horn M, Butz B, et al. Preoperative evaluation of carotid artery stenosis: comparison of contrastenhanced MR angiography and duplex sonography with

41. 42. 43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55. 56.

57. 58.

digital subtraction angiography. AJNR Am J Neuroradiol 2003; 24:1117–1122. Choi CJ, Kramer C. MR imaging of atherosclerotic plaque. Radiol Clin N Am 2002; 40:887–898. Yuan C, Kerwin W. MRI of atherosclerosis. J Magn Reson Imaging 2004; 19:710–719. Hatsukami T, Ross R, Polissar N, et al. Visualization of fibrous cap thickness and rupture in human atherosclerotic carotid plaque in vivo with high-resolution magnetic resonance imaging. Circulation 2000; 102:959–964. Yuan C, Zhang S, Pollisar N, et al. Identification of fibrous cap rupture with magnetic resonance imaging is highly associated with recent transient ischemic attack or stroke. Circulation 2002; 105:181–185. Provenzale J. Dissection of the internal carotid and vertebral arteries: imaging features. AJNR Am J Neuroradiol 1995; 165:1099–1104. Duran M, Shoenberg S, Yuh W, et al. Cerebral arteriovenous malformations: morphologic evaluation by ultrashort 3D gadolinium-enhanced MR angiography. Eur Radiol 2002; 12:2957–2964. Tsuchiya K, Katase S, Yoshino A, et al. MR digital subtraction angiography of cerebral arteriovenous malformations. AJNR Am J Neuroradiol 2000; 21:707–711. Mori H, Aoki S, Okubo T, et al. Two-dimensional thickslice MR digital subtraction angiography in the assessment of small to medium-size intracranail arteriovenous malformations. Neuroradiology 2003; 45:27–33. Tsuchiya K, Aoki C, Fujikawa A, et al. Three-dimensional MR digital subtraction angiography using parallel imaging and keyhole data sampling in cerebrovascular diseases: initial experience. Eur Radiol 2004; 14:1494–1497. Gauvrit J, Oppenheim C, Nataf F, et al. Three-dimensional dynamic magnetic resonance angiography for the evaluation of radiosurgically treated cerebral arteriovenous malformations. Eur Radiol 2006; 16:583–591. Borden J, Wu J, Shucart W. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg 1995; 82:166–179. Kwon B, Han M, Kang H, et al. MR imaging findings of intracranial dural arteriovenous fistulas: relations with venous drainage patterns. AJNR Am J Neuroradiol 2005; 26:2500–2507. Saraf-Lavi E, Bowen B, Quencer R, et al. Detection of spinal dural arteriovenous fistulae with MR imaging and conrast-enhanced MR angiography: sensitivity, specificity, and prediction of vertebral level. AJNR Am J Neuroradiol 2002; 23:858–867. Farb R, Kim J, Willinsky R, et al. Spinal dural arteriovenous fistula localization with a technique of first-pass gadolinium-enhanced MR angiography: initial experience. Radiology 2002; 222:843–850. Thurnher M, Castillo M. Imaging in acute stroke. Eur Radiol 2005; 15:408–415. Yoneda Y, Tokui K, Hanihara T, et al. Diffusion-weighted magnetic resonance imaging: detection of ischemic injury 39 minutes after onset in a stroke patient. Ann Neurol 1999; 45:794–797. Abe O, Aoki S, Shirouzu I, et al. MR imaging of ischemic penumbra. Eur J Radiol 2003; 46:67–78. Kidwell C, Alger J, Saver J. Beyond mismatch: evolving paradigms in imaging the ischemic penumbra with multimodal magnetic resonance imaging. Stroke 2003; 34:2729–2735.

7 Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology Jaroslaw Krejza Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A., and Department of Nuclear Medicine, Medical University of Gdansk, Poland

INTRODUCTION

Bioeffects and Safety of US

Since the introduction of echoencephalography in the early 1950s, ultrasonic techniques have evolved dramatically. Ultrasound (US) imaging is now considered an integral part of the evaluation of patients with cerebrovascular disease (CVD), because it is noninvasive, relatively inexpensive, accurate, and readily accessible. This chapter provides the basics of US and summarizes diagnostic and therapeutic applications of US in interventional neuroradiology.

Diagnostic US can produce heat that may be hazardous to sensitive organs (3). Nonthermal effects, such as pressure changes and mechanical disturbances, in tissue have not been demonstrated in humans (4). US used in therapy, however, can cause both substantial temperature increase and mechanical damage in the tissue (4–6).

TECHNICAL ASPECTS OF US IMAGING Basics of US In US imaging, pulsed waves emitted by a transducer pass into the body and reflect off the boundaries between different tissue types. These reflections, or echoes from the reflected waves, are then assembled into a picture on a video monitor. The frequency, density, focus, and aperture of the US beam can vary. Higher frequencies produce more clarity but penetrate less deeply into the body. Lower frequencies penetrate more deeply but produce lower resolution, or clarity. US entering tissue may be transmitted, absorbed, reflected, and/or scattered (1). The transmission properties of a tissue depend on its density and elasticity. Density and speed of propagation determine a tissue’s acoustic impedance. In homogeneous tissues, US waves propagate until all their energy is dissipated as heat. In nonhomogeneous tissues, reflection, scattering, transmission, or a combination of these processes occurs when waves encounter a layer with different acoustic impedance. The larger the difference in acoustic impedance, the more the waves reflected (1,2). Reflection further depends on the angle of insonation, and stronger echoes are received when the angle of insonation is zero (2). Strongly reflective hyperechoic interfaces, such as bone or air, prevent imaging of weaker echoes from deeper tissue and cast an acoustic shadow. Hypoechoic or poorly reflective tissues, including fluids, are called sonolucent.

Thermal Effects

Generally, denser tissue absorbs more heat from US. Therefore, the fluid does not heat very much, soft tissues heat somewhat more, and bone heats the most. The heating rate in the bone surface can be up to 50 times faster than in soft tissues. This heating effect is of interest, particularly in regard to the transcranial Doppler ultrasonography (TCD) and its therapeutic applications. Diagnostic US systems now display numbers that provide crude measures of a risk to patients from the heat and/or mechanical effects. The thermal index (TI) is an estimate of risk from heat. When the TI is above 1, it is recommended that the risks of US be weighed against the benefits (4). The consensus is to minimize exposure, particularly in pulsed Doppler applications, as a significant temperature increase can occur at the bone–soft tissue interface. Nevertheless, short-term continuous TCD monitoring did not increase temperature at the temporal window in vivo (7). Nonthermal Effects

US can also produce various mechanical effects such as cavitation, pressure amplitude, force, torque, and acoustic streaming (3). Cavitation occurs when US passes through an area that contains a cavity, such as a gas bubble. US can cause the bubble to expand and contract rhythmically. When bubbles pulsate, they send secondary US waves in all directions. These secondary waves can actually improve US imaging. If the bubbles contract toward the point of collapsing, they can build up very high temperatures and pressures for a few tens of nanoseconds. These high temperatures and pressures can potentially

136

Krejza

produce free radicals and other toxic compounds that, although considered unlikely, could theoretically cause genetic damage (8). The rapid contraction of bubbles can also cause microjets of liquid that can damage cells. The safety guidelines for diagnostic US are designed to prevent cavitations to occur (3,4). Apart from cavitation, US produces changes in pressure, force, torque, and streaming. These changes, in turn, can cause audible sounds, electrical changes in cell membranes that make them more permeable to large molecules, movement and redistribution of cells in liquid, and cell damage (9). In liquids, US causes a type of stirring action called acoustic streaming. As the acoustic pressure of US increases, the flow of liquid speeds up. When the streaming liquid comes near a solid object, shearing may occur, which can damage platelets and lead to abnormal blood clotting. Effects of US Contrast Agents

These agents usually take the form of stable gas-filled microbubbles (MBs), which can potentially produce cavitations and/or microstreaming, the risk of which increases with the mechanical index (MI) value.

Diagnostic US Techniques Gray-Scale Imaging

B-mode, or brightness mode, displays structures proportionally to the intensity of returning echoes. The US beam is swept quickly through the field of view, and the image is continuously renewed, allowing a real-time visualization of the underlying tissue anatomy. In M-mode imaging, used to evaluate the motion of well-defined surfaces such as blood vessel walls, a vertical time-base trace driven from left to right across the display is simultaneously generated. The echoes are displayed vertically as the depth of US penetration increases. Doppler Display Modes and Blood Velocity Measurements

The difference in frequency between emitted and reflected ultrasonic echoes is the Doppler frequency shift. The magnitude of the shift depends on the US transmission velocity in the tissue (C), the relative velocity of blood (V), and the US emitted frequency (Fo). The observed frequency shift (DF) is expressed as DF ¼ 2VFo/C. The shift is measured only for the component of motion along the axis of the US beam. Therefore, absolute velocity measurements require that a correction be made for the angle () between the vessel and the beam as follows: V ¼ DFC/(2Fo cos ). Doppler modes are used to measure flow velocity. The frequency shift is proportional to the velocity of moving blood. The simplest Doppler US instrument has two identical piezoelectric crystal transducers. One crystal continuously emits toward the region of interest, and the other continuously receives reflected echoes. Flow toward the transducer produces an increase in the received frequency, whereas flow away from the transducer causes a drop. Continuous Doppler systems can

measure a wide range of velocities but provide no information about the depth of the reflecting tissue, because any moving object in the beam’s pathway reflects echoes. The depth or position insensitivity of continuous-wave Doppler is overcome to a large extent by using pulsed-wave Doppler. In this technique, a single transducer generates US pulses and detects returning echoes. Assuming that the speed of transmission of US in tissues is a constant, the time delay between the emitted pulse and the reflected echo enables the sampled structure’s depth to be determined. However, anatomy is not displayed, and the pulse duration and repetition frequency impose limits on the maximum velocity that can be measured. This technique is used for conventional TCD. Anatomy is displayed in duplex imaging, which combines pulsedwave Doppler with two-dimensional real-time grayscale imaging. The gray-scale image of a selected vessel is displayed, allowing precise placement of the Doppler sample volume in the vessel to measure flow velocity throughout the cardiac cycle. Optimal angle correction for velocity calculations can be performed as the course of the vessel in relation to the US beam is visually depicted. Color duplex is the most commonly used technology today for extracranial carotid imaging. It is also used for transcranial color-coded duplex sonography (TCCS). Color is superimposed on a conventional gray-scale image to enhance the image of the Doppler frequency shift. Red indicates flow toward the transducer, whereas blue represents flow away from the transducer. High flow velocities are depicted with increasing brightness. As a result, the presence of flow, its direction, and hemodynamic disturbances can be quickly assessed. The color map in color Doppler US can be displayed as the integrated power of the Doppler signal, which is related to the number of red blood cells that produce the Doppler shift. Advantages of this power mode include independence from the angle of insonation, absence of aliasing, and the ability to detect very low flows. Piezoelectric crystals are arranged into an array inside a transducer. Linear transducers (7.5–16 MHz) for carotid imaging produce rectangular fields of view, while phased-array transducers (1–3.5 MHz) used in TCCS produce wedge-shaped fields of view.

DIAGNOSTIC US IMAGING IN INTERVENTIONAL NEURORADIOLOGY TCD Since the early 1980s, TCD imaging has permitted insonation of the basal brain arteries. TCD technology substantially evolved during the mid to late 1990s, and TCCS is increasingly used today. Both TCD and TCCS have specific advantages. TCD is based on pulsedwave Doppler measurements of blood flow velocity. Its 2-MHz, relatively small, transducers are easy to use, particularly when prolonged monitoring is performed. Experience with this technology is extensive, but the angle of insonation cannot be assessed and exact placement of a sample volume in the insonated artery cannot be controlled, leading to an error in both velocity

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

measurement and vessel identification. TCCS combines two-dimensional real-time gray-scale imaging with pulsed-wave Doppler and color-coded display of velocity information (10). It is performed with phased-array, 1.6- to 3.5-MHz transducers that are slightly larger and less easy to manipulate than their

137

TCD counterparts. In contrast to TCD, however, TCCS enables the sonographer to outline intracranial parenchymal structures, to acquire a Doppler sample at a specific site of an insonated artery, and to image segments of the basal cerebral arteries in color (Fig. 1). These advantages permit more rapid studies, provide

Figure 1 (A) Typical color image of the M1 segment of the right MCA superimposed on a sector-shaped conventional gray-scale image. The sample volume is precisely placed on a green color spot related to an aliasing artifact, which indicates the site of highest flow acceleration in the segment, and the angle between the course of the vessel in relation to the US beam is measured by an electronic cursor. This approach allows to obtain the angle-corrected flow velocity measurements from the waveform displayed below the gray-scale image. In this 57-year-old patient, the follow-up TCCS study, a year after clipping of MCA aneurysm, shows blood flow velocities in the right M1 MCA (A), M2 MCA (B), A1 ACA (C), and P1 PCA (D) within normal reference range of 110 cm/sec, 48 cm/sec, 71 cm/sec, and 58 cm/sec, respectively. Angiographic study (E) performed at the same day showed normal caliber of the vessel. (F) Shows complex spatial relationship between US beam, courses of M1 MCA and A1 ACA, and the site of temporal window. Abbreviations: MCA, middle cerebral artery; US, ultrasound; TCCS, transcranial color-coded duplex sonography; ACA, anterior cerebral artery; PCA, posterior cerebral artery.

138

Krejza

more information, and improve the sonographer’s confidence as well as the test’s accuracy (11). Both TCD and TCCS are noninvasive and enable bedside testing. Measurements are also highly reproducible. Inadequate ultrasonic windows, present in 10% to 20% of patients (12), and limited accuracy constitute the major disadvantages. Reference Values

Reference values have been established for both TCD and TCCS (13,14). Because of differences in correction of insonation angles and depth of insonation, TCD reference values cannot be used for TCCS measurements. Normal reference data for TCD and TCCS velocities have been published elsewhere, and are presented in Table 1 for TCCS (13,14). Velocities are highest during the first decade of life and drop during the fifth and sixth decades. Women tend to have higher velocity values up to the age of 60. This tendency may be partially explained by the effect of the hormonal fluctuations that affect the reactivity and tone of the cerebral vasculature (15,16). Other factors that affect flow velocities include intracranial pressure, Hct, fibrinogen, cardiac rhythm disorders, and medications (17,18). Referring of blood flow velocities obtained from a patient to the age and sex increases accuracy of the TCCS in detecting flow abnormalities (19).

Detection of Intracranial Arterial Occlusion For current acute ischemic stroke therapy to be effective, it must be initiated in the first few hours after stroke (20,21). The only current proven therapy for acute ischemic stroke is thrombolysis with tissue plasminogen activator (tPA) within three hours of stroke (21,22). While this treatment appears to be effective in all major ischemic stroke subtypes, recent trials have suggested that some therapies may offer a benefit for one mechanism but not for others (23,24). If the stroke mechanism could be determined in the first few hours after stroke, then patients with specific subtypes could be selected for specific potential therapies in clinical trials and ultimately in clinical practice. The clinical diagnosis of stroke subtype during the

first 24 hours is frequently inaccurate, since clinicians often rely only on the history, physical examination, noncontrast CT, and ECG. Thus, for mechanismdirected therapy to be implemented, additional diagnostic information is required (Fig. 2). In the setting of acute stroke, rapid TCD testing can be an attractive approach to early stroke subtype diagnosis that subsequently influences patient management (25). Early diagnosis of an acute large intracranial artery occlusion with TCCS is made on the basis of the absence of Doppler signal in the artery (26). The suitability of the acoustic window, however, must be proven by the visualization of at least one ipsilateral cerebral artery (26,27). Flow disturbances in other intracranial arteries can further increase the diagnostic accuracy of TCCS. Occlusion of the M1 MCA, for example, is frequently associated with increased velocities in the ipsilateral ACA due to the increased flow through leptomeningeal collaterals (28,29). Carotid occlusion leads to the development of collateral flow through the ophthalmic artery and the anterior (ACoA) and posterior (PCoA) communicating arteries, while basilar artery (BA) occlusion increases the flow through PCoA (29,30). Intracranial occlusion of the vertebral artery (VA), located proximal to the origin of the posterior inferior cerebellar artery, may lead to reverse flow in the ipsilateral distal VA. Occlusions of the intracranial ICA, VA, and BA reduce upstream velocities, except in the BA, if adequate collateral flow through cerebellar arteries is present. If the results of the TCCS study are inconclusive, MR angiography (MRA) or CT angiography (CTA) can be used for making the diagnosis of intracranial occlusion. Perfect sensitivity, specificity, PPV, and NPV of TCCS in diagnosis of M1 MCA occlusion using predefined criteria were found in a study of 30 patients with ischemic stroke of less than 24 hours’ duration (31). Another study has shown that MCA occlusions located in the main stem or branches in 20 of 23 patients with acute ischemic stroke of less than five hours’ duration can be rapidly (5–7 minutes) detected using contrast-enhanced TCCS (32). Other authors, on the basis of small series, also suggest high reliability of

Table 1 Mean and Normal Reference Limits of Vps, Vmean, and Ved Blood Flow Velocities in MCA, ACA, and PCA Cerebral Arteries Arteries Velocities (cm/sec) All subjects

Women

Men

N

MCA

ACA

PCA

Vps Vmean Ved

304

105 (52–166) 68 (32–112) 45 (17–77)

76 (34–121) 50 (18–82) 33 (10–57)

69 (37–103) 46 (21–72) 30 (11–51)

Vps Vmean Ved

193

107 (48–168) 71 (31–115) 47 (17–75)

77 (39–124) 51 (23–83) 33 (11–57)

70 (40–107) 47 (24–74) 31 (12–51)

Vps Vmean Ved

111

100 (54–158) 64 (31–102) 43 (17–72)

74 (32–123) 48 (16–82) 32 (8–57)

67 (35–104) 45 (20–73) 30 (12–53)

Abbreviations: Vps, peak-systolic velocity; Vmean, mean velocity; Ved, end-diastolic velocity; MCA, middle cerebral artery; ACA, anterior cerebral artery; PCA, posterior cerebral artery.

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

139

Figure 2 Identification of arteries of the circle of Willis with TCCS enables for an operator to detect an isolated occlusion. (A) Angiography shows not patent A1 ACA on the right side. Also TCCS shows no flow in the A1, while flow is clearly seen in the MCA (B), ICA (C), and PCA (D). In contrast, the A1 ACA was erroneously identified in three sequential conventional TCD studies in this patient. Abbreviations: TCCS, transcranial color-coded duplex sonography; ACA, anterior cerebral artery; MCA, middle cerebral artery; ICA, internal carotid artery; PCA, posterior cerebral artery; TCD, transcranial Doppler ultrasonography.

contrast-enhanced TCCS in detection of M1 MCA occlusion (33,34). Moreover, high sensitivity (94%) and specificity (93%) of contrast-enhanced TCCS were reported in 30 patients with ischemic stroke within 12 hours after symptom onset (35). The accuracy of TCCS/TCD in detection of occlusion of M2 MCA has not yet been studied. A TCD study, however, has shown that occlusion of more than three MCA branches is associated with decreased velocities in M1 MCA (36). Another TCD study has investigated the diagnostic accuracy of intracranial occlusion assessment using predefined criteria (37). The corresponding sensitivities were 93% for M1 MCA, 56% for the VA, and 60% for the BA, with specificities of 96–98%. In summary, TCD/TCCS can detect angiographic MCA occlusions with high (>90%) accuracy, and ICA siphon, VA, and BA occlusions with fair to good (70–90%) accuracy. Furthermore, TCD-detected occlusions are associated with poor neurological recovery, disability, or death after 90 days (38,39), whereas normal results predict early improvement (40,41). In patients with acute ICA territory stroke, TCD findings, stroke severity at 24 hours, and CT lesion size were independent predictors of outcome after 30 days (38). When combined with carotid duplex

sonography, the presence and total number of arteries with suspected steno-occlusive lesions (especially intracranial) by TCD in patients with transit ischemic attack (TIA) or ischemic stroke were associated with an increased risk of further vascular events and death within six months (42). TCD-detected M1 MCA occlusions within six hours of stroke onset may be an independent predictor of spontaneous hemorrhagic transformation, with a positive predictive value of 72% (43). Occluded intracranial arteries recanalize in most cases, and TCD/TCCS provides a means to monitor the process. Recanalization additionally confirms the diagnosis of a previous occlusion (44). A multicenter TCCS study assessed M1 MCA occlusion and recanalization in patients with acute ischemic anterior circulation stroke who were treated with intravenous tPA or aspirin or heparin (45). MCA recanalization rates were 50% and 78% two hours after therapy and 24 hours after the onset of stroke in 10 cases treated with IV tPA, and 0% and 8% in 12 conservatively treated patients. A recent study showed that delayed (>6 hours) spontaneous recanalization was independently associated with hemorrhagic transformation (46). The use of contrast enhancement improves the quality of imaging and markedly increases the

140

Krejza

diagnostic confidence of TCCS, in particular in posterior circulation (47,48). TCD/TCCS is probably useful for the evaluation of patients with suspected occlusion, particularly in the ICA siphon and the MCA. The relative value of TCD/TCCS compared with MRA or CTA remains to be determined; however, if the results of the TCD/TCCS study are inconclusive, MRA can be used for diagnosis. Perfusion harmonic imaging (PHI) can detect ischemic lesions earlier than CT and distinguish the stroke subtype and severity of cerebral ischemia (49,50). There is growing interest in PHI for diagnosis, predicting recovery, differentiating stroke pathogenesis, and monitoring therapy. PHI is based on the nonlinear emission of harmonics by resonant MBs pulsating in an US field. The emission at twice the driving frequency, termed the second harmonic, can be detected and separated from the fundamental US frequency. The advantage of the harmonic over the fundamental frequency is that MBs resonate with harmonic frequencies, whereas adjacent tissue does so very little (51). In this way, PHI may enhance the signal-to-noise ratio and the ability of gray-scale scanners to differentiate MBs in the tissue vascular space from the echogenic surrounding avascular tissue. PHI is able to identify abnormal contrast enhancement in most patients with stroke. In one study, 84% (n ¼ 21) of stroke patients were correctly classified on the basis of PHI (75% sensitivity and 100% specificity in predicting size and localization of the infarction). Particularly large ischemic areas affecting both the area of the lentiform nucleus supplied by the lenticulostriate arteries and the convex surface of the brain supplied by the superficial MCA can be identified and differentiated from isolated perforator ischemia or infarctions that exclusively affect the areas supplied by the superficial branches of the MCA. Cortical infarctions in the territory of the superficial MCA can be identified if the adjacent white matter was affected as well. By contrast, lacunar infarctions could not be depicted. PHI provides a bedside tool to locate acute cerebral ischemia, in particular a large space occupying and striatocapsular MCA infarctions. A normal study may imply a minor or lacunar stroke with minimal tissue damage. The widespread availability of TCCS makes this technique a practical alternative to MRI, SPECT, and PET. Larger trials are required to establish value of PHI with respect to the extent, severity, and short-term outcome of hemispheric stroke. Major limitations of PHI are as follows: time-consuming analysis of data, problems with adequate and symmetric transparency of temporal bone windows, limited sector-shaped view of brain parenchyma, and restricted access to cortical areas of the brain. By contrast, the white matter is easily and reliably depicted because of the favorable insonation depth, the median localization in the US sector, and the marked increase in optic intensity after echo contrast application. Intracranial Atherosclerotic Stenosis

Ischemia related to intracranial artery stenosis is believed to account for 6% to 10% of strokes in Whites

and up to 29% in African-Americans and Asians (52). The most common mechanisms for ischemic stroke from intracranial stenosis are hemodynamic compromise of collateral blood flow and thromboembolism (53). The high rate of recurrent ischemia in patients managed medically suggests that angioplasty and stenting can be effective when implemented in a timely fashion (54,55). Thus, early detection of the stenosis has important implications for stroke prevention. TCD has been studied more often than TCCS, and available data suggest that when compared with contrast angiography, TCD is approximately 80% to 90% sensitive and over 95% specific in detecting stenotic lesions of the ICA siphon and M1 MCA (56–59). In expert hands, both the sensitivity and specificity of TCCS for the same arterial segments are more than 98% (60,61). Both techniques are less accurate when evaluating lesions of the PCA, VA V4 segment, and proximal BA, with the respective sensitivity and specificity of 70% and 85% for TCD (56,62) and 70% and 98% for TCCS (60). For these lesions, CT or MRA may be more useful, particularly in patients with acute distal basilar artery occlusion. There is no consensus in the literature today regarding specific criteria for the severity of stenosis. The investigators of the Stroke Outcomes and Neuroimaging of Intracranial Atherosclerosis (SONIA) study, an NIH-funded investigation assessing the accuracy of TCD and MRA in patients with symptomatic intracranial stenosis, opted for a mean velocity of 100 cm/sec for the 50% narrowing of MCA, 90 cm/sec for the carotid siphon and supraclinoid segment, and 80 cm/sec for the distal VA and proximal BA as the minimal cutoff points for enrollment in the study (56). Higher peak-systolic cutoff velocities for 50% narrowing, ranging from 180 to 220 cm/sec, have been proposed for TCCS (60,61). The major diagnostic problem, however, remains in patients with insufficient temporal windows (12). The use of sonographic contrast agent may further improve TCCS detectability of intracranial stenosis in these patients. Substantial efforts have concentrated on establishing a particular threshold of flow velocity, which can be considered as ‘‘diagnostic’’ for a specific degree of vessel narrowing. However, flow velocity in an artery is affected by many factors, which limit the diagnostic reliability of any isolated threshold of blood flow velocity. Factors decreasing the flow velocity, such as (1) increased intracranial pressure, (2) decreased cardiac output, (3) advanced age of a patient, and (4) thromboembolic occlusion of peripheral vessels, may lead to false-negative results. Falsepositive results may arise from (1) increasing velocity in cerebral arteries supplying collateral channels in the presence of severe narrowing or occlusion of other cerebral arteries, or supplying arteriovenous malformations (AVMs), (2) dilation of the cerebral resistance vessels and disturbed autoregulation in the case of stroke, or brain trauma, and (3) systemic diseases such as anemia (sickle-cell anemia) and hyperthyreosis, which may raise the CBF and flow velocity in all cerebral arteries (63).

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

The accuracy of transcranial sonography can be improved if several Doppler parameters are taken into account in defining the status of a vessel (64). The use of an interhemispheric index might be helpful in detecting the narrowing of a vessel (65). This index, however, is not useful when dealing with multiple lesions. High-grade MCA stenoses may also be suspected because of the presence of increased velocities in the ipsilateral ACA, which result from leptomeningeal collaterals (66). Intracranial arterial stenotic lesions in the internal carotid distribution, however, are dynamic and can evolve over time, with increasing or decreasing flow velocities and appearance of new collateral patterns, the latter suggesting further hemodynamic compromise distal to the stenotic lesion (67,68). In two recent studies in small, highly selected populations using peak-systolic or mean flow velocities and variable noninvasive criteria for change in degree of stenosis, progression of MCA stenosis was associated with new ipsilateral stroke or TIA or major vascular events (67,69). In summary, TCS/TCCS can be the first-line modality in the detection of MCA/ICA stenosis in patients with sufficient temporal windows, though data are insufficient to establish reliable criteria for greater than 50% stenosis or for progression of stenosis in intracranial arteries. The use of sonographic contrast agents can increase TCCS detectability of the artery in patients with an insufficient temporal window. MRA or CTA should be used instead of TCCS in patients without the windows. Also, MRA or CTA can be used to verify the results of TCD/TCCS before referral of a patient to intra-arterial treatment. Catheter angiography, however, remains a first-line diagnostic modality in patients who cannot be conclusively studied with TCCS, CTA, or MRA. Diagnosis and Monitoring of Cerebral Vasospasm

Symptomatic vasospasm (VSP) contributes significantly to the morbidity and mortality of patients after subarachnoid hemorrhage (SAH), and evidence indicates that early treatment can positively influence outcome (see chap. 14). Proper timing for intervention is often uncertain, because the diagnosis and monitoring of VSP are difficult when based solely on neurological examination, because other complications common in this patient population, such as recurrent hemorrhage, hydrocephalus, metabolic disorders, and seizures, can also produce similar neurological symptoms. Digital subtraction angiography (DSA) remains the standard criterion for defining the anatomy of intracranial arteries to diagnose VSP, but is impractical in screening and monitoring of VSP because it requires significant time, requires moving the patient to the angiographic suite, is invasive, and carries a small but definite risk of stroke, renal injury, and other complications. Alternative vascular tests, such as MRA and CTA, are less expensive and safer, but they are substantially less accurate, cannot be performed at the bedside, and have often limited accessibility. Furthermore, the risk associated with transport from intensive care unit and placement of

141

the patient in an environment where monitoring is difficult at best should not be underestimated. TCD is employed extensively for diagnosis and monitoring of cerebral VSP, but recent systematic meta-analysis of published reports revealed that specificity of TCD in the diagnosis of MCA spasm is high, at the expense of low sensitivity (70,71). For the ACA, PCA, ICA, BA, and VA, the accuracy of TCD has either not been estimated or is known to be low. Opinions are that some published data are of low methodological quality, and thus bias cannot be ruled out. It has been suggested that mean velocities less than 120 cm/sec or greater than 200 cm/sec, a rapid rise in flow velocities, or a higher Lindegaard ratio (VMCA/VICA) (6  0.3) can reliably predict the absence or presence of clinically significant angiographic MCA VSP, although prediction of neurological deterioration is problematic (71–73). Unfortunately, almost 60% of patients have velocities that fall between these thresholds. Consequently, the accuracy of conventional TCD in diagnosis of VSP remains questionable. A variety of factors, such as technical issues (the specific insonation site and the angle between the artery and the US beam cannot be determined in TCD), vessel anatomy, age, increased ICP, mean arterial pressure, Hct, arterial CO 2 content, collateral flow patterns, and response to therapeutic interventions, influence flow velocities and must be taken into account when interpreting TCD results. The other problems in the diagnosis of VSP are related to the common presence of impaired autoregulation and diffuse VSP. Although corresponding data concerning the accuracy of TCCS in the diagnosis of VSP are scarce, published reports strongly suggest that the accuracy of the ‘‘color’’ technique in the detection of the condition is high (Fig. 3) (19,70,74). TCCS is most reliable in detecting angiographic VSP of M1 MCA. The best predictive Doppler parameter is peak-systolic velocity, and an average threshold of 182 cm/sec corresponds to maximal efficiency of discrimination between states of spasm and nonspasm (efficiency, sensitivity, specificity, PPV, and NPV were 92%, 86%, 93%, 73%, and 97%, respectively) (70). In the presence of VSP, the use of the VMCA/VICA ratio [Lindegaard index (75)] might be able to identify patients with hyperemia, especially on triple-H therapy, whereas corresponding TCCS data (74) showed that the overall accuracy of the VMCA/VICA ratio in the diagnosis of mild and moderate-to-severe MCA narrowing is better than the respective accuracy of velocity measurements alone. Value 3.6 of the ratio is the most efficient threshold in the diagnosis of mild (up to 25% narrowing) M1 MCA spasm, while the threshold of 4.4 is the most efficient in the diagnosis of moderate-tosevere spasm (more than 25% artery narrowing) (74). The thresholds are higher than the upper normal reference limits of the VMCA/VICA ratio, calculated on the basis of the mean velocity (76). This ratio varies in healthy subjects from 0.86 to 3.14, for VACA/VICA from 0.54 to 2.55, and for VPCA/VVA from 0.76 to 2.90 (76). Standardization of flow velocities with respect to age and sex further increases the performance of TCCS (19). Neural networks also can be employed to improve

142

Krejza

Figure 3 Fifty-six-year old women six days after SAH. Angiography shows vasospasm in M2 MCA and A1 ACA (A) on the right side and in the distal segment of M1 MCA of the left side (B). Based on increased mean velocities in these spastic segments–217 cm/s in the M2 (C) and 393 cm/s in the M1 MCA (D), and referencing these velocities to velocities in carotid arteries in the neck (E, F) (flow velocity ratios: 8.0 on the right side and 9.3 on the left side) imaging TCD study diagnosed severe VSP. Note that the aliasing artifact (blue spot ) enabled proper placement of a sample volume to measure the highest velocity in the M1 segment. The proper identification of the artery and the site of highest velocity acceleration is important, because in this patient conventional nonimaging TCD study, performed on the same day, detected only slight velocity increase (90 cm/sec) in these arteries. Abbreviations: SAH, subarachnoid hemorrhage; VSP, vasospasm; MCA, middle cerebral artery; ACA, anterior cerebral artery; TCD, transcranial Doppler ultrasonography.

the performance of TCCS, and it has been shown that classification accuracy amounted to 92% in moderateto-severe spasm detection, and to 87% in the assessment of VSPs of other grades (77). Thus, it could be recommended that patients with suspicion of MCA

VSP should be investigated first with TCD, especially with TCCS. DSA should be reserved for patients who cannot be conclusively investigated with TCCS. TCCS diagnosis of ACA VSP using a mean velocity threshold of 75 cm/sec resulted in the

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

sensitivity and specificity values of 71% and 85%, respectively (78). Visualization of the normal and particularly the narrowed ACA is more difficult than that of the MCA (10). False-negative results for ACA may be explained by collateral flow through the ACoA and by problems with angiographic differentiation of frequently occurring ACA hypoplasia from vessel narrowing. The VACA/VICA ratio can be helpful in the differentiation of ACA spasm from the normal status of the artery. In practice, however, diagnosis of unilateral spasm of the ACA is not obligatory, because its hemodynamic consequences for the downstream flow are generally not a cause of concern. On the contrary, bilateral ACA spasm may reduce flow to the postcommunicating ACA segments, and TCCS can detect increases in velocity involving at least one artery. Very few data have been provided on TCD diagnosis of spasm of the PCA and BA (79). A recent study evaluating the reliability of TCD assessment of BA VSP found a 100% sensitivity and a 95% specificity by using a ratio of peak mean velocity in BA to the velocity in extracranial vertebral artery (VA) over 2 as diagnostic criterion (80). Normal reference ranges of the velocity ratio VPCA/VVA (0.76–2.90) can also be helpful in interpreting abnormal velocity results (81). TCD is not useful for the detection of VSP directly affecting the convexity or vertically oriented branches of the intracranial arteries distal to the basal cisterns (82,83), although the presence of VSP at these sites may be inferred in some cases by indirect Doppler waveform observations (e.g., decreased diastolic flow, increased pulsatility, side-to-side differences in pulsatility indexes). Data on TCCS in this context are lacking. TCD/TCCS is useful in monitoring the temporal course of angiographic VSP after SAH. Although no adequate study has been conducted, TCD is thought to be valuable in the day-to-day evaluation of SAH patients in VSP and to assess the effect and durability of neuroradiological interventions (84,85). TCD has been used to detect angiographic VSP following prophylactic transluminal balloon angioplasty in SAH patients at a high risk of developing VSP (86), as a noninvasive surrogate endpoint, or to demonstrate biological effects of treatments for vasoconstriction or VSP in uncontrolled trials of pharmacological therapies for eclampsia and SAH (87–89). Data are insufficient to make a recommendation regarding the use and method(s) of autoregulation testing to predict the risk of delayed cerebral ischemia. The follow-up TCD/TCCS studies to assess VSP dynamics should begin at admission, when the probability of VSP is still relatively low, in order to determine reference velocities for further comparison. In many patients, the rate of velocity increases and the maximal velocity can identify patients at greatest risk of symptomatic VSP. Such patients should receive daily TCD/TCCS studies, while those with normal velocities and no substantial velocity rise can be monitored every two to three days throughout the period of the high risk of VSP. In interpreting Doppler results, the global and local ICP increase and disturbed autoregulation should be taken into account. In every

143

patient, the velocity ratios (VMCA/VICA, VACA/VICA, and VPCA/VICA) should also be calculated. Increased impedance indexes may suggest the presence of localized or generalized increased ICP or hydrocephalus, necessitating appropriate diagnostic evaluation and treatment. A patient whose neurological condition is deteriorating and who has a normal or nondiagnostic TCD/TCCS study should undergo angiography to detect TCD/TCCS occult VSP if another cause for deterioration is not identified. In addition, patients with VSP whose condition does not improve or continues to deteriorate in spite of aggressive conservative management should be considered for urgent endovascular treatment. TCD/TCCS can demonstrate the effectiveness of the treatment by showing a decrease in flow velocities and velocity ratios. TCD/TCCS can be helpful in proper timing for aneurysm clipping or coiling, and postoperative management. If flow velocities are very high (mean velocity in MCA above 150 cm/sec) or there is evidence of altered autoregulation and low CBF in the first or second week after SAH, operative results may be poor, in particular if the patient develops hypotension during the procedure. After the procedure, transfer from the intensive care unit or mobilization of postoperative patients is inadvisable in the presence of high velocities and should be postponed until flow velocity in the affected vessel begins to decline. Surveillance of Coiled Intracranial Aneurysms

Endovascular detachable coil treatment is being increasingly used as an alternative to craniotomy and clipping for many ruptured intracranial aneurysms. Since the long-term risk of reopening and possibly for rebleeding after endovascular coiling is somewhat higher than after surgical treatment, the persistence of aneurysm occlusion after coil embolization is of concern. A significant problem of endovascular therapy is the known instability of initial coil packing, and complete occlusion of the aneurysm is not always possible without running a high risk of inadvertent vessel occlusion or coil migration. Consequently, an initially occluded aneurysm can recanalize, which may be associated with higher risks of regrowth and rebleeding. DSA is currently used as the primary imaging technique for the immediate and long-term evaluation of endovascular therapy of intracranial aneurysms. Diagnostic DSA is performed at least three times: at the end of coiling procedure and at 6 and 18 months after the procedure. If reopening resulting in moderate-to-extensive residual flow is seen in subsequent DSA studies, then re-embolization therapy is usually undertaken. DSA, however, is a costly and invasive procedure. Furthermore, the estimation of aneurysm occlusion can sometimes be difficult because of X-ray attenuation of metal coils and artifacts caused by the densely packed coils (Fig. 4). TCCS may be more cost effective in the surveillance of coiled intracranial aneurysms. TCCS can identify large and medium-sized intracranial aneurysms located in the proximal segments of the circle of Willis (90–92). Typically, an aneurysm is imaged as a

144

Krejza

Figure 4 TCCS can identify large- and medium-sized intracranial aneurysms in the proximal segments of the circle of Willis. Angiography (A) shows large ICA aneurysm, which is depicted with power TCCS as red pulsatile color structure (B). Abbreviations: TCCS, transcranial color-coded duplex sonography; ICA, internal carotid artery.

pulsatile colored structure adjacent to the large parent artery. Aneurysms can display various flow patterns. The most typical color-coded feature is the presence of two areas with inversely directed flow: half of the aneurysm is coded blue and the other half is coded red. The colors correspond to the direction of inflowing and outflowing blood. Between these two areas, a black separation zone with undetectable blood flow can be recognized. Flow velocities are usually low, without turbulence and spontaneous velocity fluctuations. Aneurysms located in the basal arterial trunks can be recognized more easily than those situated in the periphery. The use of sonographic contrast material can improve the reliability of TCCS in the detection of aneurysms. Preliminary reports suggest that TCCS with contrast enhancement is highly specific and sensitive in the detection of clinically relevant residual flow within an aneurysm after endovascular coiling. Schuknecht et al. (93) reported that TCCS, performed immediately after embolization, reliably confirmed complete occlusion of 42 out of 43 aneurysms. In four other aneurysms, a slight residual flow was recognized with TCCS. Furthermore, in three (ophthalmic and basilar tip and cavernous carotid aneurysm) of 26 aneurysms reexamined 6 to 20 months after treatment, reappearance of color flow signal adjacent to the coils was detected, which was in agreement with DSA. The persistent occlusion in the other 23 cases was demonstrated by TCCS, which was either confirmed by angiography in 13, or by MRA in 10. Turner et al. (94) reported that TCCS confirmed complete occlusion in 19 of 20 aneurysms, while minor residual flow was detected with TCCS in 4 of 10 aneurysms. In the detection of clinically relevant residual flow, the results of standard TCCS were less consistent. While moderate residual flow was detected in eight of nine aneurysms (sensitivity 89%, specificity 97%), more extensive residual flow was detected only in

three of seven aneurysms (sensitivity 43%, specificity 100%). The use of contrast slightly improves the sensitivity of TCCS in the detection of moderate (100%) residual flow, but substantially improves sensitivity in aneurysms with extensive residual flow (86%) (94,95). These studies suggest that TCCS could be used to selectively monitor intracranial aneurysms, which would reduce the requirement for long-term invasive monitoring. The detection of neck refilling is improved with contrast enhancement. In our study (unpublished data), which is based on 107 patients with coiled aneurysms, we found that standard TCCS can be very specific in detecting moderate-to-severe residual blood flow in relatively large (over 10-mm-diameter) aneurysms located at the basilar tip, at the ICA bifurcation, and at the AcoA area. Thus, TCCS has great potential to replace DSA in the surveillance of coiled intracranial aneurysms in selected locations (Fig. 5). Vascular Malformations

TCCS studies of AVMs show a focal accumulation of vascular convolutions as a color mosaic with abnormal Doppler waveforms (96,97). TCCS and TCD can also detect the hemodynamic abnormalities in feeding and draining vessels. Typically, flow velocities in feeding vessels are high, ranging from 140 to 200 cm/sec, and impedance indexes are low, indicating a drop in distal resistance (98,99). Draining veins are enlarged, channeling pulsatile arterialized blood away from the AVM nidus (Fig. 6). The diagnostic accuracy of TCD and TCCS in detecting AVMs is not known. Large (>4 cm) and medium-sized (2–4 cm) radiologically proven AVMs are regularly detected (100). Because more than onethird of small (<2 cm) AVMs can be missed, TCD is not considered a reliable diagnostic tool in this setting. Nevertheless, both TCD and TCCS are useful in monitoring the effects of therapeutic procedures (101,102).

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

145

Figure 5 TCCS can detect clinically relevant residual flow within an aneurysm after endovascular coiling. In a patient with ICA bifurcation aneurysm (A), TCCS shows coils as hyperechoic structure (B–D), while TCCS in power mode identifies residual flow in the neck and body of the aneurysm. Abbreviations: TCCS, transcranial color-coded duplex sonography; ICA, internal carotid artery.

The Intracerebral Venous System: A Neurosonological Study with TCCS

The straight sinus, basal cerebral veins, cavernous sinus, and superior and inferior sagittal sinuses can be insonated with TCCS. Normal peak-systolic velocities usually range from 5 to 35 cm/sec, varying from one sinus to the other (103). In sinus thrombosis, flow velocities may decrease, or they may markedly increase to above 100 cm/sec, most likely indicating increased collateral circulation (104,105). Follow-up studies show gradual normalization after a period of months (105). Venous flow velocities are also affected by ICP changes (106). Experience with cerebral vein insonation, however, remains limited.

Current View on Microemboli Detection The physical and technical aspects of ultrasonic detection of microembolic signals (MES) by TCD have recently been reviewed (56,107,108). Particulate (solid, fat) and gaseous materials in flowing blood have different acoustic impedance properties than surrounding red blood cells. The Doppler US beam is both reflected and scattered at the interface between the embolus and blood, resulting in an increased intensity of the

received Doppler signal. MES are usually observed within the spectral envelope, while artifacts, which can resemble MES, extend outside the envelope and are bidirectional. The hierarchy of backscatter of the US, in descending order, is gaseous emboli, solid emboli, and normal-flowing blood (including transient red blood cell aggregates). In clinical practice, however, it is difficult to determine whether a given MES corresponds to a large platelet embolus or to a small atheroma due to a considerable overlap between MES characteristics. Several techniques have been proposed to resolve these issues (Fig. 7) (109). Technical limitations present considerable difficulties. For example, using a higher-decibel threshold improves reproducibility, but it can decrease sensitivity (110). In an attempt to establish a general consensus among investigators, a committee of experts has defined MES characteristics (107). Manual saving of suspicious signals by the recording operator and subsequent offline analysis is the standard practice today, but it is cumbersome and time consuming. Automated systems for embolus detection have been developed (111,112). Their accuracy remains limited. Although the optimal duration of insonation needed to achieve maximum sensitivity is unknown, most centers monitor for 30 minutes to 1 hour (113). Longer periods of

146

Krejza

Figure 6 TCCS shows an AVM in the base of the skull as an area of color mosaic with abnormal low-resistance Doppler waveforms (A, B). TCCS in power mode (C) shows the arterial convolution of AVM in red color, which can help identifying the feeder and drainer vessels. Angiography shows that the AVM is supplied from right (D) and left side (E). Abbreviations: AVM, arteriovenous malformation; TCCS, transcranial color-coded duplex sonography.

insonation, or repeat studies, may be needed in some instances. MES have been detected in patients with asymptomatic and symptomatic high-grade ICA stenosis, prosthetic cardiac valves, myocardial infarction, atrial fibrillation, aortic arch atheroma, fat embolization syndrome, and retinal or general cerebral vascular disease. In addition, these signals occur in coronary catheterization, coronary angioplasty, direct current cardioversion, cerebral angiography, carotid endarterectomy (CEA), carotid angioplasty, and cardiopulmonary bypass. TCD can be used to localize the embolic source or monitor the effects of antithrombotic treatment in patients with atherosclerotic CVD (114). In patients with high-grade carotid stenosis, sources of asymptomatic MES may include ulcerated plaques (115) and microscopic platelet aggregates and fibrin clots (116). Asymptomatic cerebral microembolization was associated with an increased risk of further cerebral ischemia in this setting (115). Comparison between studies on MES detection is difficult, however, because of differences in: 1) diagnostic criteria, 2) detection threshold, 3) instruments and instrument settings, 4) nature and severity of disease, 5) time between last symptom and detection of microembolic signals, and 6) incidence of microembolic signals (56,107). Nevertheless, TCD is

probably useful to detect cerebral MES in a wide variety of cardiovascular or cerebrovascular disorders. However, current data do not support the use of TCD for diagnosis or for monitoring response to antithrombotic therapy in ischemic CVD. Furthermore, data have not shown that detection of MES leads to improved patient outcomes. US Monitoring of Neurointerventional Procedures

Carotid angioplasty and stenting. MES have been associated with a higher neurological complication rate and are a potential cause of periprocedural stroke after CEA. MES have also been observed during carotid angioplasty and stenting (CAS). The high frequency of MES, however, is not associated with a chronic cognitive impairment as shown by TCD monitoring, although CAS is accomplished with increased dislocation of microemboli compared with the surgical approach (117). Consequently, cerebral protection devices are increasingly used. TCD monitoring can help assess the efficacy of cerebral protection devices deployed during stenting (118), though a recent study showed that the frequency of procedure-related silent cerebral lesions appeared to be independent of the number of MES during the procedure (119). Also, a systematic microscopic analysis of debris captured by

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

Figure 7 TCD can detect MES in major intracranial arteries. The upper image shows MES during carotid angioplasty and stenting, performed without a protective device. The MES are also present at clam release during CEA (lower image). Also note that flow velocities rebound substantially after the clamp release. Abbreviations: TCD, transcranial Doppler ultrasonography; MES, microembolic signals; CEA, carotid endarterectomy.

the filter device has no predictive value for potential cerebral ischemia after carotid artery stent placement (120). In CAS, TCD monitoring provides insight into the pathogenesis of procedure-related cerebral events. Microemboli during poststent dilation, particulate macroembolism, massive air embolism, and angioplasty-induced asystole are associated with adverse outcome, as are male gender and prior cerebral ischemia (121). TCD monitoring can also predict early cerebral outcome after carotid bifurcation CAS (122). In CAS, in addition to such obviously adverse events as particulate macroembolism and massive air embolism, multiple MES (>5 showers) at postdilation after stent deployment and angioplasty-induced asystole and hypotension with a significant reduction in MCA blood flow velocities are associated with periprocedural cerebral deficits. In combination with the presence of preprocedural cerebral symptoms, these four TCD-monitoring variables reasonably differentiate between patients with and without adverse cerebral outcome (122). Additionally, color-coded duplex sonography of carotid arteries can demonstrate hemodynamic improvement after ICA stenting (123). A study suggests that asymptomatic MES correlate with clinical risk (124). However, outcome studies

147

are required to determine whether MES detection may allow prediction of stroke risk and monitoring of the effectiveness of therapy. Percutaneous angioplasty in posterior arteries. TCD monitoring is also useful to detect MES during and after percutaneous transluminal angioplasty (PTA) in the posterior circulation (125). In the patients with subclinical microemboli released from the dilated vessels for three days after vertebral and subclavian PTA, anticoagulant or antiplatelet therapies may prevent embolic complications after the procedure. Monitoring of endovascular treatment of intracranial aneurysms. Selective occlusion of intracranial aneurysms with detachable coils has an overall estimated procedure-related permanent complication rate of 3.7% to 6.8%. Thromboembolic events with partially or completely persisting neurological deficits are reported in 2.4% to 5.2% of endovascular-treated patients (126). Intraprocedural systemic administration of heparin is widely used in several institutes to reduce the risk of thromboembolism. Acute clotting at the thrombogenic interventional materials is considered the most important source of thromboembolism during endovascular embolization of cerebral aneurysms. Potential clinical complications can be avoided by early recognition of thrombus at the coil–parent artery interface and by administering appropriate medical therapy (127). TCD monitoring during and immediately after coiling can help identify patients at high risk of thromboembolic complications. Those with high risk can be selected for a therapy. A relatively low-dose intra-arterial abciximab infusion can immediately dissolve an acute thrombus that forms during intracranial aneurysm coil placement (128). An inhibition of the platelet function by acetylsalicylic acid might be yet another effective strategy to minimize the rate of thromboembolism (129). However, even if the anticoagulation strategy is the most important factor to decrease the rate of embolic events during aneurysm treatment, the strategy has to balance the risks of thromboembolism and bleeding (129).

Duplex Sonography of Carotid and Vertebral Arteries Sonographic Assessment of Vascular Pathology

The earliest atherosclerotic changes include intimal thickening secondary to lipid deposits and lipidladen macrophage infiltration of the arterial wall (130). As the process advances, atheromatous plaques begin to protrude into the arterial lumen. Initially, these plaques are covered with a fibrous cap that gives them mechanical stability. Fibrous plaques are rarely associated with neurological symptoms. On US, they appear smooth, isoechoic, and homogenous. During subsequent stages, increasing amounts of extracellular lipids and cholesterol esters are deposited, calcification occurs, and intraplaque hemorrhages develop, giving the plaque a heterogeneous appearance on US examination (108). When the mechanical support of the plaque surface erodes, embolization of plaque contents may occur. In addition, plaque surface ulcers

148

Krejza

Figure 8 Atherosclerotic plaques on US images. (A) Homogenous, soft, and fibrous plaque at the ICA origin without velocity increase. In another patient, (B) the soft plaque in the ICA is irregular with hypoechoic area within the plaque as a potential source of thromoembolism. In the third patient, (C, D) plaques are more complex with calcification. In the left, ICA (C) plaques cause over 70% stenosis with subsequent peak-systolic velocity increase to 270 cm/sec, while on the left side irregular plaques cause less than 50% ICA narrowing. Abbreviations: US, ultrasound; ICA, internal carotid artery.

may develop and serve as foci for thrombus formation. These plaques appear heterogeneous with variable echodensities, calcific shadows, and surface irregularities. Morphological and physiological features readily assessed with US and associated with an increased risk of cerebral infarction include intraplaque echolucency, surface ulceration, and most importantly degree of stenosis (Fig. 8). Degree of Stenosis

The North-American Symptomatic Carotid Endarterectomy Trial (NASCET), the European Carotid Surgery Trial (ECST), and the Asymptomatic Carotid Artery Surgery (ACAS) trial demonstrated the benefit of endarterectomy in symptomatic and asymptomatic patients with moderate and severe carotid stenosis (131–133). In NASCET, however, endarterectomy was only marginally beneficial when the degree of stenosis was between 50% and 70%, underscoring the importance of accurately measuring the severity of stenosis. Ultrasonography must therefore be able to distinguish between a carotid stenosis of 50% and of 70% in symptomatic patients and to identify 60% diameter stenosis in asymptomatic patients. Noninvasive evaluation of the extracranial ICA with color duplex US and MRA is increasingly regarded a replacement of DSA, but the role of ultrasonic

quantification of narrowing is not undisputed (134). In a recent study comparing duplex US to DSA, duplex ultrasonography misclassified 28% of patients considered candidates for CEA (135). Other studies suggest that this classification can be performed more accurately if high standards of ultrasonography are maintained (136–138). Various diagnostic criteria have been proposed for determining the percentage of stenosis (see reviews of the criteria and tables) (108,138,139). These include peak-systolic velocity, end-diastolic velocity, and the ratio of peak-systolic velocities in the ICA to the mid-CCA (ICA/CCA ratio). The peaksystolic velocity has traditionally been felt to provide the closest angiographic correlation and is easily obtained; however, many laboratories rely also on the end-diastolic velocity or the VICA/VCCA ratio to obtain improved diagnostic accuracy and to correct for factors that may alter the carotid blood flow. Such factors include low cardiac output, valvular disease, acute elevations in blood pressure, anemia, and abnormal collateral flow. Any of these conditions may lead to flow alterations across a carotid plaque and to either over- or underestimation of the true degree of stenosis. In these instances, the VICA/VCCA ratio often helps in correcting for hemodynamic disturbances, but it has limitations (108). The diagnostic impression is further confirmed by B-mode and color flow

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

imaging, which allow visual inspection of the degree of stenosis caused by the plaque. A quantitative crosssectional analysis of plaque stenosis derived from color flow images was recently proposed, which further increases diagnostic accuracy (140).

Table 3 Diagnostic Parameters for Internal Carotid Artery Stenosis of 60% or More Reference

Clinical Utility of Carotid Duplex

Characterization of plaque morphology and determination of degree of vessel stenosis are the most common clinical applications of carotid US and were reviewed in many publications (108). In interventional neuroradiology, carotid US is also used in the following areas. Table 2 Diagnostic Parameters for Internal Carotid Artery Stenosis of 70% or More Reference

Diagnostic criteria

Sensitivity Specificity Accuracy (%) (%) (%)

139

Vps > 230

80

90

NR

144

Vps > 325 ICA/CCA Vps ratio > 4

83 91

100 87

88 88

145

Ved > 100 Vps > 210 Vps > 130 and Ved > 100

77 89 81

85 94 98

80 93 95

146

Vps > 270 Ved > 110 Vps > 270 and Ved > 110

96 91 96

86 93 91

88 93 93

147

Ved > 70 ICA/CCA Ved ratio > 3.3 Vps > 210 ICA/CCA Vps ratio > 3

92 100

60 65

77 79

94 91

77 78

83 83

148

Vps > 130 and Ved > 100

87

97

95

149

Vps > 230 Ved > 70 ICA/CCA Vps ratio > 3.2

86 82 87

90 89 90

89 87 89

All velocities in cm/sec. Abbreviations: NR, not reported; Vps and Ved, peak-systolic and end-diastolic velocities, respectively; CCA and ICA, common and internal carotid arteries, respectively.

Diagnostic criteria

Sensitivity Specificity Accuracy (%) (%) (%)

150

Vps > 260 and Ved > 70

84

94

90

151

ICA/CCA Vps ratio > 2 Ved > 40 ICA/CCA Ved ratio > 2.4 Vps > 170

97

73

76

97 100

52 80

86 88

98

87

92

89

92

NR

Reference Values

A wide range of criteria have been proposed to identify the clinically relevant degrees of ICA stenosis. They are summarized in Tables 2–4. The diagnostic accuracy of duplex US ranges between 85% and 95% and varies among laboratories. A survey of diagnostic criteria showed that at least nine different diagnostic parameters are currently used to measure the severity of stenosis (141). These differences among laboratories illustrate the fact that US testing is equipment and operator dependent, and they emphasize the necessity for each laboratory to develop its own diagnostic criteria on the basis of DSA correlations (142,143).

149

152

Vps > 245 and Ved > 65

All velocities in cm/sec. Abbreviations: NR, not reported; Vps and Ved, peak-systolic and end-diastolic velocities, respectively; CCA and ICA, common and internal carotid arteries, respectively. Table 4 Diagnostic Parameters for Internal Carotid Artery Stenosis of 50% or More Reference

Diagnostic criteria

Sensitivity Specificity Accuracy (%) (%) (%)

153

Vps > 120

79

84

82

145

Vps > 130

97

97

97

154

ICA/CCA Vps ratio > 1.6

95

92

93

155

Ved > 50 Vps > 150 ICA/CCA Vps ratio > 2

91 98 96

86 84 89

89 92 93

149

ICA/CCA Vps ratio > 1.6 Vps > 130

93

83

88

92

90

91

All velocities in cm/sec. Abbreviations: Vps and Ved, peak-systolic and end-diastolic velocities, respectively; CCA and ICA, common and internal carotid arteries, respectively.

Monitoring after revascularization procedures. While the practice of serial follow-up examinations after endarterectomy is intuitively appealing, the value of routine postoperative surveillance is uncertain. The incidence of restenosis (defined as a reduction in diameter of more than 50%) varies between 2% and 20% at one to three years after surgery (156), but the incidence of recurrent symptoms is low (157,158). Restenosis within two years of surgery is usually secondary to intimal hyperplasia and carries a benign prognosis, since the risk of distal embolism is low and lesions often regress (158,159). Late restenosis is most likely secondary to recurrent atherosclerosis, and the associated risk of ipsilateral hemispheric or retinal symptoms may not be different than that of the original primary lesion (159). In a small percentage of cases, postoperative testing shows evidence of thrombus formation at the endarterectomy site, intimal flaps, and occlusion.

150

Krejza

Figure 9 Carotid US of common carotid artery dissection. Gray-scale image (A) shows luminal irregularities and an intimal flap. Color image (B) confirms absence of flow beneath the flap. In the ipsilateral internal carotid artery (C), the flow pattern is abnormal with disappearance of flow during diastole (high-resistance flow pattern). Note, compensatory high-velocity flow pattern in the VA (D). Abbreviation: US, ultrasound; VA, vertebral artery.

Intracarotid and intravertebral stent placement is being performed with increasing frequency. Repeat testing after stent placement usually reveals an improvement of the intraluminal hemodynamic pattern. It is unclear, though, whether the diagnostic criteria presented in Tables 2–4 are applicable in the detection and monitoring of in-stent stenosis. Arterial dissection. The US features of dissection are less specific than those observed with angiography and usually reflect flow abnormalities seen in highgrade stenosis secondary to any etiology: high flow velocities, high resistance flow patterns, or complete absence of flow (108,160). Despite the advantage of US in displaying luminal irregularities, an intimal flap is infrequently seen, possibly because the size of the flap lies beyond the resolution of US (Fig. 9) (161). US is helpful in monitoring the course of natural repair. Vertebral dissections follow a similar course. VA flow disturbances are nonspecific and show the same patterns as any stenotic lesion associated with intraluminal hemodynamic change. Such patterns include absence of a flow signal, bidirectional or dampened flow, and elevated flow velocities with associated turbulence (162,163). Pitfalls of Carotid Ultrasonography

Carotid occlusion. US diagnosis of a carotid occlusion remains unreliable, as a minimally patent arterial

lumen with a trickle of flow can be missed. In the case of symptomatic atherosclerotic disease, such a differentiation is vital because CEA or CAS is clearly indicated in a patent vessel, but is generally not possible in the case of occlusion. Early reports suggested a diagnostic accuracy of 85% for ICA occlusion, but in more recent studies, which were based on color duplex imaging, the accuracy was shown to exceed 96% (164,165). Difficulties arise from the presence of calcific plaque formation and the low flow volume in near occlusions. In addition, arterial tortuosity may cause angle artifacts, further compromising sensitivity. Diagnostic confusion may also arise when an external carotid artery branch overlies the ICA occlusion and is incorrectly identified as a patent residual lumen. In some patients with ICA occlusion, the external carotid artery assumes a low-resistance pattern as it provides collateral flow to the brain. Tapping the fingers over the temporalis muscle and the identification of vascular branches may help differentiate the external from the internal carotid artery. The ICA occlusion and high-grade stenosis also lead to diagnostic difficulties in determining the degree of stenosis on the contralateral side. Increased contralateral flow velocities may be secondary to collateral flow and lead the sonographer to overestimate the degree of true stenosis. In such cases, the use of peaksystolic velocity alone is insufficient and misleading. The overestimation is proportional to the degree of

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

151

contralateral stenosis. Increasing the number of diagnostic criteria in the setting of contralateral stenosis improves the diagnostic accuracy. The VICA/VCCA ratio may accurately reflect the degree of stenosis in this setting. Calcification. Heavily calcified plaques often cast an acoustic shadow that prevents duplex examination. Doppler velocities can then only be measured proximal and distal to the lesion, and elevated flow velocities at the level of the stenotic plaque can be missed (108). If the width of the acoustic shadow does not exceed 1 cm, it may be inferred from normal distal flow velocities that a high-grade lesion is not present. Tortuosity. With the aging process, the ICA can become elongated and develop loops or kinks, which may cause increases in flow velocity, suggesting a focal area of stenosis. Color duplex examinations are particularly helpful in these cases. High bifurcation. In patients with high CCA bifurcation, the mandible interferes with the US evaluation. A posterior approach in these instances often allows a better evaluation of the artery. Extracranial VA

Atherosclerotic plaques of the extracranial VA are usually localized at the artery’s origin from the subclavian artery, and they also tend to involve the vertebrobasilar junction. In addition, the VA is also susceptible to dissection at the V1 and V3 segments. Intraluminal flow characteristics can be readily assessed with extracranial US. However, velocities are usually measured only in the V2 intravertebral segment. Interrogation at this point allows determination of flow direction and pattern, but it gives only indirect evidence about proximal or distal stenotic lesions. Insonation of the VA origin is technically difficult because of its deep intrathoracic location, which does not always allow for optimal angle correction. Normal values range between 19 and 98 cm/ sec for peak-systolic velocity, and 6 and 30 cm/sec for end-diastolic velocity (108,166). For the normal VA origin, a peak-systolic velocity of 69 cm/sec and enddiastolic velocity of 16 cm/sec have recently been reported (167). Compared with the ICA and MCA, flow is slower in the vertebrobasilar trunk. There are no established US criteria for VA stenosis. Hemodynamically significant VA disease can be inferred when a focal flow velocity increase of 50% or more is detected. The presence of a highresistance pattern suggests high-grade distal stenosis (168). However, because the resistance pattern is highly variable, it is an unreliable finding, which is further confounded by the frequent presence of congenital variants in the vertebrobasilar circulation, including intradural VA hypoplasia (169). Flow in a hypoplastic vessel may be dampened, mimicking a high-resistance pattern with almost absent diastolic flow. This characteristic confuses the interpreter and affects the test’s accuracy. Experience with extracranial VA US remains limited, and the technique is not used as often as for ICA disease. The subclavian steal syndrome is usually a result of high-grade stenosis or

Figure 10 The subclavian steal syndrome as a result of highgrade stenosis of the proximal subclavian artery. The ipsilateral VA acts as a collateral vessel, channeling blood distal to the obstruction. Flow direction in the VA is reversed, which is shown in red color compared with the flow direction in the common carotid artery shown in blue color (upper image). Duplex US shows abnormal waveform pattern (lower image). Abbreviation: US, ultrasound.

occlusion of the proximal subclavian artery. As perfusion pressure and blood flow in the arm drop, the ipsilateral VA acts as a collateral vessel, channeling blood distal to the obstruction. Flow direction in the VA is reversed. The syndrome can be diagnosed with US with high sensitivity (Fig. 10) (170,171).

THERAPEUTIC USE OF US IN ACUTE STROKE The current treatment of acute ischemic stroke requires intravenous delivery of a large dose of a serine protease, such as tPA, urokinase, or streptokinase, within three hours of symptom onset. Proteases work by converting plasminogen to the natural thrombolytic agent, plasmin (20,21,172). Plasmin lyses thrombotic vascular occlusions by degrading fibrinogen and fibrin contained in a blood clot. If therapeutic recanalization of the occluded artery is prompt, a favorable outcome is anticipated in about 60% of those with an ischemic

152

Krejza

stroke compared with spontaneous thrombolysis, which occurs in up to 20% of patients with variable clinical recovery. With tPA treatment, a faster recanalization results in moderate clinical improvement. Small increments of delay in treatment decrease chances for timely return of flow and favorable outcome, supporting the concept ‘‘Time is brain.’’ If a thrombolytic agent is delivered three hours after onset of symptoms, the risk of hemorrhagic transformation increases substantially; however, by using catheter-directed arterial delivery of the thrombolytic drug, the treatment window can be extended to six hours with relatively low risk of hemorrhage (173–175). This procedure involves a much smaller dose of the trombolytic agent and is directly delivered to the thrombus in the clotted artery. After six hours, there is no effective pharmacological thrombolytic treatment, because if late reperfusion occurs, the area of ischemic stroke may convert into the much more severe hemorrhagic stroke with worsened outcomes. However, thrombolytic agents alone, even if given in the desired time windows, have limited success in recanalyzing thrombotically occluded arteries (176,177). Major reasons for incomplete recovery include a severe initial ischemic insult and slow and incomplete thrombolysis (178,179). Successful thrombolysis depends on the delivery of tPA to the thrombus through residual blood flow around the arterial obstruction (180,181). As such, there is a strong need to enhance the effectiveness of thrombolytic agents by shortening the time to reperfusion. Experimental and limited clinical studies suggested sufficient penetration and thrombolytic effects of either low-frequency (182,183) or diagnostic (184) US through the skull in vitro and, hence, encouraged empiric assessments of US for thrombolysis even with standard US equipment (185).

Mechanism of US-Accelerated Thrombolysis The mechanisms for US-accelerated thrombolysis in externally applied exposures are unclear. Early studies demonstrated that the effect was primarily nonthermal and did not involve mechanical fragmentation (186). Enhancement has been shown to decrease with increasing frequency and increase as a function of ontime as the duty cycle is varied (187). US increases the uptake of tPA into a clot, suggesting that enzyme transport is important (180). It also increases the binding of tPA to fibrin by maximizing access of the enzyme to potential binding sites on the fibrin matrix (188). Furthermore, it can reversibly increase fluid permeation through fibrin (189), a finding shown to depend on reversible increases in the number of fibers per unit area and concomitant decreases in fiber diameter during US exposure. Degassing reduces the effect of US on flow through fibrin and associated structure changes. These and other clues implicate gas concentration as an important factor and suggest that acoustic cavitation (see the section ‘‘Technical Aspects of US Imaging’’) may be a dominant mechanism. In stable cavitation, the stiffness of the gas in the MBs controls the radial pulsations of MBs driven by the US field. If a

bubble is induced to grow by US to a diameter larger than the pore size of the fibrin lattice surrounding it, stretching of clot fibers may occur. Microstreaming around the MBs may cause damage to nearby cells or fibers, or act to stir fresh fibrinolytic enzyme into otherwise inaccessible regions in a clot. In inertial cavitation, the radial motion of a bubble is controlled by the inertia of the rapidly moving liquid surrounding it. For symmetrical collapse, hot spots can form that can produce hydroxyl free radicals capable of attacking nearby fibers (8). For asymmetrical collapse, microjets may form that can damage nearby fibers in the manner of pitting on a ships’ propeller. If MBs collapse sufficiently violently to produce broadband acoustic emissions, additional inertial cavitation may produce localized stresses, hot spots, or microjets that may further alter the structure of clot fibers. In all cases, US-driven MBs might exteriorize new binding sites along fibers to allow fibrinolytic enzymes increased access [see editorial by Polak (190)]. Besides cavitation, other effects, which depend on the level of US energy applied, may play an important role in vitro when the diagnostic range of US is used. At very low energies, US promotes microstreaming of blood close to the occluding thrombus and enhances the mixing of tPA, effectively increasing the concentration of the agent that is in contact with the thrombus. The pressure waves that are generated may also increase the permeation of tPA into the interior of the fibrin network. At slightly higher US energies, the binding of tPA to the cross-linked fibrin and fibrin elements within a matrix is enhanced, in vitro (191), and the fibrin cross-links are weakened, further increasing the binding of tPA. It is also possible that the heat generated by US is additionally responsible for accelerating thrombolysis (192). Experiments have confirmed that the temperature elevation generated by US of sufficient power can increase the dissolution rate of thrombi. A major limitation of TCD, however, is attenuation of US by the bones of the cranium; consequently, diagnostic imaging and the therapeutic use of US may not be possible in 10% to 15% of patients (12). The US beam becomes attenuated, and 90% of energy is deposited in the bone–soft tissue interface. Consequently, only 10% of the maximum output intensity hits the thrombus, which comes to an effective energy of about 0.07 W/cm2. Solid data concerning comparatively low levels of energy and their effects on thrombolysis are lacking. In summary, the mechanism responsible for the effect of US on thrombus dissolution is not completely known. The excessive heat deposition at the bone–soft tissue interface is a major limitation in applying higher US power through the temporal window, in particular at higher-frequency US. To overcome this problem, several strategies were developed: first, to use endovascular wires and transducers to deliver US locally; second, to use lower-frequency and subsequently higher-power US for transcutaneous US-enhanced thrombolysis; third, to use US contrast to induce and increase the number of cavitations at the site where the US beam of a high mechanical index is targeted.

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

Endovascular Ultrasound Thrombolysis Intravascular devices such as vibrating wires at frequencies of 20 to 25 kHz delivering very high power levels of US of up to 20 W locally have been shown to disrupt the clots in vitro (193–195). This approach has been used to fragment, mechanically, thrombi into small particles, resulting in reperfusion in patients with obstructed peripheral arteries (196–198). These arteries require great angiographic skill besides of disadvantages such as unknown effects of distal embolization of fragments, damage or perforation of the vessel wall, heating, and ultrasonic wire breakage. Miniaturized transducers also have been attached to catheters for direct endovascular use, offering the potential of localized US thrombolysis while avoiding attenuation of intensity through the skull and reducing insonation of the surrounding tissue. Tachibana and Tachibana demonstrated enhanced clot lysis in vitro using a microtransducer operating at 225 kHz. Similar in vitro results were demonstrated for combined application of US (170 kHz, 0.5 W/cm2) and thrombolytic infusion (199). The specialized US thrombolytic infusion catheter (EKOS Corporation, Bothell, Washington, U.S.) combines the use of a miniature US transducer on the tip of the catheter with infusion of a thrombolytic agent through the catheter. After a bolus of tPA is injected, an infusion of tPA is started with simultaneous US given for up to one hour. Human trials showed great promise. Only large vessels can be effectively treated with US, but tPA may lyse peripheral fragments in the area. The delays involved with angiography and demands for very skilled operators, which apply to all mechanical devices, limit the potential of endovascular use of microcatheters for acute stroke treatment to specialized centers; thus a broader applicability seems unrealistic.

Transcutaneous US-Enhanced Thrombolysis Noninvasive external application of US has greater potential for wider therapeutic application because it requires neither angiography nor selective catheterization, eliminates the risk of vessel damage by the catheter, and can be used for vessels too small or inaccessible for catheterization. Frequencies used include 20 kHz (200), 40 kHz (201), 170 kHz (202,203), 300 kHz (183), 1 MHz (186,187,196, 203,204), and 2 MHz (205), at intensities from 0.25 to 10 W/cm2. In vitro studies have shown various levels of moderate thrombolytic improvements averaging 30% to 40% and required one to three hours of insonation to get the effect (206). Several studies confirmed lytic activity during transtemporal delivery of US using a transducer similar to a regular TCD transducer. Lower frequencies penetrate the skull more efficiently than higher frequencies. Standard physical therapy devices used a 1-MHz frequency for delivery, while TCD devices used a 2-MHz frequency to measure flow velocity.

153

In human clinical trials, recovery rates higher than expected with simple tPA treatment have been described (185,205,207,208). Better rates of recanalization have been seen with those treated with continuous US as well as tPA. Several reports showed the bleeding rate with this technique to be similar to that with simple tPA therapy. The largest of these studies, the CLOTBUST phase II study (205), used a standard TCD aimed by a skilled sonographer at MCA thrombus in 126 randomized acute ischemic stroke patients. The flow in the artery was observed, and intravenous tPA was given. Continuous full-power TCD was used for two hours, and flow was assessed intermittently. The US beam is quite narrow; thus it requires a highly skilled sonographer to target the occluded segment and keep the beam on target using specially designed head frame. The study showed that the technique is safe and that TCD enhances recanalization. Using lower frequencies (20 kHz to 1 MHz) than those used for diagnostic purposes, tPA-mediated clot degradation was shown to be as much as 50% more efficient when US was applied transcranially (182,184). As mentioned above, the CLOTBUST study using 2-MHz transcranial probes suggested enhancement of tPA activity with acceleration of arterial reperfusion, but so far did not demonstrate clinical improvement. Although encouraging, these data lack confirmation of vascular and brain tissue effects through criterion standard imaging procedures and are in contrast to experimental studies using diagnostic US plus tPA (209). Consequently, Transcranial Low-Frequency USMediated Thrombolysis in Brain Ischemia (TRUMBI) trial (206), a phase II, prospective, nonrandomized study at six German university stroke centers, was scheduled to test safety and practicability of thrombolytic therapy in acute stroke patients with combined application of tPA plus low-frequency US. A secondary objective was to compare clinical recovery and rates of recanalization, reperfusion, and infarct size as evidenced by serial MRI. Patients were alternately allocated a standard 0.9-mg/kg tPA treatment and a combination of tPA treatment with transcranial insonation of low-frequency pulse-wave mode US (NeuroFlowTM, Walnut Technologies, Andover, Massachusetts, U.S.) for 60 to 90 minutes. The study was prematurely stopped because 5 of 12 patients from the tPA-only group, but 13 of 14 patients treated with the tPA plus US, showed signs of bleeding in MRI. Within three days of treatment, five symptomatic hemorrhages occurred within the tPA plus US group (r-tPA) thrombolysis in humans using low-frequency US (6). Two reasons were considered to be responsible for the increased risk of hemorrhage, the thermal effect and disruption of the blood-brain barrier (BBB). A study by Fatar at al. (210) showed that brain temperature increases within two to five minutes of insonation. The brain temperature increase and cooling time, however, were in proportion to power level, and even with the highest intensity of 7 W/cm2 for 30 minutes, the maximum elevation of mean brain

154

Krejza

temperature was 0.98C, with the highest cooling time of 40 minutes. However, no deleterious side effects of this treatment were found in histological examination. Another study by Reinhard et al. (211) showed abnormal permeability of the BBB after insonation with low-frequency US generated by the NeuroFlow. It indicates that the observed excessive bleeding rate with low-frequency sonothrombolysis also involving atypical locations (such as the intraventricular or subarachnoid space) might in fact be attributable to primary disruption of the BBB. In comparison with routine 2-MHz Doppler devices, as used in the CLOTBUST study without hemorrhagic side effects, the applied device had a wider sonification field but comparable power. Transient disruption of the BBB by focused US has been described recently in animals when it is applied in the presence of preformed gas bubbles (9). Ultrastructural animal studies have, among other mechanisms, proposed endothelial injury with high power, but partial opening of tight junctions already with low-power insonation (212). A clue to the mechanism of BBB disruption is that it occurs distant to the target volume: standing waves near the bone at the borderzone of the large insonation field may have occurred during continuous insonation and lead to local heating or mechanical effects disrupting the BBB. Therefore, small-field insonation should likely be preferred for sonothrombolysis in acute ischemic stroke.

Microbubble-Augmented US Thrombolysis MBs are small air- or gas-filled microspheres with specific acoustic properties that make them useful as US contrast agents for sonographic examinations. Two agents (Definity, Bristol-Meyers-Squibb Medical Imaging, Inc., Princeton, New Jersey, U.S., and Optison, Amersham Health, Princeton, New Jersey, U.S.) are commercially available with FDA approval for use in clarifying the outlines of the ventricles in cardiac US imaging. In diagnostic US, MBs create an acoustic impedance mismatch between fluids and body tissues, increasing the reflection of sound. Experimental studies have shown that US-accelerated thrombolysis may be further enhanced by administration of MBs (202,213,214). Low-frequency US with high power has been demonstrated to produce cavitation and fluid motion into the thrombus (214,215). MBs, by acting as cavitation nuclei, lower the amount of energy needed for cavitation. Application of high-acoustic-pressure US has been shown to induce nonlinear oscillations of MBs, leading to a continuous absorption of energy until the bubbles explode, releasing the absorbed energy (216). Thus, US-mediated MBs destruction may further accelerate the clot-dissolving effect of US. The synergistic effect of US and MBs on sonothrombolysis has been demonstrated in clinical studies in patients with arteriovenous dialysis graft thrombosis (217). Molina et al. investigated the effects of galactose-based MBs on the beginning, degree, and time to maximum completeness of MCA recanalization after application of tPA intravenously plus 2-hour

continuous 2-MHz TCD monitoring plus three boluses of 400 mg/dL of the galactose-based MBs (Levovist), given at 2, 20, and 40 minutes after tPA administration. (218). They showed that administration of MBs further enhances US-augmented systemic thrombolysis in acute ischemic stroke, leading to a more complete arterial recanalization and to a trend toward better short- and long-term outcome. Further research is required to evaluate possible combinations of thrombolytic drugs, MBs, and various modes of US delivery. Once these combinations can be assessed, some new techniques should be ready for application in humans. Currently, human CLOTBUST studies are progressing rapidly and involve not only thrombolytic drugs but the addition of MBs and of dedicated machines to make US delivery easier and more reliable.

REFERENCES 1. Hoeks APG, Reneman SR. Biophysical principles of vascular diagnosis. J Clin Ultrasound 1995; 23:71–79. 2. Taylor KJW, Holland S. Doppler US. I. Basic principles, instrumentation, and pitfalls. Radiology 1990; 174: 297–307. 3. Nyborg WL. Biological effects of ultrasound: development of safety guidelines. Part II: General review. Ultrasound Med Biol 2001; 27:301–333. 4. Barnett SB, Ter Haar GR, Ziskin MC, et al. International recommendations and guidelines for the safe use of diagnostic ultrasound in medicine. Ultrasound Med Biol 2000; 26:355–366. 5. Wu JR, Winkler AJ, Oneill TP. Effect of acoustic streaming on ultrasonic heating. Ultrasound Med Biol 1994; 20:195–201. 6. Daffertshofer M, Gass A, Ringleb P, et al. Transcranial lowfrequency ultrasound-mediated thrombolysis in brain ischemia: increased risk of hemorrhage with combined ultrasound and tissue plasminogen activator: results of a phase II clinical trial. Stroke 2005; 36:1441–1446. 7. Mariak Z, Krejza J, Swiercz M, et al. Human brain temperature in vivo: lack of heating during color transcranial Doppler ultrasonography. J Neuroimaging 2001; 11:308–312. 8. Kondo T, Kodaira T, Kano E. Free-radical formation induced by ultrasound and its effects on strand breaks in DNA of cultured fm3a cells. Free Radical Res Comm 1993; 19:S193–S200. 9. Hynynen K, McDannold N, Sheikov NA, et al. Local and reversible blood-brain barrier disruption by noninvasive focused ultrasound at frequencies suitable for trans-skull sonications. Neuroimage 2005; 24:12–20. 10. Krejza J, Mariak Z, Melhem ER, et al. A guide to the identification of major cerebral arteries with transcranial color Doppler sonography. AJR Am J Roentgenol 2000; 174:1297–1303. 11. Krejza J, Rudzinski W, Mariak Z. [Transcranial colorcoded Doppler sonography: getting started]. Neurol Neurochir Pol 2001; 35(suppl 5):101–109. 12. Krejza J, Swiat M, Pawlak M, et al. Suitability of temporal bone acoustic window: conventional TCD versus transcranial color-coded duplex sonography. J Neuroimaging 2007 (in press). 13. Ringelstein EB, Kahlscheuer B, Niggemeyer E, et al. Transcranial Doppler sonography: anatomical landmarks and normal velocity values. Ultrasound Med Biol 1990; 16:745–761. 14. Krejza J, Mariak Z, Walecki J, et al. Transcranial color Doppler sonography of basal cerebral arteries in 182

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

healthy subjects: age and sex variability and normal reference values for blood flow parameters. AJR Am J Roentgenol 1999; 172:213–218. Krejza J, Mariak Z, Huba M, et al. Effect of endogenous estrogen on blood flow through carotid arteries. Stroke 2001; 32:30–36. Belfort MA, Saade GR, Snabes M. Hormonal status affects the reactivity of the cerebral vasculature. Am J Obstet Gynecol 1995; 172:1273–1278. Brass LM, Pavlakis SG, DeVivo D, et al. Transcranial Doppler measurements of the middle cerebral artery. Effect of hematocrit. Stroke 1988; 19:1466–1469. Brown MM, Marshall J. Regulation of cerebral blood flow in response to changes in blood viscosity. Lancet 1985; 1:604–609. Krejza J, Mariak Z, Lewko J. Standardization of flow velocities with respect to age and gender improves accuracy of transcranial color Doppler ultrasonography in the diagnosis of middle cerebral artery spasm. AJR Am J Roentgenol 2003; 181:245–252. Albers GW, Clark WM, Madden KP, et al. ATLANTIS trial: results for patients treated within 3 hours of stroke onset. Stroke 2002; 33:493–495. Marler JR, Brott T, Broderick J, et al. Tissue-plasminogen activator for acute ischemic stroke. N Engl J Med 1995; 333:1581–1587. Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke: The PROACT II study: a randomized controlled trial. JAMA 1999; 282:2003–2011. Madden KP, Karanjia PN, Adams HP, et al. Accuracy of initial stroke subtype diagnosis in the TOAST study. Neurology 1995; 45:1975–1979. Caplan LR, Mohr JP, Kistler JP, et al. Should thrombolytic therapy be the first-line treatment for acute ischemic stroke? Thrombolysis: not a panacea for ischemic stroke. N Engl J Med 1997; 337:1309–1313. Alexandrov AV, Bladin CF, Norris JW. Intracranial blood flow velocities in acute ischemic stroke. Stroke 1994; 25:1378–1383. Maurer M, Shambal S, Berg D, et al. Differentiation between intracerebral hemorrhage and ischemic stroke by transcranial color-coded duplex-sonography. Stroke 1998; 29:2563–2567. Martin PJ, Pye I, Abbott R, et al. Color-coded ultrasound diagnosis of vascular occlusion in acute ischemic stroke. J Neuroimaging 1995; 5:152–156. Kluytmans M, van der Grond J, van Everdingen KJ, et al. Cerebral hemodynamics in relation to patterns of collateral flow. Stroke 1999; 30:1432–1439. Demchuk AM, Christou I, Wein TH, et al. Specific transcranial Doppler flow findings related to the presence and site of arterial occlusion. Stroke 2000; 31:140–144. Stolz E, Mendes I, Gerriets T, et al. Assessment of intracranial collateral flow by transcranial color-coded duplex sonography using a temporal and frontal axial insonation plane. J Neuroimaging 2002; 12:136–143. Kenton AR, Martin PJ, Abbott RJ, et al. Comparison of transcranial color-coded sonography and magnetic resonance angiography in acute stroke. Stroke 1997; 28: 1601–1606. Goertler M, Kross R, Baeumer M, et al. Diagnostic impact and prognostic relevance of early contrast-enhanced transcranial color-coded duplex sonography in acute stroke. Stroke 1998; 29:955–962. Gerriets T, Goertler M, Stolz E, et al. Feasibility and validity of transcranial duplex sonography in patients with acute stroke. J Neurol Neurosurg Psychiatry 2002; 73:17–20.

155

34. Nabavi DG, Droste DW, Kemeny V, et al. Potential and limitations of echocontrast-enhanced ultrasonography in acute stroke patients. Stroke 1998; 29:949–954. 35. Postert T, Federlein J, Braun B, et al. Contrast-enhanced transcranial color-coded real-time sonography: a reliable tool for the diagnosis of middle cerebral artery trunk occlusion in patients with insufficient temporal bone window. Stroke 1998; 29:1070–1073. 36. Zanette EM, Fieschi C, Bozzao L, et al. Comparison of cerebral angiography and transcranial Doppler sonography in acute stroke. Stroke 1989; 20:899–903. 37. Demchuk AM, Burgin WS, Christou I, et al. Thrombolysis in brain ischemia (TIBI) transcranial Doppler flow grades predict clinical severity, early recovery, and mortality in patients treated with intravenous tissue plasminogen activator. Stroke 2001; 32:89–93. 38. Camerlingo M, Casto L, Censori B, et al. Prognostic use of ultrasonography in acute non-hemorrhagic carotid stroke. Ital J Neurol Sci 1996; 17:215–218. 39. Baracchini C, Manara R, Ermani M, et al. The quest for early predictors of stroke evolution. Can TCD be a guiding light? Stroke 2000; 31:2942–2947. 40. Toni D, Fiorelli M, Zanette EM, et al. Early spontaneous improvement and deterioration of ischemic stroke patients. Stroke 1998; 29:1148. 41. Kushner M, Zanette EM, Bastianello S, et al. Transcranial Doppler in acute hemispheric brain infarction. Neurology 1991; 41:109–113. 42. Wong KS, Li H, Chan YL, et al. Use of transcranial Doppler ultrasound to predict outcome in patients with intracranial large-artery occlusive disease. Stroke 2000; 31:2641–2647. 43. Alexandrov AV, Black SE, Ehlieh LE, et al. Predictors of hemorrhagic transformation occuring spontaneously and on anticoagulants in patients with acute ischemic stroke. Stroke 1197; 28:1198–1202. 44. Ringelstein B, Biniek R, Weiller C, et al. Type and extent of hemispheric brain infarctions and clinical outcome in early and delayed middle cerebral artery recanalization. Neurology 1992; 42:289–298. 45. Gerriets T, Postert T, Goertler M, et al. DIAS I: Duplex sonographic assessment of the cerebrovascular status in acute stroke: a useful tool for future stroke trials. Stroke 2000; 31:2342–2345. 46. Molina CA, Mantaner J, Abilleira S, et al. Timing of spontaneous recanalization and risk of hemorrhagic transformation in acute cardioembolic stroke. Stroke 2001; 32:1079–1084. 47. Stolz E, Nuckel M, Mendes I, et al. Vertebrobasilar transcranial color-coded duplex ultrasonography: improvement with echo enhancement. AJNR Am J Neuroradiol 2002; 23:1051–1054. 48. Gerriets T, Seidel G, Modrau B, et al. Contrast-enhanced transcranial color-coded duplex sonography: efficiency and validity. Neurology 1999; 52:1133–1137. 49. Seidel G, Algermissen C, Christoph A, et al. Harmonic imaging of the human brain. Visualization of brain perfusion with ultrasound. Stroke 2000; 31:151–155. 50. Pohl C, Tiemann K, Schlosser T, et al. Stimulated acoustic emission detected by transcranial color Doppler ultrasound: a contrast-specific phenomenon useful for the detection of cerebral tissue perfusion. Stroke 2000; 31:1661–1666. 51. Deng CX, Lizzi FL. A review of physical phenomena associated with ultrasonic contrast agents and illustrative clinical applications. Ultrasound Med Biol 2002; 28:277–286. 52. Sacco R, Kargman D, Gu Q, et al. Race-ethnicity and determinants of intracranial atherosclerotic cerebral

156

53. 54. 55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

Krejza infarction. The Northern Manhattan Stroke Study. Stroke 1995; 26:14–20. Mohr JP, Albers GW, Amarenco P, et al. Etiology of stroke. Stroke 1997; 28:1501–1506. Song JK, Eskridge JM. Intracranial angioplasty and thrombolysis. Neurosurg Clin N Am 2000; 11:49–65. Gress DR, Smith WS, Dowd CF, et al. Angioplasty for intracranial symptomatic vertebrobasilar ischemia. Neurosurgery 2002; 51:23–27. Babikian VL, Feldmann E, Wechsler LR, et al. Transcranial Doppler ultrasonography: year 2000 update. J Neuroimaging 2000; 10:101–115. Rorick MB, Nichols FT, Adams RJ. Transcranial Doppler correlation with angiography in detection of intracranial stenosis. Stroke 1994; 25:1931–1934. Ley-Pozo J , Ringelstein EB. Noninvasive detection of occlusive disease of the carotid siphon and middle cerebral artery. Ann Neurol 1990; 28:640–647. Babikian V, Sloan MA, Tegeler CH, et al. Transcranial Doppler validation pilot study. J Neuroimaging 1993; 3:242–249. Baumgartner RW, Mattle HP, Schroth G. Assessment of >/¼50% and <50% intracranial stenoses by transcranial color-coded duplex sonography. Stroke 1999; 30:87–92. Kimura K, Yasaka M, Wada K, et al. Diagnosis of middle cerebral artery stenosis by transcranial color-coded realtime sonography. AJNR Am J Neuroradiol 1998; 19:1893– 1896. Mull M, Aulich A, Hennerici M. Transcranial Doppler ultrasonography versus arteriography for assessment of the vertebrobasilar circulation. J Clin Ultrasound 1990; 539–549. Babikian VL, Gomes J, Krejza J. Assessment of cerebrovascular pathophysiology. In: Pinsky RM, ed. Cerebral Blood Flow: Mechanisms of Ischemia, Diagnosis and Theraphy. Berlin, Heidelberg: Springer, 2002:201–216. Christou I, Felberg RA, Demchuk AM, et al. A broad diagnostic battery for bedside transcranial Doppler to detect flow changes with internal carotid artery stenosis or occlusion. J Neuroimaging 2001; 11:236–242. Bay-Hansen J, Ravn T, Knudsen GM. Application of interhemispheric index for transcranial Doppler sonography velocity measurements and evaluation of recording time. Stroke 1997; 28:1009–1014. Mattle HP, Grolimund P, Huber P, et al. Transcranial Doppler sonographic findings in middle cerebral artery disease. Arch Neurol 1988; 45:289–295. Arenillas JF, Molina CA, Montaner J, et al. Progression and clinical recurrence of symptomatic middle cerebral artery stenosis: a long-term follow-up transcranial Doppler ultrasound study. Stroke 2001; 32:2898–2904. Schwarze JJ, Babikian VL, DeWitt LD, et al. Longitudinal monitoring of intracranial arterial stenoses with transcranial Doppler ultrasonography. J Neuroimaging 1994; 4:182–187. Wong KS, Li H, Lam WWM, et al. Progression of middle cerebral artery occlusive disease and its relationship with further vascular events after stroke. Stroke 2002; 33:532–536. Mariak Z, Krejza J, Swiercz M, et al. Accuracy of transcranial color Doppler ultrasonography in the diagnosis of middle cerebral artery spasm determined by receiver operating characteristic analysis. J Neurosurg 2002; 96:323–330. Lysakowski C, Walder B, Costanza MC, et al. Transcranial Doppler versus angiography in patients with vasospasm due to a ruptured cerebral aneurysm: a systematic review. Stroke 2001; 32:2292–2298. Vora YY, Suarez-Almazor M, Steinke DE, et al. Role of transcranial Doppler monitoring in the diagnosis of

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

cerebral vasospasm after subarachnoid hemorrhage. Neurosurgery 1999; 44:1237–1248. Grosset DG, Straiton J, McDonald I, et al. Use of transcranial Doppler sonography to predict development of a delayed ischemic deficit after subarachnoid hemorrhage. J Neurosurg 1993; 78:183–187. Krejza J, Kochanowicz J, Mariak Z, et al. Middle cerebral artery spasm after subarachnoid hemorrhage: detection with transcranial color-coded duplex US. Radiology 2005; 236:621–629. Lindegaard KF, Nornes H, Bakke SJ, et al. Cerebral vasospasm after subarachnoid haemorrhage investigated by means of transcranial Doppler ultrasound. Acta Neurochir Suppl (Wien) 1988; 42:81–84. Krejza J, Szydlik P, Liebeskind DS, et al. Age and sex variability and normal reference values for the V-MCA/ V-ICA index. AJNR Am J Neuroradiol 2005; 26:730–735. Swiercz M, Krejza J, Mariak Z. Detection of middle cerebral artery spasm with learning vector quantization artificial neural networks. Biocybern Biomed Eng 2004; 24:3–16. Proust F, Callonec F, Clavier E, et al. Usefulness of transcranial color-coded sonography in the diagnosis of cerebral vasospasm. Stroke 1999; 30:1091–1098. Soustiel JF, Bruk B, Shik B, et al. Transcranial Doppler in vertebrobasilar vasospasm after subarachnoid hemorrhage. Neurosurgery 1998; 43:282–293. Soustiel JF, Shik V, Schreiber V, et al. Basilar vasospasm diagnosis. Investigation of a modified ‘‘Lindegaard Index’’ based on imaging studies and blood velocity measurements of the basilar artery. Stroke 2002; 33:72–78. Krejza J, Szydlik P, Liebeskind DS, et al. Age and sex variability and normal reference values for the V(MCA)/ V(ICA) index. AJNR Am J Neuroradiol 2005 Apr; 26(4): 730–735. Sloan MA, Haley EC, Kassel NF, et al. Sensitivity and specificity of transcranial Doppler ultrasonography in the diagnosis of vasospasm following subarachnoid hemorrhage. Neurology 1989; 39:1514–1518. Newell D. Distribution of angiographic vasospasm after subarachnoid hemorrage: implications for diagnosis by TCD. Neurosurgery 1990; 27:574–577. Wardlaw JM, Offin R, Teasdale GM, et al. Is routine transcranial Doppler ultrasound monitoring useful in the management of subarachnoid hemorrhage? J Neurosurg 1998; 88:272–276. Murayama Y, Malisch T, Guglielmi G, et al. Incidence of cerebral vasospasm after endovascular treatment of acutely ruptured aneurysms: report on 69 cases. J Neurosurg 1997; 87:830–835. Muizelaar JP, Zweinenberg M, Rudisill NA, et al. The prophylactic use of transluminal balloon angioplasty in patients with Fisher grade 3 subarachnoid hemorrhage: a pilot study. J Neurosurg 1999; 91:51–58. Naidu S, Payne AJ, Moodley J, et al. Randomised study assessing the effect of phenytoin and magnesium sulphate on maternal cerebral circulation in eclampsia using transcranial Doppler ultrasound. Br J Obstet Gynaecol 1996; 103:111–116. Bailey IC, Lyttle JA, Matthew B, et al. Effect of calcitoningene-related peptide in patients with delayed postoperative cerebral-ischemia after aneurysmal subarachnoid hemorrhage. Lancet 1992; 339:831–834. Findlay JM, Kassell NF, Weir BKA, et al. A randomized trial of intraoperative, intracisternal tissue-plasminogen activator for the prevention of vasospasm. Neurosurgery 1995; 37:168–176. Baumgartner RW, Mattle HP, Kothbauer K, et al. Transcranial color-coded duplex sonography in cerebral aneurysms. Stroke 1994; 25:2429–2434.

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology 91. Martin PJ, Gaunt ME, Naylor AR, et al. Intracranial aneurysms and arteriovenous malformations: transcranial colour-coded sonography as a diagnostic aid. Ultrasound Med Biol 1994; 20:689–698. 92. Wardlaw JM, Cannon JC. Color transcranial ‘‘power’’ Doppler ultrasound of intracranial aneurysms. J Neurosurg 1996; 84:459–461. 93. Schuknecht B, Fandino J, Yuksel C, et al. Endovascular treatment of cerebral vasospasm: assessment of treatment effect by cerebral angiography and transcranial colour Doppler sonography. Neuroradiology 1999; 41:453–462. 94. Turner CL, Higgins JNP, Kirkpatrick PJ. Assessment of transcranial color-coded duplex sonography for the surveillance of intracranial aneurysms treated with Guglielmi detachable coils. Neurosurgery 2003; 53:866–871. 95. Turner CL, Higgins JNP, Gholkar A, et al. Intracranial aneurysms treated with endovascular coils: detection of recurrences using unenhanced and contrast-enhanced transcranial color-coded duplex sonography. Stroke 2005; 36:2654–2659. 96. Krejza J, Mariak Z, Bert RJ. Transcranial colour Doppler sonography in emergency management of intracerebral haemorrhage caused by an arteriovenous malformation: case report. Neuroradiology 2000; 42:900–904. 97. Klotzch C, Henkes H, Nahser HC, et al. Transcranial color-coded duplex sonography in cerebral arteriovenous malformations. Stroke 1995; 26:2298–2301. 98. Bartels E, Rodiek SO, Lumenta C, et al. Evaluation of arteriovenous malformations with transcranial colorcoded duplex ultrasonography. J Ultrasound Med 1998; 17:166–169. 99. Diehl RR, Henkes H, Nahser HC, et al. Blood flow velocity and vasomotor reactivity in patients with arteriovenous malformations. A transcranial Doppler study. Stroke 1994; 25:1574–1580. 100. Baumgartner RW, Mattle HP, Schroth G. Transcranial colour-coded duplex sonography of cerebral arteriovenous malformations. Neuroradiology 1996; 38:734–737. 101. Chioffi F, Pasqualin A, Beltramello A, et al. Hemodynamic effects of preoperative embolization in cerebral arteriovenous malformations: evaluation with transcranial Doppler sonography. Neurosurgery 1992; 31:877–884. 102. Manchola IF, De Salles AAF, Kok Foo T, et al. Arteriovenous malformation hemodynamics: a transcranial Doppler study. Neurosurgery 1993; 33:556–562. 103. Ries S, Steinke W, Neff KW, et al. Echocontrast-enhanced transcranial color-coded sonography for the diagnosis of transverse venous thrombosis. Stroke 1997; 28:696–700. 104. Becker G, Bogdahn U, Gehlberg C, et al. Transcranial color-coded real-time sonography of intracranial veins. J Neuroimaging 1995; 5:87–94. 105. Valdueza MJ, Schultz M, Harms L, et al. Venous transcranial Doppler ultrasound monitoring in acute dural sinus thrombosis. Stroke 1995; 26:1196–1199. 106. Schoser BGH, Riemenschneider N, Hansen HC. The impact of raised intracranial pressure on cerebral venous hemodynamics: a prospective venous transcranial Doppler ultrasonography study. J Neurosurg 1999; 91:744–749. 107. Ringelstein EB, Droste DW, Babikian VL, et al. Consensus on microembolus detection by TCD. Stroke 1998; 29:725–729. 108. Forteza AM, Krejza J, Koch S, et al. Ultrasound Imaging of Cerebrovascular Disease. In: Babikian VL, Wechsler L, Higashida RT, eds. Imaging Cerebrovascular Disease. Philadelphia: Butterworth-Heinemann, 2003:3–35. 109. Smith JL, Evans DH, Bell PRF, et al. A comparison of four methods for distinguishing Doppler signals from gaseous and particulate emboli. Stroke 1998; 29:1133–1138. 110. Markus HS, Molloy J. Use of a decibel threshold in detecting Doppler embolic signals. Stroke 1997; 28:692–695.

157

111. Cullinane M, Reid G, Dittrich R, et al. Evaluation of new online automated embolic signal detection algorithm, including comparison with panel of international experts. Stroke 2000; 31:1335–1341. 112. Markus H, Loh A, Brown MM. computerized detection of cerebral emboli and discrimination from artifact using Doppler ultrasound. Stroke 1993; 24:1667–1672. 113. Molloy J, Khan NH, Markus HS. Temporal variability of asymptomatic embolization in carotid artery stenosis and optimal recording protocols. Stroke 1998; 29:1129–1132. 114. Goertler M, Blaser T, Krueger S, et al. Cessation of embolic signals after antithrombotic prevention is related to reduced risk of recurrent arterioembolic transient ischaemic attack and stroke. J Neurol Neurosurg Psychiatry 2002; 72:338–342. 115. Molloy J, Markus HS. Asymptomatic embolization predicts stroke and TIA risk in patients with carotid artery stenosis. Stroke 1999; 30:1440–1443. 116. Stork JL, Kimura K, Levi CR, et al. Source of microembolic signals in patients with high-grade carotid stenosis. Stroke 2002; 33:2014–2018. 117. Jordan WD, Voellinger DC, Doblar DD, et al. Microemboli detected by transcranial Doppler monitoring in patients during carotid angioplasty versus carotid endarterectomy. Cardiovasc Surg 1999; 7:33–38. 118. Reimers B, Corvaja N, Moshiri S, et al. Cerebral protection with filter devices during carotid artery stenting. Circulation 2001; 104:12–15. 119. Rosenkranz M, Fiehler J, Niesen W, et al. The amount of solid cerebral microemboli during carotid stenting does not relate to the frequency of silent ischemic lesions. AJNR Am J Neuroradiol 2006; 27:157–161. 120. Maleux G, Demaerel P, Verbeken E, et al. Cerebral ischemia after filter-protected carotid artery stenting is common and cannot be predicted by the presence of substantial amount of debris captured by the filter device. AJNR Am J Neuroradiol 2006; 27:1830–1833. 121. Ernst JMPG, Suttorp MJ, Tenberg JM, et al. Transcranial Doppler monitoring in angioplasty and stenting of the carotid bifurcation. J Endovasc Ther 2003; 10:702–710. 122. Ackerstaff RGA, Suttorp MJ, van den Berg JC, et al. Prediction of early cerebral outcome by transcranial Doppler monitoring in carotid bifurcation angioplasty and stenting. J Vasc Surg 2005; 41:618–624. 123. Lu CJ, Kao HL, Sun Y, et al. The hemodynamic effects of internal carotid artery stenting: a study with color-coded duplex sonography. Cerebrovasc Dis 2003; 15:264–269. 124. Markus HS, Thomson ND, Brown MM. Asymptomatic cerebral embolic signals in symptomatic and asymptomatic carotid-artery disease. Brain 1995; 118:1005–1011. 125. Sawada M, Hashimoto N, Nishi S, et al. Detection of embolic signals during and after percutaneous transluminal angioplasty of subclavian and vertebral arteries using transcranial Doppler ultrasonography. Neurosurgery 1997; 41:535–540. 126. Pelz DM, Lownie SP, Fox AJ. Thromboembolic events associated with the treatment of cerebral aneurysms with Guglielmi detachable coils. AJNR Am J Neuroradiol 1998; 19:1541–1547. 127. Workman MJ, Cloft HJ, Tong FC, et al. Thrombus formation at the neck of cerebral aneurysms during treatment with Guglielmi detachable coils. AJNR Am J Neuroradiol 2002; 23:1568–1576. 128. Song JK, Niimi Y, Fernandez PM, et al. Thrombus formation during intracranial aneurysm coil placement: treatment with intra-arterial abciximab. AJNR Am J Neuroradiol 2004; 25:1147–1153. 129. Ries T, Buhk JH, Kucinski T, et al. Intravenous administration of acetylsalicylic acid during endovascular

158

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

Krejza treatment of cerebral aneurysms reduces the rate of thromboembolic events. Stroke 2006; 37:1816–1821. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis: a report from the committee on vascular-lesions of the council on arteriosclerosis, American-Heart-Association. Circulation 1995; 92:1355–1374. Barnett HJM, Taylor W, Eliasziw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. N Engl J Med 1998; 339:1415–1425. Warlow C. MRC-European-Carotid-Surgery-Trial: interim results for symptomatic patients with severe (70-99percent) or with mild (0-29-percent) carotid stenosis. Lancet 1991; 337:1235–1243. Walker MD, Marler JR, Goldstein M, et al. Endarterectomy for asymptomatic carotid-artery stenosis. JAMA 1995; 273:1421–1428. New G, Roubin GS, Oetgen ME, et al. Validity of duplex ultrasound as a diagnostic modality for internal carotid artery disease. Catheter Cardiovasc Interv 2001; 52:9–15. Johnston DCC, Goldstein LB. Clinical carotid endarterectomy decision making: noninvasive vascular imaging versus angiography. Neurology 2001; 56:1009–1015. Benedetti-Valentini F, Gossetti B, Irace L, et al. Carotid surgery without angiography is possible and safe. J Cardiovasc Surg 2000; 41:601–605. Thusay MM, Khoury M, Greene K. Carotid endarterectomy based on duplex ultrasound in patients with and without hemispheric symptoms. Am Surg 2001; 67:1–6. Sabeti S, Schillinger M, Mlekusch W, et al. Quantification of internal carotid artery stenosis with duplex US: comparative analysis of different flow velocity criteria. Radiology 2004; 232:431–439. Hunink MG, Polak JF, Barlan MM, et al. Detection and quantification of carotid artery stenosis: efficacy of various Doppler velocity parameters. AJR Am J Roentgenol 1993; 160:619–625. Bonig L, Weder B, Schott D, et al. Prediction of angiographic carotid artery stenosis indexes by colour Dopplerassisted duplex imaging. A critical appraisal of the parameters used. Eur J Neurol 2000; 7:183–190. Perkins JMT, Galland RB, Simmons MJ, et al. Carotid duplex imaging: variation and validation. Br J Surg 2000; 87:320–322. Hoskins PR. Accuracy of maximum velocity estimates made using Doppler ultrasound systems. Br J Radiol 1996; 69:172–177. Criswell BK, Langsfeld M, Tullis MJ, et al. Evaluating institutional variability of duplex scanning in the detection of carotid artery stenosis. Am J Surg 1998; 176:591–597. Moneta GL, Edwards JM, Chitwood RW, et al. Correlation of North-American Symptomatic Carotid Endarterectomy Trial (NASCET) angiographic definition of 70-percent to 99-percent internal carotid-artery stenosis with duplex scanning. J Vasc Surg 1993; 17:152–159. Faught WE, Mattos MA, Vanbemmelen PS, et al. Colorflow duplex scanning of carotid arteries: new velocity criteria-based on receiver operator characteristic analysis for threshold stenoses used in the symptomatic and asymptomatic carotid trials. J Vasc Surg 1994; 19:818–828. Neale ML, Chambers JL, Kelly AT, et al. Reappraisal of duplex criteria to assess significant carotid stenosis with special reference to reports from the north-american symptomatic carotid endarterectomy trial and the European Carotid Surgery Trial. J Vasc Surg 1994; 20:642–649. Carpenter JP, Lexa FJ, Davis JT. Determination of duplex Doppler ultrasound criteria appropriate to the North

148.

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.

162.

163. 164.

165.

166.

167.

168.

American Symptomatic Carotid Endarterectomy Trial. Stroke 1996; 27:695–699. Hood DB, Mattos MA, Mansour A, et al. Prospective evaluation of new duplex criteria to identify 70% internal carotid artery stenosis. J Vasc Surg 1996; 23:254–261. Huston J, James EM, Brown RD, et al. Redefined duplex ultrasonographic criteria for diagnosis of carotid artery stenosis. Mayo Clin Proc 2000; 75:1133–1140. Moneta GL, Edwards JM, Papanicolaou G, et al. Screening for asymptomatic internal carotid-artery stenosis: duplex criteria for discriminating 60-percent to 99-percent stenosis. J Vasc Surg 1995; 21:989–994. Carpenter JP, Lexa FJ, Davis JT. Determination of sixty percent or greater carotid artery stenosis by duplex Doppler ultrasonography. J Vasc Surg 1995; 22:697–705. Jackson MR, Chang AS, Robles HA, et al. Determination of 60% or greater carotid stenosis: a prospective comparison of magnetic resonance angiography and duplex ultrasound with conventional angiography. Ann Vasc Surg 1998; 12:236–243. Ballard JL, Fleig K, Delange M, et al. The diagnosticaccuracy of duplex ultrasonography for evaluating carotid bifurcation. Am J Surg 1994; 168:123–126. Paivansalo M, Leinonen S, Turunen J, et al. Quantification of carotid artery stenosis with various Doppler velocity parameters. Rofo 1996; 164:108–113. Winkelaar GB, Chen JC, Salvian AJ, et al. New duplex ultrasound scan criteria for managing symptomatic 50% or greater carotid stenosis. J Vasc Surg 1999; 29:986–993. Oszkinis G, Pukacki F, Juszkat R, et al. Restenosis after carotid endarterectomy: incidence and endovascular management. Intervent Neuroradiol 2008 (in press). Brott T, Toole JF. Medical compared with surgical-treatment of asymptomatic carotid-artery stenosis. Ann Inter Med 1995; 123:720–722. Moore WS, Kempczinski RF, Nelson JJ, et al. Recurrent carotid stenosis: results of the asymptomatic carotid atherosclerosis study. Stroke 1998; 29:2018–2025. Frericks H, Kievit J, van Baalen JM, et al. Carotid recurrent stenosis and risk of ipsilateral stroke: a systematic review of the literature. Stroke 1998; 29:244–250. Steinke W, Rautenberg W, Schwartz A, et al. Noninvasive monitoring of internal carotid-artery dissection. Stroke 1994; 25:998–1005. Sturzenegger M, Mattle HP, Rivoir A, et al. Ultrasound findings in carotid-artery dissection: analysis of 43 patients. Neurology 1995; 45:691–698. Pelkonen O, Tikkakoski T, Leinonen S, et al. Extracranial internal carotid and vertebral artery dissections: angiographic spectrum, course and prognosis. Neuroradiology 2003; 45:71–77. Hoffmann M, Sacco RL, Chan S, et al. Noninvasive detection of vertebral artery dissection. Stroke 1993; 24:815–819. Gortler M, Niethammer R, Widder B. Differentiating subtotal carotid-artery stenoses from occlusions by color-coded duplex sonography. J Neurol 1994; 241:301–305. Ferrer JME, Samso JJ, Serrano JR, et al. Use of ultrasound contrast in the diagnosis of carotid artery occlusion. J Vasc Surg 2000; 31:736–741. Trattnig S, Hubsch P, Schuster H, et al. Color-coded Doppler imaging of normal vertebral arteries. Stroke 1990; 21:1222–1225. Kuhl V, Tettenborn B, Eicke BM, et al. Color-coded duplex ultrasonography of the origin of the vertebral artery: normal values of flow velocities. J Neuroimaging 2000; 10:17–21. Kimura K, Yasaka M, Moriyasu H, et al. Ultrasonographic evaluation of vertebral artery to detect vertebrobasilar axis occlusion. Stroke 1994; 25:1006–1009.

Chapter 7: Ultrasonographic Imaging and Physiological Techniques in Interventional Neuroradiology 169. Touboul PJ, Bousser MG, Laplane D, et al. Duplex scanning of normal vertebral arteries. Stroke 1986; 17:921–923. 170. Kotval PS, Babu SC, Shah PM. Doppler diagnosis of partial vertebral subclavian steals convertible to full steals with physiologic maneuvers. J Ultrasound Med 1990; 9:207–213. 171. Yip PK, Liu HM, Hwang BS, et al. Subclavian steal phenomenon: a correlation between duplex sonographic and angiographic findings. Neuroradiology 1992; 34:279–282. 172. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue-plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995; 274:1017–1025. 173. Kase CS, Furlan AJ, Wechsler LR, et al. Cerebral hemorrhage after intra-arterial thrombolysis for ischemic stroke: the PROACT II trial. Neurology 2001; 57:1603–1610. 174. Furlan A, Higashida RT, Wechsler L, et al. Intraarterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. JAMA 1999; 282:2003–2011. 175. Zaidat OO, Suarez JI, Santillan C, et al. Response to intraarterial and combined intravenous and intra-arterial thrombolytic therapy in patients with distal internal carotid artery occlusion. Stroke 2002; 33:1821–1826. 176. Lincoff AM, Topol EJ. Illusion of reperfusion. Does anyone achieve optimal reperfusion during acute myocardialinfarction. Circulation 1993; 88:1361–1374. 177. Every NR, Parsons LS, Hlatky M, et al. A comparison of thrombolytic therapy with primary coronary angioplasty for acute myocardial infarction. N Engl J Med 1996; 335:1253–1260. 178. Francis CW. Ultrasound-enhanced thrombolysis. Echocardiography 2001; 18:239–246. 179. Alexandrov AV, Burgin WS, Demchuk AM, et al. Speed of intracranial clot lysis with intravenous tissue plasminogen activator therapy: sonographic classification and shortterm improvement. Circulation 2001; 103:2897–2902. 180. Francis CW, Blinc A, Lee S, et al. Ultrasound accelerates transport of recombinant tissue-plasminogen activator into clots. Ultrasound Med Biol 1995; 21:419–424. 181. Blinc A, Kennedy SD, Bryant RG, et al. Flow-through clots determines the rate and pattern of fibrinolysis. Thromb Haemost 1994; 71:230–235. 182. Akiyama M, Ishibashi T, Yamada T, et al. Low-frequency ultrasound penetrates the cranium and enhances thrombolysis in vitro. Neurosurgery 1998; 43:828–833. 183. Behrens S, Daffertshofer M, Spiegel D, et al. Low-frequency, low-intensity ultrasound accelerates thrombolysis through the skull. Ultrasound Med Biol 1999; 25:269–273. 184. Spengos K, Behrens S, Daffertshofer M, et al. Acceleration of thrombolysis with ultrasound through the cranium in a flow model. Ultrasound Med Biol 2000; 26:889–895. 185. Alexandrov AV, Demchuk AM, Felberg AR, et al. High rate of complete recanalization and dramatic clinical recovery during tPA infusion when continuously monitored with 2-MHz transcranial Doppler monitoring. Stroke 2000; 31:610–614. 186. Francis CW, Onundarson PT, Carstensen EL, et al. Enhancement of fibrinolysis in vitro by ultrasound. J Clin Invest 1992; 90:2063–2068. 187. Blinc A, Francis CW, Trudnowski JL, et al. Characterization of ultrasound-potentiated fibrinolysis in vitro. Blood 1993; 81:2636–2643. 188. Siddiqi F, Odrljin TM, Fay PJ, et al. Binding of tissueplasminogen activator to fibrin: effect of ultrasound. Blood 1998; 91:2019–2025. 189. Siddiqi F, Blinc A, Braaten J, et al. Ultrasound increases flow-through fibrin gels. Thromb Haemost 1995; 73: 495–498.

159

190. Polak JF. Ultrasound energy and the dissolution of thrombosis. N Engl J Med 2004; 351:2154–2155. 191. Sakharov DV, Barrett-Bergshoeff M, Hekkenberg RT, et al. Fibrin-specificity of a plasminogen activator affects the efficiency of fibrinolysis and responsiveness to ultrasound: comparison of nine plasminogen activators in vitro. Thromb Haemost 1999; 81:605–612. 192. Krejza J, Weigele JB, Alokaili R, et al. Sonothrombolysis in acute ischemic stroke for patients ineligible for rt-PA. Neurology 2006; 66:154. 193. Hong AS, Chae JS, Dubin SB, et al. Ultrasonic clot disruption: an in vitro study. Am Heart J 1990; 120:418–422. 194. Rosenschein U, Bernstein JJ, Disegni E, et al. Experimental ultrasonic angioplasty: disruption of atherosclerotic plaques and thrombi in vitro and arterial recanalization in vivo. J Am Coll Cardiol 1990; 15:711–717. 195. Steffen W, Cumberland D, Gaines P, et al. Catheterdelivered high-intensity, low-frequency ultrasound induces vasodilation in vivo. Eur Heart J 1994; 15:369–376. 196. Luo H, Steffen W, Cercek B, et al. Enhancement of thrombolysis by external ultrasound. Am Heart J 1993; 125:1564–1569. 197. Siegel RJ, Cumberland DC, Crew JR. Ultrasound recanalization of diseased arteries: from experimental studies to clinical-application. Surg Clin N Am 1992; 72:879–895. 198. Rosenschein U, Rozenszajn LA, Kraus L, et al. Ultrasonic angioplasty in totally occluded peripheral arteries. Initial clinical, histological, and angiographic results. Circulation 1991; 83:1976–1986. 199. Tachibana K, Tachibana S. Prototype therapeutic ultrasound emitting catheter for accelerating thrombolysis. J Ultrasound Med 1997; 16:529–535. 200. Sehgal CM, Leveen RF, Shlanskygoldberg RD. Ultrasoundassisted thrombolysis. Invest Radiol 1993; 28:939–943. 201. Suchkova V, Siddiqi FN, Carstensen EL, et al. Enhancement of fibrinolysis with 40-kHz ultrasound. Circulation 1998; 98:1030–1035. 202. Tachibana K, Tachibana S. Albumin microbubble echocontrast material as an enhancer for ultrasound accelerated thrombolysis. Circulation 1995; 92:1148–1150. 203. Olsson SB, Johansson B, Nilsson AM, et al. Enhancement of thrombolysis by ultrasound. Ultrasound Med Biol 1994; 20:375–382. 204. Lauer CG, Burge R, Tang DB, et al. Effect of ultrasound on tissue-type plasminogen-activator induced thrombolysis. Circulation 1992; 86:1257–1264. 205. Alexandrov AV, Molina CA, Grotta JC, et al. Ultrasoundenhanced systemic thrombolysis for acute ischemic stroke. N Engl J Med 2004; 351:2170–2178. 206. Daffertshofer M, Hennerici M. Ultrasound in the treatment of ischaemic stroke. Lancet Neurol 2003; 2:283–290. 207. Eggers J, Koch B, Koenig I, et al. Transcranial ultrasound accelerates recanalization and improves outcome in stroke patients with acute middle cerebral artery occlusion. Circulation 2003; 108:602. 208. Cintas P, Le Traon AP, Larrue V. High rate of recanalisation of middle cerebral artery occusion during 2-MHz transcranial color-coded Doppler continuous monitoring without thrombolytic drug. Stroke 2002; 33:626–628. 209. Pfaffenberger S, Devcic-Kuhar B, Kollmann C, et al. Can a commercial diagnostic ultrasound device accelerate thrombolysis? An in vitro skull model. Stroke 2005; 36:124–128. 210. Fatar M, Stroick M, Griebe M, et al. Brain temperature during 340-kHz pulsed ultrasound insonation: a safety study for sonothrombolysis. Stroke 2006; 37:1883–1887. 211. Reinhard M, Hetzel A, Kruger S, et al. Blood-brain barrier disruption by low-frequency ultrasound. Stroke 2006; 37:1546–1548.

160

Krejza

212. Sheikov N, McDannold N, Jolesz F, et al. Brain arterioles show more active vesicular transport of blood-borne tracer molecules than capillaries and venules after focused ultrasound-evoked opening of the blood-brain barrier. Ultrasound Med Biol 2006; 32:1399–1409. 213. Mizushige K, Kondo I, Ohmori K, et al. Enhancement of ultrasound-accelerated thrombolysis by echo contrast agents: dependence on microbubble structure. Ultrasound Med Biol 1999; 25:1431–1437. 214. Culp WC, Porter TR, Lowery J, et al. Intracranial clot lysis with intravenous microbubbles and transcranial ultrasound in swine. Stroke 2004; 35:2407–2411. 215. Cintas P, Nguyen F, Boneu B, et al. Enhancement of enzymatic fibrinolysis with 2-MHz ultrasound and microbubbles. J Thromb Haemost 2004; 2:1163–1166.

216. Dijkmans PA, Knaapen P, Sieswerda GTJ, et al. Quantification of myocardial perfusion using intravenous myocardial contrast echocardiography in healthy volunteers: comparison with positron emission tomography. J Am Soc Echocardiogr 2006; 19:285–293. 217. Xie F, Tsutsui JM, Lof J, et al. Effectiveness of lipid microbubbles and ultrasound in declotting thrombosis. Ultrasound Med Biol 2005; 31:979–985. 218. Molina CA, Ribo M, Rubiera M, et al. Microbubble administration accelerates clot lysis during continuous 2-MHz ultrasound monitoring in stroke patients treated with intravenous tissue plasminogen activator. Stroke 2006; 37:425–429.

8 Techniques and Devices in Interventional Neuroradiology Jeffrey M. Katz Department of Radiology, New York Presbyterian Hospital, Weill Medical College of Cornell University, New York, U.S.A.

Y. Pierre Gobin and Howard A. Riina Departments of Radiology and Neurosurgery, New York Presbyterian Hospital, Weill Medical College of Cornell University, New York, U.S.A.

INTRODUCTION The endovascular surgeons’ armamentarium is expanding at an exponential rate. New tools and materials are being developed and introduced into clinical practice at a dizzying pace, while current devices are continuously being modified for improved safety, efficacy, and navigation in the tortuous cerebrovascular environment. With these advancements, standard angiographic and interventional techniques are evolving and innovative methods and procedures are being introduced into the endovascular arena, enabling more effective treatment of a larger variety of neurovascular and spinal diseases that were previously too challenging to access or complex to successfully cure. In this chapter, we discuss techniques commonly used in neurointerventional practice. While a comprehensive review of all the devices available to the endovascular surgeon is beyond the scope of this chapter, we provide illustrative examples of the devices used in the endovascular treatment of intracranial, extracranial/cervical, and spinal diseases in our clinical practice.

PREOPERATIVE EVALUATION All endovascular procedures begin with preoperative patient assessment and, when appropriate, educating the patient about the intended procedure (1). The indications for the interventional procedure should be reviewed by comprehensive patient evaluation and discussions with the patient’s referring physician. The patient’s existing clinical condition, past medical and surgical histories, medications, and allergy history, including any previous exposure to contrast dye and whether or not the patient had any adverse effects from that experience, need to be ascertained. If a contrast allergy is suspected, premedication with prednisone 50 mg orally at 24, 12, and 1 hour as well as diphenhydramine 50 mg orally or intravenously 1-hour preoperatively is the standard protocol to prevent an allergic reaction. Physical examination

should include an assessment of vital signs, the pulmonary and cardiac status, a complete neurological examination, documenting any preprocedural deficits, an evaluation of the patient’s pulses, including the bilateral dorsalis pedis and posterior tibial pulses for transfemoral approaches, and Allen’s test for patients and procedures requiring radial artery access. Crucial laboratory data include platelet count, prothrombin time, international normalized ratio, partial thromboplastin times to evaluate for a bleeding diathesis, and blood urea nitrogen and creatinine to look for renal insufficiency or failure. For patients with renal insufficiency, prehydration with sodium bicarbonate (130 mEq/L IV solution at 3.5 mL/kg bolus over 1 hour, then 1.2 mL/kg/hr during the procedure and for 6 hours after the procedure) and administration of N-acetylcysteine (600 mg orally at 24 and 12 hours before and after the procedure) can prevent contrastinduced nephropathy (2,3). Noninvasive imaging and previous angiography should be reviewed for preoperative clinical assessment and treatment planning. Depending on the patient’s age, clinical status, and anesthesia requirements, further investigation and testing may be required, including electro- and echocardiography, chest X-ray, pulmonary function tests, and referral to specialists in internal medicine, cardiology, and pulmonology may be required for presurgical clearance. The need for general endotracheal anesthesia versus monitored sedation during endovascular procedures depends on several factors related to the patient (e.g., age, anxiety level), the procedure, and operator preference. In general, embolization procedures, especially of aneurysms and arteriovenous malformations/fistulas (AVMs/AVFs), and intracranial angioplasty for vasospasm or atherosclerotic disease are performed under general endotracheal anesthesia at our institution, whereas we perform carotid artery stenting and tumor embolizations under monitored sedation. Many other procedures can be performed safely using either method, and the choice of anesthesia relies highly on operator preference.

162

Katz et al.

Of paramount importance in the preoperative patient visit is a succinct but comprehensive discussion of the intended procedure, as well as the risks, benefits, and alternatives of the suggested therapy. At this time, informed written consent should be obtained from the patient or from the patient’s representative (health care proxy or closest relative) if the patient is incapacitated, and the discussion documented in the medical record. In certain situations, emergency consent by two physicians is necessary to allow the rapid institution of lifesaving therapy, as is sometimes the case with several endovascular procedures (e.g., acute stroke therapy). Particular requirements for consent vary between regions and institutions and must be learned and strictly followed during the consenting process.

VASCULAR ACCESS Patient Monitoring Once the patient is brought to the angiography suite and placed in the supine position, an 18- or 20-gauge IV line is started. Continuous electrocardiography, pulse oximetry, and an intermittent automated blood pressure cuff are connected to the patient during all endovascular procedures. An arterial line is inserted for procedures requiring strict blood pressure monitoring and control, including post-subarachnoid hemorrhage (SAH) vasospasm therapy, ruptured aneurysm embolization, and revascularization of critically stenosed or occluded extra- and intracranial vessels. Intracranial pressure monitoring is typically utilized in patients with higher-grade SAH during aneurysm embolization or vasospasm therapy. Once the patient is on a monitor, either light sedation or general endotracheal anesthesia is administered.

Vascular Access Sites Common Femoral Artery

The choice of arterial access site depends mostly on operator preference, experience, and patient anatomy. For common femoral artery (CFA) access, the right CFA is the preferred access point because most interventionalists are right handed. Both groins are prepared and draped in case right CFA access is unsuccessful or bilateral arterial access is required. The latter is sometimes necessary when numerous devices that need to be inserted intravascularly cannot simultaneously fit within one introducer sheath. Complications of the transfemoral approach include retroperitoneal and groin hemorrhage, arteriovenous fistula and pseudoaneurysm formation, femoral nerve injury, dissection, and especially in children, arterial thrombosis or embolization to the distal lower extremity. Radial Artery

Radial artery access is becoming more popular for endovascular procedures because it offers several

potential advantages over CFA access. The transradial approach obviates the requirement for prolonged bed rest and eliminates the risk of occult hematoma formation, making it safer in anticoagulated patients (4), and it is sometimes necessary when transfemoral access of tortuous aortic arch vessels is technically impossible or in the 2% to 10% of patients who are incapable of undergoing transfemoral angiography because of peripheral vascular disease, aortic occlusion, or morbid obesity (5,6). Several series have reported successful and safe utilization of this approach for cerebral angiography (4–6). The transradial approach requires the performance of Allen’s test with pulse oximetry to demonstrate adequate collateral circulation to the hand from the ulnar artery and a cocktail of transintroducer sheath injections, including verapamil (2.5 mg), cardiac lidocaine (2%, 1 mL), and nitroglycerin (0.1 mg), to prevent vasospasm, as well as intravenous heparin (80 IU/kg) to prevent thrombosis. The transbrachial approach, offering the early ambulation and anatomical benefits of the transradial approach, has disadvantages that preclude its routine use, including potential median nerve injury and brachial artery thrombosis. Direct Puncture

In certain circumstances, alternative vascular access sites are required. During the treatment of vascular malformations of the head and neck, including lymphangiomas, capillary angiomas, venous angiomas, and AVMs, lesion access is by direct puncture. The latter method employs butterfly needles or long angiocaths inserted percutaneously under fluoroscopic or ultrasound guidance. Once intravascular position is confirmed by contrast injection under fluoroscopy, sclerotherapy is performed. Direct puncture of the external carotid artery (ECA) or its branches can be successfully performed in patients with facial AVMs previously treated with surgical ECA ligation (7). Arterial access is performed with a micro access set (AngioDynamics, Inc., Queensbury, New York, U.S.; see discussion in the next section) using fluoroscopic guidance or cervical landmarks.

Vascular Access Technique Standard Approach

In the transfemoral approach, the CFA is found at the medial aspect of the femoral head, a landmark that can be visualized under fluoroscopy and is useful in obese patients. To prevent retroperitoneal hematoma, the CFA should be accessed 2 to 3 cm below the inguinal ligament (1). Once the artery is palpated, the overlying skin is anesthetized by a 1-cm wheal of 1% lidocaine using a long 21-gauge needle at the skin surface. The needle is then advanced through the wheal, and 10 mL of lidocaine is carefully infused subcutaneously on either side of the artery and along a track to the artery. The syringe should be aspirated slightly before each injection to ensure extravascular

Chapter 8: Techniques and Devices in Interventional Neuroradiology

163

Figure 1 Photograph of standard arterial sheath introducer set. (A) 6-Fr introducer sheath. (B) 19-guage single-wall needle. (C) Dilator. (D) 0.035-inch J-wire with tip straightener.

infiltration. For single-wall punctures, a 19-gauge single-wall needle (Fig. 1) is used at a 458 angle to the skin as the artery is fixed in position by placing the left third, fourth, and fifth fingers above the puncture site and the index finger just below. A small skin incision can be made just prior to needle insertion, but we prefer to incise the skin only when sheath introduction necessitates it. With the bevel up, the needle is advanced toward the artery, and a transmitted pulse can commonly be felt against the right thumb positioned over the needle hub. A pulsatile efflux of bright red blood is an indication of successful needle insertion into the artery lumen. Only when this efflux is seen should the guidewire be inserted (commonly a 0.035-inch wire with a 3-mm J tip; Fig. 1), the needle removed with firm pressure held over the puncture site to prevent bleeding, and the introducer sheath with dilator advanced into the arterial lumen by the Seldinger technique. Alternatively, double-wall arterial puncture can be performed using a 19-gauge styleted needle (e.g., Pulse-Vu Needle, AngioDynamics, Inc.) over a straight 0.035-inch wire (e.g., Bentson wire, Merit Medical Systems, Inc., South Jordan, Utah, U.S.). If guidewire insertion meets resistance, fluoroscopically guided guidewire advancement should be performed to avoid subintimal insertion and vessel dissection. Contrast injection into the sheath can also ensure intra-arterial (IA) catheterization. In patients with multiple past transfemoral catheterizations, groin fibrosis may significantly increase resistance to sheath insertion. In these cases, progressive dilatation with dilators (5- to 8-Fr dilator, Cook, Inc., Bloomington, Indiana, U.S.) of increasing diameter over a stiffer introducer wire (i.e., 0.035-inch Terumo glidewire, Scimed/Boston Scientific, Fremont, California, U.S.) may enable successful sheath insertion. Once the sheath is in place, it should be secured with a Tegaderm (3M Health Care, St. Paul, Minnesota, U.S.) or suture and connected to continuous pressurized heparin saline flush.

Figure 2 Photograph of 5-Fr Micro Access Kit (Angiodynamics Inc.). (A) 21-guage echo tip needle. (B) 10-cm micro access dilator assembly. (C) 0.018-inch stainless steel guidewire with platinum tip (inset).

Micro Access Approach

For transradial and transbrachial approaches, patients on anticoagulation, and children and infants, a micro access set (AngioDynamics, Inc.; Fig. 2) is recommended. This system uses a 21-gauge needle and a 0.018-inch guidewire and comes with a 4- or 5-Fr sheath dilator. Once the micro puncture sheath is inserted, a standard 0.035-inch J tip guidewire is introduced through the sheath, which is then exchanged for a standard length introducer sheath.

Introducer Sheaths Standard introducer sheaths (Fig. 1) used in endovascular procedures are 11 cm long and range in size from 4 and 5 Fr, used commonly for diagnostic catheterization, to 9 Fr, required for mechanical embolectomy using the Merci Retriever (Concentric Medical, Inc., Mountain View, California, U.S.). Sheaths of 6 and 7 Fr are routinely used for most embolization and revascularization procedures. Larger sheath sizes (sheath size refers to the sheath inner diameter, whereas catheter size refers to the catheter outer diameter) are necessary when more devices are required for a particular procedure. For instance, balloon-assisted aneurysm coil embolization typically requires a 7-Fr sheath to simultaneously introduce the balloon catheter and the embolization microcatheter rather than using two 6-Fr guide catheters through bilateral CFA access. When more catheter support is needed, as in carotid artery angioplasty and stenting (CAS) procedures, a standard sheath can be exchanged over a

164

Katz et al.

guidewire for a long introducer sheath. The Vista Brite Tip (Cordis Corp., Miami Lakes, Florida, U.S.) introducer sheath comes in 35-cm (5 and 6 Fr) and 90-cm lengths (6-8 Fr) and provides excellent guide catheter support. The Shuttle Select Tuohy-Borst Introducer (Cook, Inc.) can also be used for the same purpose. It is available in 6- to 8-Fr sizes with 90-cm length and three different tip curves. Offers enhanced flexibility and trackability that can be useful when navigating tortuous vessel anatomy.

Anticoagulation Systemic heparinization during endovascular procedures should be initiated following sheath insertion and is administered in bolus doses. An 80-IU/kg loading dose is given to target an activated clotting time (ACT) of two to three times baseline. The ACT should be checked periodically during the procedure and additional heparin boluses given to ensure targeted anticoagulation to prevent embolic complications. At the end of the procedure, protamine sulfate can be administered, slowly to avoid systemic hypotension and anaphylaxis, at a dose of 0.5 to 1 mg/ 100 units of heparin remaining in the patient, with a maximum of 50 mg given over 10 minutes. Protamine administration, however, is not always necessary—for instance, following embolization procedures—and in many circumstances we allow the patient’s anticoagulation to correct spontaneously.

CATHETERS There are three main types of catheters: diagnostic catheters, guide catheters, and microcatheters (balloon catheters are discussed later in the section on balloons). Safe catheter use requires that (1) the catheter be advanced over a guidewire, especially in atherosclerotic patients; (2) the guidewire be withdrawn to just within the catheter tip when turning the catheter; (3) the catheter and guidewire tips never be out of fluoroscopic visualization during navigation; (4) the cause of resistance, if encountered, be investigated and the catheter not forced; and (5) diagnostic and guide catheters be continuously flushed with heparinized saline to avoid thrombosis at the catheter tip and inadvertent embolic complications.

Fremont, California, U.S.) catheters that are excellent for diagnostic angiography of uncomplicated aortic arches. Tortuous great vessels and bovine origin of the left common carotid artery can be challenging to access. When catheterizing with simple-curve catheters, maneuvers including having the patient cough while advancing the catheter, turning the patient’s head away from the targeted vessel, and breath holding in deep inspiration can unbend a tortuous vessel enabling catheter advancement (1). The use of digital roadmap technique and the transradial approach can also be helpful in this regard. The Simmons II catheter (Terumo Glidecath, Scimed/Boston Scientific, Fremont, California, U.S., and SIM 2, Merit Medical Systems, Inc.; Fig. 3) is a complex-curve device that is useful for accessing tortuous great vessels and bovine aortic arches. While complex-curve catheters facilitate entry into the proximal segment of a tortuous vessel, their secondary curve may inhibit distal catheterization. In this circumstance, the use of an exchange length guidewire (0.035 or 0.038 inch, 260 cm; e.g., Terumo Glidewire, Scimed/ Boston Scientific, Fremont, California, U.S.) may be used to exchange for a simple-curve catheter. For spinal angiography, we prefer the Cobra 1 and 2 (Terumo Glidecath, Scimed/Boston Scientific, Fremont, California, U.S.; Figs. 3 and 4) catheters because their shape enables access of posterior intercostal and lumbar arteries with relative ease and stability for spinal embolization procedures. For lower lumbar arteries, a Simmons 1 shape (Figs. 3 and 5) is preferable. Multiholed pigtail catheters (Merit Medical Systems, Inc.; Figure 6) are used for arch aortograms when the delineation of precise arch anatomy is required.

Guide Catheters Guide catheters are stiffer than diagnostic catheters, which improves support but makes navigation more

Diagnostic Catheters A wide variety of catheters are available from different endovascular companies, and in general, catheter selection is based on personal experience and preference. For diagnostic angiography, catheters are divided into simple-curve and complex-curve end-hole catheters, where simple-curve devices have only a primary (distal) curve and complex-curve devices have both primary and secondary curves that necessitate reforming of the distal curve once the catheter is positioned in an appropriate aortic arch vessel (typically, the left subclavian artery) (1). Examples of simple-curve catheters include the TempoVert (Cordis Corp.) and the Terumo Glidecath Angled Taper (Scimed/Boston Scientific,

Figure 3 Photograph of several diagnostic catheters (Terumo Glidecath, Boston Scientific, Fremont, California, U.S.). (A, B) Complex-curve Simmons 1 and 2 diagnostic catheters, respectively. (C) Simple-curve Cobra 2 diagnostic catheter.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

Figure 4 Spinal angiogram, AP view, demonstrating angiographic appearance of the Cobra 2 catheter. Note the significant tumor blush arising from branches of the right T10 posterior intercostal artery.

165

Figure 6 Arch aortogram, AP view, in a patient with Takayasu’s Arteritis demonstrating the angiographic appearance (inset) and use of the pigtail diagnostic catheter. Note the absence of the bilateral common carotid arteries and left vertebral artery and the compensatory hypertrophied ascending and deep cervical arteries.

distally in the targeted vessel. Delicate locations during catheter exchange include where the initial catheter is withdrawn below the origin of the tortuous vessel, where the catheter tip leaves the sheath, and where the guide catheter enters the tortuous vessel (1). A long (125 cm) 4- or 5-Fr diagnostic catheter can be placed within the larger guide catheter to facilitate catheter navigation or exchange. In addition, as previously mentioned, a long sheath can be inserted for added support to facilitate the insertion and stabilize the position of a guide catheter. Our guide catheter preference is the Envoy catheter (Cordis Corp.; Fig. 7) for its good support balanced with sufficient navigability and smooth device exchange.

Microcatheters Flow-Guided Microcatheters

Figure 5 Spinal angiogram, AP view, demonstrating the angiographic appearance of the Simmons 1 catheter. Note the artery of Adamkiewicz ( ) arising from the left T9 posterior intercostal artery.

complex. When advancing a guide catheter through a tortuous great vessel, certain tricks can be employed to ease navigation. The guide catheter can be advanced coaxially over an exchange length guidewire placed

Microcatheter selection is again based on experience and preference rather than science. Microcatheters can be divided into flow-guided and guidewire-directed devices. Flow-guided microcatheters, such as the Balt Magic (Balt/Boston Scientific, Fremont, California, U.S.) and the Spinnaker Elite (Target Therapeutics/ Boston Scientific, Fremont, California, U.S.) microcatheters are manufactured with a 100- to 120-cm, 3-Fr proximal shaft designed for support and pushability, a supple 25- to 30-cm, 2.5-Fr mid-shaft segment enabling navigation through tortuous vessels, and an extra supple 10- to 30-cm, 1.2-Fr, 1.5-Fr, or 1.8-Fr shapeable distal segment that allows for flow-guided vessel selection. Flow-guided microcatheters are designed for embolization of AVMs and AVFs with liquid acrylic adhesive material and are intended for

166

Katz et al.

Figure 8 Illustration of various microcatheter shapes. (A) Straight. (B) 458 curve. (C) 908 curve. (D) J-shaped curve. (E) S-shaped curve. (F). C-shaped curve. Source: Courtesy of Boston Scientific, Fremont, California, U.S.

WIRES Guidewires Figure 7 Photograph of the Envoy guide catheter (6 Fr, Cordis Corp.).

navigation through the circulation by contrast injection and flow, although microguidewire support and assistance may be needed. Guidewire-Directed Microcatheters

There are multiple guidewire-directed microcatheters. The choice depends on the main quality required, on distal trackability or support (e.g., for aneurysm embolization), and also on individual preferences. Selection of a particular shape, as with adding curves to microwires, largely depends on the vessel being selected and the curves being traversed during the catheterization process. Preshaped microcatheters (Fig. 8) come with 458 or 908 angled tips, or J-, S-, or C-shaped curves, range in distal diameter from 1.7 to 2.3 Fr, and generally have two radiopaque distal markers. The same microcatheters are available with straight tips, and all microcatheters can be shaped or reshaped with steam. The angle a targeted vessel (or aneurysm) takes from its parent artery will dictate which curve is selected. Examples of different microcatheters used in our practice include the Prowler 10, 14, and Plus (Cordis Corp.) microcatheters that are excellent for polyvinyl alcohol (PVA) particle embolization of tumors and ECA supply to dural AVF and the Echelon 10 and 14 (Micro Therapeutics, Inc., Irvine, California, U.S.) and the Excelsior SL-10 and 1018 (Target Therapeutics/Boston Scientific, Fremont, California, U.S.) microcatheters used for aneurysm coil embolization and IA infusions.

Most guidewires and catheters used in endovascular practice are hydrophilically coated to enhance coaxial navigation of the catheter over the wire. This coating must be prewetted with heparinized saline prior to introduction into the patient by flushing the device directly (catheter) or through an injection port on the plastic wire container (Fig. 9). Many microwires require prewetting in a bowl of heparinized saline for 15 to 30 seconds prior to shaping. Similar preparation is also required for coils and balloon catheters. Guidewire navigation and rotation is aided by the use of a torque device (Fig. 10) placed approximately 2 to 3 cm from the catheter hub. Guidewires used with diagnostic and

Figure 9 Photograph of 0.038-inch Terumo Glidewire (Boston Scientific, Fremont, California, U.S.) with distal curve inside circular plastic container with Luer lock hub for flushing with heparinized saline before use.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

Figure 10 Photograph of torquer device over microguidewire (Concentric Medical, Inc.).

guide catheters are, typically, 0.035- or 0.038-inchdiameter wires and are available in standard 150-cm or exchange 260-cm lengths with angled or straight/ shapeable tips (e.g., Terumo Glidewire and Glidewire LT, Scimed/Boston Scientific, Fremont, California, U.S.). In certain situations, a stiffer wire, such as the Ampaltz 0.038-inch 145-cm Super Stiff wire (Scimed/ Boston Scientific, Fremont, California, U.S.), may be needed to facilitate catheter advancement into the lumens of complex vessels or to increase guidecatheter/ sheath support in patients with severely tortuous great vessel anatomy (the so-called ‘‘buddy wire’’ method).

Microguidewires As with microcatheters, there is a panoply of microguidewires available, and subtle differences as well as operator preference dictate microguidewire selection for different purposes. Microguidewire properties, such as flexibility, coating, and torque ability, help define which microguidewire is best for which situation. In general, stiffer wires are more torqueable but are also more likely to perforate the vessel and may be difficult to navigate past the vessel origin. Common microguidewires used in our practice listed from stiffest to softest include the Transend-10, -14, and EX-14 (Target Therapeutics/Boston Scientific, Fremont, California, U.S.), the Silver- Speed-10 and -14 (Micro Therapeutics, Inc., the Agility-14 standard (Cordis Corp.), the Agility-10 and -14 soft (Cordis Corp.), and the Synchro-10 and -14 (Target Therapeutics/Boston Scientific, Fremont, California, U.S.) microguidewires. The Mirage wire (Micro Therapeutics, Inc.) has a 0.008inch diameter and can be used to push a flow-guided catheter around a tight curve and gain superselectivity during AVM embolization. In many instances, however, it is difficult to predict which microcatheter/ microguidewire combination will be successful in navigating difficult curves, and multiple attempts with various combinations may be necessary. As with all devices used in interventional neuroradiolgy, gaining experience with a selected handful of microguidewires

167

can be very rewarding when complicated anatomy is encountered. Giving the microguidewire tip a shape is indispensable for navigation, with shapes ranging from slight to 908 angles to J-, S-, and C-shaped curves. If the target is at a limited angle from the parent vessel, then a limited angle or slight J-shaped curve is needed. For acute angles, such as accessing the ophthalmic artery, an S-shaped curve may be useful. Even with experience, trial and error with different shapes, angles, and microcatheters may be needed to catheterize a challenging target. To maintain control when navigating a microcatheter over a microguidewire, it is important to look for slack and tension buildup in the guide catheter and microcatheter. These irregularities are anticipated when the operators’ movements are not being effectively translated into microcatheter/microguidewire movement. In addition, at the conclusion of a maneuver, it is important to slightly direct the device in the opposite direction to keep it in a steady position. For instance, when a microcatheter is advanced to the correct position within an aneurysm dome, it is necessary at the end of forward motion to withdraw slightly on the microcatheter to abolish the forward thrust that could inadvertently move the device too distally.

ENDOVASCULAR INFUSIONS Wada Testing Provocative testing of the central nervous system with IA amobarbital infusion followed by neurological and/ or neuropsychological assessment is called WADA testing. During preoperative evaluations for epilepsy surgery, the patient is connected to continuous electroencephalographic monitoring, and 80 to 140 mg of amobarbital diluted to 10 mL in saline is injected through the diagnostic catheter at 1 to 2 mL/sec into the internal carotid artery with the patient’s arms elevated to watch for intended contralateral hemiplegia. The patient typically recovers complete neurological function within 15 minutes, and then the contralateral hemisphere can be tested. During AVM embolization, 25 to 50 mg of amobarbital in a 1-mL aliquot can be injected through a microcatheter into a feeding artery and, following neurological assessment, predict dysfunction prior to permanent vessel occlusion (8). Similarly, 10 to 50 mg of 2% cardiac lidocaine can be injected selectively into extra-axial arteries to examine the peripheral nervous system prior to arterial embolization of extra-axial tumors or vascular malformations, and sequential testing with amobarbital followed by lidocaine may be useful for testing dangerous extracranial to intracranial anastamoses (8).

Vasodilator Infusions Both catheter-induced vasospasm and post-SAH vasospasm can be effectively managed with IA verapamil. On selective catheterization of the targeted vessel, 5 mg of verapamil diluted to 20 mL in saline is slowly

168

Katz et al.

infused at 1 to 2 mL/min through the microcatheter at a dose of 5 mg per injected vessel (though higher doses may be given if needed). Slow infusion limits cardiac side effects of hypotension and bradycardia. The safety and efficacy (though not durability) of chemical infusion for post-SAH vasospasm has been demonstrated in retrospective studies for both verapamil (9) and nicardipine (10). The dose used for nicardipine is also 5 mg per injected vessel. One to two milligrams of verapamil infused through a guiding catheter is usually sufficient to treat catheterinduced vasospasm, and pretreatment with verapamil can provide effective prophylaxis for vasospasm susceptible vessels (e.g., ECA). The IA infusion of papaverine will also dilate spastic vessels; however, the effect is short-lived and numerous side effects, including hypotension, intracranial hypertension (occurs with verapamil and nicardipine, but to a lesser extent), seizures, and even increased vasospasm, limit broad appeal (10). The advantages of chemical infusion over angioplasty for vasospasm therapy include decreased risk and the ability to treat distal and small artery vasospasm not reachable by balloon. However, balloon angioplasty is more durable.

Thrombolytic Infusions Intra-arterial Thrombolysis

IA thrombolysis by direct thrombolytic infusion following superselective catheterization offers treatment to acute ischemic stroke (AIS) patients who are ineligible for intravenous thrombolysis. Since the landmark PROACT II trial (11), both IA urokinase and recombinant tissue plasminogen activator (r-tPA) have been increasingly used for endovascular stroke therapy at many centers. On the basis of data showing a trend toward lower ICH complications with urokinase infusion compared with IA r-tPA (12), we preferentially infuse urokinase during IA stroke therapy at a dose of 250,000 to 750,000 IU (maximum of 1,000,000 IU) through a microcatheter placed within the clot over 30 to 120 minutes. Prior to infusion, the microguidewire is passed several times through the clot to increase the surface area for thrombolysis. The use of IA glycoprotein (GP) IIb/IIIa inhibitors to treat intracerebral artery occlusion refractory to IA thrombolysis also appears effective and safe (13). Our anecdotal experience with IA infusion of the GP IIb/IIIa inhibitor abciximab, including for the treatment of procedure-related thromboembolic complications, has been superb. Transvenous Thrombolysis

Targeted transvenous thombolytic therapy for medically refractory dural venous sinus thrombosis involves direct thrombolytic infusion through a microcatheter positioned in the thrombosed sinus. Sinus access is typically from the common femoral vein, though subclavian venous access may be required if the transfemoral route is obstructed (e.g., by femoral or inferior vena caval clot). Complete sinus recanalization customarily requires the microcatheter (e.g., Prowler Select 170-cm microcatheter is needed to navigate

from the common femoral vein to the superior sagittal sinus; Cordis Corp.) to be kept within the sinus for continuous infusion (e.g., intrasinus r-tPA at a rate of 1 to 2 mg/hr over 24 to 48 hours, with or without a bolus, although no standard protocol exists). Chemical thrombolysis may be done alone, but is usually performed in conjunction with mechanical thrombectomy using the microguidewire, balloon angioplasty, or rheolytic thrombectomy device (AngioJet, Possis Medical, Inc., Minneapolis, Minnesota, U.S.; requires an 8-Fr guide catheter).

BALLOONS Balloons used in endovascular surgery may be divided into high- or low-pressure devices and have four major uses, including angioplasty for extracranial or intracranial atherosclerotic stenosis (high-pressure balloons), angioplasty for vasospasm, balloon-assisted aneurysm remodeling, and balloon test occlusion (low-pressure balloons). Endovascular balloons are designed as balloon microcatheters to be advanced either coaxially or in a monorail fashion over a microguidewire into the desired location and are available in numerous sizes (balloon size is shorthanded as balloon diameter  balloon length). Careful sizing of the balloon’s diameter to less than the diameter of the target vessel is critical for minimizing the risk of artery dissection or rupture. All balloons must be prepared prior to patient introduction by vigilantly purging all air from the balloon by hand suction and then training the balloon by inflation with contrast solution. Balloon preparation is intended to minimize the risk of air embolism if the balloon ruptures.

High-Pressure Balloons The Aviator (Cordis Corp.) and Viatrac 14 Plus (Guidant Corp., Indianapolis, Indiana, U.S.) balloon catheters are examples of high-pressure balloons used for extracranial angioplasty. The Aviator balloon is a 3.3-Fr catheter designed for rapid monorail exchange, is available in 4- to 5.5-mm diameters, and has a nominal filling pressure of 10 atm with a rated burst pressure (RBP) of 14 atm. The Viatrac 14 Plus (Figs. 11 and 12) is also a 3.3-Fr catheter allowing rapid monorail exchange, is available in 4- to 7-mm diameters, and has a nominal filling pressure of 8 atm with an RBP of 14 atm. Both systems are efficiently used with the RX Acculink Carotid Stent System (Guidant Corp.). For intracranial atherosclerosic stenosis, the Maverick (Scimed/Boston Scientific, Fremont, California, U.S.; Fig. 13) balloon catheter (3.2 Fr) is a coaxially navigated dual lumen high-pressure coronary angioplasty device designed with enough flexibility and trackability to pass through the tortuous intracranial anatomy and is available in 1.5- to 4.0-mm balloon diameters. The inner catheter lumen allows for microguidewire employment (<0.014 inch), while the outer lumen is used for inflation (with a 50:50 contrast in normal saline solution) to nominal pressure of 6 atm and an RBP of 12 to 14 atm (depending on the balloon size).

Chapter 8: Techniques and Devices in Interventional Neuroradiology

Figure 11 Photograph of Viatrac 14 Plus high-pressure angioplasty balloons. (A) 4.5 mm  20 mm balloon and (B) 4.0 mm  30 mm balloon (Guidant Corp.).

169

Figure 13 Cerebral angiogram, unsubtracted AP view, demonstrating the angiographic appearance of the Maverick angioplasty balloon (Boston Scientific, Fremont, California, U.S.) inflated in the right middle cerebral artery.

During extracranial or intracranial angioplasty for atherosclerotic stenosis, balloon diameter is selected to approximate the vessel diameter just proximal and distal to the stenosis. The microguidewire is advanced past the stenosis, and the balloon is tracked over a microguidewire and positioned with the proximal and distal radiopaque markers straddling the stenosis. Careful inflation with an insufflator device (Fig. 14) is visualized under fluoroscopy (Fig. 15), and then the balloon is deflated. Several inflations may be necessary to achieve the desired lumen diameter, and the balloon position may be readjusted proximally or distally between inflations (once the balloon is fully deflated). Angioplasty of an atherosclerotic plaque creates microdissections in the nondiseased arterial wall adjacent to the plaque (14). For concentric extracranial plaques and in-stent restenosis, a cutting balloon (e.g., Ultra2 Monorail, Boston Scientific, Fremont, California, U.S./Scimed) may be more useful to create controlled dissections within the lesion.

Low-Pressure Balloons

Figure 12 Cervical angiogram, unsubtracted lateral view, demonstrating the angiographic appearance of the Viatrac 14 Plus angioplasty balloon (Guidant Corp.) inflated in the left internal carotid artery.

Hyperglide and Hyperform (Micro Therapeutics Inc.; Fig. 16) balloon catheters are examples of low-pressure balloons designed for intracranial use. These devices are single-lumen balloon catheters necessitating that a 0.010-inch microguidewire be positioned 10 mm beyond the balloon catheter tip to ensure central lumen occlusion and enable balloon inflation. The balloon is inflated by 1-cc syringe hand injection of a

170

Katz et al.

specific nominal and maximal inflation volumes). The balloon is deflated by careful retraction of the 1-cc syringe plunger and not by microguidewire withdrawal. The catheter can be flushed by withdrawing the microguidewire proximal to the proximal marker. Balloon-Assisted Aneurysm Remodeling

Balloon-assisted remodeling using these low-pressure devices is sometimes required during coil embolization of wide-neck aneurysms. The balloon is positioned over the neck of the aneurysm, while a second microcatheter is positioned into the dome of the aneurysm. The balloon is inflated during coil placement and then slowly deflated while monitoring coil position. If the coil mass is stable within the aneurysm dome, the coil is detached and the balloon reinflated for the insertion of the next coil. Balloon Angioplasty for Vasospasm

Figure 14 Photograph of an insufflator device used to inflate high-pressure angioplasty balloons. The device is filled with 10 to 20 cc of 50% contrast solution and purged of air bubbles prior to use.

Angioplasty for post-SAH vasospasm (Fig. 17) uses the same technique as for intracranial atherosclerosic angioplasty, except that a low-pressure balloon is employed to decrease the risks of vessel dissection and rupture. When using the smallest balloon catheter (Hyperform 4 mm  7 mm, Micro Therapeutics, Inc.) to perform distal angioplasty for vasospasm, we recommend using an 80% contrast solution to improve visibility, and to further enhance inflation safety, we use the Cadence Precision Injector (Micro Therapeutics Inc.), a screw syringe that enables precise control over balloon expansion. Balloon Test Occlusion

recommended 50:50 solution of 60% contrast in normal saline and requires nominal inflation volumes of, for example, 0.06 mL and 0.27 mL for the 4-mm and 7-mm Hyperform balloons, respectively, to achieve full balloon diameter (see individual device package insert for

Balloon test occlusion (BTO; Fig. 18) may be performed to evaluate a patient’s ability to tolerate permanent artery occlusion, for example, during the treatment of a giant fusiform aneurysm. The balloon is advanced to the target artery by standard technique, and the patient

Figure 15 Cerebral angiogram, AP view, left internal carotid artery injection. (A) Before angioplasty, a critical left middle cerebral artery stenosis is seen. (B) Unsubtracted AP view during Maverick balloon (Boston Scientific, Fremont, California, U.S.) inflation. Note the proximal and distal balloon markers that enable precise balloon positioning across the stenosis. (C) After angioplasty and stent placement, complete recanalization was obtained.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

171

Figure 16 Illustration of (A) Hyperglide and (B) Hyperform lowpressure balloons used during balloon aneurysm remodeling. Source: Courtesy of Micro Therapeutics Inc.

is examined preinflation, and then serially for 30 minutes during inflation. Any neurological deficit during balloon inflation is a failed test and a bypass graft procedure should be contemplated. Predictive value can be enhanced by qualitative cerebral blood flow (CBF) imaging during BTO by 99mTc SPECT scanning, where the radionucleotide is injected once the patient tolerates test occlusion for 15 minutes. Following a successful 30-minute BTO, the balloon is deflated and the procedure is concluded. The patient is then transferred for nuclear imaging. Quantitative CBF can be obtained using xenon CT or CT perfusion imaging, but is cumbersome to perform because of the requirement for injection during balloon inflation. Other Balloon Indications

Other balloon applications include permanent balloon artery occlusion (e.g., detachable silicone balloon, Target Therapeutics/Boston Scientific, Fremont, California, U.S.), infrequently used, for mechanical recanalization during acute stroke therapy, and microcatheter guidance through a sharp arterial curve. As an example of the latter, we recently employed a 7  7-mm Hyperform balloon in the supraophthalmic carotid artery to ease microcatheter/microguidewire navigation into an acutely angled ophthalmic artery. The same technique is helpful when advancing into acutely angled lenticulostriate arteries with the balloon positioned in the middle cerebral artery.

EMBOLIC PROTECTION DEVICES During CAS, the rate of embolic debris may be as high as 80% to 90% with a consequent 5% to 9% neurological event rate (15). The advent of improved pharmacological prophylaxis and embolic protection devices has

Figure 17 Cerebral angiogram, AP view. (A, B) Left internal carotid artery injection showing severe left A1 and A2 anterior cerebral artery vasospasm before (A) and after (B) balloon angioplasty, with significant luminal diameter improvement following serial balloon inflations. (C, D) Unsubtracted AP view of a 4  7-mm Hyperform balloon seen in the left A1 (C) and A2 (D) segments. Vasospasm of the left middle cerebral artery was treated in similar fashion.

lowered this neurological event rate to 0% to 2% (15). Various devices are under investigation, including (1) filter protection devices composed of a 0.014-inch microguidewire mounted with a polyurethane filter that is advanced beyond the stenosis, deployed in a straight vessel segment, and then retrieved by resheathing at the conclusion of CAS and (2) balloon occlusive devices, where a balloon is advanced past the stenosis and inflated to block flow during CAS or an aspiration balloon guide catheter is inflated proximal to the stenosis followed by initiation of flow reversal from the common carotid artery to the common femoral vein. In the latter method, a second balloon is inflated in the ECA to prevent reflux of debris from the ECA into the internal carotid artery circulation through collateral anastamoses. The advantage of filter devices is that CBF is not interrupted and also angiography can continue to be performed during the procedure. Several embolic protection devices being investigated

172

Katz et al.

Figure 18 Cerebral angiogram during balloon test occlusion. (A) Unsubtracted AP view during Hyperglide balloon (*, dilineates balloon margins) inflation across the vertebral basilar junction, performed to assess tolerance of bilateral vertebral artery occlusion as a treatment option for this (B) large complex basilar artery fenestration aneurysm (PA view, 3D rotational projection, postcomputerized reconstruction). (C) Right and (D) left vertebral artery injection demonstrates no anterograde flow into the basilar artery during balloon inflation. The patient developed vertigo at 25 minutes postinflation, thus failing the test.

include the ACCUNET (Guidant Corp., Menlo Park, California, U.S.; Fig. 19), ANGIOGUARD (Cordis Corp.), Emboshield (Abbott Laboratories, Abbott Park, Illinois, U.S.), and FilterWire EZ (Boston Scientific, Fremont, California, U.S.) filter systems, the PercuSurge balloon (Medtronic, Inc., Santa Rosa, California, U.S.) for distal occlusion, and the Parodi Anti-Embolism System (AnteriA Medical Science, San Francisco, California, U.S.) for proximal occlusion and flow reversal (15).

STENTS Stents are useful in varied circumstances, such as the treatment of atherosclerotic stenosis, arterial dissection, and wide-necked aneurysms. The utilization of all stents require pre- and postmedication with aspirin (81–325 mg/day) and clopidogrel [75 mg/day; at least 300 mg (up to 600 mg) given preoperatively, with therapy continued for 4 to 12 weeks]. Stents may be divided into balloon expandable (stainless steel) stents deployed on a balloon catheter (e.g., Palmaz Genesis Aviator-Medium Biliary Stent, Cordis Corp., as well as drug-eluting stents and stent grafts; see below) and over-the-wire self-expanding stents made of nitinol, a shape-memory alloy. Self-expanding stents widen to a

preset diameter (straight or tapering), which is limited by the vessel wall when delivered into a vessel of smaller diameter, and enables these stents to conform to vessels with tapering lumens typical of the neurovasculature. Stent radial force implants the struts into the vessel wall facilitating neo-endothelialization that regenerates a smooth vessel lumen. Oversizing the stent compared with the angiographic vessel diameter can enhance this process (16). Typically, stents are oversized approximately 0.5 to 1.0 mm more than the target vessel diameter, while balloons are undersized by approximately the same degree. Examples of nitinol stents used for CAS include the RX Acculink Carotid Stent System (Guidant Corp.; Fig. 20) available in tapered (7- to 10-mm and 6- to 8-mm taper diameters) and straight configurations that range in size from 5- to 10-mm diameter by 20- to 40-mm length, and the Precise stent (Cordis Corp.; designed as a biliary stent) ranging in size from 5- to 8-mm diameter by 20- to 40-mm length. Intracranial stent deployment for atherosclerotic or dissection-related stenosis requires the use of smaller stents, of which most were designed for the coronary arteries. Examples of these stents include the Driver, S7, and S660 balloon expandable stents (Medtronic, Inc.). The benefit of these stents is sufficient

Chapter 8: Techniques and Devices in Interventional Neuroradiology

Figure 19 (A) Photograph of Accunet distal embolic protection device. (B) Cervical angiogram, AP view, left common carotid artery injection demonstrating the angiographic appearance of a deployed Accunet device. Source: (A) Courtesy of Guidant Corp.

Figure 20 Photograph of Acculink 6  30-mm to 8  30-mm tapered stent (Guidant Corp.). Inset shows the Acculink delivery system.

173

radial force to maintain luminal patency following angioplasty. However, a major drawback is diminished flexibility and trackability to navigate these stents through the tortuous intracranial anatomy. Neointimal hyperplasia, the same problem encountered following coronary angioplasty and stenting, plagues intracranial (and to a lesser extent extracranial) stenting procedures. This process can result in restenosis as well as occlusion of small penetrating arteries. To counteract this process, drugeluting stents, including sirolimus (CYPHER, Cordis Corp.) and Paclitaxel (TAXUS, Boston Scientific, Fremont, California, U.S.) eluting stents, are approved for use in the coronary circulation, and coated stents, with heparin (Hepacoat, to prevent thrombosis; Cordis Corp.) or titanium-nitrous-oxide (TiNOX), preliminarily hold promise at lowering restenosis rates. For neuroendovascular procedures, drug-eluting stents are currently most useful in the treatment of vertebral artery stenosis, both because this artery has a particularly high-restenosis rate as well as because the vertebral arteries have a similar diameter to the coronary arteries (*2–3 mm) for which these stents have thus far been designed. The Neuroform stent (Boston Scientific, Fremont, California, U.S./Target Therapeutics, Inc., Natick, Massachusetts, U.S.; Fig. 21) is a self-expanding nitinol stent with open-cell design intended for intracranial use during endovascular coil embolization of wide-necked aneurysms. The Food and Drug Administration (FDA) has approved it as a humanitarian use device for this indication. The Neuroform Microdelivery System offers simplified stent deployment compared with a significantly more complex earlier iteration. During the

Figure 21 Photograph of a 3.5  10-mm Neuroform stent.

174

Katz et al.

Figure 22 Illustration of aneurysm coil embolization techniques. (A) Coils positioned through a microcatheter in a narrow-necked aneurysm. (B, C) For wide-necked aneurysms, coil placement with balloon remodeling followed by stent deployment is preferred. (D) Additional coils may be placed following stent deployment by positioning the microcatheter through the stent struts into the residual aneurysm neck or dome. Source: Courtesy of Boston Scientific, Fremont, California, U.S.

treatment of wide-necked aneurysms, the Neuroform stent may be deployed (1) prior to coil placement with the microcatheter delivered through the stent struts; (2) prior to coil placement, with the microcatheter positioned within the aneurysm dome and wedged between the stent and the vessel wall, the so-called ‘‘jail method’’; (3) alone without coil placement; or (4) following coil occlusion performed with or without balloon remodeling (Fig. 22). The last technique is our preferred approach. If possible, coil embolization with concurrent stent placement should be limited to the treatment of unruptured aneurysms because the required anticoagulation and antiplatelet therapy may carry an excessively high hemorrhagic complica-

tion risk for recently ruptured aneurysms (17). If complete occlusion of a ruptured aneurysm following coil embolization necessitates stent deployment, our practice is to place a stent during a second-stage procedure 4 to 12 weeks following the acute SAH. The Neuroform stent is not radiopaque and instead has four distal and four proximal radiopaque platinum markers (Fig. 23A). Under fluoroscopy, the delivery system has four indicators (Fig. 23B). Once the stent is in position, the stent stabilizer (the most proximal marker) is advanced to the proximal stent marker and the microcatheter is then gently retracted unsheathing the stent into position. The Neuroform stent has a relatively low radial force compared with

Figure 23 Cerebral angiogram, unsubtracted AP view. (A) The Neuroform stent has (1) four distal and (2) four proximal radiopaque markers. (B) The Neuroform3 Microdelivery System has four radiopaque indicators representing (1) the microdelivery catheter tip, (2) the compressed distal stent markers, (3) the compressed proximal stent markers, and (4) the stent stabilizer tip.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

Figure 24 (A, B) Cerebral angiogram, AP view, right vertebral artery injection, before (A) and after (B) stent graft occlusion of a large basilar artery fenestration aneurysm, showing complete aneurysm occlusion following the deployment of two stent grafts from the right vertebral artery into the left basilar fenestration arm. (Note: same patient as in Fig. 18). (C) Photograph of an unexpanded stent graft (JoStent, Abbott Laboratories).

coronary stents, making it impractical for the treatment of intracranial atherosclerotic or dissectionrelated stenosis. The new Wingspan stent (Boston Scientific, Fremont, California, U.S.) is intended to overcome this limitation. The FDA has recently approved it as a humanitarian use device for the treatment of symptomatic and greater than 50% intracranial stenosis resistant to medical therapy. Future directions for stent-assisted aneurysm embolization involves replacing standard open-cell stents with ‘‘covered’’ stents (also called stent grafts) that would obviate the need for coil embolization of some giant and wide-necked aneurysms and expand treatment options for large fusiform aneurysms and cavernous-carotid fistulas. Current stent grafts are balloon expandable stents composed of a polytetrafluoroethylene membrane sandwiched between two stainless steel stents (JoStent Graftmaster Coronary Stent, Abbott Laboratories; Fig. 24). Early experience treating intracranial vertebral and carotid artery aneurysms has been successful (18,19). The major limitations of covered stents are limited flexibility, which makes navigation through the tortuous intracranial vasculature difficult, and the potential for occlusion of small penetrating arteries.

EMBOLIC MATERIAL Microcoils Dr. Guglielmi first introduced electrolytically detachable platinum microcoils in 1990. Since then, advancements have led to the availability of a plethora of coils of

175

Figure 25 Illustration of coil shapes. (A) Standard helical coil. (B) 2D (2-diameter) coil. (C) 3D coil. Source: Courtesy of Boston Scientific, Fremont, California, U.S.

varying sizes and shapes, degrees of stiffness, and coatings from numerous manufacturers that have allowed the successful treatment of more varied and complex aneurysms. Coil size is shorthanded as coil diameter (approximate range, 2–24 mm)  coil length (approximate range, 1–60 cm). Coil shapes (names and styles vary by manufacturer) include standard helical, two-diameter, and three-dimensional (Fig. 25). Examples of some novel shapes include the Cyclone, Eight, and Pretzel configurations available for Sapphire, NXT, and Topaz coils (Micro Therapeutics, Inc.). Coil stiffness availability includes ultrasoft, soft, standard, firm, and extrafirm. Most current coils are manufactured with stretch-resistant technology. Coating options include bare platinum [e.g., Guglielmi detachable coils (GDC), Boston Scientific, Fremont, California, U.S.], bioactive platinum coils with polyglycolic-polylactic acid (PGLA) copolymer covering (Matrix and Matrix2, Boston Scientific, Fremont, California, U.S.; Fig. 26) or interwoven microfilaments (Nexus, Micro Therapeutics, Inc.), and hydrogel coils (HydroCoil, MicroVention, Inc., Aliso Viejo, California, U.S.). Tight packing of GDC coils within an aneurysm causes impedance and stagnation of intraaneurysm blood flow facilitating thrombosis. The electric current applied to detach the coil augments this prothrombotic effect. The PGLA component of Matrix and Nexus coils incites an inflammatory cascade that hastens the conversion of intra-aneurysmal thrombus to mature fibrocellular scar tissue and accelerates neoendothelialization of the aneurysm neck. Whether this modification actually enhances long-term aneurysm occlusion rates over bare platinum embolization is not known. The HydroCoil is a platinum coil coated with a hydrogel polymer called Intelligel, which is designed to swell over 20 minutes when the coil contacts blood (Fig. 27). The HydroCoil-10, -14, and -18 swell to volumes 5, 7, and 11 times the volume of a standard platinum coil,

176

Katz et al.

Figure 26 Illustration of coil coating. (A) Bare platinum coil (GDC) and (B) PGLA-coated coil (Matrix). (C) Magnified view showing the PGLA coating over the platinum base of the Matrix coil. Abbreviations: GDC, Guglielmi detachable coils; PGLA, polyglycolic-polylactic acid. Source: Courtesy of Boston Scientific, Fremont, California, U.S.

The technique of aneurysm coil embolization involves first the careful placement of a framing coil (typically a standard 2D or 3D coil is most suitable) into the aneurysm dome under fluoroscopic guidance using roadmap or subtraction fluoroscopy technique. Once coil position within the aneurysm, and without loop herniation into the parent vessel, is confirmed by fluoroscopy, the coil is detached. The coil pusher wire is then carefully withdrawn under fluoroscopic guidance while ensuring that the coil was fully detached. If residual aneurysm opacification is seen, then additional coils of various shapes and successively diminishing sizes are added to the coil mass until complete occlusion (if possible) is obtained. The first coil has to be as large as possible in relation to the size of the aneurysm, and the following coils are delivered in decreasing sizes, so that the aneurysm is filled with coils from the outside to the center. Good choice and placement of the first coil is fundamental and sometimes will alone determinate the success of the procedure. The most hazardous coils are the first (framing) coil, which has the greatest chance of causing aneurysm rupture, and the last coil at the neck, which has the highest likelihood of herniation into the parent vessel. Other types of coils include fibered coils intended for arterial and venous embolization in the treatment of large AVF and AVM (e.g., VortX, VortX Diamond, and fibered platinum coils, Boston Scientific, Fremont, California, U.S.; Fig. 28), including vein of Galen malformations. Fibered coils are constructed of a platinum base coil with attached dense polyester fibers intended to increase their thrombogenicity. Liquid coils (Berenstein Liquid Coil, Boston Scientific, Fremont, California, U.S.) are platinum microcoils that are injected with saline instead of those delivered with a pusher wire and are employed for artery or vein occlusion. The supplied plunger is attached to a 3-mL saline-filled syringe and then inserted into the hub of a microcatheter. Gentle saline infusion pushes

Figure 27 (A) Photograph of a hydrated Hydrocoil and (B) electronmicrograph of the dehydrated Intelligel coating of the Hydrocoil. Source: Courtesy of MicroVention Inc.

respectively, which is intended to increase coil-packing density. The coils are detached by hydrostatic pressure rather than electrolysis. The coils must be prehydrated with lactated Ringer’s solution rather than heparinized saline. We have found these coils to be most useful during the treatment of giant aneurysms.

Figure 28 Photograph of the fibered VortX coil. Source: Courtesy of Boston Scientific, Fremont, California, U.S.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

177

the coil to the tip of the microcatheter and then a small saline bolus injects the coil into the target vessel. We commonly use these coils during tumor embolization to block a proximal nonfeeding artery branch to avoid PVA particle embolization of that artery’s normal distal territory [e.g., occlusion of the anterior middle meningeal artery (MMA) branch during PVA embolization of a posterior MMA branch feeder of a tentorial meningioma].

Polyvinyl Alcohol Particles Mechanical vessel occlusion is efficiently obtained with PVA particles (Fig. 29) during the endovascular treatment of tumors, dural AVF, and epistaxis. PVA particles are mixed in contrast and injected selectively through a microcatheter under subtraction fluoroscopic guidance. PVA particles are available in 45- to 150-mm to 1000- to 1180-mm sizes (Contour PVA particles, Boston Scientific, Fremont, California, U.S.). We generally select 150- to 250-mm PVA particles for most tumor embolization procedures. During PVA embolization, the microcatheter is advanced into a feeding branch (avoiding wedge flow), and then PVA is infused in a pulsatile fashion until particles are seen accumulating at the microcatheter tip. Care must be taken to avoid particle accumulation in the microcatheter hub, and flushing the microcatheter with heparinized saline between successive infusions (away from other devices) is required. The particles are adherent to the endothelium, wedge in the vessels, and accumulate within the tumor bed inciting necrosis. Tumor necrosis may be accompanied by significantly increased surrounding edema; therefore, for large tumors with mass effect, we give a bolus dose of dexamethasone at the time of tumor embolization to prevent adverse sequelae. Arterial occlusion with PVA may be temporary with recanalization occurring over weeks to months as the periparticle thrombus resorbs (20). Thus, PVA embolization is most appropriate for preoperative tumor embolization, epistaxis (where the etiology of the hemorrhage is temporary), and occasionally is useful for embolization of small dural feeders of dural AVF.

Figure 30 Photograph of PVA particles in solution demonstrating the rough surface and aggregation tendency typical of PVA.

Embosphere Microspheres Embospheres (Biosphere Medical, Inc., Rockland, Massachusetts, U.S.) are an alternative particulate embolic agent composed of trisacryl gelatin microspheres. Unlike PVA particles, which have uneven surfaces and variable sizes (Fig. 30), microspheres are spherical particles of uniform size and shape. These characteristics are ideal for improving distal penetration of vascular lesions and tumor vascular beds. In a direct comparison with PVA particles, preoperative meningioma embolization with Embospheres resulted in significantly less procedural blood loss (21). Following tumor and vascular malformation embolization, the microspheres do not degrade and cause only a moderate surrounding inflammatory response (22). Embospheres, unlike PVA particles, do not aggregate; therefore, a 100-mm Embosphere bead will reliably occlude a 100-mm vessel, whereas a 100-mm PVA particle is less predictable and may bind to a 200-mm vessel wall causing obstruction following aggregation with other PVA particles. This absence of aggregation also enhances delivery because the microspheres do not clump within the microcatheter.

N-Butyl Cyanoacrylate

Figure 29 Photograph of 500- to 710-mm PVA particles.

The liquid acrylic adhesive N-butyl cyanoacrylic acid (NBCA) (Trufill NBCA Liquid Embolic, Cordis Corp.) is used in the embolization of AVM and pial AVF. NBCA preparation requires the use of clean gloves and should be performed in a location absent of all ionic solutions (e.g., blood, contrast, and saline). The

178

Katz et al.

adhesive is first diluted 25% to 33% with ethiodol (Cordis Corp.), an oily medium that increases the polymerization time of the NBCA monomers and enhances the character of the polymer mass as it contacts blood (23). The solution can be further mixed with tantalum powder (Cordis Corp.) to increase radiopacity. Typically, a flow-guided microcatheter is advanced into a feeding artery until wedge flow is obtained. The microcatheter is flushed with 3 mL of 5% dextrose in water, and then the NBCA solution is carefully injected under subtracted fluoroscopy into the AVM nidus or fistula, being cautious to avoid reflux at the microcatheter tip. Anions in blood, primarily hydroxyl anions, initiate the exothermic polymerization of the compound, which forms an occlusive cast in the feeding artery and nidus/fistula. Once the adhesive is sufficiently delivered, the microcatheter is briskly withdrawn and discarded. Further embolization can be performed in the same session or in staged procedures. In our experience, however, the majority of AVMs will require surgery or radiation therapy for definitive cure.

Onyx

MECHANICAL EMBOLECTOMY The Merci Retriever Mechanical embolectomy devices are used in the treatment of AIS patients who are either ineligible for or failed IV r-tPA, and may be used in conjunction with IA thrombolytic infusion. The Merci Retriever (X5 and X6 Retriever, Concentric Medical) is the only FDA-approved device available for mechanical clot extraction during AIS treatment. Merci device safety and efficacy in the treatment of AIS patients were demonstrated in the MERCI Trial (29,30). The device is a nitinol wire with five helical loops of diminishing diameter at the distal tip (Fig. 31). The newer L-series retrievers have cylindrical rather than tapered loops as well as bound suture material to enhance clot capture. The L-series retrievers are FDA approved for endovascular foreign body retrieval and are used off-label for mechanical embolectomy in AIS (Fig. 32). An 8- or 9-Fr balloon guide catheter (BGC, Merci BGC, Concentric Medical, Inc.; Fig. 31) is advanced into a proximal vessel (proximal internal carotid artery or subclavian artery just proximal to the vertebral artery take-off) and inflated to cease anterograde flow

Onyx (Micro Therapeutics, Inc.) is a liquid embolic agent composed of ethylene vinyl alcohol copolymer dissolved in dimethyl sulfoxide (DMSO), which has been available in Europe for the treatment of AVM and difficult aneurysms and was recently approved by the FDA for use in the United States for AVM embolization. The benefit of Onyx over NBCA for AVM embolization is that Onyx polymerizes by desiccation and is not adhesive, which allows for controlled and extensive filling of the nidus with less risk of premature polymerization, venous occlusion, and catheter gluing (24,25). For aneurysm treatment, a balloon is placed over the neck of the aneurysm, and Onyx is slowly injected at 0.1 mL/min (using the Cadence Precision Injector, Micro Therapeutics, Inc.) until the aneurysm is filled. With the balloon deflated, the DMSO diffuses out over approximately 10 minutes and the Onyx solidifies. The balloon is then re-inflated and the microcatheter is carefully removed (26).

Absolute Ethyl Alcohol Absolute ethyl alcohol is sometimes used in the treatment of vascular malformations of the head and the neck not amenable to surgical resection. Access is by selective IA or transvenous approach or by direct percutaneous puncture. Alcohol has low viscosity, facilitating deep penetration of the nidus, and is extremely cytotoxic, causing fibrinoid necrosis of the endothelium and vessel thrombosis (27,28). Ethanol is carefully infused into the lesion and requires 5 to 10 minutes for its thrombotic effect. Complete occlusion typically requires a staged embolization approach (27). General anesthesia is required because ethanol embolization is painful. The main risk of intravascular ethanol infusion is acute pulmonary hypertension.

Figure 31 Photograph of the Merci mechanical embolectomy system (Concentric Medical, Inc.). (A) Tapered loops of the X6 Retriever. (B) 18X microcatheter. (C) Balloon guide catheter with balloon inflated.

Chapter 8: Techniques and Devices in Interventional Neuroradiology

179

Figure 32 (A) Cerebral angiogram, unsubtracted AP view, demonstrating the angiographic appearance of the L-series Merci Retriever. (B) Cervical angiogram, unsubtracted lateral view following right common carotid artery injection with balloon inflated, demonstrates a thrombus halo surrounding the Merci Retriever.

during clot removal to prevent distal embolization of clot fragments. The microcatheter (18X and 18L Merci Microcatheter, Concentric Medical, Inc.) is advanced coaxially over a microguidewire (0.014 inch) through the clot. The microguidewire is carefully removed and replaced with the Merci Retriever. Two to four retriever loops are released distal to the clot, then the device and microcatheter are withdrawn into the clot and the device is torqued counterclockwise for two revolutions. The microcatheter is then retracted to unsheath the remaining loops within the clot, the BGC is inflated, and the device is torqued clockwise for up to five revolutions. The retriever/microcatheter unit is then withdrawn into the BGC lumen while aspirating, the balloon is deflated, and the entire system is removed from the patient. Alternatively, the BGC can be kept in place if more passes with the retriever are needed.

Other Embolectomy Devices Additional thrombus retrieval devices include the Neuronet Endovascular Snare (Guidant Corp.), composed of a nitinol basket attached eccentrically to a microguidewire, and the Amplatz Goose Neck Snare (Microvena Corp., White Bear Lake, Minnesota, U.S.), both designed for endovascular foreign body retrieval. These devices are under investigation for mechanical embolectomy in AIS and are used routinely at some centers for this indication. Other mechanical embolectomy technology currently in development for intracranial use or being employed investigationally for AIS therapy include laser thrombectomy with the Endovascular Photo Acoustic Recanalization (EPAR) Laser System (Endovasix, Inc., San Francisco, California, U.S.) and the LaTIS Neuro Laser Thrombolysis System (LaTIS, Inc., Coon Rapids, Minnesota, U.S.), and obliterative thrombectomy with the AngiJet Rheolytic Thrombectomy System

(Possis Medical, Inc.) (31). The latter device has been used successfully in the endovascular treatment of dural venous sinus thrombosis (32–34). The EKOS MicroLysUS infusion catheter (EKOS Corp., Bothel, Washington, U.S.) is a novel 2.5-Fr microcatheter with a distal 2-mm, 2.1-MHz sonographic ring transducer at the tip that uses ultrasonic pulse waves to create local cavitation at the thrombus surface to increase the surface area for IA thrombolysis (31).

MANAGEMENT OF COMPLICATIONS Neuroendovascular procedures carry significant risks, although our techniques have progressed so much in recent years that in experienced hands the vast majority of procedures are uncomplicated. Complications include thromboembolism, vessel perforation, dissection, vasospasm, and device fracture. However, if managed effectively, these occurrences can remain technical in nature, avoiding patient morbidity and mortality. Postintervention angiography is intended to evaluate technical success and to assess for any of the above complications.

Ischemic Complications Intracranial thromboembolic complications, identified by the lack of distal branch filling or the new onset of focal neurological deficits in the absence of hemorrhage, may be immediate or delayed (e.g., embolic stroke following aneurysm coil embolization may occur days postoperatively), and if not recognized acutely will result in cerebral infarction. Cerebral hypoperfusion, which may also result in new focal neurological deficits, is sometimes unavoidable, for example, following parent artery occlusion for large fusiform aneurysms; however, certain measures can minimize injury and morbidity. Postoperative hypertensive and

180

Katz et al.

hypervolemic therapy to increase collateral flow to an ischemic area may decrease infarction volume (useful for the last example), and judicious use of aspirin therapy may prevent delayed embolic stroke following aneurysm coil embolization. Aspirin and clopidogrel treatment, as already mentioned, is required following stent placement to prevent in-stent thrombosis. Intravenous GP IIb/IIIa antagonists (35), IA GP IIb/IIIa antagonists, and IA tPA are each useful alone or in combination for treating thromboembolic complications that are recognized during angiography, and mechanical embolectomy may be needed for larger vessel occlusions. Extracranial embolic complications are avoided by continuous pressurized heparin-saline flush and by the use of proper angiographic technique.

Hemorrhagic Complications Intracranial hemorrhage may evolve secondary to vessel perforation, angiographically noticed as contrast extravasation into the subarachnoid space and clinically by acute blood pressure elevation (Cushing response), or may be a delayed effect of hyperperfusion or reperfusion injury following revascularization procedures (angioplasty/stenting and AIS therapy; Fig. 33). If contrast extravasation is seen, it is important

Figure 33 Axial, noncontrast head CT scan showing a large hyperdensity in the right basal ganglia representing massive reperfusion intracerebral hemorrhage with midline shift following mechanical embolectomy for an acute right middle cerebral artery occlusion. The patient had MRI diffusion abnormality in the right striatum prior to embolectomy. However, a large perfusiondiffusion mismatch prompted endovascular treatment. Mechanical recanalization was complete without the use of thrombolytic drugs.

not to withdraw the microcatheter/wire, because the device likely plugs the perforation hole and prevents massive hemorrhage. If a balloon is in place, as in some aneurysm treatments, it should be inflated to tamponade the rupture. Damage may then be minimized by either NBCA infusion into the puncture hole (infuse through the microcatheter gently into the subarachnoid space), or coil embolization of the aneurysm, or parent vessel. Post-revascularization injury is best prevented by proper patient selection (especially of AIS patients) and aggressive postoperative blood pressure control. If a patient is at high risk for hyperperfusion injury, we monitor the patient in the intensive care unit and keep the systolic blood pressure between 100 and 110 mmHg. For lower-risk patients, systolic blood pressure below 140 mmHg is usually adequate.

Other Complications Vessel dissections may be managed by stent placement during the procedure, or more conservatively, as most dissections heal on their own, with aspirin or anticoagulation therapy to prevent embolization from the intimal flap. As already discussed, verapamil infusion is very effective for the treatment of catheterinduced vasospasm. Device fractures (e.g., coils; Fig. 34) may be efficiently managed by retrieval using the Merci Retriever or Amplatz Goose Neck Snare.

Figure 34 Cerebral angiogram, unsubtracted AP view, showing a coil fracture during the embolization of this large left middle cerebral artery aneurysm. The coil was successfully retrieved with the Amplatz Goose Neck Snare (Microvena Corp.).

Chapter 8: Techniques and Devices in Interventional Neuroradiology

GROIN CLOSURE When the endovascular procedure is concluded, the catheters and sheaths are removed, and hemostasis is achieved by either manual compression or the use of a closure device. Manual compression requires pressure with the two fingers placed over the arterial pulse and a third finger positioned over the puncture hole. Pressure is gradually alleviated over 10 to 15 minutes and the site inspected for further bleeding. Complete hemostasis may necessitate up to 30 to 60 minutes of compression, especially for patients on systemic anticoagulation and antiplatelet therapy. In children and neonates, care must be taken not to stop arterial flow and cause thrombosis when applying manual pressure, and compression may be performed with a pulse oximeter connected to the ipsilateral foot to monitor extremity perfusion. Following manual compression, the patient is maintained on strict bed rest, with the accessed leg immobilized for the number of hours equal to the size of the introducer sheath in French (i.e, 6 hours for a 6-Fr sheath). Hemostasis following radial artery puncture can be aided by the use of a compression clamp. After one hour of compression, the clamp should be slowly released while watching for continued bleeding. As already mentioned, one of the benefits of the transradial approach is the lack of postprocedure ambulation restriction. Several closure devices are commercially available, which all carry the advantage of early ambulation, enabling the patient to walk one hour after groin closure. Examples include the Perclose Closure System (Abbott Laboratories), which employs suture closure, and the Angio-Seal vascular closure device (St. Jude Medical, Inc., St. Paul, Minnesota, U.S.), which uses a bioabsorbable anchor and collagen sponge to sandwich and seal the arteriotomy site. Angio-Seal is available in a 6-Fr device for sheaths less than or equal to 6 Fr and an 8-Fr device for sheaths less than or equal to 8 Fr. In our opinion, the Angio-Seal is simpler to use and has fewer device failures compared with Perclose. The use of any closure device requires placement under sterile conditions (we typically change gloves and reprep the groin area with rubbing alcohol). Although not recommended by the manufacturer, we also administer one dose of antibiotic (e.g., a cephalosporin for skin flora coverage) because the device employs foreign body implantation. Closure devices are too large and not indicated in pediatric patients and should be cautiously deployed in atherosclerotic and diabetic patients. In addition, iliac artery angiography before device groin closure may minimize postclosure complications. Finally, in our experience, thin adult patients often complain of persistent groin pain following Angio-Seal closure, and we recommend that the device be avoided in slender patients.

CONCLUSION In recent years, the explosion of novel and enhanced neurointerventional devices has resulted in tremendous progress in endovascular techniques, making procedures simpler, safer, and with broader indications

181

and rendering interventions that were once unimaginable not only possible but commonplace. One caveat is that during this rapid evolution, insufficient outcome analyses have been performed and many new devices and techniques remain unproven. Our challenge in the years to come is to demonstrate both the technical and clinical efficacy of current and new endovascular devices and procedure through further prospective studies and randomized clinical trials.

REFERENCES 1. Osborn AG. Diagnostic neuroangiography: basic technique. In: Osborn AG, ed. Diagnostic Cerebral Angiography. 2nd ed. Philadelphia: Lippincott Williams and Wilkins, 1999:421–431. 2. Merten GJ, Burgess WP, Gray LV, et al. Prevention of contrast-induced nephropathy with sodium bicarbonate: a randomized controlled trial. JAMA 2004; 291(19):2328–2334. 3. Tepel M, Van Der Giet M, Schwarzfeld C, et al. Prevention of radiographic-contrast-agent-induced reductions in renal function by acetylcysteine. N Engl J Med 2000; 343:180–184. 4. Iwasaki S, Yokoyama K, Takayama K, et al. The transradial approach for selective carotid and vertebral angiograpghy. Acta Radiologica 2002; 43:549–555. 5. Matsumoto Y, Hongo K, Toriyama T, et al. Transradial approach for diagnostic selective cerebral angiography: results of a consecutive series of 166 cases. AJNR Am J Neuroradiol 2001; 22:704–708. 6. Levy EI, Boulos AS, Fessler RD, et al. Transradial cerebral angiograpghy: an alternative route. Neurosurgery 2002; 51:335–342. 7. Gobin YP, Pasco A, Merland JJ, et al. Percutaneous puncture of the external carotid artery or its branches after surgical ligation. AJNR Am J Neuroradiol 1994; 15:79–82. 8. Mathis JM, Barr JD. Pharmacologic testing as an adjunct to neuroendovascular procedures. Neurosurg Clin N Am 2000; 11(1):21–26. 9. Feng L, Fitzsimmons BF, Young WL, et al. Intraarterially administered verapamil as adjunct therapy for cerebral vasospasm: safety and 2-year experience. AJNR Am J Neuroradiol 2002; 23:1284–1290. 10. Badjatia N, Topcuoglu MA, Pryor JC, et al. Preliminary experience with intra-arterial nicardipine as a treatment for cerebral vasospasm. AJNR Am J Neuroradiol 2004; 25:819–826. 11. Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in acute cerebral thromboembolism. JAMA 1999; 282:2003–2011. 12. Kidwell CS, Saver JL, Carneado J, et al. Predictors of hemorrhagic transformation in patients receiving intraarterial thrombolysis. Stroke 2002; 33:717–724. 13. Deshmukh VR, Fiorella DJ, Albuquerque FC, et al. Intraarterial thrombolysis for acute ischemic stroke: preliminary experience with platelet glycoprotein IIb/IIIa inhibitors as adjunctive therapy. Neurosurgery 2005; 56(1):46–55. 14. Kaneko E, Skinner MP, Raines EW, et al. Detection of dissection and remodeling of atherosclerotic lesions in rabbits after balloon angioplasty by magnetic-resonance imaging. Coron Artery Dis 2000; 11(8):599–606. 15. Boulos AS, Levy EI, Bendok BR, et al. Evolution of neuroendovascular intervention: a review of advancement in device technology. Neurosurgery 2004; 54:438–453. 16. Guterman LR, Fessler RD, Hopkins LN. Cervical carotid revascularization. Neurosurg Clin N Am 2000; 11(1):39–48.

182

Katz et al.

17. Katz JM, Tsiouris AJ, Biondi A, et al. Advances in endovascular aneurysm treatment: are we making a difference?. Neuroradiology 2005; 47(9):695–701. 18. Felber S, Henkes H, Weber W, et al. Treatment of extracranial and intracranial aneurysms and arteriovenous fistulae using stent grafts. Neurosurgery 2004; 55:631–638. 19. Burbelko MA, Dzyak LA, Zorin NA, et al. Stent-graft placement for wide-neck aneurysm of the vertebrobasilar junction. AJNR Am J Neuroradiol 2004; 25:608–610. 20. Choi IS, Tantivatana J. Neuroendovascular management of intracranial and spinal tumors. Neurosurg Clin N Am 2000; 11(1):167–185. 21. Bendszus M, Klein R, Burger R, et al. Efficacy of trisacryl gelatin microspheres versus polyvinyl alcohol particles in the preoperative embolization of meningiomas. AJNR Am J Neuroradiol 2000; 21:255–261. 22. Beaujeux R, Laurent A, Wassef M, et al. Trisacryl gelatin microspheres for therapeutic embolizations II: preliminary clinical evaluation in tumors and arteriovenous malformations. AJNR Am J Neuroradiol 1996; 17:541–549. 23. Kerber CW, Wong W. Liquid acrylic adhesive agents in interventional neuroradiology. Neurosurg Clin N Am 2000; 11(1):85–99. 24. Gobin YP, Murayama Y, Milanese K, et al. Head and neck hypervascular lesions: embolization with ethylene vinyl alcohol copolymer: laboratory evaluation in swine and clinical evaluation in humans. Radiology 2001; 221:309–317. 25. Jahan R, Murayama Y, Gobin YP, et al. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients. Neurosurgery 2001; 48:984–995. 26. Molyneux AJ, Cekirge S, Saatci I, et al. Cerebral Aneurysm Multicenter European Onyx (CAMEO) trial: results

27.

28.

29.

30.

31. 32.

33.

34.

35.

of a prospective observational study in 20 European centers. AJNR Am J Neuroradiol 2004; 25(1):39–51. Do YS, Yakes WF, Shin SW, et al. Ethanol embolization of arteriovenous malformations: interim results. Radiology 2005; 235(2):674–682. Mourao G, Hodes JE, Gobin YP, et al. Curative treatment of scalp arteriovenous fistulas by direct puncture and embolization with absolute alcohol. J Neurosurgery 1991; 75(4):634–637. Gobin YP, Starkman S, Duckwiler GR, et al. MERCI 1: a phase 1 study of mechanical embolus removal in cerebral ischemia. Stroke 2004; 35:2848–2854. Smith WS, Sung G, Starkman S, et al. Safety and efficacy of mechanical embolectomy in acute ischemic stroke: results of the MERCI trial. Stroke 2005; 36:1432–1440. Katz JM, Gobin YP, Segal AZ, et al. Mechanical embolectomy. Neurosurg Clin N Am 2005; 16(3):463–474. Chow K, Gobin YP, Saver J, et al. Endovascular treatment of dural sinus thrombosis with rheolytic thrombectomy and intra-arterial thrombolysis. Stroke 2000; 31:1420–1425. Opatowsky MJ, Morris PP, Regan JD, et al. Rapid thrombectomy of superior sagittal sinus and transverse sinus thrombosis with a rheolytic catheter device. AJNR Am J Neuroradiol 1999; 20:414–417. Dowd CF, Malek AM, Phatouros CC, et al. Application of a rheolytic thrombectomy device in the treatment of dural sinus thrombosis: a new technique. AJNR Am J Neuroradiol 1999; 20:568–570. Alexander MJ, Duckwiler GR, Gobin YP, et al. Management of intraprocedural arterial thrombus in cerebral aneurysm embolization with abciximab: technical case report. Neurosurgery 2002; 50:899–901.

9 Balloon Occlusion, Wada, and Pharmacological Testing Linda J. Bagley Departments of Radiology and Neurosurgery, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania, U.S.A.

INTRODUCTION Despite vast recent technical advances in interventional neuroradiology, parent vessel sacrifice remains the only viable therapeutic option for many complex intracranial lesions. As such, there continues to be a role for occlusion testing in the preprocedural management of certain aneurysms and neoplasms with associated vascular encasement or potential vascular compromise. Functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG), and diffusiontensor imaging-based MR tractography in certain cases provide noninvasive means of function and tract localization; however, there continue to be significant limitations of these modalities for memory localization and when performed in the presence of tumors or vascular malformations, which may be associated with distorted brain architecture, altered venous drainage patterns, and/or reorganization of function. Selective and superselective pharmacological tests are thus employed in the preprocedural assessment of certain patients prior to arteriovenous malformation (AVM) embolization, epilepsy surgery, and select tumor resections. This chapter discusses the indications, techniques, and potential complications and limitations of arterial and venous occlusion, Wada, and pharmacological testing. Illustrative case examples are also provided.

ARTERIAL AND VENOUS OCCLUSION TESTING Indications Arterial and/or venous occlusion tests are appropriate for patients scheduled for endovascular, neurosurgical, and/or otorhinolaryngological procedures in which vascular occlusion is indicated as therapy or in which there is a significant risk of intraprocedural vascular occlusion (1–7). Such patients may include those with aneurysms not amenable to coil embolization, stent placement, or microsurgical repair (Fig. 1) (2,3), those with tumors encasing or in intimate contact with the carotid or vertebral arteries (Fig. 2) (4), and those with tumors involving a dural venous sinus

or necessitating a surgical approach that could result in sacrifice of a dural venous sinus.

Technique---Venous Occlusion Testing A guiding catheter is typically placed within the internal jugular vein, and a nondetachable balloon microcatheter is advanced into the transverse or sigmoid sinus. Arterial catheterization as well as angiography is performed during balloon occlusion of the transverse or sigmoid sinus that is at risk. The venous phase of the angiogram is then assessed for drainage of the affected hemisphere through collateral pathways. Larger balloon sizes (often up to 1.5 cm) are required than those employed in arterial testing. Anatomic data are obtained, but functional data are not, as symptoms of venous ischemia are more likely to be insidious at the onset.

Techniques---Arterial Occlusion Testing A number of techniques have been described for temporary arterial occlusion testing. In virtually all cases, diagnostic angiograms are initially performed (1, 5). This technique allows assessment of the cervical vasculature for atherosclerotic disease and of the intracranial circulation for integrity of the circle of Willis and potential collateral flow. Following diagnostic angiography and prior to nondetachable balloon inflation, patients are routinely anticoagulated, typically with 5000 to 10,000 units of heparin with elevation of the activated coagulation time (ACT) to two to three times the baseline value. Heparinized saline flush is also employed (1,5,8). Most cases are performed with conscious sedation. Both double lumen balloon catheters and nondetachable balloon microcatheters (with maximal balloon diameters of 5–8 mm) have been used in these procedures. In some institutions, a second catheter is utilized to perform diagnostic angiography during the occlusion (confirming occlusion and assessing adequacy of collateral circulation in the arterial, parenchymal, and venous phases). During carotid occlusion testing, larger

184

Bagley

Figure 1 (A, B) Lateral views of digital subtraction angiogram obtained during right common carotid artery injection demonstrate a lobulated, wide-necked approximately 2.5  2.0-cm right internal carotid artery aneurysm, located predominately within the cavernous segment with extension into the supraclinoid segment. There are multiple surgical aneurysm clips about the supraclinoid internal carotid artery.

Figure 2 (A, B) Axial gadolinium-enhanced T1-weighted images demonstrate an enhancing extra-axial posterior fossa mass (a meningioma), encasing the right vertebral artery and compressing the brain stem. (C) Lateral projection of digital subtraction angiogram obtained during common carotid artery injection reveals hypervascular tumor blush in the posterior fossa. (D) AP projection of left vertebral artery angiogram demonstrates no reflux of contrast into the distal right vertebral artery. The left anterior inferior cerebellar artery is large and supplies the posterior inferior cerebellar artery territory. (E) AP projection of right vertebral artery angiogram demonstrates mild displacement and narrowing of the distal right vertebral artery secondary to the meningioma.

Chapter 9: Balloon Occlusion, Wada, and Pharmacological Testing

balloon catheters are typically inflated in the proximal internal carotid artery, while microcatheters may be advanced to the petrous segment. Inflation times of 15 to 30 minutes have typically been reported (1,5,8). Continuous clinical monitoring of the neurological exam is performed, with any changes in the exam prompting immediate deflation of the balloon and termination of the procedure. Following deflation of the balloon, angiography is repeated to evaluate the internal carotid artery for injury/dissection and to exclude emboli in the intracranial circulation. Almost all patients failing the balloon occlusion test by clinical criteria will develop a permanent neurological deficit if a revascularization procedure is not performed prior to the intended vascular occlusion (9). Of those who tolerate the test by clinical criteria, between 5% and 20% have been reported to incur a permanent neurological deficit following permanent vascular occlusion (7,10), generally within hours to days of the permanent occlusion. Deficits may result from perfusional ischemia and/or stump emboli. Numerous adjuvant techniques, including measurement of stump pressures (11,12), induced hypotension (10,13), single-photon emission tomography (SPECT) (8,14–17), transcranial Doppler examinations (18,19), xenon CT (1,12,18,20–22), cerebral blood flow measurements (22), perfusion imaging (7,23,24), electroencephalography (EEG) (25–31), and monitoring of somatosensory-evoked potentials (SSEP) (32,33) have been employed to improve the predictive value of occlusion testing. Unfortunately, for many of these techniques, conflicting results have been reported in the literature. Measurement of Stump Pressures

Arterial pressure can be monitored distal to the occlusion by the use of a double lumen catheter with connection of the second lumen to a pressure transducer. The utility of this technique is controversial. Some authors have reported that maintenance of a stump pressure ratio (initial mean stump pressure/ preocclusion mean arterial pressure) of 60% or more during test occlusion is a useful marker of adequate collateral circulation (11). A number of studies have found a significant correlation between stump pressures and measures of cerebral perfusion, such as SPECT (8), while others have failed to demonstrate such a correlation (12). Induced Hypotension

This technique is applied in patients who have tolerated 15 to 20 minutes of arterial occlusion while normotensive (10,13). Following initial testing, systolic blood pressure is pharmacologically lowered by approximately 30% (typically with nitroprusside infusion or intravenous labetalol), and neurological testing is continued for an additional 15 to 20 minutes. Standard et al. (10) reported the identification of an additional 19% of patients in his series with limited cerebrovascular reserve and a 5% false-negative rate for patients tolerating balloon occlusion testing with hypotensive challenge (reduced compared with

185

reported false-negative rates of conventional occlusion testing of up to 20%). In contrast, Dare et al. (13) reported a 15% false-negative rate for such patients. False negatives may in part be secondary to vasodilatory effects of nitroprusside. Concerns have also been raised that induced hypotension may falsely elevate the sensitivity of this test and thus inappropriately subject these patients to revascularization procedures. SPECT

With this technique, Tc 99m hexamethylpropyleneamine oxime (HMPAO) is injected immediately following balloon inflation and vascular occlusion. SPECT imaging is performed following balloon deflation and removal of the catheter with measurement of activity ipsilateral and contralateral to the occlusion. Activity reflects cerebral blood flow within a few minutes of the injection (8,15). Hemispheric asymmetries and reductions in tracer uptake (in comparison to preocclusion testing when available) have been shown to correlate with the development of clinical deficits. MR Perfusion

MR perfusion imaging (7) may be performed with dynamic gadolinium-enhanced imaging during balloon occlusion when MR is available in the interventional suite. Gadolinium is administered by bolus injection at a dose of 0.1 mmol/kg. Cerebral blood volume (CBV) may be calculated on the basis of signal intensity changes in brain with and without contrast agent. Mean transit time (MTT) of contrast through the arterial system may also be calculated, allowing determination of regional cerebral blood volume (rCBV) (24). Authors have demonstrated greater perfusion delays in patients who have clinically failed the test. Authors have also described alterations in contrast enhancement and in brain parenchymal signal intensity in areas of hypoperfusion (7). CT Perfusion

Perfusion CT and acetezolamide challenge have also been employed in patients who have clinically tolerated test occlusion again in attempts to improve sensitivity. The technique described by Jain et al. (23) requires transport of the patient with a catheter in the internal carotid artery and reinflation of the balloon without fluoroscopic guidance, but on the basis of the volume used to inflate the balloon initially. Multiple perfusion CTs are performed with axial images through the basal ganglia obtained during rapid administration of intravenous contrast (4 cc/sec) initially during balloon reinflation and then following balloon deflation. Additionally, perfusion CT is repeated 20 minutes following administration of 1 g of acetezolamide (Bedford Kaboratories, Bedford, Ohio, U.S.) with balloon reinflation. Xenon CT

This technique also attempts to measure cerebral blood flow in patients who have clinically tolerated test occlusion (1,12,18,20–22). Again, it requires transport

186

Bagley

of the patient to the CT scanner with a catheter in the internal carotid artery and reinflation of the balloon. The patient inhales a gas mixture of 33% xenon and 67% oxygen. Baseline scans are obtained prior to balloon inflation and are repeated during balloon inflation. Xenon uptake in the middle cerebral artery distribution is used to estimate regional cerebral blood flow (rCBF). Regions with cerebral blood flow of less than 30 mL/100 g/min are judged to be at risk. Transcranial Doppler Ultrasound

Transcranial Doppler interrogation of the middle cerebral artery may be performed during test occlusion. There is an imperfect correlation between cerebral blood flow and mean velocity in the middle cerebral artery, as velocity is also affected by vascular caliber, hematocrit, viscosity, and insonation depth and angle (18). However, reductions of mean blood flow velocity and pulsatility index of less than 30% have been shown to be predictive of clinical tolerance, whereas reductions of more than 50% have been shown to correlate with clinical symptoms (19). Neurophysiological Monitoring

Many studies have documented the utility of neurophysiological monitoring (NPM) (EEG, SSEP, and brain stem–evoked potentials) in patients undergoing cerebrovascular surgery, most notably carotid endarterectomy, performed under general anesthesia (25–28). EEG changes have been reported when rCBF is less than 10 mL/100 g/min. An rCBF of 15 mL/100 g/min appears to be a critical value below which cortical SSEP amplitude is reduced, central conduction time is prolonged, and cerebral infarction is likely to occur. Liu et al. also described the use of NPM in patients undergoing endovascular procedures and demonstrated NPM changes in 26% of the patients in his series, with resultant alterations in management in 14% of the patients. Monitoring proved most beneficial in patients who were unable to cooperate with neurological testing. However, NPM changes were also observed in cooperative patients without corresponding abnormalities noted on physical neurological exam (33).

COMPLICATIONS Reported complications of balloon occlusion testing include arterial dissections, embolic infarcts, and perfusional ischemia. Reported rates of complications range from less than 1% to 15% (the majority between 1% and 7%), with higher rates being reported in earlier studies. Complication rates have decreased with improvements in catheter and balloon technology. Adequate anticoagulation during balloon inflation has reduced the number of embolic complications (1,5,6). However, adjuvant techniques intended to increase the sensitivity of the test (e.g., those requiring transport of patients with indwelling catheters, blind reinflation of the balloon, and/or induction of hypotension) have sometimes been associated with higher complication rates (due to vascular injury and creation of perfusional deficits) (1,10,20–23).

ILLUSTRATIVE CASE The patient is a 52-year-old woman with history of hypertension who presented with a three-week history of headache and right eye pain. Physical examination was notable for a right sixth nerve palsy and mild ptosis. Imaging studies were notable for an approximately 2-cm aneurysm of the cavernous right internal carotid artery (Fig. 3A–C). Prior to definitive treatment and following systemic anticoagulation, a 30-minute test occlusion of the distal cervical right internal carotid artery was performed. No changes in the patient’s neurological exam were noted during the test occlusion. Subsequently, treatment of the aneurysm with Neuroform stent placement and coil embolization was attempted, but resulted in compromise of the parent vessel (Fig. 3D, E). Ultimately, permanent occlusion of the right internal carotid artery was performed without neurological complication (Fig. 3F, G).

WADA TEST Epilepsy Epilepsy, the condition of spontaneously recurring seizures, is quite common, affecting approximately 0.5% to 1% of the population (34). It is a potentially psychosocially devastating, life-altering, and even lifethreatening disorder (due to associated increased incidences of sudden death, traumatic injuries, and suicide). While many advances have been made in the medical therapy of epilepsy, many cases, between 5% and 20%, remain medically intractable (35,36). Surgical therapies are appropriate for certain patients and include lesional resections, temporal lobectomies, callosotomies, hemispherectomies, and subpial transactions (35). The most common cause of intractable epilepsy is mesial temporal sclerosis. As such, temporal lobectomy is the most commonly performed surgical procedure for the treatment of epilepsy. Patients undergoing temporal lobectomy are at risk of developing speech/language and/or memory deficits, and therefore commonly undergo functional preoperative testing to minimize these risks.

Historical Background Amobarbital is a lipid soluble substance that can cross the blood-brain barrier and temporarily block neuronal function (37). Currently, the intracarotid amobarbital test is performed in conjunction with neuropsychological testing to determine lateralization of speech/language and memory functions, most often in patients with medically intractable epilepsy scheduled for surgical resection of epileptogenic tissue (38–40). This test is also sometimes employed in nonepileptic patients scheduled for resection of unilateral temporal or frontotemporal lesions. Juhn A.Wada first performed the intracarotid amobarbital procedure in the 1940s. Wada initially injected sodium amobarbital into the left common

Chapter 9: Balloon Occlusion, Wada, and Pharmacological Testing

187

Figure 3 (A) Axial CT scan demonstrates enlargement of the right cavernous sinus with a mildly hyperdense extra-axial mass. (B, C) AP oblique and lateral views of digital subtraction angiogram obtained during right common carotid artery injection reveal an approximately 2 cm irregular aneurysm, without discernable neck and with possible intraluminal thrombus, of the cavernous right internal carotid artery. (D, E) AP and lateral digital subtraction angiograms obtained following stent placement and partial coil embolization of the aneurysm are notable for markedly decreased flow within the internal carotid artery. (F) Lateral digital subtraction angiogram demonstrates multiple embolization coils and complete occlusion of the right internal carotid artery. (G) AP angiogram obtained during left common carotid artery injection demonstrates extensive flow across the anterior communicating artery with contrast opacification of the right anterior and middle cerebral arteries, without opacification of the previously demonstrated aneurysm.

carotid artery of a patient with frequent status epilepticus in an attempt to arrest the convulsions. In 1949, he first described the use of this procedure to determine language lateralization (38) in an attempt to improve the safety and efficacy of electroconvulsive therapy in psychiatric patients (by placement of the electrodes over the nonspeech dominant hemisphere). In 1960, Wada collaborated with Rasmussen and reported the use of the intracarotid amobarbital procedure for determination of hemispheric language dominance in epilepsy surgical patients. In 1962, Milner, Branch, and Rasmussen reported the use of this test for assessment of memory function in the isolated hemisphere. Subsequent studies have examined the correlation of memory lateralization with location of the seizure focus and prediction of surgical outcomes (41–49).

Technique The intracarotid amobarbital procedure is performed without sedation, as sedation may confound neuropsychological test results. Subsequently, diagnostic angiography is initially performed with catheter placement in the cervical internal carotid artery to assess for anatomic variation in the anterior circulation, in

particular to exclude a carotid-basilar anastomosis, but also to assess the likely distribution of amobarbital. The internal carotid artery supplying the presumed abnormal hemisphere is generally the one initially selected. Following diagnostic angiography, approximately 80 to 125 mg of sodium amobarbital are injected into the internal carotid artery over approximately three to five seconds, while the patient counts backward and attempts to maintain his or her arms in an elevated position. Efficacy of amobarbital administration may be confirmed in several ways: concurrent unilateral EEG slowing and/or development of a neurological deficit (hemiparesis and/or aphasia). Testing of language skills and visuospatial and verbal memory is then performed. Following a 30- to 45-minute delay (to allow the effects of amobarbital to diminish), the contralateral internal carotid artery is selected and the procedure is repeated. When a carotid-basilar anastomosis is present, a microcatheter can be advanced beyond it and the procedure continued. Similarly, when there is significant cross-flow through the anterior communicating artery, bifrontal impairment may result, and the neuropsychological testing may be rendered invalid secondary to impaired patient consciousness and inability to cooperate. In such

188

Bagley

cases, selective middle (MCA) and posterior cerebral artery (PCA) amobarbital administration has often been described with somewhat reduced dosages of the agent (75–80 mg). Selective MCA amobarbital administration has been used when language lateralization was of primary concern, and selective PCA administration was performed for memory assessment (50).

Predictive Value The intracarotid amobarbital test has also long been used to determine language lateralization. Multiple studies have also investigated its validity in the prediction of postoperative memory deficits (41,42,49, 51–54). Cohen-Gadol et al. (52) demonstrated statistically significant correlations between Wada test scores and hippocampal volumetry, as well as a significant inverse correlation between the disparity of the scores and changes in verbal (though not visuospatial) memory following temporal lobectomy. Andelman et al. (49) demonstrated a significant correlation between memory scores in the ipsilateral hemisphere and postsurgical memory changes. Multiple additional studies have also demonstrated a correlation between Wada memory test results and seizure-free outcome, most notably in patients with temporal lobe epilepsy. Correct lateralization of the seizure focus (and hence, prediction of outcome) has been reported in 75% to 85% of patients using various criteria (45–48).

Alternative Agents At times, there has been a global shortage of amobarbital. As such, alternative agents have been investigated. Methohexital (Brevital) has been employed, but it is quite short acting, and reinjection may be required. Jones-Gotman et al. (55) reported successful experience with etomidate, given by bolus followed by infusion, which was continued during sampling of speech measures and presentation of objects for memory testing. In the 30 hemispheres tested, contralateral hemiplegia developed in all patients, and slowing of EEG was observed in all injected hemispheres. Aphasia followed dominant hemisphere injections. All affects reversed over approximately four minutes, following termination of infusion. Propofol (56), administered in doses of 10 to 20 mg (typically less than 15 mg), has also been employed, with results comparable to those obtained with amobarbital. However, in Mikuni’s study (57), adverse reactions (including eye pain, shivering, facial contortion, confusion, involuntary movements, increased muscle tone, rhythmic movements, and tonic posturing) were noted in one-third of the patients, and in some cases precluded completion of the testing. Adverse reactions were observed more frequently in patients over 55 years and when larger doses of propofol were administered. The most severe reactions likely result from hyperexcitic phenomena, which have been shown to occur with sudden increases in cerebral propofol concentrations.

Noninvasive Testing A number of noninvasive tests have been proposed to replace the Wada test in the presurgical evaluation of patients with intractable epilepsy. Most notably, fMRI has been proposed as a replacement for the Wada test (58–74); fMRI has the advantages of being less timeconsuming (typically requiring *30–60 minutes), of posing minimal risk to the patient, and of being significantly less expensive (58). Numerous studies have examined the validity of fMRI in language lateralization. Binder et al. (59) found 96% concordance between fMRI and Wada test results for language dominance. Yetkin et al. and Lehericy et al. have reported similar results (67,68). However, the most promising results have been achieved in patients with left hemispheric language dominance (the majority of patients studied) and temporal lobe epilepsy. False categorization of language dominance by fMRI was reported to be approximately 9% in Woermann’s series, ranging from 3% in left-sided temporal lobe epilepsy to 25% in left-sided extratemporal epilepsy (60). Sabbah et al. (64) reported 95% concordance of the Wada test and fMRI in a group of epileptic patients with suspected atypical language lateralization. Additional studies have examined the ability of fMRI to predict memory following temporal lobectomy (69–73) and to lateralize temporal lobe epilepsy [and hence predict surgical outcome (70)]. Rabin et al. (69) examined fMRI activation during complex visual scene encoding (which is believed to engage both visuospatial and verbal memory systems) and reported a correlation between activation asymmetries in the mesial temporal regions on fMRI with hemispheric memory dominance. Additionally, Rabin et al. reported correlation of these activation asymmetries with postsurgical memory as well as an inverse correlation between absolute activation in the hemisphere to be resected and postsurgical memory. Jokeit et al. (70) reported 90% accuracy of lateralization of seizure focus in patients with unilateral temporal lobe epilepsy on the basis of memory-induced mesial temporal lobe activation asymmetries on fMRI.

MEG MEG has also been proposed as an alternative to the Wada test (75–80). MEG is based on the principle that all electrical currents are associated with magnetic fields. Cohen (80) developed a superconducting quantum interference device to measure magnetic fields generated by intracranial currents. Approximately 10,000 to 100,000 neurons must be simultaneously generating current to produce a magnetic field strong enough to be detected with present technology. Magnetic fields are minimally distorted by intervening tissue, and hence MEG may provide precise localization of the source of electrical current (e.g., seizure foci or functional cortex). Data can then be coregistered with conventional MRI. MEG has been used to localize visual, auditory, and somatosensory cortex, and to lateralize language (79). Papanicolaou et al. (75) demonstrated a high degree of (though not absolute)

Chapter 9: Balloon Occlusion, Wada, and Pharmacological Testing

concordance (87%) between MEG and Wada data for determination of hemispheric language dominance. MEG tended to detect more activity in the nondominant hemisphere (similar to fMRI) than predicted by the Wada test.

ILLUSTRATIVE CASE The patient is a 41-year-old right-handed man with history of fall at the age of 2 associated with brief coma with subsequent development of medically intractable partial complex and generalized tonic-clonic epilepsy in adulthood. EEG was suggestive of a left temporal lobe seizure focus. MRI was notable for left mesial temporal sclerosis (Fig. 4A) and positron emission tomography (PET) was notable for mild left temporal hypometabolism. A Wada test was performed as part of the presurgical evaluation for scheduled left temporal lobectomy. Digital subtraction angiography performed following selection of the left internal carotid artery revealed a persistent left trigeminal artery

189

(Fig. 4B, C). A microcatheter was subsequently advanced beyond the origin of the trigeminal artery into the supraclinoid left internal carotid (Fig. 4D, E), and 125 mg of sodium amobarbital was instilled. The patient developed transient right hemiparesis and aphasia. Neuropsychological testing revealed left hemispheric language dominance and right hemispheric memory dominance. The patient subsequently underwent right temporal lobectomy with histological confirmation of mesial temporal sclerosis.

PHARMACOLOGICAL TESTING Indications The development of permanent neurological deficits following embolizations of cerebral and spinal AVMs has been reported in 5% to 10% of the cases, largely due to ischemic sequelae in eloquent tissue. Ischemic injury to the cranial nerves, to the eye, or to cerebral cortex (via external to internal carotid artery or external

Figure 4 (A) Coronal T2-weighted image demonstrates atrophy and hyperintensity of the left hippocampus, indicative of mesial temporal sclerosis. (B, C) AP and lateral angiograms obtained during left internal carotid artery injection reveal a persistent trigeminal artery. (D, E) AP and lateral angiograms obtained with a microcatheter placed in the supraclinoid left internal carotid artery beyond the origin of the trigeminal artery.

190

Bagley

to vertebral artery anastamoses) may complicate external carotid artery embolizations. As such, superselective angiography and provocative pharmacological tests are often employed to predict the safety of arterial embolization of cerebral and spinal AVMs, dural AVMs, facial and oral vascular malformations, and tumors (81–84). Similar to Wada testing, superselective provocative pharmacological testing may also be employed prior to planned surgical intervention.

by placement of a coil or similar agent and repeating the testing, utilizing a different embolic agent (such as larger particles) or abandonment of embolization of that particular pedicle. Complications of provocative pharmacological testing are uncommon. Injection of lidocaine into the external circulation often produces some degree of discomfort in the patient . While rare, injection of lidocaine into the cerebral circulation has been reported to be complicated by seizures and cardiorespiratory depression (86).

Technique Typically, provocative testing for arteries with extraaxial destinations is performed with lidocaine and provocative testing for arteries with intra-axial destinations is performed with barbiturates, such as amobarbital, though both agents have also been utilized in both situations (81–86). Patients undergoing such procedures usually receive a small amount of intravenous sedation. Neurological assessments are made immediately before and after anesthetic injection and completed within a few minutes. In some cases NPM supplements the physical exam (particularly when the territory involved is the brain stem or thalamus), whereas in patients under general anesthesia it may completely replace it. Amobarbital is generally administered in doses of 30 to 75 mg, and lidocaine is typically administered in doses ranging from 10 to 40 mg. Some authors have reported the administration of both agents (lidocaine and amobarbital) prior to embolization, as these agents produce anesthetic effects via different mechanisms, and concomitant administration may thus improve the sensitivity of provocative testing. Amobarbital acts at the gamma aminobutyric acid A (GABA-A) receptor through inhibition of postsynaptic neurons in cortical and deep gray matter and in the hippocampus. As cerebral white matter contains a paucity of GABAergic synapses, it is largely unaffected by the drug. Lidocaine blocks voltagegated sodium channels, which are present on all nerve cell membranes and thus inactivates neurons in gray and white matter (85). Fitzsimmons et al. reported detection of clinically significant deficits with superselective lidocaine administration in four patients that did not develop a deficit with superselective amobarbital testing (85). With high-flow lesions, such as AVMs, hemodynamic factors may produce false-negative and false-positive tests with amobarbital or lidocaine. The anesthetic agent may bypass adjacent normal tissue, but liquid embolic agents will polymerize and may initially occlude high-flow channels, leading to redistribution of the embolic agent (82). Deveikis et al. (86) advocated the use of amobarbital injections in the external carotid circulation (to supplement lidocaine test injections) to improve the sensitivity for detection of external carotid to internal carotid or vertebral artery anastomoses. When a deficit is produced with pharmacological testing, therapeutic options include advancing the microcatheter closer to the nidus and repeating the testing, protecting the normal territory

Noninvasive Alternatives While pharmacological testing provides specific information about the function of tissue within a vascular territory, fMRI (87) and tractography (88–91) may noninvasively localize eloquent tissue prior to planned surgical intervention. As above, fMRI has been shown to localize (though imperfectly) language and memory. It has also been utilized to localize motor, sensory, and visual cortex. Similarly, thinsection diffusion tensor imaging has been employed to localize sensory, motor and visual cortex, and pathways (88–91). As diffusion of water perpendicular to the neural axis is limited by cell membranes and myelin sheaths, and diffusion along the nerve axis is less limited, the primary direction of the principal eigenvector of the diffusion tensor at each location can be determined and corresponds to the trajectory of the corresponding fiber. Thus, sensory and motor tracts may be mapped in three dimensions and in relation to AVMs and their draining veins (88). Additionally, tractography may be fused with conventional MRI images and utilized with neuronavigational systems (90). This technique offers advantages over pharmacological testing as drug distribution may be altered by arteriovenous shunting and as such testing yields no information about displacement of cortex and tracts or relationship of draining veins to eloquent cortex and tracts. This technique has also been attempted with tumors, but its utility is decreased by the presence of vasogenic edema. The presence of hemorrhage may also degrade images due to susceptibility artifacts.

SUMMARY Arterial and venous occlusion, Wada, and pharmacological tests provide vital information in the management of complex aneurysms, neoplasms, AVMs, and epilepsy. A variety of techniques have been employed to improve the accuracy of these tests. Complications are rare, but their incidence may be increased with increasing complexity of the tests. A number of noninvasive alternatives (predominantly employing MRI) to these tests have been developed and have demonstrated promising though imperfect results. As such, the neurointerventionalist must be knowledgeable of the indications, limitations, and potential complications of provocative testing and maintain expertise in its performance.

Chapter 9: Balloon Occlusion, Wada, and Pharmacological Testing

REFERENCES 1. Mathis JM, Barr JD, Jungreis CA, et al. Temporary balloon test occlusion of the internal carotid artery: experience in 500 cases. AJNR Am J Neuroradiol 1995; 16:749–754. 2. Van der Schaaf IC, Brilstra EH, Buskens E, et al. Endovascular treatment of aneurysms in the cavernous sinus: a systematic review of balloon occlusion of the parent vessel and embolization with coils. Stroke 2002; 33:313–326. 3. Vazquez Anon V, Aymard A, Gobin YP, et al. Balloon occlusion of the internal carotid artery in 40 cases of giant intracavernous aneurysm: technical aspects, cerebral monitoring, and results. Neuroradiology 1992; 34:245–251. 4. Sanna M, Piazza P, DiTrapani G, et al. Management of the internal carotid artery in tumors of the lateral skull base: preoperative permanent balloon occlusion without reconstruction. Otol Neurotol 2004; 25:998–1005. 5. Meyers PM, Thakur GA, Tomsick TA. Temporary endovascular balloon occlusion of the internal carotid artery with a nondetachable silicone balloon catheter: analysis of technique and cost. AJNR Am J Neuroradiol 1999; 20:559–564. 6. Allen JW, Alastra AJ, Nelson PP. Proximal intracranial internal carotid artery branches: prevalence and importance for balloon occlusion test. J Neurosurg 2005; 102:45–52. 7. Michel E, Liu H, Remley K, et al. Perfusion MR neuroimaging in patients undergoing balloon test occlusion of the internal carotid artery. AJNR Am J Neuroradiol 2001; 22:1590–1596. 8. Tomura N, Omachi K, Takahashi S, et al. Comparison of Technetium Tc 99m hexamethylpropyleneamine oxime single-photon emission tomograph with stump pressure during the balloon occlusion test of the internal carotid artery. AJNR Am J Neuroradiol 2005; 26:1937–1942. 9. de Vries EJ, Sekhar LN, Eibling DE, et al. A new method to predict safe resection of the internal carotid artery. Laryngoscope 1990; 100:85–88. 10. Standard SC, Ahuja A, Guterman LR, et al. Balloon test occlusion of the internal carotid artery with hypotensive challenge. AJNR Am J Neuroradiol 1995; 16:1453–1458. 11. Morishima H, Kurata A, Miyasaka Y, et al. Efficacy of the stump pressure ratio as a guide to the safety of permanent occlusion of the internal carotid artery. Neurol Res 1998; 20:732–736. 12. Barker DW, Jungreis CA, Horton JA, et al. Balloon test occlusion of the internal carotid artery: change in stump pressure over 15 minutes and its correlation with xenon CT cerebral blood flow. AJNR Am J Neuroradiol 1993; 14:587–590. 13. Dare AO, Chaloupka JC, Putman CM, et al. Failure of the hypotensive provocative test during temporary balloon test occlusion of the internal carotid artery to predict delayed hemodynamic ischemia after therapeutic carotid occlusion. Surg Neurol 1998; 50:147–156. 14. Peterman SB, Taylor A, Hoffman JC. Improved detection of cerebral hypoperfusion with internal carotid balloon test occlusion and 99mTc HMPAO cerebral perfusion SPECT imaging. AJNR Am J Neuroradiol 1991; 12:1035–1041. 15. Palestro CJ, Sen C, Muzinic M, et al. Assessing collateral cerebral perfusion with technetium-99m-HMPAO SPECT during temporary internal carotid artery occlusion. J Nucl Med 1993; 34:1556–1557. 16. Ryu YH, Chung TS, Lee JD, et al. HMPAO SPECT to assess neurologic deficits during balloon test occlusion. J Nucl Med 1996; 37:551–554. 17. Suguwara Y, Kikuchi T, Ueda T, et al. Usefulness of brain SPECT to evaluate brain tolerance and hemodynamic changes during temporary balloon occlusion test and after permanent carotid occlusion. J Nucl Med 2002; 43:1616–1623.

191

18. Kofke WA, Brauer P, Policare R, et al. Middle cerebral artery blood flow velocity and stable xenon-enhanced computed tomographic blood flow during balloon test occlusion of the internal carotid artery. Stroke 1995; 26:1603–1606. 19. Eckert B, Thie A, Carvajal M, et al. Predicting hemodynamic ischemia by transcranial Doppler monitoring during therapeutic balloon occlusion of the internal carotid artery. AJNR Am J Neuroradiol 1998; 19:577–582. 20. Johnson DW, Stringer WA, Marks MP, et al. Stable xenon CT cerebral blood flow imaging: rationale for and role in clinical decision making. AJNR Am J Neuroradiol 1991; 12:201–213. 21. Linskey ME, Jungreis CA, Yonas H, et al. Stroke risk after abrupt internal carotid artery sacrifice: accuracy of preoperative assessment with balloon test occlusion and stable xenon-enhanced CT. AJNR Am J Neuroradiol 1994; 15:829–843. 22. Marshall RS, Lazar RM, Young WL, et al. Clinical utility of quantitative cerebral blood flow measurements during internal carotid artery occlusion tests. Neurosurgery 2002; 50:996–1005. 23. Jain R, Hoeffner EG, Deveikis JP, et al. Carotid perfusion CT with balloon occlusion and acetazolamide challenge test: feasibility. Radiology 2004; 231:906–913. 24. Rosen BR, Belliveau JW, Vevea JM, et al. Perfusion imaging with NMR contrast agents. Magn Reson Med 1990; 14:249–265. 25. Whittemore AD, Kauffman JL, Kohler TR, et al. Routine electroencephalographic (EEG) monitoring during carotid endarterectomy. Ann Surg 1983; 197:707–713. 26. McFarland HR, Pinkerton JA, Frye D. Continuous electroencephalographic monitoring during carotid endarterectomy. J Cardiovasc Surg 1988; 29:12–18. 27. Baker JD, Gluecklich B, Watson CS, et al. An evaluation of electroencephalographic monitoring for carotid study. Surgery 1975; 78:787–794. 28. Harada RN, Comerota AJ, Good GM, et al. Stump pressure, electroencephalographic changes and the contralateral carotid artery: another look at selective shunting. Am J Surg 1995; 170:148–153. 29. Morioka T, Matsushima T, Fujii K, et al. Balloon test occlusion of the internal carotid artery with monitoring of compressed spectral arrays (CSAs) of electroencephalogram. Acta Neurochir 1989; 101:29–34. 30. Herkes GK, Morgan M, Grinnell V, et al. EEG monitoring during angiographic balloon test carotid occlusion: experience in sixteen cases. Clin Exp Neurol 1993; 30:98–103. 31. Cloughesy TF, Nuwer MR, Hoch D, et al. Monitoring carotid test occlusions with continuous EEG and clinical examination. J Clin Neurophysiol 1993; 10:363–369. 32. Schellhammer F, Heindel W, Haupt WF, et al. Somatosensory evoked potentials: a simple neurophysiological monitoring technique in supra-aortal balloon test occlusions. Eur Radiol 1998; 8:1586–1589. 33. Liu AY, Lopez JR, Do HM, et al. Neurophysiological monitoring in the endovascular therapy of aneurysms. AJNR Am J Neuroradiol 2003; 24:1520–1527. 34. Fisher RS, Stein A, Karis J. Epilepsy for the neuroradiologist. AJNR Am J Neuroradiol 1997; 18:851–863. 35. Jack CR. Epilepsy: surgery and imaging. Radiology 1993; 189:635–646. 36. Andermann F. Brain structure and epilepsy: the impact of modern imaging. AJNR Am J Neuroradiol 1997; 18:302–306. 37. Silfevenius H, Fagerlund M, Saisa J, et al. Carotid angiography in conjunction with amytal testing of epilepsy patients. Brain Cogn 1997; 33:33–49. 38. Snyder PJ, Harris LJ. The intracarotid amobarbital procedure: an historical perspective. Brain Cogn 1997; 33:18–32.

192

Bagley

39. Trenerry MR, Loring DW. Intracarotid amobarbital procedure. The Wada test. Neuroimaging Clin N Am 1995; 5:721–728. 40. Spencer DC, Morrell MJ, Risinger MW. The role of the intracarotid amobarbital procedure in evaluation of patients for epilepsy surgery. Epilepsia 2000; 41:320–325. 41. Acharya JN, Dinner DS. Use of the intracarotid amobarbital procedure in the evaluation of memory. J Clin Neurophysiol 1997; 14:311–325. 42. Bengner T, Haettig MA, Merschhemke M, et al. Memory assessment during the intracarotid amobarbital procedure. Neurology 2003; 61:1582–1587. 43. Baxendale SA, Van Paesschen W, Thompson PJ, et al. The relation between quantitative MRI measures of hippocampal structure and the intracarotid amobarbital test. Epilepsia 1997; 38:998–1007. 44. Bramham J, Morris RG. Pre- and postoperative intracarotid amytal procedure: an assessment of validity. Epilepsy Behav 2003; 4:556–563. 45. Alpherts WCJ, Vermeulen J, van Veelen CWM. The Wada test: prediction of focus lateralization by asymmetric and symmetric recall. Epilepsy Res 2000; 39:239–249. 46. Lee GP, Park YD, Westerveld M, et al. Wada memory performance predicts seizure outcome after epilepsy surgery in children. Epilepsia 2003; 44:936–943. 47. Roman DD, Beniak TE, Nugent S. Memory performance on the intracarotid amobarbital procedure as a predicator of seizure focus. Epilepsy Res 1995; 25:243–248. 48. Perrine K, Westerveld M, et al. Wada memory disparities predict seizure laterality and postoperative seizure control. Epilepsia 1995; 36:851–856. 49. Andelman F, Kipervasser S, Neufeld MY, et al. Predictive value of Wada memory scores on postoperative learning and memory abilities in patients with intractable epilepsy. J Neurosurg 2006; 104:20–26. 50. Stabell KE, Bakke SJ, Andresen S, et al. Selective posterior cerebral artery amobarbital test: its role in presurgical memory assessment in temporal lobe epilepsy. Epilepsia 2004; 45:817–825. 51. Loring DW, Meador KJ, Lee GP, et al. Wada memory asymmetries predict verbal memory decline after anterior temporal lobectomy. Neurology 1995; 45:1329–1333. 52. Cohen-Gadol AA, Westerveld M, Alvarez-Carilles J, et al. Intracarotid amytal test and hippocampal magnetic resonance imaging volumetry: validity of the Wada test as an indicator of hippocampal integrity among candidates for epilepsy surgery. J Neurosurg 2004; 101:926–931. 53. Chiaravalloti N, Glauser G. Material-specific memory changes after anterior temporal lobectomy as predicted by the intracarotid amobarbital test. Epilepsia 2001; 42:901–911. 54. Kirsch HE, Walker JA, Winstanley BA, et al. Limitations of Wada memory asymmetry as a predictor of outcomes after temporal lobectomy. Neurology 2005; 65:676–680. 55. Jones-Gotman M, Sziklas V, Djordjevic J, et al. Etomidate speech and memory test (eSAM): a new drug and improved intracarotid procedure. Neurology 2005; 65:1723–1729. 56. Takayama M, Miyamoto S, Ikeda A, et al. Intracarotid propofol test for speech and memory dominance in man. Neurology 2004; 63:510–515. 57. Mikuni N, Takayama M, Satow T, et al. Evaluation of adverse effects in intracarotid propofol injection for Wada test. Neurology 2005; 65:1813–1816. 58. Santiago Medina L, Aguirre E, Bernal B, et al. Functional MR imaging versus Wada test for evaluation of language lateralization: cost analysis. Radiology 2004; 230:49–54. 59. Binder JR, Swanson SJ, Hammeke TA, et al. Determination of language dominance using functional MRI: a comparison with the Wada test. Neurology 1996; 46:978–984.

60. Woermann FG, Jokeit H, Luerding R, et al. Language lateralization by Wada test and fMRI in 100 patients with epilepsy. Neurology 2003; 61:699–701. 61. Worthington C, Vincent DJ, Bryant AE, et al. Comparison of functional magnetic resonance imaging for language localization and intracarotid speech amytal testing in presurgical evaluation for intractable epilepsy. Preliminary results. Stereotact Funct Neurosurg 1997; 69:197–201. 62. Gao X, Jiang C, Lu C, et al. Determination of the dominant language hemisphere by functional MRI in patients with temporal lobe epilepsy. Chin Med J 2001; 114:711–713. 63. Deblaere K, Boon PA, Vandemaele P, et al. MRI language dominance assessment in epilepsy patients at 1.0T: region of interest analysis and comparison with intracarotid amytal testing. Neuroradiology 2004; 46:413–420. 64. Sabbah P, Chassouz F, Leveque C, et al. Functional MR imaging in assessment of language dominance in epileptic patients. Neuroimage 2003; 18:460–467. 65. Sabsevitz DS, Swanson SJ, Hammeke TA, et al. Use of preoperative functional neuroimaging to predict language deficits from epilepsy surgery. Neurology 2003; 60:1788–1792. 66. Rutten GJM, Ramsey NF, van Rijen PC, et al. fMRIdetermined language lateralization in patients with unilateral or mixed language dominance according to the Wada test. Neuroimage 2002; 17:447–460. 67. Yetkin FZ, Swanson S, Fischer M, et al. Functional MR of frontal lobe activation: comparison with Wada language results. AJNR Am J Neuroradiol 1998; 19:1095–1098. 68. Lehericy S, Biondi A, Sourour N, et al. Arteriovenous brain malformations: is functional MR imaging reliable for studying language reorganization in patients? Initial observations. Radiology 2002; 223:672–682. 69. Rabin ML, Narayan VM, Kimberg DY, et al. Functional MRI predicts post-surgical memory following temporal lobectomy. Brain 2004; 127:2286–2298. 70. Jokeit H, Okujava M, Woermann F. Memory fMRI lateralizes temporal lobe epilepsy. Neurology 2001; 57:1786–1793. 71. Golby AJ, Poldrack RA, Illes J, et al. Memory lateralization in medial temporal lobe epilepsy assessed by functional MRI. Epilepsia 2002; 43:855–863. 72. Machielsen WC, Rombouts SA, Barkjof F, et al. FMRI of visual encoding: reproducibility of activation. Hum Brain Mapp 2000; 9:156–164. 73. Constable RT, Carpentier A, Pugh K, et al. Investigation of the human hippocampal formation using a randomized event-related paradigm and Z-shimmed functional MRI. Neuroimage 2000; 12:55–62. 74. Balsamo LM, Galliard WD. The utility of functional magnetic resonance imaging in epilepsy and language. Curr Neurol Neurosci Rep 2002; 2:142–149. 75. Papanicolaou AC, Simos PG, Castillo EM, et al. Magnetocephalography: a non-invasive alternative to the Wada procedure. J Neurosurg 2004; 100:867–876. 76. Wheless JW, Castillo E, Maggio V, et al. Magnetoencephalography (MEG) and magnetic source imaging (MSI). Neurologist 2004; 10:138–153. 77. Barkley GL, Baumgartner C. MEG and EEG in epilepsy. J Clin Neurophysiol 2003; 20:163–178. 78. Otsubu H, Ochi A, Elliot I, et al. MEG predicts epileptic zone in lesional extrahippocampal epilepsy: 12 pediatric surgery cases. Epilepsia 2001; 42:1523–1530. 79. Grondin R, Chuang S, Otsubo H, et al. The role of magnetoencephalography in pediatric epilepsy surgery. Childs Nerv Syst 2006; 22:779–785. 80. Cohen D. Magnetoencephalography: detection of the brain’s electrical activity with a superconducting magnetometer. Science 1972; 175:664–666.

Chapter 9: Balloon Occlusion, Wada, and Pharmacological Testing 81. Barr JD, Mathis JM, Horton JA. Provocative pharmacologic testing during arterial embolization. Neurosurg Clin N Am 1994; 5:403–411. 82. Berenstein A, Lasjaunias P. Technical aspects of surgical neuroangiography. In: Berenstein A, Lasjaunias P, eds. Surgical Neuroangiography: Endovascular Treatment of Cerebral Lesions. London: Springer-Verlag, 1992:194–266. 83. Niimi Y, Sala F, Deletis B, et al. Neurophysiologic monitoring and pharmacologic provocative testing for embolization of spinal cord arteriovenous malformation. AJNR Am J Neuroradiol 2004; 25:1131–1138. 84. Rodesch G, Hurth M, Alvarez H, et al. Spinal cord intradural arteriovenous fistulae: anatomic, clinical, and therapeutic considerations in a series of 32 consecutive patients seen between 1981 and 2000 with emphasis on endovascular therapy. Neurosurgery 2005; 57:973–983. 85. Fitzsimmons BFM, Marshall RS, Pile-Spellman J, et al. Neurobehavioral differences in superselective Wada testing with amobarbital versus lidocaine. AJNR Am J Neuroradiol 2003; 24:1456–1460. 86. Deveikis JP. Sequential injections of amobarbital sodium and lidocaine for provocative neurologic testing in the

87.

88.

89.

90.

91.

193

external carotid circulation. AJNR Am J Neuroradiol 1996; 17:1143–1147. Cannestra A, Pouratian N, Forage J, et al. Functional magnetic resonance imaging and optical imaging for dominant-hemisphere perisylvian arteriovenous malformation. Neurosurgery 2004; 55:804–814. Yamada K, Kizu O, Ito H, et al. Tractography for arteriovenous malformations near the sensorimotor cortices. AJNR Am J Neuroradiol 2005; 26:598–602. Itoh D, Aoki S, Marayma K, et al. Corticospinal tracts by diffusion tensor tractography in patients with arteriovenous malformations. J Comput Assist Tomogr 2006; 30:618–623. Hlatky R, Jackson EF, Weinberg JS, et al. Intraoperative neuronavigation using diffusion tensor MR tractography for the resection of a deep tumor adjacent to the corticospinal tract. Stereotact Funct Neurosurg 2005; 83:228–232. Kikuta K, Takagi Y, Nozaki K, et al. Early experience with 3-T magnetic resonance tractography in the surgery of cerebral arteriovenous malformations in and around the visual pathway. Neurosurgery 2006; 58:331–337.

10 Endovascular Management of Tumors and Vascular Malformations of the Head and Neck Johnny C. Pryor and Joshua A. Hirsch Department of Interventional Neuroradiology and Endovascular Neurosurgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A.

Robert W. Hurst Department of Radiology, Neurology, and Neurosurgery, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.

INTRODUCTION

VASCULAR TUMORS

Vascular lesions of the head and neck are a wideranging and often confusing constellation of lesions, and diagnosis and treatment are ideally approached through multidisciplinary cooperation. Mulliken et al. were the first to present a rational classification of hemangiomas and vascular malformations (1–5). With modifications over time, this classification has evolved into a clinically useful system that helps us understand the etiology and pathophysiology of these lesions and gives us a context within which to understand their behavior and plan a rational treatment strategy. This context also helps establish proper communications between physicians and patients. Previously, vascular lesions of the face and head were often ubiquitously referred to as hemangiomas, and this mistake is often still repeated even in modern clinical practice. Mulliken et al. demonstrated that craniofacial vascular lesions can be generally differentiated into vascular tumors or malformations of vascular structures. Craniofacial vascular malformations can involve any or all of the vascular components, including arteries, capillaries, veins, and lymphatic tissue. These malformations may be simple or complex and may represent angiodysplastic syndromes that are congenital and possibly have a genetic basis. Therefore, while most craniofacial vascular lesions present in childhood, many of those that do not, have their origins or predispositions determined in utero. Certain vascular pathologies may arise as a result of exogenous causes such as traumatic arteriovenous fistulas or erosion of a nonvascular tumor into a significant vascular structure; however, even these seemingly unrelated occurrences may have an underlying genetic predisposition. Understanding the underlying pathophysiology allows us to predict clinical behavior and propose rational treatment strategies.

Vascular tumors are differentiated from vascular malformations by their patterns of cellular growth. Hemangiomas are a vascular neoplasm of endothelial cell origin and proliferation of these cells leads to tumor growth. Two types of hemangiomas can be differentiated clinically: cavernous hemangiomas, which occur in adults and are associated with thrombosis pathologically, and capillary hemangiomas, which are found in children and have a propensity to spontaneously involute.

CAPILLARY HEMANGIOMA Capillary hemangiomas are five times more common in girls and are usually detected in the first three months of life. Though they are only identified at birth in about a third of patients, hemangiomas typically display a proliferative phase in the first six months of life that is characterized by rapid growth of these lesions. Gradual involution is usually detected by about the child’s first birthday and is the hallmark of capillary hemangiomas. Involution is typically complete by age seven and will occur in over 95% of patients presenting in infancy. The triggers and mechanisms of this spontaneous involution remain incompletely understood, though an apoptotic mechanism seems important (6). Cavernous hemangiomas proliferate over a different time course and continue to proliferate into and potentially throughout adulthood. Since spontaneous involution can be expected in over 95% of children presenting with capillary hemangiomas, conservative treatment is warranted in the vast majority cases. Nevertheless, no means is currently available to differentiate those lesions fated to progress from the vast majority, which will go on to complete resolution. In addition, some patients may

196

Pryor et al.

require more urgent therapeutic intervention based on the clinical scenario. Such situations include lesions that impinge on or compromise anatomically critical structures and those which cause functional impairment or prevent normal development of critical functions. Examples of complications related to capillary hemangioma that may mandate urgent treatment include congestive heart failure, which may develop as a result of high intralesional flow with arteriovenous shunting. In addition, compromise of airway, feeding, or visual function are indications for immediate treatment. Moreover, hemorrhage or coagulation abnormalities such as consumptive coagulopathy, which in association with rapid tumor growth and thrombocytopenia is known as Kasabach-Merritt phenomenon, is well described with capillary hemangiomas (7). Kasabach-Merritt syndrome is also called the ‘‘hemangioma thrombocytopenia syndrome’’ or thrombocytopenic coagulopathy. It should be emphasized that while KasabachMerritt syndrome has rarely been associated with capillary hemangiomas, it is most often associated with tumors having lymphatic-like vessels such as kaposiform hemangioendothelioma or tufted angioma (8,9). Treatment of Kasabach-Merritt syndrome begins with corticosteroid therapy with the possible addition of dipyridamole (10). The efficacy of high-dose steroids is a matter of debate, especially considering the balance of positive response versus complications. Moreover, interferon, the second line of treatment if the platelet count does not quickly rise, is not without complications. For example, interferon therapy for hemangiomas has been associated with spastic diplegia (11). Lastly, the course of interferon treatment is prolonged and typically extends over a year. Should the platelet count not respond to interferon, other treatment options include chemotherapeutic agents, such as Vinblastine or Vincristine, radiation therapy, or surgical resection. While radiation therapy has been shown to be effective in selected patients, worries about its potential long-term side effects, such as inhibition of regional growth, scarring, and malignancy, limit its application (12). Vincristine has also been shown to be effective in treating patients with respiratory compromise (13). Endovascular embolization with small polyvinyl alcohol (PVA) particles can help stabilize the situation in selected patients and may also be used to buy time for medical therapy to work or as an adjunct to surgical resection (3). Embolization may obliterate arteriovenous shunting through the tumor vasculature and thereby reverse high-output heart failure. Embolization may also markedly shrink capillary hemangiomas to relieve mass effect on the airway or esophagus. Mass effect on the eyelid, which impairs sight from one eye and may permanently interfere with the development of binocular vision in the neonate, represents an additional indication for endovascular treatment (3). Embolization with absolute ethanol is also reported to quickly reverse thrombocytopenia (14). However, one should be aware of potential complications with absolute ethanol, such as significant swelling,

necrosis of normal tissue, hemolysis, renal failure and pulmonary hypertension. In addition, the full effect of absolute alcohol infusion is not immediately angiographically apparent and additional closure of the malformation may occur within a relatively short time.

JUVENILE ANGIOFIBROMA Although a histologically benign neoplasm, juvenile angiofibroma (JAF) often grows aggressively, spreading by local invasion throughout the nasal fossa and anterior skull base. Nearly all patients are male and typically present between the ages of 14 and 17. Symptoms are determined by the tumor size and extent, with unilateral nasal obstruction being the most common initial symptom. Far more importantly, epistaxis, often associated with major blood loss, is the symptom that brings most patients to medical attention. Tumor occlusion of the sinus ostia may obstruct drainage and cause sinusitis. JAF originates from the posterolateral nasopharynx in the region of the pterygopalatine fossa. Extension into adjacent spaces determines staging (15–18). At the time of diagnosis, JAF virtually always extends through the sphenopalatine foramen, involving both the pterygopalatine fossa and posterior nasal cavity. Lateral extension into the infratemporal fossa is also characteristic. The tumor often extends into the sphenoid sinus, while maxillary sinus involvement is much less frequent. Orbital extension via the inferior orbital fissure occurs in approximately one-third of patients (19). Intracranial involvement, present in less than 20%, makes complete resection difficult (20). Imaging studies are essential prior to angiographic evaluation and embolization of JAF to make the diagnosis, delineate the extent of tumor, and differentiate tumor from retained sinus secretions (21). On CT scan, the tumor is isodense to muscle prior to contrast administration (Fig. 1). Avid homogeneous contrast enhancement is characteristic. Remodeling and expansion of the bony walls of the pterygopalatine fossa is most often present. This results in anterior bowing of the posterior wall of the maxillary antrum, a characterisic finding on the lateral skull film. Expansion of the sphenopalatine foramen indicates extension between the pterygopalatine fossa and the nasopharynx. Generally 2 to 6 cm in diameter at the time of diagnosis, JAF also remodels and expands the nasal cavity. MR findings include characteristic ‘‘salt-andpepper’’ appearance on most sequences, with punctate dark regions representing flow voids from the tumor’s high vascularity. Intense enhancement follows contrast administration (Fig. 2). MR imaging is important to differentiate between enhancing tumor and retained sinus secretions. Fat-saturated enhanced images are often best for delineating involvement of the skull base foramina and extension into the cavernous sinus, sphenoid sinus, or intracranial cavity. Initially, arterial supply to JAF arises from the site of origin, with the most common feeders from the pterygopalatine portion of the internal maxillary artery, including the sphenopalatine, infraorbital, and

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

197

Figure 1 Juvenile angiofibroma. An 18-year-old male with epistaxis. Axial CT scan. (A, B) Small posterior nasal mass (arrowhead ) extends into expanded pterygopalatine fossa (*) widening the sphenopalatine foramen. (C, D) Higher cuts show extension anteriorly toward infraorbital foramen and posteriorly toward vidian canal and foramen rotundum.

Figure 2 Juvenile angiofibroma (same patient as in Figure 1). (A–C) Axial T2-weighted MR images demonstrate soft tissue signal intensity with ‘‘salt-and-pepper’’ pattern involving nasal fossa (arrowhead ) as well as pterygopalatine fossa. (D–F) Axial fat-saturated enhanced T1-weighted MR images demonstrating homogeneous, avid enhancement.

198

Pryor et al.

Figure 3 Juvenile angiofibroma (same patient as in Figure 1). (A) Lateral view of ECA injection demonstrates intense tumor blush in the posterior nasal region. (B) Microcatheter injection illustrates nonenlarged feeding arteries, tumor blush, and lack of arteriovenous shunting. (C) Postembolization microcatheter injection of internal maxillary artery—no residual tumor filling. (D) Postembolization CCA injection demonstrates no residual tumor supply. Abbreviations: ECA, external carotid artery; CCA, common carotid artery.

descending palatine branches (Fig. 3). Recruitment of adjacent vessels, including the accessory meningeal, ascending pharyngeal, and ascending palatine arteries, is often seen with larger tumors. Sphenoid sinus extension results in development of blood supply from extradural branches of the internal carotid artery (ICA) (Fig. 4). Supply from pial branches of the ICA although uncommon, should be sought, since it reflects intracranial extension into the anterior and middle fossae (22). Bilateral supply is frequent, particularly with large tumors, and should also be sought in each case. Early, dense, persistent contrast staining characterizes JAF angiographically. Unlike many other vascular neoplasms, including paragangliomas (see

below), significant enlargement of feeding arteries and arteriovenous shunting are uncommon. Complete surgical removal is the therapy of choice. Preoperative embolization of JAF has been shown to reduce both perioperative blood loss and the duration of surgical resection (23–25). Recent experience with endoscopic resection has improved surgical outcomes and reinforces the value of preoperative embolization in JAF (26–28). The location of JAF mandates particular attention to the possibility of orbital or intracranial anastomoses from vessels that also feed tumor. Supply to cranial nerves is also of concern, as in all cases of embolization involving the skull base. Presurgical embolization is usually accomplished using PVA particles of 150 to 350 mm in diameter (29–31).

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

199

Figure 4 Juvenile angiofibroma. A 17-year-old male with severe epistaxis and nasal obstruction. (A) Coronal CT and (B) MR demonstrate avidly enhancing JAF expanding the posterior right nasal cavity with extension through the right sphenopalatine foramen into the pterygopalatine fossa (*, normal left pterygopalatine fossa). Tumor has eroded the nasal septum to cross the midline. Sphenoid sinus extension is present on the right, denoted by (s). (C) AP view of angiogram demonstrates vascular tumor blush involving the same region with supply from the pterygopalatine branches of the internal maxillary artery. (D) Lateral view of angiogram demonstrates vascular tumor blush and internal maxillary supply (internal maxillary artery, arrow). (E) Lateral view shows microcatheter route through the internal maxillary artery (arrow) into the pterygopalatine location from which embolization was performed. (F) Postembolization, CCA injection shows delayed ECA filling with residual tumor supply from ascending pharyngeal artery (arrowhead ) and cavernous branch of the ICA (arrow). The latter supply angiographically indicates the sphenoid sinus extension seen on CT and MR. Abbreviations: ECA, external carotid artery; CCA, common carotid artery; ICA, internal carotid artery.

PARAGANGLIOMAS Paragangliomas, also known as glomus tumors, are highly vascular neoplasms, which arise from chemoreceptor cells of the paraganglia or glomus bodies (32). In the head and neck, the carotid body location is most common. Temporal bone paragangliomas are next in frequency; glomus tympanicum tumors involve the middle ear and glomus jugulare tumors, the jugular fossa. Jugulotympanic paragangliomas involve both temporal bone locations. Paragangliomas associated with the vagus nerve (glomus vagale) and those involving the larynx are less common (33). The usual age of onset is in the fifth decade, but a wide age distribution is reported with early onset in familial cases. A female predominance of nearly 3:1 is present in most head and neck locations, except for carotid body tumors. As many as 50% of all paragangliomas have been found to be hereditary and may be associated with familial paraganglioma, neurofibromatosis type 1, von

Hippel-Lindau disease, the Carney triad, and, rarely, with multiple endocrine neoplasia type 2 (34). Multiplicity is common in hereditary forms, affecting up to one-third of patients. Although less common in nonfamilial cases, multiple tumors may occur in up to 10%, usually with vagal or carotid body locations. The high frequency of heritability and multiple lesions requires imaging evaluation to detect undiagnosed lesions in any patient with a paraganglioma involving the head and neck. Genetic counseling should be offered to all patients with a family history of paraganglioma, while patients positive for paternal paraganglioma locus gene should undergo regular radiological screening with MRI (35). The typically slow, locally invasive growth of paragangliomas causes bony destruction and infiltration of adjacent structures. Lymph node involvement and metastases are rare, reported in less than 5% of cases (36). Clinical manifestations related to mass effect depend on the tumor location.

200

Pryor et al.

Symptoms associated with paragangliomas of the temporal bone reflect the extent of tumor spread. The Fisch classification evaluates tumors on a scale of A through D, with A representing localized tympanic tumor, B mastoid extension, C erosion of the carotid canal (with subtypes), and D indicating intracranial extension (37,38). Typical symptoms include hearing loss, pulsatile tinnitus, and cranial nerve palsy. The latter most often involves the lower cranial nerves or CN VII within the temporal bone. Carotid body tumors present as slow-growing, painless neck masses at the bifurcation of the common carotid artery (CCA). While the vast majority of paragangliomas give histological evidence of catecholamines, clinical hypersecretion occurs in less than 5% of cases and causes signs and symptoms identical to hyperfunctioning adrenal pheochromocytoma: episodic hypertension, headache, nausea, excessive perspiration, and nausea. When a catecholamine-secreting tumor is suspected because of paroxysmal symptoms, biochemical documentation of catecholamine and fractionated metanephrine hypersecretion should precede imaging evaluation (34). Laboratory or clinical evidence of hypersecretion suggests an increased risk of blood pressure alterations during angiography, embolization, or surgical resection. Noninvasive imaging must detect and characterize the lesion, as well as determine multiplicity. CT scanning can show demineralization, erosion, and destruction of the temporal bone structures in jugulotympanic paragangliomas. Features include destruction of the jugular plate, indicating involvement of middle ear structures as well as the jugular fossa. Encasement or displacement of the facial nerve should also be evaluated. Anterior extension with destruction of the vertical segment of the petrous carotid canal may indicate ICA encasement, involvement, and vascular supply. MR is the mainstay of noninvasive imaging. Both T1- and T2-weighted sequences typically demonstrate the characteristic ‘‘salt-and-pepper’’ appearance of high-velocity flow voids within the tumor. Following gadolinium administration, intense enhancement is present, frequently with residual flow voids. The extent of the tumor should be evaluated, including intraluminal jugular vein growth, ICA encasement, dural enhancement, and invasion of posterior fossa structures. MRA using 3D TOF or contrast-enhanced sequences may be useful to initially delineate arterial supply. Angiography may also provide diagnostic information when the diagnosis is in doubt (39). Preoperative angiographic evaluation, guided by MRI, is however necessary for all but very small tumors confined to the middle ear or involving the carotid body. Angiographic features of paraganglioma reflect the high vascularity with enlarged feeding arteries, rapid appearance of dense vascular tumor blush, and early venous drainage. Paragangliomas often consist of multiple compartments, each receiving separate arterial supply. Multiple superselective catheterizations may therefore be necessary to opacify or embolize the entire tumor. Evaluation of the normal intracranial venous

outflow is also necessary, should jugular vein or sigmoid sinus sacrifice be required. The angiographic differential diagnosis for vascular lesions of the temporal bone includes other highly vascular neoplasms including metastases from thyroid, or renal cell carcinoma, and hemangiopericytoma. In addition, nonneoplastic lesions, including aberrant ICA, dural arteriovenous fistulas, and petrous or cavernous ICA aneurysms, while easily differentiated by noninvasive imaging, may present a highly vascular angiographic appearance in the skull base (31,40). Complete surgical resection, usually with preoperative embolization of major external carotid artery (ECA) feeding pedicles, is the mainstay of therapy. In very large or unresectable tumors, radiotherapy, conventional or stereotactic, is a viable therapeutic option (41–43). Preoperative embolization confined to the ECA supply usually gives the most favorable risk/benefit ratio (44). Occlusion testing of the ICA may also be necessary when carotid encasement is present. Superselective angiography is necessary to delineate cranial nerve supply or dangerous anastomoses with the ICA or vertebral artery. Preoperative embolization is usually performed using particles of PVA from 150 to 350 mm in diameter. This agent has been shown to be both safe and effective in reducing intraoperative blood loss in paragangliomas (45,46). Safe and effective embolization requires knowledge of vascular anatomy specific to the particular tumor location. Jugulotympanic paragangliomas virtually always receive major supply from ascending pharyngeal artery branches to the most common areas of origin (47). The middle ear receives supply from the inferior tympanic branch, while the neuromeningeal branch supplies both the jugular fossa and hypoglossal canal (Fig. 5). Additional ECA supply, including the stylomastoid branch of the occipital artery, is also extremely common in jugulotympanic paragangliomas. In addition, supply from the temporal branch of the middle meningeal artery (MMA) may also occur in tumors with anterior extension. Extradural ICA supply from caroticotympanic or cavernous branches may also occur, particularly with anterior extension of temporal bone paragangliomas. Pial supply from branches of either the internal carotid or vertebrobasilar system indicates transdural invasion and usually a worse prognosis. Carotid body tumors characteristically widen the CCA bifurcation, splaying apart the proximal ICA and ECA. Early dense tumor blush between the two vessels reflects tumor vascularity, usually originating from the proximal ECA. Ascending pharyngeal artery supply may also be present, typically to the superior aspect of the tumor. Multiple short feeders originating in proximity to the CCA bifurcation mandate meticulous fluoroscopic monitoring during embolization to prevent emboli from entering the ICA. Anterior displacement of the cervical ICA is often seen in cases of glomus vagale tumors. The musculospinal branch of the ascending pharyngeal artery typically supplies vagal paragangliomas inferior to the skull base (Fig. 6). Supply from deep and

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

201

Figure 5 Glomus jugulare. (A) Axial CT scan demonstrates enlargement and bony erosive changes involving right jugular fossa (arrowheads). (B) Right CCA angiogram demonstrates highly vascular glomus jugulare tumor (arrow). (C) Lateral view of microcatheter injection of ascending pharyngeal artery demonstrates major tumor supply. (D) Postembolization angiogram demonstrates no residual vascularity. Abbreviation: CCA, common carotid artery.

Figure 6 Carotid body and glomus vagale. (A) Sagittal enhanced fat-saturated T1-weighted MR demonstrates left-sided carotid body tumor and glomus vagale tumor (* ). (B) Left CCA angiogram demonstrates intense tumor blush with characteristic splaying of carotid bifurcation due to carotid body tumor and anterior displacement of the cervical ICA due to glomus vagale tumor (*). (C) Microcatheter angiogram during embolization, demonstrates dangerous anastomosis with the vertebral artery (arrow). No embolization was performed from this site (arrowhead, microcatheter tip). (D) Postembolization left CCA angiogram demonstrates decreased tumor blush characteristic displacements of the ICA remain. Abbreviations: CCA, common carotid artery; ICA, internal carotid artery.

202

Pryor et al.

anterior cervical arteries as well as the inferior thyroidal artery may also be present.

MENINGIOMAS Meningiomas, the most common nonglial primary intracranial tumor, account for approximately 15% of primary intracranial tumors in the general population. Believed to arise from arachnoid granulation cells, meningiomas most often present between 20 and 60 years of age. A female predominance of nearly 3:1 is present, with meningiomas comprising more than 50% of primary intracranial tumors in women (48). Meningiomas most often arise over the cerebral convexities, particularly in the parasagittal area (Fig. 7). Other common locations include the sphenoid wing, parasellar region, olfactory groove, and tentorium. Approximately 10% of meningiomas involve the posterior fossa (49,50). Multiple meningiomas may occur in up to 10% of cases and may rarely appear in children, in which case initiating factors such as prior

radiation or underlying genetic abnormalities such as neurofibromatosis type 2 must be considered (51–54). Because meningiomas arise from the dura, meningeal arterial branches, usually originating from the ECA, are the initial source of blood supply to the majority of the tumors. Nevertheless, meningiomas often recruit blood supply from adjacent meningeal arteries or invade the dura, receiving supply from pial vessels. However, the arterial supply to the tumor’s meningeal site of origin continues and aids prediction of primary blood supply on the basis of tumor location. The MMA, which supplies dura of the sphenoid wing, cerebral convexities, and much of the anterior fossa, is the vessel most often providing arterial supply to meningiomas. MMA supply is often bilateral in meningiomas of the parasagittal region or those crossing the midline. Meningiomas of the olfactory groove most often receive supply from dural branches of the ICA, including ethmoidal artery supply from the ophthalmic artery. Tumors originating from the tentorium and clivus may receive supply from cavernous ICA branches in addition to the MMA (Fig. 8) (55). Embolization of

Figure 7 Meningioma. (A). Sagittal T1-weighted enhanced MR demonstrates parasagittal meningioma with considerable mass effect. (B) Lateral left CCA injection demonstrates meningioma supply originating from MMA and ACA. (C) Microcatheter injection of left MMA shows tumor supply with early venous drainage (arrow, catheter tip). No orbital communication is present. (D) Postembolization angiogram demonstrates decrease in tumor supply. Abbreviations: CCA, common carotid artery; MMA, middle meningeal artery; ACA, anterior cerebral artery.

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

203

Figure 8 (A) Sagittal T1-weighted enhanced MR demonstrates extra-axial mass on the superior surface of the tentorium. (B) Lateral view of CCA angiogram shows dense contrast blush involving the mass. (C) Superselective injection of MMA demonstrates enlargement of tentorial branch supplying the meningioma. Note ‘‘spokewheel’’ appearance of tumor vasculature (arrowhead, microcatheter tip). (D) Postembolization angiogram demonstrates no residual tumor filling. Abbreviations: CCA, common carotid artery; MMA, middle meningeal artery.

these ICA dural branches is often associated with increased risk and is not commonly performed. Posteromedially located posterior fossa meningiomas often receive supply from meningeal branches of the vertebral artery. Those more laterally located in the posterior fossa commonly receive transmastoid branches of the occipital artery as well as contributions from the ascending pharyngeal artery. Similar to the case with ICA dural branches, preoperative embolization of meningeal arteries originating directly from the vertebral artery or of pial branches does not provide as favorable a risk/benefit ratio as for embolization of supply that originates from branches of the ECA (56,57). Angiographic evaluation of meningiomas is guided by noninvasive imaging, but provides additional information currently impossible to obtain noninvasively. The angiogram should be designed to

acquire particular information to aid the safest and most complete resection possible. Vascular supply to the tumor, including dural supply, pial supply, and any evidence of transosseous supply, must be identified. Anatomic variants, particularly those involving the MMA and the ophthalmic artery, may impact the ability to safely embolize and must be identified (Figs. 9 and 10). Examination of the venous phase is also exceedingly important. In particular, depiction of large cortical and deep-draining veins often impacts the surgical approach. In addition, patency of dural venous sinuses adjacent to or involved by the tumor, particularly the sigmoid, transverse, and superior sagittal sinuses, must be evaluated. The angiographic appearance of meningioma typically demonstrates a ‘‘spokewheel’’ pattern of intratumoral arteries centered on the primary meningeal

204

Pryor et al.

Figure 9 (A) Axial FLAIR and (B) coronal-enhanced T1-weighted images demonstrate an enhancing duralbased lesion crossing the falx near the vertex. Several areas of punctuate flow void are present within the lesion. An enhancing ‘‘dural tail’’ is present (arrow). Abbreviation: FLAIR, fluid-attenuated inversion-recovery.

Figure 10 (Same patient as in Figure 9). (A) AP and (B) lateral views of the right CCA angiogram demonstrate an enlarged parietal branch of the MMA (arrows) supplying the meningioma. In addition, a frontal dural branch (open arrowheads), normally a branch of the MMA, originates from the ophthalmic artery. (C) Lateral view of the right ECA angiogram also demonstrates the enlarged parietal branch (arrows) of the MMA with tumor blush. Embolization of this branch was accomplished prior to surgical resection. Abbreviations: CCA, common carotid artery; MMA, middle meningeal artery; ECA, external carotid artery.

feeding vessel. A contrast stain or blush outlining the tumor may be present depending on the tumor vascularity. Some tumors show little vascularity, while others are highly vascular and manifest rapid prominent venous drainage. Preoperative embolization of feeding arteries to meningiomas is usually confined to larger tumors or those manifesting high vascularity. Designed to decrease blood loss and engender tumor necrosis, quantitative measurement of the success of embolization has proved difficult. Nevertheless, evidence of reduction in perioperative transfusion associated with embolization has been presented (58). In addition, there is also evidence of decreased blood loss correlating with loss of previously present enhancement on

MR and following the use of relatively small embolic particles (59,60). The skull base origin of the meningeal arteries most commonly involved in meningioma supply means that close inspection of the selective microcatheter angiogram for dangerous anastomotic connections assumes paramount importance. Testing with intraarterial sodium amytal or lidocaine may aid in the clinical detection of anastomoses or cranial nerve supply, which might increase risk of neurological deficit (see chap. 9). Preoperative embolization of meningiomas is usually performed using particles of PVA, although liquid embolic agents have also been used in highly vascular tumors. Complications associated with the procedure have been found to be low (58).

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

VASCULAR MALFORMATIONS In contrast to neoplasms, such as hemangiomas, vascular malformations are nonproliferative in nature and enlarge by recruiting flow rather than increasing the number of cells. Vascular malformations may be comprised of any or all vessel types singly or in combination, including arterial, capillary, venous, and lymphatic malformations. Vascular malformations are typically present at birth and often grow proportionally with the patient. They do not spontaneously regress and are usually apparent throughout life. A spectrum of flow characterizes the lesions. Venous, capillary, and lymphatic vascular malformations are low-flow lesions, while arteriovenous malformations (AVMs) or fistulas demonstrate very high flow. Venous malformations, composed of saccular venous channels are the most common vascular malformation of the head and neck requiring treatment. They represent up to 50% of patients seen at centers treating such lesions (61). Capillary malformations such as port-wine stains involve pathological capillary and venular-sized vessels in the dermis and may be the most common cutaneous vascular anomaly in the general population. Although they are often confused with the harmless vascular birthmarks that usually occur in the neck, eyelids, glabella, or lips of up to 40% of newborns, ‘‘stork bites’’ typically disappear spontaneously within a year leaving no trace and are not considered to be dermatopathological lesions. AVMs are characterized by a ‘‘short circuit’’ or nidus between the feeding arteries and draining veins with no intervening capillaries, and their presentation may vary from congestive heart failure at birth to mass or hemorrhage in early adulthood. Lymphatic malformations are composed of clusters of channels derived from defective lymphatic vessels within the cutaneous and subcutaneous compartments. They are found most commonly in the cervicofacial area and are usually microcystic in that location. Combined malformations involving arterial, capillary, venous, and lymphatic components also exist and are more often found unilaterally, especially on the limbs. Differentiation between high- and low-flow lesions has important imaging and treatment implications. High-flow lesions (AVMs) are usually well demonstrated on angiography and are frequently amenable to treatment via transarterial catheterbased embolization techniques. Low-flow lesions are not readily demonstrable angiographically, although occasionally angiography may be used to exclude rapid flow indicative of an AVM and thereby confirm the correct diagnosis. Low-flow lesions are usually imaged noninvasively and treated via percutaneous injections and sclerotherapy (62).

AVM AVMs of the head and neck are far less common than low-flow vascular anomalies. Nevertheless, these lesions deserve attention because of their frequent presentation with cosmetic defects or life-threatening hemorrhage.

205

Kohout et al. reported an extensive retrospective analysis of 81 patients with AVM of the head and neck evaluated over a period of 20 years (63). The age of presentation varied widely, from 2 to 66 years with a male to female ratio of 1:1.15. A vascular anomaly was apparent at birth in 59% of patients (82% in men, 44% in women). Ten percent of patients noted onset in childhood, 10% in adolescence, and 21% in adulthood. Eight patients first noted the malformation at puberty, and six others experienced exacerbation during puberty. Fifteen AVMs appeared or expanded during pregnancy. Clinical presentation was categorized according to the authors’ modification of Schobinger clinical staging: 27% in stage I (quiescence), 38% in stage II (expansion), and 38% in stage III (tissue destruction). There was a single patient with stage IV malformation (cardiac decompensation). Stage I lesions remained stable for long periods. Expansion (stage II) was usually followed by pain, bleeding, and ulceration (stage III). Once present, these symptoms and signs inevitably progressed until the malformation was resected. Sites of occurrence could be categorized in anatomic patterns. Sixty-nine percent occurred in the midface, 14% in the upper third of the face, and 17% in the lower third. The most common sites were the cheeks (31%), ear (16%), nose (11%), and forehead (10%) (63). Angiographically, high-flow malformations such as AVMs or fistulas are characterized by enlarged arteries and veins with early filling of the draining veins (Fig. 11). Depending on the complexity of the lesion, the relationship of the feeding arteries and draining veins with the nidus may be difficult to define. Superselective angiography is often necessary to best define the anatomy, particularly in complex lesions. Angiographic evaluation of head and neck AVMs must include all vessels likely to provide arterial supply. Persky et al. found the inferior alveolar artery characteristically supplied AVMs of the mandible (Fig. 12). Supply to regions of soft tissue extension depended on additional sites involved (e.g., labial and submental arteries to the lower lip and floor of mouth, occipital artery to the ear lobe, and the masseteric branch of the internal maxillary artery to the masseter muscle). AVMs isolated to the mandible often had contralateral supply from branches of the lingual, facial, and inferior alveolar arteries (64). They found that supply to maxillary AVMs consisted of distal branches of the internal maxillary artery. Maxillary arterial vascular malformations with soft tissue extension were supplied by their corresponding arterial systems, and the ophthalmic artery was most commonly recruited. Both the internal maxillary and facial artery systems supplied combined maxillary/ mandibular arterial vascular malformations. MR typically demonstrates enlarged, serpiginous vessels with flow voids and phase shifts. Use of fat-saturation technique with T2 and gadoliniumenhanced T1 images are often beneficial in providing enhanced detail (65). MR is often superior because CT image quality can be compromised by dentition and metal artifacts such as dental appliances or fillings.

206

Pryor et al.

Figure 11 Facial AVM. (A) Axial T1-weighted MR demonstrates left facial swelling with flow voids (arrowheads). (B) Lateral left ECA injection demonstrates AVM supplied by left facial artery (arrowhead, early venous drainage). (C) Microcatheter injection of left facial artery demonstrating AVM supply and early venous drainage (arrowhead ). (D) Postembolization angiogram demonstrates no residual filling, patient underwent surgical resection. Abbreviations: AVM, arteriovenous malformation; ECA, external carotid artery.

Although the lesion may appear compact on crosssectional imaging or angiography, the appearance may be deceptive because the surrounding tissue is often predisposed to develop shunting. This event may lead to recurrence or recanalization via collaterals through the adjacent tissue. AVM involvement of bony structures of the face may be documented on CT, MR, and plain or Panorex films. CT findings include radiolucencies, often having the appearance of a honeycomb or soap bubbles, with small rounded and irregular lacunae. Root resorption has been observed, creating an appearance of teeth floating in the adjacent alveolar osseous erosion. The lesion is often well demarcated and may mimic the appearance of odontogenic cysts (66). Recognition of bony involvement, particularly the roots of the teeth, is important because catastrophic bleeding may occur when teeth become loose, are pulled, or are lost in an uncontrolled fashion.

Kohout et al. found bony involvement in 22 patients: 11 in the nasomaxillary region and 8 in the mandible. In 7 patients, the bone was the primary site; in 15 other patients, the bone was involved secondarily. Nasomaxillary AVMs invariably encompassed the overlying soft tissues, whereas mandibular AVMs were confined to bone in 50% of cases. In both sites, AVM extensively permeated and expanded the bone, crossing the midline in many patients. Persky et al. retrospectively reviewed 26 patients with AVM involving the mandible and/or maxilla (64). They also found that soft tissue involvement was near universal with maxillary lesions, while that associated with mandibular lesions was more limited. MR imaging of low-flow lesions such as venous vascular malformations are characterized by significant enhancement, intermediate T1 signal, and heterogeneous high T2 signal and may demonstrate venous lakes and/or phleboliths, which appear as foci of low

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

207

Figure 12 Mandibular AVM. (A) Lateral CCA injection demonstrates intraosseous mandibular AVM supplied by inferior alveolar artery. (B) Microcatheter injection of the inferior alveolar artery demonstrates AVM with intraosseous venous drainage. (C) Post-NBCA embolization, glue cast of arterial supply, AVM nidus, and proximal venous drainage (arrowheads). (D) Postembolization angiogram demonstrates minimal residual AVM. Abbreviations: CCA, common carotid artery; AVM, arteriovenous malformation; NBCA, N-butyl cyanoacrylate.

signal (67). These lesions may also demonstrate septations or be associated with satellite nodules (68). Treatment options for venous and other slow flow malformations include surgical resection, laser treatment of skin or mucous membranes, or percutanous sclerotherapy with a variety of agents, including ethanol, sodium tetradecyl sulfate, OK-432, bleomycin, as well as others (69). In performing percutaneous sclerotherapy, one must be careful to have free flow of sclerotic agent within the vascular channels. Extravasation of these agents into the soft tissue may produce significant damage to skin or mucous membranes manifested by blistering or deep ulceration. In addition, severe damage to adjacent structures such as the cranial nerves or the orbital contents may result, especially with ethanol. Hematuria is a frequent sequel and local swelling is often pronounced. Swelling can be controlled by elevating the head of the bed and early ambulation along with ice packs. Powdered contrast

agents such as Amipaque are no longer commercially available to mix with ethanol and thereby allow visualization, making precise delivery impossible. In our practice, we mix sodium tetradecyl sulfate (Sotradecol 3%, Bioniche Pharma U.S.A., Inc., Bogart, Georgia, U.S.) 2:1 with Iopamidol 76% (Isovue-370, Bracco Diagnostics, Inc., Princeton, New Jersey, U.S.) to make a 1% injection. This mixture can be ‘‘foamed’’ by passing between syringes so that the material is more stable at the site of injection. Ethanol (dehydrated alcohol, American Regent, Inc., Shirley, New York, U.S.) is approximately 98% pure without bacteriostatic agents and is injected without alteration. We typically use small (22- or 24-gauge) intravenous catheters (Jelco, Medex, Inc., Carlsbad, California, U.S.) and approach the lesion through adjacent normal skin. Ultrasound guidance may be useful. When free flow of blood is demonstrated, one may image the lesion and identify routes of venous outflow by gently injecting contrast.

208

Pryor et al.

A piece of connecting tubing can be connected to prevent inadvertent loss of position of the angiocatheter due to hand movement or torquing. Bleeding from the puncture site may be controlled by injecting a very thin slurry of collagen (Avitene, Davol, Inc., Cranston, Rhode Island, U.S.) through the angiocatheter as it is being withdrawn. One must take into consideration that thrombosis and inflammation will continue to occur for some period of time after the injection of material has ceased that may lead to unwanted tissue damage, particularly with ethanol. We believe it prudent to stop injections while the lesion still demonstrates filling to decrease the risk of complications, although patients should be warned that further embolization sessions may be needed. Lymphatic malformations also respond well to sclerotherapy, with macrocystic lesions responding particularly well to OK-432, which modulates the immune response (70). Surgery and/or laser therapy may also play a role in the treatment of lymphatic malformations. Capillary vascular malformations have shown good results when treated with laser therapy, although other treatment options may be considered in special cases (71). Treatment of high-flow lesions such as AVMs can include surgery, laser, embolization, sclerotherapy, or a combination of options. A multidisciplinary team approach is required for the assessment and treatment of these lesions, which typically involves preoperative angiography with superselective embolization, followed by resection of the lesion (72). Embolization options include particles, coils, N-butyl cyanoacrylate (NBCA; TruFill Liquid Embolic System, Cordis Neurovascular, Inc., Miami Lakes, Florida, U.S.), Onyx (Onyx Liquid Embolic System, Micro Therapeutics, Inc., Irvine, California, U.S.), or other materials. Sclerosing agents can be injected intraarterially or via direct percutaneous puncture using an intravenous angiocatheter and include, primarily, ethanol and Sotradecol. Important considerations in selection of appropriate interventional methods include: briskness of flow; relationship to nerves; proximity to the surface of the skin or mucous membranes; relation to the vermillion border of the lips, the ear, or hairbearing areas; and pigmentation of the skin. Platinum coils may be placed to slow flow in extremely highflow lesions or to treat fistulas that are difficult to define angiographically or are in particularly dangerous areas, such as in close relation to the spinal cord. While coils are useful, Kohout et al. have emphasized that proximal ligation of arterial feeding vessels by any technique frequently resulted in rapid clinical progression and acquisition of diffuse collateral arterial supply and is to be avoided if possible (63,73). Until very recently, our preferred agents of choice have been ethanol and NBCA; however, our recent experience has led us to rely on Onyx for embolizing AVMs that are not near the surface of the skin. One should be aware that Onyx is opacified with microparticulate tantalum in suspension and will cause permanent visible staining of the skin if injected near the surface. However, the performance characteristics of Onyx allow excellent penetration and prolonged injec-

tions producing essentially anatomic injections of deep malformations. Practitioners need to be aware, however, of proximity to important anatomical structures, such as blood supply to cranial nerves and the eyes, as well as dangerous anastomoses to vessels communicating with the internal carotid, vertebral, or other critical arteries. We have found that penetration of embolic material into malformations, especially large, complicated ones, seems to be significantly improved with Onyx compared with NBCA. Our recent experience also suggests that the decrease in flow following embolization with Onyx is much more consistent and predictable than with sclerosing agents such as ethanol, which may be because ethanol does not remain in contact with the vessel wall for a sufficient period of time to damage the endothelium and promote complete thrombosis and scarring. This issue can be corrected at least partially by manually occluding the venous outflow, if the venous drainage is accessible. This maneuver is also helpful to prevent ethanol penetrating and damaging sensitive structures, such as the orbital veins. Direct percutaneous attack can be used when microcatheters cannot be navigated distal enough in very tortuous, normal anatomy to allow safe deposition of material. Ultrasound may be useful in directing percutaneous catheter placement. Superficial bleeding can usually be controlled by direct pressure or via percutaneous injection of Avitene slurry through an intravenous angiocatheter, though at times these techniques need to be augmented or supplemented with other open endovascular or surgical procedures. Local control of bleeding at the time of embolization may rarely be necessary but is especially difficult, particularly in bony lesions, because pressure is difficult to apply within the tooth socket without the presence of the tooth. Urgent control of bleeding from AVMs with mandibular involvement by local pressure may be facilitated by retaining the tooth. Using a tea bag to tamponade bleeding may help, presumably because of the vasoconstricting properties of tannic acid. Avitene or materials that promote clotting may also be useful.

Epistaxis While a common clinical problem, the vast majority of epistaxis is minor in magnitude, originates from the anterior nasal septum, and is self-limited or ceases with a short period of nasal compression. Epistaxis originating posteriorly in the nasal cavity, however, is inaccessible to direct pressure or cauterization and may be life threatening. In many cases, permanent hemostasis is achieved by the use of nasal packing to provide tamponade within the nasal cavity. When nasal packing fails, surgical ligation of the distal branches of the internal maxillary artery supplying the nasal fossa has been advocated. Both prolonged nasal packing and internal maxillary ligation have been associated with patient discomfort, complications, and recurrence (74,75). In such cases, endovascular techniques, specifically arterial embolization, may rapidly relieve most cases of epistaxis and permit removal of packing without surgical intervention.

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

A relatively broad differential diagnosis must be considered in patients presenting with severe epistaxis. The vast majority of cases, considered idiopathic, arise from the effects of longstanding hypertension and arteriosclerosis on the vessels of the nasal mucosa. Often exacerbated by low humidity, patients often present in the fall, when home heating systems lacking humidifiers are turned on. Nevertheless, in any case of severe epistaxis, it is imperative that a specific etiology be identified, if present, so that the most effective treatment can be directed to the underlying cause of the bleeding. Conditions that should be considered in the differential diagnosis for severe epistaxis include 1. 2. 3. 4. 5.

idiopathic epistaxis, which often occurs in elderly patients with hypertension or evidence of generalized atherosclerosis; tumor, e.g., JAF, nasal polyps, malignant nasal, or sinus tumors; coagulopathy, including medications, uremia, or hepatic failure; hereditary hemorrhagic telangiectasia (HHT); and trauma, with damage to ECA or ICA, particularly in the cavernous or petrous segments, including aneurysms or pseudoaneurysms.

Pre-embolization evaluation should exclude specific etiologies, including neoplasm and traumatic

209

lesions. A careful medication history should be taken with particular attention to newly added medications, such as antiplatelet agents. HHT, or Rendu-Osler-Weber syndrome, should be given particular consideration in patients with multiple episodes of epistaxis or a family history. A number of features of the physical examination, history, and angiogram suggest the diagnosis (Fig. 13). HHT is a common disorder that affects multiple organ systems, the manifestations of which arise from inherited abnormalities of vascular structure (76,77). These genetically determined abnormalities result in the development of arteriovenous communications of varying sizes. The nasal mucosa, skin, lung, gastrointestinal tract, and central nervous system are most frequently affected. In the nasal mucosa and skin, the usually tiny lesions are referred to as telangiectasias. In the central nervous system, both AVMs and direct arteriovenous fistulas have been described, while in the lung most lesions are direct arteriovenous fistulas. Gastrointestinal tract lesions include telangiectasias, AVMs, as well as angiodysplasias. The four major diagnostic criteria for HHT include 1. 2.

epistaxis, e.g., spontaneous, recurrent nose bleeds; telangiectasias, which are usually multiple and occur at characteristic sites, including the lips, oral cavity, fingers, and nose;

Figure 13 Hereditary hemorrhagic telangiectasia. A 72-year-old man presented to the emergency room with severe epistaxis that had occurred intermittently for over 24 hours. He reported a family history of epistaxis affecting his mother, brother, and several cousins. (A, B) Physical examination revealed multiple red, slightly raised lesions on his lips, tongue, conjunctiva, fingertips, and nail beds. Unsubtracted (C) and subtracted (D) late arterial phase RICA angiographic examination demonstrates telangiectasias as multiple intravascular contrast collections involving the mucosa of the lips, tongue, and nasal fossa (arrows). (E) Unsubtracted and (F) subtracted lateral right internal maxillary injection demonstrates telangiectasia (arrow) with early venous drainage (arrowhead ) indicating arteriovenous shunting from these multiple lesions (*, microcatheter tip). (G) Lateral view of CCA injection after embolization of internal maxillary arteries was performed bilaterally using PVA particles. Embolization was followed by removal of nasal packing with no recurrence of epistaxis. Abbreviations: CCA, common carotid artery; PVA, polyvinyl alcohol.

210

Pryor et al.

Figure 14 Petrous carotid aneurysm. Patient presented with severe epistaxis. (A) Axial CT scan (bone windows) demonstrates sharply marginated expansile skull base lesion (arrowheads) with extension into the middle ear (arrow) and sphenoid sinus. (B) Right CCA angiogram identifies an aneurysm of the petrous ICA as source of epistaxis. (C) Right common carotid angiogram after coil embolization of aneurysm with parent vessel occlusion (arrows, coil mass). Abbreviation: ICA, internal carotid artery.

3.

4.

visceral lesions, including gastrointestinal telangiectasia (with or without bleeding), pulmonary AVM, hepatic AVM, cerebral AVM, and spinal AVM; and family history of a first-degree relative with HHT according to these criteria.

The diagnosis of HHT is considered ‘‘definite’’ if three of the above criteria are present, ‘‘possible’’ if two criteria are present, and ‘‘unlikely’’ if only one is present (78). Nasal telangiectasias are responsible for the most common manifestation of HHT, epistaxis. This symptom occurs in over 90% of affected patients, usually beginning before the third decade. Severity, while variable, tends to increase with age, often leading to chronic anemia and requiring multiple episodes of treatment. The lesions of the nasal fossa can be identified angiographically, suggesting the diagnosis. Embolization has been shown to be a safe and effective treatment for prolonged epistaxis and can be repeated if necessary (79). Pulmonary arteriovenous fistulas have been identified in 10% to 15% of HHT patients. Conversely, it is estimated that over half of the patients with pulmonary AV fistulas have HHT. The direct rightto-left shunts may initially manifest as neurological deficits caused by cerebral emboli. Most often diagnosed by CT angiography, these lesions are currently treated with endovascular techniques. Recent genetic studies have confirmed that HHT is inherited as an autosomal dominant trait whose penetrance and expressivity are variable. Mutations involving either of two genes, endoglin or ALK-1, may cause HHT. Two disease subtypes, HHT1 and HHT2, result from mutations of endoglin or ALK-1, respectively. The variability of disease severity in different family members suggests, however, that other factors in addition to the specific mutations modify the HHT phenotype (80). Normally most patients with epistaxis are given a trial of nasal packing. If this treatment is unsuccessful, endovascular embolization should be considered in management of most cases. Embolization of the arterial supply to the nasal fossa is most often successful in idiopathic cases as well as in specific

etiologies of epistaxis directly related to disease of the nasal fossa vessels, including uncorrectable coagulopathy and HHT. The procedure is associated with very low complication rates (81). In the angiographic evaluation of epistaxis, visualization of both the ICA and ECA is necessary. ICA evaluation excludes rare vascular lesions, such as petrous or cavernous aneurysms, which might cause epistaxis and whose presence would change therapy (Fig. 14). Sources of collateral supply to the nasal fossa should be identified, including the facial arteries and ethmoidal branches of the ophthalmic artery. In idiopathic epistaxis, no significant vascular abnormalities are normally identified. Most cases of idiopathic epistaxis respond to bilateral embolization of the pterygopalatine branches of the internal maxillary arteries, which give the majority of supply to the nasal fossa. PVA particles (150–350 mm) are usually the preferred embolic material. In most cases, proximal blockage of the internal maxillary arteries using coils or other devices is not necessary, is usually ineffective in preventing collateral formation, and interferes with retreatment. Embolization of the distal facial artery may also be necessary to block collateral vascularization, particularly in patients who have undergone prior ligation of the internal maxillary artery or its major branches. Ethmoidal branches of the ophthalmic artery may also provide collateral supply to the nasal fossa, particularly in patients who have undergone prior surgery for maxillary ligation. Embolization of these vessels is not routinely performed because of the risk to vision. Typically, patients undergo embolization with nasal packing in place. Following initial embolization of the internal maxillary branches, packing is removed in the angiographic suite before removal of the femoral sheath. If hemorrhage persists, additional embolization of additional collaterals may then be performed.

CONCLUSION Vascular lesions of the face and head are a complex group of pathologies that are undergoing explosive growth in understanding and treatment. Diagnosis

Chapter 10: Endovascular Management of Tumors and Vascular Malformations of the Head and Neck

and treatment are best carried out by a multidisciplinary team, including specialists bringing unique interests, knowledge, and skills.

REFERENCES 1. Finn MC, Glowacki J, Mulliken JB. Congenital vascular lesions: clinical application of a new classification. J Pediatr Surg 1983; 18(6):894–900. 2. Glowacki J, Mulliken JB. Mast cells in hemangiomas and vascular malformations. Pediatrics 1982; 70(1):48–51. 3. Lasjaunias P, Berenstein A. Surgical Neuroangiography. Vol. 2: Endovascular Treatment of Craniofacial Lesions. Berlin: Springer-Verlag, 1987:317–340. 4. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69(3):412–422. 5. Mulliken JB, Zetter BR, Folkman J. In vitro characteristics of endothelium from hemangiomas and vascular malformations. Surgery 1982; 92(2):348–353. 6. Razon MJ, Kraling BM, Mulliken JB. Increased apoptosis coincides with onset of involution in infantile hemangioma. Microcirculation 1998; 5(2–3):189–195. 7. Kasabach H, Merritt K. Capillary hemangioma with extensive purpura: report of a case. Am J Dis Child 1940; 59:1063–1070. 8. Enjolras O, Wassef M, Mazoyer E, et al. Infants with Kasabach-Merritt syndrome do not have ‘‘true’’ hemangiomas. J Pediatr 1997; 130(4):631–640. 9. Sarkar M, Mulliken JB, Kozakewich HP, et al. Thrombocytopenic coagulopathy (Kasabach-Merritt phenomenon) is associated with Kaposiform hemangioendothelioma and not with common infantile hemangioma. Plast Reconstr Surg 1997; 100(6):1377–1386. 10. Wananukul S, Nuchprayoon I, Seksarn P. Treatment of Kasabach-Merritt syndrome: a stepwise regimen of prednisolone, dipyridamole, and interferon. Int J Dermatol 2003; 42(9):741–748. 11. Barlow CF, Priebe CJ, Mulliken JB, et al. Spastic diplegia as a complication of interferon alfa-2a treatment of hemangiomas of infancy. J Pediatr 1998; 132(3 pt 1):527–530. 12. Frevel T, Rabe H, Uckert F, et al. Giant cavernous haemangioma with Kasabach-Merritt syndrome: a case report and review. Eur J Pediatr 2002; 161(5):243–246. 13. Moore J, Lee M, Garzon M, et al. Effective therapy of a vascular tumor of infancy with vincristine. J Pediatr Surg 2001; 36(8):1273–1276. 14. Yang YH, Lee PI, Lin KH, et al. Absolute ethanol embolotherapy for hemangioma with Kasabach-Merritt syndrome. Zhonghua Min Guo Xiao Er Ke Yi Xue Hui Za Zhi 1998; 39(1):51–54. 15. Antonelli AR, Cappiello J, Di Lorenzo D, et al. Diagnosis, staging, and treatment of juvenile nasopharyngeal angiofibroma (JNA). Laryngoscope 1987; 97(11):1319–1325. 16. Paris J, Guelfucci B, Moulin G, et al. Diagnosis and treatment of juvenile nasopharyngeal angiofibroma. Eur Arch Otorhinolaryngol 2001; 258(3):120–124. 17. Onerci M, Ogretmenoglu O, Yucel T. Juvenile nasopharyngeal angiofibroma: a revised staging system. Rhinology 2006; 44(1):39–45. 18. Fisch U. The infratemporal fossa approach for nasopharyngeal tumors. Laryngoscope 1983; 93(1):36–44. 19. Cruz AA, Atique JM, Melo-Filho FV, et al. Orbital involvement in juvenile nasopharyngeal angiofibroma: prevalence and treatment. Ophthal Plast Reconstr Surg 2004; 20(4):296–300.

211

20. Bales C, Kotakpa M, Loevner LA, et al. Craniofacial resection of advanced juvenile nasopharyngeal angiofibroma. Arch Otolaryngol Head Neck Surg 2002; 128(9):1071–1078. 21. Lloyd G, Howard D, Lund VJ, et al. Imaging for juvenile angiofibroma. J Laryngol Otol 2000; 114(9):727–730. 22. Gill G, Rice D, Ritter F. Intracranial and extracranial nasopharyngeal angiofibroma. Arch Otolaryngol 1976; 102:371–377. 23. Davis K, Debrun G. Embolization of juvenile angiofibromas. Semin Interv Radiol 1987; 4:309–320. 24. Economou T, Abemayor E, Ward F. Juvenile nasopharyngeal angiofibroma: an update of the UCLA experience 1960–1985. Laryngoscope 1988; 98:170–175. 25. Roberson G, Price AC, Davis JM, et al. Therapeutic embolization of juvenile angiofibroma. AJR Am J Roentgenol 1979; 133:657–663. 26. Sciarretta V, Pasquini E, Farneti G, et al. Endoscopic sinus surgery for the treatment of vascular tumors. Am J Rhinol 2006; 20(4):426–431. 27. Enepekides DJ. Recent advances in the treatment of juvenile angiofibroma. Curr Opin Otolaryngol Head Neck Surg 2004; 12(6):495–499. 28. Scholtz AW, Appenroth E, Kammen-Jolly K, et al. Juvenile nasopharyngeal angiofibroma: management and therapy. Laryngoscope 2001; 111(4 pt 1):681–687. 29. Gruber A, Balvinski G, Killer M, et al. Preoperative embolization of hypervascular skull base tumors. Minim Invasive Neurosurg 2000; 43(2):62–71. 30. Turowski B, Zanella FE. Interventional neuroradiology of the head and neck. Neuroimaging Clin N Am 2003; 13 (3):619–645. 31. Hurst RW. Interventional neuroradiology of the head and neck. Neuroimaging Clin N Am 1996; 6(2):473–495. 32. Zak F, Lawson W. The Paraganglionic Chemoreceptor System: Physiology, Pathology, and Clinical Medicine. New York: Springer-Verlag, 1982. 33. Pellitteri PK, Rinaldo A, Myssiorek D, et al. Paragangliomas of the head and neck. Oral Oncol 2004; 40(6):563–575. 34. Young WF Jr.. Paragangliomas: clinical overview. Ann N Y Acad Sci 2006; 1073:21–29. 35. Dundee P, Clancy B, Wagstaff S, et al. Paraganglioma: the role of genetic counselling and radiological screening. J Clin Neurosci 2005; 12(4):464–466. 36. Lee JH, Barich F, Karnell LH, et al. National Cancer Data Base report on malignant paragangliomas of the head and neck. Cancer 2002; 94(3):730–737. 37. Fisch U. Infratemporal fossa approach for glomus tumors of the temporal bone. Ann Otol Rhinol Laryngol 1982; 91 (5 pt 1):474–479. 38. Saringer W, Kitz K, Czerny C, et al. Paragangliomas of the temporal bone: results of different treatment modalities in 53 patients. Acta Neurochir (Wien) 2002; 144(12):1255– 1264; discussion 1264. 39. van den Berg R. Imaging and management of head and neck paragangliomas. Eur Radiol 2005; 15(7):1310–1318. 40. Lasjaunias P, Berenstein A. Surgical Neuroangiography, vol 2. Berlin, Heidelberg. New York: Springer-Verlag, 1987:127–148. 41. Gujrathi CS, Donald PJ. Current trends in the diagnosis and management of head and neck paragangliomas. Curr Opin Otolaryngol Head Neck Surg 2005; 13(6):339–342. 42. Boedeker CC, Ridder GJ, Schipper J. Paragangliomas of the head and neck: diagnosis and treatment. Fam Cancer 2005; 4(1):55–59. 43. Myssiorek D. Head and neck paragangliomas: an overview. Otolaryngol Clin North Am 2001; 34(5):829–836. 44. Valavanis A. Preoperative embolization of the head and neck: indications, patient selection, goals, and precautions. AJNR Am J Neuroradiol 1986; 7:943–952.

212

Pryor et al.

45. Persky MS, Setton A, Niimi Y, et al. Combined endovascular and surgical treatment of head and neck paragangliomas—a team approach. Head Neck 2002; 24(5):423–431. 46. Tikkakoski T, Luotonen J, Leinonen S, et al. Preoperative embolization in the management of neck paragangliomas. Laryngoscope 1997; 107(6):821–826. 47. Moret J, Lasjaunias P. Vascular architecture of tympanojugular glomus tumors. In: Vignaud J, ed. The Ear: Diagnostic Imaging. New York, Paris, Barcelona: Masson, 1986:289–303. 48. Lantos PL, Vandenberg SR, Kleihues P. Tumors of the nervous system. In: Graham DI, Lantos PL, eds. Greenfield’s Neuropathology. New York: Oxford University Press, Inc., 1997:583–586. 49. Demaerel P, Wilms G, Lammens M, et al. Intracranial meningiomas: correlation between MR imaging and histology in 50 patients. J Comput Assist Tomogr 1991; 15: 45–51. 50. Russell D, Rubinstein L. Pathology of Tumours of the Nervous System. Baltimore, Hong Kong, London, Sydney: Williams and Wilkins, 1989:449–479. 51. Caroli E, Russillo M, Ferrante L. Intracranial meningiomas in children: report of 27 new cases and critical analysis of 440 cases reported in the literature. J Child Neurol 2006; 21(1):31–36. 52. Perry A, Gutmann DH, Reifenberger G. Molecular pathogenesis of meningiomas. J Neurooncol 2004; 70(2):183–202. 53. Ragel BT, Jensen RL. Molecular genetics of meningiomas. Neurosurg Focus 2005; 19(5):E9. 54. Salvati M, Caroli E, Brogna C, et al. High-dose radiationinduced meningiomas. Report of five cases and critical review of the literature. Tumori 2003; 89(4):443–447. 55. Sekhar LN, Swamy NK, Jaiswal V, et al. Surgical excision of meningiomas involving the clivus: preoperative and intraoperative features as predictors of postoperative functional deterioration. J Neurosurg 1994; 81(6):860–868. 56. Gruber A, Killer M, Mazal P, et al. Preoperative embolization of intracranial meningiomas: a 17-years single center experience. Minim Invasive Neurosurg 2000; 43(1):18–29. 57. Tamiya T, Ono Y, Matsumoto K, et al. Peritumoral brain edema in intracranial meningiomas: effects of radiological and histological factors. Neurosurgery 2001; 49(5):1046– 1051; discussion 1051–1052. 58. Dean BL, Flom RA, Wallace RC, et al. Efficacy of endovascular treatment of meningiomas: evaluation with matched samples. AJNR Am J Neuroradiol 1994; 15 (9):1675–1680. 59. Grand C, Bank WO, Baleriaux D, et al. Gadoliniumenhanced MR in the evaluation of preoperative meningioma embolization. AJNR Am J Neuroradiol 1993; 14 (3):563–569. 60. Wakhloo AK, Juengling FD, Van Velthoven V, et al. Extended preoperative polyvinyl alcohol microembolization of intracranial meningiomas: assessment of two embolization techniques. AJNR Am J Neuroradiol 1993; 14(3):571–582. 61. Vikkula M, Boon LM, Mulliken JB, et al. Molecular basis of vascular anomalies. Trends Cardiovasc Med 1998; 8 (7):281–292. 62. Konez O, Burrows PE. An appropriate diagnostic workup for suspected vascular birthmarks. Cleve Clin J Med 2004; 71(6):505–510.

63. Kohout MP, Hansen M, Pribaz JJ, et al. Arteriovenous malformations of the head and neck: natural history and management. Plast Reconstr Surg 1998; 102(3):643–654. 64. Persky MS, Yoo HJ, Berenstein A. Management of vascular malformations of the mandible and maxilla. Laryngoscope 2003; 113(11):1885–1892. 65. Kakimoto N, Tanimoto K, Nishiyama H, et al. CT and MR imaging features of oral and maxillofacial hemangioma and vascular malformation. Eur J Radiol 2005; 55(1):108–112. 66. Noreau G, Landry PP, Morais D. Arteriovenous malformation of the mandible: review of literature and case history. J Can Dent Assoc 2001; 67(11):646–651. 67. Baker LL, Dillon WP, Hieshima GB, et al. Hemangiomas and vascular malformations of the head and neck: MR characterization. AJNR Am J Neuroradiol 1993; 14(2): 307–314. 68. Smith JK, Castillo M, Wilson JD. MR characteristics of low-flow facial vascular malformations in children and young adults. Clin Imaging 1995; 19(2):109–117. 69. Berenguer B, Burrows PE, Zurakowski D, et al. Sclerotherapy of craniofacial venous malformations: complications and results. Plast Reconstr Surg 1999; 104(1):1–11; discussion 12–15. 70. Peters DA, Courtemanche DJ, Heran MK, et al. Treatment of cystic lymphatic vascular malformations with OK-432 sclerotherapy. Plast Reconstr Surg 2006; 118(6):1441–1446. 71. Lam SM, Williams EF III. Practical considerations in the treatment of capillary vascular malformations, or port wine stains. Facial Plast Surg 2004; 20(1):71–76. 72. Erdmann MW, Jackson JE, Davies DM, et al. Multidisciplinary approach to the management of head and neck arteriovenous malformations. Ann R Coll Surg Engl 1995; 77(1):53–59. 73. Riles TS, Berenstein A, Fisher FS, et al. Reconstruction of the ligated external carotid artery for embolization of cervicofacial arteriovenous malformations. J Vasc Surg 1993; 17(3):491–498. 74. Kotecha B, Fowler S, Harkness P, et al. Management of epistaxis: a national survey. Ann R Coll Surg Engl 1996; 78 (5):444–446. 75. Spafford P, Durham JS. Epistaxis: efficacy of arterial ligation and long-term outcome. J Otolaryngol 1992; 21 (4):252–256. 76. Guttmacher AE, Marchuk DA, White RI Jr. Hereditary hemorrhagic telangiectasia. N Engl J Med 1995; 333 (14):918–924. 77. Jaskolka J, Wu L, Chan RP, et al. Imaging of hereditary hemorrhagic telangiectasia. AJR Am J Roentgenol 2004; 183(2):307–314. 78. Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (RenduOsler-Weber syndrome). Am J Med Genet 2000; 91(1):66–67. 79. Andersen PJ, Kjeldsen AD, Nepper-Rasmussen J. Selective embolization in the treatment of intractable epistaxis. Acta Otolaryngol 2005; 125(3):293–297. 80. Begbie ME, Wallace GM, Shovlin CL. Hereditary haemorrhagic telangiectasia (Osler-Weber-Rendu syndrome): a view from the 21st century. Postgrad Med J 2003; 79 (927):18–24. 81. Duncan IC, Fourie PA, Grange CE, et al. Endovascular treatment of intractable epistaxis—results of a 4-year local audit. S Afr Med J 2004; 94(5):373–378.

11 Dissections of the Carotid and Vertebral Arteries Qaisar A. Shah Department of Neurology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A., and Department of Neurology, University of Minnesota, Minneapolis, Minnesota, U.S.A.

Scott E. Kasner and Robert W. Hurst Department of Neurology; Departments of Radiology, Neurology, and Neurosurgery, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.

INTRODUCTION Arterial dissection occurs when intraluminal blood penetrates the vessel wall, usually through a tear in the intima, and extends through and between the tissue layers of the wall. Dissections occurring in the head and neck may involve either the internal carotid or vertebral arteries and may affect either the extradural or intradural portions of the vessels. Dissections involving the extradural portions of the internal carotid artery (ICA) and vertebral artery (VA) are more common than intradural dissections, with extradural ICA dissection occurring more commonly than that of VA. Conversely, among intradural dissections involvement of the vertebral arteries is considerably more common than is dissections of the ICA or its intracranial branches.

EPIDEMIOLOGY Cervicocerebral arterial dissection is an uncommon cause of stroke in the general population but accounts for 15% to 20% of all strokes in patients under 55 years (1). The true incidence of dissection is difficult to ascertain because some patients may remain asymptomatic or have minor symptoms that are never diagnosed. One community-based study reported the overall incidence of cervicocerebral dissection to be 2.6/100,000/yr (2). Other estimates place the annual incidence of ICA dissection as high as 3.5/100,000/yr and that for VA dissection at 1 to 1.5/100,000/yr (3). Dissections of the extradural ICA accounted for 70% to 80% of all cervical arterial dissections, exceeding the extradural VA dissection by fourfold, according to a prospective study of 200 patients (4). The mean age of presentation for extradural ICA dissection is 44 years and there is no sex predilection, while extradural VA dissection presents at a mean age of 39 years with a female preponderance (5). The vast majority of extradural ICA dissections involve the cervical portion of the ICA with petrous or

cavernous involvement representing less than 5% of extradural ICA dissections. Sixty-five percent of VA dissections involve the suboccipital segment (V3) with nearly 15% of all VA dissections extending into the intradural segment (V4). The proximal VA segment (V1) is the second most common extradural VA location involved. Bilateral VA dissections are also relatively common and have been reported with coexistent ICA dissections (6,7). Intradural arterial dissections are far less common than those involving the extradural portions of the vessels. VA dissection makes up over 80% of intradural dissections with intradural (V4) involvement often representing distal extension of a dissection involving the V3 or suboccipital segment. Intradural VA dissection may also occur in isolation or with extension into the basilar artery. Two-thirds of patients with intradural VA dissection are male and on an average are nine years older than those with intradural ICA dissection (51.8 vs. 43.8 years) (8). Intradural dissection of the anterior circulation represents less than 20% of intradural dissections (9). They may involve individual vessels (ICA, MCA, or ACA) or extend from the intradural ICA into the MCA or ACA (10). Isolated middle cerebral artery (MCA) or anterior cerebral artery (ACA) dissection is rare and usually associated with direct arterial injury or head trauma.

PATHOGENESIS Arterial dissections are often characterized as either spontaneous or traumatic. Traumatic dissections occur secondary to overt head and neck trauma, while spontaneous dissections occur without obvious traumatic injury, although in some cases subclinical trauma may be implicated. For example, dissection has been attributed to seemingly trivial traumatic events, such as nose blowing, head turning while backing an automobile, or prolonged telephone conversations (5).

214

Shah et al.

A number of connective tissue disorders appear to be the risk factors for dissection, perhaps by increasing susceptibility of the artery wall to injury from otherwise subclinical trauma. Conditions associated with an increased incidence of dissection include fibromuscular dysplasia (FMD), Marfan’s syndrome, Ehlers-Danlos syndrome (type IV), osteogenes i s i m p e r f e c t a , c y s t i c m e d i a l n e c r o s i s , an d pseudoxanthoma elasticum (11). Conditions other than connective tissue disorders associated with dissection include recent infection, migraine, and hyperhomocystienemia (12,13). Redundancy and loops of the cervical ICA have also been associated with an increased incidence of ICA dissections (14). With the exception of FMD, potentially predisposing conditions are identified in only a small minority of patients with dissection. Dissection of the carotid or vertebral arteries sets in motion a sequence of events the understanding of which is essential for proper management. Arterial walls are composed of three layers: an internal or endothelial layer (tunica intima), a middle or muscular layer (tunica media), and an external or connective tissue layer (tunica adventitia). Most commonly, a tear or disruption of the intima initiates the damage. Disruption of the intima exposes the subintimal components of the wall, so that platelets adhere to the site of injury, which then serves as a nidus for thrombus formation. The extent of injury to the wall and the hemodynamic features of the injury site determine the subsequent effects of the initial injury. Penetration and extension of blood into the vessel wall results in an intramural hematoma. The hematoma extends between the layers of the vessel wall and, particularly if located between the intimal and medial layers, may constrict the residual lumen with narrowing or complete occlusion. Flow impairment may then cause ischemia as a result of hypoperfusion of neural structures supplied by the damaged vessel. Continued flow through the injured lumen is also potentially detrimental since embolization of thrombus from the injury site may occur. Intramural hematoma may also impair the structural integrity of the wall, resulting in aneurysmal expansion of the vessel. Aneurysm formation is more likely when the intramural hematoma extends between the medial and adventitial layers of the vessel wall. The term ‘‘dissecting aneurysm’’ is applied when the walls of the dilated segment are composed of the incomplete remaining elements of the vessel wall. Hara and Yamamoto note that dissecting aneurysm was originally a pathological term defined as ‘‘a lesion produced by penetration of the circulating blood into the substance of the wall of a vessel, with subsequent extension of the effused blood for varying distances between its layers.’’ In contrast, the term ‘‘fusiform aneurysm’’ refers to the morphology of the aneurysm and makes no reference to its etiology. Consequently, some dissecting aneurysms, particularly those located intradurally, are referred to as fusiform aneurysms on the basis of their morphology

as identified on imaging or angiographic studies (15). Other dissecting aneurysms, particularly those involving the extradural portions of the vessels, may not have fusiform morphology and are connected to the true lumen through a relatively narrow neck. Complete disruption of the arterial wall permits extravasation of blood into adjacent structures. If the hemorrhage is into soft tissue, as is usually the case in extradural locations, a pseudoaneurysm may form. Unlike dissecting aneurysms, the walls of pseudoaneurysms are not composed of layers of the vessel wall. The lumen of a dissecting aneurysms may retain communication with flowing blood in the arterial lumen. The increased diameter of the injured segment exposes the damaged and weakened wall to increased wall tension compared with the more normal vessel. These factors may permit delayed growth of dissecting aneurysms. In addition, slow flow within the dilated segment combined with the absence of intimal lining may permit intra-aneurysmal clot formation and subsequent embolization. The pathophysiology of dissection injury is similar whether extracranial or intracranial locations are involved. It is the environment surrounding the vessel at the site of injury that is often the major determinant of the subsequent course. Extracranially, the internal carotid and vertebral arteries are surrounded primarily by soft tissue that usually gives physical support to the artery wall and limits, but does not always prevent, delayed symptomatic hemorrhage. Consequently, the most common symptoms of extradural dissection are ischemic and arise from hypoperfusion or emboli. Extradural dissection results in symptomatic hemorrhage only rarely and in specific locations, such as within the sphenoid sinus, middle ear, or other cavities of the skull base (16). Intradurally, however, the vessels and their branches course within the subarachnoid space, surrounded only by cerebrospinal fluid. There is no external structural support for the vessel wall and no significant restriction or confinement of hemorrhage if the wall is breached completely. In addition, normal thinning of the media and adventitia as well as defects of the internal elastic lamina characterize the intradural segment of the VA, thus further impairing the structural integrity of the vessel wall (17). Consequently, the chance of aneurysmal enlargement and rupture is significantly increased in intradural as compared with extradural dissections. In addition, hemorrhage within the subarachnoid space gives rise to the well-known sequence of events associated with the very high morbidity and mortality of aneurysmal subarachnoid hemorrhage (SAH). Extradural cervical arteries are more prone to dissection than the intradural segments because of their greater mobility, lack of protection by the skull, and susceptibility to mechanical damage by neighboring bony structures. Sites where a relatively mobile segment of a vessel must traverse a fixed location, such as skull base foramina, can also serve to concentrate mechanical forces. These factors determine the characteristic locations affected by dissection.

Chapter 11: Dissections of the Carotid and Vertebral Arteries

Extradural ICA dissection usually begins several centimeters distal to the relatively fixed common carotid bifurcation from where the ICA originates (12). The dissection then involves the relatively mobile cervical ICA, stopping at the skull base where the artery is secured by the bony walls of the carotid canal. The VA is mobile at both its most proximal (V1) and distal (V3) extradural segments. The vessel’s location is fixed at the origin, within the foraminal segment (V2), and at the site of dural perforation. Extradural VA dissection often involves the V3 or suboccipital segment, the most mobile segment of the vessel; it may also involve the intradural (V4) portion of the VA, often by extension from V3. In contrast, the foraminal or V2 portion of the vessel, because of its lack of mobility and protection by the bony walls of the transverse foramina, is only rarely affected. The preforaminal or V1 segment is involved in approximately 10% of VA dissections. V1 dissections often terminate as the relatively mobile V1 segment enters the fixed transverse vertebral foramen to become V2 (12).

CLINICAL MANIFESTATIONS Clinical features of extradural dissections include headache, neck pain, transient ischemic attack (TIA), ischemic stroke, pulsatile tinnitus, or the less common cranial nerve palsy. Intradural dissection most often presents with either ischemic stroke or SAH.

Extradural ICA Dissection Clinical manifestations of extradural ICA dissection arise either from local effects at the injury site or from distal ischemia. Although nonspecific, headache is the most common symptom reported in at least half of the patients. Silbert et al. reported headache, typically anterior and ipsilateral to the dissection, in 68% of patients with ICA dissection (18). Constant pain of a sharp, aching, or pressing quality is most frequent and typically precedes other neurological signs or symptoms of ICA dissection, with a median interval of approximately four days. Ten percent of patients with ICA dissections have eye, facial, or ear pain without headache (19,20). Ophthalmological manifestations are the second most common group of findings associated with extradural ICA dissection. Ipsilateral Horner’s syndrome, complete or partial, occurs in over half of the patients (21). Ocular ischemic syndromes, such as amaurosis, ischemic optic neuropathy, and central retinal artery occlusion, are less frequent and may result from emboli or distal extension of dissection (22). Extradural ICA dissection may cause ischemia by embolization of thrombus from the site of dissection or by hypoperfusion due to luminal compromise. The former appears to be more common with extracranial ICA dissection. TIAs, hemispheric strokes, or both occur in up to 49% of patients (4). Less frequent signs or symptoms should also be sought in suspected ICA dissection. Pulsatile tinnitus is noted in nearly 25% of cases. Lower CN palsies (CN

215

VII to XII) have been reported in 10% of patients, ocular motility disorders as a result of CN III, IV, or VI dysfunction in 4% of cases, and dysgeusia, likely from involvement of the chorda tympani has been reported in 2% of patients (23,24). A combination of ipsilateral cranial nerve findings and hemispheric deficits may mimic a brain stem stroke and have been referred to as ‘‘false localizing signs’’ (25,26).

Extradural VA Dissection Headache and/or neck pain are the most common clinical features of extradural VA dissection. Headaches posterior and ipsilateral to the dissection affect 83% of patients, while 43% present with neck pain (4,18). Similar to the case with ICA dissection, neurological signs are often delayed, developing a median of 14.5 hours after the onset of headache. The most common neurological symptoms of VA dissection include dizziness, vertigo, double vision, ataxia, and dysarthria. As with ICA dissection, cerebral ischemia may be caused by either thromboembolism or hypoperfusion. Lateral medullary infarction is the most common stroke caused by extradural VA dissection, although cerebellar, basilar tip, or posterior cerebral artery strokes also occur. TIAs are less common with extradural VA dissection than with extradural ICA dissection (4). Extradural VA dissection has been found to be associated with cervical spine fractures, particularly those with a rotational component. While a high index of suspicion should be maintained in patients with this type of injury, there is disagreement as to the clinical significance of vertebral dissection in this setting and, consequently, the need for treatment (27,28).

Intradural ICA Dissection Stroke is far more frequent following intradural ICA dissections than with extradural ICA dissections. In most series, ischemia predominates, although SAH may also result (29–31). Ohkuma et al. reported that among 49 patients with intradural ICA dissection, 63% presented with ischemic symptoms and 36% with SAH (9). However, Chaves et al. reported a series of spontaneous intradural ICA dissections where ischemia developed in 90% of patients while SAH was present in only 10% of cases (32). Ischemia is likely a more common presentation than SAH in isolated dissections of the MCA or ACA, although numbers are small and presentation may vary between neurosurgical and neurological series (33–35). Severe unilateral headache almost always heralds the onset of intradural ICA dissection, usually followed by evidence of ischemia or SAH within minutes to hours. The appearance of neurological signs is therefore more frequent and more rapid than in extradural dissections (29,33,35). Migraine-type headaches have also been associated with intradural dissection but the possibility of SAH must be thoroughly investigated before a diagnosis of migraine can be accepted in this setting (32).

216

Shah et al.

in evaluating extradural VA dissection, which often requires angiographic evaluation. Noninvasive imaging can only suggest the diagnosis of intradural dissection, therefore digital subtraction angiography remains necessary to conclusively evaluate virtually all suspected intradural dissections. Noninvasive imaging modalities are more fully discussed elsewhere in this volume. Angiographic evaluation of suspected ICA dissection begins with injection of the common carotid artery (CCA). Imaging the CCA bifurcation depicts the most proximal extent of ICA dissection since the CCA bifurcation is usually spared in ICA dissection (41). Complete angiographic evaluation of any suspected dissection requires visualization of the intracranial circulation. Features to be sought include intradural extension of an extradural dissection, evidence of emboli, and collateral routes of intracranial supply. Filming into the venous phase demonstrates slow flow or stagnation of contrast within damaged portions of vessels, including aneurysms. The most common angiographic findings in extradural ICA dissection are luminal stenosis and occlusion (Table 1). The irregular stenosis seen in dissection differs in both location and configuration from that caused by atherosclerotic disease. Dissection usually spares the carotid bulb and irregularly narrows the ICA, beginning at 4 to 6 cm distal to the origin and stopping at the skull base where the luminal configuration characteristically returns to normal. Similarly, dissection-related occlusion of the ICA usually spares the proximal portion of the vessel and has a tapered distal extent in the acute phase (Fig. 1). Other angiographic findings, which may be identified in carotid dissection, include an intimal flap, double lumen, intraluminal filling defects, and dissecting aneurysms (42). Angiographic features of conditions that predispose to ICA dissection should also be sought. The most common is FMD, a condition reported in up to 20% of ICA dissections, a higher prevalence than that found in either the general population or in patients with VA dissection (43) (Fig. 2). Angiographic examination of suspected extradural VA dissection may begin with injection of the proximal subclavian artery. After excluding proximal

Hypoperfusion due to arterial narrowing has been suggested to be more prominent than emboli as a mechanism of cerebral ischemia in intradural ICA dissection than in extracranial ICA dissection (32).

Intradural VA Dissection Intradural VA dissection is associated with headache in approximately 55% of patients with other neurological symptoms, including vertigo, tinnitus, nausea, and vomiting occurring less commonly. Cerebrovascular events include SAH or ischemia of the cerebellum and brain stem. As is the case with extradural VA dissections, lateral medullary syndrome is the most frequent ischemic stroke syndrome, occurring in 26% to 43% of patients who develop infarcts (36). However, unlike intradural ICA dissections, the majority of intradural VA dissections are associated with SAH (37). A nationwide study in Japan evaluated 357 patients with intracranial dissection without reported trauma. Over 90% of cases involved the VA. SAH was the presenting symptom in 60% of patients. The remaining patients presented with ischemia or infarction due to stenosis, occlusion, or emboli from the dissection site. In addition, recurrent symptoms were more frequent in patients with SAH (14%) than in patients with no hemorrhage (4.2%) (8). Other series have reported SAH recurrence rates of over 70%, associated with mortality exceeding 50% in the absence of effective treatment of the dissection (38). Correlating with the high rate of SAH, aneurysmal dilatation has been demonstrated in 45% to 76% of intradural VA dissections, more than that observed in intradural ICA dissections (39,40).

ANGIOGRAPHIC DIAGNOSIS OF DISSECTION Dissection may be diagnosed noninvasively using ultrasonography, computed tomography angiography (CTA), magnetic resonance imaging (MRI), and magnetic resonance angiography (MRA). Each has been proposed as a reliable method to identify arterial dissection, but each has its limitations. In most cases of extradural ICA dissection, MRI can provide the pertinent information necessary for diagnosis and medical management. However, MRI is less effective

Table 1 Angiographic Features in Extracranial ICA and VA Dissections (%) Features Reference 112

Artery

Number of arteries

43 43 23

spontaneous ICA ICA VA ICA

78 46 200

42 45

ICA VA

76 26

Normal Stenosis Aneurysm

65

76 5 7 17

41 37 17 47 54

40

Intimal flap

Occlusion

29

17

9 22

1

41 56 72 (occlusion or stenosis > 80%) 29 42

Classification of features varies between reports. Abbreviations: FMD, fibromuscular dysplasia; ICA, internal carotid artery; VA, vertebral artery.

Slow ICA-MCA Branch flow occlusion 24

FMD

11 18 7 23

Figure 1 Angiographic findings in extradural dissections. (A) Lateral CCA angiogram of the neck demonstrates characteristic flameshaped occlusion from dissection beginning several centimeters above the CCA bifurcation. (B) Lateral view of the head [of patient (A)] shows external carotid artery collaterals reconsitituting the intracranial ICA via the ophthalmic artery. (C) Lateral and (D) AP views show characteristic narrowing of ICA lumen beginning above the CCA bifurcation and returning to normal as the vessel enters the petrous bone. (E) Dissection with narrowing and ‘‘tell tale pouch’’ representing mild aneurysmal dilatation. (F) AP and (G) lateral views of ICA dissection with dissecting aneurysm (arrow shows narrowing of residual lumen). (H) Surgical specimen from different case with identical angiographic findings as (F) and (G) demonstrates resected dissecting aneurysm. (I) Extradural VA dissection. Abbreviations: CCA, common carotid artery; ICA, internal carotid artery; VA, vertebral artery.

218

Shah et al.

Figure 2 Angiographic findings in intradural VA dissection. (A) AP and (B) lateral views of left intradural VA dissection with intradural dissecting aneurysm. In a patient with bilateral intradural VA dissection: (C) AP and (D) lateral views of right VA injection with ‘‘string of pearls’’ configuration. (E) AP view of left intradural dissecting aneurysm. (F) CTA demonstrates bilateral dissection. Abbreviation: VA, vertebral artery.

VA damage, selective VA catheterization is performed to further evaluate the vessel. Evaluation of both VAs is important to exclude bilateral dissection, found in 12% to 25% of cases, which is more than that noticed in ICA dissections (43). The most common angiographic feature of extradural VA dissection is also irregular stenosis, either with or without occlusion. Involvement usually centers at the C12 level with intradural extension in up to 15% of cases (44,45). Dissecting aneurysms are less commonly associated with extradural VA dissection than with either extradural ICA dissection or intradural VA dissection. Angiographic findings of intradural VA dissection include segmental narrowing, referred to as the ‘‘string sign’’ or the ‘‘pearl and string sign’’ if narrowed segments alternate with adjacent segments of vessel dilatation. Additional angiographic findings include aneurysms, either fusiform or saccular, double lumen, and tapered narrowing with occlusion of the

vessel. An angiographic pattern of isolated stenosis tends to be associated with an ischemic presentation, while aneurysms, including the ‘‘pearl and string sign’’ are more frequently present with SAH (39) (Fig. 3). A number of authors have reviewed the type and frequency of angiographic findings in intradural VA dissection (Table 2). Terminology varies, however, and firm conclusions are difficult because of the small numbers reported. While aneurysmal dilatation, including the ‘‘pearl and string sign,’’ is commonly identified in intradural VA dissections, stenosis is more common with intradural ICA dissection. Although rare, dissections involving the ICA branches may demonstrate either stenosis or aneurysmal dilatation (46). Suspicion for intradural VA dissection, even without a history of trauma, must remain high when SAH involving the posterior fossa is present without identification of an aneurysm. In such cases, both intradural vertebral arteries must be visualized angiographically to exclude dissection as a cause of the SAH.

Chapter 11: Dissections of the Carotid and Vertebral Arteries

Figure 3 Lateral common carotid angiogram in patient with dissection (arrowhead ) associated with characteristic changes of FMD (arrow). Abbreviation: FMD, fibromuscular dysplasia.

TREATMENT Some dissections are believed to occur without producing any symptoms and therefore may remain completely unrecognized. Consequently, it is possible that some dissections have a benign prognosis in the absence of therapy. Unfortunately, at present there is no reliable method to identify low-risk patients, and observation without therapy cannot be recommended. In the vast majority of dissections, medical treatment represents the first line of therapy.

Medical Treatment of Extradural Dissection Dissections presenting with acute ischemic stroke may be candidates for thrombolysis with intravenous tissue plasminogen activator (IV-tPA), if treated within three hours of symptom onset, unless there are other medical contraindications (47). Patients who develop dissection secondary to severe trauma are usually excluded from receiving IV-tPA, but many patients with dissection have trivial or no trauma, thus IV-tPA is not contraindicated.

219

In a small case series suggesting the use of IVtPA in the acute setting, one patient showed symptomatic intracranial hemorrhage and 36.4% of patients showed excellent recovery with 90-day modified Rankin score (MRS) of 0–1 (48,49). Between three to six hours after symptom onset, intra-arterial (IA)-tPA has been advocated, with efficacy supported by several case series (50–54). Some authors have recommended IA-tPA even within three hours, though others recommend IV-tPA in this context (see below) (49). After the first few hours, antithrombotic therapy is usually recommended, though there has been no clear agreement on the optimal medical management to prevent delayed or recurrent stroke. In the first few days, there appears to be a relatively high risk of ischemic symptoms, and anticoagulation is often recommended despite the lack of controlled trials supporting its use (55,56). A large systematic review of 49 observational studies comprising 683 patients suggested no significant benefit for anticoagulation over antiplatelet therapy with regard to the outcome of death or disability from the initial stroke (57). Nevertheless, treatment with antiplatelet agents was associated with a higher risk of subsequent stroke than treatment with anticoagulation (4.2% vs. 0.9%), arguing in favor of anticoagulation. Heparin or low molecular weight heparin followed by warfarin for three months is usually the mainstay of treatment (58–61). Follow-up imaging studies are recommended after three months of therapy, and if there is normal luminal configuration therapy is often discontinued. If repeat studies show residual stenosis or irregularities, treatment should be continued for three more months with repeat imaging studies (62,63). The use of anticoagulation therapy beyond 12 months is not recommended despite persistent irregularity or stenosis of the lumen, as there is a low risk of stroke beyond the one-year mark (49,60,64). Spontaneous recanalization of ICAs with dissection-related occlusion occurs in 47% to 85% of cases. If warfarin is contraindicated because of systemic trauma or other medical issues, antiplatelet agents can be used. Patients who are treated with medical management have uneventful recovery in 80% to 85% of cases (65,66) (Fig. 4).

Surgical Treatment of Extradural Dissections Surgical options for extradural dissection include carotid ligation, aneurysm resection with carotid reconstruction, and extracranial to intradural ICA bypass (supraclinoid or petrous ICA) (67). However, surgical

Table 2 Angiographic Features in Intracranial VA Dissection (%) Features Reference 113 39 114 115

Number of arteries

String sign

Pearl and string sign

Aneurysm

Double lumen

Occlusion

14 24 41 21

21 58 (or occlusion)

14

7 42

29

21

32 5

33

5

Classification of features varies between reports.

68 38

14

220

Shah et al.

Figure 4 A 39-year-old female presented with left hemiparesis two days following motor vehicle accident. (A) T2 axial image shows deep watershed infarct on the right. (B, C) Axial T1-weighted MRI shows hyperintense mural hemorrhage involving the right ICA (arrow). (D) AP right CCA angiogram shows irregular cervical ICA narrowing characteristic of dissection. Abbreviations: ICA, internal carotid artery; CCA, common carotid artery.

morbidity may be high. Perioperative stroke rates may occur in 10% of cases, peripheral cranial nerve injury occurs in more than half, and mortality in 2% of patients. Recent studies suggest that extradural dissecting aneurysms, if asymptomatic, generally do not warrant surgical intervention, as they tend to resolve spontaneously or at least remain stable. However, symptomatic dissecting aneurysms may be resected followed by reconstruction of the ICA with saphenous vein graft or primary reanastomosis (68–70). Increasing experience and evidence supports the conclusion that in many cases in which medical treatment is not appropriate or is unsuccessful, endovascular procedures may be preferable to a surgical approach for treatment of dissection-related injury. Nevertheless, additional study is needed to more rigorously document the indications and optimal management of these patients.

Endovascular Treatment of Extradural Dissections While medical therapy is currently the mainstay and initial management in most cases of extradural carotid or VA dissections, recognition of the role of endo-

vascular treatment and its timely application have become increasingly important. As in all neuroendovascular techniques, recognition of proper indications for intervention is an essential and a meticulous technique to ensure patient benefit (71). In the acute period following extradural ICA or VA dissection, symptoms arise most often from intracranial emboli and may require emergent thrombolysis. In cases where IV thrombolytic treatment is ineffective or contraindicated, intra-arterial thrombolysis plays a role as noted above. Less frequently, acute symptomatic arterial stenosis or occlusion may require revascularization of the dissected artery using angioplasty and stenting. The need for endovascular treatment most often occurs with symptomatic extradural ICA dissections (Fig. 5). Nevertheless, extradural vertebral dissections with stenosis or occlusion may also require emergent treatment when VA asymmetry or intracranial emboli are present. A number of investigators have confirmed not only the dismal outcome of untreated acute vertebrobasilar embolic occlusion, but also the potential for significant benefit if thrombolysis and reopening is accomplished prior to irreversible infarction (72,73).

Chapter 11: Dissections of the Carotid and Vertebral Arteries

221

Figure 5 Simplified schematic illustration of the pathophysiological process of carotid artery dissection proceeding from the acute stage to either spontaneous healing (1), formation of false lumen (2), residual stenosis of varying degree or complete occlusion (3), and formation of a pseudoaneurysm (4). A stent is used in cases not responding to medical therapy either to relieve a hemodynamically significant stenosis, to occlude a false lumen, or to serve as a scaffold to enable coil embolization of a wide-necked pseudoaneurysm. Source: From Ref. 75.

In general, indications for endovascular treatment of extradural dissections include 1. 2. 3. 4. 5. 6.

patients with ischemia in whom IV thrombolysis for ischemia is contraindicated because of systemic hemorrhage, recent surgery, or trauma; patients with ischemia and contralateral ICA stenosis or occlusion; patients in whom there is a need for elective occlusion of the contralateral ICA or VA for other pathology; patients in whom there is a need to avoid flow increase through the anterior communicating artery because of an associated aneurysm; when intradural extension of dissection occurs with consequent risk of SAH (74,75); and for treatment of dissecting aneurysms under specific circumstances (see below).

Intra-arterial thrombolysis may be needed acutely when dissections result in symptomatic intracranial emboli or arterial narrowing with superimposed clot causing impairment of flow. In such cases, thorough investigation as to the etiology of cerebral ischemia is necessary to identify dissection, if

present. Identification of salvageable tissue on neuroimaging studies is becoming more widely available and can assist in the selection of the most appropriate patients for acute treatment. Intra-arterial thrombolysis is performed using the techniques for acute stroke (see chap. 16). Particular attention must, however, be directed to determine the presence and extent of dissection-related injury and consider specific treatment of the dissected vessel, if necessary. As noted, small series and case reports support the use of intra-arterial thrombolysis in dissection, particularly outside the conventional time constraints of IV thrombolysis or in specific cases in which IV thrombolysis is contraindicated (49–54). In cases of symptomatic stenosis as a result of dissection, angioplasty using a stent can be used to exclude a false lumen, relieve hemodynamically significant stenosis, and restore the true lumen to more normal size, thereby increasing flow (76,77). The technique may also be successful when the dissected vessel is completely occluded. In cases of complete occlusion, however, the potential for distal embolization on reopening the vessel may be substantial, depending on the clot burden within the occluded

222

Shah et al.

segment of the vessel. Careful consideration must be given to relative risks and benefits of reopening an occluded vessel. The technique of stent angioplasty begins with angiographic confirmation of the location and the extent of the dissection-related stenosis. A microcatheter and microguidewire (0.014–0.018 inch) are then maneuvered through the true arterial lumen using road mapping angiography. The microcatheter is advanced to a position distal to the damaged segment. An exchange length (260–300 cm) microguidewire is used to exchange the microcatheter for the stent catheter. The stent is then deployed within the stenotic segment of the vessel. Because of the relatively low levels of radial force needed for restoration of lumen diameter in dissected vessels, primary angioplasty is not normally necessary and may place excessive stress on the already damaged arterial wall. The stent maintains sufficient radial force on the damaged segment of the artery wall, placing the layers separated by the intramural hematoma in contact with one another. The result is obliteration of the false lumen, restoration of the normal luminal diameter, and resolution of the stenosis. The stent length should be chosen to cross the entire damaged segment of the vessel when possible. In some cases of long segment dissection, multiple overlapping stents may be required. In such cases, the initial stent is usually placed at the proximal margin of the dissection to eliminate the inflow zone of the false lumen. Following initial stent placement, additional angiography is done to confirm the need for additional stents. Both self-expanding and balloon-

mounted stents have been found to be suitable for this application (75). Following stent placement, patients are maintained on an appropriate antiplatelet regimen to prevent stent thrombosis. In cases of dissection-associated aneurysm, stent placement has also been found useful, both alone as well as to provide a scaffold to permit coil embolization (see below). The dynamic nature of carotid and VA dissections and their ability to change over very short periods of time has been emphasized (12). Consequently, the need for endovascular treatment of extradural internal carotid or vertebral dissections may also develop after the acute phase of the injury (78). Because medical treatment of extradural dissections is usually quite effective, situations which merit delayed intervention, while uncommon, usually represent failure of medical therapy (79). They are usually manifested by new onset, fluctuation, or recurrent neurological dysfunction of the vascular distribution of the damaged artery (80). As is the case in the acute phase, delayed ischemic symptoms usually result either from recurrent emboli or development of symptomatic stenosis with poor collateral circulation (Fig. 6). In either situation, persistent abnormality within the damaged segment of the artery is usually identified. Specific features include failure of the dissected segment to heal with the development of luminal stenosis, or persistent clot formation despite medical treatment. In either case, angiographic evaluation must determine whether intracranial emboli are present, the status of the dissected vessel, and collateral routes to the affected vascular distribution. Any areas

Figure 6 A 24-year-old male after gunshot wound to left neck. (A) Initial lateral and (B) AP angiograms demonstrate irregularity (arrow) consistent with LICA dissection. Six months later, while on anticoagulation therapy, patient presented with aphasia lasting four hours followed by complete resolution. CT scan (C, D) shows hyperdensity within left MCA branches (arrow) representing emboli. (E) Angiogram demonstrated interval growth of dissecting aneurysm. (F) Patient underwent carotid occlusion (lateral angiogram). Abbreviations: LICA, left internal carotid artery; MCA, middle cerebral artery.

Chapter 11: Dissections of the Carotid and Vertebral Arteries

of damage not present or recognized at the time of the initial evaluation must also be sought and identified. When intracranial embolic occlusion occurs after the acute phase of dissection, thrombolysis, either intravenous or intra-arterial, may be necessary as an initial step. Should symptoms develop in a setting of appropriate medical therapy, additional endovascular treatment, such as stent angioplasty, may be required to minimize the chance of recurrence. Extradural Dissection-Associated Aneurysms

As noted earlier, controversy often surrounds the terminology applied to aneurysms associated with arterial dissection. While the underlying pathology is similar regardless of the morphology, the morphological features of a dissecting aneurysm have major implications for endovascular techniques that may be useful for its treatment. Extradural aneurysms associated with dissection represent a radiological finding that often engenders considerable concern and uncertainty as to management. While extradural ICA aneurysms may arise from other causes, including atherosclerosis and infection, a significant percentage results from prior dissection. Older series have recommended nearly universal treatment of extradural carotid artery aneurysms to prevent neurological deterioration (81). Several recent studies, however, have found that extradural dissecting aneurysms of either the ICA or VA rarely enlarge over time. These studies have also concluded that the vast majority of such aneurysms remain quiescent and asymptomatic in the face of appropriate medical management and, in most cases, require no additional treatment (82,83). Nevertheless, documented examples confirm that dissection-associated aneurysms can, under certain circumstances, cause neurological deterioration as a result of embolization or expansion with compression of cranial nerves (84–87). In addition, when located within or adjacent to the skull base, these aneurysms may pose a significant risk of potentially fatal bleeding (16). In general, dissection-associated extradural aneurysms should be seriously considered for treatment when 1. 2. 3.

they are found to be enlarging or causing symptoms related to mass effect, their location exposes the patient to risk of hemorrhage into an adjacent sinus or skull base cavity such as the middle ear, and symptoms occur, which are attributable to emboli from an extradural aneurysm in a patient on appropriate medical management.

Therapy for dissecting extradural aneurysms is usually feasible using endovascular techniques. Often the simplest, safest, and most effective option may be permanent occlusion of the involved carotid or VA. Feasibility of vessel sacrifice is of course dependent on collateral flow from adjacent circulations. Prior to sacrifice of the vessel, an occlusion test is usually performed as outlined elsewhere (see chap. 9).

223

A number of techniques have been described to treat extradural aneurysms with preservation of the involved vessel. Morphology of the aneurysm is the primary feature that dictates the endovascular technique that may be successful in treating the lesion. In cases of extradural aneurysms with relatively narrow necks, treatment has been accomplished using coil embolization alone. More recently, stent-assisted coiling has been recommended (88). Similar to the treatment of wide-necked or fusiform intracranial aneurysms, a stent is placed across the dissected segment and neck of the aneurysm. The stent provides support to the dissected vessel wall and also prevents herniation of coils into the parent vessel. Other authors have indicated that a significant percentage of dissection-related aneurysms will resolve after endovascular stent placement alone without the necessity for additional coil embolization (89,90). The presence of the stent across the aneurysm neck likely impairs inflow and promotes intra-aneurysmal clot formation, with subsequent thrombosis and closure of the aneurysm (Fig. 7). Recently, placement of covered stent grafts has been reported to be successful in obliterating dissection-associated aneurysms in a small number of cases (91–93). The technique has shown promise at moderately long-term follow-up and may be especially useful in cases of patients symptomatic of mass effect (94). Despite the frequent success of endovascular treatment, a role for surgical treatment remains for some cases of extradural aneurysms of the ICA. This treatment is ideally accomplished for lesions located proximally in the neck, and is adopted most often as a result of difficulty with placement of endovascular devices across the lesion. Difficulty in crossing the lesion may be due to associated arterial disease such as FMD or excessive tortuosity of the vessel. The latter feature is encountered with some frequency and has in fact been noted as a potentially predisposing factor to carotid dissection (14). The use of more flexible stents has, however, been reported to make even relatively tortuous vessels amenable to endovascular treatment (95). Lastly, aggressive evaluation and consideration of endovascular or surgical treatment should be entertained in situations in which intradural extension of extradural dissection is suspected. This situation most often affects the region of dural penetration of the vertebral arteries (44). As discussed below, intradural dissection, whether primary or by extension from an extradural injury, may represent a significant risk of morbidity and mortality as a result of either ischemia or SAH.

Medical Treatment of Intradural Dissection Treatment of intradural dissection is dictated by the initial clinical event. Patients who present with ischemic stroke are usually treated with antiplatelet or anticoagulant therapy. Nevertheless, because systemic medical therapy may present excessive risks with intradural dissections, careful diagnosis prior to medical treatment is necessary. As with extracranial

224

Shah et al.

Endovascular Treatment of Intradural Dissections

Figure 7 Acute neck pain following motor vehicle accident. Initial MRI (not shown) demonstrated wall hematoma. Two months after accident, left Horner’s syndrome was noted after transient right arm weakness. (A) Lateral and (B) AP views of LICA dissecting aneurysm. (C) Unsubtracted view of stent in place. (D) Six-month follow-up with resolution of aneurysm. No symptom recurrence. Abbreviation: LICA, left internal carotid artery.

dissections, surgical or endovascular approaches may be considered in patients whose symptoms recur despite medical therapy. Patients with intradural dissection who present with SAH usually require aggressive treatment with either surgical or endovascular repair, since conservative management may result in poor outcome (38,96).

Surgical Treatment of Intradural Dissection Surgical procedures utilized for treatment of intradural dissection have included proximal occlusion of the parent artery, trapping of the lesion, vascular reconstruction, surgical wrapping, or clipping of the aneurysm. Because the vast majority of intradural dissecting aneurysms are fusiform, the role of clipping, the standard surgical technique for addressing saccular aneurysms, is limited (96).

Because of the propensity for vessel rupture and recurrent hemorrhage, intradural dissections require anatomic correction by either open surgical or endovascular techniques at a higher rate than is the case with extradural dissections in which the initial therapy is most often medical. Nevertheless, indications for treatment as well as alternatives remain controversial. Conservative treatment or medical therapy has most often been advocated in cases without aneurysm or evidence of hemorrhage that present with ischemia. In patients presenting with SAH, poor outcome with conservative management has been emphasized and a number of surgical and endovascular alternatives have been reported (38,96). Endovascular treatment has assumed a major role in the management of intradural dissecting aneurysms. Permanent endovascular occlusion using coils has been shown to be a useful therapeutic endovascular technique for the treatment of fusiform and acute intradural dissecting aneurysms of the vertebrobasilar system (97–99). Techniques vary, using combinations of detachable balloons and coils, and are often preceded by test occlusion of the involved vessel to determine whether endovascular occlusion can be safely accomplished. Leibowitz et al. reported long-term outcomes for unilateral intradural VA aneurysms treated by permanent occlusion. They found better clinical outcomes than the patients whose aneurysms involved the basilar artery or both vertebral arteries, where complete thrombosis cannot be achieved by using permanent vessel occlusion (100). The authors also reviewed prior series supporting the usefulness of this relatively simple endovascular occlusion technique. Other investigators have emphasized endovascular trapping of the diseased segment proximally and distally to ensure closure of the dissection site by the coil mass, thereby preventing regrowth or rehemorrhage (Fig. 8) (101). Endovascular occlusion of the intradural VA has also been found useful in cases in which dissection involves isolated vertebrobasilar branches or extends more distally into the basilar artery (Fig. 9) (102). Nevertheless, obvious limitations of the technique exist in more extensive dissections where vessel preservation is essential (103). The appearance of newer endovascular techniques with the potential to preserve vessel patency has not excluded permanent vessel occlusion from the endovascular therapeutic armamentarium. Reports of reformation of aneurysms following stent supported embolization designed to preserve arterial patency have led to suggestions that parent vessel occlusion remains the first option for treatment in patients who will tolerate sacrifice of the parent vessel along with its diseased segment (104). Nevertheless, techniques that promise to treat intradural dissections while preserving vessel patency are of increasing clinical importance and interest. Advantages include maintaining maximum intracranial flow, a particular advantage in older patients who might have coexisting vascular disease, but also

Chapter 11: Dissections of the Carotid and Vertebral Arteries

225

Figure 8 A 36-year-old female with acute onset of headache, no SAH. (A) AP view of left VA angiogram at presentation demonstrates mild irregular fusiform dissection of intradural VA (arrow). (B–D) MRI at time of presentation shows minimal enlargement of intradural left VA (arrow). (E–G) CT scan 10 years after initial presentation, when patient developed progressive left hemiparesis, shows hyperdense mass in the region of previous abnormality. (H) Angiography of right VA (J, K) demonstrates growth of thrombus filled fusiform aneurysm. (I) AP plain film following coil embolization with packing of aneurysmal segment and left VA occlusion. Abbreviations: SAH, subarachnoid hemorrhage; VA, vertebral artery.

potentially beneficial in younger patients who would be expected to live longer with the results of treatment. In addition, vessel-preserving treatment extends the advantages of endovascular therapy to patients who will not tolerate vessel occlusion including those with more extensive lesions. The usually fusiform morphology and structurally incompetent wall of intradural dissecting aneurysms eliminates any major role for selective aneurysm embolization using coils alone. The use of stents, either alone or more often followed by coil placement through the interstices of the stent, has emerged as a significant advance in managing intradural dissec-

tions while preserving the affected vessel (105). The technique is identical to that utilized in stentsupported coil embolization of saccular aneurysms. It has shown high rates of success in the limited number of reported cases with low incidence of delayed vessel occlusion or ischemic stroke (Fig. 10) (88). Stent-supported coiling of intradural dissecting aneurysms extends endovascular options to portions of vessels where sacrifice of a dissected segment is not possible without neurological deficit. This observation is particularly true when intradural vertebral dissections extend to involve the basilar artery (106). The usefulness of stenting has been demonstrated in

226

Shah et al.

Figure 9 (A) Lateral right VA angiogram showing dissection of right PICA (arrow) in patient who presented with SAH. (B) Unsubtracted and subtracted (C) images following coil occlusion of right PICA. Abbreviations: VA, vertebral artery; PICA, posterior inferior cerebellar artery; SAH, subarachnoid hemorrhage.

Figure 10 Enlarged intradural VA-dissecting aneurysm in a patient who had already undergone left VA occlusion for an enlarged intradural VA aneurysm ( , coil mass within occluded left VA–dissecting aneurysm). (A) Subtracted AP view after placement of two overlapping stents across the aneurysm neck (arrowheads, markers at the ends of the overlapping stents). (B) Microcatheter (arrows) crosses the interstices of the stent to deploy coils within the aneurysm. (C) Aneurysm coiled, follow-up confirmed aneurysm occlusion with normal flow through the parent vessel.

both aneurysmal and occlusive basilar dissection (107). In addition, an increasing number of reports document the effectiveness of the technique in intradural dissections of the anterior circulation (108). Recent reports indicate that intradural fusiform aneurysm treatment may also be accomplished using covered stent grafts (94,109). Long-term outcome and specific indications for this technique await additional experience.

PROGNOSIS AND OUTCOME Most patients with stroke due to extradural ICA or VA dissection sustain relatively mild deficits with ultimate resolution of their symptoms, though a significant minority (5–10%) suffer disabling stroke. The recurrence rate of thromboembolic episodes after dissection is 0.6% to 10.4% (79). There may be a tendency

for VA dissections to cause more severe strokes than ICA dissections. Traumatic dissections appear to have a worse prognosis than spontaneous dissections in terms of persistent neurological symptoms (110). Patients with intradural dissection have poor outcome as compared to those with extracranial dissection. Intradural dissection associated with dissecting aneurysm and associated SAH carries high risk of morbidity and mortality (20–50%), exposes the patient to the risk of recurrent hemorrhage, and thus requires urgent medical, endovascular, or surgical intervention as discussed earlier (99). In general, the rate of recurrent arterial dissection is low. Recurrence in the same vessel is rare, though may possibly occur as a result of a vascular defect created by scar tissue formation. The recurrence rate may also be higher in patients with a family history of arterial dissection (111).

Chapter 11: Dissections of the Carotid and Vertebral Arteries

Patients who have had cervicocerebral arterial dissections should be advised to avoid activities that may cause sudden rotation or extension of the neck. However, no substantial data exist to define the limits of activity for these patients. There is no apparent reason to manage physical therapy differently during rehabilitation following stroke resulting from dissection. Dissection of the carotid and vertebral arteries represents a significant etiology of stroke in young patients. The understanding, recognition, and diagnosis of this disorder have rapidly advanced in recent years and the development of endovascular techniques has made a major contribution to patients for whom medical treatment is not suitable.

18.

19.

20.

21.

22.

23.

REFERENCES 1. Anson J, Crowell RM. Cervicocranial arterial dissection. Neurosurgery 1991; 29(1):89–96. 2. Schievink WI, Mokri B, Whisnant JP. Internal carotid artery dissection in a community. Rochester, Minnesota, 1987–1992. Stroke 1993; 24(11):1678–1680. 3. Schievink WI, Mokri B, Piepgras DG. Spontaneous dissections of cervicocephalic arteries in childhood and adolescence. Neurology 1994; 44(9):1607–1612. 4. Thanvi B, Munshi SK, Dawson SL, et al. Carotid and vertebral artery dissection syndromes. Postgrad Med J 2005; 81(956):383–388. 5. Caplan LR, Biousse V. Cervicocranial arterial dissections. J Neuroophthalmol 2004; 24(4):299–305. 6. Friedman DP, Flanders AE. Unusual dissection of the proximal vertebral artery: description of three cases. AJNR Am J Neuroradiol 1992; 13(1):283–286. 7. Houser OW, Mokri B, Sundt TM Jr., et al. Spontaneous cervical cephalic arterial dissection and its residuum: angiographic spectrum. AJNR Am J Neuroradiol 1984. 5(1):27–34. 8. Yamaura A, Ono J, Hirai S. Clinical picture of intracranial non-traumatic dissecting aneurysm. Neuropathology 2000; 20(1):85–90. 9. Ohkuma H, Suzuki S, Ogane K. Dissecting aneurysms of intracranial carotid circulation. Stroke 2002; 33(4):941–947. 10. Manz HJ, Vester J, Lavenstein B. Dissecting aneurysm of cerebral arteries in childhood and adolescence. Case report and literature review of 20 cases. Virchows Arch A Pathol Anat Histol 1979; 384(3):325–335. 11. Hademenos GJ, Alberts MJ, Awad I, et al. Advances in the genetics of cerebrovascular disease and stroke. Neurology 2001; 56(8):997–1008. 12. Schievink WI. Spontaneous dissection of the carotid and vertebral arteries. N Engl J Med 2001; 344(12):898–906. 13. Rubinstein SM, Peerdeman SM, van Tulder MW, et al. A systematic review of the risk factors for cervical artery dissection. Stroke 2005; 36(7):1575–1580. 14. Barbour PJ, Castaldo JE, Rae Grant AD, et al. Internal carotid artery redundancy is significantly associated with dissection. Stroke 1994; 25(6):1201–1206. 15. Hara M, Yamamoto I. Introduction to symposium: intracranial dissecting aneurysms. Neuropathology 2000. 20(1):83–84. 16. Moonis G, Hwang CJ, Ahmed T, et al. Otologic manifestations of petrous carotid aneurysms. AJNR Am J Neuroradiol 2005; 26(6):1324–1327. 17. Sato T, Sasaki T, Suzuki K, et al. Histological study of the normal vertebral artery—etiology of dissecting

24.

25.

26.

27. 28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

38.

227

aneurysms. Neurol Med Chir (Tokyo) 2004; 44(12):629–635; discussion 636. Silbert PL, Mokri B, Schievink WI. Headache and neck pain in spontaneous internal carotid and vertebral artery dissections. Neurology 1995; 45(8):1517–1522. Biousse V, Woimant F, Amaranco P, et al. Headache in 67 patients with extracranial internal carotid artery dissection. Cephalalgia 1991; 11:232–233. Biousse V, D’Anglejan-Chatillon J, Massiou H, et al. Head pain in non-traumatic carotid artery dissection: a series of 65 patients. Cephalalgia 1994; 14(1):33–36. Biousse V, Touboul PJ, D’Anglejan-Chatillon J, et al. Ophthalmologic manifestations of internal carotid artery dissection. Am J Ophthalmol 1998; 126(4):565–577. Guillon B, Biousse V, Massiou H, et al. Orbital pain as an isolated sign of internal carotid artery dissection. A diagnostic pitfall. Cephalalgia 1998; 18(4):222–224. Baumgartner RW, Arnold M, Baumgartner I, et al. Carotid dissection with and without ischemic events: local symptoms and cerebral artery findings. Neurology 2001. 57(5):827–832. Knibb J, Lenthall R, Bajaj N. Internal carotid artery dissection presenting with ipsilateral tenth and twelfth nerve palsies and apparent mass lesion on MRI. Br J Radiol 2005; 78(931):659–661. Wessels T, Rottger C, Kaps M, et al. Upper cranial nerve palsy resulting from spontaneous carotid dissection. J Neurol 2005; 252(4):453–456. Mokri B, Silbert PL, Schievink WI, et al. Cranial nerve palsy in spontaneous dissection of the extracranial internal carotid artery. Neurology 1996; 46(2):356–359. Parbhoo AH, Govender S, Corr P. Vertebral artery injury in cervical spine trauma. Injury 2001; 32(7):565–568. Schellinger PD, Schwab S, Krieger D, et al. Masking of vertebral artery dissection by severe trauma to the cervical spine. Spine 2001; 26(3):314–319. Guridi J, Gallego J, Monzon F, et al. Intracerebral hemorrhage caused by transmural dissection of the anterior cerebral artery. Stroke 1993; 24(9):1400–1402. Wisoff HS, Rothballer AB. Cerebral arterial thrombosis in children. Review of literature and addition of two cases in apparently healthy children. Arch Neurol 1961; 4:258–267. Wolman L. Cerebral dissecting aneurysms. Brain 1959; 82:276–291. Chaves C, Estol C, Esnaola MM, et al. Spontaneous intracranial internal carotid artery dissection: report of 10 patients. Arch Neurol 2002; 59(6):977–981. Adams HP Jr., Aschenbrener CA, Kassell NF, et al. Intracranial hemorrhage produced by spontaneous dissecting intracranial aneurysm. Arch Neurol 1982; 39(12):773–776. Kurino M, Yoshioka S, Ushio Y. Spontaneous dissecting aneurysms of anterior and middle cerebral artery associated with brain infarction: a case report and review of the literature. Surg Neurol 2002; 57(6):428–436. Nagumo K, Nakamori A, Kojima S. [Spontaneous intracranial internal carotid artery dissection: 6 case reports and a review of 39 cases in the literature]. Rinsho Shinkeigaku 2003; 43(6):313–321. Sturzenegger M. Headache and neck pain: the warning symptoms of vertebral artery dissection. Headache 1994; 34(4):187–193. Caplan LR, Baquis GD, Pessin MS, et al. Dissection of the intracranial vertebral artery. Neurology 1988; 38 (6):868–877. Mizutani T, Aruga T, Kirino T, et al. Recurrent subarachnoid hemorrhage from untreated ruptured vertebrobasilar dissecting aneurysms. Neurosurgery 1995; 36(5):905– 911; discussion 912–913.

228

Shah et al.

39. Shin JH, Suh DC, Choi CG, et al. Vertebral artery dissection: spectrum of imaging findings with emphasis on angiography and correlation with clinical presentation. Radiographics 2000; 20(6):1687–1696. 40. Yamaura A, Watanabe Y, Saeki N. Dissecting aneurysms of the intracranial vertebral artery. J Neurosurg 1990; 72(2):183–188. 41. Hirth K, Sander S, Hormann K. Common carotid artery dissection: a rare cause for cervical pain. J Laryngol Otol 2002; 116(4):309–311. 42. Pelkonen O, Tikkakoski T, Leinonen S, et al. Extracranial internal carotid and vertebral artery dissections: angiographic spectrum, course and prognosis. Neuroradiology 2003; 45(2):71–77. 43. Dziewas R, Konrad C, Drager B, et al. Cervical artery dissection—clinical features, risk factors, therapy and outcome in 126 patients. J Neurol 2003; 250(10):1179–1184. 44. Komiyama M, Ishiquro T, Matsusaka Y, et al. Simultaneous dissection of intra- and extracranial vertebral artery. Report of two cases and review of literature. Acta Neurochir (Wien) 2002; 144(7):729–733. 45. Saeed AB, Shuaib A, Al-Sulaiti G, et al. Vertebral artery dissection: warning symptoms, clinical features and prognosis in 26 patients. Can J Neurol Sci 2000; 27(4):292–296. 46. Lin CH, Jeng JS, Yip PK. Middle cerebral artery dissections: differences between isolated and extended dissections of internal carotid artery. J Neurol Sci 2005; 235(1–2):37–44. 47. The National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med 1995; 333(24):1581–1587. 48. Derex L, Nighoghossian N, Tuejman F, et al. Intravenous tPA in acute ischemic stroke related to internal carotid artery dissection. Neurology 2000; 54(11):2159–2161. 49. Saver JL, Easton JD. Dissections and trauma of cervicocerebral arteries. In: Barnett HJM, Mohr JP, Stein BM, Yatsu FM , eds. Stroke: Pathophysiology, Diagnosis, and Management. New York: Churchill Livingstone, 1998:769–786. 50. Furlan A, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study: a randomized controlled trial. Prolyse in acute cerebral thromboembolism. JAMA 1999; 282(21):2003–2011. 51. Gomez CR, May AK, Terry JB, et al. Endovascular therapy of traumatic injuries of the extracranial cerebral arteries. Crit Care Clin 1999; 15(4):789–809. 52. Molina CA, Alvarez-Sabin J, Schonewille W, et al. Cerebral microembolism in acute spontaneous internal carotid artery dissection. Neurology 2000; 55(11):1738–1740. 53. Price RF, Sellar R, Leung C, et al. Traumatic vertebral arterial dissection and vertebrobasilar arterial thrombosis successfully treated with endovascular thrombolysis and stenting. AJNR Am J Neuroradiol 1998; 19(9):1677–1680. 54. Sampognaro G, Turgut T, Conners JJ III, et al. Intraarterial thrombolysis in a patient presenting with an ischemic stroke due to spontaneous internal carotid artery dissection. Catheter Cardiovasc Interv 1999; 48(3):312–315. 55. Leys D, Lucas C, Gobert M, et al. Cervical artery dissections. Eur Neurol 1997; 37(1):3–12. 56. Sturzenegger M. Spontaneous internal carotid artery dissection: early diagnosis and management in 44 patients. J Neurol 1995; 242:231–238. 57. Lyrer P, Engelter S. Antithrombotic drugs for carotid artery dissection. Cochrane database of systematic reviews [electronic resource] 2000(4):CD000255. http:// www.cochrane.org/reviews/en/ab000255.html 58. Biousse V, D’Anglejan-Chatillon J, Toubol PJ, et al. Time course of symptoms in extracranial carotid artery dissections. A series of 80 patients. Stroke 1995; 26(2):235–239.

59. Hart RG, Easton JD. Dissections of cervical and cerebral arteries. Neurol Clin N Am 1983; 1:155–182. 60. Kasner SE, Hankins LL, Bratina P, et al. Magnetic resonance angiography demonstrates vascular healing of carotid and vertebral artery dissections. Stroke 1997; 28:1993–1997. 61. Lucas C, Moulin T, Deplanque D, et al. Stroke patterns of internal carotid artery dissection in 40 patients. Stroke 1998; 29(12):2646–2648. 62. Bogousslavsky J, Despland PA, Regli F. Spontaneous carotid dissection with acute stroke. Arch Neurol 1987; 44(2):137–140. 63. Mokri B, Sundt TM Jr., Houser OW, et al. Spontaneous dissection of the cervical internal carotid artery. Ann Neurol 1986; 19:126–138. 64. Jacobs A, Lanfermann H, Neveling M, et al. MRI- and MRA-guided therapy of carotid and vertebral artery dissections. J Neurol Sci 1997; 147(1):27–34. 65. Beletsky V, Nadareishvili Z, Lynch J, et al. Cervical arterial dissection: time for a therapeutic trial?. Stroke 2003; 34(12):2856–2860. 66. Gonzales-Portillo F, Bruno A, Biller J. Outcome of extracranial cervicocephalic arterial dissections: a follow-up study. Neurol Res 2002; 24(4):395–398. 67. Sandmann W. Surgery of the cervical vascular system. In: Heberhold C, Panje WR , ed. Head and Neck Surgery, vol 3. 2nd ed. Stuttgart: Georg Thieme Verlag, 1998:51–67. 68. Candon E, Marty-Ane C, Pieuchot P, et al. Cervical-topetrous internal carotid artery saphenous vein in situ bypass for the treatment of a high cervical dissecting aneurysm: technical case report. Neurosurgery 1996; 39(4):863–866. 69. Morgan MK, Sekhon LH. Extracranialintracranial saphenous vein bypass for carotid or vertebral artery dissections: a report of six cases. J Neurosurg 1994; 80(2):237–246. 70. Schievink WI, Mokri WI, O’Fallon WM. Recurrent spontaneous cervical-artery dissection. N Engl J Med 1994; 330 (6):393–397. 71. Cothren CC, Moore EE, Ray CE Jr., et al. Carotid artery stents for blunt cerebrovascular injury: risks exceed benefits. Arch Surg 2005; 140(5):480–485; discussion 485–486. 72. Arnold M, Nedelchev K, Schroth G, et al. Clinical and radiological predictors of recanalisation and outcome of 40 patients with acute basilar artery occlusion treated with intra-arterial thrombolysis. J Neurol Neurosurg Psychiatry 2004; 75(6):857–862. 73. Brandt T, von Kummer R, Muller-Kuppers M, et al. Thrombolytic therapy of acute basilar artery occlusion. Variables affecting recanalization and outcome. Stroke 1996; 27(5):875–881. 74. Biondi A, Katz JM, Vallabh J, et al. Progressive symptomatic carotid dissection treated with multiple stents. Stroke 2005; 36(9):e80–e82. 75. Malek AM, Higashida RT, Phatouros CC, et al. Endovascular management of extracranial carotid artery dissection achieved using stent angioplasty. AJNR Am J Neuroradiol 2000; 21(7):1280–1292. 76. Cohen JE, Gomori JM, Umansky F. Endovascular management of symptomatic vertebral artery dissection achieved using stent angioplasty and emboli protection device. Neurol Res 2003; 25(4):418–422. 77. Fateri F, Groebli Y, Rufenacht DA. Intraarterial thrombolysis and stent placement in the acute phase of blunt internal carotid artery trauma with subocclusive dissection and thromboembolic complication: case report and review of the literature. Ann Vasc Surg 2005; 19(3):434–437. 78. Hemphill JC III, Gress DR, Halbach VV. Endovascular therapy of traumatic injuries of the intracranial cerebral arteries. Crit Care Clin 1999; 15(4):811–829.

Chapter 11: Dissections of the Carotid and Vertebral Arteries 79. Kremer C, Mosso M, Georgiadis D, et al. Carotid dissection with permanent and transient occlusion or severe stenosis: long-term outcome. Neurology 2003; 60(2):271–275. 80. Touze E, Gauvrit JY, Moulin T, et al. Risk of stroke and recurrent dissection after a cervical artery dissection: a multicenter study. Neurology 2003; 61(10):1347–1351. 81. Zhang Q, Duan ZQ, Xin SJ, et al. Management of extracranial carotid artery aneurysms: 17 years’ experience. Eur J Vasc Endovasc Surg 1999; 18(2):162–165. 82. Guillon B, Brunereau L, Biousse V, et al. Long-term follow-up of aneurysms developed during extracranial internal carotid artery dissection. Neurology 1999. 53(1):117–122. 83. Touze E, Randoux B, Meary E, et al. Aneurysmal forms of cervical artery dissection: associated factors and outcome. Stroke 2001; 32(2):418–423. 84. Kadyrov NA, Friedman JA, Nichols DA, et al. Endovascular treatment of an internal carotid artery pseudoaneurysm following transsphenoidal surgery. Case report. J Neurosurg 2002; 96(3):624–627. 85. Peeters A, Goffette P, Dorban S, et al. An old dissecting aneurysm of the internal carotid artery presenting as acute stroke. Acta Neurol Belg 2003; 103(3):179–182. 86. Saito R, Ezura M, Takahashi A, et al. Combined neuroendovascular stenting and coil embolization for cervical carotid artery dissection causing symptomatic mass effect. Surg Neurol 2000; 53(4):318–322. 87. Sturzenegger M, Huber P. Cranial nerve palsies in spontaneous carotid artery dissection. J Neurol Neurosurg Psychiatry 1993; 56(11):1191–1199. 88. Ahn JY, Chung SS, Lee BH, et al. Treatment of spontaneous arterial dissections with stent placement for preservation of the parent artery. Acta Neurochir (Wien) 2005; 147(3):265–273; discussion 273. 89. Albuquerque FC, Han PP, Spetzler RF, et al. Carotid dissection: technical factors affecting endovascular therapy. Can J Neurol Sci 2002; 29(1):54–60. 90. Biffl WL, Moore EE, Offner PJ, et al. Blunt carotid arterial injuries: implications of a new grading scale. J Trauma 1999; 47(5):845–853. 91. Heye S, Maleux G, Vanderberghe R, et al. Symptomatic internal carotid artery dissecting pseudoaneurysm: endovascular treatment by stent-graft. Cardiovasc Intervent Radiol 2005; 28(4):499–501. 92. Redekop G, Marotta T, Weill A. Treatment of traumatic aneurysms and arteriovenous fistulas of the skull base by using endovascular stents. J Neurosurg 2001; 95(3):412–419. 93. Saket RR, Razavi MK, Sze DY, et al. Stent-graft treatment of extracranial carotid and vertebral arterial lesions. J Vasc Interv Radiol 2004; 15(10):1151–1156. 94. Saatci I, Cekirge HS, Ozturk MH, et al. Treatment of internal carotid artery aneurysms with a covered stent: experience in 24 patients with mid-term follow-up results. AJNR Am J Neuroradiol 2004; 25(10):1742–1749. 95. Pride GL Jr., Reploge RE, Rappard G, et al. Stent-coil treatment of a distal internal carotid artery dissecting pseudoaneurysm on a redundant loop by use of a flexible, dedicated nitinol intracranial stent. AJNR Am J Neuroradiol 2004; 25(2):333–337. 96. Uhl E, Schmid-Elsaesser R, Steiger HJ. Ruptured intracranial dissecting aneurysms: management considerations with a focus on surgical and endovascular techniques to preserve arterial continuity. Acta Neurochir (Wien) 2003; 145(12):1073–1083; discussion 1083–1084.

229

97. Anxionnat R, de Melo Neto JF, Bracard S, et al. Treatment of hemorrhagic intracranial dissections. Neurosurgery 2003; 53(2):289–300; discussion 300–301. 98. Aymard A, Gobin YP, Hodes JE, et al. Endovascular occlusion of vertebral arteries in the treatment of unclippable vertebrobasilar aneurysms. J Neurosurg 1991; 74(3):393–398. 99. Rabinov JD, Hellinger FR, Morris PP, et al. Endovascular management of vertebrobasilar dissecting aneurysms. AJNR Am J Neuroradiol 2003; 24(7):1421–1428. 100. Leibowitz R, Do HM, Marcellus ML, et al. Parent vessel occlusion for vertebrobasilar fusiform and dissecting aneurysms. AJNR Am J Neuroradiol 2003; 24(5):902–907. 101. Iihara K, Sakai N, Murao K, et al. Dissecting aneurysms of the vertebral artery: a management strategy. J Neurosurg 2002; 97(2):259–267. 102. Yoshimoto Y, Hoya K, Tanaka Y, et al. Basilar artery dissection. J Neurosurg 2005; 102(3):476–481. 103. Kurata A, Ohmomo T, Miyasaka Y, et al. Coil embolization for the treatment of ruptured dissecting vertebral aneurysms. AJNR Am J Neuroradiol 2001; 22(1):11–18. 104. MacKay CI, Han PP, Albuquerque FC, et al. Recurrence of a vertebral artery dissecting pseudoaneurysm after successful stent-supported coil embolization: case report. Neurosurgery 2003; 53(3):754–759; discussion 760–761. 105. Lylyk P, Cohen JE, Ceratto R, et al. Combined endovascular treatment of dissecting vertebral artery aneurysms by using stents and coils. J Neurosurg 2001; 94(3):427–432. 106. Ramgren B, Cronquist M, Rommer B, et al. Vertebrobasilar dissection with subarachnoid hemorrhage: a retrospective study of 29 patients. Neuroradiology 2005. 47(2):97–104. 107. Willing SJ, Skidmore F, Donaldson J, et al. Treatment of acute intracranial vertebrobasilar dissection with angioplasty and stent placement: report of two cases. AJNR Am J Neuroradiol 2003; 24(5):985–989. 108. Irie K, Negoro M, Hayakawa M, et al. Treatment of a spontaneous intracranial dissecting aneurysm with stentassisted coil embolization. Neuroradiology 2003. 45(11):825–829. 109. Chiaradio JC, Guzman L, Padilla L, et al. Intravascular graft stent treatment of a ruptured fusiform dissecting aneurysm of the intracranial vertebral artery: technical case report. Neurosurgery 2002; 50(1):213–216; discussion 216–217. 110. Ast G, Woimant F, Georges B, et al. Spontaneous dissection of the internal carotid artery in 68 patients. Eur J Med 1993; 2(8):466–472. 111. Schievink WI, Mokri B, Piepgras DG, et al. Recurrent spontaneous arterial dissections: risk in familial versus nonfamilial disease. Stroke 1996; 27(4):622–624. 112. Mokri B. Dissections of cervical and cephalic arteries. In: Meyer FB, ed. Sundt’s Occlusive Cerebrovascular Disease. Philadelphia: W.B. Saunders Co., 1994. 113. Yoshimoto Y, Wakai S. Unruptured intracranial vertebral artery dissection. Clinical course and serial radiographic imagings. Stroke 1997; 28(2):370–374. 114. Hosoya T, Adachi M, Yamaguchi K, et al. Clinical and neuroradiological features of intracranial vertebrobasilar artery dissection. Stroke 1999; 30(5):1083–1090. 115. Naito I, Iwai T, Sasaki T. Management of intracranial vertebral artery dissections initially presenting without subarachnoid hemorrhage. Neurosurgery 2002; 51(4):930–937; discussion 937–938.

12 Direct Carotid Cavernous Fistula Uday S. Kanamalla, Charles A. Jungreis, and Jeffrey P. Kochan Temple University Hospital, Temple University School of Medicine, Philadelphia, Pennsylvania, U.S.A.

INTRODUCTION

CLINICAL FEATURES

Carotid cavernous fistula (CCF) is an abnormal communication between the internal carotid artery (ICA) and the cavernous sinus. Direct CCF represents one specific form of CCF with high-pressure arterial blood entering the low-pressure venous cavernous sinus, most commonly via a single hole in the cavernous segment of the ICA (Fig. 1). In contrast, the other variety of CCF has been coined an ‘‘indirect’’ CCF and is characterized by a nidus of dural arterioles (1–3). Indirect CCF is described more fully in another chapter.

The onset of symptoms and signs of a direct CCF is usually acute and most commonly occurs within a few days of the instigating trauma. The length of time between the onset of the first symptoms and radiographic diagnosis of a CCF, however, has been reported to be variable and ranges up to 18 months (1,8). The severity and acuity of clinical features do not directly correlate to the size of the fistula, but are affected by the venous drainage. The most common symptoms and signs are orbital (Table 1) and include pulsatile tinnitus with a periorbital bruit, proptosis, chemosis, and injection (1–3,9,10). Less common presentations include intracranial hemorrhage (11). Cerebral ischemia, related to vascular steal, appears to be an exceedingly rare phenomenon, though it may occur in patients with an incomplete circle of Willis. The orbital symptoms and signs appear to relate not only to the degree of shunt but also to the pattern of venous drainage. The symptoms of CCF may be present in the contralateral eye secondary to drainage of the fistula through the intercavernous veins to the opposite side. Improvement or resolution of orbital symptoms following successful occlusion of the fistula tends to occur in about 80% of patients (12). The symptoms typically tend to resolve significantly in hours or days, though total resolution could take weeks or months, if it occurs at all. Duration of symptoms is also an important prognostic factor, with prolonged symptoms resolving more slowly. A relatively common, though dangerous, clinical scenario occurs in the setting of major trauma. Several days or a week after the trauma, an acute onset of severe proptosis, chemosis, injection, and pain associated with a pulsatile bruit over the orbit develops. Intraocular pressures elevate dramatically, and the patient becomes ophthalmoplegic in the affected eye. Despite the severe signs and symptoms, intervention within the next day or two typically results in rapid improvement (Fig. 3).

ETIOLOGY While most indirect CCFs are of spontaneous origin and uncertain etiology, direct CCFs most often occur as the result of closed head injury associated with a basal skull fracture (1–3). The ICA is fixed between the foramen lacerum and the anterior clinoid process by dural attachments. Shearing forces from head trauma, sometimes with accompanying penetrating injury from bony spicules, can cause the ICA to be torn between its points of dural attachment. In most cases, the laceration is single and unilateral. Sometimes the holes in the ICA are multiple, and sometimes bilateral CCFs occur. Direct CCF can also result from penetrating trauma, including iatrogenic trauma, such as during transsphenoidal surgery (4). Collagen deficiency diseases, such as Ehlers-Danlos syndrome, ruptured cavernous aneurysms, dissections, osteogenesis imperfecta, and fibromuscular dysplasia, have also been associated with the development of a CCF (5,7). The etiology of spontaneous direct CCF remains speculative and has most commonly been attributed to rupture of a preexisting cavernous aneurysm (2,3) (Fig. 2). Direct CCF is more common in the younger population as opposed to indirect CCF, which tends to occur in the older population. However, trauma at any age remains the most common etiology of direct CCF.

232

Kanamalla et al.

Figure 1 Direct CCF embolized with detachable coils. (A) Axial MR shows large varices especially in relation to the left cavernous sinus. (B) Axial MR shows dilated superior ophthalmic veins bilaterally. (C) Lateral angiographic view during a selective ICA injection shows rapid flow into the varices. The posterior venous drainage is poor, and most of the flow is anterior into the superior ophthalmic vein. (D) Postembolization with detachable coils shows a minimal remnant. No arteriovenous shunting persists and the ICA is preserved. (E) Unsubtracted angiographic view showing coil nest postemobilization. Abbreviations: MR, magnetic resonance; ICA, internal carotid artery.

Figure 2 Direct CCF secondary to rupture of a cavernous ICA aneurysm embolized with a detachable balloon. (A) Lateral angiographic view during a selective ICA injection shows rapid flow into the varices. The venous drainage is mainly posterior into the IPS. (B) Postembolization. The cavernous aneurysm is now apparent. The fistula is closed. Abbreviations: CCF, carotid cavernous fistula; ICA, internal carotid artery; IPS, inferior petrosal sinus.

Table 1

Symptoms and Signs Associated with Direct CCF (1) Number of cases

Symptoms Diplopia Eye redness Proptosis Headache Bruit Diminished vision Facial numbness Ocular pain Signs Dilated episcleral veins Diminished vision Elevated intraocular pressure Sixth nerve paresis Chemosis Third nerve paresis Papilledema Total number of cases ¼ 14.

14 13 10 9 8 7 6 5 14 12 10 10 10 4 4

ANATOMY AND PATHOPHYSIOLOGY The cavernous sinus has been regarded as a contiguous network of anatomically separated sinusoids rather than actual veins (13). A rent in the wall of the intracavernous carotid artery, or rupture of one of its branches that traverses and is surrounded on all sides by the sinus cavity, produces an arteriovenous fistula without concomitant venous injury in contradistinction to fistulas elsewhere in the body. The superior and inferior ophthalmic veins provide normal venous drainage from the orbit to the cavernous sinus. The superficial middle cerebral veins drain the brain through the sphenoparietal sinus to the cavernous sinus. The cavernous sinus, in turn, normally drains through the superior and inferior petrosal sinuses (IPSs) to the jugular bulb and via emissary veins to the pterygoid venous plexus. When a fistula develops between the ICA and the cavernous sinus, the high flow and pressure

Chapter 12: Direct Carotid Cavernous Fistula

233

Additionally, during treatment, one should always remember the possibility of redirection of flow into dangerous venous channels by the treatment itself. This kind of acute venous diversion into the remaining venous pathways following selective occlusion of one venous channel can result in aggravation of symptoms and increase the morbidity and mortality associated with the disease. For example, if the IPS is occluded during treatment, diversion of flow into the superior ophthalmic vein may increase proptosis and may also increase cortical venous drainage with associated increased risk of intracerebral or subarachnoid hemorrhage (11,14).

CLASSIFICATION

Figure 3 64-year-old female before and after embolization of a CCF. (Top) Photograph before treatment. Note the bilateral proptosis, chemosis, and injection. The pupils are dilated. (Bottom) Photograph 10 days after treatment. It shows significant resolution. Abbreviation: CCF, carotid cavernous fistula.

within the venous drainage pathways increase and there is reversal of flow within the normal tributaries to the cavernous sinus. Furthermore, the venous drainage pathways dilate to accommodate the increased flow. It is this abnormal venous diversion of flow that results in the characteristic signs and symptoms associated with direct CCF. The reversed and increased flow into the superior and inferior ophthalmic veins causes orbital venous hypertension. Visual deterioration results from a combination of reduced arterial perfusion and venous hypertension with accompanying glaucoma. Retinal perfusion suffers. Intraocular pressures rise as a result of venous hypertension. Rubeosis irdis, a neovascularity of the iris induced by prolonged ischemia, may also contribute to overall ocular necrosis (10). Obtrusive diplopia and ophthalmoplegia occur as a result of cranial nerve compression secondary to mass effect in the cavernous sinus from distended vessels. Edema of extraocular contents, including the muscles, can also contribute to diplopia and appears to be related to vascular engorgement and enlargement of the extraocular muscles. The symptoms and signs vary depending on which veins drain the fistula and how distended they become. For example, posterior drainage via the superior and IPSs can result in pulsatile tinnitus. Intracranial hemorrhage is a dreaded complication. This is due to reversal of venous drainage into the sphenoparietal sinus, with resultant cerebral cortical venous hypertension.

CCFs can be classified according to three criteria: (1) pathogenically into spontaneous or traumatic fistulas, (2) hemodynamically into high-flow or lowflow fistulas, and (3) angiographically into direct or indirect (dural) fistulas. Some CCFs are hybrids of the above. An angiographic classification provides an objective method for grouping CCF, determining prognosis, and planning the therapeutic management. All CCFs can be placed into one of four angiographic categories based on whether the CCF is direct and on the anatomical origin of the arteries supplying the fistula (Fig. 1) (2,3). Type A fistulas are direct shunts between the ICA and cavernous sinus. Types B, C, and D are indirect or dural shunts. Type B is a fistula between meningeal branches of the ICA and the cavernous sinus. Type C is a dural shunt between meningeal branches of the external carotid artery (ECA) and the cavernous sinus. Type D, the most common type, is a dural shunt between meningeal branches of both the ICA and ECA and the cavernous sinus. Bilateral CCFs represent a special case of the above. The angiographic criteria for differentiating a fistula into high-flow or low-flow categories are quite subjective. High-flow fistulas fill the cavernous sinus and efferent veins within a fraction of a second, and the intracranial branches of the ICA fill partially or are not visualized at all. In contrast, an angiogram of a low-flow fistula will reveal slower drainage into the venous system and filling of the intracranial branches of the ICA. Note that the terms ‘‘high flow’’ and ‘‘low flow’’ are relative to each other. Both are high flow compared with normal.

INDICATIONS FOR TREATMENT The indications for treatment are not absolute and depend on the general physical condition of the patient, the severity of the symptoms, and the anatomy of the fistula, which, in turn, determines the treatment modality. Type A direct fistulas will rarely resolve spontaneously and almost always require treatment. Progressive visual loss, uncontrollable elevations of intraocular pressure, an intolerable bruit or headache, or enlargement of traumatic

234

Kanamalla et al.

aneurysm beyond the cavernous sinus are all strong indications for treatment. Additionally, treatment is also appropriate because of corneal exposure, obtrusive diplopia, or cosmetically offensive proptosis. Finally, the presence of cortical venous drainage, where there is concern for intracranial hemorrhage, constitutes an indication for therapy (2–8). Halbach et al. have identified certain high-risk features that represent indications for urgent treatment. These include development of intracranial hemorrhage, epistaxis, increased intraocular pressures, decreased visual acuity, rapidly progressive proptosis, and cerebral ischemia (11).

IMAGING STUDIES Conventional catheter angiography is usually required. It helps confirm the diagnosis, helps determine the type of fistula, and provides the therapeutic capability. CT or MRI can establish the diagnosis of CCF. Their primary role, though, is to evaluate the brain parenchyma for associated injuries in the setting of trauma and to detect possible ischemic changes. Additionally, thin-section CT with coronal reconstructions can help evaluate skull base fractures. The findings of CCF on cross-sectional imaging include proptosis, swelling of extraocular muscles, and dilation of the superior ophthalmic vein with distention of the ipsilateral cavernous sinus (Fig. 4). However, CT or MRI

will not typically help with differentiation of direct from the indirect types. Orbital ultrasound can also be performed to demonstrate findings of thickened extraocular muscles as well as dilated superior ophthalmic veins. In the right clinical setting, this procedure can also help confirm the clinical diagnosis. Other tests that are often performed include complete ophthalmologic workup inclusive of visual acuity, pupillary function, intraocular pressure measurement, fundoscopy (direct and indirect), and gonioscopy. For optimal angiography, high-resolution digital subtraction is essential. The goals of the diagnostic angiogram are to evaluate the location and size of the fistula and the venous drainage pathways. Additionally, associated traumatic vascular injuries, ICA pseudoaneurysms, and cavernous sinus varices need to be excluded. To help differentiate a direct from an indirect fistula, an angiographic search should be carried out for an ECA supply to the fistula. The angiographic evaluation of CCF should, therefore, include selective catheterization and angiography of the ICAs and ECAs bilaterally. Because of the very high flow, it may not be possible to identify the morphology of the fistula in terms of exact location or size on selective angiograms without specific maneuvers to slow the flow across the fistula. The maneuvers could consist of ipsilateral ICA compression while injecting into the artery to slow the flow. Alternatively, the contralateral ICA or vertebral artery (so-called Allcock maneuver) can be

Figure 4 Direct CCF embolized with a detachable balloon. (A) Axial CT showing dilated left superior ophthalmic vein. (B) PA angiographic view during a right ICA injection shows arteriovenous shunting with cross-filling to the opposite side, including the left cavernous sinus and left superior ophthalmic vein. (C) Lateral angiographic view during a right ICA injection again demonstrates rapid arteriovenous shunting. (D) Lateral angiographic view with balloon in position closing the fistula with preservation of the ICA. (E) Unsubtracted view showing balloon postembolization. Abbreviations: CCF, carotid cavernous fistula; CT, computed tomography; PA, posteroanterior; ICA, internal carotid artery.

Chapter 12: Direct Carotid Cavernous Fistula

injected while providing ipsilateral ICA compression (Dr. Allan Fox, personal communication). Often, because of the high flow, injection of the contralateral ICA or the vertebral artery will opacify the CCF even without compression. The venous drainage from the cavernous sinus of a direct CCF can extend anteriorly into the ophthalmic veins, inferiorly into the pterygoid venous plexus, posteriorly into the petrosal sinuses, superiorly into the cortical veins via the sphenoparietal sinus, and finally into the contralateral cavernous sinus. The pattern of venous drainage dictates the predominant symptom or sign seen in the patient. Most often a pattern of mixed venous drainage will be seen.

TREATMENT OPTIONS AND CONSIDERATIONS While there are reasons to intervene very early, in general the treatment of CCF is rarely an emergent procedure, but it is often urgent. That is, treatment can usually be undertaken semielectively when the patient is otherwise stable. The treatment of CCF has evolved over the past 40 years. The earliest surgical treatments of proximal occlusion of the ICA or trapping have largely been abandoned because of the high risk of stroke and blindness, often without obliterating the fistula (10). Today, the primary treatment modality is endovascular therapy. Approaches to occlude CCFs have been described, with an increased focus on preserving ICA flow. Maintained patency of the ICA after treatment of traumatic CCF is increasing largely because of improvements in the technology (better devices), but will probably never reach 100% (3,15–17). The goal of treatment is to obliterate the fistula, which can be accomplished with a wide variety of techniques and agents, each with its advantages and disadvantages. The treatment of CCF takes into consideration the speed of flow through the fistula, its arterial supply, and the routes of venous drainage. Equally important is to take into consideration the patient’s general physical condition prior to formulating a therapeutic plan. For example, in the setting of acute CCF in a multitrauma patient, the more critical injuries of the patient must be addressed first.

Medical Therapy In the acute setting of vision loss and/or paralysis of cranial nerves, glucocorticosteroids (e.g., dexamethasone) may be used while waiting for definitive diagnostic studies and treatments. Similarly, in patients with elevated intraocular pressures, pharmacologic management with topical ß-adrenergic blockers and oral acetazolamide (Diamox) is performed as adjunctive therapy, until definitive therapy for the fistula is undertaken.

Surgical Therapy Surgical therapy is presently considered only as a last resort. The earlier treatments in the form of proximal

235

ICA occlusion or ICA trapping have largely been abandoned. Surgical therapy is only considered in cases where there has been a failure of endovascular therapy or is used in conjunction with endovascular therapy. One form of surgical therapy is assistance in endovascular access to the superior ophthalmic vein. In patients without arterial or venous access to the fistula, direct access to the superior ophthalmic vein following surgical exposure of the vein in the orbit has been used with successful transvenous embolization of the fistula (18). Direct surgical exposure of the cavernous sinus via craniotomy followed by direct puncture for embolization and closure of the fistula remains an alternative when all other routes are exhausted (19). There are also reports in the literature of using a superficial temporal artery to middle cerebral artery bypass prior to sacrifice of the ipsilateral ICA in patients who are unable to tolerate ICA occlusion (2). Adjunctive surgical procedures in the form of a lateral canthotomy have been performed as a temporizing measure for orbital symptoms such as severe proptosis, markedly elevated intraocular pressures, and rapidly declining visual acuity.

Endovascular Therapy This therapy is performed transarterially and/or transvenously. Large series have shown the effectiveness of transarterial balloon embolization, which had emerged as the treatment of choice for this disease (3,15–17). In the United States, however, detachable silicone balloons have been withdrawn from the market, which has led to the use of various alternative embolization agents, including platinum microcoils and acrylic (N-butyl cyanoacrylate, or NBCA) (20–26). More recently, closure of fistula via stent graft and stent/balloon assist has been reported (26–34). Currently, our initial treatment attempt is with endovascular coils via the transarterial approach. Transarterial Approach

Conceptually, the ideal goal of treatment is to occlude the fistula on the venous side, thereby preserving the ICA. However, sometimes this treatment proves to be impossible, and sacrifice of the ICA is required to close the fistula. If the ICA requires sacrifice, then it must be occluded both above and below the fistula or flow to the fistula will persist (Fig. 5). This methodology of occluding the ICA is called ‘‘trapping’’ and can be done surgically, endovascularly or by a combination thereof. In treating 54 traumatic CCFs with detachable balloons, Debrun et al. had to sacrifice the ICA close to the fistula in 20 (37%) of their cases. ICA occlusion at the level of the fistula can be performed with detachable balloons or with platinum coils. Interestingly, since the ICA flow above the level of the fistula is usually reversed into the fistula, test occlusion of the ICA is not usually required. However, in a patient in whom some distal ICA flow above the fistula remains antegrade, tolerance to ICA occlusion

236

Kanamalla et al.

Figure 5 Direct CCF secondary to a gunshot wound. The right ICA required sacrifice to close the fistula completely. Endovascular coils were utilized to trap the fistula. (A) Lateral view after embolization. The detachable coils are in the right ICA and trap the fistula. Detachable coils are both above and below the fistula. Note also that some larger ‘‘free’’ coils are in the proximal ICA to provide additional stability to the coil nest. (B) Frontal angiographic view after right carotid sacrifice during left ICA angiogram showing excellent cross-filling without flow into the fistula. (C) Lateral angiographic view after embolization during a vertebral artery angiogram. The distal right ICA is filled via the posterior communicating artery, but the coils prevent retrograde flow to the fistula. Abbreviations: CCF, carotid cavernous fistula; ICA, internal carotid artery.

using the ‘‘balloon occlusion test’’ prior to occlusion of the ICA may be required. In the uncommon event of failure to tolerate the occlusion, alternate methods of therapy will need to be strongly considered. Selective embolization of the fistula via the transarterial approach with detachable balloons or coils is presently considered the method of choice for the treatment of most single-hole CCFs. When using balloons, the procedure involves the detachment of single or multiple silicone balloons into the cavernous sinus (venous side) close to the fistula to occlude the abnormal fistulous communication. Thromboembolic complications from the procedure, though rare, are well described in the literature and could be the result of catheter or balloon manipulation causing endothelial damage or the result of inadvertent balloon detachment or balloon migration. Transarterial balloon embolization fails in at least 5% to 10% of cases (20,26,35). Failure occurs because the fistula orifice may be too small to allow entry of the balloon catheter, the venous compartment of the fistulous communication may be too small to allow balloon inflation, or sharp margins of the adjacent bony fragments or foreign bodies may rupture the balloon during inflation (36). Also, in some patients who have subtotal occlusion after initial balloon embolization, navigation of additional balloons into the fistula may be unsuccessful owing to the presence of balloons partially blocking the fistula orifice. Lastly, the ability to microcatheterize and hold position within the fistula may also be limited depending on the position of the fistula ostium. Typically, it is more difficult to catheterize when the fistula orifice is just beyond the posterior genu along the inferior wall of the C4 segment of the cavernous carotid artery because the angle of entry from the ICA is very acute. The development of steerable microcatheters and microguidewires has allowed successful use of other embolic agents such as platinum coils or liquid embolic material. It is technically easier to guide a microcatheter/microguidewire combination through

a small fistula than it is to guide a detachable balloon, and the former also allows for more precise placement of embolic agents within the cavernous sinus close to the fistula orifice. Care must be taken to ensure that these embolic agents are not deposited within the carotid artery. An intracavernous venogram should always be obtained to verify that the microcatheter is positioned properly and to accurately delineate the cavernous sinus prior to deposition of coils. The occluding coils should be placed as close to the fistula orifice as possible. Also, fewer coils are needed if the fistula can be blocked at its orifice. The use of standard nondetachable or ‘‘free’’ platinum coils in the successful treatment of CCF has been described in the literature (20). Since the development of detachable platinum coils, though, embolization using free coils by themselves is rarely performed. Technical pitfalls associated with embolization using free platinum coils, including difficulty in retrieval, the relative stiffness of the coils, the risk of perforation, and the difficulty of packing them tightly, have encouraged the use of the newer-generation detachable coils. With detachable coils, if the microcatheter recoils during placement of the coil, the coil can be repositioned. Nondetachable balloon assistance during coil insertion can also be helpful. The development of detachable or retrievable platinum coils has significantly reduced the risks associated with standard free platinum coils (20–26). The advantage of using detachable platinum coils is the ability to control their placement and to easily retrieve, reposition, or exchange them if necessary. Technical pitfalls are possible with detachable platinum coils also. The soft platinum coils exert little force on the surrounding structures and might be more easily displaced in a high-flow fistula. Again, use of stent- or balloon-assisted placement of the detachable platinum coils may allow for tighter packing and more complete closure of the fistula. Liquid adhesives such as NBCA or IBCA have also been used in conjunction with balloons or coils (26).

Chapter 12: Direct Carotid Cavernous Fistula

Initial placement of detachable balloons or coils within the fistula helps significantly reduce the flow, allowing for a safer use of liquid adhesives to complete the occlusion of the fistula. When liquid adhesives are used, reflux into the carotid artery and devastating strokes can occur. This risk increases when closure of the fistula is nearly complete and the pressure gradient between the carotid artery and cavernous sinus is lowered. Real-time digital subtraction and careful slow injections of small volumes of embolic material can avert this potential complication. Newer liquid agents such as Onyx (Onyx Liquid Embolic System, Micro Therapeutics, Inc., Irvine, California, U.S.) hold interesting promise but remain unproven. In younger patients and in those with a straighter course of the vessel, it is possible to deliver covered stents to bridge the site of fistula and, thereby, close the fistula with preservation of the ICA (33,34). Delivery of stents to the small-caliber, tortuous intracranial arteries requires low-profile, flexible stents with high elasticity as well as good pushability properties for the delivery catheter. Stent designs are continually being improved in order to overcome problems in delivery, deployment, and prevention of stent thrombosis. The Jomed covered stent (JoMed International AB, Helsingborg, Sweden) is a surgical steel endoprosthesis with an expandable PTFE graft, and this device is manually compressed over a conventional angioplasty balloon. The highly stretchable, ultrathin graft material allows radial expansion, and its surface and edges are very smooth. Once compressed on an angioplasty balloon, the device is still relatively stiff and its introduction into an intracranial artery is problematic. The use of a stiff guidewire and distally advanced guiding catheter with firm backup are mandatory. Stent thrombosis after successful deployment constitutes an important complication, especially in covered stents. Adequate anticoagulant and antiplatelet treatment before and after the procedure are essential to prevent thromboembolic complications, therapeutic requirements that might not be desirable in the multitrauma patient.

237

angiographic localization of the fistula site and confirmation of fistula closure. While the transvenous approach may be effective, it is not without risk. The most common complication associated with the procedure is perforation of intracranial veins during catheterization. Although arterialization of venous structures can occur in long-standing shunts, the draining veins and dural sinuses in CCFs are often still thin-walled and can be perforated by a catheter or guidewire. With the increased pressure and flow of the arterialized blood in these structures, small perforations can result in rapidly fatal subarachnoid hemorrhage. Alternatively, venous thrombosis may occur following injury to the veins. This could be significant if the venous drainage is altered but the fistula remains open. For example, if the posterior drainage (IPS) is occluded without closure of the fistula, aggravation of ocular symptoms (superior ophthalmic vein) or hemorrhage (cortical drainage) may occur. Similarly, if the superior ophthalmic vein is occluded, fistula flow may be diverted into cortical veins, increasing the risk of hemorrhagic complications. In our experience, closure of the distal superior ophthalmic vein is often associate with severe aggravation of ocular symptoms if persistent fistula exists. As a technical note, therefore, care must be taken not to occlude the access vein before the other venous channels and fistula are closed. If cortical venous drainage is noted, occlusion of the cavernous sinus at the origin of the vein or, alternatively, direct placement of coils into the sphenoparietal sinus should be performed as a first step. Also, if a vein is occluded, drainage of the pons and brain stem may be impaired. Debrun et al. (15), in his report of transfemoral venous occlusion of CCF with detachable balloons, noted that the percentage of success was low because partitions within the cavernous sinus precluded placement of a balloon near the fistula orifice. With the development of softer and smaller catheters and steerable guidewires, the risks of these complications have been reduced. Transorbital Approach

Transvenous Approach

When transarterial routes are unsuccessful, or if the venous anatomy is opportune, transvenous embolization can be performed (15,26,35,37). Access to the cavernous sinus through the ipsilateral jugular vein and IPS is the most common transvenous approach. Other venous routes that have been used on occasion to access the cavernous sinus include the contralateral IPS, pterygoid plexus of veins, superior ophthalmic vein, and also cortical veins via the sphenoparietal sinus. These alternate venous routes were used because of nonvisualization of the IPS or inability to adequately gain access to the cavernous sinus via the ipsilateral IPS. Following transvenous access to the cavernous sinus, detachable balloons, detachable platinum coils, or liquid embolic agents have been used successfully. In transvenous cases, an arterial catheter (4-5 French) typically has to be placed into the ipsilateral common carotid artery or ICA for

As the technology has improved, the success rate for treatment using other approaches has increased. The transorbital approach entails direct cut-down under ultrasound guidance into the superior ophthalmic vein with catheterization and embolization of the fistula. We found this approach more difficult and cumbersome than the others and have abandoned it.

FOLLOW-UP Angiographic residual flow through the fistula is not a definite indication for further treatment. Subtotal occlusion of a fistula may not indicate failure. On the contrary, the goals of treatment are to alleviate the clinical symptoms, to control the intraocular pressures, and to eliminate cortical venous drainage. We use a follow-up angiogram at three to six months after treatment for evaluation in conjunction with a thorough ophthalmologic exam.

238

Kanamalla et al.

SUMMARY Direct CCF results from a tear in the cavernous ICA. Endovascular occlusion of the fistula is the preferred method of treatment. With improvement in catheter techniques as well as embolic agents and stents, the treatment of these complex lesions is generally safe and effective.

REFERENCES 1. Newton TH, Hoyt WF. Dural arteriovenous shunts in the region of the cavernous sinus. Neuroradiology 1970; 1:71–81. 2. Barrow DL, Spector RH, Braun IF, et al. Classification and treatment of spontaneous carotid-cavernous sinus fistulas. J Neurosurg 1985; 62:248–256. 3. Debrun GM, Vinuela F, Fox AJ, et al. Indications for treatment and classification of 132 carotid-cavernous fistulas. Neurosurgery 1988; 22:285–289. 4. Kocer N, Kizilkilie O, Albayrum S, et al. Treatment of iatrogenic internal carotid artery laceration and carotid cavernous fistula with endovascular stent-graft placement. AJNR Am J Neuroradiol 2002; 23:442–446. 5. Heishima G, Cahan LD, Mehringer CM, et al. Spontaneous arteriovenous fistulas of cerebral vessels in association with fibromuscular dysplasia. Neurosurgery 1986; 18:454–458. 6. Farley MK, Clark RD, Fallor MK, et al. Spontaneous carotid-cavernous fistula and the Ehlers-Danlos syndrome. Ophthalmology 1983; 90:1337–1342. 7. Hollister DW. Heritable disorders of connective tissue: Ehlers-Danlos syndrome. Pediatr Clin North Am 1978; 25:575–591. 8. Debrun GM, Lacour P, Fox AJ, et al. Traumatic carotid cavernous fistulas: etiology, clinical presentation, diagnosis, treatment, results. Semin Intervent Radiol 1987; 4:242–248. 9. Jorgensen JS, Guthoff R. Ophthalmoscopic findings in spontaneous carotid cavernous fistula: an analysis of 20 patients. Graefes Arch Clin Exp Ophthalmol 1988; 226:34–36. 10. Sanders MD, Hoyt WF. Hypoxic ocular sequelae of carotid-cavernous fistulae. Study of the causes of visual failure before and after neurosurgical treatment in a series of 25 cases. Br J Ophthalmol 1969; 53:82–97. 11. Halbach VV, Hieshima GB, Higashida RT, et al. Carotidcavernous fistulae: indications for urgent therapy. AJNR Am J Neuroradiol 1987; 8:627–633. 12. Serbinenko FA. Balloon catheterization and occlusion of major cerebral vessels. J Neurosurg 1974; 41:125–145. 13. Day AL, Rhoton Al Jr. Aneurysms and arteriovenous fistulae of the intracavernous carotid artery and its branches. Youmans JR, ed. Neurological Surgery. 2nd ed. Philadelphia: WB Saunders, 1982:1764–1785. 14. Turner DM, VanGilder JC, Mojtahedi S, et al. Spontaneous intracerebral hematoma in carotid cavernous fistula. J Neurosurg 1983; 59:680–686. 15. Debrun G, Lacour P, Vinuela F, et al. Treatment of 54 traumatic carotid-cavernous fistulas. J Neurosurg 1981; 55:678–692. 16. Higashide RT, Halbach VV, Tsai FY, et al. Interventional neurovascular treatment of traumatic and vertebral lesions, results in 234 cases. AJR Am J Roentgenol 1989; 153:577–582. 17. Lewis A, Tomsick TA, Tew JJ. Management of 100 consecutive direct carotid-cavernous fistulas: results of treatment with detachable balloons. Neurosurgery 1995; 36:239–244.

18. Uflacker R, Lima S, Ribas G, et al. Carotid-cavernous fistulas: embolization through the superior ophthalmic vein approach. Radiology 1986; 159:175–179. 19. Isamat F, Ferre E, Twose J. Direct intracavernous obliteration of high-flow carotid-cavernous fistulas. J Neurosurg 1986; 65:770–775. 20. Halbach VV, Higashida RT, Barnwell SL, et al. Transarterial platinum coil embolization of carotid-cavernous fistula. AJNR Am J Neuroradiol 1991; 12:429–433. 21. Guglielmi G, Vineula F, Duckwiler G, et al. High flow, small-hole arteriovenous fistula; treatment with electrodetachable coils. ANJR Am J Neuroradiol 1995; 16:325–328. 22. Siniluoto T, Seppanen S, Kuurne T, et al. Transarterial embolization of a direct carotid cavernous fistula with Guglielmi detachable coils. AJNR Am J Neuroradiol 1997; 18:519–523. 23. Bavinski G, Killer M, Gruber A, et al. Treatment of posttraumatic carotico-cavernous fistulae using electrolytically detachable coils; technical aspects and preliminary experience. Neuroradiology 1997; 39:81–85. 24. Jansen O, Dorfler A, Forstinb M, et al. Endovascular therapy of arteriovenous fistulae with electrolytically detachable coils. Neuroradiology 1999; 41:951–957. 25. Goto K, Hieshima GB, Higashida RT, et al. Treatment of direct carotid cavernous sinus fistulae: various therapeutic approaches and results in 148 cases. Acta Radiol Suppl 1986; 369:576–579. 26. Moron FE, Klucznik RP, Mawad ME, et al. Endovascular treatment of high-flow carotid cavernous fistulas by stentassisted coil placement. AJNR Am J Neuroradiol 2005; 26:1399–1404. 27. Lee CY, Yim MB, Kim IM, et al. Traumatic aneurysm of the supraclinoid internal carotid artery and an associated carotid-cavernous fistula: vascular reconstruction performed using intravascular implantation of stents and coils—case report. J Nuerosurg 2004; 100:115–119. 28. Ahn JY, Lee BH, Joo JY. Stent-assisted Guglielmi detachable coils embolization for the treatment of a traumatic carotid cavernous fistula. J Clin Neurosci 2003; 10:96–98. 29. Men S, Ozturk H, Hekimoglu B. Traumatic carotidcavernous fistula treated by combined transarterial and transvenous coil embolization and associated cavernous internal carotid artery dissection treated with stent placement: case report. J Neurosurg 2003; 99:584–586. 30. Redekop G, Marotta T, Weill A. Treatment of traumatic aneurysm and arteriovenous fistulas of the skull base by using endovascular stents. J Neurosurg 2001; 95:412–419. 31. Weber W, Henkes H, Berg-Dammer E, et al. Cure of a direct carotid cavernous fistula by endovascular stent deployment. Cerebrovasc Dis 2001; 12:272–275. 32. Kim SH, Qureshi AI, Boulos AS, et al. Intracranial stent placement for the treatment of a carotid-cavernous fistula associated with intra-cranial angioplasty: case report. J Neurosurg 2003; 98:1116–1119. 33. Mollers MO, Reith W. Intracranial arteriovenous fistula caused by endovascular stent-grafting and dilatation. Neuroradiology 2004; 46:323–325. 34. Vannien RI, Manninen HI, Rinne J. Intrasellar latrogenic carotid pseudoaneurysm: endovascular treatment with a polytetrafluroethtlene-covered stent. Cardiovasc Intervent Radiol 2003; 26:298–301. 35. Halbach VV, Higashida RT, Hieshima GB, et al. Transvenous embolization of direct carotid cavernous fistulas. AJNR Am J Neuroradiol 1988; 9:741–747. 36. Horton JA, Jungreis CA, Stratemeier PH. Sharp vascular calcifications and acute balloon rupture during embolization. AJNR Am J Neuroradiol 1991; 12:1070–1073. 37. Chung GF, Tomsick TA. Transvenous embolization of a direct carotid cavernous fistula through the pterygoid plexus. AJNR Am J Neuroradiol 2003; 23:1156–1159.

13 Endovascular Management of Intracranial Aneurysms Darren Orbach, Tibor Becske, and Peter Kim Nelson Departments of Neurology, Neurosurgery, and Radiology, New York University Medical Center, New York, New York, U.S.A.

DEFINITIONS AND EPIDEMIOLOGY Intracranial arterial aneurysms historically have been thought to result from a developmental abnormality in vasculogenesis or angiogenesis, resulting in an error in the normal cycle of cell birth, apoptosis, and maintenance of the normal extracellular matrix (1) and ultimately leading to fatigue of the viscoelastic elements of the vessel wall and outward ballooning of the affected vascular segment with an increasing propensity to rupture as they enlarge. Although computer models have recently been used to elucidate the growth and rupture mechanism of aneurysms (2,3), it remains an area of active investigation. Arterial dissections, pseudoaneurysms, mycotic aneurysms, and flow-related aneurysms associated with arteriovenous malformations (AVMs) are specifically excluded from this discussion in that these lesions result from postnatal insults to a biologically and morphologically normal cerebral artery. While the techniques brought to bear in treating these conditions can be similar to those used for congenital aneurysms, the etiology and overall management of these entities are sufficiently disparate that these latter conditions merit their own attention. Intracranial aneurysms constitute a common cerebrovascular abnormality resulting in a significant health problem worldwide. The estimated adult prevalence varies from 0.2% to 9% (4,5), with the incidence of those discovered angiographically ranging from 0.5% to 1% of the population (6). A recent 30-year retrospective study found the incidence of intracranial aneurysms to be nearly equal to that of primary brain tumors in the same population (7). While patients with intracranial aneurysms may present with neurologic deficit secondary to mass effect or thromboembolism, the most common clinical presentation is subarachnoid hemorrhage (SAH). This condition, which commonly afflicts young and middle-aged adults, has an annual U.S. incidence of over 25,000 cases, leaving over 18,000 of these patients (i.e., 72%) dead or severely disabled, and generating a cost of greater than $1.75 billion/yr (8). Comparing the high prevalence of cerebral aneurysms with the relatively low incidence of SAH, it should be apparent that only a small fraction of aneurysms ever rupture. However,

the neurologic devastation wrought by a rupture is sufficient enough for the treatment of asymptomatic aneurysms to be often advocated.

DECISION MATRIX---TO TREAT OR NOT TO TREAT Ruptured intracranial aneurysms are almost always treated urgently because of the high risk of rebleed. Moreover, there is evidence that unruptured aneurysms in a patient with a history of SAH are at increased risk of rupture, and elective treatment for these patients is widely advocated. However, the decision to treat an unruptured intracranial aneurysm in a patient with no prior history of SAH must be based on consideration of the patient’s clinical presentation, within the context of what is known of the natural history of aneurysms and the risk/benefit ratio associated with a specific treatment. These latter two considerations are themselves complex, involving such factors as location and size of the aneurysm, presence of a connective tissue disorder such as polycystic kidney disease, the presence of hypertension, the patient’s age, the case volume of the treating institution and operator, etc. (9–12). The 1998 International Study of Unruptured Intracranial Aneurysms (ISUIA) investigated both the natural history of previously unruptured intracranial aneurysms as a function of different patient characteristics (in its retrospective arm) and the morbidity and mortality of surgical treatment (in its prospective arm) (13). The investigators initially concluded that for patients with aneurysms measuring less than 10 mm in diameter, if there was no prior history of SAH in the patient, the risk of rupture was less than 0.05% per year, and such aneurysms are referred to as incidental. In the setting of previous SAH (from another aneurysm), the risk was 11 times higher, and such aneurysms are referred to as additional. For aneurysms of 10-mm or greater diameter, the annual risk of rupture approached 1%, whether or not there was a prior history of SAH. Thus, in asymptomatic patients with small aneurysms, the adverse effects associated with surgical treatment exceeded the morbidity and mortality of the condition.

240

Orbach et al.

Several neurosurgical groups criticized the ISUIA soon after its publication (14), stressing that the retrospective arm of the study was biased toward aneurysm locations that were less likely to rupture. Moreover, by dint of the study subjects having been specifically chosen as a group in which not to intervene, these aneurysms were thought to possibly be more stable than randomly selected aneurysms from the community would be. In fact, a long-term study of unselected unoperated patients from Finland, published after ISUIA (15), reported a significantly higher rupture rate. Others (16) pointed out that the annual incidence of 16,000 cases of SAH in the United States could not be explained by a rupture rate of 0.05% for aneurysms smaller than 10 mm; the prevalence of unruptured aneurysms would have to be orders of magnitude higher than it actually is. Along these lines, a biophysical model of aneurysm rupture (17) predicted that, rather than a sharp threshold of increasing risk at 10 mm as suggested by ISUIA, there was actually a continuously increasing risk of rupture, scaling with the diameter (18). Nevertheless, the ISUIA and other studies (19) led to the issuance, in 2000, of a set of formal Recommendations for the Management of Patients with Unruptured Intracranial Aneurysms by the American Heart Association (20), which recommended no treatment for small, incidental aneurysms in asymptomatic patients. Treatment was recommended for those with large aneurysms, those with worrying morphology, and those with a history of SAH. In 2003, the full report from the prospective arm of ISUIA was published (13), comparing the natural history of unruptured aneurysms and the morbidity and mortality of neurosurgical clipping and endovascular coiling in various groups of patients. As in the 1998 report, the investigators found that the size of the aneurysm was an important determinant of its natural history, with aneurysms smaller than 7 mm in patients without prior SAH having an approximately 0.1% annual risk. Interestingly, a family history of rupture seemed to pose no additional hazard. The investigators pointed out that this cohort was nearly asymptomatic, and the presence of symptoms may be indicative of a more dangerous aneurysm. Larger aneurysm size was a definite risk factor for rupture in previously unruptured patients, but less clearly so for patients with prior SAH, in whom even small aneurysms had a relatively high rupture rate. Multivariate analysis showed that relative to those with aneurysms smaller than 7 mm, those with 7- to 12-mm aneurysms had a relative risk (RR) of rupture of 3.3 and those with larger than 12 mm had an RR of 17. Three locations were associated with altered risks of rupture relative to the internal carotid artery: the basilar tip (RR 2.3), the cavernous carotid (RR 0.15), and the posterior communicating artery (RR 2.1). Patient age was not predictive of rupture likelihood. However, surgical morbidity and mortality were closely tied to age; endovascular morbidity appeared to be less so. Other factors linked to poor postsurgical outcome included aneurysms larger than 12 mm, posterior circulation location, previous cerebral ischemic disease,

and aneurysmal symptoms other than rupture (such as mass effect). Conversely, only location in the posterior circulation and sizes greater than 12 mm were linked to postcoiling outcome. This second report, too, was criticized (21) on the basis that the mechanism for stratification of patients into untreated, surgically clipped, and coiled cohorts was not elucidated, and thus selection bias may have contributed to the disparate results between cohorts. A 2005 meta-analysis examining the risk of aneurysmal rupture as a function of location (22) found that, excluding the ISUIA data, across all sizes of aneurysm, the annualized risk of rupture was 1.3% for all anterior circulation aneurysms and 3.4% for the posterior circulation. There was no statistical difference among different locations within the anterior circulation, including the posterior communicating artery. When the ISUIA data was included, the annualized risk was 0.5% for the anterior circulation and 1.8 for the posterior circulation, lending credence to the notion that ISUIA rates of rupture were markedly lower than those published in other studies. The authors conclude that the ISUIA inclusion of posterior communicating segment aneurysms with those of the posterior circulation as possessing higher risk of rupture is misguided. A recent study using three-dimensional CTA to follow the natural history of unruptured cerebral aneurysms (23), with a mean follow-up of 17.7 months, found that aneurysmal growth or new bleb formation, thought to be precursors to rupture, were likelier the larger the initial size of the aneurysm. Thus, 2.4% of 2- to 4-mm aneurysms, 9.1% of 5- to 9-mm aneurysms, and 50% of 10- to 20-mm aneurysms grew . Location was an important factor as well, with 2 out of 5 basilar tip aneurysms and 0 of 43 middle cerebral artery aneurysms growing. Giant intracranial aneurysms (>2.5 cm) may become clinically apparent through progressive focal neurologic deficit secondary to mass effect, SAH, or thromboembolic events resulting from dislodgment of intraluminal thrombus contained within the aneurysm fundus. These aneurysms are disproportionately represented in most endovascular treatment seriesand, as discussed below, are often associated with increased treatment failure as defined by degree and persistence of occlusion. However, when successful, endovascular treatment may be beneficial not just in reducing the risk of rupture and thromboembolism, but also in ameliorating symptoms of mass effect. In a review of 26 patients with neurologic deficits related to aneurysmal mass effect treated by neurointerventional management, Halbach et al. (24) observed clinical improvement in 9 of 11 patients harboring giant aneurysms despite incomplete occlusion of seven aneurysms. As is the case for intradural aneurysms, for patients with intracavernous aneurysms, the decision to treat depends to a large extent on the clinical presentation. Signs and symptoms usually relate to cranial neuropathy secondary to the mass effect of the aneurysm. Kupersmith et al. (25) concluded that aneurysms of the cavernous carotid artery are rarely associated

Chapter 13: Endovascular Management of Intracranial Aneurysms

with life-threatening symptoms. In this and other studies (25–27), spontaneous clinical improvement in symptomatic cavernous carotid aneurysms has been observed following conservative management. Therapeutic intervention is usually reserved for patients with acute incapacitating cranial neuropathy or significant neuropathy failing to resolve with conservative management, intractable pain, signs of transient ischemic attack, SAH, or potential for life-threatening epistaxis as a result of bony erosion into the sphenoid sinus. In addition, cavernous aneurysms arising from the anterior genu of the carotid siphon (ring aneurysms), with MR characteristics suggestive of extension into the subarachnoid space, are usually treated aggressively because of the risk of SAH (26).

ENDOVASCULAR STRATEGIES Although microsurgical clipping has been historically advocated in the treatment of aneurysmal SAH and incidentally discovered aneurysms, endovascular therapy has evolved over the past decade and can now be offered as a viable alternative to direct open surgical treatment in many cases. Endovascular methods typically promote occlusive thrombosis of the aneurysm and may be broadly classified into two essential strategies: deconstructive and reconstructive approaches. Deconstructive procedures involve sacrifice of the parent vessel from which the aneurysm arises, while reconstructive approaches involve selective occlusion of an aneurysm with preservation of the parent artery (i.e., ‘‘reconstruction’’ of an aneurysm-free lumen).

DECONSTRUCTIVE APPROACH Most surgical series of carotid ligation for treatment of intracranial aneurysms were characterized by high morbidity and mortality related to ischemic stroke (28). In a survey of the surgical literature for carotid occlusion therapy, Scott et al. (29) reported a morbidity rate of 33% among 909 patients retrospectively reviewed. By comparison, most neurointerventional series of balloon occlusion therapy have reported transient neurologic deficits ranging from 7% to 10%, with permanent deficit seen in 1.5% to 5% of patients (30–32). These results have led to wider acceptance of endovascular occlusion in the treatment of symptomatic giant, fusiform or wide-necked aneurysms of the proximal carotid artery and vertebrobasilar circulation (31,33–35). Endovascular occlusion of the ICA offers several advantages to surgical clamping or ligation. Occlusion may be performed easily on an awake patient permitting continuous neurologic assessment during balloon test occlusion (BTO) prior to permanent vessel sacrifice. The collateral support to the compromised vascular territory may be evaluated angiographically at the time of BTO, providing an anatomic basis for confidence in the occlusion test (Figs. 1D, E and 2C,–F). The possibility of residual flow within the aneurysm may be evaluated for each potential site of parent vessel occlusion. This is particularly important

241

in aneurysms of the supraclinoid segment of the ICA and those affecting the vertebrobasilar circulation, which may be reconstituted by collateral vessels after occlusion therapy (36).

SAFETY AND EFFICACY Deconstructive procedures have been demonstrated to be highly effective and safe in the treatment of properly selected patients. Higashida et al. (32) reported results in 127 patients treated with detachable balloons and found the incidence of permanent morbidity to be 5.5%, with a mortality rate of 3.9%. Importantly, in this and other studies, following carotid occlusion, the size of the aneurysm often decreased (27), usually accompanied by resolution of the patient’s symptoms. Fox et al. (30) reported complete thrombosis in 50 of 65 cases of large and giant surgically incurable aneurysms of the anterior circulation treated by parent artery occlusion. Morbidity due to cerebral ischemia was seen in 12.3% of patients, with only one case progressing to permanent stroke and no deaths. While it is often assumed that post-occlusion stroke in patients previously passing BTO is related to cerebrovascular insufficiency, other mechanisms may account for the observed ischemic events. Anon et al. (37) have reported an angiographically documented middle cerebral embolus in one of two patients suffering from permanent neurologic complications after ICA occlusion, among 32 patients treated for cavernous carotid aneurysms who successfully passed BTO. Their data, as well as that of others (38), suggest the possibility that mechanisms other than insufficient collateral support may explain the development of neurologic deficits in patients after ICA sacrifice. While deconstructive approaches to aneurysms involving the supraclinoid ICA may be effective, proximal parent vessel occlusion is less reliable at inducing complete aneurysm thrombosis because of the possibility of persistent aneurysm filling through collaterals at the ophthalmic segment and circle of Willis. In the series of Fox et al. (30), 37 patients with cavernous aneurysms were found to have complete obliteration of the aneurysm as a result of proximal artery occlusion. However, among 21 supraclinoid aneurysms in their series, complete thrombosis occurred in only 10. Similar results have been observed for aneurysms arising within the vertebrobasilar circulation. Aymard et al. (39) reported complete thrombosis in only 13 of 21 inoperable posterior circulation aneurysms by endovascular occlusion of the vertebral arteries. Complications in these procedures were typically because of thromboembolic phenomena or inadequate collateral blood flow resulting in transient or permanent cerebral ischemic events.

THERAPEUTIC VESSEL OCCLUSION: TECHNIQUES Deconstructive procedures commonly employed in the management of intracranial aneurysms involve parent vessel occlusion with balloons (Fig. 1) or coils

242

Orbach et al.

Figure 1 (A) Axial T2-weighted and (B) lateral arterial phase angiographic images illustrating a giant cavernous segment aneurysm of the LICA. BTO of the LICA was performed prior to sacrifice of the parent vessel. (C) Lateral arterial phase image of the right common carotid artery after inflation of the test balloon confirming flow arrest in the LICA. (D) Frontal arterial phase image of the intracranial RICA runoff, and (E) lateral arterial phase image of a left vertebral/basilar runoff performed during BTO of the LICA, demonstrating collateral support of the left hemispheric circulation through the circle of Willis. The patient tolerated test occlusion of the LICA and the aneurysm was subsequently treated deconstructively by permanent occlusion of the parent LICA with detachable balloons (F). A second diagnostic catheter is useful in angiographically assessing potential collateral support to a compromised circulation during BTO. Abbreviations: LICA, left internal carotid artery; BTO, balloon test occlusion.

(Fig. 3). Refinements in this technique have included the deposition of coils within the aneurysm proper to further secure occlusion of the aneurysm fundus, particularly in situations where the aneurysm may be reconstituted by collateral circulation. The protocol often includes BTO in the initial angiographic assessment of potential collateral supply to the cerebral vascular territory placed at ischemic risk by the parent vessel occlusion. Such collateral supply may be derived through (1) the circle of Willis, (2) externalto-internal carotid artery anastomoses, at the ophthalmic, cavernous, and petrous ICA segments (Fig. 4), or (3) leptomeningeal collateral support over the convexities, depending on the proposed level of occlusion. The uninvolved cervical and cerebral vasculature should also be evaluated for signs of disease, i.e., additional aneurysms, vessel dissection, fibromuscular dysplasia, or other conditions that might mitigate against occlusion therapy. Patients in whom the collateral circulation, as demonstrated angiographically, appears adequate,

tolerance testing is performed by temporary balloon occlusion of the parent vessel at or near the site of proposed permanent vessel sacrifice (36). In cases of aneurysms arising from the ICA, a double lumen occlusion balloon catheter may be introduced alone or via a coaxial system through a larger guiding catheter positioned within the common carotid artery. The balloon occlusion catheter is introduced into the cervical segment of the ICA and then inflated under digital subtraction fluoroscopy, employing a roadmapped image of the ICA. In addition to systemic heparinization, local perfusion of the occluded segment proximal to the balloon is established through the guide catheter component of the coaxial system, while the segment distal to the occlusive balloon is perfused with heparinized saline through the distal lumen of the occlusion catheter. Following balloon inflation, vessel occlusion is verified either by administration of contrast through the distal lumen of the balloon occlusion catheter (resulting in a stagnant column of contrast within the occluded segment distal

Chapter 13: Endovascular Management of Intracranial Aneurysms

243

Figure 2 Angiographically failed BTO of the LICA. (A) Frontal arterial phase DSA image of the LICA demonstrating a recurrent widenecked aneurysm previously treated by surgical clipping. (B) A balloon was subsequently inflated within the LICA for test occlusion. Cerebral angiography of the RICA, performed during BTO of the LICA (C–F), was notable for angiographically inadequate collateral support across the anterior communicating artery characterized by poor cross opacification of the left hemispheric circulation and delayed angiographic progression. Note the increasing discrepancy in angiographic phase between right and left hemispheres evident from (C) to (F) suggesting reduced flow within the left anterior and middle cerebral territories. The left posterior communicating artery was atretic. Abbreviations: LICA, left internal carotid artery; BTO, balloon test occlusion; RICA, right internal carotid artery.

to the inflated balloon) or by angiographic assessment of ipsilateral common carotid runoff from the proximal guiding catheter. While the potential for collateral reconstitution of the compromised cerebrovascular territory may be inferred from dilution of the stagnant contrast column by wash-in of unopacified blood, this determination is better made using a second diagnostic catheter to angiographically assess collateral support during test occlusion of the index vessel. If carotid occlusion is not tolerated, sensorimotor dysfunction usually develops within a short time. While not universally accepted, a variety of semiquantitative measures of cerebral blood flow (PET, SPECT, and XeCT) have been used to increase the rigor of occlusion testing (38,40–42). Hypotensive challenge (29) with or without transcranial Doppler ultrasound monitoring has been advocated for assessment of marginal cases, although its power in predicting delayed ischemic complications has been questioned (43). In

recent years, the use of hexamethylpropyleneamine oxide (42) has made it technically simpler, easier, and safer to perform cerebral blood flow measurement during test occlusion by eliminating the need for blind inflation of the occlusion balloon and transfer of the patient with indwelling catheters for imaging purposes. Unfortunately, no method offers perfect specificity and sensitivity in predicting the outcome of permanent parent vessel occlusion, and all may be associated with complications. With experience, however, a false positive predictive morbidity rate of less than 5% should be achievable. Following uneventful test occlusion, the balloon is deflated, and permanent vessel sacrifice performed with detachable balloons or coils. The advantages of balloon occlusion include the immediacy of flow arrest and an overall reduction in procedure time. The primary disadvantage is the possibility of premature balloon deflation during placement of the initial

244

Orbach et al.

Figure 3 Reformatted CT angiogram demonstrating a mostly thrombosed giant fusiform aneurysm involving the distal left vertebral, VB junction, and proximal basilar arteries. Both posterior communicating arteries were atretic, so (A) surgical bypass from the right external carotid to posterior artery was performed (B) prior to deconstructive coil occlusion of the aneurysm and parent left vertebral and basilar arteries (C–F). (C) Unsubtracted image of the skull illustrating the coil mass. (D) Frontal arterial phase image of the bypass (white arrowhead ) runoff demonstrating retrograde opacification of the upper basilar artery to the level of the AICA segment (white arrow). (E) Frontal arterial phase DSA image of the right vertebral artery confirming occlusion of the aneurysm and VB segment with preservation of the right posterior inferior cerebellar artery (notched black arrowhead ) and radiculomedullary contribution to the anterior spinal artery (black arrow). (F) DSA left vertebral artery confirming occlusion of the distal left vertebral artery and aneurysm. Abbreviations: VB, vertebrobasilar; AICA, anterior inferior cerebellar artery.

balloon prior to secured flow arrest, which may lead to embolization of the deflated balloon. The likelihood of unintended balloon embolization can be reduced by timely placement of tandem balloons, or deployment of the initial balloon under conditions of flow arrest, employing a balloon occlusion guide catheter to arrest flow more proximally while the target vessel is permanently occluded. The risks of coil occlusion include increased time of the procedure, the possibility of coil embolization into the distal intracranial circulation, and the potential for thromboembolic events during the procedure prior to the full arrest of antegrade flow in the sacrificed vessel. These difficulties can in part be addressed by using an occlusion balloon guide catheter to arrest flow in the parent vessel proximal to the site of coil deposition during the procedure. It should be pointed out that permanent occlusion of major

vessels such as the ICA or vertebral arteries may require a large number of coils, increasing the expense of the procedure. For balloon occlusion of the ICA, a detachable balloon is usually advanced to the site of proposed deployment (36). In treating cavernous segment aneurysms, a common strategy is to use a trapping method, in which the first balloon is advanced, often with the aid of a microguidewire, to the segment of ICA just distal to the aneurysm neck. A second balloon is then placed either across the aneurysm neck or just proximal to the aneurysm, effectively trapping the lesion between the two inflated balloons. In cases complicated by cavernous aneurysm rupture (with accompanying carotid-cavernous fistula), proximal occlusion may result in cerebral ischemia or stroke as a result of vascular ‘‘steal’’ unless the accompanying fistula is first occluded (44).

Chapter 13: Endovascular Management of Intracranial Aneurysms

245

Figure 4 Skull base collaterals potentially confounding adequate BTO of the internal carotid artery. (A) Midarterial phase DSA image from a left common carotid runoff (lateral projection) after inflation of a balloon (black arrow ) within the midcervical segment of the left ICA during BTO. Injection of contrast at the left common carotid artery bifurcation reveals collateral reconstitution of the ICA beyond the occlusion balloon (notched black arrowhead ) through the vidian artery (partially obscured by the petrous bone, white arrowhead ) and the ophthalmic artery (white arrow ) through a meningoophthalmic branch (black arrowhead ). (B) DSA image of the left ICA (lateral projection) demonstrating angiographically identifiable meningohypophyseal (arrow) and vidian (notched arrowhead ) arteries. (C) Midarterial DSA image (frontal projection) during super-selective microcatheter angiography of the neuromeningeal trunk of the left ascending pharyngeal artery (microcatheter tip labeled with solid arrowhead ). Distal branches of the right ascending pharyngeal artery (notched arrowhead ) are retrogradely opacified across the midline through collaterals of the retroclival arcade (small double arrows). The left internal carotid artery (large single arrow) is opacified through collaterals between the neuromeningeal trunk and branches of the MHA. This ECA/ICA collateral route could provide a hemodynamically significant conduit for reconstituting the intracranial ICA circulation during BTO, unless the balloon was inflated over or distal to the MHA. Abbreviations: ICA, internal carotid artery; ECA, external carotid artery; BTO, balloon test occlusion; MHA meningohypophyseal artery.

After detachment of the initial balloon, additional balloons are usually deployed within the cervical segment of the ICA, with the most proximal occlusion balloon usually detached at the internal carotid origin effectively creating a stumpectomy. The need for placement of a stumpectomy balloon is somewhat controversial. At issue is the relative risk of thromboembolic events arising from turbulent flow and thrombosis at the unsecured internal carotid bulb, which may occur through external-to-internal carotid collaterals (45), versus cardiodynamic instability attributed to mechanical overdistension of baroreceptors with deployment of a balloon at this location. For giant aneurysms involving the supraclinoid segment of the ICA or the vertebrobasilar circulation, an understanding of the potential pathways for collateral reconstitution of the lesion (through the ophthalmic or posterior communicating arteries in supraclinoid ICA aneurysms or through collaterals at the cervical interspaces in distal vertebral artery aneurysms) will determine the level of parent vessel occlusion required. Deconstructive treatment of aneurysms arising from distal cerebral vessels is more complex and technically more challenging. Positioning the balloon microcatheter for temporary occlusion of the distal circulation is associated with increased risk of vessel rupture and relies far more on anatomic studies in the evaluation of the leptomeningeal circulation at the cortical level to assess potential reconstitution of the distal

vascular territory occluded by the balloon. The risk of future deficit may be assessed by superselective injection of sodium amytal, which will usually disclose a potential neurologic deficit related to the entire territory of the occluded vessel. This deficit generally provides an overestimation of the potential neurologic dysfunction, but it may nevertheless prompt a more conservative approach, depending on the degree of eloquence of the territory at risk. As opposed to balloon or macrocoil (0.035 inch) occlusion of larger proximal vessels, it is often preferential to use microcoils for distal occlusive procedures. In this context, refinements in microcoil technology, reflected in part by the development of the Guglielmi detachable coil (GDC) system (see below), have improved the accuracy of coil positioning and retrieval, dramatically increasing the safety of this approach. Deconstructive endovascular approaches can also be used in concert with surgical exploration in treating giant aneurysms associated with critical mass effect. Mounayer et al. (46) report on a giant anterior communicating aneurysm that presented with symptoms of bifrontal mass effect, whose treatment consisted of endovascular occlusion of both anterior cerebral arteries using a Trispan device and several coils (with leptomeningeal MCA collaterals supplying the affected territories), followed several days later by surgical aneurysmectomy. Others have reported similarly on the utility of temporary intraoperative

246

Orbach et al.

endovascular flow arrest in providing for a bettercontrolled surgical approach to large paraclinoid aneurysms (47).

RECONSTRUCTIVE APPROACH Reconstructive repair of aneurysms, i.e., approaches that aim to exclude the aneurysm sac from the arterial circulation while preserving the parent vessel, can take one of two general forms: (1) endosaccular, where the strategy is aimed at filling the aneurysm sac with embolic materials thereby promoting aneurysmal thrombosis, and (2) endoluminal, where the goal is not merely to treat the aneurysm sac, but rather to restore the vessel wall deficiency to its preaneurysmal state. Early endovascular reconstructive approaches were exclusively endosaccular. It quickly became apparent that selective occlusion of an aneurysm with preservation of the parent artery (i.e., ‘‘reconstruction’’ of an aneurysm-free lumen) could be performed by direct deposition of occlusion balloons within the aneurysm lumen. Despite several early reports of success using reconstructive balloon occlusion therapy (24,48–50), this technique as well as those employing pushable microcoils and liquid embolic agents (51) for intracranial aneurysm occlusion have been eclipsed by the safety and efficacy of detachable coil systems, exemplified by the GDC (52–54). The GDC was introduced for clinical trial to a limited number of centers in 1991 and received the U.S. FDA approval for treatment of patients with surgically unmanageable aneurysms in 1995. The recommended indications were subsequently liberalized in 2003, following publication of results from the International Subarachnoid Aneurysm Trial (ISAT) (55), to permit its use more broadly in select patients with ruptured cerebral aneurysms. In this study, clinical, neuropsychological, and angiographic follow-up was obtained, on a cohort of patients with ruptured aneurysms thought to be equally suitable for clipping or coiling, and randomly assigned to one or the other. Enrollment was prematurely terminated after determination of a demonstrably lower risk of death or dependency in the coiled subgroup (absolute risk reduction of 6.9%). The rebleeding rate was 2.6% in the coiled group versus 1% in the clipped group, but the morbidity was nevertheless lower. The publication of ISAT led to the release of a position statement by the American Society of Interventional and Therapeutic Neuroradiology (ASITN) (56), endorsing a neurointerventional consultation, wherever feasible, for every case of SAH. A follow-up of the ISAT cohort, published in 2005 (57), reported that the relative risk reduction of coiling versus clipping had in the interim increased to 7.4% and was maintained out to seven years of followup. The risk of epilepsy was significantly lower in the coiled group as well (RR 0.52). Despite the fact that the risk of rebleeding in the coiled group was higher (7 vs. 2 patients), as was the percentage of incompletely occluded aneurysms on follow-up angiograms (34% vs. 18%), the investigators concluded that the

follow-up study bolstered the initial ISAT conclusions regarding the potential superiority of endovascular treatment of acutely ruptured cerebral aneurysms, in cases where either treatment is a valid option. The advantage of detachable coils in treating cerebral aneurysms is related to four distinct properties: (1) the method of its delivery, attached to a delivery mandrel, enables the operator to determine and to change the placement of the coil prior to final deployment, thereby increasing the number of possible strategies for achieving ideal occlusion, (2) the intrinsic compliance of the coil helix permits adaptation to a relatively wide range of aneurysm morphologies, (3) the low thrombogenicity of the platinum coil strand allows coil retrieval and repositioning of suboptimally deployed coils, without undue concern for shearing of clot matter from the manipulated coil, and (4) availability in a wide selection of sizes and shapes from several manufacturers. There are in general two size classes of aneurysm coils, one employing platinum wire of 0.015-inch gauge for delivery through a 1.9- to 2.5-French (Fr) microcatheter and a second constructed of 0.010-inch gauge platinum coils, which may be delivered through 1.7-Fr microcatheters. The original platinum coils were available in various lengths ranging from 2 to 30 cm and were constructed to assume a helical configuration of 2- to 20-mm diameter in the unconstrained state, depending on the specific type of coil. Since their release in 1991, more sophisticated coil designs have become available, varying in their conformability, configurational complexity (2D, 3D, 360), and resistance to stretching during manipulation or retrieval. Later generations have included hybrid coils incorporating various bioactive substances and volumetrically expansile coatings (hydrogels). In most designs, the coil proper is attached proximally to a delivery mandrel (measuring *175 cm in length) through an intermediate coupling representing the potential detachment zone. By convention, the detachment zone for all coil makes is 3 cm distal to a radiopaque marker located on the delivery mandrel, which, when aligned with the proximal marker of the aneurysm microcatheter, serves to ensure proper location of the detachment zone outside the tip of the microcatheter. Once acceptably placed within the aneurysm, coils may be separated from the delivery stylette by electrolysis or temperature-dependent dissolution of a polymeric linkage, employing direct current (DC) of 1 to 2 mA, or by one of several hydraulic systems.

ANEURYSM COIL PROTOCOL At most neurointerventional centers, reconstructive treatment of cerebral aneurysms with aneurysm coils is performed under general anesthesia, which insures more controlled patient management and better imaging quality, particularly important for navigating difficult vascular anatomy and the superselective catheterization of small and complex aneurysms. A radial artery line is placed for continuous monitoring of

Chapter 13: Endovascular Management of Intracranial Aneurysms

arterial blood pressure and the activated coagulation time (ACT). In those patients with compromised cardiopulmonary function, or in patients presenting with acute SAH, central venous access may be helpful. This enables both monitoring of central venous pressure and placement of a Swan-Ganz catheter for monitoring of pulmonary wedge pressure in patients requiring subsequent ‘‘triple-H’’ (hypertensive, hypervolemic, hemodilutional) therapy. In the setting of acute SAH, it has been shown that the morbidity and mortality rate (40%) secondary to medical complications in the aftermath approaches that due to the primary neurologic event (58), and scrupulous attention to peri- and postprocedural neurocritical care is mandatory. After patient preparation, cerebral angiography is performed to obtain the anatomic information required for optimal planning of endovascular treatment. The angiographic assessment, which is frequently aided by 3D volumetric rendering of a rotational angiographic data set, includes (1) the size, shape, and orientation of the aneurysm, including correlation with transaxial imaging to assess the presence of intra-aneurysmal thrombus; (2) the size of the aneurysm neck in relation to that of the aneurysm fundus and parent vessel; (3) the size and morphology of the parent vessel from which the aneurysm arises; (4) potential collateral blood supply to the vascular territory supplied by the affected artery, including the presence of any anatomic variations; and (5) angiographic evidence of vasospasm or other vascular anomalies or disease states often associated with

247

intracranial aneurysms (AVMs, fibromuscular dysplasia, vasculitis, additional aneurysms, etc.). After obtaining the best working position, a road-map image of the parent vessel and aneurysm is obtained, and the aneurysm fundus selectively catheterized with a coaxially introduced 2-marker microcatheter (Fig. 5). The tip of the catheter may be shaped to conform to the anatomic configuration of the aneurysm–parent vessel complex or may be introduced unshaped, depending on the strategy employed by the operator. Regardless of approach, however, care must be taken to ensure that neither the microcatheter tip nor the guidewire forcibly interact with the wall of the aneurysm. The first coil selected should be the largest coil deployed and serves to frame the theoretical boundary of the aneurysm. Appropriately sized smaller coils are subsequently delivered to fill the interstices of the aneurysm fundus, and any coil inadvertently herniating into the parent vessel, or suboptimally positioned within the aneurysm, may be withdrawn and repositioned into a more optimal configuration prior to its detachment. As successive coils are deployed within the aneurysm, it may become increasingly difficult to visualize each new coil fluoroscopically as it is positioned within the deposited coil mass. Precise placement of additional coils, however, may continue through alignment of the delivery mandrel’s radiopaque marker, with the proximal platinum marker on the microcatheter (outside the aneurysm), 3 cm from its tip.

Figure 5 (A) Reformatted rotational angiographic image of the intracranial LICA circulation demonstrating a large left middle cerebral artery aneurysm (white arrow) (other aneurysms have been clipped). (B, C) Successive real-time road-mapped images of the intracranial LICA circulation (frontal projection) during deployment of the first platinum coil. In (C), the coil has been fully delivered into the aneurysm as indicated by the alignment of the delivery mandrel marker with the proximal microcatheter marker band (notched arrow). (D) Posttreatment rotational angiographic image confirming near complete coil occlusion of the aneurysm. Abbreviation: LICA, left internal carotid artery.

248

Orbach et al.

Using live digital subtraction techniques, the deployment of each new coil can be individually followed. After the desired coil position has been achieved, DC is applied to the proximal end of the delivery wire, resulting in separation of the coil from its guidewire over an interval of 2 to 60 seconds, depending primarily on the specific manufacturer’s design. The delivery mandrel is then removed, and if required, additional coils are deployed through the indwelling microcatheter, usually until a dense coil mesh within the aneurysm has been achieved. The deposited coil mass results in stagnation of flow within the aneurysm fundus, promoting thrombosis and isolation of the aneurysm wall from the arterial circulation. The extent to which DC-induced electrothrombosis participates in the formation of the aneurysmal clot during coil deployment has not been delineated in vivo. Initial evidence from in vitro models and animal experiments suggested that to some degree thrombus does form around the activated platinum anode in proportion to the current magnitude (54), and further, that aneurysmal thrombosis is followed by growth of a neo-endothelium over the base of the aneurysm (59). However, subsequent findings in humans (60) and bifurcation aneurysm models in experimental animals (61) have raised questions regarding the stability of the coil-thrombus complex and its histologic evolution. Given that thromboembolic complications constitute the major risk to patients undergoing coilendovascular occlusion of intracranial aneurysms, anticoagulation during treatment is employed at most centers. Although the precise regimen varies across institutions, generally anticoagulation with intravenous heparin or heparinization combined with antiplatelet therapy is begun during the endovascular procedure and is continued for 12 to 48 hours. However, a controlled study examining the efficacy of anticoagulation and/or antiplatelet therapy in this setting has not been done.

FOLLOW-UP AFTER TREATMENT Radiographic follow-up of treated aneurysms usually involves plain-film assessment of the coil pack configuration in conjunction with serial angiograms, usually at yearly intervals during the first years, supplemented by MRA (62,63) for long-term assessment of treatment stability. In most cases, 3D time-offlight images are adequate, with the addition of a saturation band in cases of high perianeurysmal signal allowing the distinction between slow flow and blood products. Contrast-enhanced MRA is further useful (64), increasing signal-to-noise ratio and helping distinguish slow flow at the coiled aneurysm base, particularly in the setting of giant aneurysms. The application of MRA in the evaluation of aneurysm recanalization was reported to have a sensitivity of 97%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 94.7%.

CLINICAL RESULTS: RECONSTRUCTIVE APPROACH WITH ANEURYSM COILS It must be borne in mind that most series that appear in the literature describing outcome after endovascular aneurysm treatment include patients unsuitable for surgical clipping, and are thus biased by inclusion of a higher proportion of posterior circulation aneurysms and of patients with higher Hunt and Hess grade than are comparable surgical series. A 2004 study (65) reviewed the clinical outcome among patients who presented with Hunt and Hess grade IV or V SAH and who were coiled between days one and seven from their ictus. Patients with large intracranial hematomas, with evidence of large volume infarcts, with extensive midline shift, or with evidence of brain stem damage were excluded. No balloons or stents were used. Outcome was assessed at six months and was favorable (i.e., at most moderate disability) in 48% of the cohort, a figure that is comparable to results from early neurosurgical clipping and significantly improved over the natural course of such patients (66). A retrospective 1998 study (67), comparing the frequency of vasospasm after SAH of similar severity treated either with surgical clipping (19 patients) or endovascularly (18 patients), showed a notably higher tendency toward vasospasm in the surgical group, with 14 of the 19 patients requiring triple-H therapy and three requiring mechanical angioplasty and intraarterial pharmacologic treatment of vasospasm. Among the endovascular group, four patients developed clinical signs of vasospasm, all of whom responded to elevation of blood pressure. Koivisto et al. (68) prospectively examined the question of endovascular versus surgical treatment of acutely ruptured aneurysms (i.e., within 72 hours of ictus) by randomly assigning 109 consecutive patients to either modality. Clinical and neuropsychological follow-up was performed at 3 and 12 months postprocedure, MRI of the brain at 12 months, and follow-up angiography after clipping for the surgical group and at 3 and 12 months for the endovascular group. The following conclusions emerged: (1) outcome was overwhelmingly dependent on the severity of the clinical presentation of the initial SAH, regardless of treatment modality; (2) there were no significant differences in outcome between the two groups on clinical or neuropsychological grounds, with no important clinical improvement between months 3 and 12, but with meaningful neuropsychological improvement during that time; and (3) while there were no significant differences in the rate of clinical vasospasm between the groups, MRI showed more frequent ischemic lesions in the territory of the ruptured aneurysm for the surgical group than for the endovascular group. A retrospective study of 327 patients who presented with SAH and who were treated with GDCs (69) found no statistical difference in outcome when patients were stratified according to time lag between the ictus and the procedure. The authors thus recommended treatment as early as possible to circumvent rebleeding.

Chapter 13: Endovascular Management of Intracranial Aneurysms

Cloft and Kallmes (70), in a meta-analysis, demonstrated that the risk of endovascular intraprocedural aneurysmal perforation was significantly higher in the setting of acutely ruptured aneurysms (4.1%) than in cases of elective coiling (0.5%). The associated risk of permanent neurologic morbidity or death was significantly higher as well (38% vs. 29%, respectively). Interestingly, the morbidity and mortality rate for perforations caused by coils (39%) was similar to those caused by microcatheters (33%), while perforations caused by microguidewires had no associated morbidity or mortality. There was a tendency toward higher morbidity and mortality among patients receiving intraprocedural IV heparin than among those not receiving heparin, but this did not achieve statistical significance. However, in cases of perforation, the investigators recommend reversal of anticoagulation, placement of a ventriculostomy (if not already in place), and use of an additional microcatheter to complete the embolization, leaving the perforating device in place in the interim. Needless to say, in cases where the aneurysm location makes surgical approach difficult or impossible, endovascular therapy may be the only viable option. A retrospective study (71) of 150 basilar tip aneurysms treated with GDCs revealed, after a mean angiographic follow-up period of 9.8 months for unruptured aneurysms and 13.7 months in cases of rupture, a rebleed rate of 3.3% for the ruptured group and 4.1% for the unruptured group, with permanent deficits secondary to stroke of 5% and 9%, respectively, and a periprocedural mortality rate of 2.7%. These rates contrast with mortality rates of 23% in conservatively managed ruptured basilar tip aneurysms and 12% in unruptured cases over similar lengths of follow-up. The authors conclude that endovascular therapy is certainly warranted in cases of rupture, while appropriate management of cases of unruptured basilar tip aneurysms still needs clarification. Similarly, Tateshima et al. (72) reported their single-center experience with 75 basilar tip aneurysms treated with GDCs and found that immediately postprocedure, 85.3% of the aneurysms were completely occluded, 9.3% partially occluded, and 5.3% could not be treated because of technical anatomic difficulties. The procedure-related morbidity and mortality were 4.1% and 1.4%, respectively, and 91.3% of the patients treated remained either neurologically intact or unchanged from their state at presentation. Johnston et al. (73) calculated total costs, in both morbidity and financial terms, in comparing endovascularly treated versus surgically clipped aneurysms at academic medical centers. They followed the rate of adverse outcomes (defined as in-hospital deaths and discharge from the hospital to a nursing home or rehabilitation center), length of stay, and cost of hospitalization. They took into account the fact that the endovascular patients occasionally required recatheterization and repeated treatment because of incomplete aneurysm occlusion, while the surgical patients rarely required follow-up treatment. The cost of endovascular treatment was calculated as the sum of the costs for all admissions for a given patient required to treat a given

249

aneurysm, and adverse outcomes for the endovascular patients were tallied as the worst outcome associated with any one of the endovascular procedures performed to treat that aneurysm. Even so, the results significantly favored endovascular therapy: the rates of adverse outcomes were 18.5% for the surgical group and 10.6% for the endovascular group, the rates of inhouse deaths were 2.3% and 0.4%, the lengths of stay were 9.6 and 4.6 days, and the mean charges were $43,000 and $30,000, respectively. These results are all the more remarkable in that this was a cohort rather than a randomized study and, as pointed out above, during this era, patients directed to endovascular therapy tend to be poor surgical candidates, either because of concomitant systemic medical conditions or because of surgically inaccessible aneurysms. A similar, more recent comparison of neurologic outcome (measured as change in Rankin scale) was performed by Johnston et al. (74) for aneurysms that were retrospectively judged by blinded neuroradiologists and neurosurgeons to be approachable via either technique. Twenty-five percent of the surgically treated group had a change in Rankin scale of 2 or more (signifying significant disability), while only 8% of the endovascular group did so, and again, total length of stay and hospital charges were greater for the surgical group (7.7 vs. 5 days, and $38,000 vs. $33,400, respectively). Most importantly, at a mean follow-up of 3.9 years postprocedure, 34% of the surgically treated patients reported a persistent new deficit since treatment, while only 8% of the endovascular group did so. Importantly, it was only after one year that 50% of the surgical cohort reported returning to their baseline condition, while for the endovascular group the comparable elapsed time was 27 days.

RECURRENCE IN WIDE-NECKED ANEURYSMS: TESTING THE ENDOSACCULAR APPROACH Despite the encouraging results from ISAT and other case series (75), the surprisingly low reported frequencies of up-front aneurysm occlusion (76–79) and the prevalence of posttreatment recurrences (76–81) among coil-treated aneurysms have become serious obstacles to the widespread acceptance of endovascular treatment as definitive therapy. This is particularly (and paradoxically) true in the case of large, complex aneurysms for which the treatment was initially indicated. Aneurysm recurrence following coiling has generally been thought to proceed through two mechanisms: (1) recanalization (acute or delayed) of the coiled aneurysm fundus, resulting from an underlying instability of the intra-aneurysmal coil-thrombus complex, and/or (2) progressive absolute aneurysm growth from either an unsecured niche of an incompletely coiled aneurysm or an intrinsic (initially occult) deficiency in the wall of the perianeurysmal parent artery. Such features as completeness of aneurysm lumen obliteration, tightness of packing of the coils, aneurysmal geometry and location (e.g., dome-to-neck ratio of at least 2 and location of the aneurysm away from highflow arterial bifurcation sites) are thought to be strongly

250

Orbach et al.

prognostic of long-term success (82). Nevertheless, there have been reports of recurrent aneurysm with SAH; in a notable example, 18 months posttreatment, at the site of an unruptured aneurysm that was angiographically occluded after coiling and that remained occluded at six-month angiographic follow-up (83). The extent of the problem, first suggested in early reports (76), has since been elucidated in multiple studies. Raymond et al. (81), in a retrospective analysis of data collected from 466 patients with 501 aneurysms, observed a strong correlation between aneurysm dimensions and neck size and the prevalence of posttreatment recurrence. Among 383 patients with follow-up angiograms, recurrence, defined as major by the authors, was found in 20.7% (at a mean angiographic follow-up of 17 months). When analyzed by aneurysm morphology, recurrence (of all degrees) was observed in 50.6% of large aneurysms as opposed to 21.3% of small aneurysms (<10 mm), and 52.3% of wide-necked aneurysms (>4 mm) versus 23.7% of aneurysms with small neck size. More recently, a 2005 study reviewing long-term follow-up (mean 36 months) on 705 patients with ruptured aneurysms treated with GDCs (84) reported a technical feasibility rate of 96.9%, with failures related either to vessel tortuosity or to coil herniation outside the aneurysm neck. Complete occlusion (>99% of the volume) was achieved in 72% of cases, subtotal (95–99% embolization) in 25%, and incomplete (<95%) in 2.4%. For completely embolized aneurysms, 72.4% remained stable, 10.5% showed a ‘‘small recurrence’’ and were followed with MRA and DSA, and 2.8% showed significant coil compaction and recanalization and were retreated. For cases with subtotal occlusion, 42.1% showed no change, 29.2% had spontaneous thrombosis resulting in complete occlusion, and 7.6% had significant recanalization and were retreated. For patients with incomplete initial occlusion, 37.5% had repeat coiling to attempt improvement, 31.3% had spontaneous improvement from thrombosis, and 25% died. The only parameter examined that was predictive of occlusion outcome was the initial size of the aneurysm, with total occlusion achieved in 74% of aneurysms smaller than 10 mm, and 50% in aneurysms larger than 15 mm. In terms of the proximate postcoiling outcome, Kole et al. (78) found 27% of aneurysms with large remnants immediately after coiling. The same authors further reported an increased remnant size in 19.1% of patients with a mean angiographic follow-up of 18.2 months, 14.5% requiring aneurysm recoiling. This last statistic is of interest because two deaths occurred among the patients undergoing retreatment, illustrating an often-ignored source of risk to which patients with unresolved aneurysms are exposed. Ironically, our perception of incomplete treatment or aneurysm recurrence after endosaccular coil therapy has been partly obfuscated by the rapid evolution and assimilation of new coil technologies (3D coils, polymer- and hydrogel-coated coils, and ultrasoft coils) (85–88), liquid embolic agents (89–92), and adjunctive techniques (balloon remodeling) (93–95), each purported to improve treatment efficacy, whether

or not these new innovations actually contribute to a more secure endovascular outcome. In a provocative study examining the impact of endovascular advancements during the decade following introduction of the GDC, Murayama et al. (79) reported results from 11 years experience in 818 patients treated at University of California Los Angeles (UCLA). Analyses of treatment outcomes were stratified by aneurysm morphology and date of treatment. Angiographic results from those who were treated in the first half of the decade (1990–1995) were compared with those from the second half (1996–2002), after modified specialty coils and newer adjunctive techniques became available. The early group comprised 230 patients harboring 251 aneurysms, while the more recently treated group represented 588 patients harboring 665 aneurysms: 49.4% of the patients presented with acute SAH, while 41.8% had unruptured aneurysms. Angiographic follow-up ranged from three months to eight years (mean 11 months). Ironically, despite the intervening improvements in devices and technique, recurrence rates among the larger aneurysm subgroups treated since 1996 [37.7% (large) to 52.9% (giant)] were not statistically different from those in the first half-decade [33.3% (large) to 63% (giant)]. Furthermore, recurrences were also found in 18.2% of small wide-necked aneurysms for the years 1996 to 2002. These results, unfortunately, did not capture the effects of more recent innovations designed to improve volumetric packing of the aneurysm sac (with hydrogelcoated coils or liquid embolic agents) or to increase the effectiveness of the coil mass in promoting maturation of the intra-aneurysmal thrombus (with bioactive coatings). However, subsequently publicized, unpublished results from the Microvention-sponsored HEAL (AANS-CNS Joint Section-ASITN 2005 Annual Meeting, New Orleans, Louisiana, U.S.) and Boston Scientific– sponsored MATRIX-ACTIVE (American Society of Neuroradiology, 2004 Annual Meeting, Seattle, Washington, U.S.) trials, in addition to early published reports of clinical experience with polyglycolic acid– lactide copolymer (PGLA)-coated coils (87), have been less than wholly reassuring, implicating a potential deficiency in coil-dependent endosaccular approaches to the treatment of these aneurysm subtypes, at least in terms of achieving a stable anatomic result. Additionally, early experience in the MATRIX Registry seemed to confirm the fundamental importance of the aneurysmal coil mass (the density and completeness of coil packing), and by implication its effect on intraaneurysmal blood flow to the ultimate recanalization outcome, supporting the importance of a stable hemostatic environment in enabling the bioinducible potential of the PGLA jacket.

CONDITIONS FAVORING SUCCESSFUL ENDOVASCULAR ANEURYSM TREATMENT Three conditions seem necessary for durable endosaccular occlusion of cerebral aneurysms with coils. (1) Effective hemostasis must be established throughout the aneurysm and sustained over some critical

Chapter 13: Endovascular Management of Intracranial Aneurysms

interval. (2) The thrombus formed within the aneurysm as a result of coil-induced intra-aneurysmal hemostasis must mature during this critical interval (ideally undergoing organization into a fibrointimal scar). (3) The uniformity and stability of the coil-thrombus complex at the aneurysm base must be biomechanically sufficient to support neointimal overgrowth of the aneurysm neck defect. While these requirements are usually satisfied during endosaccular coiling of the idealized smallnecked aneurysm (favorable neck/fundus ratio), reconstruction of the larger aneurysm neck (which in certain dysplastic aneurysms encompass >1808 of the cross-sectional vessel circumference) (Fig. 1) is technically more challenging and frequently confounded by inadequate coiling of the aneurysm base. This situation has led to the adoption of adjunctive methods such as balloon remodeling and the complimentary use of endoluminal devices for more robust reconstruction of the neck defect in such aneurysms (94–101). The long-term consequences of subtotal aneurysm coiling have not been studied scientifically; however, they may be inferred from several single-center (81) and multicenter (55,102) series, which suggest that incomplete coiling of the aneurysm neck increases the likelihood of recurrence and may be a factor in delayed (Sluzewski, 2005) posttreatment rehemorrhage— estimated to range between 0.2% and 0.3% per year for previously ruptured aneurysms treated with coils. The collective factors responsible for recurrence (103–106), theoretically, can be divided into factors operational in incompletely occluded aneurysms and factors responsible for recanalization of aneurysms initially occluded at the time of coil treatment. For complex-neck, large, and giant aneurysms, recanalizations frequently result from unintentional or deliberate undercoiling of the neck region and involve a sequence of events leading to lysis and remodeling of incompletely organized intra-aneurysmal thrombus and coil compaction, with or without true continued growth of the aneurysm. Once formed, the fate of the intraaneurysmal coil-thrombus complex depends on a number of factors, including the coagulative disposition of the specific patient, the coil composition (i.e., surface texture, coatings, and charge density), the completeness of aneurysm packing, and, importantly, the degree of sustained aneurysmal hemostasis. Although coil packing density has been correlated anecdotally to stable aneurysm occlusion, with bare metallic- or polymer-coated coils, coil packing densities of treated aneurysms usually are significantly less than 40% by volume, even in ideally packed aneurysms (85). It is therefore likely that the effectiveness of coils in treating aneurysms is dependent in large part on the degree by which intra-aneurysmal flow is reduced, and, correspondingly, to the inherent stability of the intraaneurysmal coil-thrombus complex.

BALLOON REMODELING In an effort to facilitate coil treatment of wide-necked aneurysms, Moret et al. (18,94,107,108) developed a method whereby a small balloon occlusion microcatheter

251

is intermittently inflated within the parent artery across the aneurysm neck to provide structural support against which to deploy coils within the aneurysm through a second microcatheter positioned within the aneurysm fundus. In providing this barrier at the aneurysm neck, the inflated balloon allows successively deployed coils to assume an increasingly stable conformation within the aneurysm, while preventing their herniation into the parent vessel. Flow is reestablished in the parent artery after each balloon deflation (Fig. 6). In a review of their experience, Moret et al. (94) reported stable angiographic occlusion in 20 of 21 completely occluded broad-necked aneurysms (anterior and posterior circulations) among patients with at least four-month angiographic follow-up (mean, 13 months). These results approach those obtained with small-necked aneurysms coiled directly. Furthermore, when analyzing common complications related to the GDC technique (thromboembolic events, etc.), these authors concluded that their complication rate employing balloon-assisted coiling was no greater than that experienced with primary (unassisted) GDC treatment. Similar results have been reported by Malek et al. (109), Lefkowitz et al. (93), and Nelson and Levy (95), who demonstrated angiographic aneurysmal occlusion at a mean follow-up of 19 months, in 17 of 20 patients who had undergone balloon-assisted embolization of wide-necked aneurysms. Although promising, several technical concerns have arisen. The procedure may be performed via a single groin site, employing a 6-Fr sheath (Shuttle, Cook Inc., Bloomington, Indiana, U.S.) or a single 7- or 8-Fr guide catheter, or using two guide catheters introduced through separate groin sites. This may lead to increased frequency of groin complications, particularly in centers employing a regimen of 24 to 48 hours of anticoagulation after coil treatment. While groin sheaths may safely be left in place over the 24 to 48 hour posttreatment period that the patient is anticoagulated, several hemostatic devices are available to facilitate early groin sheath removal and may be useful to avoid complications of the groin puncture site in patients who are anticoagulated or who are uncooperative following the procedure and in whom it is desirable to remove the sheath prior to normalization of the partial thromboplastin time to avoid iliofemoral vessel injury or extensive hemorrhage. Additionally, the increased number of guidewires and microcatheters necessary for balloonassisted coiling potentially complicate the technique, making vessel dissection, rupture, and thromboembolic complication likelier. While intracranial balloon angioplasty has been used extensively in the treatment of vasospasm with acceptable complication rates (110), inflation of the microballoon in the vicinity of an aneurysm neck is known to be associated with increased risk of vessel or aneurysm rupture. Precise control of balloon placement and inflation is therefore paramount in using this technique safely. The duration of balloon inflation must be matched with the distal collateral supply to the occluded vascular territory to avoid the development of cerebral ischemia during treatment of the aneurysm.

252

Orbach et al.

Figure 6 (A) Lateral DSA image of the LICA illustrating a large wide-necked ophthalmic segment aneurysm. (B, C) Unsubtracted images during coiling of the aneurysm with balloon assistance, which facilitates curvilinear reconstruction of the anterosuperior vessel wall across the neck deficiency, evident after balloon deflation (D); usually not possible in the absence of an assist balloon. (E) Posttreatment angiography of the LICA confirms occlusion of the aneurysm. Abbreviation: LICA, left internal carotid artery.

In a retrospective review of 49 patients undergoing temporary occlusion of the MCA during aneurysm surgery, Lavine et al. (111) concluded that temporary occlusion times of 10 minutes or less were safely tolerated in the majority of patients. Additionally, the risk of delayed coil herniation into the parent artery must be considered, particularly in those cases in which coils placed subsequent to the framing coil are deployed without further balloon protection. Finally, the risk of thromboembolism associated either with temporary flow arrest in the parent artery or with the increased exposed coil surface at the aneurysm neck may require more aggressive anticoagulation during and after the procedure.

ACHIEVING DENSE ANEURYSM PACKING: ALTERNATIVES TO BALLOON REMODELING Refinement and improvement of the armamentarium for endovascular treatment of aneurysms proceed at a rapid clip: bridging devices other than balloons have found use in improving the effectiveness and safety of aneurysm coil packing (112) and biologically active materials are being incorporated into the surface of the coils to promote healing and degree of occlusion (6–8,85,88,113). Novel microcatheter approaches to

otherwise inaccessible aneurysms, such as retrograde via a major communicating vessel (114) and the incorporation of 3D digital subtraction angiography in the assessment of aneurysmal conformation (115), have broadened the range of aneurysms that are amenable to endovascular treatment. Three-dimensional coil shapes may be useful in this regard as well. A 2004 study (116) on the use of 3D coils in the treatment of wide-necked (>4 mm) aneurysms (excluding giant aneurysms) showed similar success in achieving complete occlusion of narrownecked (72%) and wide-necked (68%) aneurysms, so long as one or more 3D coil was used in first framing the aneurysm. Moreover, of the 160 aneurysms coiled using this technique, balloon remodeling was needed in only two cases. In terms of bioactive agents, Kallmes et al. (117), in a murine model, showed increased fibroblast proliferation and collagen formation in coils that had been coated with fibroblasts containing multiple copies of the gene-encoding fibroblast growth factor (FGF), as compared with control coils (at 14 and 35 days). Others (118) have pointed out that coated coils may be an excellent means by which to deliver targeted gene therapy aimed at preventing aneurysm rupture, either by way of antibody-tethered viruses expressing gene products that promote vascular

Chapter 13: Endovascular Management of Intracranial Aneurysms

healing or by way of radiation-inducible promoters, injected intravenously, being activated by radioactive coils. Moreover, it has been demonstrated in a canine model (119) that the implantation of radioactive (86) P ions onto standard platinum coils inhibits recanalization of coiled aneurysms, and clinical feasibility of the technique has been suggested by a small clinical trial (120). Clearly, questions about long-term stability and potential adverse effects of such treatment remain to be clarified. A drastically different approach toward endosaccular aneurysm obliteration relies on the use of liquid embolic agents (89–92). Biophysical models have demonstrated that coils fill only 30% to 40% of the volume of aneurysms (121,122) [possibly increased to as high as 73% with hydrogel-coated coils (85)], and there may be residual flow even within tightly packed aneurysms. Moreover, in recurrent aneurysms, coils have been demonstrated to compact over time, occupying less volume than when originally placed. Liquid embolic agents theoretically permit more efficient and complete endosaccular aneurysm filling than is possible with coils. Several are under active development, such as calcium alginate (123), a nontoxic emulsifier and Onyx. Onyx is a liquid embolic agent originally developed for use in AVMs. The authors of the CAMEO study (90) reported that with Onyx they achieved a higher-occlusion rate, better clinical outcome, and rate of adverse effects comparable to more traditional endovascular techniques in treating large or giant aneurysms. This was corroborated in a 2005 study of Onyx in wide-necked large or giant carotid aneurysms (124), where total occlusion was achieved in 81% of aneurysms and was maintained at follow-up angiography in 91%. Complications included two stenoses and one occlusion of the involved parent internal carotid artery. Four cases had presented with cranial neuropathy secondary to mass effect, and two of these patients improved following treatment. Of the initially asymptomatic patients, 14/15 remained so at follow-up, while one experienced a transient ischemic attack. A recent report (125) of chemical meningitis in two patients following placement of both hydrogel and Matrix coils raises the specter of an exuberant inflammatory response to this particular combination. Both patients responded well to immunomodulation with steroids.

SURGERY FOLLOWING PARTIAL EMBOLIZATION AND VICE VERSA Although technically challenging, it is possible to surgically clip aneurysms that have been partially coiled. In a recent study (126), patients with SAH had initially undergone endovascular embolization after presentation with Hunt and Hess grade III and IV hemorrhages or medical comorbidities that precluded surgery. The average time interval between embolization and surgery was 11 months. Patients had intraoperative angiography, and all were followed with serial contrast-enhanced MRA. All aneurysms were found at surgery to have some coils extruded through

253

the aneurysm wall into the extravascular space, and all were associated with a translucent membrane that made removal impossible; as a result, the coils were left in place. Eighteen patients underwent attempted clipping, in 15 of whom complete aneurysm exclusion was successful, while 3 patients required aneurysm wrapping as an alternative. The authors recommend a dome-to-neck ratio of at least 2 to make postembolization clipping feasible. Others have reported similar results (127,128). However, balloon remodeling and/or stenting as a second treatment after coil embolization is preferable to surgery, if possible. A report (129) on the morbidity and success rate of repeat coiling for aneurysms found to be incompletely embolized at sixmonth follow-up showed no complications, and complete occlusion was achieved in 76% of cases. Conversely, others (130–132) have reported on the feasibility of endovascular treatment of an aneurysm remnant after partial surgical clipping. Repeat surgery is often technically difficult and poses significant morbidity and mortality risk because of perianeurysmal scarring (133). In this setting, endovascular occlusion may represent a potentially valuable option.

ENDOLUMINAL RECONSTRUCTION: THE EMERGENCE OF ADJUNCTIVE STENTSUPPORTED COIL EMBOLIZATION OF CEREBRAL ANEURYSMS The feasibility of a combined endoluminal-endosaccular aneurysm treatment was first established in a dog aneurysm model (100) and, subsequently, confirmed by several case reports (99,101) and small clinical series (96,97), documenting results initially with balloonexpandable stainless steel stents and following the introduction of Neuroform (Boston Scientific-Target Therapeutics, Fremont, CA) and Leo (Balt, Montmorency, France), with self-expanding microstents. A mathematical model (134) demonstrated the drastic change in intra-aneurysmal flow pattern that followed placement of an asymmetric stent, with inflow dramatically reduced in comparison to the pattern seen following coil embolization. The rationale for adjunctive stenting in the treatment of wide-necked cerebral aneurysms relies on three effects. (1) The uncoupling of momentum exchange between the parent artery and aneurysm. This effect enhances the flow disruptive influence of the intra-aneurysmal coil mass, diminishing intra-aneurysmal flow and increasing the mean circulation time through the aneurysm fundus (135–138). The net effect is the induction of more profound hemostasis within the aneurysm, contributing hypothetically to intra-aneurysmal conditions in which recanalization is less likely. (2) The subintimal incorporation of the stent into the parent vessel wall. The mural integration of the stent into the parent artery (Fig. 2) modifies the viscoelastic properties of the perianeurysmal vascular segment, reinforcing the parent artery at

254

Orbach et al.

the neck margins and potentially reducing the likelihood of recurrent aneurysm growth from the neck region. (3) Neck-bridging barrier effects. These effects create a structural boundary across the aneurysm neck. The stent-imposed scaffolding facilitates more complete endosaccular treatment, together with neck region coils, provides a more organized substrate to support neointimal growth over the aneurysm neck. Many groups deploy a stent over the aneurysm neck and then introduce the microcatheter through the stent interstices into the aneurysm sac to place coils, which is in accordance with advantage 3 listed above. However, it should be noted that advantages 1 and 2 above can be achieved even if the stent is deployed after aneurysm packing by coils. This approach may be particularly valuable in the setting of SAH, where pretreatment with antiplatelet agents is not desirable. The strategy in these cases, therefore, is to secure the aneurysm acutely and, subsequently, deploy a stent several weeks later, following adequate premedication.

NEUROFORM: CLINICAL RESULTS While early experience with stent-supported coil endosaccular treatment of wide-necked aneurysms has been promising (96,97,139), long-term angiographic evaluation of the synergy expected from such combined endovascular therapy is lacking. Examination of the existing literature on the topic reveals significant variation in the reporting of outcome milestones and imaging follow-up in addition to the wide variety of cardiac stents used prior to the introduction of the flexible self-expanding Neuroform (Boston Scientific) and Leo (Balt) devices. In a review of 50 patients with wide-necked cerebral aneurysms undergoing attempted Neuroformsupported coiling, Lylyk et al. (140) reported immediate occlusion rates between 84.6% and 87.5%, depending on whether the stent preceded coiling or was placed after aneurysm coiling, respectively. Their results are particularly encouraging when compared to the immediate occlusion rate of 40.4% for large aneurysms reported by Murayama et al. (79), and in line with initial occlusion rates reported for balloon remodeling techniques in other series of broad-necked aneurysms (94,95). Fiorella et al. (97) reported, in their 20-month prospective study of 61 aneurysms undergoing stent-supported coiling with Neuroform, complete or near complete aneurysm occlusion in 28 (45.9%) and partial (presumably analogous to subtotal) occlusion in 33 (54%) of patients. Follow-up angiography or MRA was available at a median reevaluation period of four months. This series included a higher percentage of smaller aneurysms with small necks but unfavorable neck/fundus ratios and a larger number of ruptured aneurysms for which they advocated a conservative, staged approach in which the aneurysm is deliberately undercoiled during the initial therapeutic setting. Nelson et al. (141) recently reported their experience with Neuroform-supported endosaccular

treatment of select complex wide-necked aneurysms and assessed the anatomic stability of this combined endoluminal-endosaccular treatment over mediumterm clinical and angiographic follow-up in 16 patients. The cohort included one giant aneurysm (>25 mm diameter) and 15 large aneurysms (ranging in size from 11 to 22 mm), all of which exhibited neck sizes equaling or exceeding 7 mm in linear dimension along the parent vessel (ranging from 7 to 14 mm). There were no treatment-related deaths or clinically evident neurologic complications. A single technical complication occurred involving transient nonocclusive stentassociated thrombus, which resolved without clinical sequelae after the administration of intra-arterial abciximab. Follow-up angiography was between 11 and 24 months posttreatment. With the exception of three patients, all treated aneurysms were occluded at angiographic reevaluation. Clinical follow-up averaged 29 months. All patients had excellent clinical outcomes with the exception of two patients. The first experienced recurrence of third and sixth nerve palsies, while the second had minor residual short-term memory deficit related to an index SAH, which prevented fulltime return to work. There was one suicide-related death 22 months after treatment.

ADVANCED STENT TECHNIQUES For complex aneurysms, where the parent vessel– aneurysm interface cannot be sufficiently resolved by 2D fluoroscopy/angiography, staged approaches using balloon remodeling to coil the aneurysm prior to stent placement may provide more confidence in coiling neck domains overlapping the parent vessel (Fig. 7). Balloon remodeling will likely continue to be useful in facilitating more complete initial occlusion in complex ruptured aneurysms, permitting later staged stent placement once antiplatelet agents can be given safely. Overlapping stents may be useful to increase metallic coverage over the aneurysm–parent vessel interface. By constraining full expansion of the internal stent through oversizing, one further increases the mesh coverage throughout the region of stent overlap beyond that anticipated because of the summation of struts from like-sized stents. The additional stent coverage further enhances the hemodynamic modification of the intra-aneurysmal circulation. A technique being adopted in bifurcation aneurysms is the use of the Y-stent configuration (142), whereby a Neuroform stent is first placed across one arm of the bifurcation and then a second stent is telescoped through the interstices of the first stent, into the second arm of the bifurcation. The aneurysm is then coiled through the interstices of the doublestent scaffold.

STENT COMPLICATIONS Use of stents and other adjuncts to coiling has been reported to increase the risk of complications. A recent large study examining follow-up from 1811

Chapter 13: Endovascular Management of Intracranial Aneurysms

255

Figure 7 (A) Images from an intracranial CT angiogram demonstrating a fusiform mid-basilar artery aneurysm (patient #1). (B) Frontal and lateral oblique angiographic images depicting the aneurysm. (C) The aneurysm was treated with GDC under balloon remodeling conditions (4 mm  20 mm balloon), following which (D) overlapping 4.0 mm  20 mm (outside) and 4.5 mm  20 mm (inside) Neuroform stents were deployed across the aneurysm neck. (E) Immediate post-treatment angiography (lateral oblique projection) disclosed minor persistent contrast opacification throughout the intra-aneurismal coil interstices (white arrow) which is not apparent on the 12 month (F) follow-up angiogram (frontal and lateral views). (G) Sequential source images through the aneurismal segment from an MRA obtained 25 months after treatment, illustrating cylindrical reconstruction of the parent basilar artery (small white arrows: susceptibility artifact related to the intra-aneurismal coil mass).

aneurysms coiled at a single center, including both ruptured and unruptured aneurysms (143), reported an overall complication rate of 17.7%. Large aneurysm diameter and neck-to-dome ratio were associated with a higher complication rate than was the case for small, narrow-necked aneurysms. Use of ‘‘modified’’ technique (balloon remodeling or stenting) resulted in an increase in the complication rate from 15.7% to 20%. Of course, it would be expected that aneurysms requiring the use of the modified technique were larger and more complex than others, perhaps explaining some of the increases in complications. The Barrow group (144), in their initial report of their Neuroform experience, described several

types of complications. In two cases, the stent migrated away from its intended location after microcatheter manipulations, in one case migrating into the aneurysm lumen. In one case, GDC coil stretching occurred followed by coil fracture, as the coils were deployed through the stent interstices, likely because of friction between the coil and the stent matrix. In all cases in which stenting and coiling were at the same sitting, the coils were successfully passed through the stent interstices. In a single case in which two overlapping stents were placed, in which transstent coiling was attempted two weeks after stent deployment, only the microwire and not the microcatheter could be passed through the interstices. This group had

256

Orbach et al.

reported (144) technical problems involving the stent stabilizer that have been resolved in later-generation Neuroforms. Additional concerns arising when stents are placed within the intracranial circulation include the potential for perforator occlusion [Lopes, 2003 (41)] and delayed in-stent stenosis (145). Fiorella (Congress of Neurological Surgeons 2005 Annual Meeting, Boston, Massachusetts, U.S.) and Woo (WFITN 2005 Annual Meeting, Venice, Italy) have reported, from the combined experiences at the Barrow Neurologic Institute and Cleveland Clinic, several isolated cases of Neuroform-associated delayed in-stent stenosis found on follow-up angiography to have resolved spontaneously. These findings implicate a dynamic series of local histovascular events set in motion by stent implantation and its subsequent subintimal integration into the vessel wall (some of which may be important in mediating aneurysm healing).

treatment of acutely ruptured aneurysms without antiplatelet pretreatment, further emphasizing the inherent thrombogenic hazard of primary stenting. The investigators felt that overlapping stents were significantly more thrombogenic than single stents, and as a result, they add treatment with aspirin for 24 hours following stent placement to their normal antiplatelet regimen with clopidogrel in cases of stent overlap. Benitez et al. (96) described 4 thrombotic events (3 stent related) among their Neuroform series of 49 aneurysms in 48 patients. Nelson et al. (141) reported on a patient harboring a large basilar apex aneurysm requiring Y-stent reconstruction of the neck prior to aneurysm coiling, who developed nonocclusive in-stent thrombus within the left P1 segment despite what was felt to be adequate antiplatelet coverage.

CONCLUSIONS THROMBOEMBOLISM Thromboembolic complications of endovascular aneurysm treatments may be attributed to multiple factors, which deserve particular consideration when selecting patients for stent-supported coil treatment. Attention should be given to (1) the inherent thrombogenicity of the devices; (2) specific patient factors such as comorbidities (age, diabetes, etc.), coagulopathic states, and resistance to antiplatelet drugs; (3) procedure-related factors, such as the duration and complexity of the procedure, timing and dosing of the antiplateletanticoagulant regimen, mechanical compromise of the parent vessel lumen (either from encroachment of the coil mass or malapposition of the stent), dislodgement of existing intra-aneurysmal thrombus, or perianeurysmal injury of the parent vessel (by balloon remodeling or stent deployment); and (4) attributes of the lesion and its microenvironment contributing to thromboembolism, such as the vessel size, aneurysm morphology, local platelet activation, and the hemodynamic characteristics in the vicinity of the aneurysm. The prevalence of thrombotic events in coiltreated aneurysms has been estimated to range from 2.5% to 61% depending on the individual case series cited and the method of surveillance employed, with permanent deficits ranging from 2.5% to 5.5% (146–148). Soeda et al. (148), using diffusion MR surveillance, noted that in 73% of cases in which balloon remodeling was used, there was a diffusion-positive lesion, although most were clinically silent; the incidence of diffusion-positive lesions was similarly high in cases of aneurysms with wide necks. Interestingly, 78% of the diffusion-positive lesions in the posterior fossa were in brain parenchyma whose arterial supply lay proximal to the aneurysms, potentially implicating catheter and wire manipulation rather than aneurysmal thrombus as the etiology. Four stent-related thrombotic events (2 clinically consequential) were reported by Fiorella et al. (144) in their series of 21 Neuroform-treated aneurysms in 19 patients. Three of their four events occurred during

Early experience with stent-supported coil embolization of cerebral aneurysms has engendered significant interest in endoluminal solutions for cerebral aneurysms. Available devices have primarily been used within the context of increasing the effectiveness of existing coil-based endosaccular approaches. For how long this complementary therapeutic approach will evolve before being supplanted by more sophisticated combined or stand-alone endosaccular and endoluminal devices is uncertain. However, it is clear that stent-supported endosaccular approaches represent potentially promising endovascular solutions to that set of complex aneurysms for which it was hoped the GDC would provide safe definitive therapy. Combined surgical and endovascular approaches to very complex aneurysms have been successfully reported (149–152), and this multipronged approach will likely yield results in complex cases superior to that which might have been achievable through the use of either technique in isolation. A promising potential avenue for endovascular treatment of aneurysms is the use of covered stents or stent grafts, which perhaps best encapsulates the concept of luminal reconstruction. Several case reports and one series of patients (153) with internal carotid aneurysms proximal to the level of the supraclinoid segment have been published, with excellent results and low morbidity. Significant technical hurdles remain, chiefly involving the rigidity of the coronarycovered stents now in use and the inability to navigate them distally into the tortuous cerebral vasculature. Additionally, long-term patency rates remain to be determined, with the potential development of longterm delayed stenosis, a known complication of coronary use. A recent study detailing the first rigorous very long-term follow-up on a cohort of patients with intracranial aneurysms that were treated surgically (154) makes clear that the presence of cerebral aneurysms is not an isolated phenomenon but rather represents a manifestation of a widespread vasculopathy.

Chapter 13: Endovascular Management of Intracranial Aneurysms

At follow-up angiography of 140 aneurysms in 112 patients at an average of 9.3 years postoperatively, 2.9% of the clipped aneurysms were found to have regrown (with an annual risk of regrowth of 0.26%), while formation of de novo aneurysms was seen in 8% of patients (with an annual risk of de novo formation of 0.89%). The majority of patients who developed SAH did so over nine years after surgery, and the cumulative rate of aneurysm recurrence was approximately 10% at nine years. Patients with multiple aneurysms were no more likely than those with single aneurysms to have recurrences or de novo aneurysms, nor was gender a factor. The authors thus propose that even in the setting of aneurysms considered to be fully treatable, where technically unassailable surgical clipping has taken place, repeat angiography 9 to 10 years postoperatively might be a reasonable diagnostic procedure. Other studies (155) have shown that patients with completely clipped, unruptured asymptomatic aneurysms continue to have a higher mortality rate than age-matched cohorts from the general population, for up to 10 years of follow-up; for a 52-year-old cohort, survival at 10 years was 91% in the general population and 76% in the clipped cohort. Thus, patients with aneurysms remain vulnerable, perhaps throughout their lives, to the development of other aneurysms, and lifelong surveillance is necessary. Clearly, much work remains in elucidating the biological mechanisms underlying aneurysm formation. This would allow us to aim our treatment at these foundational causes rather than coping with treatments aimed largely at improving vascular morphology, as we do today.

REFERENCES 1. Berenstein A, Lasjaunias P, Ter Brugge KG. Aneurysmal vasculopathy: segmental vulnerability. In: Surgical Neuroangiography. Berlin: Springer-Verlag, 2004:375–380. 2. Hoi Y, Meng H, Woodward SH, et al. Effects of arterial geometry on aneurysm growth: three-dimensional computational fluid dynamics study. J Neurosurg 2004; 101:676–681. 3. Meng H, Feng Y, Woodward SH, et al. Mathematical model of the rupture mechanism of intracranial saccular aneurysms through daughter aneurysm formation and growth. Neurol Res 2005; 27:459–465. 4. Inagawa T, Hirano A. Autopsy study of unruptured incidental intracranial aneurysms. Surg Neurol 1990; 34:361–365. 5. McCormick WF, Nofzinger JD. Saccular intracranial aneurysms: an autopsy study. J Neurosurg 1965; 22:155–159. 6. Atkinson JL, Sundt TM Jr., Houser OW, et al. Angiographic frequency of anterior circulation intracranial aneurysms. J Neurosurg 1989; 70:551–555. 7. Menghini VV, Brown RD Jr., Sicks JD, et al. Incidence and prevalence of intracranial aneurysms and hemorrhage in Olmsted County, Minnesota, 1965 to 1995. Neurology 1998; 51:405–411. 8. Vespa PM, Gobin YP. Endovascular treatment and neurointensive care of ruptured aneurysms. Crit Care Clin 1999; 15:667–684. 9. Barker FG II, Amin-Hanjani S, Butler WE, et al. Agedependent differences in short-term outcome after surgical or endovascular treatment of unruptured intracranial

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

257

aneurysms in the United States, 1996-2000. Neurosurgery 2004; 54:18–28; discussion 28–30. Barker FG II, Amin-Hanjani S, Butler WE, et al. In-hospital mortality and morbidity after surgical treatment of unruptured intracranial aneurysms in the United States, 1996-2000: the effect of hospital and surgeon volume. Neurosurgery 2003; 52:995–1007; discussion 1007–1009. Cross DT III, Tirschwell DL, Clark MA, et al. Mortality rates after subarachnoid hemorrhage: variations according to hospital case volume in 18 states. J Neurosurg 2003; 99:810–817. Hoh BL, Rabinov JD, Pryor JC, et al. In-hospital morbidity and mortality after endovascular treatment of unruptured intracranial aneurysms in the United States, 1996–2000: effect of hospital and physician volume. AJNR Am J Neuroradiol 2003; 24:1409–1420. International Study of Unruptured Intracranial Aneurysms Investigators. Unruptured intracranial aneurysms– risk of rupture and risks of surgical intervention. N Engl J Med 1998; 339:1725–1733. Ausman JI. The New England Journal of Medicine report on unruptured intracranial aneurysms: a critique. Surg Neurol 1999; 51:227–229. Juvela S, Porras M, Poussa K. Natural history of unruptured intracranial aneurysms: probability of and risk factors for aneurysm rupture. J Neurosurg 2000; 93:379–387. Winn HR, Jane JA Sr., Taylor J, et al. Prevalence of asymptomatic incidental aneurysms: review of 4568 arteriograms. J Neurosurg 2002; 96:43–49. Dickey P, Kailasnath P. The diameter-cube hypothesis: a new biophysical model of aneurysm rupture. Surg Neurol 2002; 58:166–173; discussion 173–180. Aletich VA, Debrun GM, Misra M, et al. The remodeling technique of balloon-assisted Guglielmi detachable coil placement in wide-necked aneurysms: experience at the University of Illinois at Chicago. J Neurosurg 2000; 93:388–396. Johnston SC, Gress DR, Kahn JG. Which unruptured cerebral aneurysms should be treated? A cost-utility analysis. Neurology 1999; 52:1806–1815. Bederson JB, Awad IA, Wiebers DO, et al. Recommendations for the management of patients with unruptured intracranial aneurysms: a statement for healthcare professionals from the Stroke Council of the American Heart Association. Stroke 2000; 31:2742–2750. Ausman JI. The Unruptured Intracranial Aneurysm Study-II: a critique of the second study. Surg Neurol 2004; 62:91–94. Clarke G, Mendelow AD, Mitchell P. Predicting the risk of rupture of intracranial aneurysms based on anatomical location. Acta Neurochir (Wien) 2005; 147:259–263; discussion 263. Matsubara S, Hadeishi H, Suzuki A, et al. Incidence and risk factors for the growth of unruptured cerebral aneurysms: observation using serial computerized tomography angiography. J Neurosurg 2004; 101:908–914. Halbach VV, Higashida RT, Dowd CF, et al. The efficacy of endosaccular aneurysm occlusion in alleviating neurological deficits produced by mass effect. J Neurosurg 1994; 80:659–666. Kupersmith MJ, Hurst R, Berenstein A, et al. The benign course of cavernous carotid artery aneurysms. J Neurosurg 1992; 77:690–693. Linskey ME, Sekhar LN, Hirsch WL Jr., et al. Aneurysms of the intracavernous carotid artery: natural history and indications for treatment. Neurosurgery 1990; 26:933–937; discussion 937–938. Strother CM, Eldevik P, Kikuchi Y, et al. Thrombus formation and structure and the evolution of mass effect in intracranial aneurysms treated by balloon embolization:

258

28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

Orbach et al. emphasis on MR findings. AJNR Am J Neuroradiol 1989; 10:787–796. Kak VK, Taylor AR, Gordon DS. Proximal carotid ligation for internal carotid aneurysms. A long-term follow-up study. J Neurosurg 1973; 39:503–513. Scott M, Skwarok E. The treatment of cerebral aneurysms by ligation of the common carotid artery. Surg Gynecol Obstet 1961; 113:54–61. Fox AJ, Vinuela F, Pelz DM, et al. Use of detachable balloons for proximal artery occlusion in the treatment of unclippable cerebral aneurysms. J Neurosurg 1987; 66:40–46. Higashida RT, Halbach VV, Dowd C, et al. Endovascular detachable balloon embolization therapy of cavernous carotid artery aneurysms: results in 87 cases. J Neurosurg 1990; 72:857–863. Higashida RT, Halbach VV, Dowd CF, et al. Intracranial aneurysms: interventional neurovascular treatment with detachable balloons–results in 215 cases. Radiology 1990; 178:663–670. Berenstein A, Ransohoff J, Kupersmith M, et al. Transvascular treatment of giant aneurysms of the cavernous carotid and vertebral arteries. Functional investigation and embolization. Surg Neurol 1984; 21:3–12. Debrun G, Fox A, Drake C, et al. Giant unclippable aneurysms: treatment with detachable balloons. AJNR Am J Neuroradiol 1981; 2:167–173. Serbinenko FA. Balloon catheterization and occlusion of major cerebral vessels. J Neurosurg 1974; 41:125–145. Allen JW, Alastra AJ, Nelson PK. Proximal intracranial internal carotid artery branches: prevalence and importance for balloon occlusion test. J Neurosurg 2005; 102:45–52. Vazquez Anon V, Aymard A, Gobin YP, et al. Balloon occlusion of the internal carotid artery in 40 cases of giant intracavernous aneurysm: technical aspects, cerebral monitoring, and results. Neuroradiology 1992; 34:245–251. Linskey ME, Jungreis CA, Yonas H, et al. Stroke risk after abrupt internal carotid artery sacrifice: accuracy of preoperative assessment with balloon test occlusion and stable xenon-enhanced CT. AJNR Am J Neuroradiol 1994; 15:829–843. Aymard A, Gobin YP, Hodes JE, et al. Endovascular occlusion of vertebral arteries in the treatment of unclippable vertebrobasilar aneurysms. J Neurosurg 1991; 74:393–398. Eckard DA, Purdy PD, Bonte FJ. Temporary balloon occlusion of the carotid artery combined with brain blood flow imaging as a test to predict tolerance prior to permanent carotid sacrifice. AJNR Am J Neuroradiol 1992; 13:1565–1569. Jawad K, Miller D, Wyper DJ, et al. Measurement of CBF and carotid artery pressure compared with cerebral angiography in assessing collateral blood supply after carotid ligation. J Neurosurg 1977; 46:185–196. Monsein LH, Jeffery PJ, van Heerden BB, et al. Assessing adequacy of collateral circulation during balloon test occlusion of the internal carotid artery with 99mTc-HMPAO SPECT. AJNR Am J Neuroradiol 1991; 12:1045–1051. Chaloupka JC, Dare AO, Putman CM. Limitations of the hypotensive challenge with balloon test occlusion for predicting delayed hemodynamic ischemia: the Yale experience. Presented at the 34th Annual Meeting of the American Society of Neuroradiology, Seattle 1996. Gurian JH, Vinuela F, Gobin YP, et al. Aneurysm rupture after parent vessel sacrifice: treatment with Guglielmi detachable coil embolization via retrograde catheterization: case report. Neurosurgery 1995; 37:1216–1220; discussion 1220–1211. Barnett HJ, Peerless SJ, Kaufmann JC. ‘‘Stump’’ on internal carotid artery–a source for further cerebral embolic ischemia. Stroke 1978; 9:448–456.

46. Mounayer C, Undren P, Piotin M, et al. Neck-bridge device for combined endovascular and surgical treatment of a giant anterior communicating artery aneurysm. Neuroradiology 2005; 47:295–299. 47. Arnautovic KI, Al-Mefty O, Angtuaco E. A combined microsurgical skull-base and endovascular approach to giant and large paraclinoid aneurysms. Surg Neurol 1998; 50:504–518; discussion 518–520. 48. Higashida RT, Halbach VV, Barnwell SL, et al. Treatment of intracranial aneurysms with preservation of the parent vessel: results of percutaneous balloon embolization in 84 patients. AJNR Am J Neuroradiol 1990; 11:633–640. 49. Moret J, Boulin A, Mawad M. Endovascular therapy of berry aneurysms by endosaccular balloon occlusion. Neuroradiology 1991; 33(suppl):135–136. 50. Romodanov AP, Shcheglov VI. Intravascular occlusion of saccular aneurysms of the cerebral arteries by means of a detachable balloon catheter. In: Krayenbuhl H, Sweet WH, eds. Advances and Technical Standards in Neurosurgery, vol 2. Springer-VerlagNew York1982:25–49. 51. Kinugasa K, Mandai S, Terai Y, et al. Direct thrombosis of aneurysms with cellulose acetate polymer. Part II: Preliminary clinical experience. J Neurosurg 1992; 77:501–507. 52. Guglielmi G, Vinuela F. Intracranial aneurysms. Guglielmi electrothrombotic coils. Neurosurg Clin N Am 1994; 5:427–435. 53. Guglielmi G, Vinuela F, Dion J, et al. Electrothrombosis of saccular aneurysms via endovascular approach. Part 2: Preliminary clinical experience. J Neurosurg 1991; 75:8–14. 54. Guglielmi G, Vinuela F, Sepetka I, et al. Electrothrombosis of saccular aneurysms via endovascular approach. Part 1: Electrochemical basis, technique, and experimental results. J Neurosurg 1991; 75:1–7. 55. Molyneux A, Kerr R, Stratton I, et al. International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: a randomised trial. Lancet 2002; 360:1267–1274. 56. Derdeyn CP, Barr JD, Berenstein A, et al. The International Subarachnoid Aneurysm Trial (ISAT): a position statement from the Executive Committee of the American Society of Interventional and Therapeutic Neuroradiology and the American Society of Neuroradiology. AJNR Am J Neuroradiol 2003; 24:1404–1408. 57. Molyneux AJ, Kerr RS, Yu LM, et al. International Subarachnoid Aneurysm Trial (ISAT) of neurosurgical clipping versus endovascular coiling in 2143 patients with ruptured intracranial aneurysms: a randomised comparison of effects on survival, dependency, seizures, rebleeding, subgroups, and aneurysm occlusion. Lancet 2005; 366:809–817. 58. Solenski NJ, Haley EC Jr., Kassell NF, et al. Medical complications of aneurysmal subarachnoid hemorrhage: a report of the multicenter, cooperative aneurysm study. Participants of the Multicenter Cooperative Aneurysm Study. Crit Care Med 1995; 23:1007–1017. 59. Mawad ME, Mawad JK, Cartwright J Jr., et al. Long-term histopathologic changes in canine aneurysms embolized with Guglielmi detachable coils. AJNR Am J Neuroradiol 1995; 16:7–13. 60. Molyneux AJ, Ellison DW, Morris J, et al. Histological findings in giant aneurysms treated with Guglielmi detachable coils. Report of two cases with autopsy correlation. J Neurosurg 1995; 83:129–132. 61. Reul J, Weis J, Spetzger U, et al. Long-term angiographic and histopathologic findings in experimental aneurysms of the carotid bifurcation embolized with platinum and tungsten coils. AJNR Am J Neuroradiol 1997; 18:35–42. 62. Derdeyn CP, Graves VB, Turski PA, et al. MR angiography of saccular aneurysms after treatment with Guglielmi

Chapter 13: Endovascular Management of Intracranial Aneurysms

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

detachable coils: preliminary experience. AJNR Am J Neuroradiol 1997; 18:279–286. Kahara VJ, Seppanen SK, Ryymin PS, et al. MR angiography with three-dimensional time-of-flight and targeted maximum-intensity-projection reconstructions in the follow-up of intracranial aneurysms embolized with Guglielmi detachable coils. AJNR Am J Neuroradiol 1999; 20:1470–1475. Anzalone N, Righi C, Simionato F, et al. Three-dimensional time-of-flight MR angiography in the evaluation of intracranial aneurysms treated with Guglielmi detachable coils. AJNR Am J Neuroradiol 2000; 21:746–752. Bergui M, Bradac GB. Acute endovascular treatment of ruptured aneurysms in poor-grade patients. Neuroradiology 2004; 46:161–164. [Epub 2003 Dec 20]. Bailes JE, Spetzler RF, Hadley MN, et al. Management morbidity and mortality of poor-grade aneurysm patients. J Neurosurg 1990; 72:559–566. Yalamanchili K, Rosenwasser RH, Thomas JE, et al. Frequency of cerebral vasospasm in patients treated with endovascular occlusion of intracranial aneurysms. AJNR Am J Neuroradiol 1998; 19:553–558. Koivisto T, Vanninen R, Hurskainen H, et al. Outcomes of early endovascular versus surgical treatment of ruptured cerebral aneurysms. A prospective randomized study. Stroke 2000; 31:2369–2377. Baltsavias GS, Byrne JV, Halsey J, et al. Effects of timing of coil embolization after aneurysmal subarachnoid hemorrhage on procedural morbidity and outcomes. Neurosurgery 2000; 47:1320–1329; discussion 1329–1331. Cloft HJ, Kallmes DF. Cerebral aneurysm perforations complicating therapy with Guglielmi detachable coils: a metaanalysis. AJNR Am J Neuroradiol 2002; 23:1706–1709. Eskridge JM, Song JK. Endovascular embolization of 150 basilar tip aneurysms with Guglielmi detachable coils: results of the Food and Drug Administration multicenter clinical trial. J Neurosurg 1998; 89:81–86. Tateshima S, Murayama Y, Gobin YP, et al. Endovascular treatment of basilar tip aneurysms using Guglielmi detachable coils: anatomic and clinical outcomes in 73 patients from a single institution. Neurosurgery 2000; 47:1332–1339; discussion 1339–1342. Johnston SC, Dudley RA, Gress DR, et al. Surgical and endovascular treatment of unruptured cerebral aneurysms at university hospitals. Neurology 1999; 52:1799–1805. Johnston SC, Wilson CB, Halbach VV, et al. Endovascular and surgical treatment of unruptured cerebral aneurysms: comparison of risks. Ann Neurol 2000; 48:11–19. Johnston SC, Zhao S, Dudley RA, et al. Treatment of unruptured cerebral aneurysms in California. Stroke 2000; 32:597–605. Fernandez Zubillaga A, Guglielmi G, Vinuela F, et al. Endovascular occlusion of intracranial aneurysms with electrically detachable coils: correlation of aneurysm neck size and treatment results. AJNR Am J Neuroradiol 1994; 15:815–820. Hayakawa M, Murayama Y, Duckwiler GR, et al. Natural history of the neck remnant of a cerebral aneurysm treated with the Guglielmi detachable coil system. J Neurosurg 2000; 93:561–568. Kole MK, Pelz DM, Kalapos P, et al. Endovascular coil embolization of intracranial aneurysms: important factors related to rates and outcomes of incomplete occlusion. J Neurosurg 2005; 102:607–615. Murayama Y, Nien YL, Duckwiler G, et al. Guglielmi detachable coil embolization of cerebral aneurysms: 11 years’ experience.[see comment]. J Neurosurg 2003; 98:959–966. Malisch TW, Guglielmi G, Vinuela F, et al. Intracranial aneurysms treated with the Guglielmi detachable coil:

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

259

midterm clinical results in a consecutive series of 100 patients. J Neurosurg 1997; 87:176–183. Raymond J, Guilbert F, Weill A, et al. Long-term angiographic recurrences after selective endovascular treatment of aneurysms with detachable coils. Stroke 2003; 34:1398–1403. Debrun GM, Aletich VA, Kehrli P, et al. Selection of cerebral aneurysms for treatment using Guglielmi detachable coils: the preliminary University of Illinois at Chicago experience. Neurosurgery 1998; 43:1281–1295; discussion 1296–1297. Hodgson TJ, Carroll T, Jellinek DA. Subarachnoid hemorrhage due to late recurrence of a previously unruptured aneurysm after complete endovascular occlusion. AJNR Am J Neuroradiol 1998; 19:1939–1941. Gallas S, Pasco A, Cottier JP, et al. A multicenter study of 705 ruptured intracranial aneurysms treated with Guglielmi detachable coils. AJNR Am J Neuroradiol 2005; 26:1723–1731. Cloft HJ, Kallmes DF. Aneurysm packing with HydroCoil Embolic System versus platinum coils: initial clinical experience. AJNR Am J Neuroradiol 2004; 25:60–62. Dawson RC, Krisht AF, Barrow DL, et al. Treatment of experimental aneurysms using collagen-coated microcoils. Neurosurgery 1995; 36:133–139; discussion 139–140. Kang HS, Han MH, Kwon BJ, et al. Short-term outcome of intracranial aneurysms treated with polyglycolic acid/ lactide copolymer-coated coils compared to historical controls treated with bare platinum coils: a single-center experience. AJNR Am J Neuroradiol 2005; 26:1921–1928. Murayama Y, Tateshima S, Gonzalez NR, et al. Matrix and bioabsorbable polymeric coils accelerate healing of intracranial aneurysms: long-term experimental study. Stroke 2003; 34:2031–2037. Mawad ME, Cekirge S, Ciceri E, et al. Endovascular treatment of giant and large intracranial aneurysms by using a combination of stent placement and liquid polymer injection. J Neurosurg 2002; 96:474–482. Molyneux AJ, Cekirge S, Saatci I, et al. Cerebral Aneurysm Multicenter European Onyx (CAMEO) trial: results of a prospective observational study in 20 European centers. AJNR Am J Neuroradiol 2004; 25:39–51. Murayama Y, Vinuela F, Tateshima S, et al. Endovascular treatment of experimental aneurysms by use of a combination of liquid embolic agents and protective devices. AJNR Am J Neuroradiol 2000; 21:1726–1735. Raymond J, Metcalfe A, Desfaits AC, et al. Alginate for endovascular treatment of aneurysms and local growth factor delivery. AJNR Am J Neuroradiol 2003; 24:1214–1221. Lefkowitz MA, Gobin YP, Akiba Y, et al. Ballonn-assisted Guglielmi detachable coiling of wide-necked aneurysma: Part II–clinical results. Neurosurgery 1999; 45:531–537; discussion 537–538. Moret J, Cognard C, Weill A, et al. The ‘‘remodelling technique’’ in the treatment of wide neck intracranial aneurysms. Interv Neurorad 1997; 3:21–35. Nelson PK, Levy DI. Balloon-assisted coil embolization of wide-necked aneurysms of the internal carotid artery: medium-term angiographic and clinical follow-up in 22 patients. AJNR Am J Neuroradiol 2001; 22:19–26. Benitez RP, Silva MT, Klem J, et al. Endovascular occlusion of wide-necked aneurysms with a new intracranial microstent (Neuroform) and detachable coils. Neurosurgery 2004; 54:1359–1367; discussion 1368. Fiorella D, Albuquerque FC, Deshmukh VR, et al. Usefulness of the Neuroform stent for the treatment of cerebral aneurysms: results at initial (3-6-mo) follow-up. Neurosurgery 2005; 56:1191–1202. Lanzino G, Wakhloo AK, Fessler RD, et al. Efficacy and current limitations of intravascular stents for intracranial

260

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

Orbach et al. internal carotid, vertebral, and basilar artery aneurysms. J Neurosurg 1999; 91:538–546. Mericle RA, Lanzino G, Wakhloo AK, et al. Stenting and secondary coiling of intracranial internal carotid artery aneurysm: technical case report. Neurosurgery 1998; 43:1229–1234. Szikora I, Guterman LR, Wells KM, et al. Combined use of stents and coils to treat experimental wide-necked carotid aneurysms: preliminary results. AJNR Am J Neuroradiol 1994; 15:1091–1102. Wanke I, Doerfler A, Schoch B, et al. Treatment of widenecked intracranial aneurysms with a self-expanding stent system: initial clinical experience. AJNR Am J Neuroradiol 2003; 24:1192–1199. Vinuela F, Duckwiler G, Mawad M. Guglielmi detachable coil embolization of acute intracranial aneurysm: perioperative anatomical and clinical outcome in 403 patients. J Neurosurg 1997; 86:475–482. Kawanabe Y, Sadato A, Taki W, et al. Endovascular occlusion of intracranial aneurysms with Guglielmi detachable coils: correlation between coil packing density and coil compaction. Acta Neurochir 2001; 143:451–455. Sluzewski M, Menovsky T, van Rooij WJ, et al. Coiling of very large or giant cerebral aneurysms: long-term clinical and serial angiographic results. AJNR Am J Neuroradiol 2003; 24:257–262. Sluzewski M, van Rooij WJ, Slob MJ, et al. Relation between aneurysm volume, packing, and compaction in 145 cerebral aneurysms treated with coils. Radiology 2004; 231:653–658. Tamatani S, Ito Y, Abe H, et al. Evaluation of the stability of aneurysms after embolization using detachable coils: correlation between stability of aneurysms and embolized volume of aneurysms. AJNR Am J Neuroradiol 2002; 23:762–767. Levy DI. Embolization of wide-necked anterior communicating artery aneurysm: technical note. Neurosurgery 1997; 41:979–982. Mericle RA, Wakhloo AK, Rodriguez R, et al. Temporary balloon protection as an adjunct to endosaccular coiling of wide-necked cerebral aneurysms: technical note. Neurosurgery 1997; 41:975–978. Malek AM, Halbach VV, Phatouros CC, et al. Balloonassist technique for endovascular coil embolization of geometrically difficult intracranial aneurysms. Neurosurgery 2000; 46:1397–1406; discussion 1406–1497. Newell DW, Elliott JP, Eskridge JM, et al. Endovascular therapy for aneurysmal vasospasm. Crit Care Clin 1999; 15:685–699. Lavine SD, Masri LS, Levy ML, et al. Temporary occlusion of the middle cerebral artery in intracranial aneurysm surgery: time limitation and advantage of brain protection. J Neurosurg 1997; 87:817–824. Turk AS, Rappe AH, Villar F, et al. Evaluation of the TriSpan neck bridge device for the treatment of widenecked aneurysms: an experimental study in canines. Stroke 2001; 32:492–497. Abrahams JM, Diamond SL, Hurst RW, et al. Topic review: surface modifications enhancing biological activity of Guglielmi detachable coils in treating intracranial aneurysms. Surg Neurol 2000; 54:34–40; discussion 40–41. Moret J, Ross IB, Weill A, et al. The retrograde approach: a consideration for the endovascular treatment of aneurysms. AJNR Am J Neuroradiol 2000; 21:262–268. Anxionnat R, Bracard S, Ducrocq X, et al. Intracranial aneurysms: clinical value of 3D digital subtraction angiography in the therapeutic decision and endovascular treatment. Radiology 2001; 218:799–808.

116. Vallee JN, Pierot L, Bonafe A, et al. Endovascular treatment of intracranial wide-necked aneurysms using. AJNR Am J Neuroradiol 2004; 25:298–306. 117. Kallmes DF, Williams AD, Cloft HJ, et al. Platinum coilmediated implantation of growth factor-secreting endovascular tissue grafts: an in vivo study. Radiology 1998; 207:519–523. 118. Ribourtout E, Raymond J. Gene therapy and endovascular treatment of intracranial aneurysms. Stroke 2004; 35:786– 793; [Epub 2004 Feb 12]. 119. Raymond J, Leblanc P, Morel F, et al. Beta radiation and inhibition of recanalization after coil embolization of canine arteries and experimental aneurysms: how should radiation be delivered? Stroke 2003; 34:1262–1268. 120. Raymond J, Roy D, Leblanc P, et al. Endovascular treatment of intracranial aneurysms with radioactive coils. Stroke 2003; 34:2801–2806; [Epub 2003 Nov 06]. 121. Piotin M, Iijima A, Wada H, et al. Increasing the packing of small aneurysms with complex-shaped coils: an in vitro study. AJNR Am J Neuroradiol 2003; 24:1446–1448. 122. Piotin M, Mandai S, Murphy KJ, et al. Dense packing of cerebral aneurysms: an in vitro study with detachable platinum coils. AJNR Am J Neuroradiol 2000; 21:757–760. 123. Soga Y, Preul MC, Furuse M, et al. Calcium alginate provides a high degree of embolization in aneurysm. Neurosurgery 2004; 55:1401–1409; discussion 1409. 124. Weber W, Siekmann R, Kis B, et al. Treatment and followup of 22 unruptured wide-necked intracranial aneurysms of the internal carotid artery with Onyx HD 500. AJNR Am J Neuroradiol 2005; 26:1909–1915. 125. Meyers PM, Lavine SD, Fitzsimmons BF, et al. Chemical meningitis after cerebral aneurysm treatment using two second-generation aneurysm coils: report of two cases. Neurosurgery 2004; 55:1222. 126. Veznedaroglu E, Benitez RP, Rosenwasser RH. Surgically treated aneurysms previously coiled: lessons learned. Neurosurgery 2004; 54:300–303; discussion 303–305. 127. Gurian JH, Martin NA, King WA, et al. Neurosurgical management of cerebral aneurysms following unsuccessful or incomplete endovascular embolization. J Neurosurg 1995; 83:843–853. 128. Thornton J, Dovey Z, Alazzaz A, et al. Surgery following endovascular coiling of intracranial aneurysms. Surg Neurol 2000; 54:352–360. 129. Slob MJ, Sluzewski M, van Rooij WJ, et al. Additional coiling of previously coiled cerebral aneurysms: clinical and angiographic results. AJNR Am J Neuroradiol 2004; 25:1373–1376. 130. Cekirge HS, Islak C, Firat MM, et al. Endovascular coil embolization of residual or recurrent aneurysms after surgical clipping. Acta Radiol 2000; 41:111–115. 131. Lubicz B, Leclerc X, Gauvrit JY, et al. Endovascular treatment of remnants of intracranial aneurysms following. Neuroradiology 2004; 46:318–322; [Epub 2004 Mar]. 132. Pierot L, Boulin A, Visot A, et al. Postoperative aneurysm remnants: endovascular treatment as an alternative to further surgery. Neuroradiology 1999; 41:315–319. 133. Drake CG, Friedman AH, Peerless SJ. Failed aneurysm surgery. Reoperation in 115 cases. J Neurosurg 1984; 61:848–856. 134. Rudin S, Wang Z, Kyprianou I, et al. Measurement of flow modification in phantom aneurysm model: comparison of coils and a longitudinally and axially asymmetric stent– initial findings. Radiology 2004; 231:272–276. 135. Barath K, Cassot F, Rufenacht DA, et al. Anatomically shaped internal carotid artery aneurysm in vitro model for flow analysis to evaluate stent effect. AJNR Am J Neuroradiol 2004; 25:1750–1759. 136. Canton G, Levy DI, Lasheras JC, et al. Flow changes caused by the sequential deployment of stents across the neck of sidewall cerebral aneurysms. J Neurosurg 2005; 103:891–902.

Chapter 13: Endovascular Management of Intracranial Aneurysms 137. Lieber BB, Gounis MJ. The physics of endoluminal stenting in the treatment of cerebrovascular aneurysms. Neurol Res 2002; 24 (suppl 1):S33–S42. 138. Lieber BB, Livescu V, Hopkins LN, et al. Particle image velocimetry assessment of stent design influence on intraaneurysmal flow. Ann Biomed Eng 2002; 30:768–777. 139. Lylyk P, Cohen JE, Ceratto R, et al. Endovascular reconstruction of intracranial arteries by stent placement and combined techniques. J Neurosurg 2002; 97:1306–1313. 140. Lylyk P, Ferrario A, Pasbon B, et al. Buenos Aires experience with the Neuroform self-expanding stent for the treatment of intracranial aneurysms. J Neurosurg 2005; 102:235–241. 141. Nelson PK, Sahlein D, Shapiro M, et al. Recent steps toward a reconstructive endovascular solution for the orphaned, complex-neck aneurysm. Neurosurgery 2006; 59:S377–S392. 142. Chow MM, Woo HH, Masaryk TJ, et al. A novel endovascular treatment of a wide-necked basilar apex aneurysm by using a Y-configuration, double stent technique. AJNR Am J Neuroradiol 2004; 25:509–512. 143. Henkes H, Fischer S, Weber W, et al. Endovascular coil occlusion of 1811 intracranial aneurysms: early angiographic and clinical results. Neurosurgery 2004; 54:268– 280; discussion 280–285. 144. Fiorella D, Albuquerque FC, Han P, et al. Preliminary experience using the Neuroform stent for the treatment of cerebral aneurysms.[see comment]. Neurosurgery 2004; 54:6–16; discussion 16–17. 145. Fiorella D, Albuquerque FC, Deshmukh VR, et al. In-stent stenosis as a delayed complication of neuroform stentsupported coil embolization of an incidental carotid terminus aneurysm. AJNR Am J Neuroradiol 2004; 25:1764–1767. 146. Pelz DM, Lownie SP, Fox AJ. Thromboembolic events associated with the treatment of cerebral aneurysms with Guglielmi detachable coils. AJNR Am J Neuroradiol 1998; 19:1541–1547.

261

147. Qureshi AI. Editorial comment–thromboembolic events during neuroendovascular procedures [comment]. Stroke 2003; 34:1728–1729. 148. Soeda A, Sakai N, Sakai H, et al. Thromboembolic events associated with Guglielmi detachable coil embolization of asymptomatic cerebral aneurysms: evaluation of 66 consecutive cases with use of diffusion-weighted MR imaging [see comment]. AJNR Am J Neuroradiol 2003; 24:127–132. 149. Hacein-Bey L, Connolly ES Jr., Mayer SA, et al. Complex intracranial aneurysms: combined operative and endovascular approaches. Neurosurgery 1998; 43:1304–1312; discussion 1312–1303. 150. Hoh BL, Carter BS, Budzik RF, et al. Results after surgical and endovascular treatment of paraclinoid aneurysms by a combined neurovascular team. Neurosurgery 2001; 48:78–89; discussion 89–90. 151. Martin NA. The combination of endovascular and surgical techniques for the treatment of intracranial aneurysms. Neurosurg Clin N Am 1998; 9:897. 152. Ng PY, Huddle D, Gunel M, et al. Intraoperative endovascular treatment as an adjunct to microsurgical clipping of paraclinoid aneurysms. J Neurosurg 2000; 93:554–560. 153. Saatci I, Cekirge HS, Ozturk MH, et al. Treatment of internal carotid artery aneurysms with a covered stent: experience in 24 patients with mid-term follow-up results. AJNR Am J Neuroradiol 2004; 25:1742–1749. 154. Tsutsumi K, Ueki K, Morita A, et al. Risk of aneurysm recurrence in patients with clipped cerebral aneurysms. Results of long-term follow-up angiography. Stroke 2001; 32:1191–1194. 155. Britz GW, Salem L, Newell DW, et al. Impact of surgical clipping on survival in unruptured and ruptured cerebral aneurysms: a population-based study. Stroke 2004; 35:1399–1403.

14 Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm Jonathan L. Brisman Department of Cerebrovascular and Endovascular Neurosurgery, Winthrop University Hospital, Mineola, Long Island, New York, U.S.A.

David W. Newell and Joseph M. Eskridge Department of Neurosurgery; Department of Interventional Neuroradiology, Seattle Neuroscience Institute, Seattle, Washington, U.S.A.

INTRODUCTION Endovascular therapy to treat symptomatic vasospasm after aneurysmal subarachnoid hemorrhage (aSAH) has become a mainstay in many centers. Cerebral vasospasm, defined as reversible vasoconstriction of the intracranial vasculature, is found in approximately 30% to 70% of patients after aSAH, although perhaps only one-third to one-half of these patients will develop symptoms and/or delayed ischemic neurologic deficits (DINDs). DINDs remain the leading cause of stroke, morbidity, and mortality after aSAH (1). The Fisher grade (Table 1) (2), scoring the amount of blood seen on the initial head computed tomography (CT) scan, remains a good predictor of the severity of vasospasm to be anticipated and the incidence of CT demonstrable infarction and associated morbidity and mortality. Whether patients presenting with aSAH are more likely to develop vasospasm, if treated by endovascular coiling versus craniotomy and clipping, is a matter of recent debate with evidence supporting both claims (3–5) and no prospective study as of yet performed. Medical therapy, including the administration of nimodipine for 21 days post bleed, regardless of the presence of vasospasm, and ‘‘triple-H’’ (hypervolemia, hypertension, and hemodilution) therapy once vasospasm has been identified have improved outcomes after aSAH and averted vasospasm-induced Table 1 Fisher’s Grade Grade 1 Grade 2 Grade 3 Grade 4

No detectable blood on CT Diffuse think SAH on CT with vertical layers <1-mm thick Local clot or thick SAH with vertical layers >1-mm thick Intracerebral or intraventricular blood without findings of a grade 3

Abbreviations: CT, computed tomography; SAH, subarachnoid hemorrhage.

DINDs in some patients. Some patients, however, will suffer devastating cerebral ischemia despite these efforts. Neurointerventional techniques, including intra-arterial administration of vasodilators such as papaverine and transluminal balloon angioplasty (TBA), have gained good results and have emerged as a more aggressive approach for such patients (6). These endovascular techniques (intra-arterial infusion of medication and angioplasty) have their own associated risks and benefits, and controversy exists over the best method (7). At what point to intervene with endovascular treatment has also been controversial with some waiting until aggressive maximal medical therapy has failed and others advocating prophylactic balloon angioplasty in patients at high risk of developing symptomatic vasospasm (8,9).

PATHOPHYSIOLOGY Critical to an understanding of the endovascular approach to vasospasm is a working knowledge of the proposed mechanism leading to vasoconstriction. The exact etiology remains unknown. Causation is likely multifactorial, with the process somehow triggered by the coating of the vessels with the breakdown products of blood, such as oxyhemoglobin, and the release of factors including serotonin, angiotensin, prostaglandins, and thromboxane. Free radicals generated in the blood vessel wall are also believed to play a major role in causing vasospasm. Both vascular smooth muscle cells and endothelium mediate cerebral vascular autoregulation, a process whereby the brain attempts to maintain adequate blood flow despite varying alterations in blood pressure. Vasoconstriction is the result of smooth muscle cell contractility, which is a calcium-dependent function. The counterpart to vasoconstriction is vasodilatation, a process influenced by the endothelium and its production of substances such as nitric oxide. In essence, vasospasm seems to be the result of both a

264

Brisman et al.

heightened vasoconstriction by the smooth muscle cells and a decreased vasodilatation, which is perhaps the result of diminished nitric oxide activity and increased endothelin activity. Histologic response of the blood vessels to a prolonged exposure to blood components has been documented and includes thickening of the media, intimal edema, subintimal cellular proliferation with muscle cells and fibroblasts, thrombus formation, and ultimately necrosis and fibrosis.

HISTORICAL ASPECTS In 1984, Zubkov et al. published the first report of balloon angioplasty to reverse angiographic vasospasm caused by aSAH with resultant clinical improvement (10). Although, earlier reports had described the endovascular infusion of vasodilators for vasospasm, these techniques had not been incorporated into routine clinical usage. As endovascular technology developed, softer silicone balloons and newer microcatheters permitted the more widespread use of these devices by interventional neuroradiologists who were already employing such devices to treat aneurysms and intracranial vascular malformations. In 1989, two groups in North America fueled the mainstream application of these techniques with promising results by treating vasospasm via the endovascular route (11,12). Other reports followed and the instillation of intra-arterial papaverine (IAP) locally at the site of intracranial vasospasm gained popularity for some time. IAP, however, was found to be associated with multiple side effects and its action was found to be short-lived. This dissatisfaction led to the experimental use of other selectively infused vasodilators, such as nicardipine, verapamil, and milrinone, and paved the way for the growth of transluminal angioplasty as a first-line therapy for refractory vasospasm (13,14).

DIAGNOSIS AND MEDICAL MANAGEMENT Cerebral vasospasm generally occurs between days 3 and 12 after aSAH, with a peak incidence on days 6 to 8 (1,15). The condition is marked by blood vessel narrowing at the base of the brain in the large vessels of and near the circle of Willis. Distal vasospasm may also be seen, but less commonly so. The diagnosis of cerebral vasospasm is made based on the clinical condition of the patient in concert with a series of ancillary tests. The importance of distinguishing between angiographic or radiologic vasospasm and clinically symptomatic vasospasm cannot be underscored enough, as the latter is treated much more aggressively and the risks of endovascular therapy are usually higher than with medical therapy alone. An estimated one-third to one-half of all patients who develop angiographic vasospasm will not develop symptoms (15,16). When cerebral blood flow (CBF) is greatly reduced and compensatory mechanisms, such as autoregulation, collateral flow, and increased oxygen extraction, are exhausted, clinically symptomatic ischemia and infarction may develop (15). The Fisher four point grading scale (Table 1) has been shown to be a good prognostic indicator for the

development of vasospasm, but the correlation is not exact and all patients with aSAH should be monitored carefully for the development of vasospasm (2,17). Any patient who has suffered aSAH and who develops a new focal neurologic deficit or decrease in level of arousal should be strongly considered to have vasospasm, particularly once other causes such as metabolic derangements, aneurysmal rebleeding, and hydrocephalus have been ruled out. Frequently, these conditions may coexist or such symptoms may be difficult to detect in patients who present in poor clinical grade, and it may therefore be wise to assume that symptomatic vasospasm is present until proven contrary. We use transcranial Doppler (TCD) velocities, obtained daily in all patients with aSAH, to detect and monitor the progress of vasospasm (Fig. 1). Mean velocities are recorded in all vascular territories and measured in cm/sec. Mild vasospasm as detected by TCDs is responded to with permissive hypertension (allowing the patient’s blood pressure to rise on its own and withdrawing the patient’s usual antihypertensive regimen, if one exists) initially and artificial hypertensive therapy if the TCD velocities rise into the moderate to severe range (1,15,17,18). The second component of triple-H therapy, hypervolemia, is important in that it facilitates the use of pressor agents to maintain the blood pressure within the desired range. Therefore, patients experiencing incremental rise in TCD velocities are hydrated with saline and albumin to achieve a central venous pressure of 8 to 12 mm Hg. Swan-Ganz catheters are routinely used in patients with significant cardiac histories; a cardiology consult is prudent to assist with the fluid management at this point. If the patient remains neurologically stable, TCD velocity alterations are treated with medical therapy alone and then gradually tapered in response to decreasing TCD velocities. A baseline single-photon emission computerized tomography (SPECT) is obtained on admission, and may be repeated in patients who deteriorate, to decipher between cerebral ischemia and other causes for the clinical decline. When the significance of the TCD results are in question, computed tomographic angiography and at times diagnostic angiography are often employed, particularly if the confirmation or repudiation of vasospasm would change the degree or duration of the therapy employed (15).

ENDOVASCULAR THERAPY Intra-arterial Antispasmodics Papaverine

Papaverine, a benzylisoquinoline alkaloid and a potent smooth muscle relaxant, is believed to act by the inhibition of phosphodiesterase activity in smooth muscle cells (15). First used to treat cerebral vasospasm in 1992, papaverine has been demonstrated to reproducibly result in vasodilatation in both

Chapter 14: Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm

265

Figure 1 TCD velocities accurately predict angiographic vasospasm. Anteroposterior left ICA angiograms on admission (left ) and several days after aSAH shows A1 and M1 segment severe stenosis (long arrows) on day 5, which correlates with markedly elevated TCD velocities (in the severe vasospasm range). Abbreviations: TCD, transcranial Doppler; ICA, internal carotid artery; aSAH, aneurysmal subarachnoid hemorrhage.

experimental animal models (19,20) and clinical practice (21–23). Perhaps the largest experience in endovascular management of aSAH-induced vasospasm has been with this agent (21,23–26). Administration of IAP is technically straightforward. Once angiography documents the degree and location of vasospasm, a microcatheter is navigated into the spastic vessels for the anterior circulation and just proximal to the area of spasm in the posterior circulation. Although dosage recommendations vary and should be based somewhat on the degree of vasospasm encountered, administration of 100 to 300 mg of the drug diluted in normal saline over 30 to 60 minutes is standard. This procedure and dosage may be repeated if different territories are involved (22,26). Because of its short duration of action, repeat procedures are often necessary (14.8–42% in three reports) (7,25,27). Reported clinical success with papaverine has, like the other endovascular approaches to vasospasm, been variable. Reported angiographic response is quite high with success in various studies seen anywhere from 57% to 90% of the time (21–23). In one study that tried to quantify blood vessel responsiveness on the basis of angiography, the authors found an average increase in vessel diameter of 26.5% in 34 patients undergoing 81 treatments (21). The success of papaverine has also been documented using indirect methods quantifying the increase in vessel diameter and CBF. Studies using

TCD ultrasound, Xenon 133, and SPECT have each lent supporting evidence that intra-arterially administered papaverine relieves vasospasm with resultant augmentation of blood flow (7,23). Polin et al. studied the effect of intra-arterially administered papaverine using TCD examinations on the day before and after treatment. They found that 41% of patients displayed greater than 20 cm/sec improvement in vasospasm parameters, with six individuals showing more than 50 cm/sec change (24). In a study comparing IAP to TBA, we showed that TCD velocities measured in the internal carotid artery (ICA) and middle cerebral arteries (MCAs) reliably showed vasodilatation after IAP (7). SPECT scanning in the same study showed increased perfusion in 31% of 37 patients in whom this test was performed after IAP (7). Oskouian et al. used TCD velocities to compare the vasodilatation seen after IAP versus that observed with a combination of TBA and IAP (28). Fandino et al., after studying jugular bulb venous oxygen saturation (SvjO2) and arteriovenous differences in lactate in 10 patients before and after IAP, demonstrated that IAP improved global perfusion. All 10 patients had an early improvement in neurologic function, nine of whom had improvement in the SvjO2; interestingly, no significant difference in the lactate was observed (29). In another study, researchers implanted thermal diffusion microprobes into the white matter of affected vascular territories in

266

Brisman et al.

eight patients with vasospasm. IAP resulted in a significant improvement in the CBF, with the increase being proportional to the degree of vasospasm and hypoperfusion prior to treatment (30). Hoh et al. conducted a literature review looking at the improvement in CBF after IAP and found an increase in CBF in 60% of patients and in 31% of vascular territories treated (16). Despite the success of IAP at reversing vasospasm, its use has been significantly hampered by the fact that the results appear to be temporary in most cases (7,21). Recurrent vasospasm with persistent neurologic deficits prompts repeat endovascular instillations. In one study reviewing the collective literature on 401 patients undergoing IAP for vasospasm, Hoh and Ogilvy found a total of 663 treatments or an average of 1.7 intra-arterial sessions per patient (16). This result is similar to the observations of TCD velocities before and after IAP (7). While we detected an average improvement in TCD velocity of 20% after papaverine infusion, these levels returned to baseline one day later (7). These findings correlate well with the study by Vajkoczy et al. in which CBF was found to return to pretreatment levels after just three hours (30). And although angiographic reversal of vasospasm is often accompanied by clinical improvement, this is not necessarily the result. In one study, angiographic success was demonstrated in 78% of patients while neurologic improvement was evident in only 26% of cases (31). Clinical improvement was found in 148 out of 348 patients (43%) in one systematic review of the literature on IAP (16). Timing of IAP relative to development of symptoms appears to be important, with patients in whom symptoms have been present for a longer time appearing less likely to respond (15). Numerous complications related to IAP have been described, with systemic hypotension and increased intracranial pressure (ICP) being the most common and the most serious. Fortunately, both of these adverse side effects seem to be related to the rate of infusion and therefore can be minimized with careful attention to how quickly the drug is administered (32). Frequent communication with the anesthesiologist and a heightened attention to the blood pressure and ICP (when such monitoring is present) can avoid serious complications. Systemic hypotension and/or increased ICP can be devastating to a patient with aSAH, in whom compromised cerebral perfusion because of vasospasm or increased ICP from hydrocephalus or brain swelling make even small fluctuations in the cerebral perfusion pressure dangerous. Additional, but less commonly reported side effects of IAP include pupillary dilatation (33), aneurysm perforation (33), tachycardia, respiratory depression (34), exacerbation of vasospasm (35,36), seizures (37), and severe neurologic deterioration associated with gray-matter destruction seen on MRI (15,16,38). Given the lack of sustained effect of papaverine and the host of potential side effects, many have moved away from IAP as a first-line therapy for medically refractory vasospasm. Its use remains paramount for patients with small vessel vasospasm not amenable to

balloon angioplasty and to temporarily open blood vessels to allow microcatheter navigation beyond the area of stricture to permit more definitive therapies (i.e., balloon angioplasty or aneurysm coiling) (15,31). Nicardipine, Verapamil, and Milrinone

Because of the myriad of side effects of IAP and its short-acting nature, some have explored the intraarterial administration of two additional calcium channel blockers (verapamil and nicardipine) and one phosphodiesterase inhibitor (milrinone) to treat medically refractory vasospasm (13,14,39,40). Intraarterial nicardipine (0.5–6 mg/vessel) resulted in significant improvement in angiographic vessel caliber and TCD velocities in 18 patients (44 vessels) thus treated. These radiographic results were accompanied by clinical improvement in eight patients (42%), with only one instance of transient elevation of ICP and no other adverse events (13). In another study, intraarterial verapamil was administered to treat vasospasm after SAH in 29 patients. Intra-arterial verapamil was successfully and safely employed in one study in which 10 patients were evaluated angiographically and clinically. An average dose of 3 mg/patient resulted in successful angiographic response in all 10 patients with an average vessel dilatation of 44
9%; only 6 of these 10 patients, however, had intra-arterial verapamil as the sole endovascular therapy. When intraarterial verapamil was used as the sole antispasmodic, 5 out of 17 (29.4%) patients showed clinical improvement treated with intra-arterial verapamil alone (14). There were no adverse sequelae related to the treatment. Milrinone, a potent inotrope, was given intra-arterially (dosage: 2.5–15 mg) and then continued intravenously (0.5–0.75 (mg/kg/min) for up to two weeks in seven patients presenting with cerebral vasospasm (40). Vasodilatation was demonstrated in all patients, with increased CBF in six of six patients in which it was measured and with no adverse effects. On the basis of these studies, it seems that further validation of the safety and efficacy of both these agents for this usage are warranted.

Balloon Angioplasty Overview

Currently, we use TBA of cerebral vessels as the preferred technique for the treatment of medically refractory vasospasm (7,15,31), reserving papaverine for vasospasm in the distal segments of the intracranial vasculature not suitable for TBA or to dilate proximal blood vessels sufficiently to allow passage of the balloon. TBA, pioneered in 1984 by Zubkov, carries somewhat higher risks of a major complication such as vessel rupture compared with intra-arterial infusion of vasodilators (7,9,41). In distinction with intra-arterial instillation of antispasmodics such as papaverine, however, the vasodilatation achieved with TBA is usually more sustained. When performed by experienced operators, side effects are less frequent (7), and the procedure can safely and reproducibly

Chapter 14: Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm

lead to vasodilatation of vessels in spasm with angiographic and clinical improvement in a large proportion of patients (6,7,15,28,31,42,43). Numerous pathophysiologic mechanisms underlying the successful effects of balloon angioplasty have been hypothesized and supported clinically as well as in experimental animal models (44–48). Using electron microscopy to study intracranial blood vessels of two patients who underwent TBA treatment for aSAH and subsequently died, Honma et al. found stretching and disruption of fibers throughout the vessel wall, including nonmuscular as well as muscular elements. These changes resulted in a persistent luminal dilatation seen at autopsy that was similar in both patients and concentrated in the medial layer. In a single case report, Zubkov documented their findings at postmortem analysis using both light microscopy and electron microscopy. Similar findings included stretching of the internal elastic lamina and muscular layer as well as proliferation of connective tissue. The subendothelial layer was thicker than that seen in non-angioplastied vessels and the internal elastic lamina displayed a corrugated pattern. Such findings have been replicated in animal studies. The ICA in 12 dogs was evaluated histologically after TBA and a loss of endothelial cells associated with flattening of the intima and internal elastic lamina was observed (48). Electron microscopy showed thinning of the internal elastic lamina with occasional areas of rupture and crowding of the smooth muscle cells in the tunica media. The increase in vessel lumen diameter (documented by both angiography and histologic analysis) persisted to seven days despite efforts to induce vasospasm by bathing the vessels in clotted blood. Using electron microscopy to study autopsied MCAs in which TBA was performed, Yamamoto et al. documented torn and stretched collagen fibers and postulated that disrupted connective tissue within blood vessels was responsible for the sustained effect of TBA (47). In the same study, a very similar disruption in the normal structure of collagen fibers was seen in two segments of MCA harvested from human autopsy studies and subjected to balloon angioplasty postmortem. Macdonald et al. found that in a rabbit model of vasospasm, TBA led to endothelial proliferation and smooth muscle cell layer thinning, which persisted at three to four weeks post-angioplasty. They postulated that this long-lasting histologic finding explained the durability seen with TBA (45). Since Zubkov’s initial description of the use of balloon angioplasty to treat 105 vessels in 33 patients, numerous other reports have surfaced documenting its angiographic and clinical success for the treatment of vasospasm after aSAH (6–12,28,41–43,49–53). Improvements in endovascular technology, such as newer softer balloons and better catheters, and increased operator experience have led to increased popularity of TBA and its use as a primary modality to treat medically refractory vasospasm in many centers. Properly executed TBA can be expected to result in clinical improvement, usually immediately noticeable, in 60% to 80% of patients, sustained angiographic

267

dilatation in nearly all accessible vessels, and improvements in CBF as evidenced by SPECT and TCD studies (7,15). Clinical and Radiographic Results

Numerous studies have been published documenting the clinical success associated with TBA. Although the results of each study must be interpreted based on patient selection, timing of intervention, and outcome measures, immediate neurologic improvement was noticed in upwards of 60% of patients. In one literature review on this topic, Hoh and Ogilvy collected all the patients treated with TBA for vasospasm; clinical improvement was noted in 328 out of 530 patients treated (16). In 1992, we reported a 72% clinical improvement rate for 41 patients treated with TBA, where success was defined by increase in Glasgow Coma Score by two points or significant improvement in speech or motor deficit (6). That clinical efficacy rate remained fairly constant (61% and 74% from two additional studies from our group) despite additional operator experience, suggesting a plateau after which increased operator experience plays less of a role, patient specific factors become more germane (7,41). Reports on TBA from other groups have not been as optimistic. One recent study reviewed the results of TBA for the patients forming the cohort evaluating the benefit of tirilizad in aSAH and found no benefit of TBA when compared with controls. The methodological flaws of this study can explain the discrepancy between this report and the majority of other reports documenting the utility of TBA. CT scan results of postplasty were assessed without the benefit of preplasty scans. This assessment is unacceptable given the finding that 22 out of 29 CT scans post-angioplasty revealed infarcts. Additionally, there was variable assessment of the severity of vasospasm and inconsistent technique related to the multicenter nature of the study (15 different centers participating) (42). Clinical success of TBA, however, does not necessarily accompany angiographic success. Our group and others have found a nearly 100% angiographic response to TBA (6,11,12,54). Patient specific factors such as pre-angioplasty clinical condition and the metabolic status of the brain that may be beyond the scope of basic imaging may help explain this discrepancy. It is therefore not surprising that the rate of clinical success with TBA more closely parallels successful increase in CBF as quantified by TCD (6,7,12,29,42), Xenon CT, Xenon clearance, and SPECT. Reports using TCD (16), regional blood flow as measured by SPECT (6), and CBF measured by Xenon clearance (28) have found improvements of 69%, 80%, and 58%, respectively, after TBA. Major complications associated with TBA have been recorded and include vessel rupture (41,55), vessel occlusion (12), hemorrhagic infarction of the vascular territory undergoing angioplasty (11), arterial dissection, and hemorrhage from unsecured aneurysms (12). In a literature review on this topic, a 5.0% major complication rate with a 1.1% vessel rupture rate was noted (16). Vessel rupture, the most serious

268

Brisman et al.

complication, is often lethal. Large series in the modern era, however, have now shown that using soft compliant balloons with gentle inflation techniques (as described below) can minimize, if not eliminate, this complication. Two recent reports at high-volume treatment centers have experienced no neurologic complications associated with TBA in a total of 115 patients treated (8,43). On the basis of the above data and on our own study in which we compared TBA to IAP, we have adopted the practice of using TBA as the first-line endovascular therapy for those with symptomatic vasospasm failing medical treatment. In our study, we noticed marked and more sustained improvements in TCD velocities using TBA compared with IAP (7). To our knowledge, no prospectively controlled randomized trial has been performed comparing the two modalities. We still make use of IAP in patients in whom we find angiographic distal spasm in which TBA would not be safe or as an adjunct to facilitate TBA to open severely stenotic proximal vessels to permit passage of the balloon. A comparison of the two major techniques (IAP and TBA) for endovascular treatment of medically refractory postSAH vasospasm appears in Table 2. Technique

Our technique for performing TBA for refractory vasospasm has undergone changes that reflect our increased experience with treating this entity, an increased understanding of the pathophysiology of this disease process, and the newer-endovascular devices available (15,31). We perform TBA under general anesthesia with full paralysis once a CT scan of the head documents that there is no obvious cause for the neurologic decline other than vasospasm, such as hydrocephalus or aneurysmal rebleeding, and there is no large territorial infarction. Completed infarction, depending on the size, is a relative contraindication to TBA, particularly if it is associated with a fixed neurologic deficit present for several hours or more. Such conditions markedly increase the chance that TBA will result in a reperfusion hemorrhage. With the currently available technology, TBA can

only be performed on the accessible large vessels at the base of the brain and is not effective for distal small-vessel vasospasm, which usually means any second order branch of an intracranial arterial tree (A2, M2, P2 segments). Once the decision is made to proceed with angioplasty, all areas of vasospasm visualized angiographically are treated, not just those yielding clinical symptoms. We treat the anterior circulation first. Angioplasty is performed with the patient systemically heparinized, achieved by giving an intravenous bolus of heparin of 5000 to 7000 units, once the vascular sheath is placed, followed by additional boluses of 1000 to 2000 units to maintain an activated coagulation time above 300 seconds. A 6-French (Fr) guide catheter is typically used with a 6-Fr vascular sheath placed within the common femoral artery. A rotating hemostatic valve attached to the guiding catheter prevents untoward movements of the balloon microcatheter. Using biplane fluoroscopy and careful guidewire manipulation, the guide catheter is advanced so that the tip sits within the distal cervical ICA or distal dominant vertebral artery, depending on which part of the circulation is being treated. The chosen balloon is preloaded with a microguidewire and introduced into the rotating valve. Through the use of fluoroscopic road-mapping technique in the anteroposterior and lateral projections, the balloon microcatheter is advanced to the vessel in vasospasm. It is important to maintain a high-quality road-map image throughout the procedure as well as to advance the guide catheter distally enough such that the tip is visible at all times. The most important factor about the choice of balloon is that it is soft and pliable, which usually is associated with balloons made of silicone or polyethylene as opposed to latex. Such balloons, referred to as ‘‘compliant’’ balloons, in distinction to balloons used for angioplasty of atheromatous disease, are less likely to result in vessel rupture. Currently, we prefer the Endeavor or Sentry balloon (Target Therapeutics/ Boston Scientific, Fremont, California, U.S.) (Fig. 2), although the Hyperglide balloon (Microtherapeutics, Inc., Irvine, California, U.S.) has been used with equal success.

Table 2 Comparison of TBA and IAP for Medically Refractory Vasospasm after SAH Technique Mechanism of action

Angiographic success (%)

Clinical success (%) Complications

TBAa

Intravascular balloon inflation, disruption of IEL and smooth muscle of media

83–100

62b

Major: 5% High angiographic Rupture: 1.1%b and clinical success, sustained effect

IAP

Slow intra-arterial infusion, inhibits phosphodiesterase in smooth muscle cells

57–90

43b

Major: 9.9%

a

Advantages

Moderate angiographic and clinical success, easy technique, can Rx.distal vasospasm

Disadvantages More serious complications for inexperienced operators, cannot Rx. distal vasospasm spasm, cannot Rx. distal Effect often transient, more frequent, albeit, less serious complications.

Preferred primary modality for refractory vasospasm. From Ref. 16. Abbreviations: TBA, transluminal balloon angioplasty; IAP, intra-arterial papaverine; SAH, subarachnoid hemorrhage; IEL, internal elastic lamina; Rx., treat.

b

Chapter 14: Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm

269

Figure 2 (A) The soft compliant nature of the Sentry balloon (3.5 mm  10 mm pictured) allows for safe angioplastic dilatation of vasospastic vessels. (B) We prefer using a 3-cc syringe for inflation, with the microguidewire removed, as shown here.

The Sentry balloon (usually 3.5 mm  10 mm, though other sizes exist) is a single-lumen balloontipped microcatheter with an end hole that accepts 0.010- to 0.014-inch guidewires, allowing the balloon microcatheter to be steered to the desired vessel of interest. The balloon is prepared by flushing with contrast material prior to preloading with the microguidewire. Once the balloon has been placed in the most distal region of the vessel to be treated, the wire is removed. With the microguidewire removed, the balloon can be gently inflated and then permitted to spontaneously deflate. Although the Sentry balloon is designed to undergo controlled inflation up to 4 atmospheres with the microguidewire inside, we have found this technique useful to avoid untoward overinflation for durations longer than desired. Through the use of biplane fluoroscopy and magnification road map, the Sentry balloon is navigated using a microguidewire. The microguidewire is then removed and a 3 cc syringe filled with contrast is attached to the end of the balloon microcatheter. Angioplasty is performed as a four-step process, successively increasing the diameter of the balloon inflation. We start with the anterior circulation and begin by placing

the balloon in the M1 segment and slowly inflate the balloon under fluoroscopic guidance. Gentle inflation of the balloon to approximately 25% of the maximal balloon volume and diameter is followed by successive inflations to 50%, 75%, and ultimately 100% (15). Particular care is given to avoid either overinflation or too rapid inflation, especially with the first inflation, as these maneuvers are most likely to result in vessel injury and/or rupture. To achieve an effect, however, we have found it necessary to hold the inflation for at least one to two seconds. TBA is started distally (generally speaking slightly beyond the M1 segment) and then moved proximally, with attempts made to overlap areas of inflation so as to ‘‘smooth’’ out the angiographic result and avoid leaving areas of residual stenosis. It is critical not to inflate the balloon to a diameter that exceeds the diameter of the native vessel; the importance of careful evaluation, when possible, of the original angiogram cannot be underscored. Once the MCA segment angioplasty has been satisfactorily performed, we move the catheter proximally toward the supraclinoid ICA and repeat the process (Fig. 3). Of note, we have found that it is sometimes easy to maintain partial inflation of the

Figure 3 (A) Vasospasm of the right supraclinoid ICA and M1 segment of the MCA is evident on this oblique projection of a right ICA injection angiogram. (B) Following balloon angioplasty of both segments (right internal carotid angiogram, same obliquity), markedly increased vessel diameters are achieved. Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery.

270

Brisman et al.

balloon as it is navigated into the supraclinoid ICA, as there is a tendency for the completely inflated balloon to slide past a focal area of stenosis. This phenomenon is believed to result from the pooling of subarachnoid blood in dural folds in the vicinity of the supraclinoid ICA. TBA of the A1 segment is technically feasible (56), but should only be undertaken by experienced operators in situations where there is reason to believe that the resultant vasospasm is truly symptomatic. Once study of the initial angiogram confirms that the A1 segment of interest is truly spastic and not simply congenitally aplastic or hypoplastic, TBA is performed in the same stepwise sequence as outlined for the MCA above. Most likely, augmenting flow by dilating the supraclinoid ICA will improve flow into the anterior cerebral artery territory, and therefore if any question exists, it is best to avoid A1 dilatation. We limit angioplasty to the MCA prior to its bifurcation into the M2 segments, ICA, and A1 segments. For the posterior circulation, the general technique is identical. The guide catheter is placed in the cervical segment of the vertebral artery and the balloon microcatheter is navigated into the basilar artery using road-mapping technique. It is here that we initiate TBA of the vertebrobasilar system. As with the A1 segments, careful attention must be given to the size of the P1 segments on the original angiogram to rule out congenital hypoplasia. Angioplasty proceeds from distal to proximal in a stepwise fashion, ending just proximal to the takeoff of the posterior inferior cerebellar artery. We generally do not dilate both the vertebral arteries. Timing

On the basis of the results of treatment for other forms of neurologic deficit, such as subdural hematomas, spinal cord compression, and cerebral ischemia of all varieties, it would seem intuitive that the sooner one treats a patient with symptomatic vasospasm, the more likely there will be a good outcome. Multiple authors have documented the applicability of this phenomenon to TBA (8,41,43). TBA and IAP cannot return function to infarcted brain and may even lead to reperfusion hemorrhage. In fact some authors support TBA, pointing out that some series with low clinical responses to TBA may have included patients in whom significant completed infarction had already developed (8,51). Optimal timing for endovascular intervention remains controversial. Whereas some are evaluating the utility of prophylactic angioplasty for those deemed highly likely to develop vasospasm (9), most await failure of medical (triple-H) therapy. Two studies looking at the issue of timing of endovascular therapy and its effect on benefit in patients with symptomatic vasospasm were published in 1998. Bejjani et al. described their results with TBA in 31 patients with refractory symptomatic vasospasm and found an increased likelihood of good recovery in patients treated within 24 hours of symptom onset (43). We found similar results in our early experience with TBA, but noted a 12-hour cutoff after the

development of symptoms beyond which outcomes appeared worse (41). Currently, we proceed to TBA if a patient demonstrates neurologic deficits despite two hours of maximal triple-H therapy. This management schema is derived from the results of Rosenwasser’s study in which he analyzed the effect of timing of TBA after aSAH in 84 patients. In that study, patients treated within two hours of symptom onset (n ¼ 51) had a 70% rate of early favorable outcome compared with those patients (n ¼ 33) treated after two hours (8). Given the fact that TBA is ineffective once infarction has developed and that the time window between development of symptoms and permanent ischemia is still not well defined, a study was designed to test the appropriateness of prophylactic angioplasty in select patients. In a pilot study, Muizelaar et al. studied the effectiveness and safety of performing prophylactic angioplasty in the patients most susceptible to vasospasm, those suffering Fisher grade 3 aSAH. Out of 13 patients enrolled in this study and treated with prophylactic TBA, none suffered symptomatic vasospasm and none developed severe vasospasm by TCD criteria (9). There was one procedural-related death associated with arterial rupture during TBA, prompting concern over the risk/benefit profile associated with prophylactic TBA. Because of the small series size, statistically powered recommendations cannot be made and a larger prospective randomized study is therefore underway to further define the role of prophylactic angioplasty in such patients.

Endovascular Treatment of Vasospasm in the Face of a Ruptured Unsecured Aneurysm: A Special Case On occasion, a patient who suffers aSAH may not be brought to medical attention until several days after the ictus. Some of these patients will be found to have significant vasospasm at the time of initial angiographic evaluation. This complication poses a unique management dilemma. If the patient is salvageable, aggressive treatment of both the aneurysm and vasospasm is indicated. In 1994, we reported on five patients who presented with severe vasospasm and ruptured aneurysms (57). All five were treated with craniotomy and clip ligation and then brought immediately thereafter to the angiography suite for balloon angioplasty of the narrowed vessels. This treatment resulted in improved TCD velocities and SPECT compared with pretreatment with good clinical outcome in four out of five patients (57). As endovascular coiling has become more popular and accepted as a primary modality to secure ruptured aneurysms, the idea that the aneurysm could be secured and the vasospasm treated in the same sitting came to the fore. In one report, 12 patients presenting with symptomatic vasospasm and ruptured aneurysms were treated with aneurysm coiling and simultaneous TBA (6 patients), IAP (2 patients), or both modalities (4 patients) (58). Recently, we reported on intentional endosaccular partial dome-coiling of two MCA aneurysms presenting with severe vasospasm followed by delayed clipping

Chapter 14: Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm

271

Figure 4 (A) Axial noncontrast CT scan revealing a small SAH in the left sylvian fissure. (B) Anteroposterior left internal carotid angiogram showing a bilobed MCA bifurcation aneurysm with significant M1 vasospasm. (C) Angiogram, similar view, immediately after partial dome-coiling and balloon angioplasty of the M1 segment. (D) Angiogram, similar view, six days later shows a normal M1 lumen diameter. (E, F) Postclipping angiogram digitally subtracted (E) and unsubtracted (F), similar views, shows successful obliteration of the aneurysm. Abbreviations: CT, computed tomography; SAH, subarachnoid hemorrhage; MCA, middle carotid artery. Source: From Ref. 59.

(Fig. 4) (59). Because of the morphology of the aneurysm, it was felt that clipping would ultimately provide the most definitive long-term cure of the lesion, but partial dome-coiling was deemed safer than craniotomy and clipping to prevent an early rehemorrhage in these patients who presented in the heat of vasospasm.

Predicting Outcome after Endovascular Therapy for Vasospasm and Long-term Effects Although both IAP and TBA appear to have a welldocumented radiographic and clinical success, it is not known why some patients respond and others do not. Why does one patient proceed to infarction and poor outcome despite good angiographic success? Rabinstein et al. sought to determine predictors of outcome after endovascular treatment of symptomatic vasospasm by studying 81 consecutive patients undergoing IAP, TBA, or both after aSAH (60). Using a logistic regression analysis they found that advanced age and

poor clinical grade on admission were associated with poor outcome in patients undergoing endovascular therapy for vasospasm. Perhaps because of the known temporary angiographic and clinical effect of IAP, the long-term effects of the intra-arterial administration of papaverine on the cerebral vasculature have not been studied. Long-term effects of TBA, however, have been studied in both animal models and humans. Using a canine model of vasospasm, Megyesi et al. demonstrated that the vascular response to TBA, both functionally and morphologically, was almost completely resolved by three weeks after angioplasty (61). We reported on our analysis of 28 consecutively treated patients with TBA for vasospasm after aSAH and evaluated using TCDs an average of 44 months after this procedure. No new neurologic events were noted in 21 patients with clinical follow-up. Normal TCD velocities were recorded in all patients as were dynamic autoregulatory studies, suggesting unimpaired vasomotor functioning (62).

272

Brisman et al.

CONCLUSIONS Endovascular therapy has proven to be an important adjunct in the treatment of medically refractory vasospasm. Although both intra-arterial administration of vasodilators and balloon angioplasty have been used as primary interventional modalities, most centers now favor the use of balloon angioplasty as a first-line agent because of its reproducible angiographic and clinical results, sustained effect, and decreased neurotoxicity. Low complication rates are now the norm for TBA when performed by experienced operators using modern day compliant balloons. Intra-arterial vasodilators continue to play an adjunct role in special circumstances and future studies on newer intra-arterially administered agents may increase their use. The results of the ongoing trial on the prophylactic use of balloon angioplasty for Fisher grade 3 aSAH as well as the continued refinements in balloon and catheter technology should further define and advance the role of interventional procedures in the treatment of this condition.

REFERENCES 1. Suarez JI, Tarr RW, Selman WR. Aneurysmal subarachnoid hemorrhage. N Engl J Med 2006; 354:387–396. 2. Fisher CM, Kistler JP, Davis JM. Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery 1980; 6:1–9. 3. Hoh BL, Topcuoglu MA, Singhal AB, et al. Effect of clipping, craniotomy, or intravascular coiling on cerebral vasospasm and patient outcome after aneurysmal subarachnoid hemorrhage. Neurosurgery 2004; 55:779–786. 4. Hohlrieder M, Spiegel M, Hinterhoelzl J, et al. Cerebral vasospasm and ischaemic infarction in clipped and coiled intracranial aneurysm patients. Eur J Neurol 2002; 9:389–399. 5. Rabinstein AA, Pichelmann MA, Friedman JA, et al. Symptomatic vasospasm and outcomes following aneurysmal subarachnoid hemorrhage: a comparison between surgical repair and endovascular coil occlusion. J Neurosurg 2003; 98:319–325. 6. Newell DW, Eskridge J, Mayberg M, et al. Endovascular treatment of intracranial aneurysms and cerebral vasospasm. Clin Neurosurg 1992; 39:348–360. 7. Elliott JP, Newell DW, Lam DJ, et al. Comparison of balloon angioplasty and papaverine infusion for the treatment of vasospasm following aneurysmal subarachnoid hemorrhage. J Neurosurg 1998; 88:277–284. 8. Rosenwasser RH, Armonda RA, Thomas JE, et al. Therapeutic modalities for the management of cerebral vasospasm: timing of endovascular options. Neurosurgery 1999; 44:975–979. 9. Muizelaar JP, Zwienenberg M, Rudisill NA, et al. The prophylactic use of transluminal balloon angioplasty in patients with Fisher Grade 3 subarachnoid hemorrhage: a pilot study. J Neurosurg 1999; 91:51–58. 10. Zubkov YN, Nikiforov BM, Shustin VA. Balloon catheter technique for dilatation of constricted cerebral arteries after aneurysmal SAH. Acta Neurochir (Wien) 1984; 70:65–79. 11. Higashida RT, Halbach VV, Cahan LD, et al. Transluminal angioplasty for treatment of intracranial arterial vasospasm. J Neurosurg 1989; 71:648–653. 12. Newell DW, Eskridge JM, Mayberg MR, et al. Angioplasty for the treatment of symptomatic vasospasm following subarachnoid hemorrhage. J Neurosurg 1989; 71:654–660.

13. Badjatia N, Topcuoglu MA, Pryor JC, et al. Preliminary experience with intra-arterial nicardipine as a treatment for cerebral vasospasm. AJNR Am J Neuroradiol 2004; 25:819–826. 14. Feng L, Fitzsimmons BF, Young WL, et al. Intraarterially administered verapamil as adjunct therapy for cerebral vasospasm: safety and 2-year experience. AJNR Am J Neuroradiol 2002; 23:1284–1290. 15. Scroop R, Britz GW, West A, et al. Endovascular therapy for vasospasm associated with subarachnoid hemorrhage. In: Le Roux P, Winn HR, Newell DW, eds. Management of Cerebral Aneurysms. Philadelphia: Saunders, 2004: 489–498. 16. Hoh BL, Ogilvy CS. Endovascular treatment of cerebral vasospasm: transluminal balloon angioplasty, intraarterial papaverine, and intra-arterial nicardipine. Neurosurg Clin N Am 2005; 16:501–516. 17. Zervas NT, Ogilvy CS. Cerebral vasospasm: current clinical management and results. Clin Neurosurg 1999; 45:167–176. 18. Wijdicks EF, Kallmes DF, Manno EM, et al. Subarachnoid hemorrhage: neurointensive care and aneurysm repair. Mayo Clin Proc 2005; 80:550–559. 19. Macdonald RL, Zhang J, Sima B, et al. Papaverinesensitive vasospasm and arterial contractility and compliance after subarachnoid hemorrhage in dogs. Neurosurgery 1995; 37:962–967. 20. Fujiwara N, Honjo Y, Ohkawa M, et al. Intraarterial infusion of papaverine in experimental cerebral vasospasm. AJNR Am J Neuroradiol 1997; 18:255–262. 21. Milburn JM, Moran CJ, Cross DT III, et al. Increase in diameters of vasospastic intracranial arteries by intraarterial papaverine administration. J Neurosurg 1998; 88:38–42. 22. Kassell NF, Helm G, Simmons N, et al. Treatment of cerebral vasospasm with intra-arterial papaverine. J Neurosurg 1992; 77:848–852. 23. Firlik KS, Kaufmann AM, Firlik AD, et al. Intra-arterial papaverine for the treatment of cerebral vasospasm following aneurysmal subarachnoid hemorrhage. Surg Neurol 1999; 51:66–74. 24. Polin RS, Hansen CA, German P, et al. Intra-arterially administered papaverine for the treatment of symptomatic cerebral vasospasm. Neurosurgery 1998; 42:1256–1264. 25. Numaguchi Y, Zoarski GH, Clouston JE, et al. Repeat intraarterial papaverine for recurrent cerebral vasospasm after subarachnoid haemorrhage. Neuroradiology 1997. 39:751–759. 26. Mathis JM, Jensen ME, Dion JE. Technical considerations on intra-arterial papaverine hydrochloride for cerebral vasospasm. Neuroradiology 1997; 39:90–98. 27. Liu JK, Tenner MS, Gottfried ON, et al. Efficacy of multiple intra-arterial papaverine infusions for improvement in cerebral circulation time in patients with recurrent cerebral vasospasm. J Neurosurg 2004; 100:414–421. 28. Oskouian RJ Jr., Martin NA, Lee JH, et al. Multimodal quantitation of the effects of endovascular therapy for vasospasm on cerebral blood flow, transcranial Doppler ultrasonographic velocities, and cerebral artery diameters. Neurosurgery 2002; 51:30–41. 29. Fandino J, Kaku Y, Schuknecht B, et al. Improvement of cerebral oxygenation patterns and metabolic validation of superselective intra-arterial infusion of papaverine for the treatment of cerebral vasospasm. J Neurosurg 1998; 89:93–100. 30. Vajkoczy P, Horn P, Bauhuf C, et al. Effect of intra-arterial papaverine on regional cerebral blood flow in hemodynamically relevant cerebral vasospasm. Stroke 2001; 32:498–505.

Chapter 14: Endovascular Treatment of Post-Subarachnoid Hemorrhage Vasospasm 31. Srinivasan J, Eskridge J, Grady MS, et al. Endovascular therapy for vasospasm. Clin Neurosurg 2002; 49:261–273. 32. McAuliffe W, Townsend M, Eskridge JM, et al. Intracranial pressure changes induced during papaverine infusion for treatment of vasospasm. J Neurosurg 1995; 83:430–434. 33. Andaluz N, Tomsick TA, Tew JM Jr., et al. Indications for endovascular therapy for refractory vasospasm after aneurysmal subarachnoid hemorrhage: experience at the University of Cincinnati. Surg Neurol 2002; 58:131–138. 34. Barr JD, Mathis JM, Horton JA. Transient severe brain stem depression during intra-arterial papaverine infusion for cerebral vasospasm. AJNR Am J Neuroradiol 1994; 15:719–723. 35. Clyde BL, Firlik AD, Kaufmann AM, et al. Paradoxical aggravation of vasospasm with papaverine infusion following aneurysmal subarachnoid hemorrhage. Case report. J Neurosurg 1996; 84:690–695. 36. Tsurushima H, Kamezaki T, Nagatomo Y, et al. Complications associated with intra-arterial administration of papaverine for vasospasm following subarachnoid hemorrhage–two case reports. Neurol Med Chir (Tokyo) 2000; 40:112–115. 37. Carhuapoma JR, Qureshi AI, Tamargo RJ, et al. Intraarterial papaverine-induced seizures: case report and review of the literature. Surg Neurol 2001; 56:159–163. 38. Smith WS, Dowd CF, Johnston SC, et al. Neurotoxicity of intra-arterial papaverine preserved with chlorobutanol used for the treatment of cerebral vasospasm after aneurysmal subarachnoid hemorrhage. Stroke 2004; 35:2518–2522. 39. Arakawa Y, Kikuta K, Hojo M, et al. Milrinone reduces cerebral vasospasm after subarachnoid hemorrhage of WFNS grade IV or V. Neurol Med Chir (Tokyo) 2004; 44:393–400. 40. Arakawa Y, Kikuta K, Hojo M, et al. Milrinone for the treatment of cerebral vasospasm after subarachnoid hemorrhage: report of seven cases. Neurosurgery 2001; 48:723–728. 41. Eskridge JM, McAuliffe W, Song JK, et al. Balloon angioplasty for the treatment of vasospasm: results of first 50 cases. Neurosurgery 1998; 42:510–516. 42. Polin RS, Coenen VA, Hansen CA, et al. Efficacy of transluminal angioplasty for the management of symptomatic cerebral vasospasm following aneurysmal subarachnoid hemorrhage. J Neurosurg 2000; 92:284–290. 43. Bejjani GK, Bank WO, Olan WJ, et al. The efficacy and safety of angioplasty for cerebral vasospasm after subarachnoid hemorrhage. Neurosurgery 1998; 42:979–986. 44. Chan PD, Findlay JM, Vollrath B, et al. Pharmacological and morphological effects of in vitro transluminal balloon angioplasty on normal and vasospastic canine basilar arteries. J Neurosurg 1995; 83:522–530. 45. Macdonald RL, Wallace MC, Montanera WJ, et al. Pathological effects of angioplasty on vasospastic carotid arteries in a rabbit model. J Neurosurg 1995; 83:111–117. 46. Honma Y, Fujiwara T, Irie K, et al. Morphological changes in human cerebral arteries after percutaneous transluminal angioplasty for vasospasm caused by subarachnoid hemorrhage. Neurosurgery 1995; 36:1073–1080.

273

47. Yamamoto Y, Smith RR, Bernanke DH. Mechanism of action of balloon angioplasty in cerebral vasospasm. Neurosurgery 1992; 30:1–5. 48. Megyesi JF, Findlay JM, Vollrath B, et al. In vivo angioplasty prevents the development of vasospasm in canine carotid arteries. Pharmacological and morphological analyses. Stroke 1997; 28:1216–1224. 49. Firlik AD, Kaufmann AM, Jungreis CA, et al. Effect of transluminal angioplasty on cerebral blood flow in the management of symptomatic vasospasm following aneurysmal subarachnoid hemorrhage. J Neurosurg 1997; 86:830–839. 50. Fujii Y, Takahashi A, Yoshimoto T. Effect of balloon angioplasty on high grade symptomatic vasospasm after subarachnoid hemorrhage. Neurosurg Rev 1995; 18:7–13. 51. Coyne TJ, Montanera WJ, Macdonald RL, et al. Percutaneous transluminal angioplasty for cerebral vasospasm after subarachnoid hemorrhage. Can J Surg 1994; 37:391–396. 52. Nemoto S, Abe T, Tanaka H. In: Sano K, Takakura K, Kassell N, et al., eds. Cerebral Vasospasm. Tokyo: University of Tokyo Press, 1990:437–439. 53. Takahashi A, Yashimoto T, Mizoi K. Transluminal balloon angioplasty for vasospasm after subarachnoid hemorrhage. In: Sano K, Takakura K, Kassell N, et al., eds. Cerebral Vasospasm. Tokyo: University of Tokyo Press, 1990:429–432. 54. Bracard S, Picard L, Marchal JC, et al. Role of angioplasty in the treatment of symptomatic vascular spasm occurring in the post-operative course of intracranial ruptured aneurysms. J Neuroradiol 1990; 17:6–19. 55. Linskey ME, Horton JA, Rao GR, et al. Fatal rupture of the intracranial carotid artery during transluminal angioplasty for vasospasm induced by subarachnoid hemorrhage. Case report. J Neurosurg 1991; 74:985–990. 56. Eskridge JM, Song JK, Elliott JP, et al. Balloon angioplasty of the A1 segment of the anterior cerebral artery narrowed by vasospasm. Technical note. J Neurosurg 1999; 91:153–156. 57. Le Roux PD, Newell DW, Eskridge J. et al. Severe symptomatic vasospasm: the role of immediate post-operative angioplasty. J Neurosurg 1994; 80:224–229. 58. Murayama Y, Song JK, Uda K, et al. Combined endovascular treatment for both intracranial aneurysm and symptomatic vasospasm. AJNR Am J Neuroradiol 2003; 24:133–139. 59. Brisman JL, Roonprapunt C, Song JK, et al. Intentional partial coil occlusion followed by delayed clip application to wide-necked middle cerebral artery aneurysms in patients presenting with severe vasospasm. Report of two cases. J Neurosurg 2004; 101:154–158. 60. Rabinstein AA, Friedman JA, Nichols DA, et al. Predictors of outcome after endovascular treatment of cerebral vasospasm. AJNR Am J Neuroradiol 2004; 25:1778–1782. 61. Megyesi JF, Vollrath B, Cook DA, et al. Long-term effects of in vivo angioplasty in normal and vasospastic canine carotid arteries: pharmacological and morphological analyses. J Neurosurg 1999; 91:100–108. 62. Srinivasan J, Moore A, Eskridge J, et al. Long-term follow up of angioplasty for cerebral vasospasm. Acta Neurochir Suppl 2001; 77:195–197.

15 Endovascular Management of Brain Arteriovenous Malformations John B. Weigele and Robert W. Hurst Department of Radiology; Department of Radiology, Neurology, and Neurosurgery, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A.

Riyadh N. Al-Okaili Department of Radiology, King Abdulaziz Medical City, Riyadh, Saudi Arabia

INTRODUCTION Brain arteriovenous malformations (AVMs) are relatively rare central nervous system lesions that are the cause of significant long-term morbidity and mortality. Current therapeutic options include microvascular neurosurgery, stereotactic radiation (radiosurgery), and endovascular embolization. Embolization is an important, well-established modality for brain AVM treatment that is usually combined with surgery or stereotactic radiosurgery. Embolization is, however, associated with significant risks that must be carefully balanced against the potential benefits in each patient. Embolization performed by experienced interventional neuroradiologists in appropriately selected cases improves the overall safety and efficacy of brain AVM treatment.

CLASSIFICATION AND PATHOGENESIS OF CEREBRAL VASCULAR MALFORMATIONS Cerebral vascular malformations have been studied since the 18th century. Nonetheless, clinically useful classification schemes have only been developed recently. Initially, vascular malformations were categorized by their gross pathological appearance, resulting in confusing and contradictory nomenclature that created a barrier to understanding their etiology, natural history, and clinical management (1). A new biological classification for vascular lesions was proposed in 1982 (2). Two major categories were identified: hemangiomas and vascular malformations. Lesions with growth potential shown by proliferation of endothelial cells with active DNA synthesis were defined as hemangiomas and were considered vascular neoplasms. Lesions without endothelial cell proliferation or active DNA synthesis and displaying proportionate growth were named vascular malformations and were thought to be

hamartomas rather than neoplasms. Vascular malformations were subdivided into arterial, capillary, venous, lymphatic, and combined types (2). Four categories of intracranial vascular malformations have been defined on the basis of gross and microscopic pathological data: AVM, capillary telangiectasia, cavernous malformation, and venous malformation (3–5). A mixed malformation also has been described (6). These have been considered congenital lesions, present from birth without the potential for significant cellular proliferation or de novo postnatal development. Brain AVMs have cerebral arterial feeders directly connected to the venous system without an intervening capillary bed, resulting in high-flow arteriovenous (AV) shunts. The nidus (Latin for nest) contains the direct AV connections. The vessels in the nidus vary in size and histology from relatively well-differentiated arteries and veins to thick- and thin-walled, hyalinized, malformed vessels that are neither. Dilated segments of vessels commonly occur. There is gliotic brain parenchyma within and around the nidus. Gross or microscopic calcification may be present with the vascular walls or in the gliotic parenchyma. Hemosiderin is commonly present, indicative of some degree of prior hemorrhage. The gross pathological appearance has been aptly described as a ‘‘bag of worms’’ (5). Most brain AVMs occur sporadically; however, they also are associated with a number of congenital or hereditary syndromes, including: Rendu-OslerWeber syndrome (hereditary hemorrhagic telangiectasia), Klippel-Trenaunay syndrome, Parks-Weber syndrome, Wyburn-Mason syndrome, and SturgeWeber disease (7). Rare familial cases not associated with syndromes also have been described (8). There is recent evidence that not all brain AVMs are congenital in origin (7). Although the large majority probably occurs congenitally because of the failure of capillary formation during early embryogenesis (9), some brain AVMs appear to form in response to a postnatal stimulus of angiogenesis, particularly in

276

Weigele et al.

younger patients. The de novo development of brain AVMs in a child (10) and in an adult (11) has been reported. Also, brain AVMs have reoccurred in children after complete surgical resection (12).

EPIDEMIOLOGY OF BRAIN AVMs Most of the estimates of the prevalence of brain AVMs are flawed and potentially inaccurate. The widely quoted prevalence estimates of 500 to 600/100,000 were based on biased autopsy data. Another erroneous estimate of 140/100,000 was based on an inappropriate analysis of the Cooperative Study of Intracranial Aneurysm and Subarachnoid Hemorrhage data. These estimates may represent greater than 10-fold overestimates of the true prevalence (13). A comprehensive review of the published literature performed in 2001 identified only three populationbased studies of the incidence and/or prevalence of brain AVMs, all retrospective in nature (14). The Mayo Clinics identified a total of 48 intracranial vascular malformations in the population of Olmstead County, Minnesota, over a period of 27 years, from 1965 to 1992. The brain AVM detection rate was 1.11 (95% CI, 0.7–1.5) per 100,000 person-years (15). The incidence of symptomatic brain AVMs was 1.1 (95% CI, 0.6–1.8) per 100,000 patient-years in the Leeward Islands of the Netherlands Antilles between 1980 and 1990 (16). A retrospective study in the Lothian region of Scotland found a minimum point prevalence of 15 symptomatic brain AVMs per 100,000 in unselected living adults (17,18). In 2003, an ongoing, prospective populationbased study of brain AVMs in the New York Islands (Manhattan, Staten Island, and Long Island) with 9.4 million residents reported an AVM detection rate of 1.34 (95% CI, 1.18–1.49) per 100,000 person-years. The estimated prevalence of brain AVM hemorrhage within the detected cases was 0.68 (95% CI, 0.57–0.79) per 100,000 (19). The currently available data do not suggest that there is a large reservoir of asymptomatic brain AVMs in the general population, but that most brain AVMs become symptomatic during life (20).

NATURAL HISTORY OF BRAIN AVMs The risks of treating a brain AVM must be weighed against the natural history of the disease, in particular the possibility that a brain AVM will hemorrhage or rehemorrhage if it is not treated and the associated potential clinical consequences. Unfortunately, little unbiased natural history data are available, in part, because brain AVMs are relatively rare and quite heterogeneous and also because most undergo some form of treatment. No level I or level II natural history studies have been published (21). Data on specific predictors for the clinical course of a specific brain AVM are even more limited. In many natural history studies there is a selection bias toward untreatable AVMs. In addition, natural history outcomes usually have not been correlated with the type of presentation, the analyses have differed and follow-up periods have been short (14).

Despite these limitations, some general observations can be made about the natural history of brain AVMs. Clinical presentation can occur at any age, with the mean age of presentation in the fourth decade of life. There is an essentially equal distribution between sexes (14). Brain AVMs most commonly present with intracranial hemorrhage, epilepsy, headache, or a focal neurological deficit, although they are occasionally found incidentally (14). Intracranial hemorrhage is the most common form of clinical presentation (21). In a prospective population-based study published in 1996, 65% of patients newly diagnosed with a brain AVM presented with intracranial hemorrhage (22). Intraparenchymal hemorrhage occurred in 41% of these cases, subarachnoid hemorrhage in 24%, intraventricular hemorrhage in 12%, and a combination of these types in 23% of cases. The more recent prospective population-based study of brain AVMs in the New York Islands reported that 38% of patients with newly found AVMs presented with intracranial hemorrhage (19). Hospital-based case series have been retrospectively analyzed to identify risk factors for brain AVM hemorrhage (14). These findings have not been confirmed by prospective population-based studies. The features most consistently associated with an increased risk of hemorrhage include deep venous drainage, a single draining vein, venous stenoses, and high-feeding mean arterial pressure (14). These may share the common hemodynamic mechanism of associated high intranidal pressures (23). Less consistent risk factors for hemorrhage are a small AVM size, feeding artery and intranidal aneurysms, and deep or posterior fossa locations (14). Sex and pregnancy do not appear to increase the risk of hemorrhage (24). Features that may be associated with a decreased risk of hemorrhage include a large AVM size (25), arterial stenosis and ectasia (26), dural arterial supply (27), venous recruitment (28), and angiogenesis (29). The second most common form of clinical presentation is epilepsy. In one retrospective population-based study, 19% of newly discovered AVMs presented with seizures (22). In two retrospective hospital-based studies, 18% and 27% of AVMs presented with seizures, respectively (30,31). Other less common brain AVM presentations include headache (1% and 11% in two hospital-based series) (30,31), focal neurological deficit (7% and 5% in two hospital-based series) (30,31), and as an incidental finding in an asymptomatic individual (15% in one population-based study, 0% and 3% in two hospitalbased series) (22,30,31). There are extremely limited data on the natural history of brain AVMs following the initial diagnosis. An annual 2% to 4% risk of first-ever hemorrhage from a brain AVM is widely quoted on the basis of a few hospital-based series (30,32–35). No prospective, population-based study of the clinical course of unruptured brain AVMs has been published (14). After an initial bleed, the risk for recurrent hemorrhage may be as high as 18% in the first year (31). This appears to subsequently decrease to the baseline 2% to 4% annual risk of hemorrhage over time (36). Fatality rates from brain AVM hemorrhage range from 0% to 18% during the first year (22,30,31,35,37).

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

277

Brain AVMs demonstrate AV shunting on angiography, resulting in early opacification of the draining veins and a decrease in the AV transit time (38). This shunting is the result of a direct connection between the arterial and venous sides of the cerebral circula-

tion without an intervening capillary bed. There are two types of AV connections: fistulous and plexiform (39). A fistulous nidus contains large-caliber direct AV connections (Fig. 1). A plexiform nidus consists of a conglomerate of multiple smaller and more numerous vascular channels supplied by one or more arterial feeders (Fig. 2). These are collected into one or more draining veins. A plexiform nidus can contain one or more direct fistulas (mixed plexiform-fistulous nidus; Fig. 3) (39,40). The complete angiographic evaluation of a brain AVM consists of (1) the selective evaluation of the AVM and the entire cerebral circulation using 4- or 5-French (Fr) diagnostic catheters and (2) the superselective angiographic evaluation of the feeding arterial pedicles, the nidus, and the venous drainage using microcatheters advanced into distal aspects of the arterial feeders (38). The goals of the selective angiographic evaluation are listed in Table 1. This provides an important assessment of the arterial supply to the AVM, the general characteristics of the nidus, the venous drainage of the AVM, and the rest of the intracranial circulation. Selective angiography, however, has significant limitations. Rapid AV shunting often superimposes the arterial feeders, the nidus, and the draining veins obscuring important features, such as small arterial feeders, distal feeding pedicles, nidal aneurysms, direct AV fistulas, and small accessory draining veins (38). The goals of superselective angiography are listed in Table 2. Such detailed anatomic information from superselective angiography concerning the distal

Figure 1 Large, fistulous AVM (large arrows in A and B). Note proximal arterial aneurysms (small arrows, A), venous ectasia (arrowheads, B), and venous aneurysms (small arrows in C and D). Also note nonvisualization of normal anterior and middle cerebral arterial territories due to vascular steal. (A) AP angiogram-arterial phase, (B) lateral angiogram-arterial phase, (C) AP angiogram-venous phase, and (D) lateral angiogramvenous phase. Abbreviation: AVM, arteriovenous malformation.

Figure 2 Plexiform AVM (arrowheads in A, B, and C). (A) AP angiogram, (B) lateral angiogram, (C) superselective angiogram (microcatheter tip, arrow), and (D) lateral postembolization angiogram (arrow, residual nidus). Abbreviation: AVM, arteriovenous malformation.

The reported long-term annual fatality rates are 1% to 1.5% (30,35). Ondra et al. prospectively evaluated 166 untreated symptomatic brain AVM patients over a mean follow-up period of 23.7 years. There was a 4.0% annual rate of hemorrhage and a 1.0% annual mortality rate. The combined rate of mortality and major morbidity was 2.7% per year. Over the followup period, 23% of the patients died from hemorrhage. The incidence of bleeding and death were the same whether the AVM initially presented with hemorrhage (35). Recurrent hemorrhage appears to cause a morbidity rate similar to the initial bleed (36). In another hospital-based study, 47% of patients with a first-ever hemorrhage sustained no neurological defect and 37% experienced no significant disability despite symptoms (Rankin 1) (37). Parenchymal hemorrhage had a greater likelihood (52%) of producing a neurological defect.

ANGIOGRAPHY AND ANGIOARCHITECTURE OF BRAIN AVMs Selective and Superselective Cerebral Angiography

278

Weigele et al.

Figure 3 Mixed plexiform (A, arrowheads) and fistulous (B, arrow) nidus in an 11-month old. (A) Selective lateral angiogram and (B) superselective lateral angiogram.

Figure 4 Sulcal AVM. (A) Triangular nidus (arrow) on lateral angiogram. (B) NBCA in nidus (arrow) on unenhanced axial CT image. Abbreviation: NBCA, N-butyl cyanoacrylate.

arterial feeders, the nidus, and the proximal draining veins is critical for planning and performing endovascular embolizations (38).

A sulcal AVM nidus occupies the subpial space of the sulcus. The nidus may remain contained within the sulcus or variably extend through the sulcus into the cerebral cortex, into the subcortical white matter, and into the deep white matter to the ventricular wall. Sulcal AVMs assume a conical or pyramidal shape conforming to the sulcal space (Fig. 4). Their most superficial aspect is covered by the meninges, not by parenchyma. Because of this, meningeal arterial supply to their superficial aspect is common. Pial arteries are their primary supply. These end in the nidus after providing cortical and medullary branches to adjacent gyri (terminal feeders). This terminal supply is usually amenable to safe embolization. Larger sulcal AVMs also receive supply from basal perforating arteries (38). Gyral AVMs are covered by cortex and are typically spherical (Fig. 5). The gyrus usually is enlarged

Classification of Brain AVMs

Brain AVMs are categorized into superficial (cortical) or deep types. Cortical AVMs are subcategorized into sulcal, gyral, and mixed (sulcogyral) types. Deep AVMs are subdivided into subarachnoid, deep parenchymal, plexal, and mixed types (38).

Table 1 Goals of Selective Angiographic Evaluation of Brain AVMs 1. Arterial territories supplying the AVM 2. Feeding pedicles 3. High-flow arteriopathy (stenoses, ectasias, flow-related aneurysms) 4. Nidus (size, shape, location, flow, fistulas, ectasias, aneurysms) 5. Venous drainage (territories, deep, superficial) 6. Individual draining veins 7. High-flow venous angiopathy (dural sinuses, venous stenoses, occlusions, and varices) 8. Venous drainage of normal brain parenchyma Abbreviation: AVM, arteriovenous malformation. Source: From Ref. 114.

Table 2 Goals of Superselective Angiographic Evaluation of Brain AVMs 1. Distal feeding pedicles (anatomy, aneurysms, geometry, hemodynamics) 2. Arterionidal junction 3. Nidus (compartments, direct AV fistulas, plexiform regions, intranidal ectasias, and aneurysms) 4. Venonidal junction 5. Proximal aspects of the draining veins Abbreviations: AVM, arteriovenous malformation; AV, arteriovenous. Source: From Ref. 114.

Figure 5 Gyral AVM. Axial T2-weighted MRI (A) and lateral angiogram (B) demonstrate a small gyral AVM (arrow). Abbreviation: AVM, arteriovenous malformation.

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

279

recent years has added considerably to the understanding of brain AVM angioarchitecture. Feeding Arteries

Figure 6 Deep AVM (large arrow in A and B) on lateral angiogram (A) and axial T2-weighted MRI (B). Note venous ectasia (arrowhead, A) and venous aneurysm (small arrow, A). Abbreviation: AVM, arteriovenous malformation.

and adjacent sulci are compressed. A large gyral AVM may extend into the subcortical white matter toward the ventricular wall. The arterial supply is primarily from pial branches that continue beyond the AVM to supply normal parenchyma (indirect feeders). Meningeal supply typically is absent because the overlying cortex is positioned between the nidus and the meninges. Basal perforating arteries may supply the deeper extension of a large gyral AVM (38). Mixed (sulcogyral) types usually are large AVMs that combine both sulcal and gyral features. The AVM typically involves gyri and sulci, extending into the subcortical white matter to the ventricular wall. The arterial supply combines meningeal arteries and terminal pial branches from the sulcal component, nonterminal pial branches from the gyral component, and basal perforating arteries (38). Deep AVMs are relatively rare. They can be subdivided into subarachnoid, deep parenchymal, plexal, and mixed types. Subarachnoid AVMs are found in the basal cisterns and fissures, supplied by the subarachnoid portions of the choroidal and perforating arteries. Deep parenchymal AVMs are located in deep gray and white matter such as the thalamus, basal ganglia, and corpus callosum (Fig. 6). Basal perforators, choroidal arteries, basal circumferential arteries, and medullary pial branches supply them. Plexal AVMs are intraventricular, primarily supplied by the choroidal arteries. Mixed deep AVMs are typically larger, combining subarachnoid, deep parenchymal, and plexal features. Venous drainage is predominately into the deep venous system; however, transmedullary cortical venous drainage also is seen (38).

Angioarchitecture of Brain AVMs The routine use of superselective angiography in addition to conventional selective cerebral angiography in

The classification of the arterial feeders to a brain AVM using anatomic, geometric, and hemodynamic criteria is essential for planning and performing endovascular embolization. Pial supply may be provided by extracortical (subpial), cortical, medullary, and/or corticomedullary branches. Meningeal supply may be direct or through transdural pial anastamoses (Fig. 7). Collateral supply can occur through leptomeningeal and subependymal anastamoses. Choroidal artery supply can arise from the extraventricular (fissural, parenchymal) or intraventricular portions (39,40). Geometric classification of arterial feeders defines the relationship of the distal feeder with the nidus and normal parenchyma. Three types are defined on superselective angiography: terminal, pseudoterminal, and indirect (39,40). The terminal feeder ends within the nidus distal to branches supplying normal brain. Terminal feeders are usually large, facilitating their superselective catheterization. Embolization is relatively safe if the catheter tip is positioned distal to branches to normal parenchyma (39,40). The pseudoterminal feeder appears to end in the nidus, but actually continues beyond to supply normal brain. The distal segment is not angiographically visible because of the high flow (sump effect) into the nidus. Its presence must be inferred on an anatomic basis. A wedged catheter position during superselective angiography can contribute to the nonvisualization of the distal portion. Changing hemodynamic conditions during embolization of a pseudoterminal feeder can cause the embolic material to occlude the distal portion to normal brain, resulting in an ischemic complication (39,40). The indirect feeder (feeder en passage) is a branch to the nidus

Figure 7 External carotid supply to AVM from middle meningeal (arrow, B) and occipital (arrowheads, B) branches. (A) AP angiogram and (B) lateral angiogram. Abbreviation: AVM, arteriovenous malformation.

280

Weigele et al.

Figure 8 En passage supply during embolization of a plexiform AVM. (A) AP angiogram shows superficial (arrow, A) and deep (arrowhead, A) venous drainage. (B) Lateral angiogram shows venous ectasia (arrowhead, B) and venous aneurysm (arrow, B). (C) Lateral superselective angiogram shows en passage supply (arrowheads, C) to nidus (large arrow, C) and microcatheter tip (small arrow, C). (D) Lateral angiogram shows residual nidus after embolization (arrowheads, D). Abbreviation: AVM, arteriovenous malformation.

arising from an artery that passes in proximity to the nidus while continuing on to supply normal brain (Fig. 8). Indirect feeders are typically smaller and shorter, usually originating at an acute or right angle from the parent vessel. Superselective catheterization is often feasible, but more difficult. The parent vessel may be enlarged up to the origin of the indirect feeders and smaller beyond (39,40). Feeding arteries may be characterized hemodynamically into dominant or supplementary feeders according to the amount of flow. Dominant feeders supply a large portion of the nidus, are larger, and carry more flow than supplementary feeders. Dominant and supplementary feeders can arise from the same or different vascular territories. Most cerebral AVMs contain a combination of both types of feeders (38). High-flow angiopathy results in stenoses in the feeding arteries in up to 20% of brain AVMs. These may be isolated, proximal stenoses intrinsic to the vessel wall or rarely caused by extrinsic compression. Diffuse stenoses with a moyamoya appearance are occasionally seen in younger patients (40). Arterial stenoses associated with decreased distal tissue perfusion may result in a shift in the watershed zone toward the nidus (watershed transfer), occurring in up to 30% of superficial (cortical) AVMs. Cortical arteries and leptomeningeal collaterals are recruited to supply more of the territory distal to the AVM. This shift in arterial supply may perfuse just the normal parenchyma or also include the distal aspect of the AVM nidus. Angiogenesis can occur with watershed transfer in response to chronic parenchymal ischemia. It may be mistaken for part of the nidus; however, angiogenesis has no AV shunting and is not a true AVM component (40).

AVM Nidus

The nidus is considered the region between very distal aspects of the readily identifiable arterial feeders and the proximal aspects of the draining veins. AV shunting occurs at this site and represents the primary target of embolization. Complete obliteration of the nidus results in a cure (39,40). Most brain AVMs have a compact, well-defined nidus with well-demarcated borders, discrete feeding arteries, and draining veins. A minority has diffused and ill-defined margins. Angiogenesis associated with watershed transfer may mimic a diffused nidus. Nidal sizes vary tremendously. Their shapes tend to conform to their anatomic environments. Sulcal AVMs are usually conical (Fig. 4), gyral and subcortical white matter AVMs tend to be spherical (Fig. 5), and deep AVM shapes vary with location (callosal, cisternal, etc.) (Fig. 6). Larger AVMs have more complex shapes reflecting their involvement with multiple anatomic structures (38). Superselective angiography has led to the concept of the nidal vascular compartment, referring to an intranidal vascular unit consisting of one or more feeding arteries supplying the region of AV shunting with a unique draining vein. A nidus may be composed of one or multiple vascular compartments of varying sizes and flow patterns. The AV connections within a given compartment may be plexiform, fistulous, or mixed. These compartments are often hemodynamically interconnected, so occlusion of the compartmental feeders without occlusion of the compartmental zone of AV shunting may allow the compartment to continue to fill from neighboring units. Compartmental vein occlusion can increase the risk of

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

nidal rupture. Hence, careful characterization of the compartmental angioarchitecture is essential for planning an embolization (38). Histological studies have described the nidus as a complex system of coiling and intercommunicating vascular channels emptying into tortuous thin-walled collecting veins. Three zones have been described within the nidus: arterial, intermediate, and venous. The arterial zone consists of a plexus of interconnecting thick-walled vessels. The intermediate zone is very heterogeneous, containing four types of coiled, interconnected channels ranging from 0.15 to 1.0 mm in diameter. The venous zone consists of 1- to 3-mm thin-walled vessels converging into the draining veins. AV shunting is thought to occur between the arterial and intermediate zones (41). Draining Veins

The location of a brain AVM usually predicts the pattern of venous drainage; however, there are frequent variations. Cortical AVMs (sulcal and gyral) typically drain through cortical veins into the nearby dural sinuses. Those with subcortical or ventricular extension often have both superficial (cortical) and deep (subependymal) venous drainage. Central AVMs usually drain into the deep venous system. However, unexpected patterns, such as transcerebral cortical venous drainage of a deep AVM or deep venous drainage of a cortical AVM, may be seen approximately 30% of the time. These variants may represent venous collaterals that developed after occlusion of the original venous drainage system (40). Important aspects of the nidal venous drainage include venous anatomic variations, collateral venous drainage, and high-flow angiopathy (Fig. 9). Anatomic variations in venous drainage develop in response to hemodynamic effects, such as persistence of embryonic veins and variations in the cerebral veins and dural sinuses. Collateral venous drainage is acquired as a response to obstruction including ipsilateral, contralateral, and transcerebral rerouting of venous

281

drainage. This may be due to mechanical venous compression or intrinsic venous stenoses or thromboses due to high-flow angiopathy. Insufficiently developed collateral venous drainage may result in venous hypertension, venous aneurysms, and venous ectasia (varix) proximal to the obstruction, especially in highflow AVMs (Fig. 1). Clinical symptoms may result from direct compression of the brain or cranial nerves by the varix, seizures or neurological deficits from venous hypertension (Fig. 9), and hemorrhage from AVM rupture (38,40). Aneurysms Associated with Brain AVMs

Introduction. The association of aneurysms with brain AVMs has been reported for many years although until recently little was known regarding the frequency or clinical implications of their concomitant occurrence. The publication of relatively large series has enabled study and understanding of some aspects of the association between these two cerebrovascular lesions. Although much remains to be understood, it is clear that significant clinical and therapeutic implications may arise from the relatively common association of these two cerebrovascular lesions. Classification. Classification of aneurysms associated with AVMs was first proposed by Hayashi et al. in 1981 (42). These authors included only aneurysms external to the AVM nidus in their classification. They divided extranidal aneurysms into three groups depending on whether they were located proximally or distally on vessels giving supply to the AVM or were located on vessels unrelated to the AVM supply. In 1994, the Tew’s classification (Table 3) divided AVM-associated aneurysms into four groups based on their relationship to the AVM nidus and feeding arteries (43). This classification also includes intranidal aneurysms (Type IV). This comprehensive and relatively straightforward classification has the advantage of suggesting potential mechanisms for aneurysm formation based on location. In addition, it has proven useful in attempts to relate aneurysm types to clinical behavior. Differences in definitions are often seen, however, with some series simplifying the classification into three or even two groups of aneurysms. These often overlap and have included aneurysms unrelated to the AVM supply; flow-related aneurysms, which have been subdivided into those located either proximally or distally on arteries supplying the AVM; and intranidal aneurysms (44–46). Nevertheless, definitions of AVM-associated aneurysms similar to those outlined in the Tew’s Table 3 Tew’s Classification of AVM-Associated Aneurysms Type I Type II

Figure 9 Large, high-flow AVM with left transverse sinus occlusion (A, arrow) and right sigmoid sinus stenosis (B, arrow) causing venous hypertension and cognitive impairment. (A) AP angiogram-venous phase and (B) lateral angiogram-venous phase. Abbreviation: AVM, arteriovenous malformation.

Type III Type IV

Dysplastic or remote, not related to AVM supply Proximal, arising from the circle of Willis or origin of a vessel supplying the AVM Pedicular, arising from the midcourse of a feeding pedicle Intranidal, within the AVM nidus

Abbreviation: AVM, arteriovenous malformation. Source: From Ref. 43.

282

Weigele et al.

Figure 11 Proximally located aneurysms: (A) M1-2 junction and (B) basilar tip.

Figure 10 (A) Distal flow-related aneurysm, (B) additional example of distal flow-related aneurysm, (C) intranidal aneurysm, (D) proximal flow-related aneurysm, and (E) non-flow-related aneurysm. Source: From Ref. 47.

classification have become relatively standardized with publication of ‘‘Reporting Terminology for Brain Arteriovenous Malformation Clinical and Radiographic Features for use in Clinical Trials’’ (Fig. 10) (47). Epidemiology. The reported prevalence of aneurysms associated with AVMs varies widely among series (Table 4). In an early examination of the subject, the First Cooperative Study of Intracranial Aneurysms and Subarachnoid Hemorrhage found intracranial aneurysms associated with 6.2% of 545 AVMs (48). Similarly, in a large series of 600 AVMs, Thompson et al. identified 7.5% patients whose AVMs were associated with extranidal aneurysms (49). Other series have noted prevalence of extranidal aneurysms as high as 17.6% (Figs. 11 and 12) (50). In 1994, Turjman et al., using superselective angiography, demonstrated a considerably higher prevalence, which included a group of aneurysms located within the angiographic boundaries of the AVM nidus and which filled prior to filling of significant portions of the nidus. These intranidal aneurysms were identified

Figure 12 Distal feeding artery aneurysm on anterior choroidal artery.

in 58% of 100 consecutive AVMs (51). Meisel et al. evaluated 662 AVM patients and identified 46% with associated aneurysms. Among 305 patients having both aneurysms and AVMs, 372 of the aneurysms were identified as intranidal, with 313 located on vessels supplying the AVM (46). Halim et al. evaluated 336 AVM patients, 82 from University of California at San Francisco (UCSF) and 254 from Columbia Presbyterian Medical Center (CPMC). Their evaluation also included intranidal as well as extranidal aneurysm types. They found similar overall aneurysm prevalence at both institutions with 34% in the UCSF patients and 29% from patients evaluated at CPMC (45). Redekop et al., however, found a somewhat lower frequency and identified aneurysms in association with 16.7% of 632 AVMs of which 5.5% were intranidal aneurysms (52). Overall, the reported prevalence of AVMassociated aneurysms falls generally into the range of 15% to 25%. Nevertheless, prevalence in individual series varies 10-fold, from 5.8% to 58% in different series. This variability in reported aneurysm prevalence may arise from a number of factors. These include poor interobserver agreement as to what

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

283

Table 4 Reports of AVM-Associated Aneurysms Year

48 39 50 63 51 52

1966 1987 1990 1992 1994 1998

490 414 91 400 100 632

37 (7.6%) 45 (11%) 16 (17.6%) 39 (9.8%) 58 (58%) 97/632 (15.3%)

49 46

1998 2000

45 (7.5%) 305 (46.1%)

51% 205/305 (67.2%)

67

2000

2000 2001 2002

Total: 13/222 (5.8%); Supratent 3.5%; infratent 20.8% 25 (14.5%) 30 (11%) 117 (25.3%)

1/222 (0.045%)

53 65 44

600 662 (450 Rxd) 222: 198 supra, 24 infratent 172 270 463

45

2002

Total of 33682/254

28 (34%) (UCSF)/74 (29%)(CPMC)

4852

929/4852 (19.1%)

Total

AVMs

Multiple aneurysms

Reference

Aneurysms (Pts)

On feeding vessel

Intranidal

Unrelated

15/34 (37%) 42/45 (93%) 25 63/64 (98%)

NE NE NE NE

18/34 (43%) 3/45 (7%)

71 pts (11.2%) had 123 flow-related aneurysms 30/45 (66%) 138/450 (30.7%) All by definition

35 (5.5%)

5/632 (0.8%)

NE 181/450 (40%) NE

15/45 (33%)

1/64 (2%)

34/58 (58.6%)

18/25 (72%) 14/30 (47%) 24/117 (20.5%)

All 77/463 (17%); 54/117 (46%) 16/82 (23%); 35/254 (16%)

NE 35/463 AVMs (8%); (21/117 (17.9%) 11/82 (17%); 23/254 (11%)

1/222 (0.045%)

32/463 (7%); 18/117 (15.3%) 2/82 (4%); 10/254 (5%)

Abbreviation: AVM, arteriovenous malformation; NE, not evaluated.

constitutes an AVM-associated aneurysm, particularly aneurysms within or in proximity to the AVM nidus. In addition, differing definitions, data collection methodology, and inclusion criteria, all impact on the diagnosis and classification of AVM-associated aneurysms. For example, whether aneurysms located within the AVM nidus are included can be expected to impact the overall numbers of aneurysms identified. Lastly, heterogeneity of study populations, including referral bias, contributes to the variable prevalence reported in the literature. Most series indicate that AVM-associated aneurysm prevalence is similar in men and women. The frequency of AVM-associated aneurysms has been noted, however, to increase with patient age, as well as flow rate and size of the AVM nidus. Reported multiplicity of AVM-associated aneurysms is common but also quite variable. Ezura et al., for example, found multiple aneurysms in 18 of the 25 patients (72%) to be AVM-associated aneurysms in their series, while Meisel et al. found 67% of their AVM-associated aneurysm cases to have multiple aneurysms (46,53,54). Data suggest that close to half of patients with AVM-associated aneurysms will have more than one aneurysm, and a significant proportion will have more than two aneurysms. Pathogenesis of AVM-associated aneurysms. Three major theories have addressed the association of AVMs and aneurysms. These include coincidental occurrence, an underlying congenital vascular defect responsible for both lesions, and the flow-related or hyperdynamic theory. Early suggestions were that the relationship between the two lesions was one of simple chance

occurrence. For adults without specific risk factors, the prevalence of intracranial aneurysms alone has been estimated at approximately 2.3% (54). Coincidental occurrence may explain some AVM-associated aneurysms, particularly those located on vessels unrelated to the AVM. As noted above, however, most studies report a considerably higher frequency of aneurysms occurring in association with AVMs than would be expected by chance, and this theory is currently given little credence. To date, no underlying congenital defect has been identified to explain an association between aneurysms and AVMs. While a number of genes have been found to be differentially expressed in AVM-feeding pedicles compared with normal cerebral vessels, current information suggests that this differential expression is likely a consequence of increased flow dynamics rather than an underlying cause of the AVM (55). Although AVMs are considered congenital lesions, only a very limited number are associated with, and perhaps arise from, inherited genetic defects. An increased prevalence of brain AVMs has been most closely associated with hereditary hemorrhagic telangiectasia (HHT), an autosomal dominant disorder characterized by AV shunts involving many organ systems. MRI evaluation of a population of 184 HHT patients demonstrated brain AVMs in 5.6% of these patients (56). While reports of aneurysms in HHT exist, there is no indication that an increased frequency of arterial aneurysms characterizes patients with this disorder. Conversely, a large number of inherited or congenital conditions have been associated with an

284

Weigele et al.

increased prevalence of intracranial aneurysms. Adult polycystic kidney disease and coarctation of the aorta may engender aneurysm formation, likely through the mechanism of systemic hypertension. A number of other heritable disorders, many of which involve defects in connective tissue, have also been associated with an increased prevalence of intracranial aneurysms (57). No convincing evidence has been presented to suggest an increase in brain AVMs associated with any of these conditions. Flow phenomena associated with the AVM provide a logical mechanism to explain the greater than expected prevalence of aneurysms on AVM feeding vessels. This theory, initially articulated by McKissock over 50 years ago, bases the development of aneurysms on the hyperdynamic flow resulting from AV shunt through the AVM (58). The low resistance through the AVM and consequent increased velocity in feeding arteries places increased shear stress on the vessel walls. While aneurysm formation is undoubtedly multifactorial, abnormal shear stress acting on the arterial wall has been found to play a role in formation, growth, and rupture of all types of arterial aneurysms (59–61). Indeed, no histological or imaging features have been found to distinguish flow-related aneurysms occurring in association with AVMs from those aneurysms occurring in the absence of AVMs. Support for a hyperdynamic flow mechanism comes from the observed tendency of aneurysms to arise far more commonly on vessels providing arterial supply to the AVM, than on unrelated vessels. An analysis of 78 reported cases of AVM-associated aneurysms, by Okamoto et al., showed a significant correlation of aneurysm location to vessels supplying the AVMs (62). This was reinforced by the results reported by Cunha e Sa et al. who found 98% of AVM-associated extranidal aneurysms on vessels directly supplying the AVMs (63). In addition, Redekop et al. identified flowrelated aneurysms in 11.2% of 632 AVMs, while aneurysms on vessels unrelated to the AVM were found in only 0.8% (52). These authors also found a tendency for aneurysms to occur more often on arteries feeding larger AVMs, and therefore associated with higher flow, than on arteries supplying smaller AVMs. Additional support for a hyperdynamic flow mechanism underlying many AVM-related aneurysms arises from frequently observed changes in aneurysm size following treatment of the AVM and consequent decrease in the AV shunt. Examination of 23 proximal aneurysms following complete AVM obliteration revealed disappearance of one (4.3%) and a decrease in size of four (17.4%) aneurysms. In the same patient population, four (80%) of five distally located aneurysms regressed completely and one was unchanged (52). These data emphasize the effect of AVM flow on changes in aneurysms located along feeding arteries. They also suggest that more distally located aneurysms are more sensitive to alterations in flow than those located more proximally. Attribution of aneurysm formation to flow phenomena helps to explain the formation and clinical behavior of a large proportion of AVM-associated aneurysms. In addition, this explanation aids classifi-

cation of AVM-associated aneurysms and may suggest guidelines for management as noted below. Nevertheless, hemodynamic mechanisms do not fully explain the incidence of all AVM-associated aneurysms, even those confined to direct feeding vessels. That other, perhaps individual features, are involved is evidenced by the fact that most AVMs are not associated with aneurysms. Additional controversy surrounds the etiology of intranidal aneurysms. Many believe that these lesions represent true aneurysms located in the most distal arterial branches adjacent to the AVM nidus. Other authors have suggested, however, that some of these lesions represent early filling of dilated venous pouches rather than true arterial aneurysms, while others may represent pseudoaneurysms arising as residua of prior hemorrhages (64). Clinical implications. The natural history of AVMs associated with aneurysms has been the subject of considerable controversy and little firm agreement. In large part, this uncertainty is due to the heterogeneity of AVM-associated aneurysms as well as to the heterogeneity of AVMs themselves. Some data suggest that increased rates of both initial and recurrent hemorrhage occur in patients who have AVMs with concomitant aneurysms. Piotin et al. found that 50% of their patients with AVM-associated aneurysm presented with intracranial hemorrhage. Of these, 80% had bled from their aneurysms (65). Similarly, Batjer et al. found that for patients who harbored both lesions and presented with intracranial hemorrhage, 78% had bled from the associated aneurysms (66). Cunha e Sa et al. identified the source of intracranial hemorrhage in patients with AVM-associated aneurysms as the aneurysm in 46% of their series (63). Brown et al. emphasized the long-term risk of harboring an aneurysm in association with an AVM. They found that patients with AVM-associated aneurysms had an annual hemorrhage risk of 7% at five years following diagnosis. This was significantly higher than the 1.7% annual hemorrhage rate for those AVM patients without coexisting aneurysms (50). Studies of clinical behavior also suggest that important differences may characterize different types of AVM-associated aneurysms. Intranidal aneurysms have been associated with a higher incidence of initial hemorrhage as well as with multiple episodes of recurrent bleeding. For example, Redekop et al. noted intracranial hemorrhage associated with 38% of their series of 632 AVMs. Presentation with intracranial hemorrhage occurred in 72% of patients with intranidal aneurysm; 36% without aneurysm; and 40% with flow-related or unrelated aneurysms. These authors also found an annual hemorrhage rate of nearly 10% among patients with intranidal aneurysms who were not treated (52). In contrast, Meisel did not find initial hemorrhage to be correlated with any type of aneurysm. In their experience, however, intranidal aneurysms demonstrated a significantly higher rebleeding rate before treatment (46). Similarly, Thompson et al. were unable to attribute any increased risk of an initial hemorrhagic presentation to the existence of aneurysm (49).

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

Figure 13 Distal feeding artery aneurysm on superior cerebellar artery, (A) AP and (B) lateral view.

The difficulty in evaluating risks of AVMassociated aneurysms is emphasized by Halim et al. These investigators studied the association between AVM-associated aneurysms and presentation with intracranial hemorrhage in 336 patients from two referral populations, at CPMC and UCSF. While aneurysm prevalence was similar at both institutions, they found that initial presentation with hemorrhage was associated with a coexisting aneurysm at CPMC, while the opposite trend was seen in the UCSF population (45). Despite the difficulty in formulating risk profiles associated with AVM-associated aneurysms in general, aneurysms seem to present more of a risk of hemorrhage when located closer to or within the AVM nidus. In addition, intriguing data from small number of patients suggest that AVM-associated aneurysms may behave in an especially aggressive fashion when occurring in certain locations. Westphal’s series of

285

222 patients with AVMs included 198 supratentorial AVMs and 24 located infratentorially (67). They identified aneurysms associated with 5 out of 24 (20.8%) of infratentorial AVMs, a considerably larger proportion than the 3.5% associated with their AVMs in a supratentorial location. Their findings also suggested more aggressive behavior associated with aneurysms in the infratentorial group. Of the five aneurysms associated with infratentorial AVMs, four were located distally and 75% of these were responsible for intracranial hemorrhage (Figs. 13 and 14). Data suggesting a relatively aggressive course of cerebellar feeding pedicle aneurysms are supported by Kaptain’s evaluation of 27 cerebellar AVMs, all of which were associated with aneurysms (68). Eightynine percent of this population presented with symptoms of aneurysm rupture. The vast majority of aneurysms, i.e. 29 out of 36 (81%), was located distally on arteries supplying the AVM. These authors also concluded that aneurysms associated with cerebellar AVMs frequently present with rupture. In addition, the usually distal location made the aneurysms treatable in a large proportion of cases without significant risk to the brain stem. Finally, Khaw et al. reported a higher prevalence of aneurysms associated with infratentorial AVMs in comparison to those in a supratentorial location. Infratentorial AVMs with associated aneurysms were also significantly more likely to present with intracranial hemorrhage than those without associated aneurysms (69). It is not clear to what extent aneurysms increase the risk of intracranial hemorrhage when they occur in association with AVMs. In part, the difficulty arises from identifying the incremental risk associated with the aneurysm in the face of an already high risk of hemorrhage presented by the AVM. Sufficient data

Figure 14 (A). Intranidal aneurysm responsible for AVM hemorrhage (arrow). (B) Unsubtracted and (C) subtracted microcatheter injection of feeding pedicle (arrow: aneurysm; arrowhead: catheter tip). (D) Postembolization, no filling of pedicle (arrow) or aneurysm. Residual AVM treated with radiosurgery. Abbreviation: AVM, arteriovenous malformation.

286

Weigele et al.

should be strongly considered to be the source of hemorrhage and treatment initially directed at this feature of the AVM (46). While detailed imaging and identification of the aneurysm adjacent to the hematoma may often confirm the source of the hemorrhage, this may not be possible in all cases. The specific treatment depends on the overall treatment plan for the AVM. If surgically treatable, intranidal aneurysms may be addressed by surgical resection of the entire AVM, possibly preceded by presurgical embolization. If performed, endovascular treatment should be targeted to close the feeding pedicle from which the aneurysm originates first in order to minimize the chance of subsequent hemorrhage. If the AVM is to be treated with radiosurgery, preradiosurgery embolization directed at the intranidal aneurysm may also be a reasonable plan to decrease the risk of hemorrhage during the period required for obliteration of the AVM. While the AVM should be addressed initially when identified as the source of hemorrhage, the effect of closure of the AV shunt on associated aneurysms must also be considered, regardless of which lesion is responsible for the initial presentation. Differing recommendations have been made with respect to prophylactic treatment of feeding artery aneurysms in conjunction with AVM hemorrhage. Thompson et al. found that of 45 aneurysms identified in their 600 patients (7.5%), five bled prior to treatment while two bled within three weeks

exist, however, to indicate that patients with AVMassociated aneurysms are likely at higher risk of hemorrhage than are patients with an AVM alone. While this seems particularly true for infratentorial AVMs, it apparently holds as well for AVMs in all locations associated with intranidal aneurysms, and to a lesser extent for extranidal aneurysms. Given this information, a number of management approaches to patients identified as having AVM-associated aneurysms have been suggested. Treatment approaches. Specific management recommendations for patients having both aneurysm(s) and an AVM are difficult to formulate because of the relatively sparse and often conflicting data. General guidelines have been suggested, however (Table 5). For patients presenting with hemorrhage, the first step is to determine which lesion was responsible for the hemorrhage. It is that lesion toward which the initial treatment should be directed. As noted, in posterior fossa AVMs, higher proportions of hemorrhage may result from feeding artery aneurysms than is the case in supratentorial locations (Fig. 15). When no determination can be made as to the source of hemorrhage, the greater morbidity and higher chance of repeat hemorrhage from an aneurysm dictates that the aneurysm be addressed as a first priority. In cases where the AVM is identified as the source of hemorrhage, initial treatment is directed at that lesion. Most data indicate that intranidal aneurysms associated with AVMs presenting with hemorrhage

Table 5 Management of AVM-Associated Aneurysm Patients with Hemorrhage Aneurysm hemorrhage

Aneurysm location

Rx

AVM resectable

Intranidal aneurysm Distal pedicle aneurysm Proximal aneurysm Unrelated aneurysm Intranidal aneurysm Distal pedicle aneurysm Proximal aneurysm Unrelated aneurysm

Both lesions simultaneously Aneurysm first—consider AVM resection at same time Aneurysm first—consider AVM resection at same time Aneurysm first Embolization of aneurysm & pedicle Aneurysm Aneurysm Aneurysm

AVM hemorrhage

Aneurysm location

Rx

AVM resectable AVM unresectable

Any Intranidal aneurysm

Rx AVM first—consider distal pedicle aneurysm treatment if low risk Embolization of intranidal aneurysm feeder first Radiosurgery, consider embolization of intranidal feeder first Consider aneurysm Rx first (controversial) AVM first AVM first

AVM unresectable

Distal pedicle aneurysm Proximal aneurysm Unrelated aneurysm Unknown hemorrhage source

Aneurysm location

Rx

AVM resectable

Intranidal aneurysm Distal pedicle aneurysm Proximal aneurysm Unrelated aneurysm Intranidal aneurysm Distal pedicle aneurysm Proximal aneurysm Unrelated aneurysm

Rx AVM and aneurysm Rx AVM and aneurysm Rx aneurysm first Rx aneurysm first Embolization of intranidal feeder first Rx aneurysm Rx aneurysm Rx aneurysm

AVM unresectable

Abbreviation: AVM, arteriovenous malformation.

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

287

Figure 15 Embolization of an ectatic distal arterial feeder and aneurysm supplying a cerebellar AVM presenting with hemorrhage. (A) Selective and (B) superselective lateral vertebral angiograms show a cerebellar AVM supplied by an ectatic distal feeder (arrowhead ) containing an aneurysm (arrow). (C) Postembolization with NBCA, no filling of pedicle or aneurysm. (D) Unenhanced CT image shows an NBCA-filled aneurysm (arrow) surrounded by hyperdense hemorrhage (arrowheads ). Abbreviations: AVM, arteriovenous malformation; NBCA, N-butyl cyanoacrylate.

following AVM treatment. Their experience led them to recommend treatment of aneurysms on feeding vessels prior to definitive treatment of the AVM (49). Similarly, Ezura et al. treated feeding artery aneurysms endovascularly prior to treating the AVM with either resection or radiosurgery (53). However, others suggest that decreasing flow through the AVM results in frequent regression of extranidal aneurysms without the need for direct treatment. For example, Redekop et al. reported complete spontaneous regression of aneurysms on distal feeding arteries in 80% of cases after curative therapy of the AVM (52). These authors also noted shrinkage of proximally located aneurysms in 18% of cases with complete disappearance in 4%. The effect on aneurysms appeared to be less in cases of incomplete AVM treatment. Of 16 patients with less than 50% reduction in the AVM, no aneurysms regressed, although two enlarged and bled. In cases with greater than 50% reduction in the AVM, two of three distal aneurysms disappeared and five proximal aneurysms were unchanged. Meisel et al. also reported significant regression of feeding artery aneurysms after treatment of the AVM (46). In 83 treated patients with 149 proximally located aneurysms, they found complete regression in 8% and more than 50% shrinkage in 22%. The shrinkage of proximally located aneurysms was influenced by the degree of AVM occlusion and occurred faster for those aneurysms on midline vessels, such as the anterior cerebral artery and the circle of Willis. Their

finding that no hemorrhage from untreated proximally located aneurysms occurred after partial treatment of AVMs led them to conclude that proximally located aneurysms should not be treated primarily in cases where the AVM is the source of hemorrhage. Summary. Aneurysms are found in association with a significant proportion of AVMs. They may occur within the AVM nidus, on routes of flow to the AVM, or on vessels unrelated to the AVM supply. Those on vessels providing supply to the AVM likely arise from the high flow resulting from the AVM. AVM-associated aneurysms present an increased risk of hemorrhage, particularly those that are intranidal or distally located on feeding pedicles. They may regress spontaneously following AVM treatment, with those more distally located being more affected by flow and pressure changes associated with AVM obliteration. In any case, AVM-associated aneurysms represent an additional source of potential morbidity that must be considered in formulating treatment of the AVM patient.

GRADING SYSTEMS FOR BRAIN AVMs Because of the tremendous variability of brain AVMs with respect to their anatomy and biological behavior, a number of studies have attempted to correlate specific criteria of AVM characteristics with therapeutic outcomes to guide clinical decision-making. Most of these grading systems have focused on surgical management.

288

Weigele et al.

Luessenhop and Gennarelli were the first to assign a grade to a brain AVM in an effort to predict operability and outcome. Grading was based on the number of arteries feeding the AVM and the vascular territory that was involved. Other criteria such as size, anatomic location, degree of vascular steal, and venous drainage were not included, limiting the system’s utility (70). In 1986, two new grading systems were proposed. Shi and Chen categorized AVMs on the basis of four criteria: (1) size, (2) location and depth, (3) complexity of the arterial supply, and (4) complexity of the venous drainage. Each criterion was attributed a grade of I to IV on the basis of a detailed analysis that was related to the operative risk. A composite grade was assigned on the basis of the grades for the individual criteria. This complex system did not gain widespread usage (71). Spetzler and Martin also proposed a grading system for brain AVMs in 1986 that has become the most widely utilized. The authors sought a system that was simple and applicable to all brain AVMs providing a reasonable estimate of operative morbidity and mortality. They considered a number of parameters, including the AVM size, the number of feeding arteries, the anatomic location, the operative accessibility, the amount of flow, degree of vascular steal, the eloquence of nearby brain parenchyma, and the venous drainage pattern (72). Recognizing that a grading system that attempted to incorporate all of the potential parameters would be too complex to be practical, and that many of those variables were interrelated, they proposed a simplified grading system based on three criteria: the AVM size, the venous drainage pattern, and the eloquence of the adjacent brain parenchyma. The AVM size was divided into three categories: small (<3 cm), medium (3 to 6 cm), and large (>6 cm). The venous drainage was designated as superficial only if all of the venous drainage emptied into the cortical venous system. If any or all of the venous drainage egressed through deep veins (internal cerebral, basal veins, precentral cerebellar vein) it was categorized as deep. An AVM was considered to be adjacent to eloquent brain parenchyma if it was next to the sensorimotor cortex, language areas, visual cortex, hypothalamus, thalamus, internal capsule, brain stem, cerebellar peduncles, or deep cerebellar nuclei (72). When using the Spetzler-Martin grading system, points are assigned for the AVM size, the venous drainage pattern, and the location relative to eloquent brain (Table 6). The points for each parameter are added for the total score (1–5) that corresponds to the Spetzler-Martin grade (I–V). For example, a 2-cm anterior frontal lobe AVM with cortical venous drainage (1 point for size, 0 points for venous drainage, 0 points for eloquence) is a Spetzler-Martin grade I AVM, whereas a 4-cm thalamic AVM with deep venous drainage (2 points for size, 1 point for venous drainage, 1 point for eloquence) is a Spetzler-Martin grade IV AVM. AVMs with no possibility of surgical resection (e.g., diffuse brain stem or holohemispheric involvement) are assigned a grade of VI (72).

Table 6 Spetzler-Martin grading system for brain AVMs AVM Feature Size of nidus <3 cm (small) 3–6 cm (medium) >6 cm (large) Eloquence of adjacent brain Noneloquent Eloquent Venous drainage Superficial Deep

Points 1 2 3 0 1 0 1

The assigned grade equals the sum of the points for all the three features. Abbreviation: AVM, arteriovenous malformation. Source: From Ref. 72.

Following resection of Spetzler-Martin grade I and II AVMs, the authors’ retrospective evaluation of their personal surgical experience found a low incidence of minor deficits (0%, 5% in grade I AND II, respectively) and no major neurological deficits; grade IV and V AVM resections were associated with significant incidences of both minor deficits (20%, 19% in grade IV and grade V, respectively) and major deficits (7%, 12% in grade IV and grade V, respectively) (72). A subsequent prospective evaluation confirmed the accuracy of the Spetzler-Martin grading system for predicting both new-temporary and new-permanent neurological deficits (73). A recent analysis demonstrated interobserver variability between a neuroradiologist and a neurosurgeon performing SpetzlerMartin grading in 27.7% of patients; however, this variability did not diminish the predictive value of the Spetzler-Martin scale (74).

EMBOLIZATION OF BRAIN AVMs Historical Background In a landmark publication in 1960, Luessenhop and Spence reported the first therapeutic embolization of a brain AVM (75). Because this report predated the development of selective cerebral angiography, the authors injected Silastic spheres directly into a surgically accessed cervical internal carotid artery. Their technique relied on the much greater rate of blood flow to the AVM to direct the spheres into the nidus rather than into normal cerebral branches; however, this flow-dependent embolization was unreliable and was associated with a significant risk of causing an ischemic infarct. Another problem was that the relatively large spheres lodged in the proximal feeders and did not penetrate into the nidus. The nidus remained unoccluded and could recruit deep perforating arteries that were much more difficult to control during surgery (76–79). In 1974, Serbinenko was the first to report superselective cerebral artery catheterization and embolization using a detachable balloon attached to a flexible, flow-directed catheter (80). Superselective catheterization of the target vessel with the balloon catheter was

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

not always technically feasible because it depended on the arterial geometry and hemodynamics. Similar to the problem with the flow-directed Silastic spheres, the detachable balloons occluded the proximal arterial feeders inducing the nidus to recruit new blood supply from other branches that were often more difficult to control during surgery. This early experience suggested the AVM nidus should be the target of therapeutic embolization. In 1976, Kerber set the stage for the development of modern therapeutic brain AVM embolization techniques. He reported the use of a microcatheter with a calibrated-leak balloon to superselectively catheterize cerebral arteries and to deliver a liquid embolic agent [isobutyl-2-cyanoacrylate (IBCA)] into the AVM nidus (81). However, both the catheter and the embolic agent had serious limitations. Calibrated-leak balloon catheters were difficult to use and were associated with multiple complications including vascular injuries. The catheters also could be permanently glued into the AVM (82). In addition, the IBCA transformed the AVM into a hard, incompressible mass with illdefined borders and containing embolized vessels that were difficult to surgically cut or coagulate (82,83). IBCA also was reported to be carcinogenic and associated with toxic reactions (84,85). These pioneering efforts provided the foundation and stimulus for the development of the microcatheters, guidewires, and embolic agents that are currently used for AVM embolization. This early experience also established the concept of targeting the nidus for embolic occlusion, and defined the risks of embolizing too proximally (the arterial feeders) and too distally (the venous outflow).

Embolization Indications Current therapeutic options for brain AVMs include embolization, microvascular surgery, stereotactic radiation (radiosurgery), and various combinations. The goal of any combined therapy is to decrease the overall morbidity and mortality of AVM treatment. In many centers, brain AVM embolization is most commonly performed before microsurgical resection. Embolization is also performed prior to radiosurgery. In this setting, the goal is to permanently occlude enough of the nidus so that stereotactic radiation can target the rest with a higher dose and a better chance for cure. Less frequently, embolization is used as a stand-alone curative technique, especially for small, surgically difficult lesions. Occasionally, embolization is employed for the palliation of symptoms from an otherwise untreatable AVM. Ideally, a multidisciplinary team consisting of a microvascular neurosurgeon, an interventional neuroradiologist, and a radiation therapist collectively evaluates and formulates an individualized plan for each patient. Presurgical Embolization

Microvascular surgery is the principle method to treat brain AVMs at many centers. Neurosurgical outcomes have improved with advances in stereotactic guidance, electrophysiological monitoring, barbiturate

289

anesthesia, intraoperative angiography, and aggressive perioperative blood pressure control. Many small, superficial brain AVMs can be surgically resected without preoperative embolization with minimal morbidity and mortality (73). Nonetheless, preoperative embolization results in improvements in overall treatment outcomes for many brain AVMs. Brain AVM embolization can improve surgical outcomes through several mechanisms. Often, the most valuable contribution is the elimination of deep or surgically inaccessible feeders. The deep arterial supply is approached through the nidus toward the end of the surgical resection, and gaining surgical control can be treacherous. Preoperative embolization of the deep arterial supply can allow an otherwise inoperable AVM to be successfully resected (86). Also, embolization can decrease the size of the nidus and the amount of blood flow through the AVM resulting in shorter surgical times and less blood loss. Embolized vessels also are easily identified during surgery. This provides a road map for the resection of the arterial feeders and nidus while preserving en passage arteries to nearby eloquent parenchyma (Fig. 8). In addition, the staged embolization of a large, high-flow AVM can decrease the incidence of potentially catastrophic hemorrhage caused by rapidly changing hemodynamics (e.g., normal perfusion pressure breakthrough). Finally, preoperative embolization of feeding vessel and nidal aneurysms can eliminate those angioarchitectural risk factors for perioperative hemorrhage. Proximal feeding artery aneurysms are at risk for rupture after AVM resection because elimination of the AV shunt causes a sudden increase in arterial pressure. These proximal aneurysms may be impossible to access through the craniotomy for the AVM resection, and preoperative embolization eliminates the need for a second craniotomy (86,87). There is a general consensus that many superficial Spetzler-Martin grade I and II AVMs can be surgically resected with minimal morbidity and mortality without preoperative embolization. In these cases, the additional risks of embolization may not be justified. There are, however, exceptions such as a grade I or II AVM with a deep feeder that is difficult to access surgically (88,89). Also, some experts advocate embolizing an intranidal aneurysm in a SpetzlerMartin grade I or II AVM presenting with acute hemorrhage to stabilize the nidus until surgery (89). Embolization is used frequently for Spetzler-Martin grade III AVMs, particularly for those in central and eloquent locations and those with deep feeders. Preoperative embolization (often staged) is commonly employed for those Spetzler-Martin grades IV and V AVMs considered for resection. Preradiosurgical Embolization

There is a wide variability at different institutions in the use of combined embolization and radiosurgery. Stereotactic radiosurgery primarily is employed at some centers for small brain AVMs that have a high surgical risk because they are deep-seated or are located in the eloquent cortex. Other centers

290

Weigele et al.

frequently use embolization to render large brain AVMs amenable to stereotactic radiosurgery. In this setting, embolization is used to reduce the size of the AVM to increase the probability of a radiosurgical cure (90). The rate of cure after stereotactic radiosurgery significantly decreases as the volume of AVM being treated increases (91–93). Radiosurgical cure is more likely after embolization has reduced the residual AVM volume to less than 10 cc (90,94). The goal of embolization is to shrink the nidus into a smaller single target; however, this is not always possible. Alternatively, two or more discrete areas of residual nidus can be treated with a volume-staged approach (95). Embolization is also performed prior to radiosurgery to occlude nidal aneurysms that represent a risk for hemorrhage until the radiosurgically induced obliteration occurs and also to occlude high-flow fistulas that may be refractory to radiosurgery (90). Repeat embolization or surgery also can be used to treat residual AVM persisting after radiosurgery (96,97). Curative Embolization

There is currently a limited role for curative embolization of brain AVMs (Fig. 16). Although embolization can successfully obliterate some small AVMs that have limited feeders, it rarely cures large, complex AVMs. Most of the AVMs that have a relatively high probability of cure with embolization are amenable to complete surgical removal with negligible morbidity and mortality. Therefore, justifying the risks of an attempted curative embolization is often questionable (73,98). Small deep central AVMs with limited feeders

are exceptions where curative embolization can play an important role (99). Palliative Embolization

Palliative embolization does not appear to improve on conservative medical management of most patients with incurable AVMs and may even worsen the subsequent clinical course (100). There are, however, appropriate goal-directed roles for palliative embolization in select circumstances. Palliative embolization can alleviate symptoms due to vascular steal and mechanical compression and obliterate specific aneurysms responsible for repeated hemorrhages. Embolization of meningeal supply can improve intractable headaches (101–103).

Embolization Tools and Technique Microcatheters and Guidewires

The flow-directed microcatheters currently used for embolization with liquid agents are designed for safe and reliable navigation into the very distal aspects of the intracranial circulation. They have several segments of progressive softness. The proximal segments are relatively stiff and thick-walled to transmit longitudinal motion and torque efficiently. The transitional middle segments have thinner walls and progressively increase in flexibility but remain ‘‘pushable.’’ The distal segments are small (1.3- to 1.8-Fr outer diameter), thin walled, and extremely soft and supple. They provide no intrinsic transmission of longitudinal

Figure 16 Curative AVM embolization. (A) Lateral angiogram shows occipital AVM nidus (arrow). (B) Superselective angiogram (arrow, microcatheter tip). (C, D) AP and lateral postembolization angiogram shows complete obliteration of the nidus. Abbreviation: AVM, arteriovenous malformation.

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

force. The catheter tips are slightly bulbous so blood flow will pull them forward. The microcatheters have hydrophilic surface coatings to decrease thrombogenicity, facilitate movement through small tortuous vessels, and prevent adhesion of embolic agents. Guidewires designed for use in the cerebral arteries (0.008–0.014 inch) have very flexible distal segments and soft, shapeable platinum tips. They also are covered with a hydrophilic coating to reduce friction between the catheter and guidewire. They remain ‘‘torquable’’ even after they have gone through several curves. Only the smallest guidewires, such as the 0.008-inch Mirage (EV3, Plymouth, Minnesota, U.S.), will pass through the flow-directed microcatheters commonly used with liquid embolic agents (104). Vessel Selection

Current techniques for brain AVM embolization require advancing a suitable microcatheter into the very distal aspect of an arterial feeder supplying the nidus. A guide catheter (e.g., Envoy, Cordis Endovascular, Miami Lakes, Florida, U.S.) first is placed in the distal cervical aspect of the appropriate internal carotid or vertebral artery. A 6-Fr guide catheter is preferred for easier contrast injections while the microcatheter is inserted; however, a 5-Fr guide catheter may be safer in a small vertebral artery. A rotating hemostatic valve is used for coaxial placement of a microcatheter and to continuously flush the guide catheter with heparinized saline. Intravenous heparin is administered on an individual basis to prevent thromboemboli if there are small feeders or there is slow flow. A microcatheter (1.5- or 1.8-Fr Spinnaker Elite, Boston Scientific Corporation, Natick, Massachusetts, U.S.; Marathon or Ultraflow, EV3) with a small steamshaped distal curve (e.g., 1-mm radius distal ‘‘J’’ shape) is navigated through the cerebral arteries under continuous subtracted fluoroscopic (road-map) imaging. There are two primary techniques for intracranial navigation: flow directed and guidewire assisted. Flow-directed navigation uses arterial blood flow to drag the very flexible distal catheter segment and slightly bulbous catheter tip forward. The tip will preferentially tend to advance into the vessel with the highest flow, which is usually the desired feeder to the AVM. Directional control also is facilitated by gentle injections of contrast (puffing) to redirect the curved tip into the desired branch. For guidewire-assisted navigation, a 0.008-inch Mirage (EV3) guidewire is advanced into the distal segment of the microcatheter to augment its ‘‘pushability’’ and to change shape of the catheter tip. Advancing and withdrawing the guidewire in the distal segment also changes its elasticity often prompting the catheter tip to spring forward. When necessary, the Mirage guidewire can be extended beyond the microcatheter tip to navigate difficult anatomy; however, this must be done with caution to avoid arterial perforation or dissection. Blood pressure augmentation with neosynephrine or vasodilatation with papaverine or hypercapnia also can be used to facilitate distal catheter advancements (104–106).

291

Provocative Testing (Superselective Wada Test)

Approximately 10% of brain AVM embolizations cause a permanent neurological deficit (107). Many of these deficits are caused by embolization of branches arising from an AVM feeder that supply normal brain parenchyma (108). These branches may not be seen during superselective angiography prior to embolization because of the high blood flow (sump effect) into the nidus. They can be occluded during subsequent embolization because of changing hemodynamic conditions, resulting in an infarct. Provocative testing (the superselective Wada test) is intended to prevent this complication by identifying any angiographically occult blood supply to eloquent brain parenchyma from the feeder proposed for embolization. Although some experts are strong proponents for the use of this provocative testing, others argue it is not necessary (109). A short-acting barbiturate (amobarbital) is injected intra-arterially with the microcatheter positioned at the site of intended embolization and appropriate neurological testing is carried out. Provocative testing usually is performed with the patient awake to facilitate neurological examinations after the amobarbital is injected. If a transient neurological deficit occurs, embolization is contraindicated from that catheter position. The addition of electroencephalographic monitoring to the clinical exam has been reported to increase the test’s sensitivity (110). Amobarbital principally affects the gray matter through the gamma-amino butryic acid A (GABAA) receptor. The white matter is not affected. A recent report advocates additional provocative testing with lidocaine. This inhibits both gray and white matter by blocking the voltage-dependent sodium channel and has been suggested to be able to detect deficits that might not be detected by amobarbital alone (111). Several case series have reported provocative testing that can identify situations where embolization will cause a neurological deficit; however, they are all relatively small and uncontrolled (110–113). The value of provocative testing is vigorously debated. Those who opposed to provocative testing argue that if contrast does not flow into normal branches during superselective angiography because of the sump effect, then even the provocative agent will not flow, thus yielding a false negative test result. Also, the changing hemodynamics that occur during embolization are not simulated during the provocative test. In addition, the potential effect of proximal reflux during embolization is not evaluated. Finally, the use of general anesthesia is precluded by the need for neurological testing. Some believe general anesthesia significantly enhances the safety of embolization by preventing dangerous patient motion during critical moments of the procedure. Opponents of provocative testing conclude that careful evaluation of the AVM anatomy (location with respect to eloquent brain and the angioarchitecture), as well as the use of an intranidal microcatheter position for embolization may be more reliable than provocative testing to avoid ischemic complications (109,114). Proponents

292

Weigele et al.

for provocative testing argue that superselective amobarbital testing compliments the analysis of the vascular anatomy to minimize ischemic complications and do not believe general anesthesia is essential to perform safe embolizations (110). Provocative testing also may be important from a medicolegal perspective if there is a serious complication (109). Embolic Agents

There are three general categories of currently available embolic agents: solid occlusive devices (coils, silk threads, balloons), particulates [polyvinyl alcohol (PVA) particles], and liquids (cyanoacrylates, Onyx, ethanol) (115). Solid occlusive devices primarily are used to occlude large direct AV fistulas. Particulate embolization using PVA particles has been replaced by liquid embolization with N-butyl cyanoacrylate (NBCA) at most centers (109). Onyx is a promising liquid embolic agent recently approved by the Food and Drug Administration (FDA) (116). Although absolute ethanol and silk threads have been used to embolize brain AVMs, limited results have been published (117,118). Polyvinyl alcohol particles. PVA particles were commonly used for brain AVM embolization before liquid agents such as NBCA and Onyx became more widely used. PVA particles are supplied in various size ranges from 50 to 1000 mm. They are nonradiopaque and are mixed with iodinated contrast for delivery. PVA particles are often used in combination with coils or silk threads to facilitate their retention, especially in larger AV shunts (119). Typically, larger (e.g., 3-Fr) over-the-wire microcatheters have been required for larger PVA particles (>500 mm) resulting in more proximal embolizations of the arterial feeders rather than the nidus, although recently more distal PVA embolizations have been made possible with hybrid flow-guided or over-the-wire catheters and liquid coils (89). PVA particles have several disadvantages as compared to liquid embolic agents. They can occlude the small, flow-directed microcatheters that can be most reliably advanced into the distal feeder. Also, since the particles are radiolucent, it is not possible to identify where they deposit. There is evidence they often aggregate and frequently occlude the arterial feeder rather than the nidus. The nidus can then recruit collateral blood supply and regrow. This may explain why Sorimachi et al. found 43% of brain AVMs partially embolized with PVA particles demonstrated an increase in the size of the nidus on follow-up angiograms (119). In addition, a histopathological analysis revealed PVA-embolized vessel lumens contained clumps of particles intermixed with thrombus rather than solid luminal packing with PVA. Eighteen percent of the embolized vessels were partially recanalized (120). This may explain why AVMs that appear completely obliterated on angiography after PVA embolization can reappear on follow-up exams (119). For these reasons, a permanent occlusion of some or all of an AVM nidus seems less likely with

PVA than with liquid embolic agents. This lack of permanency is undesirable for an embolization performed as an adjunct for radiosurgery where the embolized portion of the AVM is excluded from the radiation field or for an embolization performed for cure. This conclusion is supported by the observation that 15% to 20% of AVM patients undergoing PVA embolization prior to radiosurgery had recanalization two to three years later on follow-up angiography (92,94). This lack of a durable occlusion, however, may not be a significant disadvantage for presurgical embolizations with PVA particles. A prospective, randomized, multicenter trial concluded PVA and NBCA were similar in safety and effectiveness for preoperative brain AVM embolization (89). N-Butyl cyanoacrylate. Cyanoacrylates have been used for brain AVM embolization for more than 20 years. Early problems with cyanoacrylate embolization (see historical background) that prevented widespread use have been solved with the replacement of previous formulations with NBCA and with advances in microcatheter and guidewire technology (109). The FDA approved NBCA (Trufill, Cordis Endovascular) for brain AVM embolization in 2000. NBCA has become the most commonly used embolic agent for this purpose in the United States. The Cordis NBCA kit contains NBCA, ethiodol, and tantalum powder. Ethiodol is mixed with NBCA to prolong the polymerization time and also to make the solution radiopaque. Tantalum power can be added to the NBCA/ethiodol mixture to further increase its radiopacity. The liquid NBCA monomer undergoes a rapid exothermic polymerization catalyzed by nucleophiles found in blood and on the vascular endothelium to form an adhesive, nonbiodegradable solid. The vessel is permanently occluded when the polymer completely fills the lumen. NBCA provokes an inflammatory response in the wall of the vessel and surrounding tissue leading to vessel necrosis and fibrous ingrowth. These histological responses also may contribute to the permanency of NBCA occlusions (121,122). NBCA has a number of useful properties for brain AVM embolization. The liquid monomer can be injected through small (1.5- and 1.8-Fr) flowdirected microcatheters such as the Spinnaker Elite (Boston Scientific) and the Ultraflow (EV3) that can be reliably and safely positioned in the distal arterial feeder or within the nidus. This distal catheter positioning maximizes the likelihood of adequate nidal penetration to achieve a permanent occlusion and minimizes the risk of inadvertent embolization of normal branches (109). The NBCA polymerization rate can be adjusted to satisfy specific requirements. The goal of the embolization is to form a solid intranidal NBCA cast, avoiding early polymerization in the arterial feeder or late polymerization in the venous outflow. Pure NBCA polymerizes almost instantaneously at the catheter tip. Although this may be necessary to occlude a direct high-flow fistula, immediate polymerization will not allow the NBCA to penetrate optimally into a plexiform nidus (109). The addition

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

of ethiodol slows the polymerization rate, allowing better nidal penetration. The polymerization rate progressively decreases as more ethiodol is added. The objective is to use an ethiodol/NBCA mixture with a polymerization time optimally matched to the individual AVM’s angioarchitecture and hemodynamics. The AV transit time on superselective angiography is subjectively evaluated as a guide to formulating the mixture. This is far from an exact science and is highly dependent on experience. The concept of a ‘‘wedged’’ catheter position, where forward flow is controlled by the rate of injection, theoretically allows slower, more controlled injections of a more dilute NBCA/ethiodol mixture with a longer polymerization time (109). The use of ethiodol has limitations. A high ethiodol concentration also increases the viscosity of the mixture, which can conversely decrease the nidal penetration. Dilute NBCA mixtures also tend to disperse in small droplets that can incompletely cast the vessels, possibly allowing recanalization (122). Glacial acetic acid can be added as an alternative method to slow the rate of polymerization, without causing the increased viscosity of higher ethiodol/NBCA concentrations. This may result in better nidal penetration and more solid casting (123).

293

Many experts believe portions of the AVM nidus that are well cast with NBCA can be considered permanently obliterated (109). Wikholm followed 12 brain AVMs totally occluded with NBCA for 4 to 78 months and found no angiographic evidence of recanalization (124). In a recent study, six patients with complete obliteration of the AVM nidus had no angiographic evidence of recurrence at 17 to 32 months (125). NBCA embolization therefore has the potential to transform inoperable AVMs into surgically resectable lesions and to reduce the size of an AVM nidus sufficiently to make radiosurgery possible. Some small AVMs can be cured by embolization alone. A solid NBCA cast in the nidus is essential to assure permanent obliteration of the AVM (109). Debrun described the single-column flow-controlled technique for optimal nidal filling using a microcatheter that is wedged intranidally and the use of a relatively dilute NBCA mixture (Fig. 17) (109). Solid casting is important since brain AVMs that were incompletely embolized with NBCA demonstrated histological evidence of capillary regrowth in the lumen of embolized vessels after three months (121). Brain AVMs embolized with NBCA have favorable properties for surgical resection. The vessels are

Figure 17 Wedged catheter embolization of nidal pseudoaneurysm presenting with hemorrhage. (A, B) AP and lateral angiograms show plexiform cerebellar AVM. (C) Superselective superior cerebellar angiogram shows large nidal pseudoaneurysm (arrow ) and small nidal aneurysm (arrowhead ). (D) Wedged catheter injection visualizing pseudoaneurysm and superior nidus (arrow, microcatheter tip). (E) NBCA cast in pseudoaneurysm (arrow) and superior nidus (arrowheads). (F) Lateral postembolization angiogram. Abbreviations: AVM, arteriovenous malformation; NBCA, N-butyl cyanoacrylate.

294

Weigele et al.

easily compressible and transected. The embolized feeders can be readily identified and differentiated from nonembolized en passage branches to normal brain (Fig. 8). In addition, the embolization of the AVM nidus provides a distinct boundary zone between the AVM and normal brain (126). NBCA Technique The following is a general description of the brain AVM embolization protocol using NBCA at our institution. Procedures are usually performed under conscious sedation rather than general anesthesia to allow provocative testing. An appropriate microcatheter is negotiated through the cerebral vasculature into the desired AVM feeder using a combination of flow and guidewire guidance (see vessel selection). The microcatheter tip is advanced into the distal aspect of the desired feeder. Operator preference and the nidal anatomy determine whether a free or wedged catheter tip position is used. Excess slack (redundant loops) in the microcatheter is removed to facilitate its removal after the embolization. A superselective angiogram is performed with a 1-cc syringe and a gentle hand injection. The angioarchitecture is analyzed. If no normal branches are visible, provocative testing is performed with amobarbital. The NBCA/ethiodol/tantalum mixture (Trufill, Cordis Endovascular) is prepared using clean gloves on a separate sterile table to prevent contamination with ionic catalysts. For a wedged injection (109), a relatively dilute concentration of NBCA (25–33%) is made by mixing 1 cc of NBCA with 2 or 3 cc of ethiodol in a shot glass. The vial of tantalum powder included in the Trufill kit is added to increase the radiopacity of the mixture. Relative hypotension is induced (20–30% decrease in mean arterial pressure). Test injections are made with a subtracted fluoroscopic image to confirm the catheter position and to gauge the optimal injection rate. The microcatheter is irrigated with 5% dextrose to flush all of the ionic catalysts from the lumen. The dilute NBCA solution is then injected slowly into the nidus over 15 to 60 seconds during continuous subtracted fluoroscopic observation. The injection rate is modified to obtain a solid nidal cast without causing proximal reflux. If a drop of NBCA enters a draining vein, the injection is paused several seconds. The injection is then restarted and continued if additional nidal filling is observed. If another drop enters a vein, the injection is terminated. The injection is also terminated if proximal reflux occurs. The microcatheter is aspirated and briskly removed. The guide catheter is aspirated and its tip is examined fluoroscopically. A postembolization angiogram is then obtained. A nonwedged injection is performed in a similar fashion; however, a more concentrated NBCA solution is used because of the more rapid flow and the shorter arterial-venous transit time through the nidus. The injection rate is faster and the injection time is much shorter (one to three seconds). If a large direct fistula is present, maximal induced hypotension and a very high NBCA concentration are used. In this setting, coils (Liquid Coils, Target/Boston Scientific, Natick, Massachusetts, U.S.) can be injected first into

the fistula to slow the rate of flow and to form a framework for the NBCA to adhere to. At our institution, NBCA is most commonly used for preoperative embolization of large, complex AVMs. We typically occlude a maximum of 33% of the nidus during one session to minimize the risk of normal perfusion pressure breakthrough–induced hemorrhage (127). Embolization is also terminated if venous stagnation occurs to minimize the risk of postprocedural hemorrhage caused by venous outflow compromise (128). Onyx. Onyx (EV3) is a premixed, liquid embolic agent consisting of ethylene-vinyl alcohol copolymer (EVOH) and tantalum powder (for radiopacity) dissolved in dimethyl sulfoxide (DMSO). EVOH contains 48-mol/L ethylene and 52-mol/L vinyl alcohol (129). Taki et al. were the first to describe the use of EVOH, mixed with metrizimide powder (for radiopacity), dissolved in DMSO to embolize brain AVMs in 1990 (130). Subsequent studies led to a multicenter randomized trial that demonstrated noninferiority of Onyx compared to NBCA in achieving greater than or equal to 50% volume reduction for presurgical brain AVM embolization, resulting in FDA approval of Onyx for presurgical brain AVM embolization in 2005 (116,131,132). Onyx is a cohesive, nonadhesive liquid embolic agent. The copolymer holds together as it is injected, but it does not adhere to the endothelium or to the microcatheter tip. When the mixture contacts an aqueous solution such as saline or blood, the DMSO diffuses away rapidly, causing the copolymer to precipitate into a soft, spongy solid. The precipitation progresses from the outer surface inward, forming a skin with a liquid center that continues to flow as the solidification continues. During the injection, the column of Onyx advances into the path of least resistance. The rate of precipitation of the copolymer is proportional to the concentration of EVOH in the solution. There are currently two commercially available concentrations of EVOH for brain AVM embolization: Onyx 18 (6% EVOH) and Onyx 34 (8% EVOH). Onyx 18 travels farther distally and penetrates more deeply into the nidus because of its lower viscosity and slower precipitation rate. Onyx 18 is used for distal feeding pedicle injections into a plexiform nidus, whereas Onyx 34 is recommended for embolizing high-flow fistulas. Complete solidification of both formulations occurs within five minutes. DMSO was chosen as the solvent because it rapidly diffuses in aqueous solution and its physiological properties in humans are well known (130). DMSO is angiotoxic, however, with adverse effects that include vasospasm, angionecrosis, arterial thrombosis, and vascular rupture (133). These undesirable consequences are related to the volume of DMSO infused and the endothelial contact time (131). Severe angiotoxic effects do not occur when the DMSO infusion rate does not exceed 0.25 mL/90 sec (116,131). Only specifically approved microcatheters (Ultraflow, Marathon, Echelon; EV3) can be used with Onyx because the DMSO will dissolve incompatible

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

catheters. Patients may notice a garlic-like taste for several hours, and their skin and breath may have a characteristic odor due to the DMSO for one to two days after an embolization with Onyx. The cohesive and nonadhesive properties of Onyx provide several advantages compared to NBCA. Because Onyx is nonadhesive and it solidifies more slowly than NBCA, typical injections are performed over much longer time intervals (minutes) and are easier to control. This procedure results in a much more leisurely embolization, providing time to analyze progress with interval angiography, if desired, and involving less risk of refluxing the embolic agent too proximally in the arterial feeders or extending too distally into the venous outflow (Fig. 18). It is also possible that a more complete and solid casting of the nidus may be obtained with Onyx compared to NBCA. This may result in an increased rate of cure, but this remains to be proven. In addition, the catheter also can be repositioned into a second pedicle and another embolization can be performed, a maneuver that is not possible with NBCA. Finally, Onyx does not cause inadvertent gluing of the catheter tip to the vessel (116). Although Onyx is a very promising embolic agent, only limited data on its use for brain AVM have been published. Jahan et al. embolized 23 brain AVMs achieving an average 63% reduction in AVM volume with 4% permanent morbidity and no mortality. Histopathological examination of the resected specimens showed mild inflammation one day after embolization and chronic inflammatory changes after four days. Angionecrosis was seen in two patients, but the vessel wall integrity was maintained in all specimens (116). In another study, the surgical handling characteristics of Onyx were compared with NBCA in embolized swine rete mirabile (132). Onyx was softer and handled better than NBCA during surgical resection. The permanency of Onyx embolization is as yet unknown. Short-term angiographic follow-up (1–100 days) did not reveal recanalization (116);

Figure 18 Angiogram obtained during embolization with Onyx. (A) Onyx cast (arrowheads) and microcatheter tip (arrow) on lateral radiograph. (B) Lateral angiogram obtained through guide catheter during embolization.

295

however, no long follow-up studies have been published. Therefore, the role for Onyx as an adjunct to stereotactic radiosurgery or for curative embolization has not been definitively established. Onyx Technique Patients can experience pain during embolization with Onyx; therefore, general anesthesia is used more frequently than with NBCA. Some perform provocative testing at the planned embolization sites prior to inducing general anesthesia. An Onyx-compatible microcatheter is positioned in the desired location using flow-directed and guidewire-assisted navigation as described above. Better nidal penetration is usually obtained in a larger feeder. A wedged intra-nidal position is optimal. After positioning of the microcatheter, a superselective angiogram is obtained. The Onyx solution must be vigorously shaken for 20 minutes to fully suspend the micronized tantalum powder. Mixing is continued until just before the embolization. Failure to do this may result in inadequate radiopacity (129). The manufacturer provides an adapted Vortex-Genie (Scientific Industries, Inc., Bohemia, New York, U.S.) to mix the Onyx. The catheter is flushed with normal saline and the dead space is loaded with pure DMSO solvent. The Onyx mixture is drawn into a DMSO-compatible 1-cc syringe, the syringe is connected to the microcatheter and a slow, steady injection is begun at a rate of 0.25 mL/90 sec to displace the DMSO in the dead space with Onyx. Subtracted fluoroscopy is begun just before the dead space volume has been replaced by the injection. The injection is continued at 0.1 mL/min as the Onyx begins to deliver out of the microcatheter. A slow, steady injection usually results in optimal nidal penetration. Changes in the injection rate tend to cause proximal reflux. The injection rate is not allowed to exceed 0.25 mL/90 sec to prevent angiotoxicity. If proximal reflux occurs, the injection is paused 30 seconds and restarted. This allows a plug to solidify around the catheter tip that prevents further reflux and promotes forward flow. This ‘‘plug and push’’ technique can be repeated multiple times as required. Proximal reflux around the catheter tip should be limited to 1 cm to avoid causing difficulty with the catheter retrieval. As a circumferential plug forms around the catheter tip, the injection can be paused for as long as two minutes to allow the plug to solidify. This will establish proximal flow arrest when the injection is restarted so that the subsequent flow will move distally. Similarly, if Onyx begins to fill a draining vein, a pause in the injection will allow that material to solidify, and when the injection is restarted the additional Onyx usually fills new areas of the nidus. The injection should never be paused for more than two minutes to prevent Onyx from precipitating in the catheter lumen. If there is resistance, the injection should be discontinued to avoid rupturing the catheter. There are two catheter retrieval techniques. The slow ‘‘traction’’ method uses incremental catheter withdrawal (cm by cm) with sustained moderate tension on the catheter. The quick ‘‘wrist-snap’’

296

Weigele et al.

technique is to withdraw the catheter enough (3–5 cm) to create slight tension and then quickly snapping the wrist (not the entire arm) 10 to 20 cm left to right. Pulling too far or hard runs the risk of causing a catheter separation.

Postprocedural Care Patients are observed in the neurointensive care unit for 24 hours and usually discharged to home on the second postembolization day. Mild hypotension (mean arterial pressure ~90% of normal) may be induced for 24 hours if a large, high-flow AVM has been embolized. Additional embolization sessions for large, high-flow AVMs are staged every three to four weeks.

Results The literature on outcomes for brain AVM treatments primarily consists of uncontrolled, single institution case series. Many of these have demonstrated an important role for brain AVM embolization in selected patient populations. Nonetheless, they have relatively small sample sizes, and tremendous variability in selection criteria, techniques, patient evaluation, and follow-up. Multicenter, randomized, controlled outcome trials are needed to form a scientific basis for the selection of optimal therapeutic plans. Since brain AVM embolization is used mostly as an adjunct to surgery or radiosurgery, these trials will need to compare the natural history with the overall results of individual and combined treatment strategies. Presurgical Embolization

A number of case series comparing groups undergoing surgical resection of brain AVMs with and without preoperative embolization have demonstrated that selective preoperative embolization improves overall outcomes (126,134,135). Preoperative embolization and surgery is also cost-effective compared to surgery without embolization, with cost per qualityadjusted life-year savings as high as 34% (136). Pasqualin et al. demonstrated that preoperative embolization of large, high-flow AVMs was associated with less intraoperative bleeding, and there were fewer postoperative neurological deficits, seizures, and deaths in the group that underwent preoperative embolization. The frequency of major new deficits was 31% in the surgery only group versus 5% in the combined embolization and surgery group. However, the incidence of postoperative hyperemic complications did not differ between the embolized and nonembolized groups (134). Demeritt et al. compared 30 patients who underwent preoperative AVM embolization with NBCA followed by surgery with 41 patients who had surgery without preoperative embolization. The combined embolization and surgery group had a higher average Spetzler-Martin score compared to the surgery only group (89% vs. 68% in grade III and IV, respectively) and a larger average nidal maximal diameter (4.2


1.5 cm vs. 3.4
1.8 cm); however, two-week and longterm Glasgow Outcome Scale scores were better in the combined embolization and surgery group than the surgery only group (70% vs. 41%, respectively, had a two-week Glasgow Outcome Scale score of 5; 86% vs. 66%, respectively, had a long-term Glasgow Outcome Scale score of 5) (135). Similarly, Jafar et al. compared 20 patients who underwent preoperative AVM embolization with NBCA followed by surgery with 13 patients who had surgery alone. The combined group had a larger average AVM diameter (3.9 cm vs. 2.3 cm) and a higher average Spetzler-Martin grade (3.2 vs. 2.5) compared to the nonembolized group. Embolization complications included immediate and delayed hemorrhage (15%) and transient ischemia (5%). There was no embolization-related death. No difference in surgical complications was found between the embolized and nonembolized groups. The large majority of patients (91%) in both groups had good to excellent long-term neurological outcome. The authors concluded that ‘‘preoperative NBCA embolization of AVMs makes lesions of larger size and higher grade the surgical equivalent of lesions of smaller size and lower grade’’ (126). Martin found that embolization was only effective to decrease blood loss and shorten operative time when the nidal size was decreased more than 66%. Less reduction in the size of the nidus and a reduction in the rate of AV shunting were not effective (86). Preradiosurgical Embolization

Gobin et al. reported the results in treatment of 125 patients with embolization and radiosurgery. Approximately half of the AVMs had diameters greater than 4 cm and most were Spetzler-Martin grade III or greater. Embolization cured 11% and made 77% suitable for radiosurgery. Greater than 50% of the AVMs with diameters greater than 6 cm and more than 10% with diameters between 4 and 6 cm did not have sufficient nidal size reduction for subsequent radiosurgery. Overall cure rates were 76% to 78% for AVMs less than 4 cm in diameter, 59% for those 4 to 6 cm, and 7% for those over 6 cm. Therefore, adjunctive embolization was most effective for AVMs of 4 to 6 cm in diameter. There was no convincing advantage for combined embolization and radiosurgery compared to radiosurgery alone for AVMs smaller than 4 cm. Embolization and radiosurgery did not result in a significant cure rate for lesions larger than 6 cm. Preradiosurgical embolization did not provide protection from hemorrhage during the latent period until radiosurgical obliteration. There was approximately a 3% annual rate of hemorrhage during the one- to three-year follow-up period, similar to the natural history brain AVM hemorrhage rate. The absence of residual AVM nidus or AV shunting after radiosurgery does not equate with definitive evidence of permanent obliteration of the AVM. Although a negative angiogram had been considered the practical endpoint defining successful treatment, a recent retrospective review of

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

236 fradiosurgery-treated AVMs followed for a median of 6.4 years after angiographic evidence of obliteration found four cases of subsequent hemorrhage in the previous AVM site. The two cases that were resected had small regions containing tiny patent AVM vessels. In each case, there was enhancement in the treated site on postgadolinium MRI scans despite normal posttreatment angiograms. The annual risk of hemorrhage was 0.3% (137). Embolization can also be used to treat an AVM persisting after radiosurgery. Marks et al. reported six patients with brain AVMs remaining 24 to 55 months (mean 34 months) after radiosurgery. Embolization resulted in one cure, facilitated surgical resection in three, and caused sufficient volume reduction in two patients in whom repeat radiosurgery could be performed. There were no complications (96). Curative Embolization

Published embolization cure rates vary considerably because of selection bias, differing therapeutic goals and techniques. Small AVMs with few feeders have the highest probability of endovascular cure. Case series performed without specific selection of those AVMs that are most likely to be cured by embolization alone, have reported an overall durable embolization cure in 5% to 40% of patients (90,138–140). Valavanis and Yasargil had a 74% rate of curative embolization in a subgroup of patients with favorable angiographic features such as one or few dominant feeders, no perinidal angiogenesis, and a fistulous nidus versus a 40% rate of curative embolization for their entire series of 387 patients (139). Palliative Embolization

In general, palliative embolization does not appear to improve on conservative medical management and may even worsen the subsequent clinical course. Kwon et al. obtained long-term follow-up in a group of 27 patients with inoperable brain AVMs. Out of these patients, 16 were treated medically and 11 were partially embolized. There was no significant difference between the two groups with respect to clinical improvement, lack of change, and deterioration. In addition, 46% of the partially embolized group experienced hemorrhage in the follow-up period versus 25% in the nonembolized group (p ¼ 0.27) (100). Miyamoto et al. obtained 49-month (mean) follow-up of 46 patients with unresectable AVMs treated with various palliative techniques (partial embolization, radiosurgery, subtotal resection, and feeder ligation). There was a 14.6% annual rate of hemorrhage, 23% incidence of new major neurological deficits, and a 9% mortality rate (141). Nonetheless, palliative embolization can be beneficial in selected circumstances. Ischemic neurological deficits caused by vascular steal and venous hypertension in a high-flow, inoperable AVM were improved following partial embolization (102). In another patient, hemifacial spasm caused by a dilated lateral mesencephalic vein draining an inoperable temporo-occipital AVM was cured by selective trans-

297

Table 7 AVM Embolization Morbidity and Permanent Mortality Reference Year 140 109 139 144 153 89 145 143

1995 1997 1998 2002 2002 2002 2004 2006

Number of patients

Mortality rate (%)

Permanent morbidity rate (%)

1246 54 387 233 450 54 201 306

1.0 3.7 1.3 1.0 1.1 1.9 2.0 2.6

8.0 5.6 5.1 14.0 3.8 13.0 9.0 4.9

Abbreviation: AVM, arteriovenous malformation. Source: From Ref. 143.

venous embolization (101). Embolization of dural supply can alleviate intractable headaches. In patients with repeated hemorrhages, targeted embolization of angioarchitectural risk factors such as proximal and nidal aneurysms can limit additional bleeds (103). Complications

Incidence The reported incidence of overall complications from brain AVM embolization varies from 3% to 25% (88,108,134,138,142). The rates of permanent morbidity and mortality in large series range from 3.8% to 14% and 1.0% to 3.7%, respectively (Table 7) (143). Most are caused by hemorrhagic and ischemic events (103). Since complications are related to a number of technical and hemodynamic factors, this wide range in reported rates probably reflects at least, in part, differences in case selection, embolization techniques, and management strategies. In 1995, Frizel and Fisher reported a review of 32 case series of brain AVM embolizations in a total of 1246 patients over 35 years from 1969 to 1993. Overall temporary and permanent morbidity were 10% and 8%, respectively. Overall mortality was 1%. There was no significant difference in permanent morbidity and mortality in the patients treated before and after 1990 (140). In 2002, Hartmann et al. prospectively evaluated 233 patients undergoing 545 embolizations. Thirtythree patients (14%) had treatment-related neurological deficits; however, they were persistent and disabling in only five patients (2%). There were two deaths (1%). Factors statistically associated with new deficits were increasing patient age, absence of a pretreatment deficit, and the number of embolization sessions (144). In 2004, Taylor et al. reviewed 339 AVM embolizations performed in 201 patients over an 11-year period. There was a 7.7% rate of major complications per procedure. Nine percent of the patients had a permanent neurological deficit and 2% died from the embolizations (145). In 2006, Haw et al. reported the results of 513 attempted embolizations in 306 patients performed between 1984 and 2002. There were eight (2.6%) deaths, six caused by hemorrhage and two caused by ischemic strokes. The rate of death and permanent disabling morbidity was 3.9%. Three factors were

298

Weigele et al.

statistically associated with complications: an eloquent AVM location, a AV fistula, and venous deposition of the embolic agent (cyanoacrylates). Passage of the embolic agent into the draining veins caused 8 of the 12 (67%) deaths or disabling deficits. There was a reduction in complications producing death or permanent disability in the second half of the study. The authors suggest this was due to advances in equipment and techniques, as well as greater expertise and clinical judgment gained through experience (143). Types of complications Periprocedural hemorrhage There are a number of causes of periprocedural hemorrhage from brain AVM embolizations. Technical factors include catheter or guidewire-induced arterial perforations, dissections, rupture of aneurysms, vascular injuries during catheter retrieval, and accidental embolizations of venous outflow (109,143). Physiological factors include venous outflow thrombosis, hemodynamic changes in the setting of impaired cerebrovascular reactivity, and hemodynamic stresses on angioarchitectural weak sites such as feeder, nidal, and venous aneurysms. Embolization can markedly reduce flow through a fistulous nidus causing stagnation in the draining veins (Fig. 19). This can result in venous outflow thrombosis, nidal congestion, and a delayed hemorrhage or a venous ischemic infarct (128,143,146). Normal perfusion pressure breakthrough is another important physiological cause of hemorrhage following AVM treatment. The ‘‘sump effect’’ of a large shunt causes low pressure in the arterial feeders and nearby parenchymal branches. The high flow through the nidus elevates venous pressures. The result is a chronically low cerebral perfusion pressure that can impair cerebrovascular autoregulation. If the shunt is suddenly therapeutically disrupted, there is an immediate increase in arterial pressure and a decrease in venous pressure, with a resulting dramatic increase in the cerebral perfusion pressure. If cerebrovascular autoregulation is impaired, resulting parenchymal hyperperfusion can cause cerebral edema or hemorrhage. Spetzler et al. called this normal perfusion pressure breakthrough and found it was associated with large, high-flow AVMs, poor angiographic filling of normal cerebral arteries, extensive collateral flow (steal) (Fig. 1), external carotid supply, and progressive or

fluctuating neurological deficits (103,127). The risk of normal perfusion pressure breakthrough–induced hemorrhage can be minimized by stepwise reduction in the degree of shunting in large, high-flow AVMs through multiple embolizations staged every three to four weeks, facilitating the gradual recovery of normal vascular reactivity (103). Prompt surgical evacuation of an embolizationinduced cerebral hematoma results in a good outcome in most cases. Jafar and Rezai reported the emergent surgical evacuation of acute intracerebral hematomas from brain AVMs in 10 patients experiencing acute neurological deterioration, including eight cases occurring after embolization. They employed immediate intubation, hyperventilation, osmotic diuresis, barbiturate anesthesia, and surgery. The hematoma was evacuated and the AVM was totally resected if possible (8 out of 10). Postoperative cerebral perfusion pressure was maintained above 55-mm Hg with mannitol and barbiturates. Nine of the 10 patients had good to excellent outcomes (147). Ischemic stroke Technical causes of acute stroke during embolization include the showering of NBCA droplets from the catheter tip as it is removed, catheter- or guidewire-related arterial dissections and thromboemboli, the embolization of en passage or pseudoterminal supply to normal brain parenchyma distal to the nidal supply, and the inadvertent reflux of embolic material into normal branches proximal to the catheter tip. Ischemic stroke can result from retrograde thrombosis in stagnant feeding arteries propagating into branches to normal brain (Fig. 20) (148). Delayed venous thrombosis can cause a venous infarct (146). Careful attention to the angioarchitecture on superselective angiography and to an optimal embolization technique will minimize these events (143). Other complications Embolization with cyanoacrylates has caused permanent adhesion of the microcatheter tip to the embolized vessel in a small percentage of cases. This incidence has decreased significantly in recent years with the use of NBCA rather than isobutyl cyanoacrylate, a wedged microcatheter position to prevent proximal reflux, more dilute NBCA/ethiodol mixtures with slower polymerization rates, and more durable microcatheters with hydrophilic coatings. Careful technique also is important. Redundant loops should be removed before

Figure 19 Partially embolized AVM nidus with venous stagnation (C, arrowheads). Note venous stenosis (B, large arrow) and venous aneurysm (B, small arrow). (A) Lateral angiogram-arterial phase, (B) lateral angiogram-venous phase, and (C) lateral angiogram-late venous phase. Abbreviation: AVM, arteriovenous malformation.

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

Figure 20 Partially embolized AVM with arterial stagnation (arrowheads, B). (A) Lateral angiogram-arterial phase and (B) lateral angiogram-venous phase. Abbreviation: AVM, arteriovenous malformation.

embolization, and the microcatheter should be aspirated and pulled briskly during removal (109). Retained catheters usually can be transected and buried in the femoral arteriotomy without adverse sequelae; however, brain and lower extremity ischemic complications have been reported (143,149). Pulmonary emboli (PE) have been reported with both particulate and liquid embolic agents (Fig. 21). Most are asymptomatic, although respiratory distress

299

and death have occurred. The risk of PE using NBCA is increased with high-flow fistulas, with the use of ethiodol or glacial acetic acid to slow the polymerization time and by not using flow-arrest techniques. PE were found in 12 (35%) of a series of 47 pediatric patients after brain AVM embolization. The large majority (45) were asymptomatic and found incidentally on chest X rays. The most common agent was cyanoacrylate (10 out of 12), causing respiratory distress in two (150). A retrospective review of 182 patients embolized with cyanoacrylates found three cases of symptomatic PE, associated with the embolization high-flow AVMs without the use of flow-arrest techniques (151). Multistage angiographic and embolization procedures result in significant radiation doses. Temporary alopecia has been reported, which typically occurs after a short-term radiation dose of 3 to 6 Gy (152).

CONCLUSIONS Brain AVMs are very heterogeneous, rare central nervous system vascular malformations associated with significant long-term morbidity and mortality. Embolization has become an increasingly important therapeutic option, usually in combination with surgery or stereotactic radiosurgery. It is, however, associated with risks that must be considered in the context of the overall treatment plan. A multispecialty team comprised of experts in vascular neurosurgery, interventional neuroradiology, and radiosurgery optimally manages brain AVMs. Rapid advances in technology have had a profound impact on brain AVM embolization, and the innovations promise to continue. Dramatic improvements in microcatheter and guidewire technology have led to the superselective catheterization of distal cerebral arteries. This has led to a better understanding of AVM angioarchitecture and has enabled the intranidal embolization of AVMs with liquid embolic agents. Nonetheless, further work is required to improve the safety and efficacy of embolization, and more rigorous data on the natural history of brain AVMs and treatment outcomes are needed.

REFERENCES

Figure 21 NBCA pulmonary embolus (A, arrow) on chest CT scan causing pulmonary infarct (B, arrow). Abbreviation: NBCA, N-butyl cyanoacrylate.

1. Mulliken J. Classification of vascular birthmarks. In: Mulliken JB, Young AE, eds. Vascular Birthmarks: Hemangiomas and Malformations. Philadelphia: WB Saunders, 1988:24–37. 2. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69(3):412–422. 3. McCormick WF, Nofzinger JD. ‘‘Cryptic’’ vascular malformations of the central nervous system. J Neurosurg 1966; 24(5):865–875. 4. McCormick WF, Hardman JM, Boulter TR. Vascular malformations (‘‘angiomas’’) of the brain, with special reference to those occurring in the posterior fossa. J Neurosurg 1968; 28(3):241–251.

300

Weigele et al.

5. McCormick WF. The pathology of vascular (‘‘arteriovenous’’) malformations. J Neurosurg 1966; 24(4):807–816. 6. Awad IA, Robinson JR Jr., Mohanty S, et al. Mixed vascular malformations of the brain: clinical and pathogenetic considerations. Neurosurgery 1993; 33(2):179–188; discussion 188. 7. Chaloupka JC, Huddle DC. Classification of vascular malformations of the central nervous system. Neuroimaging Clin N Am 1998; 8(2):295–321. 8. Herzig R, Burval S, Vladyka V, et al. Familial occurrence of cerebral arteriovenous malformation in sisters: case report and review of the literature. Eur J Neurol 2000; 7(1):95–100. 9. Mullan S, Mojtahedi S, Johnson DL, et al. Embryological basis of some aspects of cerebral vascular fistulas and malformations. J Neurosurg 1996; 85(1):1–8. 10. Weigele JB, Chaloupka JC, Lesley WS. De novo development of multiple cerebral arteriovenous malformations. In: American Society of Neuroradiology Annual Meeting, April 21–27, 2001; Boston, Massachusetts. 11. Friedman JA, Pollock BE, Nichols DA. Development of a cerebral arteriovenous malformation documented in an adult by serial angiography. Case report. J Neurosurg 2000; 93(6):1058–1061. 12. Kader A, Goodrich JT, Sonstein WJ, et al. Recurrent cerebral arteriovenous malformations after negative postoperative angiograms [see comment]. J Neurosurg 1996; 85(1):14–18. 13. Berman MF, Sciacca RR, Pile-Spellman J, et al. The epidemiology of brain arteriovenous malformations. Neurosurgery 2000; 47(2):389–396; discussion 397. 14. Al-Shahi R, Warlow C. A systematic review of the frequency and prognosis of arteriovenous malformations of the brain in adults. Brain 2001; 124(pt 10):1900–1926. 15. Brown RD Jr., Wiebers DO, Torner JC, et al. Incidence and prevalence of intracranial vascular malformations in Olmsted County, Minnesota, 1965 to 1992. Neurology 1996; 46(4):949–952. 16. Jessurun GA, Kamphuis DJ, van der Zande FH, et al. Cerebral arteriovenous malformations in The Netherlands Antilles. High prevalence of hereditary hemorrhagic telangiectasia-related single and multiple cerebral arteriovenous malformations. Clin Neurol Neurosurg 1993; 95(3):193–198. 17. Al-Shahi R, Fang JS, Lewis S, et al. The prevalence of adults with symptomatic arteriovenous malformations of the brain. J Neurol 2000; 247(suppl 3):31 (abstr). 18. Al-Shahi R, Fang JS, Lewis SC, et al. Prevalence of adults with brain arteriovenous malformations: a community based study in Scotland using capture-recapture analysis. J Neurol Neurosurg Psychiatry 2002; 73(5):547–551. 19. Stapf C, Mast H, Sciacca RR, et al. The New York Islands AVM Study: design, study progress, and initial results. Stroke 2003; 34(5):e29–e33. 20. ApSimon HT, Reef H, Phadke RV, et al. A populationbased study of brain arteriovenous malformation: longterm treatment outcomes. Stroke 2002; 33(12):2794–2800. 21. Ogilvy CS, Stieg PE, Awad I, et al. Recommendations for the management of intracranial arteriovenous malformations: a statement for healthcare professionals from a special writing group of the Stroke Council, American Stroke Association. Circulation 2001; 103(21):2644–2657. 22. Brown RD Jr., Wiebers DO, Torner JC, et al. Frequency of intracranial hemorrhage as a presenting symptom and subtype analysis: a population-based study of intracranial vascular malformations in Olmsted County, Minnesota. J Neurosurg 1996; 85(1):29–32. 23. Duong DH, Young WL, Vang MC, et al. Feeding artery pressure and venous drainage pattern are primary deter-

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40. 41. 42.

43.

minants of hemorrhage from cerebral arteriovenous malformations. Stroke 1998; 29(6):1167–1176. Horton JC, Chambers WA, Lyons SL, et al. Pregnancy and the risk of hemorrhage from cerebral arteriovenous malformations. Neurosurgery 1990; 27(6):867–871; discussion 871–872. Han PP, Ponce FA, Spetzler RF. Intention-to-treat analysis of Spetzler-Martin grades IV and V arteriovenous malformations: natural history and treatment paradigm [see comment]. J Neurosurg 2003; 98(1):3–7. Mansmann U, Meisel J, Brock M, et al. Factors associated with intracranial hemorrhage in cases of cerebral arteriovenous malformation. Neurosurgery 2000; 46(2):272–279. Langer DJ, Lasner TM, Hurst RW, et al. Hypertension, small size, and deep venous drainage are associated with risk of hemorrhagic presentation of cerebral arteriovenous malformations. Neurosurgery 1998; 42(3):481–486. Nataf F, Meder JF, Merienne L, et al. [Therapeutic strategy for cerebral arteriovenous malformations. Proposal for classification of individual hemorrhagic risk.] Neurochirurgie 1998; 44(2):83–93. Mansmann U, Meisel J, Brock M, et al. Factors associated with intracranial hemorrhage in cases of cerebral arteriovenous malformation. Neurosurgery 2000; 46(2):272–279; discussion 279–281. Crawford PM, West CR, Chadwick DW, et al. Arteriovenous malformations of the brain: natural history in unoperated patients. J Neurol Neurosurg Psychiatry 1986; 49(1):1–10. Mast H, Young WL, Koennecke HC, et al. Risk of spontaneous haemorrhage after diagnosis of cerebral arteriovenous malformation. Lancet 1997; 350(9084):1065–1068. Brown RD Jr., Wiebers DO, Forbes G, et al. The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 1988; 68(3):352–357. Graf CJ, Perret GE, Torner JC. Bleeding from cerebral arteriovenous malformations as part of their natural history. J Neurosurg 1983; 58(3):331–337. Itoyama Y, Uemura S, Ushio Y, et al. Natural course of unoperated intracranial arteriovenous malformations: study of 50 cases. J Neurosurg 1989; 71(6):805–809. Ondra SL, Troupp H, George ED, et al. The natural history of symptomatic arteriovenous malformations of the brain: a 24-year follow-up assessment. J Neurosurg 1990; 73(3):387–391. Halim AX, Johnston SC, Singh V, et al. Longitudinal risk of intracranial hemorrhage in patients with arteriovenous malformation of the brain within a defined population. Stroke 2004; 35(7):1697–1702. Hartmann A, Mast H, Mohr JP, et al. Morbidity of intracranial hemorrhage in patients with cerebral arteriovenous malformation [see comment]. Stroke 1998; 29(5):931–934. Valavanis A. The role of angiography in the evaluation of cerebral vascular malformations. Neuroimaging Clin N Am 1996; 6(3):679–704. Yasargil M. AVM of the brain, history, embryology, pathological considerations, hemodynamics, diagnostic studies, microsurgical anatomy. In: Microneurosurgery. New York: Thieme Verlag, 1987. Berenstein A, Lasjaunias P. Surgical Neuroangiography. Berlin, Heidelberg, New York: Springer Verlag, 1991. Hamby WB. The pathology of supratentorial angiomas. J Neurosurg 1958; 15(1):65–75. Hayashi S, Arimoto T, Itakura T, et al. The association of intracranial aneurysms and arteriovenous malformation of the brain. Case report. J Neurosurg 1981; 55(6):971–975. Perata HJ, Tomsick TA, Tew JM Jr. Feeding artery pedicle aneurysms: association with parenchymal hemorrhage

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57. 58.

59.

60.

61.

and arteriovenous malformation in the brain. J Neurosurg 1994; 80(4):631–634. Stapf C, Mohr JP, Pile-Spellman J, et al. Concurrent arterial aneurysms in brain arteriovenous malformations with haemorrhagic presentation. J Neurol Neurosurg Psychiatry 2002; 73(3):294–298. Halim AX, Singh V, Johnston SC, et al. Characteristics of brain arteriovenous malformations with coexisting aneurysms: a comparison of two referral centers. Stroke 2002; 33(3):675–679. Meisel HJ, Mansmann U, Alvarez H, et al. Cerebral arteriovenous malformations and associated aneurysms: analysis of 305 cases from a series of 662 patients. Neurosurgery 2000; 46(4):793–800. Reporting terminology for brain arteriovenous malformation clinical and radiographic features for use in clinical trials. Stroke 2001; 32(6):1430–1442. Perret G, Nishioka H. Report on the cooperative study of intracranial aneurysms and subarachnoid hemorrhage. Section VI. Arteriovenous malformations. An analysis of 545 cases of cranio-cerebral arteriovenous malformations and fistulae reported to the cooperative study. J Neurosurg 1966; 25(4):467–490. Thompson RC, Steinberg GK, Levy RP, et al. The management of patients with arteriovenous malformations and associated intracranial aneurysms. Neurosurgery 1998; 43(2):202–211. Brown RD Jr., Wiebers DO, Forbes GS. Unruptured intracranial aneurysms and arteriovenous malformations: frequency of intracranial hemorrhage and relationship of lesions. J Neurosurg 1990; 73(6):859–863. Turjman F, Massoud TF, Vinuela F, et al. Aneurysms related to cerebral arteriovenous malformations: superselective angiographic assessment in 58 patients. AJNR Am J Neuroradiol 1994; 15(9):1601–1605. Redekop G, TerBrugge K, Montanera W, et al. Arterial aneurysms associated with cerebral arteriovenous malformations: classification, incidence, and risk of hemorrhage. J Neurosurg 1998; 89(4):539–546. Ezura M, Takahashi A, Jokura H, et al. Endovascular treatment of aneurysms associated with cerebral arteriovenous malformations: experiences after the introduction of Guglielmi detachable coils. J Clin Neurosci 2000; 1:14–18. Rinkel GJ, Djibuti M, Algra A, et al. Prevalence and risk of rupture of intracranial aneurysms: a systematic review. Stroke 1998; 29(1):251–256. Shenkar R, Elliott JP, Diener K, et al. Differential gene expression in human cerebrovascular malformations. Neurosurgery 2003; 52(2):465–477. Fulbright RK, Chaloupka JC, Putman CM, et al. MR of hereditary hemorrhagic telangiectasia: prevalence and spectrum of cerebrovascular malformations. AJNR Am J Neuroradiol 1998; 19(3):477–484. Schievink WI. Genetics and aneurysm formation. Neurosurg Clin N Am 1998; 9(3):485–495. McKissock W, Paterson JH. A clinical survey of intracranial angiomas with special reference to their mode of progression and surgical treatment: a report of 110 cases. Brain 1956; 79(2):233–266. Inci S, Spetzler RF. Intracranial aneurysms and arterial hypertension: a review and hypothesis. Surg Neurol 2000; 53(6):530–540; discussion 540–542. Kerber CW, Liepsch D. Flow dynamics for radiologists. II. Practical considerations in the live human. AJNR Am J Neuroradiol 1994; 15(6):1076–1086. Kerber CW, Liepsch D. Flow dynamics for radiologists. I. Basic principles of fluid flow. AJNR Am J Neuroradiol 1994; 15(6):1065–1075.

301

62. Okamoto S, Handa H, Hashimoto N. Location of intracranial aneurysms associated with cerebral arteriovenous malformation: statistical analysis. Surg Neurol 1984; 22(4):335–340. 63. Cunha e Sa MJ, Stein BM, Solomon RA, et al. The treatment of associated intracranial aneurysms and arteriovenous malformations. J Neurosurg 1992; 77(6): 853–859. 64. Marks MP, Lane B, Steinberg GK, et al. Intranidal aneurysms in cerebral arteriovenous malformations: evaluation and endovascular treatment. Radiology 1992; 183(2): 355–360. 65. Piotin M, Ross IB, Weill A, et al. Intracranial arterial aneurysms associated with arteriovenous malformations: endovascular treatment [see comment]. Radiology 2001; 220(2):506–513. 66. Batjer H, Suss RA, Samson D. Intracranial arteriovenous malformations associated with aneurysms. Neurosurgery 1986; 18(1):29–35. 67. Westphal M, Grzyska U. Clinical significance of pedicle aneurysms on feeding vessels, especially those located in infratentorial arteriovenous malformations. J Neurosurg 2000; 92(6):995–1001. 68. Kaptain GJ, Lanzino G, Do HM, et al. Posterior inferior cerebellar artery aneurysms associated with posterior fossa arteriovenous malformation: report of five cases and literature review [see comment]. Surg Neurol 1999; 51(2):146–152. 69. Khaw AV, Mohr JP, Sciacca RR, et al. Association of infratentorial brain arteriovenous malformations with hemorrhage at initial presentation. Stroke 2004; 35(3):660–663. 70. Luessenhop AJ, Gennarelli TA. Anatomical grading of supratentorial arteriovenous malformations for determining operability. Neurosurgery 1977; 1(1):30–35. 71. Shi YQ, Chen XC. A proposed scheme for grading intracranial arteriovenous malformations. J Neurosurg 1986; 65(4):484–489. 72. Spetzler RF, Martin NA. A proposed grading system for arteriovenous malformations. J Neurosurg 1986; 65(4):476–483. 73. Hamilton MG, Spetzler RF. The prospective application of a grading system for arteriovenous malformations. Neurosurgery 1994; 34(1):2–6; discussion 6–7. 74. Du R, Dowd CF, Johnston SC, et al. Interobserver variability in grading of brain arteriovenous malformations using the Spetzler-Martin system. Neurosurgery 2005; 57(4):668–675; discussion 668–675. 75. Luessenhop AJ, Spence WT. Artificial embolization of cerebral arteries. Report of use in a case of arteriovenous malformation. JAMA 1960; 172:1153–1155. 76. Wolpert SM, Stein BM. Catheter embolization of intracranial arteriovenous malformations as an aid to surgical excision. Neuroradiology 1975; 10(2):73–85. 77. Luessenhop AJ, Rosa L. Cerebral arteriovenous malformations. Indications for and results of surgery, and the role of intravascular techniques. J Neurosurg 1984; 60(1):14–22. 78. Mullan S, Kawanaga H, Patronas NJ. Microvascular embolization of cerebral arteriovenous malformations: a technical variation. J Neurosurg 1979; 51(5):621–627. 79. Stein BM, Wolpert SM. Surgical and embolic treatment of cerebral arteriovenous malformations. Surg Neurol 1977; 7(6):359–369. 80. Serbinenko FA. Balloon catheterization and occlusion of major cerebral vessels. J Neurosurg 1974; 41(2):125–145. 81. Kerber C. Balloon catheter with a calibrated leak. A new system for superselective angiography and occlusive catheter therapy. Radiology 1976; 120(3):547–550.

302

Weigele et al.

82. Debrun G, Vinuela F, Fox A, et al. Embolization of cerebral arteriovenous malformations with bucrylate. J Neurosurg 1982; 56(5):615–627. 83. Purdy PD, Batjer HH, Risser RC, et al. Arteriovenous malformations of the brain: choosing embolic materials to enhance safety and ease of excision. J Neurosurg 1992; 77(2):217–222. 84. Samson D, Marshall D. Carcinogenic potential of isobutyl2-cyanoacrylate. J Neurosurg 1986; 65(4):571–572. 85. Vinters HV, Galil KA, Lundie MJ, et al. The histotoxicity of cyanoacrylates. A selective review. Neuroradiology 1985; 27(4):279–291. 86. Martin NA, Khanna R, Doberstein C, et al. Therapeutic embolizations of arteriovenous malformations: the case for and against. Clin Neurosurg 2000; 46:295–318. 87. Rosenblatt S, Lewis AI, Tew JM. Combined interventional and surgical treatment of arteriovenous malformations. Neuroimaging Clin N Am 1998; 8(2):469–482. 88. Deruty R, Pelissou-Guyotat I, Amat D, et al. Complications after multidisciplinary treatment of cerebral arteriovenous malformations. Acta Neurochir (Wien) 1996; 138(2):119–131. 89. NBCA Trial Investigators. N-butyl cyanoacrylate embolization of cerebral arteriovenous malformations: results of a prospective, randomized, multi-center trial. AJNR Am J Neuroradiol 2002; 23(5):748–755. 90. Gobin YP, Laurent A, Merienne L, et al. Treatment of brain arteriovenous malformations by embolization and radiosurgery [see comment]. J Neurosurg 1996; 85(1):19–28. 91. Friedman WA, Bova FJ, Mendenhall WM. Linear accelerator radiosurgery for arteriovenous malformations: the relationship of size to outcome. J Neurosurg 1995; 82(2): 180–189. 92. Pollock BE, Flickinger JC, Lunsford LD, et al. Factors associated with successful arteriovenous malformation radiosurgery. Neurosurgery 1998; 42(6):1239–1244; discussion 1244–1247. 93. Karlsson B, Lindquist C, Steiner L. Prediction of obliteration after gamma knife surgery for cerebral arteriovenous malformations. Neurosurgery 1997; 40(3):425–430. 94. Mathis JA, Barr JD, Horton JA, et al. The efficacy of particulate embolization combined with stereotactic radiosurgery for treatment of large arteriovenous malformations of the brain. AJNR Am J Neuroradiol 1995; 16(2): 299–306. 95. Firlik AD, Levy EI, Kondziolka D, et al. Staged volume radiosurgery followed by microsurgical resection: a novel treatment for giant cerebral arteriovenous malformations: technical case report. Neurosurgery 1998; 43(5):1223–1228. 96. Marks MP, Lane B, Steinberg GK, et al. Endovascular treatment of cerebral arteriovenous malformations following radiosurgery. AJNR Am J Neuroradiol 1993; 14(2): 297–303; discussion 304–305. 97. Steinberg GK, Chang SD, Levy RP, et al. Surgical resection of large incompletely treated intracranial arteriovenous malformations following stereotactic radiosurgery. J Neurosurg 1996; 84(6):920–928. 98. Schaller C, Schramm J. Microsurgical results for small arteriovenous malformations accessible for radiosurgical or embolization treatment. Neurosurgery 1997; 40(4): 664–672. 99. Hurst RW, Berenstein A, Kupersmith MJ, et al. Deep central arteriovenous malformations of the brain: the role of endovascular treatment. J Neurosurg 1995; 82(2): 190–195. 100. Kwon OK, Han DH, Han MH, et al. Palliatively treated cerebral arteriovenous malformations: follow-up results. J Clin Neurosci 2000; 7(suppl 1):69–72. 101. Konan AV, Roy D, Raymond J. Endovascular treatment of hemifacial spasm associated with a cerebral arteriovenous

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

malformation using transvenous embolization: case report. Neurosurgery 1999; 44(3):663–666. Sugita M, Takahashi A, Ogawa A, et al. Improvement of cerebral blood flow and clinical symptoms associated with embolization of a large arteriovenous malformation: case report. Neurosurgery 1993; 33(4):748–751; discussion 752. Martin NA, Khanna R, Doberstein C, et al. Therapeutic embolization of arteriovenous malformations: the case for and against. Clin Neurosurg 2000; 46:295–318. Richling B, Killer M. Endovascular management of patients with cerebral arteriovenous malformations. Neurosurg Clin N Am 2000; 11(1):123–145. Djurberg HG, Tjan GT, Al Moutaery KR. Enhanced catheter propagation with hypercapnia during superselective cerebral catherisation. Neuroradiology 1998; 40(7): 466–468. Aletich VA, Debrun GM, Koenigsberg R, et al. Arteriovenous malformation nidus catheterization with hydrophilic wire and flow-directed catheter. AJNR Am J Neuroradiol 1997; 18(5):929–935. The Arteriovenous Malformation Study Group. Arteriovenous malformations of the brain in adults. N Engl J Med 1999; 340(23):1812–1818. Vinuela F, Dion JE, Duckwiler G, et al. Combined endovascular embolization and surgery in the management of cerebral arteriovenous malformations: experience with 101 cases. J Neurosurg 1991; 75(6):856–864. Debrun GM, Aletich V, Ausman JI, et al. Embolization of the nidus of brain arteriovenous malformations with n-butyl cyanoacrylate. Neurosurgery 1997; 40(1):112–120. Rauch RA, Vinuela F, Dion J, et al. Preembolization functional evaluation in brain arteriovenous malformations: the superselective Amytal test. AJNR Am J Neuroradiol 1992; 13(1):303–308. Fitzsimmons BF, Marshall RS, Pile-Spellman J, et al. Neurobehavioral differences in superselective Wada testing with amobarbital versus lidocaine. AJNR Am J Neuroradiol 2003; 24(7):1456–1460. Moo LR, Murphy KJ, Gailloud P, et al. Tailored cognitive testing with provocative amobarbital injection preceding AVM embolization. AJNR Am J Neuroradiol 2002; 23(3): 416–421. Paulsen RD, Steinberg GK, Norbash AM, et al. Embolization of rolandic cortex arteriovenous malformations. Neurosurgery 1999; 44(3):479–484; discussion 484–486. Valavanis A. The role of angiography in the evaluation of cerebral vascular malformations. Neuroimaging Clin N Am 1996; 6(3):679–704. Howington JU, Kerber CW, Hopkins LN. Liquid embolic agents in the treatment of intracranial arteriovenous malformations. Neurosurg Clin N Am 2005; 16(2):355–363. Jahan R, Murayama Y, Gobin YP, et al. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients [see comment]. Neurosurgery 2001; 48(5):984–995. Yakes WF, Krauth L, Ecklund J, et al. Ethanol endovascular management of brain arteriovenous malformations: initial results. Neurosurgery 1145; 40(6):1145–1152. Song JK, Eskridge JM, Chung EC, et al. Preoperative embolization of cerebral arteriovenous malformations with silk sutures: analysis and clinical correlation of complications revealed on computerized tomography scanning. J Neurosurg 2000; 92(6):955–960. Sorimachi T, Koike T, Takeuchi S, et al. Embolization of cerebral arteriovenous malformations achieved with polyvinyl alcohol particles: angiographic reappearance and complications. AJNR Am J Neuroradiol 1999; 20(7): 1323–1328.

Chapter 15: Endovascular Management of Brain Arteriovenous Malformations 120. Germano IM, Davis RL, Wilson CB, et al. Histopathological follow-up study of 66 cerebral arteriovenous malformations after therapeutic embolization with polyvinyl alcohol. J Neurosurg 1992; 76(4):607–614. 121. Gruber A, Mazal PR, Bavinzski G, et al. Repermeation of partially embolized cerebral arteriovenous malformations: a clinical, radiologic, and histologic study. AJNR Am J Neuroradiol 1996; 17(7):1323–1331. 122. Howington JU, Kerber CW, Hopkins LN. Liquid embolic agents in the treatment of intracranial arteriovenous malformations. Neurosurg Clin N Am 2005; 16(2): 355–363, ix–x. 123. Lieber BB, Wakhloo AK, Siekmann R, et al. Acute and chronic swine rete arteriovenous malformation models: effect of ethiodol and glacial acetic acid on penetration, dispersion, and injection force of N-butyl 2-cyanoacrylate. AJNR Am J Neuroradiol 2005; 26(7):1707–1714. 124. Wikholm G. Occlusion of cerebral arteriovenous malformations with N-butyl cyano-acrylate is permanent. AJNR Am J Neuroradiol 1995; 16(3):479–482. 125. Yu SC, Chan MS, Lam JM, et al. Complete obliteration of intracranial arteriovenous malformation with endovascular cyanoacrylate embolization: initial success and rate of permanent cure. AJNR Am J Neuroradiol 1139; 25 (7):1139–1143. 126. Jafar JJ, Davis AJ, Berenstein A, et al. The effect of embolization with N-butyl cyanoacrylate prior to surgical resection of cerebral arteriovenous malformations. J Neurosurg 1993; 78(1):60–69. 127. Spetzler RF, Wilson CB, Weinstein P, et al. Normal perfusion pressure breakthrough theory. Clin Neurosurg 1978; 25:651–672. 128. Picard L, Da Costa E, Anxionnat R, et al. Acute spontaneous hemorrhage after embolization of brain arteriovenous malformation with N-butyl cyanoacrylate. J Neuroradiol 2001; 28(3):147–165. 129. Jahan R, Murayama Y, Gobin YP, et al. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients. Neurosurgery 2001; 48 (5):984–995; discussion 995–997. 130. Taki W, Yonekawa Y, Iwata H, et al. A new liquid material for embolization of arteriovenous malformations. AJNR Am J Neuroradiol 1990; 11(1):163–168. 131. Murayama Y, Vinuela F, Ulhoa A, et al. Nonadhesive liquid embolic agent for cerebral arteriovenous malformations: preliminary histopathological studies in swine rete mirabile. Neurosurgery 1998; 43(5):1164–1175. 132. Akin ED, Perkins E, Ross IB. Surgical handling characteristics of an ethylene vinyl alcohol copolymer compared with N-butyl cyanoacrylate used for embolization of vessels in an arteriovenous malformation resection model in swine. J Neurosurg 2003; 98(2):366–370. 133. Chaloupka JC, Vinuela F, Vinters HV, et al. Technical feasibility and histopathologic studies of ethylene vinyl copolymer (EVAL) using a swine endovascular embolization model. AJNR Am J Neuroradiol 1994; 15(6):1107–1115. 134. Pasqualin A, Scienza R, Cioffi F, et al. Treatment of cerebral arteriovenous malformations with a combination of preoperative embolization and surgery. Neurosurgery 1991; 29(3):358–368. 135. DeMeritt JS, Pile-Spellman J, Mast H, et al. Outcome analysis of preoperative embolization with N-butyl cyanoacrylate in cerebral arteriovenous malformations. AJNR Am J Neuroradiol 1995; 16(9):1801–1807. 136. Jordan JE, Marks MP, Lane B, et al. Cost-effectiveness of endovascular therapy in the surgical management of

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

303

cerebral arteriovenous malformations. AJNR Am J Neuroradiol 1996; 17(2):247–254. Shin M, Kawahara N, Maruyama K, et al. Risk of hemorrhage from an arteriovenous malformation confirmed to have been obliterated on angiography after stereotactic radiosurgery [see comment]. J Neurosurg 2005; 102(5): 842–846. Fournier D, TerBrugge KG, Willinsky R, et al. Endovascular treatment of intracerebral arteriovenous malformations: experience in 49 cases. J Neurosurg 1991; 75(2): 228–233. Valavanis A, Yasargil MG. The endovascular treatment of brain arteriovenous malformations. Adv Tech Stand Neurosurg 1998; 24:131–214. Frizzel RT, Fisher WS III. Cure, morbidity, and mortality associated with embolization of brain arteriovenous malformations: a review of 1246 patients in 32 series over a 35-year period. Neurosurgery 1995; 37(6):1031–1039; discussion 1039–1040. Miyamoto S, Hashimoto N, Nagata I, et al. Posttreatment sequelae of palliatively treated cerebral arteriovenous malformations. Neurosurgery 2000; 46(3):589–594; discussion 594–595. Fox AJ, Pelz DM, Lee DH. Arteriovenous malformations of the brain: recent results of endovascular therapy. Radiology 1990; 177(1):51–57. Haw CS, terBrugge K, Willinsky R, et al. Complications of embolization of arteriovenous malformations of the brain. J Neurosurg 2006; 104(2):226–232. Hartmann A, Pile-Spellman J, Stapf C, et al. Risk of endovascular treatment of brain arteriovenous malformations. Stroke 2002; 33(7):1816–1820. Taylor CL, Dutton K, Rappard G, et al. Complications of preoperative embolization of cerebral arteriovenous malformations [see comment]. J Neurosurg 2004; 100(5): 810–812. Duckwiler GR, Dion JE, Vinuela F, et al. Delayed venous occlusion following embolotherapy of vascular malformations in the brain. AJNR Am J Neuroradiol 1992; 13(6): 1571–1579. Jafar JJ, Rezai AR. Acute surgical management of intracranial arteriovenous malformations. Neurosurgery 1994; 34(1):8–12; discussion 12–13. Sipos EP, Kirsch JR, Nauta HJ, et al. Intra-arterial urokinase for treatment of retrograde thrombosis following resection of an arteriovenous malformation. Case report. J Neurosurg 1992; 76(6):1004–1007. Ruckert RI, Bender A, Rogalla P. Popliteal artery occlusion as a late complication of liquid acrylate embolization for cerebral vascular malformation. J Vasc Surg 1999; 29(3): 561–565. Kjellin IB, Boechat MI, Vinuela F, et al. Pulmonary emboli following therapeutic embolization of cerebral arteriovenous malformations in children. Pediatr Radiol 2000; 30 (4):279–283. Pelz DM, Lownie SP, Fox AJ, et al. Symptomatic pulmonary complications from liquid acrylate embolization of brain arteriovenous malformations. AJNR Am J Neuroradiol 1995; 16(1):19–26. Wen CS, Lin SM, Chen Y, et al. Radiation-induced temporary alopecia after embolization of cerebral arteriovenous malformations. Clin Neurol Neurosurg 2003; 105 (3):215–217. Meisel HJ, Mansmann U, Alvarez H, et al. Effect of partial targeted N-butyl-cyano-acrylate embolization in brain AVM. Acta Neurochir (Wien) 2002; 144(9):879–887.

16 Endovascular Treatment of Acute Ischemic Stroke Mayur A. Paralkar Department of Medicine, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, U.S.A.

Alexandros L. Georgiadis and Adnan I. Qureshi Department of Neurology, Zeenat Qureshi Stroke Research Center, University of Minnesota, Minneapolis, Minnesota, U.S.A.

Qaisar A. Shah Department of Neurology, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, U.S.A., and Department of Neurology, University of Minnesota, Minneapolis, Minnesota, U.S.A.

INTRODUCTION: THE EVOLUTION OF STROKE TREATMENT It is well known that stroke is one of the leading causes of death and disability in Western societies (1). However, progress in the treatment of stroke has been slow. For many years, neurologists could only localize lesions and describe the resulting syndromes, i.e., establish clinicopathological correlations. Little was known about the etiology of stroke, and the prevailing presumption was that in situ thrombosis of the intracranial vasculature was the most common culprit (2). The observation that thrombotic material could embolize from the carotid artery to the intracranial vessels (3) and the description of lacunar syndromes and their underlying pathophysiology (4) were pivotal in helping neurologists advance to the next stage, which was to assess the etiology of stroke. Thus, strategies for secondary prevention could emerge, but there was still no acute treatment available. Initial attempts to use systemic thrombolysis in stroke were hampered by the propensity of cerebral tissue to bleed. It was not clear how to dose thrombolytic medications, what the therapeutic time window was, and which patients were at highest risk of complications. The first three trials of intravenous thrombolysis for acute stroke used streptokinase and were all terminated because of excessive rates of intracerebral hemorrhage (ICH) in the treated patients (5–7). They were followed by one more negative trial, the European Cooperative Acute Stroke Study (ECASS), which used recombinant tissue plasminogen activator (r-tPA) at a dose of 1.1 mg/ kg within six hours from symptom onset (8). The National Institute of Neurological Diseases and Stroke (NINDS) study, using a smaller r-tPA dose of 0.9 mg/kg only within three hours from symptom onset, demonstrated significant clinical benefit for the treated patients

with an acceptable risk of ICH (9). On the basis of the results of this study, r-tPA was approved for intravenous use in acute stroke within three hours from symptom onset by the Food and Drug Administration (FDA). The three-hour window for the administration of intravenous r-tPA was later confirmed by the Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS) trial, which demonstrated no benefit for patients treated at three to five hours after symptom onset (10). To this day, intravenous thrombolysis with r-tPA remains the only FDA-approved treatment for acute stroke. Intra-arterial thrombolysis, first reported in 1982 (11), offers several advantages. These include visualization of the actual vascular lesion, a therapeutic window that extends to six hours, administration of smaller doses of thrombolytic medication, and the possibility of combining pharmacological thrombolysis with mechanical disruption of the clot (mechanical thrombolysis). The smaller dose of thrombolytic medication makes it possible to intra-arterially treat patients who do not qualify for intravenous thrombolysis, such as those in the immediate postoperative period (12). In some cases, intra-arterial thrombolysis is performed six hours after stroke onset, such as in posterior circulation strokes associated with basilar artery thrombosis (13). In those cases, the high mortality of the untreated patients justifies intra-arterial thrombolysis at even 12 hours or more after stroke onset.

INTRA-ARTERIAL THROMBOLYSIS Intra-arterial Thrombolysis as Sole Treatment There were some early promising reports of patients treated with intra-arterial thrombolysis (14), but the

306

Paralkar et al.

first randomized trial (Prolyse in Acute Cerebral Thromboembolism, or PROACT) was not published before 1998 (15). This trial showed better recanalization rates for patients randomized to intra-arterial recombinant prourokinase, but no clinical benefit. In PROACT II, 110 patients were randomized to intraarterial recombinant prourokinase and intravenous heparin versus heparin alone (16). The patients who were treated with thrombolytic medications had a higher likelihood of partial or complete recanalization (67% vs. 18%) and of living independently at three months (40% vs. 25%). The rate of symptomatic ICH was 11% in the patients treated with thrombolytic medications versus 3% in the untreated patients. A series of 54 patients treated with intra-arterial urokinase published by Suarez et al. also showed significant clinical benefit, but with a higher rate of symptomatic ICH (17). These data indicate that intra-arterial thrombolysis is an effective treatment for patients who present at three to six hours after symptom onset. The reported rates of symptomatic ICH are consistently higher than with intravenous r-tPA given within three hours, but very similar to that shown in ATLANTIS, in which r-tPA was given at three to five hours after onset. All these studies used intravenous heparin in varying doses, a practice that is now outdated and certainly increases the risk of ICH. Moreover, those patients were treated with first (urokinase) or secondgeneration (prourokinase) thrombolytic agents. Thirdgeneration thrombolytic agents available today, such as reteplase, offer higher clot specificity and promise less hemorrhagic complications.

Intra-arterial Thrombolysis and Intravenous Antiplatelet Drugs Clot formation and lysis is a dynamic process. Thrombolytic medications lead to clot lysis but at the same time activate thrombin and platelets, thus promoting rethrombosis. Platelet glycoprotein IIb/IIIa receptor activation is an integral part of the cascade that leads to clot formation; hence the idea of combining thrombolysis with glycoprotein IIb/IIIa inhibitors. Several studies have reported promising results, but so far only on small numbers of patients who were treated with a variety of different protocols and medications (23–26). Some recent retrospective data have shown that the combined use of thrombolytic agents and glycoprotein IIb/IIIa inhibitors is an independent positive predictor of recanalization (27).

Multimodal Thrombolysis In case of failure of pharmacological thrombolysis and simple microcatheter maneuvers for clot disruption, more advanced methods of mechanical thrombolysis can be employed, as deemed most appropriate for the given patient (multimodal thrombolysis) (24,25). A multimodal thrombolysis treatment is illustrated in Figures 1–9. The various methods that can be used are discussed below. Acute Angioplasty and Stenting

Acute angioplasty, combined, if needed, with intraarterial r-tPA, has been recently shown to be superior to intra-arterial r-tPA alone in a retrospective analysis of 70 patients (28). Superiority was demonstrated in

Combined Intravenous and Intra-arterial Thrombolysis An alternative treatment algorithm for patients presenting within three hours from symptom onset is to administer intravenous r-tPA at a lower dose, so as to quasi ‘‘reserve’’ some r-tPA for intra-arterial administration. This algorithm is usually applied to patients with large deficits and is based on the observation that patients with a National Institutes of Health Stroke Scale (NIHSS) score of greater than 10 often do not benefit from intravenous thrombolysis (18). Commonly, two-third of the recommended dose is given intravenously (i.e., 0.6 mg/kg), and then a maximal dose of 0.3 mg/kg is administered intraarterially if an occlusion is visualized on angiography (19–22). Some centers have recommended screening patients with diffusion- and perfusion-weighted magnetic resonance imaging (MRI) after the intravenous thrombolytic medication is given and proceeding with angiography only in selected cases (22). Available studies have not evaluated efficacy of the combined intravenous and intra-arterial approach when compared with intravenous thrombolysis alone. One new concept currently under preliminary study is to administer the full dose of intravenous r-tPA followed by small doses of intra-arterial reteplase combined with mechanical thrombolysis.

Figure 1 Angiographic perfusion image. A 20’ cranial AP projection taken after injection, with the guide catheter placed in the proximal aortic arch. There is no filling of the intracranial right carotid artery.

Chapter 16: Endovascular Treatment of Acute Ischemic Stroke

307

Figure 2 AP view of right internal carotid artery injection. (A) Early arterial phase. There is no filling of the right middle cerebral artery. (B) Late arterial-early venous phase. There is filling of the distal M1-middle cerebral artery and of one of its segments. The flow is through pial collaterals from the anterior cerebral artery. The arrows indicate the direction of flow.

Figure 3 The microcatheter is passed through the guide catheter into the right middle cerebral artery. The arrow points at the tip of the microcatheter.

Figure 5 Reteplase is administered inside and at the edges of the clot following clot manipulation with the microcatheter.

Figure 4 Double injection, AP view: a simultaneous injection through the guide catheter and the microcatheter. The middle cerebral artery fills from the microcatheter injection. The proximal middle cerebral artery is thrombosed and does not fill (circle). The arrow points at another clot located in the area of the trifurcation.

Figure 6 Following the administration of reteplase, a snare device is introduced into the middle cerebral artery to help capture and extract the clot.

308

Paralkar et al.

especially when there was no tandem intracranial occlusion present. Acute intracranial and extracranial stenting have been reported in a recent retrospective review of 168 consecutive patients who underwent multimodal reperfusion therapy at the University of Pittsburgh Medical Center (27) to be independent predictors of recanalization. Clot Retrieval Devices

Figure 7 Right internal carotid artery injection (AP view) after reteplase administration and snare maneuvers. There is good filling of the middle cerebral artery.

Figure 8 Follow-up angiography 24 hours later. The patient had been maintained on a continuous infusion of intravenous Integrilin. Right common carotid artery injection, AP view. Patency of the middle cerebral artery is maintained. There is some proximal irregularity representing residual thrombosis (arrow).

terms of incidence of favorable clinical outcome, recanalization rates, and rate of hemorrhage. In select cases of acute stroke, stenting is an additional option. A recent retrospective review (29) showed good results in patients with acute carotid occlusion who were stented, in most cases with a selfexpanding Wallstent (Boston Scientific, Nattick, MA),

Clot retrieval devices have pushed the treatment window for acute stroke to eight hours after onset of symptoms. The Mechanical Embolus Removal in Cerebral Ischemia (MERCI) trial enrolled 151 patients who presented within eight hours from stroke onset (30). It was a prospective, single-arm, multicenter trial. The patients treated had proximal intracranial occlusions demonstrated by conventional angiography (intracranial vertebral arteries, basilar artery, intracranial carotid artery, and middle cerebral artery M1 or M2 segment). The trial showed adequate safety for the procedure with symptomatic ICH in 7.8% of the patients and good clinical outcomes in those with successful recanalization. The Merci Retrieval System (Concentric Medical, Inc., Mountain View, California, U.S.) consists of a 8-French or a 9-French (Fr) guide catheter with a balloon at its distal tip and a microcatheter through which the actual Merci retriever is expressed. The Merci retriever is a tapered wire with five helical loops of decreasing diameter at its distal end. The loops are made of memory-shaped nitinol. The Merci retriever is advanced through the clot and then pulled back to ensnare the thrombus and allow it to be removed. During this procedure the balloon at the tip of the guide catheter, which is located in the common or internal carotid artery, is inflated to minimize distal flow. A full description of the procedure can be found in the original publication of the phase 1 study (31). However, the approval of this device without data from a randomized trial has drawn criticism (32,33). Also, the MERCI trial showed good clinical outcome, defined as mRS of less than or equal to 2, in 28% of the patients as opposed to 40% in PROACT II and a mortality rate of 44%, which was significantly higher than the 27% of the PROACT II study (32). The Magnetic Resonance and Recanalization of Stroke Clots Using Embolectomy (MR RESCUE) trial is an ongoing NINDS-funded study that will assess whether MRI can be used to assess which patients can benefit from intervention with the Merci device. This study will provide us with high level of evidence regarding the safety and efficacy of the Merci retriever. Other clot retrieval devices, such as the Microvena snare (Amplatz Goose Neck, Microvena Corporation, White Bear Lake, Minnesota, U.S.), have also been used with success (34). The Microvena snare is employed through a 6-Fr guide catheter. The size of the employed snare should roughly match the size of the treated vessel. A microcatheter is advanced into

Chapter 16: Endovascular Treatment of Acute Ischemic Stroke

309

the intracranial vessel and in proximity to the clot, and the snare is expressed through the microcatheter. The microcatheter and snare are advanced into the clot, and the snare is then retracted into the microcatheter. After the entire system is pulled out by a few centimeters, a control run through the guide catheter is obtained. If there is no adequate recanalization, the procedure can be repeated (34). Angiojet Catheter

The Possis AngioJet thrombectomy catheter (Possis Medical, Inc., Minneapolis, Minnesota, U.S.) uses high-pressure, pulsed saline to fragment and draw the thrombus into the catheter lumen. The AngioJet has been used mainly in peripheral and cardiac endovascular procedures. Its use in neuroendovascular procedures has been limited (35). It could be considered especially in patients with a large clot burden. EKOS MicroLysUS Infusion Catheter

The EKOS ultrasound-emitting infusion catheter (EKOS Corporation, Bothell, Washington, U.S.) was created because of the evidence, from experimental (36) and clinical studies (37), that ultrasound can enhance the effect of thrombolytic medications on clots. The EKOS catheter was used in the Interventional Management of Stroke (IMS) II trial, the preliminary results of which were announced at the 2006 International Stroke Conference. Those results established the safety of combined intravenous thrombolysis and intraarterial thrombolysis delivered through the EKOS catheter. The results, though promising in terms of clinical outcomes, will require further validation.

Flow Grading Scales To assess the efficacy of intra-arterial treatment, scales that grade blood flow need to be employed. The Thrombolysis in Myocardial Infarction (TIMI) grading scale (38) remains the most commonly used scale. Grade 0 on the TIMI scale means absence of flow, grade 1 indicates that contrast material passes beyond the area of obstruction but does not opacify the entire distal vascular bed, grade 2 denotes complete but delayed opacification of the distal vascular bed, and grade 3 denotes normal flow. The Qureshi grading scale, first reported in 2002, has been shown to have better interobserver reliability and correlation with neurological recovery and mortality (39). This scale grades flow from 0 to 5 and accounts for both the site of occlusion and the presence of collateral flow (Fig. 9).

Figure 9 Qureshi grading scheme. The patterns of occlusion observed in patients with acute ischemic stroke. Note that ICA occlusion (highlighted) constitutes the highest severity grades because of its association with poor outcome. ICA indicates internal carotid artery; MCA, middle cerebral artery; ACA, anterior cerebral artery; VA, vertebral artery; BA, basilar artery; LSA, lenticulostriate arteries; and LMC, leptomenungeal.

to benefit and the least likely to suffer complications. Catheter technology advances will continue to make endovascular procedures safer and allow for more sophisticated interventions. Multimodal intra-arterial thrombolysis will likely be the treatment of choice for many stroke patients in the future, but safety and efficacy will have to be studied further, preferably in the setting of prospective randomized trials.

FUTURE GOALS Advances in stroke care are likely to be made at multiple fronts in the near future. There will likely be safer and more effective thrombolytic and antiplatelet medications. Neuroimaging techniques might evolve to help screen patients, so that interventions are carried out for those patients who are most likely

REFERENCES 1. http://www.cdc.gov/Stroke/Stroke_facts.htm 2. Caplan LR. Stroke: A Clinical Approach. Boston: Butterworth-Heinemann, 1993:195. 3. Fisher CM. Occlusion of the internal carotid artery. Arch Neurol Psychiatry 1951; 65:346–377.

310

Paralkar et al.

4. Fisher CM. The arterial lesions underlying lacunes. Acta Neuropathol 1969; 12:1–15. 5. Hommel M, Boissel JP, Cornu C, et al. Termination of trial of streptokinase in severe acute ischemic stroke. Lancet 1995; 345:57 (letter). 6. Donnan GA, Davis SM, Chambers BR, et al. Trials of streptokinase in acute severe ischemic stroke. Lancet 1995; 345:578–579. 7. MAST-I Group. Randomized controlled trial of streptokinase, aspirin, and combination of both in treatment of acute ischemic stroke. Lancet 1995; 346:1509–1514. 8. Hacke W, Kaste M, Fieschi C, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke. The European Cooperative Acute Stroke Study (ECASS). JAMA 1995; 274:(13)1017–1025. 9. National Institute of Neurological Disorders and Stroke r-tPA Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke: the National Institute of Neurological Disorders and Stroke r-tPA Stroke Study Group. N Engl J Med 1995; 333:1581–1587. 10. Clark WM, Wissman S, Albers GW, et al. Recombinant tissue plasminogen activator (Alteplase) for ischemic stroke 3 to 5 hours after symptom onset. The ATLANTIS study: a randomized controlled trial. Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke. JAMA 1999; 282:(21)2019–2026. 11. Zeumer H, Hacke W, Kolmann HL, et al. Local fibrinolysis in basilar artery thrombosis [in German]. Dtsch Med Wochenschr 1982; 107:728–731. 12. Chalela JA, Katzan I, Liebeskind DS, et al. Safety of intraarterial thrombolysis in the postoperative period. Stroke 2001; 32:(6)1365–1369. 13. Brandt T, von Kummer R, Mueller-Kueppers M, et al. Thrombolytic therapy of acute basilar artery occlusion. Variables affecting recanalization and outcome. Stroke 1996; 27:875–881. 14. del Zoppo GJ, Ferbert A, Otis S, et al. Local intra-arterial fibrinolytic therapy in acute carotid territory stroke: a pilot study. Stroke 1988; 19:307–313. 15. del Zoppo GJ, Higashida RT, Furlan AJ, et al. PROACT: a Phase II randomized trial of recombinant prourokinase by direct arterial delivery in acute middle cerebral artery stroke. PROACT investigators. Prolyse in Acute Cerebral Thromboembolism. Stroke 1998; 29:(1)4–11. 16. Furlan I, Higashida R, Wechsler L, et al. Intra-arterial prourokinase for acute ischemic stroke. The PROACT II study—a randomized controlled trial. Prolyse in Acute Cerebral Thromboembolism. JAMA 1999; 282:2003–2011. 17. Suarez JI, Sunshine JL, Tarr R, et al. Predictors of clinical improvement, angiographic recanalization, and intracranial hemorrhage after intra-arterial thrombolysis for acute ischemic stroke. Stroke 1999; 30:2094–2100. 18. Tomsick T, Brott T, Barsan W, et al. Prognostic value of the hyperdense middle cerebral artery sign and stroke scale score before ultraearly thrombolytic therapy. AJNR Am J Neuroradiol 1996; 17:79–85. 19. The IMS study investigators. Combined intravenous and intra-arterial recanalization for acute ischemic stroke: the Interventional Management of Stroke Study. Stroke 2004; 35:904. 20. Ernst R, Pancioli A, Tomsick T, et al. Combined intravenous and intra-arterial recombinant tissue plasminogen activator in acute ischemic stroke. Stroke 2000; 31:2552–2557. 21. Zaidat OO, Suarez JI, Santillan C, et al. Response to intraarterial and combined intravenous and intra-arterial thrombolytic therapy in patients with distal internal carotid artery occlusion. Stroke 2002; 33:(7)1821–1826.

22. Suarez JI, Zaidat OO, Sunshine JL, et al. Endovascular administration after intravenous infusion of thrombolytic agents for the treatment of patients with acute ischemic stroke. Neurosurgery 2002; 50:(2)251–259. 23. Mangiafico S, Cellerini M, Nencini P, et al. Intravenous tirofiban with intra-arterial urokinase and mechanical thrombolysis in stroke: preliminary experience in 11 cases. Stroke 2005; 36:2154–2158. 24. Abou-Chebl A, Bajzer CT, Krieger DW, et al. Multimodal therapy for the treatment of severe ischemic stroke combining GPIIb/IIIa antagonists and angioplasty after failure of thrombolysis. Stroke 2005; 36:(10)2286–2288. 25. Eckert B, Koch C, Thomalla G, et al. Aggressive therapy with intravenous abciximaband intra-arterial r-tPA and additional PTA/stenting improves clinical outcome in acute vertebrobasilar occlusion: combined local fibrinolysis and intravenous abciximab in acute vertebrobasilar stroke treatment (FAST): results of a multicenter study. Stroke 2005; 36:1160–1165. 26. Qureshi AI, Harris-Lane P, Kirmani JF, et al. Intra-arterial reteplase and intravenous abciximab in patients with acute ischemic stroke: an open-label, dose-ranging, phase I study. Neurosurgery 2006; 59:(4)789–797. 27. Gupta R, Vora NI, Horowitz MB, et al. Multimodal reperfusion therapy for acute ischemic stroke factors prediciting vessel recanalization. Stroke 2006; 37:986–990. 28. Nakano S, Iseda T, Yoneyama T, et al. Direct percutaneous transluminal angioplasty for acute middle cerebral artery occlusion: an alternative option to intra-arterial thrombolysis. Stroke 2002; 33:(12)2872–2876. 29. Jovin TG, Gupta R, Uchino K, et al. Emergent stenting of extracranial internal carotid artery occlusion in acute stroke has a high revascularization rate. Stroke 2005; 36: (11)2426–2430. 30. Gobin YP, Starkman S, Duckwiler GR, et al. MERCI 1: a phase 1 study of mechanical embolus removal in cerebral ischemia. Stroke 2004; 35:(12)2853–2854. 31. Smith WS, Sung G, Starkman S, et al. Safety and efficacy of mechanical embolectomy in acute ischemic stroke: results of the MERCI trial. Stroke 2005; 36:(7)1439–1440. 32. Becker KJ, Brott TG. Approval of the MERCI clot retriever: a critical view. Stroke 2005; 36:400–403. 33. Wechsler LR. Does the Merci retriever work? Against. Stroke 2006; 37:(5)1341–1342. 34. Wikholm G. Transarterial embolectomy in acute stroke. AJNR Am J Neuroradiol 2003; 24:892–894. 35. Bellon RJ, Putman CM, Budzik RF, et al. Rheolytic thrombectomy of the occluded internal carotid artery in the setting of acute ischemic stroke. AJNR Am J Neuroradiol 2001; 22:526–530. 36. Blinc A, Francis CW, Trudnowski JL, et al. Catheterization of ultrasound: potentiated fibrinolysis in vitro. Blood 1993; 81:2636–2643 (abstr). 37. Alexandrov AV, Molina CA, Grotta JC, et al. Ultrasoundenhanced systemic thrombolysis for acute ischemic stroke. N Engl J Med 2004; 351:(21)2170–2178. 38. Chesebro JH, Knatterud G, Roberts R, et al. Thrombolysis in Myocardial Infarction (TIMI) Trial, Phase 1: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase. Clinical findings through hospital discharge. Circulation 1987; 76:142–154. 39. Qureshi AI. New grading system for angiographic evaluation of arterial occlusions and recanalization response to intra-arterial thrombolysis in acute ischemic stroke. Neurosurgery 2002; 50:(6)1405–1414.

17 Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease Eric Sauvageau Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A., and Department of Neurological Surgery, University of South Florida College of Medicine, Tampa, Florida, U.S.A.

Robert D. Ecker Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A., and Department of Neurological Surgery, U.S. Naval Hospital, Okinawa, Japan

Junichi Yamamoto, Ramachandra P. Tummala, Elad I. Levy, and L. Nelson Hopkins Department of Neurosurgery and Toshiba Stroke Research Center, Millard Fillmore Gates Hospital, Kaleida Health, University at Buffalo, State University of New York, Buffalo, New York, U.S.A.

INTRODUCTION Stroke is the third largest cause of death and is the leading cause of permanent disability and disabilityadjusted loss of independent life years in Western countries (1–3). Approximately 700,000 people in the United States experience a stroke annually (3). This number is expected to grow because of aging and changes in the ethnic distribution of the population (4). An estimated one-fourth of strokes are attributable to ischemic events related to occlusive disease of the cervical internal carotid artery (ICA) (5). The benefit of carotid endarterectomy (CEA) in reducing the risk of stroke in patients with moderate to severe (>50%) symptomatic (6–9) or asymptomatic (>60%) (10–13) carotid stenosis has been demonstrated in randomized trials. Although CEA is one of the most common surgical procedures performed in the United States, many patients cannot safely undergo such an extensive operation because of technical or anatomical factors or underlying severe medical illnesses such as coronary artery disease and cardiac failure (14–16). In an analysis of results of the North American Symptomatic Carotid Endarterectomy Trial (NASCET), CEA was approximately 1.5 times more likely to be associated with medical complications in patients with a previous history of myocardial infarction (MI), angina, or hypertension (17). Moreover, the benefits of carotid revascularization surgery shown by the NASCET (6,9), the Asymptomatic Carotid Atherosclerosis Study (ACAS) (11), the Asymptomatic

Carotid Surgery Trial (ACST) (12), and the European Carotid Surgery Trial (ECST) (18,19) are lost if the 30-day rate of perioperative stroke or death exceeds 6% for patients with symptomatic carotid stenosis or 3% for those with asymptomatic carotid stenosis. With the recent advent of embolic protection techniques, standard surgical techniques for extracranial carotid artery (CA) stenosis in high-risk surgical patients have been challenged by catheter-based angioplasty and stenting (20). In August 2004, the U.S. Food and Drug Administration (FDA) approved the first CA stenting system (Acculink stent and Accunet embolic protection device, Guidant, Santa Clara, California, U.S.) for use in patients, with greater than or equal to 50% symptomatic and greater than or equal to 80% asymptomatic carotid stenosis, who were viewed by the treating surgeon as high-risk for CEA because of anatomical risks or medical comorbidities (21,22). The subsequent Centers for Medicare & Medicaid Services (CMS) coverage decision (23), which allowed reimbursement for patients with greater than or equal to 70% symptomatic stenosis or who were enrolled in FDA-sponsored clinical trials, gave CA stenting an entry to the clinical arena as a legitimate alternative to CEA. In this chapter, trials comparing CA stenting with CEA before and after the availability of embolic protection devices are reviewed, indications for treatment of cervical CA disease and patient selection are reviewed, endovascular technique for carotid revascularization is described, and advantages of this approach are

312

Sauvageau et al.

discussed. Procedure-related limitations and complications are also discussed.

CAROTID ANGIOPLASTY AND STENTING: THE EVOLUTION The results of major trials have validated CEA and have shown annual absolute reductions in risk for stroke of approximately 1% for asymptomatic patients (10–13) and 8% for symptomatic patients (8,9,24). However, trials evaluating CEA have systematically excluded patients considered to be ‘‘high-surgical-risk’’ candidates (Table 1) (6,8–13). These important limitations were behind the rationale for developing CA stenting as a less-invasive endovascular approach to carotid revascularization. Moreover, the publication of results obtained with coronary balloon angioplasty and stentassisted balloon angioplasty played a supporting role in the performance of studies in which endovascular and surgical approaches for the treatment of CA disease were initially compared.

Patient Selection Evidence in the literature documents a much greater risk for CEA in clinical practice than is reflected in major CEA trials in which the lowest risk patients were operated on by experienced surgeons performing a relatively high volume of procedures (25–29). Nevertheless, although surgical experience may be an important factor contributing to this significant difference in complication rates, careful patient selection has been found to be the key determinant in maintaining a low perioperative complication rate (14,15,17,30). The following conditions or characteristics have been shown to predispose patients to a high perioperative risk of stroke and death in various CEA reports (14,15,17,30). Because patients with one or more of these risk factors were generally excluded from enrollTable 1 Exclusion Criteria for CEA Trials l l l l

l l l l l l

l

l

l

Older than 79 years of age Heart, kidney, liver, or lung failure Cancer likely to cause death within 5 yr Cardiac valvular lesion or rhythm disorder likely to be associated with cardioembolic stroke Previous ipsilateral CEA Contralateral CEA within 4 mo Angina or MI within the previous 6 mo Progressive neurological signs Major surgical procedure within 30 days Severe comorbidity due to other surgical or medical illness Cerebrovascular events in the distribution of the study CA with ongoing disabling symptoms Symptoms referable to the contralateral side within the previous 45 days More severe stenosis of an intracranial lesion than of the treated lesion

Abbreviations: CEA, carotid endarterectomy; MI, myocardial infarction; CA, carotid artery.

ment in prospective CEA trials, the indications for and the results of surgery in the following subgroups have not been established. Severe Coronary Artery Disease

The coexistence of severe CA disease and symptomatic coronary artery disease represents a dilemma for the clinician (17). Surgical repair of one condition cannot be accomplished without significant risk of complication from the other. In a NASCET subgroup analysis, patients who had prior treatment of coronary artery disease had a lower CEA complication rate than those who had undiagnosed coronary artery disease (31). This difference may be the result of improved cardiac and general medical care in patients undergoing treatment for coronary artery disease, many of whom may not have previously received regular long-term medical care. Adjunct to Coronary Bypass Operation

In a series of 539 patients who underwent noninvasive testing for the detection of CA occlusive disease before undergoing coronary artery bypass grafting (CABG), carotid stenosis severity of more than 75% was an independent predictor of stroke risk during the bypass operation (32). For patients with severe coexistent disease of the carotid and coronary arteries, there is some debate whether revascularization is appropriate for both conditions. Certainly, controversy exists regarding the timing of the procedures. Surgical options include the performance of a simultaneous procedure or a staged approach in which one procedure is performed several days after the other. A combination of CEA and CABG reportedly is associated with a risk of stroke or death ranging from 7.4% to 9.4%, which is roughly 1.5 to 2.0 times the independent risk of each operation (17). In a multicenter review, the composite risk of stroke and death was higher in patients who had CEA performed in conjunction with CABG (18.7%) than in those who had CEA alone (2.1%) (15). Conversely, patients who undergo CEA before CABG also have a higher risk of perioperative complications (33,34). A meta-analysis of 56 studies regarding staged CEA and CABG published by the American Heart Association reported a composite incidence of stroke, MI, and death of 16.4% for combined carotid and coronary operations, 26.2% for CEA proceeded by CABG, and 16.4% for CABG proceeded by CEA (35). These high complication rates would clearly offset the long-term benefit from secondary stroke prevention. In this highrisk subgroup, avoiding a major operation or general anesthesia by performing CA stenting may represent a valid alternative to CEA (36). Nevertheless, controversy regarding the potential for stroke risk reduction for CA stenting preceding CABG exists (37). Congestive Heart Failure

Patients with congestive heart failure have a higher rate of perioperative stroke or death with CEA. A multicenter review of patients undergoing CEA found a perioperative stroke or death rate of 8.6% in patients

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

with congestive heart failure as opposed to 2.3% in patients without the same condition (14,15). Anatomical Features and Tandem Lesions

Anatomical variations may increase the technical difficulty of CEA and adversely influence the results. A high carotid bifurcation near the skull base, especially in a patient with a short or thick neck or a long CA stenosis that extends to the skull base, can be difficult to expose surgically. Surgical dissection of the CA in these cases can be very difficult and often extremely traumatic. Low lesions can also be technically difficult and should be avoided. The presence of tandem lesions in which the distal lesion was more severe than the proximal lesion was a NASCET exclusion criterion (6). Among symptomatic patients with ipsilateral carotid siphon stenosis, the risk of postoperative stroke or death associated with CEA in a multicenter review of 1160 procedures was 13.9% versus 7.9% in patients without distal stenosis (14). In a systemic review of 36 studies, an increased risk for perioperative stroke or death was associated with CEA in patients with stenosis of the ipsilateral siphon (16). Ipsilateral Intraluminal Thrombus

In a multicenter review of 1160 procedures, the risk of postoperative stroke or death with CEA was found to be 17.9% in symptomatic patients with intraluminal thrombus in the ipsilateral CA versus 8.1% in those without thrombus (14). In a subgroup analysis of 53 patients enrolled in the NASCET who had intraluminal clot superimposed on atherosclerotic plaque identified by angiographic procedures, the 30-day risk of stroke was 10.7% in those randomly assigned to receive medical treatment and 12% in those who underwent CEA (38). The high morbidity rate in this subgroup is related to the presence of fresh clot and the substantial risk of emboli dislodgment during surgical dissection of the CA. Contralateral Carotid Occlusion

Patients with recent symptoms referable to severe CA stenosis and coexistent contralateral CA occlusion have a high risk of ipsilateral ischemic stroke. The risk of ipsilateral stroke in medically treated patients with severe stenosis of the symptomatic CA and occlusion of the contralateral CA was 69.4% at two years in a NASCET subgroup analysis (39). Although CEA led to a significant reduction in stroke risk in this group, the perioperative risk of stroke or death in the presence of contralateral CA occlusion was as high as 14.3%. This increased risk may be related to the use of CA shunting during CEA for patients with contralateral occlusions in up to 83% of (39). Postendarterectomy Restenosis

Recurrent CA stenosis is a potential problem after CEA (40). Technically, a repeat operation is more challenging than the initial procedure because of

313

scarring around the arteries, friability of the recurrent plaque, and the necessity for complex anastomosis techniques. Among 82 patients undergoing operations for recurrent carotid stenosis at one institution, the composite rate of major morbidity and mortality was 10.8%, a rate that was five times the risk associated with primary CEA at the same institution (40). Investigators at another institution found an increased risk of cerebral ischemic events associated with CEA for recurrent stenosis (41). The 30-day rates of perioperative stroke and transient ischemic attack (TIA) were 4.8% and 4%, respectively, in the reoperation group as compared with 0.8% and 1%, respectively, in the primary endarterectomy group. The investigators also found a high rate (17%) of cranial nerve palsy with reoperation. Radiation-induced Carotid Stenosis

Accelerated radiation-induced carotid stenosis presents an increased risk for perioperative complications, primarily because of the technical pitfalls associated with a surgical approach. The presence of a long lesion, lack of well-defined dissection planes, and scarring around the vessels make the surgery more difficult (42,43), exposing these patients to a higher risk of wound infections and cranial nerve palsies. Additionally, restenosis occurs more frequently after CEA in patients with radiation-induced atheromatous disease (44,45). Specific Considerations

Recurrent nerve palsy is a risk of CEA. For a patient who already has a contralateral palsy, bilateral palsy would result in the need for tracheostomy. Additionally, those who rely significantly on their voice (such as actors, speakers, and singers) are better served by CA stenting considering the lower risk of cranial nerve injury associated with the endovascular procedure (46).

Early Trials The Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS), the first randomized comparison of endovascular versus surgical treatment in patients with carotid stenosis, was started in 1992 and completed in 1997 (47). This study was designed to compare balloon angioplasty, with or without stenting (stents developed for the CA were not introduced until after the study had begun), to CEA. As in the major CEA trials, patients considered unsuitable for surgery because of medical or surgical risk factors were excluded from enrollment. A total of 504 patients from 24 centers in Europe, Australia, and Canada were randomized, 253 to CEA and 251 to CA stenting. Endovascular treatment was technically successful (balloon inflated across the stenosis at least once or stent successfully used) in 213 of 240 treated patients. Balloon angioplasty was performed in 158 of 213 patients (74%). Only 55 (26%) patients received a stent—either the Wallstent (Schneider, Minneapolis, Minnesota, U.S.), Palmaz stent (Johnson & Johnson, Interventional Systems, Warren, New Jersey, U.S.), or the strecker

314

Sauvageau et al.

(Meditech/Boston Scientific, Natick Massachusetts, U.S.) stent. No embolic protection devices were used. The results were essentially equivalent between the stenting and endarterectomy groups at 30 days with respect to the postprocedure rate of disabling stroke or death (6.4% vs. 5.9%, respectively). No significant difference was found in the ipsilateral stroke rate with survival analysis up to three years after randomization. The rate of severe (70–99%) restenosis documented by ultrasound imaging at one year after treatment was 14% in the endovascular group and 4% in the surgical group. Patients in the surgical group were found to have a higher incidence of cranial nerve palsy and major neck hematoma. The Wallstent trial was a multicenter equivalency trial of CEA and CA stenting in which 209 symptomatic patients with 60% to 99% stenosis were enrolled; 107 patients were assigned to CA stenting and 112 to CEA (48,49). The 30-day periprocedural complication rate (any stroke or death) occurred in 12.1% of the CA stenting group and 4.5% of the CEA group. The primary endpoint of ipsilateral stroke, procedure-related death, or vascular death at one year was reached by 12.1% of patients randomized to stenting and 3.6% of those randomized to endarterectomy. The major stroke rates were 3.7% for stenting and 0.9% for endarterectomy. The study was terminated before completion because of the inferiority of CA stenting. Another randomized clinical trial featuring the Wallstent compared CA angioplasty and stenting (CAS) with CEA among patients with symptomatic severe (>70%) ICA stenosis (50). The main outcome measures included death or stroke (disabling or nondisabling) within 30 days. Twenty-three patients were randomized to CEA with patch grafting or CA angioplasty with stenting; however, only 17 patients underwent the allocated treatment before the study was suspended because of an unacceptably high morbidity rate in the stenting group. Thirty days after the treatment, none of the 10 patients who underwent CEA had a periprocedural TIA or stroke, whereas five of seven patients who underwent CA angioplasty with stenting had a periprocedural TIA or nondisabling stroke and three had disabling stroke. The Wallstent trials showed that CA stenting without embolic protection was not acceptable as an alternative to CEA for the majority of patients with symptomatic CA disease.

Embolic Protection Devices Cerebral embolization of friable atheromatous material from the aortic arch and CA has been found to occur during all stages of the CA stenting procedure and may cause periprocedural neurological deficits (51–54). Three types of embolic protection devices have been developed: those that arrest antegrade ICA flow, those that reverse ICA flow, and filters for distal embolic protection (DEP). In the United States, only four such devices have received FDA approval for carotid use and each is a distal filter: the Accunet (with the Acculink stent), the EmboShield (with the Xact stent, Abbott Vascular, Redwood City, California, U.S.), the Spider (ev3, Plymouth, Minnesota, U.S.), and

the EZ filter wire (Boston Scientific, Natick, Massachusetts, U.S.). Several non-approved, commercially made filters exist. Nonapproved flow-reversal devices include the Parodi Anti-Embolic System (Gore Neuro Protection System, W.L. Gore & Associates, Flagstaff, Arizona, U.S.). Nonapproved flow-arrest devices include the PercuSurge Guardwire (Medtronic, Minneapolis, Minnesota, U.S.), TriActiv System (Kensey Nash, Exton, Pennsylvania, U.S.) balloon occlusion catheters, and the MOMA (Invatec, Brescia, Italy) proximal occlusion device. A clear indication for proximal versus distal protection has not been established. Logically, intraluminal thrombus, soft plaque, and poor distal landing zone (tortuous poststenotic vessel) would be indications for proximal protection. The results of the European Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) prospective registry showed that gray-scale median (GSM) scores of 25 or less (representing echogenic plaque) are associated with higher embolic potential (55). Out of 155 patients, 11 patients (7.1%) with preprocedural GSM scores of 25 or less had strokes after stenting versus 4 of 263 (1.5%) patients with GSM greater than 25 (p ¼ 0.005). The authors, therefore, then validated the use of DEP in patients with GSM greater than 25 (p ¼ 0.01). However, for patients with GSM of 25 or less stenting with proximal embolic protection devices or CEA may prove safer.

Trials of Angioplasty and Stent Placement with DEP Vs. CEA in High-Risk Patients EVA-3S

Endarterectomy versus Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA-3S) is a French multicenter, noninferiority randomized trial that was designed to compare the efficacy of CA angioplasty and stent placement with or without embolic protection against CEA for secondary prevention of ischemic stroke (56). Enrollment in the study group in which CA angioplasty and stent placement were performed without protection devices was halted because unprotected treatment was associated with an excess 30-day stroke or death rate. Starting in January 2003 (57), patients presenting within four months of ischemic cerebral or retinal stroke with ipsilateral carotid stenosis of 60% or more (according to NASCET criteria) (6) were randomized into either the protected CA angioplasty and stent placement group or the CEA group. Primary endpoints included any death or recurrent stroke within 30 days and at two to four years. Secondary outcomes included MI, TIA, cranial neuropathy, functional status at the end of the study, and the degree of restenosis in treated vessels. The incidence of stroke or death at 30 days was 3.9% after CEA versus 9.6% after stenting; at six months, it was 6.1% and 11.7%, respectively (46). Cranial nerve injury was more common after CEA than stenting. There were more major local complications after stenting and more systemic complications (mainly pulmonary) after endarterectomy, but the differences were not significant.

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

EVA-3S was designed to test noninferiority of CA stenting, but technical shortcomings limit the interpretation of this study. Patients treated without DEP had a 25% 30-day stroke or death rate versus 7.9% in those treated with DEP. Comparing this 7.9% stroke or death rate with the rate in the CEA group (3.9%), the relative risk becomes 2.05 (95% confidence interval: 0.97–4.36), which is not statistically significant (SAS software, SAS Institute, Inc., Cary, North Carolina, U.S.; Chi square test, values not adjusted for multiple testing). Surgeons were fully trained and completed 25 endarterectomies in the year before enrollment. However, interventionists were certified after performing as few as five carotid stent procedures (5 carotid stents among at least 35 stent procedures of supra-aortic vessels or 12 carotid stents) or were allowed to enroll patients in the trial while they were receiving their training in carotid stenting. The 12.3% stroke or death rate among endovascular physicians tutored in CA stenting during the trial, the overall 9.6% associated with CA stenting with or without DEP, and 7.9% with DEP are higher than those of other recent trials or registries (58–63). The Carotid Revascularization Endarterectomy versus Stent Trial (CREST) (see below) requires more extensive credentialing and has shown a 4.6% 30-day stroke or death rate during the lead-in phase (64). Therefore, we must carefully consider the findings of EVA-3S and hope that future trials require stenting with DEP by more experienced interventionists. CaRESS

Carotid Revascularization using Endarterectomy or Stenting Systems (CaRESS) was a multicenter, nonrandomized, prospective study comparing CA stenting with DEP and CEA (59,65). Importantly, the choice of the procedure was left up to the treating physician. In this way, the CaRESS study may represent a more ‘‘real world’’ perspective on carotid intervention. Symptomatic patients with more than 50% stenosis and asymptomatic patients with more than 75% stenosis were considered for treatment. The primary endpoint was all-cause mortality at 30 days and 1 year. The results of this trial are summarized in Table 2. Overall, in the ‘‘real world’’ setting of the CaRESS study, CA stenting exhibited a trend toward lower morbidity and mortality than CEA but appeared slightly less durable at 30 days and 1 year. Importantly, morbidity and mortality approached the range of

315

ACAS (11) and NASCET (6,9). In reviewing the demographics of the CaRESS study, the only statistically significant difference was the inclusion of more patients who had undergone previous CEA and CA stenting in the stenting cohort. Of note, no statistically significant difference was found between CaRESS cohorts with respect to many of the high-risk criteria in other studies (including contralateral stenosis, coronary artery disease, and congestive heart failure). The lack of statistical significance in the primary endpoint suggests that the treating physicians were able to triage these high-risk groups successfully. CREST

The CREST is a randomized trial that compares the efficacy of CEA with that of CA angioplasty and stent placement performed with the aid of an embolic protection device in the prevention of stroke, MI, and death in symptomatic patients (TIA or ipsilateral nondisabling stroke within the previous 180 days) with more than 50% CA stenosis and asymptomatic patients with more than 70% stenosis. The primary endpoints are death, stroke, or MI at 30 days or ipsilateral stroke within 60 days of the procedure. The trial has multiple participating centers in North America, with the goal of enrolling 2500 patients. A credentialing phase for interventionists was included that required previous carotid stenting experience and monitoring of the performance of up to 20 procedures using the Acculink stent and Acculink embolic protection system (64). During the lead-in phase, major adverse event rates were 5.7% for symptomatic patients and 3.5% for asymptomatic patients. The 30-day composite rate of stroke and death for symptomatic patients was slightly lower than the rates reported in NASCET and ECST (61). For asymptomatic patients, 30-day stroke and death rates have been slightly higher than those reported in ACAS (61) but slightly lower than those reported in the ACST (12). Similar periprocedural morbidity was observed in women and men (66) and for those treated with or without an embolic protection device (67). For octogenarians (68,69), the 30-day stroke and death rate was 11.9%, which was significantly higher than for patients aged 79 years and younger. The study is still in the enrollment phase, but promises to provide direct evidence for the role of CA angioplasty and stent placement in the community at large. SAPPHIRE

Table 2 Summary of CaRESS Results

Death or stroke at 30 days Death or stroke at 1 yr Death, stroke, or MI at 30 days Death, stroke, or MI at 1 yr Restenosis at 1 yr

CEA (%)

CA stenting (%)

p value

3.6 13.6 4.4 14.3 3.6

2.1 10 2.1 10.9 6.3

NS NS NS NS NS

Abbreviations: CaRESS, carotid revascularization using endarterectomy or stenting systems; CEA, carotid endarterectomy; CA, carotid artery; MI, myocardial infarction; NS, not statistically significant.

The aims of the Stenting and Angioplasty with Protection in Patients at High Risk for Endarterectomy (SAPPHIRE) randomized trial were to compare CA stenting with CEA and to demonstrate statistical ‘‘noninferiority’’ of stenting to CEA (63). The study population consisted of high-risk patients with greater than or equal to 50% symptomatic stenosis and greater than or equal to 80% asymptomatic stenosis. For the endovascular group, the Smart or Precise stent (Cordis Corp., Miami Lakes, Florida, U.S.) and the Angioguard or Angioguard XP (Cordis Corp.) DEP device were used. During the study period, 747 patients were

316

Sauvageau et al.

enrolled out of which 344 underwent randomization. Primary endpoints included a composite of death/ stroke/MI within 30 days and death or ipsilateral stroke between 31 days and 1 year. Secondary endpoints included target vessel revascularization at one year, cranial nerve palsy, and complications at the surgical site or the vascular access site. SAPPHIRE had a broad endpoint by comparison with previous CA intervention trials, in particular with NASCET and ACAS in which MI and death after 30 days were not primary endpoints. Similar to coronary intervention studies, MI was included in the primary composite endpoint of SAPPHIRE and the secondary endpoint of CaRESS and is a component of the primary endpoint in CREST. Moreover, perioperative non-Q wave MI after peripheral vascular operations has been associated with a sixfold increase in mortality and a 27-fold increase in recurrent MI during the six months after the operation (70). The results of the SAPPHIRE trial are summarized in Table 3. At one year, 12.2% of patients undergoing CA stenting had reached the primary endpoint versus 20.1% of the CEA group (p value for superiority: 0.053; p value for lack of inferiority: 0.004). Target vessel revascularization occurred in 4.3% of the CEA group versus 0.6% of the CA stenting group (p ¼ 0.04). Considering secondary endpoints at one year, CA stenting was superior to CEA with respect to MI (2.5% stenting vs. 8.1% CEA; p ¼ 0.03) and major ipsilateral stroke (0% stenting vs. 3.5% CEA; p ¼ 0.02). Superiority was an unexpected finding, and one that was not necessary for the trial to succeed in its goal of providing data for regulatory approval of CA stenting in high-risk patients. Moreover, an analysis of the trial outcome that excludes MI confers noninferiority of stenting compared with CEA and does not change the results of this trial. Preliminary three-year follow-up data for SAPPHIRE has been presented (71). At three years, the overall major adverse event rate (30.3% CEA, 25.5% stenting; p ¼ 0.20) and incidence of death (24.2% CEA, 20.0% stenting; p ¼ 0.280), ipsilateral stroke (7.1% CEA vs. 6.7% stenting; p ¼ 0.945), and target lesion revascularization (7.1% CEA vs. 3.0% stenting; p ¼ 0.084), all favor CA stenting over CEA but not to statistical

significance (Table 3). The absolute percentage of stroke (all strokes within 30 days and major ipsilateral strokes from 31 to 1080 days) was calculated as follows: for all randomized patients, 3.6% CEA versus 3.5% stenting; for randomized symptomatic patients, 3.2% CEA versus 5.0% stenting; and for randomized asymptomatic patients, 3.8% CEA versus 2.9% stenting. With respect to stroke morbidity, these data suggest that asymptomatic patients are slightly better served by CA stenting and symptomatic patients by CEA. At three years, the end result is that among the high-risk patients studied and with the endpoints chosen, CA stenting was not inferior to CEA in MI, stroke, and target lesion revascularization. ICSS (CAVATAS-2)

The finding of higher rates of restenosis in the CA angioplasty group of CAVATAS resulted in the initiation of a second prospective, randomized trial—the International Carotid Stenting Study (ICSS), also known as CAVATAS-2—to compare the risks and benefits of primary CA stent placement with those of conventional CEA in patients at high risk for stroke (72). According to the study protocol, use of a cerebral protection device is recommended whenever the operator thinks that the device can be safely deployed. As of September 2006, 1024 of the planned 1500 to 2000 patients from 47 centers have been randomized to participate in the trial. SPACE

To compare the safety and prophylactic efficacy of CEA with CA angioplasty and stent placement against stroke in patients with symptomatic CA stenosis, the German Ministry of Science sponsored the StentSupported Percutaneous Angioplasty of the Carotid Artery versus Endarterectomy (SPACE) trial, a prospective, randomized, multicenter study (73). Eligibility for this study was extended to patients with severe CA stenosis [70% by duplex ultrasonography, 50% by NASCET criteria (6), or 70% by ECST criteria (18)] who had experienced amaurosis fugax, TIA, or mild stroke within 180 days of randomization. A total of

Table 3 Summary of SAPPHIRE Results 1 yr Death, stroke, or MI within 30 days and death or ipsilateral stroke between 31 days and 1 yr Target vessel revascularization within 1 yr 3 yr Overall major adverse event rate Death Ipsilateral stroke Target lesion revascularization

CEA (%)

CA stenting (%)

20.1

12.2

4.3

0.6

30.3 24.2 7.1 7.1

25.3 20.0 6.3 3.0

p value 0.053 for superiority, 0.004 for lack of inferiority 0.04 0.20 0.280 0.945 0.084

Absolute percentage of stroke (all strokes to 30 days and major ipsilateral strokes from 31–1080 days) All randomized patients 3.6 3.5 Randomized asymptomatic patients 3.2 5.0 Randomized symptomatic patients 3.8 2.9 Abbreviations: SAPPHIRE, stenting and angioplasty with protection in patients at high risk for endarterectomy; CEA, carotid endarterectomy; CA, carotid artery; MI, myocardial infarction.

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

1200 patients were randomized to CA stenting (n ¼ 605) or CEA (n ¼ 595). Primary outcome measures included 30-day incidence of ipsilateral cerebrovascular events or death. A total of 1183 patients were included in the 30-day results analysis (62). The rate of death or ipsilateral ischemic stroke from the time of randomization up to 30 days after the procedure was 6.84% with CA stenting and 6.34% with CEA. SPACE failed to prove noninferiority of CA stenting compared with CEA for the periprocedural complication rate. The failure of this trial to show non-inferiority may have resulted from an underpowered sample and higher than expected event rates in both groups of patients evaluated. Results at 6 to 24 months are pending.

Trials of Angioplasty and Stent Placement with DEP Vs. CEA in Low-Risk Patients The results encountered in the high-risk population have led to the evaluation of CA stenting as a revascularization alternative for low-risk patients. Three prospective, randomized trials in low-risk patients are under way: the Asymptomatic Carotid Stenosis, Stenting versus Endarterectomy Trial (ACT I), the Asymptomatic Carotid Surgery Trial-2 (ACST-2), and the Transatlantic Asymptomatic Carotid Interventional Trial (TACIT). ACT I is currently enrolling low surgical risk patients with asymptomatic CA stenosis (a single ICA lesion with 80% but 99% stenosis) at multiple centers in North America. The devices used in this trial are the EmboShield DEP device and the Xact stent (Abbott Vascular). The randomization scheme is 3:1 for CA stenting to CEA. The ACST-2 is also randomizing asymptomatic patients with severe CA stenosis to CEA versus stenting (74). The primary analysis will include clinical MI, stroke, and death within 30 days of either treatment, and chances of long-term (five years) stroke-free survival. The investigators will use Conformite´ Europe´ennemarked devices, usually with cerebral protection. TACIT will study all-risk patients with asymptomatic CA stenosis, assigning these patients to one of two treatment groups (74,75). The first group will be optimal medical therapy alone, consisting of antiplatelet, antilipidemic, and antihypertensive therapy, as well as strict serum glucose control and tobacco cessation efforts. The second group will provide optimal medical therapy plus CA stenting using embolic protection (with commercially available devices). The aim of the TACIT investigators is to enroll approximately 2400 patients, equally divided between the two treatment groups. The primary endpoint is a composite of 30-day mortality, all strokes within the five-year study period, and a neurocognitive component (neurocognitive decline, which is defined as vascular dementia or vascular depression) measured using predominantly depression scales. Secondary endpoints will include a detailed quality-of-life and cost-effective analysis as well as plaque characterization (detailed core lab assessments of ultrasound plaque features). As mentioned, CREST is ongoing for both highrisk and low-risk patients to compare the efficacy of

317

CA stenting and CEA in preventing stroke, MI, or death in both symptomatic and asymptomatic patients with carotid stenosis. If CA stenting is found equal to or superior to CEA or to rigorous medical therapy in the low-risk patients enrolled or scheduled for enrollment in the aforementioned trials, the results may lead to broader application of stenting for carotid intervention.

CA STENTING PROCEDURE Procedural Overview The CA stenting procedure is an evolving procedure that has been modified according to operator experience and device development. The usual sequence of a procedure performed in conjunction with the use of a filter device for embolic protection is outlined below. The procedure is performed in an angiography suite with biplane digital subtraction and fluoroscopic imaging capabilities. The patient is sedated but arousable for neurological assessment. A Foley catheter and two peripheral IV lines are inserted. Blood pressure, oxygen saturation, and cardiac rhythm are monitored during the procedure. The CA is generally approached percutaneously from the common femoral artery. The interventionist should also be familiar with radial and brachial approaches in case femoral artery access is not possible. An aortic arch angiogram is initially performed to define the atherosclerotic burden as well as the anatomical configuration of the great vessels, which allows the operator to predict the feasibility of carotid cannulation and select the devices needed for the procedure. Selective carotid angiography is then performed, and the severity of the stenosis is defined. The diameters of the common carotid artery (CCA) and ICA are measured with attention paid to determining a landing zone for the protection device. Intracranial angiography is also essential prior to intervention because the presence of tandem lesions should be considered in the management strategy. The processes of angioplasty and stenting create intimal injury that promotes thrombosis (76). Therefore, patient preparation with adequate antiplatelet and anticoagulation therapy is essential. Patients receive a dual antiplatelet regimen consisting of aspirin (325 mg daily) and a thienopyridine derivative (i.e., clopidogrel, 75 mg daily; or ticlopidine, 250 mg twice daily) for at least three days before stent treatment. A loading dose of clopidogrel (300–600 mg) administered early on the day of the procedure is an alternative for patients who are already taking aspirin. An intravenous bolus dose of heparin (50– 60 U/kg) is administered after catheterization of the CCA. An activated coagulation time of 250 to 300 seconds is maintained throughout the procedure. The heparin infusion is usually discontinued at the conclusion of the procedure. Bradycardia occurs occasionally during angioplasty. Atropine and vasopressors should be readily available should significant bradycardia and hypotension develop. Continuous intraprocedural monitoring

318

Sauvageau et al.

of heart rate, blood pressure, and neurological status is essential. After completion of the diagnostic angiogram and positioning of the catheter in the CCA, road mapping of the cervical CA is performed. An exchange-length 0.035-inch wire is positioned in the external CA. The diagnostic catheter is exchanged over the wire for a 90-cm, 6- to 10-French (Fr) sheath that is then advanced into the CCA below the bifurcation. For patients who have undergone complete diagnostic cerebral angiography before the stenting procedure, a combination of a 6-Fr, 90-cm shuttle over a 6.5-Fr head-hunter 125-cm slip-catheter (Cook, Inc., Bloomington, Indiana, U.S.) or a 5-Fr 125-cm Vitek catheter (Cook) can be used. In these cases, the shuttle is introduced primarily in the femoral artery over a 0.35-inch wire and is parked in the descending aorta. The inner obturator and wire are removed. The 125-cm catheter is then advanced into the shuttle and the target vessel is catheterized. The shuttle is brought over the wire and the catheter in the CCA. The size of the shuttle is usually dictated by the embolic protection device profile and compatibility with the stent system. An optimal angiographic view that maximizes the opening of the bifurcation and facilitates crossing of the stenosis should be sought. The lesion is crossed with the protection device. Predilation of the stenotic vessel segment is performed at the operator’s discretion. A 3- to 4-mm coaxial angioplasty balloon is advanced to the lesion over the 0.014-inch wire holding the protection device. On rare occasions, predilation needs to be performed prior to the introduction of an embolic protection device. In such cases, the balloon system is then exchanged for a stent system. The diameter of the stent should be sized to the caliber of the largest segment of the CA to be covered (usually 1 to 2 mm more than the normal caliber of the CCA). Oversizing of the stent in the ICA does not usually result in adverse events, but a tapered stent can better conform to the vessel wall. Particular attention should be paid to the selection of a stent that is long enough to cover the entire lesion. After removing the stent system, poststent dilation should be performed using a balloon with a diameter matching that of the ICA distal to the stent. A coaxial balloon is usually preferred for this purpose. The embolic protection device is then removed, using its retrieval catheter. When a balloon occlusion catheter is used for cerebral protection, the embolic debris is aspirated before deflation and retrieval of the balloon. The most common settings for difficulty in capturing deployed filter protection devices are with an open-celled stent on a significant vessel curve (in which a stent strut may impinge on the vessel intima) and when the device is parked in a tortuous distal vessel. A systematic approach will generally lead to successful recapture of the device. Advancing the guide catheter into the stent will bias the wire away from the stent wall, allowing the recaptured sheath to pass. Having the patient inhale deeply or turn his or her head opposite to the direction of the vessel curve can help straighten the curve or elongate the artery

enough for passage of the sheath. More aggressively, pressing on the stent in the patient’s neck will also change the bias of the wire. If the sheath is impeded by a stent strut, redilatation with a larger balloon or spinning the sheath with forward pressure will help flatten the strut or allow passage for the sheath. If other maneuvers fail, a 4- or 5-Fr-angled glide catheter can be passed over the DEP wire to capture the filter. A device may be used for closure of the access site (e.g., Starclose, Abbott Vascular; Perclose, Abbott Vascular; or Angio-Seal, St. Jude Medical, Minnetonka, Minnesota, U.S.) on the basis of operator preference, patient anatomy, and puncture site location.

Postintervention Follow-Up Good hydration should be maintained after the procedure. Hypotension and hypertension should be avoided. Particular attention must be paid in cases of severe stenosis and contralateral occlusion to prevent reperfusion hemorrhage. If a closure device has not been used, the arterial sheath should be removed when the activated coagulation time is less than 150 seconds. The patient is usually discharged the following morning. Patients require surveillance imaging to evaluate vessel patency. Duplex sonography evaluation should be obtained before discharge, at six months, one year, and then annually thereafter. The dual antiplatelet regimen of aspirin and clopidogrel is maintained for four to six weeks postprocedure, after which patients remain on aspirin therapy.

Markers of High-Risk CA Stenting Unfavorable anatomical and lesion characteristics are factors that may impact the risk of CA stenting. Stenotic or occluded iliac arteries or abdominal aorta are among the unfavorable access characteristics. A difficult arch (type 2 or 3, calcification or bovine configuration) can be challenging for a less-experienced operator. Occlusion of the external CA or stenosis involving the CCA may increase the risk in positioning the guiding catheter or shuttle. Tortuosity of the ICA, severity of the lesion, degree of calcification, and presence of intraluminal thrombus represent other adverse factors. Low plaque echolucency (GSM) is also an adverse prognosticator (55,77).

Durability of CA Stenting Durability of carotid revascularization with CA stenting is a concern frequently expressed by the surgical community. In a retrospective study of patients undergoing stenting for de novo (119 arteries) and postendarterectomy (76 arteries) carotid stenosis, 80% or more stenosis was detected by follow-up Doppler imaging in 5.2% of the vessels stented (78). Restenosis after endarterectomy was the major risk for in-stent restenosis. Significant (symptomatic or 80%) recurrent stenosis was detected by follow-up Doppler imaging in 6 (5%) of 112 patients in our CA stenting series (79).

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

COMPLICATION OCCURRENCE AND AVOIDANCE From puncture of the femoral artery to retrieval of the protection device and performance of the final angiogram, potential exists for complications during CA stenting that can be threatening to life, limb, or brain. Delayed neurological, cardiac, and peripheral complications can also occur and may require immediate intervention for meaningful salvage. Knowledge of these complications is essential to ensure quick recognition and effective management. Patient selection is the most important factor in minimizing complications associated with CA stenting. The experience of the interventionist and staff is the second most important factor. It is essential that all personnel are familiar with all the equipment, devices, pharmacological agents, critical care management, and the disease process treated. Many risks associated with CA stenting can be mitigated before and during the procedure. Guide catheters are flushed frequently with normal saline (0.9% NaCl) with 5000 units of heparin in each pressure bag. The air should be actively removed from each flush bag.

Access Site Complications The rate of local complications occurring during diagnostic cerebral angiography can be as high as 5% (80). The most common access problems include retrograde dissection, pseudoaneurysm, arteriovenous fistula, extravasation of blood around the sheath, tortuosity, and the inability to gain access. Femoral artery occlusion and bleeding can also occur in an acute or delayed fashion. The external CA and its branches are used to support a guidewire during the exchange of a diagnostic catheter for a guide sheath or catheter. Vessel perforation in this setting has been described by our group (81). To prevent this complication, large branches of the external CA, preferably the internal maxillary artery or occipital artery, should be used for exchange maneuvers. Carotid dissection is another complication that can be encountered during the stenting procedure. For cases of small, asymptomatic and non-flow-limiting dissections, clinical observation is recommended. Stenting is warranted if the dissection is symptomatic or flow limiting. Spasm is frequently encountered when the DEP device or the guide catheter straightens or moves a kink in the CA. This can be ignored, as it will resolve with device retrieval and often resolves after stenting and postdilatation angioplasty.

Neurological Complications Stroke can occur at any point after femoral artery access has been obtained. If a patient develops a sudden neurological change, the differential diagnosis entertained should include hemorrhage and ischemia, most often due to embolism. On the basis of findings

319

of the neurological examination, rapid access should be gained to the vessel suspected of harboring the problem. If the patient’s airway is compromised, intubation should be performed. If no vessel cutoff or slow flow is appreciated, hemorrhage must be ruled out and the patient should undergo a CT scan of the head. On the basis of transcranial Doppler ultrasound data in which protected stenting with the PercuSurge device was compared to unprotected stenting, the highest-risk maneuvers for embolism in conjuction with risk in unprotected stenting, from lowest to highest risk, were predilation angioplasty, stenting, and postdilation angioplasty (51). The intracranial complications of CA stenting can be grouped into large vessel occlusion, shower of emboli, and hemorrhage. If a clear large vessel cutoff can be seen, an immediate attempt should be undertaken to recanalize the occluded vessel. If an angiogram documents slow flow and the CT scan is negative for hemorrhage, IIb/IIIa antiplatelet agents are administered. If a hemorrhage is identified, heparin anticoagulation is reversed with protamine, the blood pressure tightly controlled, and a repeat CT scan is obtained within 6 to 12 hours. Life-threatening hematomas in neurologically salvageable patients can be evacuated.

Systemic Complications Systemic complications may also occur following CA stenting. These include seizures, MI, contrast material nephropathy, and contrast material allergy. The interventionist, treating institution, and ancillary staff should be versed in the management of all four of these conditions as in most patients, seizures, MI, contrast nephropathy, and allergic reaction can be readily treated.

ILLUSTRATIVE CASES Case 1 (Tortuous Anatomy) A 68-year-old woman with a remote history of two left hemispheric strokes was found to have severe left ICA origin stenosis on noninvasive studies. Her past medical history also included diabetes, hypertension, morbid obesity, and inactive congestive heart failure. She had a mild residual right hemiparesis and no additional ischemic symptoms since her strokes three years earlier. Angiography confirmed 80% stenosis of the left ICA origin. Because she had a short, immobile neck and a high carotid bifurcation at the level of the C2 vertebral body, the patient was selected for CA stenting. The tortuosity of the ICA prohibited safe advancement of a DEP device (Fig. 1A, B). Carotid stenting with angioplasty was performed successfully using proximal protection with a flow-reversal system (Gore Neuro Protection System) (Fig. 1C). Care was taken not to position the stent in the tortuous segment of the ICA (Fig. 1D). Occlusion time was approximately 12 minutes, and the patient tolerated the procedure well. She was discharged the next day at her baseline neurological condition.

320

Sauvageau et al.

Figure 1 (A) AP and (B) lateral projections of left CCA injection demonstrating 80% stenosis of the ICA origin. Note the sharp turn in the ICA distal to the stenosis. Such tortuosity is a relative contraindication to DEP devices. (C) AP view showing the establishment of flow reversal after occlusion in the external CA and distal CCA. An arteriovenous conduit had been created between the left ICA and the left common femoral vein prior to flow reversal. (D) AP view shows good positioning of the stent. The stent does not involve the tortuous segment of the ICA. Abbreviations: CCA, common carotid artery; ICA, internal carotid artery; DEP, distal embolic protection.

Case 2 (Intraoperative Plaque Rupture)

CONCLUSIONS

A 62-year-old man with a history of left frontal stroke resulting in aphasia and left hemiparesis was found to have bilateral ICA origin stenosis. He recovered well from his stroke and underwent successful left carotid angioplasty with stenting. He returned for endovascular treatment of the asymptomatic 90% right ICA stenosis (Fig. 2A, B). DEP was obtained with a 5-mm EmboShield. Angioplasty was done before and after placement of an Xact stent (6  30 mm to 8  30 mm). An opacity, consistent with plaque debris or thrombus, was seen just distal to the stent (Fig. 2C). Intravascular ultrasonography confirmed that this was a plaque fragment, and a second Xact stent (7  20 mm) was placed (Fig. 2D, E). The lesion was now covered completely by the stents, and the embolic protection device contained visible debris. The patient remained unchanged neurologically and was discharged the next day.

In 2007, the CAS procedure is considered ‘‘not inferior’’ to CEA for the treatment of high-risk patients with symptomatic and asymptomatic CA disease. CAS and CEA are complementary procedures. The availability of both approaches at a single center can certainly optimize patient care (82). Evaluation of CA stenting in the low-risk population is ongoing. The results of such studies may have a significant influence on CEA, considering that in the case of clinical equivalence between CA stenting and CEA, patients are likely to favor the less-invasive endovascular approach. In terms of durability, stenting is slightly inferior to endarterectomy (as shown in studies mentioned in this chapter), but this will likely not limit its widespread use if equipoise between the two therapies is proved. With future technological developments including improved embolic protection systems and refined stents with smaller delivery

Figure 2 (A) AP and (B) lateral projections of right CCA injection show 90% stenosis of ICA distal to its origin. (C) Lateral view shows resolution of the stenosis after stenting and angioplasty. The embolic protection device is positioned in the distal cervical ICA. However, opacity (arrow) is seen distal to the stent. Intravascular ultrasonography confirmed that this was a fragment of plaque. (D) AP and (E) lateral views show a good angiographic result after placement of an overlapping stent to cover this fragment. Abbreviations: CCA, common carotid artery; ICA, internal carotid artery.

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

platforms, CA stenting may become the new gold standard for carotid revascularization.

13.

DISCLOSURES L. N. Hopkins: Research Grants—Boston Scientific, Cordis, Micrus; Honoraria—Bard, Boston Scientific, Cordis, Medsn; Stock or Shareholder—APW Holding Inc., Boston Scientific, EndoTex, Micrus; Consultant/ Advisory Board—Abbott, Bard, Boston Scientific, Cordis, EndoTex, Access Closure, market Rx, Micrus. E. I. Levy: Research Grants—Boston Scientific; Cordis; Other Research Support (<10K)—Wingspan devices (Boston Scientific); Honoraria—Boston Scientific, Cordis; Other—Abbott Vascular (carotid stent training), ev3 (carotid stent training), Zimmer Spine (patent royalties). The remaining authors have no financial disclosures.

14.

ACKNOWLEDGMENT

18.

The authors of this chapter thank Paul H. Dressel for preparation of the illustrations.

15.

16.

17.

19.

REFERENCES 1. Bonita R. Epidemiology of stroke. Lancet 1992; 339:342–344. 2. Matchar DB. Cost of stroke. Stroke Clin Update 2002; 5:9–12. 3. Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics–2006 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2006; 113:e85–e151. 4. Taylor TN, Davis PH, Torner JC. Projected number of strokes by subtype in the year 2050 in the United States. Stroke 1998; 29:322 (abstr). 5. Dyken ML. Stroke risk factors. In: Norris JW, Hachinski VC, eds. Prevention of Stroke. New York City: SpringerVerlag, 1991:83–102. 6. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high-grade carotid stenosis. N Engl J Med 1991; 325:445–453. 7. North American Symptomatic Carotid Endarterectomy Trial Collaborators. Methods, patient characteristics, and progress. Stroke 1991; 22:711–720. 8. MRC European Carotid Surgery Trialists. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998; 351:1379–1387. 9. Barnett HJ, Taylor DW, Eliasziw M, et al. Benefit of carotid endarterectomy in patients with symptomatic moderate or severe stenosis. North American Symptomatic Carotid Endarterectomy Trial Collaborators. N Engl J Med 1998; 339:1415–1425. 10. Asymptomatic Carotid Atherosclerosis Study Group. Study design for randomized prospective trial of carotid endarterectomy for asymptomatic atherosclerosis. Stroke 1989; 20:844–849. 11. Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995; 273:1421–1428. 12. Halliday A, Mansfield A, Marro J, et al. Prevention of disabling and fatal strokes by successful carotid endarter-

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

321

ectomy in patients without recent neurological symptoms: randomised controlled trial. Lancet 2004; 363:1491–1502. Hobson RW II, Weiss DG, Fields WS, et al. Efficacy of carotid endarterectomy for asymptomatic carotid stenosis. For the Veterans Affairs Cooperative Study Group. N Engl J Med 1993; 328:221–227. Goldstein LB, McCrory DC, Landsman PB, et al. Multicenter review of preoperative risk factors for carotid endarterectomy in patients with ipsilateral symptoms. Stroke 1994; 25(6):1116–1121. Goldstein LB, Samsa GP, Matchar DB, et al. Multicenter review of preoperative risk factors for endarterectomy for asymptomatic carotid artery stenosis. Stroke 1998; 29(4): 750–753. Rothwell PM, Slattery J, Warlow CP. Clinical and angiographic predictors of stroke and death from carotid endarterectomy: systematic review. BMJ 1997; 315 (7122):1571–1577. Paciaroni M, Eliasziw M, Kappelle LJ, et al. Medical complications associated with carotid endarterectomy. North American Symptomatic Carotid Endarterectomy Trial (NASCET). Stroke 1999; 30(9):1759–1763. European Carotid Surgery Trialists’ Collaborative Group. MRC European Carotid Surgery Trial: interim results for symptomatic patients with severe (70–99%) or with mild (0–29%) carotid stenosis. Lancet 1991; 337:1235–1243. European Carotid Surgery Trialists’ Collaborative Group. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998; 351 (9113):1379–1387. Hanel RA, Xavier AR, Kirmani JF, et al. Management of carotid artery stenosis: comparing endarterectomy and stenting. Curr Cardiol Rep 2003; 5:153–159. Food and Drug Administration. P04–85, August 31, 2004; http://www.fda.gov/bbs/topics/news/2004/NEW01111. html (accessed August 28, 2006). Food and Drug Administration. Office of Device Evaluation, Center for Devices and Radiological Health. Approval Letter P040012, August 30, 2005; http://www.fda.gov/ cdrh/pdf4/p040012a.pdf (accessed August 28, 2006). Centers for Medicare & Medicaid Services. CMS Manual System: Pub. 100–04 Medicare claims processing: April 22, 2005. http://new.cms.hhs.gov/transmittals/downloads/ R531CP.pdf (accessed August 31, 2006). North American Symptomatic Carotid Endarterectomy Trial (NASCET) Investigators. Clinical alert: benefit of carotid endarterectomy for patients with high-grade stenosis of the internal carotid artery. Stroke 1991; 22:816–817. Brott TG, Labutta RJ, Kempczinski RF. Changing patterns in the practice of carotid endarterectomy in a large metropolitan area. JAMA 1986; 255(19):2609–2612. Hannan EL, Popp AJ, Tranmer B, et al. Relationship between provider volume and mortality for carotid endarterectomies in New York state. Stroke 1998; 29(11):2292– 2297. Karp HR, Flanders WD, Shipp CC, et al. Carotid endarterectomy among Medicare beneficiaries: a statewide evaluation of appropriateness and outcome. Stroke 1998; 29(1):46–52. Stukenborg GJ. Comparison of carotid endarterectomy outcomes from randomized controlled trials and Medicare administrative databases. Arch Neurol 1997; 54 (7):826–832. Wennberg DE, Lucas FL, Birkmeyer JD, et al. Variation in carotid endarterectomy mortality in the Medicare population: trial hospitals, volume, and patient characteristics. JAMA 1998; 279(16):1278–1281.

322

Sauvageau et al.

30. Ouriel K, Hertzer NR, Beven EG, et al. Preprocedural risk stratification: identifying an appropriate population for carotid stenting. J Vasc Surg 2001; 33:728–732. 31. Ferguson GG, Eliasziw M, Barr HW, et al. The North American Symptomatic Carotid Endarterectomy Trial: surgical results in 1415 patients. Stroke 1999; 30:1751–1758. 32. Faggioli GL, Curl GR, Ricotta JJ. The role of carotid screening before coronary artery bypass. J Vasc Surg 1990; 12(6):724–731. 33. del Sette M, Eliasziw M, Streifler JY, et al. Internal borderzone infarction: a marker for severe stenosis in patients with symptomatic internal carotid artery disease. The North American Symptomatic Carotid Endarterectomy (NASCET) Group. Stroke 2000; 31(3):631–636. 34. Harbaugh RE, Stieg PE, Moayeri N, et al. Carotidcoronary artery bypass graft conundrum. Neurosurgery 1998; 43(4):926–931. 35. Moore WS, Barnett HJ, Beebe HG, et al. Guidelines for carotid endarterectomy. A multidisciplinary consensus statement from the ad hoc Committee, American Heart Association. Stroke 1995; 26:188–201. 36. Lopes DK, Mericle RA, Lanzino G, et al. Stent placement for the treatment of occlusive atherosclerotic carotid artery disease in patients with concomitant coronary artery disease. J Neurosurg 2002; 96(3):490–496. 37. Randall MS, McKevitt FM, Cleveland TJ, et al. Is there any benefit from staged carotid and coronary revascularization using carotid stents? A single-center experience highlights the need for a randomized controlled trial. Stroke 2006; 37(2):435–439. 38. Villarreal J, Silva J, Eliasziw M, et al. North American Symptomatic Carotid Endarterectomy Trial. Prognosis of patients with intraluminal thrombus in the internal carotid artery. Stroke 1998; 29(1):276 (abstr 18). 39. Gasecki AP, Eliasziw M, Ferguson GG, et al. Long-term prognosis and effect of endarterectomy in patients with symptomatic severe carotid stenosis and contralateral carotid stenosis or occlusion: results from NASCET. North American Symptomatic Carotid Endarterectomy Trial (NASCET) Group. J Neurosurg 1995; 83:778–782. 40. Meyer FB, Piepgras DG, Fode NC. Surgical treatment of recurrent carotid artery stenosis. J Neurosurg 1994; 80 (5):781–787. 41. AbuRahma AF, Jennings TG, Wulu JT, et al. Redo carotid endarterectomy versus primary carotid endarterectomy. Stroke 2001; 32(12):2787–2792. 42. Loftus CM, Biller J, Hart MN, et al. Management of radiation-induced accelerated carotid atherosclerosis. Arch Neurol 1987; 44(7):711–714. 43. Melliere D, Becquemin JP, Berrahal D, et al. Management of radiation-induced occlusive arterial disease: a reassessment. J Cardiovasc Surg (Torino) 1997; 38(3):261–269. 44. Cazaban S, Maiza D, Coffin O, et al. Surgical treatment of recurrent carotid artery stenosis and carotid artery stenosis after neck irradiation: evaluation of operative risk. Ann Vasc Surg 2003; 17(4):393–400. 45. Leseche G, Castier Y, Chataigner O, et al. Carotid artery revascularization through a radiated field. J Vasc Surg 2003; 38(2):244–250. 46. Mas JL, Chatellier G, Beyssen B, et al. Endarterectomy versus stenting in patients with symptomatic severe carotid stenosis. N Engl J Med 2006; 355(16):1660–1671. 47. CAVATAS Investigators. Endovascular versus surgical treatment in patients with carotid stenosis in the Carotid and Vertebral Artery Transluminal Angioplasty Study (CAVATAS): a randomised trial. Lancet 2001; 357:1729–1737. 48. Alberts MJ. Results of a multicenter prospective randomized trial of carotid artery stenting vs. carotid endarter-

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

ectomy. The Publications Committee of WALLSTENT. Stroke 2001; 32:325 (abstr). Alberts MJ, McCann R, Smith TP, et al. A randomized trial of carotid stenting vs. endarterectomy in patients with symptomatic carotid stenosis: study design. The Schneider Wallstent Endoprosthesis Clinical Investigators. J Neurovasc Dis 1997; 2:228–234. Naylor AR, Bolia A, Abbott RJ, et al. Randomized study of carotid angioplasty and stenting versus carotid endarterectomy: a stopped trial. J Vasc Surg 1998; 28:326–334. Al-Mubarak N, Roubin GS, Vitek JJ, et al. Microembolization during carotid stenting with the distal-balloon antiemboli system. Int Angiol 2002; 21(4):344–348. Cremonesi A, Manetti R, Setacci F, et al. Protected carotid stenting: clinical advantages and complications of embolic protection devices in 442 consecutive patients. Stroke 2003; 34(8):1936–1941. Jaeger HJ, Mathias KD, Hauth E, et al. Cerebral ischemia detected with diffusion-weighted MR imaging after stent implantation in the carotid artery. AJNR Am J Neuroradiol 2002; 23(2):200–207. Theron JG, Payelle GG, Coskun O, et al. Carotid artery stenosis: treatment with protected balloon angioplasty and stent placement. Radiology 1996; 201(3):627–636. Biasi GM, Froio A, Diethrich EB, et al. Carotid plaque echolucency increases the risk of stroke in carotid stenting: the Imaging in Carotid Angioplasty and Risk of Stroke (ICAROS) study. Circulation 2004; 110:756–762. Mas JL, Chatellier G, Beyssen B. Carotid angioplasty and stenting with and without cerebral protection: clinical alert from the Endarterectomy Versus Angioplasty in Patients With Symptomatic Severe Carotid Stenosis (EVA-3S) trial. The EVA-3S Investigators. Stroke 2004; 35:e18–e21. EVA-3S Investigators. Endarterectomy vs. Angioplasty in Patients with Symptomatic Severe Carotid Stenosis (EVA3S) trial. Cerebrovasc Dis 2004; 18:62–65. Safian RD, Bresnahan JF, Jaff MR, et al. Protected carotid stenting in high-risk patients with severe carotid artery stenosis. J Am Coll Cardiol 2006; 47(12):2384–2389. CaRESS Steering Committee. Carotid Revascularization Using Endarterectomy or Stenting Systems (CaRESS) phase I clinical trial: 1-year results. J Vasc Surg 2005; 42:213–219. Gray WA, Hopkins LN, Yadav S, et al. Protected carotid stenting in high-surgical-risk patients: the ARCHeR results. J Vasc Surg 2006; 44(2):258–268. Hobson RW II, Brott TG, Roubin GS. Carotid stenting in the CREST lead-in phase: periprocedural stroke, myocardial infarction, and death rates are lower than reported for preceding stent trials. The CREST investigators. Circulation 2003; 108(suppl 4):IV–604 (abstr 2748). Ringleb PA, Allenberg J, Bruckmann H, et al. 30 day results from the SPACE trial of stent-protected angioplasty versus carotid endarterectomy in symptomatic patients: a randomised non-inferiority trial. Lancet 2006; 368(9543):1239–1247. Yadav JS, Wholey MH, Kuntz RE, et al. Protected carotidartery stenting versus endarterectomy in high-risk patients. N Engl J Med 2004; 351:1493–1501. Hobson RW II, Howard VJ, Roubin GS, et al. Credentialing of surgeons as interventionalists for carotid artery stenting: experience from the lead-in phase of CREST. J Vasc Surg 2004; 40:952–957. CaRESS Steering Committee. Carotid revascularization using endarterectomy or stenting systems (CARESS): phase I clinical trial. J Endovasc Ther 2003; 10:1021–1030. Howard VJ, Brott TG, Qureshi AI, et al. Gender and periprocedural stroke and death following carotid artery

Chapter 17: Endovascular Treatment of Extracranial Carotid Atherosclerotic Disease

67.

68.

69.

70.

71.

72.

73.

stenting: results from the CREST lead-in phase. The CREST Investigators. Stroke 2004; 35:253 (abstr P5). Roubin GS, Brott TG, Hopkins LN. Developing embolic protection for carotid stenting in the Carotid Revascularization Endarterectomy vs Stenting Trial (CREST). The CREST Investigators. Circulation 2003; 108(suppl 4):IV–687 (abstr 3124). Hobson RW II, Howard VJ, Roubin GS, et al. Carotid artery stenting is associated with increased complications in octogenarians: 30-day stroke and death rates in the CREST lead-in phase. J Vasc Surg 2004; 40:1106–1111. Howard G, Hobson RW II, Brott TG. Does the stroke risk of stenting increase at older ages? Thirty-day stroke death rates in the CREST lead-in phase. The CREST Investigators. Circulation 2003; 8(suppl 4):V–461 (abstr 2116). Kim LJ, Martinez EA, Faraday N, et al. Cardiac troponin I predicts short-term mortality in vascular surgery patients. Circulation 2002; 106:2366–2371. Yadav JS. Carotid Stenting for the Prevention of Stroke: Three-Year Follow-Up of the SAPPHIRE Trial and US Carotid Feasibility Demonstrates Durability of the Procedure [presentation]. The SAPPHIRE Investigators. Transcatheter Cardiovascular Therapeutics (TCT) Meeting, Washington DC, October 18, 2005. Featherstone RL, Brown MM, Coward LJ. International carotid stenting study: protocol for a randomised clinical trial comparing carotid stenting with endarterectomy in symptomatic carotid artery stenosis. The ICSS Investigators. Cerebrovasc Dis 2004; 18:69–74. Ringleb PA, Kunze A, Allenberg JR, et al. The StentSupported Percutaneous Angioplasty of the Carotid Artery vs. Endarterectomy Trial. Cerebrovasc Dis 2004; 18:66–68.

323

74. ACST-2 and TACIT to answer the asymptomatic carotid question. Interventional News 2006;20 http://www.cxvascular. com/InterventionalNews/InterventionalNews.cfm?ccs¼296& cs¼1874 (accessed August 31, 2006). 75. Rundback J. Update on the TACIT trial. Endovasc Today 2006; 5:93–94. 76. Lam JY, Chesebro JH, Steele PM, et al. Deep arterial injury during experimental angioplasty: relation to a positive indium-111-labeled platelet scintigram, quantitative platelet deposition and mural thrombosis. J Am Coll Cardiol 1986; 8(6):1380–1386. 77. El-Atrozy T, Nicolaides A, Tegos T, et al. The effect of Bmode ultrasonic image standardisation on the echodensity of symptomatic and asymptomatic carotid bifurcation plaques. Int Angiol 1998; 17(3):179–186. 78. Setacci C, Pula G, Baldi I, et al. Determinants of in-stent restenosis after carotid angioplasty: a case-control study. J Endovasc Ther 2003; 10(6):1031–1038. 79. Levy EI, Hanel RA, Lau T, et al. Frequency and management of recurrent stenosis after carotid artery stent implantation. J Neurosurg 2005; 102(1):29–37. 80. Pryor JC, Setton A, Nelson PK, et al. Complications of diagnostic cerebral angiography and tips on avoidance. Neuroimaging Clin N Am 1996; 6(3):751–758. 81. Ecker RD, Guidot CA, Hanel RA, et al. Perforation of external carotid artery branch arteries during endoluminal carotid revascularization procedures: consequences and management. J Invasive Cardiol 2005; 17(6):292–295. 82. Ecker RD, Lau T, Levy EI, et al. Thirty-day morbidity and mortality rates for carotid artery intervention by surgeons who perform both carotid endarterectomy and carotid artery angioplasty and stent placement. J Neurosurg 2007; 106: 217–221.

18 Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease Nabil M. Akkawi and Ajay K. Wakhloo Division of Neuroimaging and Intervention, University of Massachusetts Medical School, Worcester, Massachusetts, U.S.A.

INTRODUCTION In 1964, the father of interventional radiology Charles T. Dotter introduced the concept of transluminal angioplasty at the University of Oregon in Portland, United States, and applied it to the peripheral vascular system. At that time, he had already discussed its potential use for the renal and coronary circulation. However, his idea was ignored in the United States for nearly 15 years. The German-born cardiologist Andreas R. Gruentzig, who was introduced to angioplasty by Eberhart Zeitler, performed the first successful balloon angioplasty on human coronary artery in 1977 in Zurich, Switzerland. Since then percutaneous transluminal angioplasty (PTA) has rapidly evolved as a safe endovascular revascularization procedure for vascular occlusive diseases of the peripheral, renal, and coronary vascular system. In newer endovascular technologies, PTA is being replaced by primary stenting. Developments in microguidewire and balloon catheter technology have enabled the interventionalists to expand their experience gradually from the cervical parts of the carotid and vertebral arteries to the cerebrovascular circulation for treatment of symptomatic and medically refractory focal stenosis. Until recently, the introduction of stent technology to the anterior and posterior intracranial vascular systems has been limited by the lack of specific stents and delivery systems capable of safe, easy, and effective navigation through the neurovascular system. The advent of a new generation of more flexible stents and flexible delivery systems has prompted consideration of stentassisted angioplasty as an alternative approach in intracranial stenoses (1–5). In addition, recent advances in imaging, including 3D angiography, high-resolution and high-speed angiography, zoom digital fluoroscopy, biplane road mapping, and 3D road map, provide a precise and safe deployment of bioimplants. Although limited experience exists, short- and midterm angiographic and follow-up studies are encouraging. In this chapter, the rationale and indications as well as the technical aspects of endovascular revascularization for intracranial atherosclerotic occlusive disease (ICD) are presented.

DEMOGRAPHICS OF INTRACRANIAL OCCLUSIVE DISEASE As suggested by the increasing number of publications, important differences are observed in the distribution of occlusive vascular disease among African-Americans, Hispanics, and Caucasians and among men and women (6,7). Baker and Iannone described the location and severity of atherosclerosis in 173 consecutive autopsies (8). The most common sites of involvement were the internal carotid artery (ICA) origin, the distal basilar artery, and the proximal and midportion of the basilar artery, while the middle cerebral artery (MCA) was the next most frequently involved, followed by the vertebral arteries and the posterior cerebral artery (PCA). The posterior inferior cerebellar arteries (PICAs), the superior cerebellar artery (SCA), and the distal anterior cerebral artery (ACA) were frequently spared from atherosclerosis. Angiography and autopsy studies all confirm that African-Americans are more prone to develop intracranial atherosclerosis, while Caucasians are prone to develop extracranial disease (9). In the Joint Study of Extracranial Arterial Occlusion, 91% of patients with ICA occlusions were Caucasians, while only 9% were African-Americans (10). In another study, the ratio of MCA to ICA occlusive disease in Caucasians and African-Americans was 13:84 and 10:1, respectively (11). It is suggested that Japanese and Chinese populations, like African-Americans, also have a predilection for occlusive lesions of the intracranial vascular system, rather than the extracranial ICA (12,13). A female preponderance of intracranial occlusive disease, with female to male ratio of 1:4 to 5:9, has been reported in contrast to the male preponderance of extracranial disease (14–16). Flora et al. analyzed 5033 consecutive autopsies and made the following observations (17). From the fourth to sixth decades of life, the percentage of women without any intracranial atherosclerotic disease was higher than that of men, but beyond age 65 the frequency of atherosclerotic lesions was similar in both sexes. Interestingly, diabetic women had higher incidence of intracranial

326

Akkawi and Wakhloo

atherosclerosis than nondiabetic men, and after the fourth decade they had at least as much involvement as diabetic men.

INDICATIONS FOR TREATMENT In the United States, ICD causes approximately 10% of ischemic strokes, i.e., nearly 70,000 to 90,000 ischemic strokes annually (18). The estimated risk of stroke in the setting of intracranial arterial stenosis varies from approximately 7% to 40% per year with or without medical treatment (19–23). The natural history of intracranial stenosis has been determined angiographically in a follow-up period of seven years: 61% of ACA, MCA, and PCA stenotic lesions progressed compared with 20% of intracranial ICA lesions (24). Many patients with MCA stenosis have recurrent cerebral ischemic events despite standard medical therapy with antiplatelet agents or oral anticoagulants (25,26). In the extracranial/intracranial bypass trial, patients with symptomatic MCA stenosis randomized to medical therapy had annual ipsilateral stroke rates of 7.8% and total stroke rates of 9.5% (19,21). Only about one-third of patients had a warning transient ischemic attack (TIA) prior to stroke. The most common presentation was a stroke attack without a warning TIA (19). The Warfarin versus Aspirin for Symptomatic Intracranial Disease (WASID) study was a randomized clinical trial that compared warfarin (target international normalized ratio, 2 to 3) and aspirin (1300 mg/ day) for preventing stroke and vascular death in patients with stenosis (50–99%) of a major intracranial artery (27). The study was prematurely terminated because of safety concerns with warfarin. The drug was associated with significantly higher rates of adverse events, e.g., gastrointestinal bleeding, and provided no significant benefit over aspirin for preventing stroke and vascular death. The risk of stroke was still significant in both the warfarin and the aspirin arm during the mean follow-up period of 1.8 years. Of the 280 patients treated with aspirin, 15% suffered strokes in the territory of the stenotic artery. Of the 289 patients receiving warfarin, 12.1% had ischemic stroke in the same territory. Similarly, the Warfarin-Aspirin Recurrent Stroke Study (WARSS) trial showed that warfarin had no benefit over aspirin for secondary prevention in the subgroup of patients with large artery thrombotic stroke (stenosis or occlusion) (28). However, this study was not specifically designed for patients with intracranial atherosclerotic disease. Medical therapy is mainly used to reduce arterioarterial thromboembolic stroke risk in case of atherosclerotic plaque ulceration. Although reduced regional cerebral blood flow may be associated with ICD, embolic events remain the major cause of stroke in this patient population. Because of the poor response to medical treatment, endoluminal revascularization of intracranial athero-occlusive disease has been introduced. This procedure would facilitate blood flow through the affected area and prevent platelet activa-

tion related to high shear within the stenotic area. A smooth neointimal growth after stenting/angioplasty across the atherosclerotic lesion may also prevent a plaque rupture and platelet adherence. However, initial attempts of revascularization were associated with complication rates as high as 30%. Thus, most clinicians reserve this modality for patients with at least 50% symptomatic stenosis and refractory to optimal medical treatment (29,30). Today, with rapid development of endovascular tools, percutaneous transluminal balloon angioplasty and primary stenting are being discussed as an option for endoluminal revascularization.

PERCUTANEOUS TRANSLUMINAL BALLOON ANGIOPLASTY In 1964, Dotter and Judkins first described PTA for femoral artery stenosis using a coaxial catheter system (31). Later, in 1980, Sundt et al. first reported successful intracranial angioplasty of the basilar artery in two patients (32). Several case reports and case series studies followed with varying results. Angioplasty can be beneficial for patients with intracranial atherosclerosis who remain symptomatic despite aggressive antiplatelet and anticoagulation therapy. Although some investigators have reported good results after intracranial PTA (33–36), others have described high rates of morbidity and mortality (30,37,38). The major problems of intracranial balloon angioplasty are distal embolization, vessel dissection and occlusion, vasospasm and vessel rupture, as well as arterial thrombosis during or immediately after the PTA. Another major disadvantage of the procedure is the risk of restenosis at follow-up, which can result from overshooting neointimal proliferation caused by injury to smooth muscle layer or a progression of the underlying disease. A retrospective study on 36 patients with 37 symptomatic intracranial atherosclerotic lesions who underwent PTA showed a significant stroke risk reduction within the mean follow-up period of 52.9 months. The periprocedural death and stroke rate was 8.3%. The annual stroke rate in the PTAtreated territory was 3.36%. However, when residual stenosis was found to be greater than 50%, annual stroke risk increased to 4.5% (39). Restenosis has been reported to be approximately 30% at three months (40). Mori et al. investigated lesion-specific features for predicting successful PTA and a low restenosis rate (40). A short, less than 5 mm, concentric or moderately eccentric lesion, which is not totally occlusive on angiograms and less than three months old, was defined as type A lesion. A type B lesion was tubular, 5 to 10 mm long, extremely eccentric or occluded, and older than three months. A type C lesion was diffused, more than 10 mm long, extremely angulated (>90%) with excessive tortuosity of the proximal segment or occluded, and older than three months. The highest success rate for PTA (92%) and the lowest incidence of restenosis were found in type A lesions (0% at 1 year) (Fig. 1), whereas in type B

Chapter 18: Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease

327

Figure 1 A 59-year-old male with recurrent episodes of transitory ischemic attacks on antiplatelet therapy and a previous history of left hemiparesis associated with a right hemispheric hemorrhagic infarction. (A) Right ICA angiogram shows a short and concentric atherosclerotic type A lesion at the origin of the MCA (arrow). (B) PTA with a noncompliant 3  9 mm coronary PTA balloon at 6 atm pressure (arrow, Maverick, BSC, Natick, Massachusetts, U.S.). (C) Post angioplasty angiography shows an excellent revascularization with a visible shelf corresponding to the remaining compressed plaque and probably a small plaque dissection (arrow). Abbreviations: ICA, internal carotid artery; MCA, middle cerebral artery; PTA, percutaneous transluminal angioplasty.

lesions the success rate was 82% and restenosis at one year was 33%. The lowest success rate (33%) and the highest restenosis rate or occlusion were found in type C lesions (100% at 1 year) (41). Connors et al. published their experience with PTA for intracranial atherosclerotic stenosis. The report focused on clinical and angiographic outcome related to applied technique and operator experience (33). The authors distinguished three subgroups of patients. In the early period, in which eight patients were treated, the angioplasty balloon size approximated the vessel size or was smaller. Time of balloon inflation was rapid and brief (15–30 seconds). Despite good clinical outcome in seven (87.5%) of eight patients, dissection occurred in 50% of the patients and residual stenosis was found in three of the eight treated patients. In the second subgroup of 12 patients, the PTA balloon size approximated the vessel size, but was permitted to be oversized by up to 0.25 mm. Angioplasty was extremely rapid and brief. Dissection occurred in 9 of 12 patients, resulting in death of one patient due to vessel occlusion. In their most recent experience, which included 50 patients, the PTA balloon was always undersized and inflation was extremely slow and over several minutes. Dissection occurred in 7 of 50 patients (14%), necessitating regional intra-arterial thrombolysis in 2 of 50 patients (4%) with uneventful outcome in both patients. No occlusion or ischemic stroke was observed in any of these patients. Residual stenosis greater than 50% was seen in 8 of 50 patients (16%). Late restenosis was observed in 4 of 44 follow-ups (9%), and a successful reangioplasty was performed in all four patients. Unlike in the coronary or peripheral system, angioplasty of intracranial arteries poses a higher risk of vessel rupture and fatal outcome because of the lack of surrounding supportive soft tissue and the composition of the vessel wall. Thus, technical details, such as selection of PTA balloon size (length and diameter), inflation pressure, and speed and duration of inflation, may all determine the short- and longterm outcome and the periprocedural morbidity.

Balloon overdilatation may be dangerous in the basilar artery and the MCA, where different authors advocate underdilating (not more than one-half of the normal vessel diameter) because of the presence of stiffer adventitia, less elastic tissue, and a greater proportion of smooth muscle, resulting in elevated risk of perforation (42,43). However, there has been no consensus among the interventionalists on techniques of angioplasty in the neurovascular system. Recently, Yoon et al. published their experience with angioplasty of symptomatic MCA stenosis (>70%) on 32 patients (44). The procedure was successful in reducing the stenosis to less than 50% in 91% of the patients. Disabling stroke or death occurred in 6% of cases. The rate of periprocedural TIA was 19%. Five patients had asymptomatic intimal dissection. During a median follow-up period of 20 months, an ipsilateral TIA was seen in one patient, while other patients remained asymptomatic. Although only a small number of intracranial PTAs have been reported, the high risk of periprocedural morbidity associated with artery dissection and acute vessel occlusion and the risk of recoil and delayed restenosis are unacceptable. Compared with the evolution in the coronary system, primary stenting of intracranial atherosclerotic disease is gradually replacing balloon angioplasty.

PRIMARY STENTING AND STENT-ASSISTED ANGIOPLASTY The term ‘‘stent’’ is derived from the name Charles Stent (1845–1901), an English dentist who developed a mold that was used to form an impression of the teeth and oral cavity. Later, the term was used in association with a device that held a skin graft in position, a support for tubular structures that were being anastomosed. More recently, the term is used for an endovascular scaffolding to relieve and prevent vascular obstructions (45). Analogous to the development in peripheral, carotid, and coronary occlusive diseases, primary stenting will replace angioplasty (46). The purported

328

Akkawi and Wakhloo

advantages of stent placement over simple angioplasty include avoiding plaque dislodgment, intimal dissection, elastic vessel recoil, plaque regrowth, and potentially late restenosis. Compared with balloon angioplasty, stent placement also achieves a better angiographic result (Fig. 2). However, intracranial stent placement has potential hazards, such as arterial rupture, in-stent thrombosis, malpositioning and malapposition of the stent, or the inability to pass the stent to the appropriate location due to the tortuosity of the cerebrovascular system. Primary stenting of the stenotic arterial segment without predilatation is frequently feasible (Fig. 4). Only severely stenotic and long lesions may require predilatation for a safe stent placement (47). In a series of 12 patients who underwent an elective stenting of the basilar artery after episodes of vertebrobasilar ischemia, stent placement was successful in all patients; no periprocedural complications were encountered (4). On clinical follow-up between 0.5 to 16 months, all patients remained asymptomatic, except one with nonspecific symptoms and another with TIA. Symptomatic occlusion of penetrating vessels and pontine perforators potentially jailed by stent struts were not observed. Stenting of the MCA was a technical challenge primarily because of the difficulty of tracking coronary balloonexpandable stents around the internal carotid siphon (48). Using coronary stents, elective revascularization of symptomatic MCA stenosis was successful in 8 of

Figure 2 A 62-year-old female with a successful thrombolysis of a basilar trunk occlusion presents with dysarthria and vertigo, intermittently for several days after the procedure and on therapeutic heparinization. (A) Right vertebral artery angiogram shows a long and ulcerated atherosclerotic type C lesion of the vertebrobasilar system with a poor filling of the posterior circulation, the right AICA, and both SCAs. (B) Primary stenting and placement of four overlapping short balloon-expandable 2.5  13-mm coronary stents. While an improved flow to the posterior circulation is noted, the right AICA and the left SCA is still poorly visualized (arrow). (C) Eight-month follow-up angiogram shows an improved flow to the jailed AICA (arrow). Some plaque growth is seen at the level of the AICA. Straightening of the artery is noted within the stented segments (arrowheads). Abbreviations: AICA, anterior inferior cerebellar artery; SCAs, superior cerebellar arteries.

14 patients, without any complications. Stent placement failed in two cases because of the tortuosity of the ICA siphon. Four patients (30%) experienced periprocedural complications, including arterial rupture in two patients and thrombotic occlusion in two other patients (49). In a larger study, Jiang et al. used balloon-expandable coronary stents in 40 patients with MCA stenosis. They reported a proceduralrelated complication rate of 10%, including a vessel occlusion in one patient and subarachnoid hemorrhage in three. The periprocedural mortality rate was 2.5% (50). The Stenting of Symptomatic atherosclerotic Lesions in the Vertebral or Intracranial Arteries (SSYLVIA) trial was a multicenter, nonrandomized, prospective feasibility study that evaluated a new, flexible, stainless steel stent (Neurolink, Guidant Corporation, Indianapolis, Indiana, U.S.) designed for the treatment of extracranial vertebral or intracranial cerebral artery stenosis (51). The study included 61 patients with symptoms attributed to a single lesion with greater than 50% stenosis. Of these 61 patients, 43 (70.5%) had an intracranial stenosis and 18 (29.5%) had an extracranial vertebral artery stenosis. The 30-day stroke and mortality rates were 6.6% and 0%, respectively. Successful stent deployment was achieved in 58 of 61 patients (95%). At six months, 32.4% of intracranial vessels and 42.9% of extracranial vertebral arteries that were treated by stenting demonstrated a recurrent stenosis. Seven (39%) patients presenting with restenosis were symptomatic. Four of 55 patients (7.3%) suffered a stroke between 30 days and 1 year. Sixty-one percent of the patients treated remained asymptomatic during the follow-up. Predictors identified for a postoperative restenosis were vertebral ostial lesion, diabetes mellitus, residual stenosis of more than 30% after stent placement, and the diameter of the treated artery. On the basis of study results, the U.S. FDA granted the stent a humanitarian device exemption (HDE) for the treatment of high-risk patients with significant intracranial and extracranial atherosclerotic diseases who have failed medical therapy. Morphology and hemodynamics of intracranial arteries are substantially different from peripheral and coronary arteries. These vessels lack robust adventitia. External elastic lamina and vasa vasorum are nearly absent, the media are thin, and there are multiple perforators originating from the diseased vessel segment supplying healthy brain tissue. In addition, the cerebrospinal fluid rather than a supportive soft tissue is the microenvironment for the pial vessels (52,53). This environment limits the transmural pressure used for the PTA balloon and the balloon diameter, which may possibly be contributing to greater recoil (54,55). On the other hand, the flow and flow velocity as well as the pulse index in cerebral arteries are higher than the values found in the coronary system. In the past, the endovascular treatment of atherosclerotic disease has focused on angioplasty and placement of balloonexpandable stents. The advantage of a PTA balloon is its flexibility, while the major advantage of balloonexpandable stents is their high radial force. The

Chapter 18: Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease

disadvantage of a PTA balloon is the risk of postprocedural elastic recoil of the vessel and dissection. The major drawbacks of a balloon-expandable stent are its limited flexibility and the risk of injury and dissection due to high expansile force used for deployment. Thus, it was imperative to develop easy trackable neurovascular stents specifically designed for intracranial use with adequate radial force to avoid recoil and provide improved restoration of vessel diameter, while reducing the incidence of complications (Fig. 3). Recently, a new concept was reported for cerebral artery revascularization by using balloon dilatation, followed by the deployment of a self-expanding nitinol microstent, the WingSpan system (Smart Therapeutics, Boston Scientific, Freemont, California, U.S.) (56,57). The results of the WingSpan Multicenter European Study were recently presented. In 45 patients, enrolled from 12 European sites, with symptomatic intracranial atherosclerosis of greater than 50%, revascularization using the WingSpan stent was performed (Fig. 4). Of these patients, 95% had prior strokes and 29% had TIAs. Technical success with angioplasty and stenting was achieved in 98% of patients (57). The periprocedural 30-day death or ipsilateral stroke rate was 4.5% (2/44), and the 6-month death or ipsilateral stroke rate was 7.1% (3/42), with an all-cause stroke rate of 9.5% (4/42). Interestingly, implantation of the self-expanding stent showed a further lumen gain in some patients after initial underdilatation, with a mean residual stenosis of 28% at the six-month follow-up. On the basis of these data, the

329

Figure 3 Low-profile, highly flexible, self-expanding nitinol (nickel-titanium alloy) cerebrovascular stents, which are deployed through a microcatheter. Arrows indicate radiopaque markers at the end of the stent. (A) Closed cell design (Enterprise, Cordis Neurovascular J&J, Miami Lakes, Florida, U.S.). (B) Open cell design (Neuroform, BSC, Natick, Massachusetts, U.S.).

U.S. FDA granted an HDE approval for the WingSpan stent system in 2005 for treatment of symptomatic intracranial stenosis of greater than 50% and refractory to maximal medical therapy. Approval for this stent system was also obtained in Europe. Randomized trials for symptomatic intracranial atherosclerotic disease comparing the best medical treatment and stenting/ angioplasty are being designed.

Figure 4 Symptomatic 61-year-old female with arterial hypertension and an MCA atherosclerotic occlusive disease that was treated with a Wingspan stent system and the Gateway PTA balloon. (A) Illustration of the highly flexible and trackable Wingspan self-expanding stent system with Gateway PTA balloon catheter that were developed for the intracranial circulation. (B) Baseline angiogram of the right ICA shows a concentric atherosclerotic lesion of the M1 segment (arrow) with poststenotic dilatation (arrowhead ) and involvement of the MCA bifurcation (double arrow), the origin of the ACA and the ICA. (C) Postangioplasty angiogram shows no major change in lesion characteristics. The microwire is left in place for placement of the stent. (D) Poststenting angiogram shows smooth vessel boundaries and lumen increase. (E) Six-month follow-up angiogram shows smooth vessel boundaries, slight narrowing at the central part of the stented segment, and remodeling of the dilated M1 section. Abbreviations: MCA, middle cerebral artery; PTA, percutaneous transluminal angioplasty; ICA, internal carotid artery; ACA, anterior cerebral artery. Source: Fig. 4A, courtesy of BSC, Natick, Massachusetts, U.S.

330

Akkawi and Wakhloo

RESTENOSIS AFTER STENTING Stents may induce myointimal hyperplasia and restenosis (Fig. 5). The restenosis rate after stenting of ICD is approximately 10% to 13%. However, recently presented data from several centers showed that at 6- and 12-month follow-up angiography study, WingSpan self-expanding stents may have an incidence of up to 40% in-stent stenosis. To address these issues, stents coated with antiproliferative agents have been developed to reduce the in-stent stenosis and are being considered for the neurovascular system (58). Sirolimus (rapamycin), an antifungal agent, and paclitaxel, a microtubule inhibitor, have shown to prevent neointimal proliferation and restenosis in the coronary vessels when compared with bare metal stents (59). These clinical results provided the impetus to study

the effect of antiproliferative agents for the intracranial system. Levy et al. studied the effects of heparincoated and sirolimus-eluting stents, which were implanted in canine basilar arteries (60,61). The mean percentage of stenosis 12 weeks after implantation was less (12%) in the group with heparin-coated stents compared with 22% in the group with uncoated devices. Compared with bare-metal stents, the sirolimus-coated devices did not impair endothelialization and, furthermore, tended to reduce the proliferation of smooth muscle cells. However, currently no single appropriate animal model exists for the study of intracranial atherosclerotic disease and of the effects of drug-eluting devices.

STENTING PROCEDURE Patients are selected for intracranial angioplasty or stenting by using the inclusion criteria listed in Table 1. Patients are excluded from the revascularization if they meet any of the exclusion criteria listed in Table 2. An experienced physician neurologically evaluates patients prior to the procedure. After selection of the patient for the endovascular treatment, imaging studies should include brain MRI, MR angiograms, CT study, and CT angiograms to assess any preexisting ischemic condition, to rule out a hemorrhage, and to obtain information on cerebrovascular anatomy. Catheter angiography before placement of the stent or balloon angioplasty confirms the degree of vascular stenosis. The percentage of stenosis is calculated as the ratio of the smallest diameter to the diameter of the vessel distal to lesion. In cases of a poststenotic collapse, the diameter of the vessel proximal to the stenosis is used. The dimensions of the normal artery adjacent to the lesion, the length of artery, and the smallest diameter of the stenosis are measured to choose the appropriate balloon or stent size.

Table 1 Inclusion Criteria for Intracranial Stenting

Figure 5 A 57-year-old man with vertebrobasilar insufficiency associated with a high-grade left vertebral artery stenosis. (A) Left vertebral artery angiogram shows an ulcerated atherosclerotic stenotic lesion distal to the origin of the PICA (arrow). (B) Postprimary stenting angiogram shows a recanalization with smooth vessel boundaries. A remnant narrowing of the diseased arterial segment is noted (arrow) as well as some decreased flow within the jailed PICA. (C) Six-month follow-up angiogram shows nonsymptomatic but significant stenosis of the entire stented arterial segment most probably because of intimal hyperplasia (arrow). Note the improved filling of the PICA. (D) Successful angioplasty of the in-stent stenosis (arrow). (E) Left MCA cross section from another patient previously stented for a symptomatic intracranial atherosclerotic disease (arrow), who died of a right intracranial hemorrhage, shows in-stent neointimal hyperplasia (arrowhead) with approximately 65% estimated stenosis. Abbreviations: PICA, posterior inferior cerebellar artery; MCA, middle cerebral artery. Source: Fig. 5E, Movat’s pentachrome stain, courtesy of Dr. D. Lopes.

>50% stenosis of a major intracranial vessel and refractory to medical treatment Minimum vessel diameter of 2.0 mm Previous stroke TIA Neurological symptoms referable to the target lesion Presence of symptoms during the 6 mo prior to treatment Acute vessel occlusion or dissection after PTA Abbreviations: TIA, transient ischemic attack; PTA, percutaneous transluminal angioplasty.

Table 2 Exclusion Criteria for Intracranial Stenting Severe neurological deficit from stroke Chronic total occlusion History of intracranial hemorrhage, hemorrhagic stroke, major stroke, or any stroke with mass effect within 6 wk of procedure Contraindication for or resistant to antiplatelet therapy

Chapter 18: Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease

Patients receive aspirin 325 mg/day orally and Plavix 75 mg/day orally, starting three days prior to the procedure. Aggrenox can be used in patients who are allergic to Plavix. After the procedure, patients are maintained on Plavix 75 mg/day for a minimum of six weeks and aspirin 81 mg/day lifelong. After explaining the procedure and possible complications, an informed consent is obtained from all patients. Most of the procedures can be performed with conscious sedation and analgesia that helps to monitor the neurological status continuously. One of the femoral arteries is accessed with a 5-French (Fr) or a 6-Fr sheath, and rarely a 7-Fr sheath if a larger guide catheter is necessary. A 5-Fr or a 6-Fr (rarely a 7-Fr) guide catheter is then placed in the distal cervical ICA or the cervical vertebral artery. The stents are placed over a guidewire (0.014 inch) passed through the guiding catheter across the lesion (Fig. 6). In cases where a primary stenting cannot be performed, an angioplasty is carried out using a low-profile flexible

331

PTA balloon over a 0.014-inch exchange-length guidewire. PTA balloon is inflated slowly over several minutes, and most of the lesions yield at 6- to 8-atm pressure. In a number of patients with extremely tortuous vascular anatomy, a microcatheter is advanced over a more flexible guidewire to cross the lesion, and then the wire can be exchanged for a stiffer one to navigate the stent delivery system after removal of the microcatheter. Stent insertion is performed under biplane fluoroscopy and road-mapping technique. Proper stent placement is confirmed by high-resolution fluoroscopy or a plain radiogram. Special attention should be paid to avoid overdilatation of the stented arterial segment. Another potential complication during placement of a stent and angioplasty is a ‘‘snow plow’’ effect on atherosclerotic plaques. Plaque may be dislodged in side branches, leading to instant occlusion (Fig. 7). To protect major side branches, placement of protective guidewires or the ‘‘kissing balloon’’ technique has been suggested.

Figure 6 A 76-year-old male with diplopia, dysarthria, and nausea who was refractory to heparin, Coumadin, Plavix, and aspirin. After stenting, the patient has been asymptomatic for four years. (A) Right vertebral artery angiograms in frontal and (B) lateral views show an atherosclerotic stenotic lesion of the right distal vertebral artery (arrowhead ) with extension into the proximal basilar trunk (arrow). (C) Because of proximal tortuosity of the vertebral artery, the primary placement of a coronary stent over a 0.014-inch exchange wire (arrow) was carried out through a left vertebral artery approach. The PTA balloon is being slowly inflated to a maximal pressure of 6 atm. (D) Fully deployed stent (arrow). The exchange wire is kept in place for distal access if needed. (E) Control angiograms in frontal and (F) lateral views show excellent revascularization without obliteration of the right distal vertebral artery. (G) A 14-month follow-up angiograms in frontal and (H) lateral views show mild in-stent neointimal hyperplasia. Some progression of the distal plaque is noted (arrow). (I) Right vertebral artery angiogram shows significant progression of distal vertebral artery atherosclerotic disease (double arrow) but still patent vertebrobasilar junction. Abbreviation: PTA, percutaneous transluminal angioplasty.

332

Akkawi and Wakhloo

Figure 7 A 67-year-old man presenting with recurrent episodes of hypesthesia, dysarthria, and diplopia while being on Coumadin. The patient has been asymptomatic for five years following stent placement. (A) Left vertebral artery angiograms in frontal and (B) lateral views show a high-grade atherosclerotic stenosis with involvement of the PICA origin (arrow). (C) Primary stenting using coronary balloon-expandable system (arrow). (D) Control angiography during partially inflated stent-balloon system (double arrows) shows ‘‘snowplowing’’ of the plaque into the PICA origin. Further dilatation was terminated. (E) Control angiograms show patent flow within the PICA, the plaque (arrowhead ) (F), and some residual narrowing of the stented arterial segment. (G) Six-month follow-up studies in frontal and (H) lateral views show some mild restenosis and narrowing of the PICA origin (arrows). (I) Twenty-month follow-up studies in frontal and (J) lateral views show no further progression of the stented atherosclerotic lesion (arrow) and remodeling of the PICA origin with excellent distal flow (arrowhead ). Abbreviation: PICA, posterior inferior cerebellar artery.

CONCLUSION Endovascular treatment of symptomatic intracranial atherosclerotic disease has rapidly evolved and has adopted technology and experience from the peripheral and coronary systems. Angioplasty and stenting appear to be both safe and highly effective. Patient selection, periprocedural medications, use of appropriate endovascular tools, and proper training are vital for positive clinical results. Advances in biomaterials engineering such as polymer technology, microfabricated stents, and delivery systems will in future allow a safe navigation of the stent through the intracranial arterial circulation. The natural history of intracranial stenosis will have to be better understood to determine acceptable rates of risks for revascularization. Advances in technology and endovascular tools will ensure a low periprocedural morbidity and mortality. New generations of drug-eluting intracranial stents will further improve the long-term outcome currently hampered by a high rate of in-stent stenosis. Nevertheless, prevention, which includes a healthy diet, cessation of smoking, and control of blood pressure, statins, antiplatelet, and nonsteroidal anti-inflammatory drugs, will remain the long-term goal of combating ICD.

REFERENCES 1. Al-Mubarak N, Gomez CR, Vitek JJ, et al. Stenting of symptomatic stenosis of the intracranial internal carotid artery. AJNR Am J Neuroradiol 1998; 19:1949–1951. 2. Dorros G, Cohn JM, Palmer LE. Stent deployment resolves a petrous carotid artery angioplasty dissection. AJNR Am J Neuroradiol 1998; 19:392–394. 3. Feldman RL, Trig L, Gaudier J, et al. Use of a coronary Palmaz Schatz stent in the percutaneous treatment of a intracranial carotid artery stenosis. Cathet Cardiovasc Diagn 1996; 38:316–319. 4. Gomez CR, Misra VK, Liu MW, et al. Elective stenting of symptomatic basilar artery stenosis. Stroke 2000; 31:95–99. 5. Mori T, Kazita K, Mori K. Cerebral angioplasty and stenting for intracranial vertebral atherosclerotic stenosis. AJNR Am J Neuroradiol 1999; 20:787–789. 6. Wityk RJ, Lehman D, Klag M, et al. Race and sex differences in the distribution of cerebral atherosclerosis. Stroke 1996; 27:1974–1980. 7. Caplan LR, Gorelick PB, Hier DB. Race, sex and occlusive cerebrovascular disease: a review. Stroke 1986; 17:648–655. 8. Baker AB, Iannone A. Cerebrovascular disease. I. The large arteries of the circle of Willis. Neurology 1959; 9:321–332. 9. Bauer RB, Sheehan S, Wechsler N, et al. Arteriographic study of the sites, incidence and treatment of arteriosclerotic cerebrovascular lesions. Neurology 1962; 12:698–711.

Chapter 18: Stenting and Angioplasty for Intracranial Atherosclerotic Occlusive Disease 10. Fields WS, Lemak N. Joint study of extracranial arterial occlusion. X. Internal carotid artery occlusion. JAMA 1976; 235:2734–2738. 11. Heyden S, Heyman A, Goree JA. Nonembolic occlusion of the middle cerebral and carotid arteries: a comparison of predisposing factors. Stroke 1970; 1:363–369. 12. Feldmann E, Daneault N, Kwan E, et al. Chinese-white differences in the distribution of occlusive cerebrovascular disease. Neurology 1990; 40:1541–1545. 13. Kieffer SA, Takeya Y, Resch JA, et al. Racial differences in cerebrovascular disease: angiographic evaluation of Japanese and American populations. Am J Roentgenol 1967; 101:94–99. 14. Kunitz S, Gross C, Heyman A, et al. The pilot stroke data bank: definition, design, and data. Stroke 1984; 15:740–746. 15. Hinton R, Mohr J, Ackerman R, et al. Symptomatic middle carotid artery stenosis. Ann Neurol 1979; 5:152–157. 16. Ford CS, Howard VJ, Howard G. The sex difference in manifestations of carotid bifurcation disease. Stroke 1986; 17:877–881. 17. Flora G, Baker A, Loewenson R, et al. A cooperative study of cerebral atherosclerosis in males and females. Circulation 1968; 38:859–869. 18. Sacco RL, Kargman DE, Gu Q, et al. Race-ethnicity and determinants of intracranial atherosclerotic cerebral infarction: the Northern Manhattan Stroke Study. Stroke 1995; 26:14–20. 19. Bogousslavsky J, Barnett HJ, Fox AJ, et al. Atherosclerotic disease of the middle cerebral artery. Stroke 1986; 17:1112–1120. 20. Chimowitz MI, Kokkinos J, Strong J, et al. The warfarinaspirin symptomatic intracranial disease study. Neurology 1995; 45:1488–1493. 21. The EC/IC Bypass Study Group. Failure of extracranialintracranial arterial bypass to reduce the risk of ischemic stroke. Results of an international randomized trial. N Engl J Med 1985; 313:1191–1200. 22. Marzewski DJ, Furlan AJ, St Louis P, et al. Intracranial internal carotid artery stenosis: long-term prognosis. Stroke 1982; 13:821–824. 23. Wechsler LR, Kistler JP, Davis KR, et al. The prognosis of carotid siphon stenosis. Stroke 1986; 17:714–718. 24. Akins PT, Pilgram TK, Cross DT III, et al. Natural history of stenosis from intracranial atherosclerosis by serial angiography. Stroke 1998; 29:433–438. 25. Arenillas JF, Molina CA, Montaner J, et al. Progression and clinical recurrence of symptomatic middle cerebral artery stenosis: a long-term follow-up transcranial Doppler ultrasound study. Stroke 2001; 32:2898–2904. 26. Wong KS, Li H, Lam WW, et al. Progression of middle cerebral artery occlusive disease and its relationship with further vascular events after stroke. Stroke 2002; 33:532–536. 27. Chimowitz MI, Lynn MJ, Howlett-Smith H, et al. Comparison of warfarin and aspirin for symptomatic intracranial arterial stenosis. N Engl J Med 2005; 352:1305–1316. 28. Mohr JP, Thompson JL, Lazar RM, et al., for WarfarinAspirin Recurrent Stroke Study Group. A comparison of warfarin and aspirin for the prevention of recurrent ischemic stroke. N Engl J Med 2001; 345:1444–1451. 29. Levy EI, Horowitz MB, Koebbe CJ, et al. Transluminal stent-assisted angioplasty of the intracranial vertebrobasilar system for medically refractory, posterior circulation ischemia: early results. Neurosurgery 2001; 48:1215–1223. 30. Takis C, Kwan ES, Pessin MS, et al. Intracranial angioplasty: experience and complications. AJNR Am J Neuroradiol 1997; 18:1661–1668. 31. Dotter CT, Judkins MP. Transluminal treatment of arteriosclerotic obstruction. Description of a new technique

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45. 46.

47.

48.

49.

50.

51.

333

and a preliminary report of its application. Circulation 1964; 30:654–670. Sundt TM Jr., Smith HC, Campbell JK, et al. Transluminal angioplasty for basilar artery stenosis. Mayo Clin Proc 1980; 55:673–680. Connors JJ III, Wojak JC. Percutaneous transluminal angioplasty for intracranial atherosclerotic lesions: evolution of technique and short-term results. J Neurosurg 1999; 91:415–423. Clark WM, Barnwell SL, Nesbit G, et al. Safety and efficacy of percutaneous transluminal angioplasty for intracranial atherosclerotic stenosis. Stroke 1995; 26:1200–1204. Honda S, Mori T, Fukuoka M, et al. Successful percutaneous transluminal angioplasty of the intracranial vertebral artery 1 month after total occlusion. Neurol Med Chir (Tokyo) 1994; 34:551–554. Mori T, Mori K, Fukuoka M, et al. Percutaneous transluminal cerebral angioplasty for total occlusion of middle cerebral arteries. Neuroradiology 1997; 39:71–74. Higashida RT, Tsai FY, Halbach VV, et al. Cerebral percutaneous transluminal angioplasty. Heart Dis Stroke 1993; 2:497–502. Alazzaz A, Thornton J, Aletich VA, et al. Intracranial percutaneous transluminal angioplasty for arteriosclerotic stenosis. Arch Neurol 2000; 57:1625–1630. Marks MP, Marcellus ML, Do HM, et al. Intracranial angioplasty without stenting for symptomatic atherosclerotic stenosis: long-term follow-up. AJNR Am J Neuroradiol 2005; 26:525–530. Mori T, Mori K, Fukuoka M, et al. Percutaneous transluminal cerebral angioplasty: serial angiographic followup after successful dilatation. Neuroradiology 1997; 39:111–116. Mori T, Fukuoka M, Kazita K, et al. Follow-up study after intracranial percutaneous transluminal cerebral balloon angioplasty. AJNR Am J Neuroradiol 1998; 19:1525–1533. Ahuja A, Guterman LR, Hopkins LN. Angioplasty for basilar artery atherosclerosis: case report. J Neurosurg 1992; 77:941–944. Nakatsuka H, Ueda T, Ohta S, et al. Successful percutaneous transluminal angioplasty for basilar stenosis: technical case report. Neurosurgery 1996; 39:161–164. Yoon W, Seo JJ, Cho KH, et al. Symptomatic middle cerebral artery stenosis treated with intracranial angioplasty: experience in 32 patients. Radiology 2005; 237:620–626. Dorland’s Illustrated Medical Dictionary. 28th ed. Philadelphia: WB Saunders Co., 1994:1577. Yadav JS, Roubin GS, King P, et al. Angioplasty and stenting for restenosis after carotid endarterectomy: initial experience. Stroke 1996; 27:2075–2079. Lylyk P, Vila JF, Miranda C, et al. Endovascular reconstruction by means of stent placement in symptomatic intracranial atherosclerotic stenosis. Neurol Res 2005; 27(suppl 1):S84–S88. Gomez CR, Misra VK, Campbell MS, et al. Elective stenting of symptomatic middle cerebral artery stenosis. AJNR Am J Neuroradiol 2000; 21:971–973. Kim JK, Ahn JY, Lee BH, et al. Elective stenting for symptomatic middle cerebral artery stenosis presenting as transient ischaemic deficits or stroke attacks: short term arteriographical and clinical outcome. J Neurol Neurosurg Psychiatry 2004; 75:847–851. Jiang WJ, Wang YJ, Du B, et al. Stenting of symptomatic M1 stenosis of middle cerebral artery: an initial experience of 40 patients. Stroke 2004; 35:1375–1380. SSYLVIA Study Investigators. Stenting of Symptomatic Atherosclerotic Lesions in the Vertebral or Intracranial Arteries (SSYLVIA): study results. Stroke 2004; 35:1388–1392.

334

Akkawi and Wakhloo

52. Volk EE, Prayson RA, Perl J II. Autopsy findings of fatal complication of posterior cerebral circulation angioplasty. Arch Pathol Lab Med 1997; 121:738–740. 53. Klugherz BD, Llanos G, Lieuallen W, et al. Twenty-eightday efficacy and pharmacokinetics of the sirolimuseluting stent. Coron Artery Dis 2002; 13:183–188. 54. Terada T, Higashida RT, Halbach VV, et al. Transluminal angioplasty for arteriosclerotic disease of the distal vertebral and basilar arteries. J Neurol Neurosurg Psychiatry 1996; 60:377–381. 55. McKenzie JD, Wallace RC, Dean BL, et al. Preliminary results of intracranial angioplasty for vascular stenosis caused by atherosclerosis and vasculitis. AJNR Am J Neuroradiol 1996; 17:263–268. 56. Henkes H, Miloslavski E, Lowens S, et al. Treatment of intracranial atherosclerotic stenoses with balloon dilatation and self-expanding stent deployment (WingSpan). Neuroradiology 2005; 47:222–228. 57. Bose A. The Wingspan Study. Presented at the American Society of Neuroradiology 43rd Annual Meeting, Toronto, Canada, May 21–27, 2005.

58. Stone GW, Ellis SG, O’Shaughnessy CD, et al., for TAXUS V ISR Investigators. A polymer-based, paclitaxel-eluting stent in patients with coronary artery disease. N Engl J Med 2004; 350:221–231. 59. Ong AT, van Domburg RT, Aoki J, et al. Sirolimus-eluting stents remain superior to bare-metal stents at two years: medium-term results from the Rapamycin-Eluting Stent Evaluated at Rotterdam Cardiology Hospital (RESEARCH) registry. J Am Coll Cardiol 2006; 47:1356–1360. 60. Levy EI, Hanel RA, Howington JU, et al. Sirolimus-eluting stents in the canine cerebral vasculature: a prospective, randomized, blinded assessment of safety and vessel response. J Neurosurg 2004; 100:688–694. 61. Levy EI, Boulos AS, Hanel RA, et al. In vivo model of intracranial stent implantation: a pilot study to examine the histological response of cerebral vessels after randomized implantation of heparin-coated and uncoated endoluminal stents in a blinded fashion. J Neurosurg 2003; 98:544–553.

19 Endovascular Management of Dural Arteriovenous Fistulas J. Marc C. van Dijk Department of Neurosurgery, University Medical Center, Groningen, Groningen, The Netherlands

Robert A. Willinsky Department of Medical Imaging, Toronto Western Hospital, Toronto, Ontario, Canada

INTRODUCTION Excluding traumatic lesions and carotid-cavernous fistulas, three basic categories of dural arteriovenous (AV) shunts can be discriminated: the pediatric dural sinus malformation with AV shunting, the infantile high-flow dural arteriovenous fistula (DAVF), and the adult-type DAVF. Since the dural lesions in the pediatric age group are beyond the scope of this chapter, the focus will be on the adult-type DAVF. The adult-type DAVF is a rare neurovascular lesion. A recent population-based study mentions a crude detection rate of 0.16 per 100,000 adults per year, and from angiographic studies it has been estimated that dural lesions represent just 10% to 15% of all intracranial AV shunts (1,2). Nevertheless, as a unique neuropathological entity, the subject DAVF deserves full attention. A DAVF consists of one or more true fistulas, i.e., direct AV connections without an intermediate capillary network or even a nidus. The fistula itself is confined to the leaflets of the dura mater, which unquestionably differentiates a DAVF from an arteriovenous malformation (AVM) that has a (sub)pial localization within the brain or spinal cord. In addition, considering the general acknowledgment that a DAVF is acquired (3–6), in contrast to the congenital nature of a vascular malformation (7), the term ‘‘dural arteriovenous fistula’’ is clearly preferable to ‘‘dural arteriovenous malformation.’’ The anatomical setting of the fistula within the dura mater explains the fact that both cranial and spinal DAVFs are recognized. Although their underlying pathophysiology is the same, their clinical presentation and behavior (and therefore their classification) are quite distinct. Consequently, cranial and spinal DAVFs are discussed in separate sections of this chapter.

PATHOPHYSIOLOGY The cause of DAVFs is still a matter of debate, although many causes have been posited in the literature. The

development of DAVFs has been described following surgery (even in remote parts of the body) or head trauma and in relation to sinus thrombosis, without revealing a definite etiological pathway. Over the decades, two main hypotheses of pathogenesis have been advanced. The first postulates that DAVFs normally exist within the dura mater as ‘‘dormant’’ channels between the meningeal arterial circulation and the venous system. Histological and radiological studies have demonstrated that these communications are indeed present in the dura of normal individuals (8). According to this theory, the channels open because of the venous hypertension associated with sinus thrombosis or sinus outflow obstruction (9). Similarly, the existence of thin-walled venous pouches close to small meningeal arteries has been reported. Rupture of these fragile pouches might be responsible for direct AV communications within the dura (10). The second hypothesis claims that the development of DAVFs is a direct consequence of neovascularization processes within the dura mater, attributable to the release of angiogenetic factors. These factors, e.g., vascular endothelial growth factor and basic fibroblast growth factor, can be either directly produced by the organization of a venous sinus thrombosis or indirectly induced by the increased intraluminal venous pressure through a tissue hypoxia pathway. Major support for the second hypothesis arose from the positive staining of excised dural fistulas for venous thrombosis (11) and angiogenetic factors (12), as well as by the demonstration that the development of DAVFs in a rat model can be reproduced using a combination of venous sinus thrombosis and venous hypertension (3,6). Histopathological studies prove that DAVFs are located within the meningeal wall of a venous sinus, although originally they had been described to reside within the thrombosed lumen of a dural sinus. The location within the wall clarifies the existence of the type of DAVF that drains directly into the cortical venous network, without venous drainage into the involved sinus (13). On close examination, the fistula consists of small venules with a diameter of

336

van Dijk and Willinsky

approximately 30 mm. These vessels have been named ‘‘crack-like vessels,’’ since they resemble cracks in the wall of the dural sinus after histological staining. Further immunohistochemical assessment of the crack-like vessels revealed a layer of endothelial and smooth muscle cells and no internal elastic lamina, thus confirming their venous origin (14). In spinal DAVFs, a small series of high-resolution microangiography after en bloc resection of the fistula has been reported. The ex vivo surgical removal included the involved dural root sleeve, proximal nerve root, and adjacent spinal dura. The subsequent detailed imaging showed that in all of the lesions, the artery split into daughter vessels that coalesced one to three times to form a network of arterial loops in the dura that invariably emptied into a medullary vein without an intervening capillary plexus. Several collateral vessels arising from adjacent intercostal or lumbar arteries were commonly present in the dura and converged at the site of the fistulous point to join a single medullary vein, providing an anatomical explanation for the presence of a multiple segment arterial supply (15).

CRANIAL DURAL AV FISTULAS Early descriptions of cranial DAVFs date from the first decades of the 20th century, typically in the form of single case reports, such as by To¨nnis in 1936 (16). The first presentation of the concept of dural fistulas originates from a publication by Fincher in 1951 (17). Despite the development of cerebral angiography in the late 1920s (18), it took nearly 40 years before the DAVF emerged as a distinct entity because of the advances in angiography, such as magnification, subtraction techniques, and selective arterial catheterization. In those days, all DAVFs were regarded as benign in comparison with the pial AVMs (19). In the early 1970s, pioneers like Aminoff, Newton, and Djindjian expanded the anatomical and clinical understanding of the dural fistulas. In 1972, the perception grew that the pattern of venous drainage was related to the clinical signs and symptoms; nevertheless, not until 1975 were the particular risks associated with cortical venous reflux (CVR) recognized (20,21). Djindjian and Merland proposed the first classification based on this concept, emphasizing that DAVFs with a free outflow into a sinus are relatively harmless and that the presence of CVR is related to severe complications (22). Dominated by the publication of three large reviews in the literature, the 1980s were an important decade for the cranial DAVFs. In 1984, on the basis of a series of 223 cases, Malik concluded that venous outflow was a key factor in the occurrence of hemorrhage or neurological symptoms and stressed the importance of location of the fistula within or outside a major venous sinus (23). Lasjaunias performed a meta-analysis on 195 cases, concluding that focal neurological deficits were dependent on the territory of draining veins and that CVR was related to a high risk of intradural bleeding (24). Studying 377 cases, Awad did the third major review. In this report, the term

‘‘aggressive’’ was introduced for DAVFs presenting with a hemorrhage or a focal neurological deficit. Awad not only emphasized the importance of CVR but also pointed out venous ectasias and galenic drainage as additional risk factors (25). During the 1990s and the beginning of the 21st century, the accumulation of information was further consolidated in the classification schemes by Cognard and Borden. The formation of multidisciplinary neurovascular study groups has elaborated the understanding of the cranial DAVF and underscored the technical and clinical significance of the CVR. By reporting large series, many authors have confirmed the clinical behavior first delineated in the 1980s. In addition, the ongoing development of endovascular techniques has proven invaluable in the treatment of these lesions.

Cranial Classification A number of classification schemes have been proposed on the basis of different aspects of DAVFs. At first, the anatomical location of the fistula was considered the key discriminating feature. In 1973, Aminoff recommended the arrangement in an anteroinferior group and a posterosuperior group (26). Subsequently, other large studies pointed out an association between location and clinical presentation of a DAVF by noticing that cavernous sinus and transverse sinus lesions behave differently than tentorial or anterior fossa fistulas. Over the years, however, it became evident that not location but location-related venous drainage pattern was crucial in determining the clinical presentation. Cranial DAVFs in some locations have a higher likelihood of developing CVR because of the local venous anatomy, e.g., the absence of a venous sinus in the direct vicinity. Although no location of a cranial DAVF is immune from aggressive behavior, it was recognized that certain regions raise the index of suspicion for the development of CVR. Following Djindjian’s original format based on the venous drainage pattern, numerous classification schemes for cranial DAVFs have been put forward, of which the Borden and Cognard classifications are most frequently applied (Table 1) (27,28). Both are used in everyday practice and have their own advantages: the three-step Borden classification is very simple to apply and requires only little knowledge of cerebral angiography; the Cognard classification is certainly theoretically superior, since it incorporates the additional effect of the flow direction in the dural sinus, but its multiple steps require a more advanced comprehension of DAVFs. The importance of judging the flow direction in the sinus is evident: retrograde flow can prohibit the cortical veins to drain into the involved sinus and can subsequently lead to venous congestion of the brain, without the occurrence of CVR. Both classifications have been validated (29).

Cranial Clinical Features The expressions ‘‘benign’’ and ‘‘aggressive’’ have been put forward in the literature and used in this chapter in relation to the typical clinical signs and symptoms of

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas Table 1 Classification Schemes for DAVFs Borden classification 1 2 3

Venous drainage directly into dural venous sinus or meningeal vein Venous drainage into dural venous sinus with CVR Venous drainage directly into subarachnoid veins (CVR only)

Cognard classification I IIa IIb IIa þ b III IV V

Venous drainage into dural venous sinus with antegrade flow Venous drainage into dural venous sinus with retrograde flow Venous drainage into dural venous sinus with antegrade flow and CVR Venous drainage into dural venous sinus with retrograde flow and CVR Venous drainage directly into subarachnoid veins (CVR only) Type III with venous ectasias of the draining subarachnoid veins Venous drainage into the perimedullary plexus

Abbreviation: CVR, cortical venous reflux. Table 2 Benign and Aggressive Clinical Features of DAVFs Benign features

Aggressive features

Pulsatile bruit Orbital congestion

Intracranial hemorrhage Nonhemorrhagic focal neurological deficit Dementia Papilledema Death

Cranial nerve palsy Chronic headache Asymptomatic

cranial DAVFs. Features such as nonhemorrhagic neurological deficits (NHND), hemorrhage, and death are regarded aggressive, whereas complaints of chronic headache, pulsatile bruit, and orbital symptoms, including cranial nerve deficits, e.g., due to cavernous sinus fistulas, are considered benign, even though these signs and symptoms might be deemed intolerable by the patient (Table 2). Benign DAVFs

There is a proven relationship between the so-called benign symptoms and the absence of CVR in the venous drainage pattern. Borden type 1 and Cognard type I and IIa DAVFs are thus regarded as benign fistulas. Large series have acknowledged that the benign DAVFs never present with grave pathology (25,30,31). Most frequently, benign fistulas include those involving the cavernous sinus or the transversesigmoid sinus. Although DAVFs can occur at any age, typically the patient is over 50 years old and complains about a disturbing pulse-synchronous tinnitus. The bruit can be very loud and audible by the physician, indicative of a high turbulent flow through a venous sinus in direct contact with the petrous bone. Another characteristic complaint is the ‘‘red eye,’’ indicative of a DAVF involving the cavernous sinus with subsequent proptosis, conjunctival injection, and chemosis of one or both eyes. The orbital symptoms may be

337

progressive, and the congestion can lead to raised intraocular pressure that may end in decrease of visual acuity. However, (transient) cranial nerve dysfunction may as well worsen the visual function or diplopia, leading to ophthalmoplegia or orbital edema with extraocular muscle swelling. Hence, in cavernous sinus DAVFs, close interaction with the ophthalmologist is essential to decide when clinical symptoms are no longer to be considered benign and palliative endovascular treatment is necessary. Even more important than the clinical presentation is the subsequent natural course of benign DAVFs. In a retrospective evaluation of 205 cases, Cognard reported the probable benign course of 111 fistulas without CVR (28), but significant follow-up could only be obtained in 66% of the cases, and it is difficult to draw conclusions on clinical stability from this study, since clinical events that occur before and after presentation are not well differentiated. Two years later, the authors mentioned seven patients in their experience who initially had a DAVF without CVR but showed a worsening in the venous drainage pattern during a mean seven-year follow-up. Five patients were embolized with particles, one patient had proximal ligation of the occipital and middle meningeal artery, and one had conservative management. In all cases, the change in venous pattern was accompanied by a worsening of the clinical symptoms (32). Davies reported the first large prospectively collected series of patients with a benign DAVF concerning a cohort of 55 cases without CVR over a mean follow-up period of 33 months. These cases made clear that the vast majority of DAVFs without CVR behave in a benign fashion, with stability over time. One patient, nevertheless, died after palliative endovascular treatment, without angiographic conversion into a lesion with CVR. This unusual course of a predicted benign disease was explained as the result of venous hypertension due to functional obstruction of the superior sagittal sinus (31). In 2002, Satomi further confirmed the benign disease course of DAVFs without CVR in a prospective study of 117 cases. Observation resulted in a tolerable, stable disease in 67 of 68 patients. One patient had a seizure during follow-up because of a hemorrhage after the development of CVR. Treatment was performed in 44 patients, aimed at the palliation of unbearable symptoms or in reversing ophthalmological phenomena. In 15 patients, multiple treatment sessions (up to 4) were required to obtain a satisfactory result. Treatment resulted in a tolerable disease in all but one patient. Both in the observational and in the palliatively treated group, approximately 66% of the patients were cured after mean 28 months, demonstrating that a benign DAVF in essence is a self-limiting disease. The remaining one-third of patients had symptoms that were well tolerated. In conclusion, the restricted management resulted in a stable and tolerable disease course in more than 98% of the cases (33). One should, however, bear in mind that benign DAVF is a dynamic disease and progression of the venous thrombosis may very well result in rerouting of the venous drainage. From this point of view, two

338

van Dijk and Willinsky

Figure 1 Venous congestive encephalopathy secondary to a DAVF. A 58-year-old male presented with ataxia. The axial T2-weighted MR (A) shows a central hyperintensity within the right cerebellar hemisphere surrounded by diffuse hypointensity. Note the prominent flow voids (arrows). The gadolinium-enhanced MR (B) shows diffuse peripheral enhancement. A selective arteriogram (C) of a dural artery (arrows) arising from the right vertebral artery shows a DAVF inferior to the straight sinus with marked reflux into the cortical veins of the right cerebellar hemisphere. Abbreviation: DAVF, dural arteriovenous fistula.

cautionary points have to be made: the possible significance of retrograde flow in the sinus, and the small but noteworthy risk of conversion into an aggressive DAVF. First, the importance of retrograde flow in the sinus without CVR is stressed in the Cognard classification. Type IIa was, according to the describing author, related to papilledema and raised intracranial pressure in 8 of 27 patients (30%). In the series by Davies and Satomi, however, this phenomenon was seen in a much smaller percentage. Theoretically, retrograde flow can indeed prohibit the cortical veins from draining into the involved sinus and lead to venous congestion of the brain. Among other authors, Hurst has related this venous congestion of the brain to global neurological deficits, such as dementia (34). The congestion has been shown to project as T2 hyperintensity on MRI (Figs. 1 and 2) (35). Willinsky additionally pointed out the presence of tortuous, engorged veins on the cerebral angiography in cases of venous congestion and labeled this the pseudophlebitic pattern (36). The other point of concern in benign DAVFs is the chance of secondary development of CVR. In the series by Davies and Satomi, this phenomena occurred in approximately 2% of the cases. This finding mandates close clinical follow-up and renewed radiological evaluation with any sudden or unexpected change in symptoms. Aggressive DAVFs

Cranial DAVFs with CVR on angiography are associated with grave pathology. In the classification schemes, Borden type 2/3 and Cognard type IIb/IIa+b/III/IV/V DAVFs are therefore labeled aggressive. Initially, the term ‘‘aggressive’’ was based on the frequent clinical presentation with intracranial hemorrhage, NHND, or even death. Nevertheless, presentation with pulsatile tinnitus or orbital congestions, mimicking a benign DAVF, is also possible (25,28). The hemorrhage may be subdural, subarachnoid, or intracerebral, since the

Figure 2 Resolution of T2 hyperintensity after treatment. A 58-year-old female presented with a progressive quadriplegia. Axial T2-weighted MR (A) shows diffuse central hyperintensity within the medulla (arrow). AP view of the right vertebral arteriogram (B) shows a DAVF at the foramen magnum with drainage into a large perimedullary vein (arrow). Selective arteriogram (C) of the dural branch feeding the fistula better defines the fistula. Embolization was not done because it proved impossible to obtain a safe position to inject liquid adhesives. Surgical disconnection was done with a single clip to close the intradural draining vein. The patient’s clinical condition gradually improved and at six months the T2 hyperintensity on the MR had resolved (D). Abbreviation: DAVF, dural arteriovenous fistula.

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas

339

refluxing cortical veins traverse each of these different compartments. As stated before, location of the DAVF is not correlated with hemorrhage; however, because of local venous anatomy, some locations are more prone to present with a hemorrhage, e.g., anterior cranial fossa DAVFs. The NHND is typically focal and directly related to regional venous congestion (24), although more global neurological deficits are also encountered analogous to the Cognard type IIa fistulas with retrograde flow in the sinus (see above). Cranial DAVFs can also be responsible for spinal neurological disorders in case of perimedullary CVR from a Cognard type V lesion (37,38). The natural course of aggressive DAVFs has been under debate for a long time. In the literature, contradictory reports have been published, with postpresentation annual hemorrhage rates varying from 1.8% in a series by Brown (39) to almost 20% in a series by Davies (40). Brown followed patients for a mean of 6.6 years but did not select for the presence of CVR, so it is very likely that he underestimated the annual risk because of a large proportion of benign DAVFs in the cohort. On the other hand, Duffau in another series found a rebleeding percentage up to 35% within two weeks after the presenting hemorrhage in case of Cognard type III/IV fistulas (41). Davies calculated an annual mortality of 19.3%, with a 19.2% annual rate of hemorrhage and a 10.9% annual rate of NHND during the disease course of DAVFs with persistent CVR. Van Dijk, in his study, recalculated these rates on the basis of a larger population and four times the follow-up time. During a combined follow-up time of 86.9 patient-years, an annual mortality rate of 10.4% was yielded, with an annual hemorrhage rate of 8.1% and an annual NHND rate of 6.9% (42). These data mandate prompt and curative management of the aggressive DAVFs.

In case of an aggressive DAVF, unenhanced CT may show hypodensities, representing areas of edema or venous ischemia. Abnormally enlarged pial veins can sometimes be visualized as increased densities in comparison to the brain parenchyma. Contrast-enhanced CT shows enhancement of the refluxing cortical venous network. A promising technique is the development of 3D CT angiography, especially in emergencies. The disadvantage of 3D CT angiography is its poor characterization of hemodynamic details.

Cranial Diagnostic Imaging

Digital Subtraction Angiography

CT

In the absence of CVR with congestion of the brain, benign DAVFs are nearly always occult on CT imaging.

MRI/MRA

On MRI, it is difficult to detect a benign DAVF, although irregular or stenotic venous sinus might raise suspicion. MRA is more sensitive, but still has limitations in depicting the fistula. Aggressive DAVFs are better visualized by MRI, characterized by flow voids on the cortex corresponding to dilated pial vessels. The brain parenchyma can show T2 hyperintensity in the white matter secondary to the venous congestion of the brain, especially in the deep white matter (43). The differential diagnosis for T2 hyperintensity includes sinus thrombosis (with venous infarction or venous congestion), demyelinization, and neoplasm. However, T2 hyperintensity in the parenchyma in combination with a surplus of pial vessels is highly suggestive of a vascular malformation (Fig. 1). This T2 hyperintensity resolves after treatment (Fig. 2). A new technique that in time will potentially replace conventional angiography is the real-time autotriggered elliptic centric-ordered 3D gadoliniumenhanced MRA (ATECO-MRA) (44). ATECO-MRA is effective in demonstrating DAVFs, especially those with CVR (Fig. 3). The technique is also ideally suited to follow up DAVFs. However, the protocol for following untreated DAVFs needs to be validated.

Conventional angiography is important to confirm the diagnosis of a DAVF and to plan the treatment. Selective contrast injections into the divisions of the external carotid artery will show rapid AV shunting through

Figure 3 Gadolinium-enhanced MRA detects DAVF and its cortical venous drainage. A 58-year-old male presented with conjunctival injection and chemosis in the right eye. The collapsed axial image from the gadoliniumenhanced MR angiogram (A) shows an early draining cortical vein (short arrow) and drainage into the right superior ophthalmic vein (long arrow). The lateral view from the right external carotid arteriogram (B) confirms that there is a cavernous sinus DAVF with CVR (arrow). Abbreviations: DAVF, dural arteriovenous fistula; CVR, cortical venous reflux.

340

van Dijk and Willinsky

the fistula into the cranial venous system. Contrast injection selectively into the internal carotid or vertebral arteries may reveal a delayed cerebral circulation time, compatible with venous congestive encephalopathy (36). In the venous phase of the angiogram, tortuous dilated collateral veins may be evident over the surface of the brain indicating long-standing venous hypertension. This finding has been referred to as the pseudophlebitic pattern and correlates with a greater risk of future hemorrhage or NHND (35). Careful analysis of the venous phase of the cerebral circulation is critical in planning treatment. Beyond doubt, the main goal in the imaging of cranial DAVFs is to detect the existence of CVR. Global nonselective angiography should be avoided as subtle CVR might be missed. Another important objective is to look for venous sinus outflow obstruction, which can result in extracranial drainage through collateral routes, including the orbital system. Venous stenosis or obstruction is a common finding in patients with retrograde flow into the cortical and cerebellar veins. Careful analysis of the venous phase of the angiogram may reveal distinct pathways that separately drain the fistula and the brain. This compartmentalization of the involved dural sinus may allow closure of the fistula and preservation of that compartment of the fistula that drains the brain (45). Finally, in diagnostic imaging the existence of multiple DAVFs within one patient should be considered, since it is reported with a frequency of 7% to 8% (46).

Cranial Therapeutic Considerations Awareness of the natural history of a given disease is essential to every clinician, because an active treatment is supposed to ameliorate the outcome of the disease. If treatment of a DAVF is indicated, the

endovascular approach is the first choice. Surgery is used either in combination with the endovascular techniques or when the endovascular technique fails. Surgery can allow access for an endovascular approach or be used to directly occlude the diseased sinus or disconnect the CVR. The endovascular treatment can be accomplished from a transarterial approach, a transvenous route, or a combination of both. The transvenous approach is now the preferred treatment in most DAVFs because of its effectiveness in obliterating the fistula or eliminating the CVR. Management of Benign DAVFs

In benign cranial DAVFs, reports show that 98% of patients have an excellent natural history, indicating that, in general, observation with gadoliniumenhanced MRA reevaluation is the best available management. A three-year follow-up catheter DSA is advised in patients with stable clinical signs and symptoms. If there is any sudden or unexpected change in the clinical status, either worsening or improvement (even disappearance), repeat catheter angiography is needed to exclude the development of CVR or progressive thrombosis with retrograde flow in the venous sinus (47). In patients who either suffer an intolerable bruit or have severe orbital symptoms, e.g., compromise of vision, palliative arterial endovascular embolization could be considered to reduce symptoms. Arterial embolization often reduces the symptoms, but one should realize that as a rule it is not effective in obtaining a complete angiographic obliteration of the fistula. In some cases, arterial embolization with liquid adhesives can permanently close the fistula (Fig. 4). With arterial particle embolization, an early improvement of

Figure 4 Embolization of a cavernous sinus DAVF with liquid adhesives. A 64-year-old male presented with chemosis, conjunctival injection, and raised intraocular pressure in the right eye. Lateral right external carotid arteriogram (A) shows a DAVF draining exclusively into the superior ophthalmic vein (open arrow). The fistula is into a small compartment in the anterior aspect of the cavernous sinus (long arrow). A selective injection (B) into a dural branch of the middle meningeal artery (small arrow) delineates the fistula site (long arrow). The cast of the liquid adhesive (arrow in C) shows that there is good penetration through the fistula. The control right external carotid arteriogram (D) confirms that the fistula is closed. Abbreviation: DAVF, dural arteriovenous fistula.

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas

symptoms is expected, but symptoms typically recur over time. In cases where palliative arterial treatment is deemed to be insufficient, transvenous coil embolization of the diseased venous sinus is very effective in resolving symptoms and frequently results in total angiographic obliteration of the DAVF (Fig. 5). Sacrifice of the venous sinus should only be done when it exclusively drains the fistula and does not participate in the venous drainage of the brain. Targeted embolization of a compartment of the sinus may be effective in obliterating the fistula with preservation of that part of the venous sinus that is not participating in the fistula (Fig. 6) (48). If necessary, sacrifice of the sinus is performed by a transvenous approach. Only in exceptional cases, catheterization of the venous sinus can be reached using a transarterial approach, e.g., in traumatic DAVFs due to the large connection between the dural artery and the adjacent vein (Fig. 7). In most spontaneous DAVFs, the feeding arterial network is too small and tortuous to allow catheterization of the sinus from the arterial approach. Management of Aggressive DAVFs with Direct CVR

Aggressive cranial DAVFs have the potential to cause severe complications as a result of their natural course, thus mandating aggressive treatment. Treatment procedures aimed at the selective disconnection of CVR have been reported for DAVFs with direct CVR without dural sinus drainage (Borden type 3/Cognard type

341

III and IV). Pathophysiologically, a DAVF is considered a venous disease, and a permanent cure of a cranial DAVF with direct CVR can thus be obtained by a selective intradural division of the venous outlet of the fistula, analogous to the well-known treatment of a spinal DAVF. Primarily, this procedure has been described as a neurosurgical technique; however, with the advances in interventional neuroradiology, the same result has been demonstrated with endovascular means as well (49–53). A transvenous approach is the most likely endovascular technique to disconnect the CVR (Fig. 8). The transvenous approach requires a road-map technique from an arterial injection that demonstrates the fistula. Only in a small percentage of patients can transarterial embolization with liquid adhesives obliterate the fistula. This technique is performed with a wedged catheter and a liquid adhesive that has a long polymerization time. The arterial catheter must be wedged close to the fistula site to allow the slow push of the liquid adhesive through the fistula into the proximal venous outlet. Too proximal embolization allows persistent arterial shunting and recruitment of collateral flow; too distal embolization may result in venous occlusion and venous infarction. Particle embolization from the arterial approach may transiently reduce the flow through the fistula and should usually be used in high-flow, complex DAVFs in combination with a curative transvenous approach or with surgery.

Figure 5 Transvenous facial approach to bilateral cavernous DAVFs. A 52-year-old female presented with diplopia, bilateral chemosis, and bilateral conjunctival injection. Lateral right (A) and left (B) external carotid arteriograms show bilateral cavernous sinus DAVFs draining into both superior ophthalmic veins. Note the subtle CVR (arrow in A) on the right. Using a femoral venous approach, a left transfacial route (C) was used to reach the contralateral cavernous sinus. On the AP view, note the selective catheter position in the contralateral cavernous sinus (D). The right cavernous sinus venogram (E) during the coil embolization of the right cavernous sinus shows the reflux into cortical veins (arrow in E). Coil embolization of both cavernous sinuses was done (F). The control AP external right and left arteriograms (G) show that the DAVFs were closed. The patient’s orbital symptoms improved over time; however, her ophthalmoplegia had not resolved at six months. Abbreviations: DAVF, dural arteriovenous fistula; CVR, cortical venous reflux.

342

van Dijk and Willinsky

Figure 6 Targeted transvenous embolization of a cavernous DAVF. A 54-year-old male presented with chemosis, conjunctival injection, and proptosis in the left orbit. His raised intraocular pressure in the right eye could not be relieved with topical medication. Lateral left internal (A) and external (B) carotid arteriograms showed shunting into the left cavernous sinus and drainage into the superior ophthalmic vein and inferior petrosal sinus (arrow). Transvenous catheterization (C) was done via the inferior petrosal sinus. The site of the fistula was identified to be in the anterior lateral compartment of the cavernous sinus. Four Hydrocoils (MicroVention, Aliso Viejo, California, U.S.) were used to close the fistula site (arrow in D). Control left internal (E) and external (F) carotid arteriograms show closure of the fistula. The patient’s signs and symptoms resolved completely within a few weeks. Abbreviation: DAVF, dural arteriovenous fistula.

Management of Aggressive DAVFs with Dural Sinus Drainage and CVR

In case of combined dural sinus drainage and CVR (Borden type 2, Cognard type IIb or IIa+b), the obliteration of the whole fistula including excision or packing of the involved dural sinus has been advocated (54–57). On the other hand, the drawback of permanent occlusion of an involved sinus may be impairment of the venous drainage of the normal brain, resulting in (hemorrhagic) venous infarction and or leading to chronic complications of venous hypertension, e.g., dementia. In this perspective, Mironov (58) reported the treatment of two Borden type 2 cases, in which he endovascularly disconnected the CVR without changing the venous drainage of the dural sinus. In this way, the fistula itself was not obliterated, but instead converted into a benign DAVF (without CVR). This conversion is clinically important, since Davies and Satomi in their series demonstrated that benign DAVFs follow a disease course without any grave neurological events and that in the majority, benign DAVF is a self-limiting disease (31,33). The ‘‘converted’’ benign DAVFs have been demonstrated to follow the same benign clinical

disease course (59). The endovascular disconnection may be done either using a transarterial approach with liquid adhesives (Fig. 9) or a transvenous approach with coils (Figs. 10 and 11). The transvenous route is preferred because it is more likely to be successful compared with the injection of liquid adhesives from the arterial feeders. If the endovascular treatment fails to eliminate the CVR, then there are a number of surgical options. A burr hole can be placed over the diseased sinus, followed by a direct puncture to allow packing of the sinus with a microcatheter placed through a small sheath. The burr hole can be localized using a roadmap technique from an arterial injection. Packing of the sinus must be done under fluoroscopic guidance and control angiography can be done from the arterial side to determine when the fistula is closed or if the CVR has been eliminated. Alternatively, direct packing of the sinus can be done after surgical exposure of the entire sinus. Surgical resection of a sinus is often associated with high morbidity and mortality, even in experienced hands (60). However, the surgical technique of selective disconnection of CVR, leaving the actual fistula in the wall of the dural sinus untouched,

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas

343

Figure 7 Transarterial approach to venous packing of a traumatic DAVF. A 22-year-old male complained of a pulsatile tinnitus after trauma. The left lateral external carotid arteriogram (A) shows an osseous fistula draining toward the pterygoid plexus. The selective middle meningeal arteriogram (B) shows a high-flow fistula within the sphenoid bone draining into an extracranial venous pouch (arrow). Selective catheterization (C) of the venous pouch allowed deposition of coils into the pouch (arrow in D). Liquid adhesives were injected into the pouch from the fistula site after the coils had significantly reduced the flow (arrow in D). A control left external carotid arteriogram (E) shows that the fistula is closed. Abbreviation: DAVF, dural arteriovenous fistula.

Figure 8 Transvenous disconnection of a Borden type 3. A 62-year-old male presented with a brain stem hemorrhage (A). The left lateral internal carotid arteriogram (B) shows a DAVF fed by an inferior marginal tentorial artery (single arrow) draining into a cortical vein (double arrows). Note the venous pouch that likely represented a pseudoaneurysm at the site of the previous bleed. Using a transvenous approach catheterization of the venous pouch was feasible (C). Coils were deposited within the cortical vein and the venous pouch. A control internal carotid arteriogram (D) showed that the fistula was closed. A six-month control catheter angiogram confirmed the stability of the embolization. Abbreviation: DAVF, dural arteriovenous fistula.

344

van Dijk and Willinsky

Figure 9 Conversion of an aggressive Borden type 2 DAVF to a benign Borden type 1 DAVF by arterial embolization with liquid adhesives. A 72-year-old male presented with pulsatile tinnitus and a bruit 12 months after a motor vehicle accident. The lateral left external carotid arteriogram (A) shows shunting into an irregular transverse sinus and reflux into the vein of Labbe´ (arrow). Selective catheterization of the posterior branch of the left middle meningeal artery allowed deposition of the liquid adhesive (NBCA) (arrow in B) at the site of the CVR. Control occipital (C) and external carotid (D) arteriograms at the time of the embolization shows persistence of the shunting into the transverse sinus with elimination of the CVR. A repeat catheter angiogram two years later showed no change from the immediate post embolization study. Abbreviations: DAVF, dural arteriovenous fistula; NBCA, N-butyl cyanoacrylate; CVR, cortical venous reflux.

Figure 10 Conversion of an aggressive Borden type 2 DAVF to a benign Borden type 1 DAVF in the cavernous sinus by transvenous embolization with coils. A 77-year-old female presented with bilateral conjunctival injection and chemosis. The left external carotid arteriogram (A) shows a DAVF of the cavernous sinus draining into cortical veins (arrow) and the inferior petrosal sinus. Using a transvenous approach through the ipsilateral inferior petrosal sinus (small arrows in B), we were able to catheterize the sphenoparietal sinus (arrow). We packed the sphenoparietal sinus and the adjacent cavernous sinus with coils (C). A control left external arteriogram (D) shows that the CVR was eliminated and there was less flow through the fistula. On the venous phase of the lateral internal carotid arteriograms before (E) and after (F) treatment, we see that the cortical veins over the temporal lobe (arrows) could drain the brain after the cortical venous disconnection. The patient’s symptoms resolved and she has been stable for five years. Abbreviations: DAVF, dural arteriovenous fistula; CVR, cortical venous reflux.

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas

345

Figure 11 Conversion of an aggressive Borden type 2 DAVF to a benign Borden type 1 DAVF by transvenous embolization of a parallel channel. A 37-year-old female developed superior sagittal sinus thrombosis (arrow in A) one year after renal transplantation. Three years later, the patient developed a bruit and her MR showed prominent, tortuous flow voids (arrows) over the surface of the brain (B). Multiple intracranial DAVFs were found on a cerebral angiogram. The lateral left external arteriogram (C) shows a DAVF of the transverse sinus with CVR (arrows) and occlusion of the ipsilateral sigmoid sinus. A transvenous approach via the contralateral transverse sinus allowed selective catheterization of a parallel channel that connected to the cortical veins that participated in the reflux. Venography in this parallel channel (D) shows the veins that were draining the fistula. This parallel channel was embolized (arrow in E) with a combination of platinum coils and Hydrocoil (MicroVention). A control left external arteriogram (F) shows that the CVR was eliminated, although the fistula persists. On the venous phase of the lateral internal carotid arteriograms before (G) and after (H) treatment, we see that these cortical veins can participate in the venous drainage of the brain after disconnection. Abbreviations: DAVF, dural arteriovenous fistula; CVR, cortical venous reflux.

is relatively simple. Coagulation and division of the refluxing cortical veins as they enter the subarachnoid space is sufficient to convert the aggressive type DAVF into a benign one. In case of complex, highflow DAVFs, preoperative embolization with liquid adhesives or particles is helpful to reduce blood loss during surgery and may improve the results of a surgical disconnection or packing of the sinus. The permanent application of aneurysm clips has been increasingly avoided during the past decade because of local distortions of the magnetic field in MRI (59). Transvenous Approaches to the Cavernous Sinus: Special Considerations

There are a number of transvenous endovascular approaches to the cavernous sinus. From the femoral venous route, these include opacified ipsilateral or contralateral inferior petrosal sinus, unopacified ipsilateral or contralateral inferior petrosal sinus, superior petrosal sinus, pterygoid plexus, and facial vein. Using the contralateral inferior petrosal sinus, one needs to cross the midline through the basal venous plexus or the intracavernous venous sinus. Surgical exposure of the superior ophthalmic vein can allow transvenous navigation into the cavernous sinus. In

Meyers’ report of 117 DAVFs of the cavernous sinus, the transvenous access was achieved through the inferior petrosal sinus or superior ophthalmic vein in 76% of cases (61). Klisch reported that 60% of their cavernous DAVFs were treated transvenously through the inferior petrosal sinus, with complete obliteration of the fistula in 78% of cases. Intracranial surgical exposure of the superficial middle cerebral vein allows access to the cavernous sinus through the sphenoparietal sinus. On the basis of the venous drainage, Klisch divided the cavernous DAVFs into four compartments: anterior, posterior, lateral, and inferior (62). Recognition of the different types of cavernous DAVFs based on these compartments may help to plan the approach, and the treatment can be targeted to the involved compartment (Fig. 6) (48). Transvenous Approaches to the Transverse Sinus: Special Considerations

If the affected sinus is isolated from the circulation, then sinus occlusion is effective to close the fistula and eliminate the CVR. The risk of venous infarction is low since the cortical veins do not drain into the affected sinus. In some patients, the isolated sinus can be reached transvenously through an occluded segment.

346

van Dijk and Willinsky

If the transvenous route is unsuccessful, a burr hole or direct surgical exposure can allow packing of the isolated sinus. If the affected sinus drains both the fistula and the cortical veins, occlusion of the sinus carries a significant risk of venous infarction. It is now clear that a number of transverse/sigmoid sinus DAVFs have a parallel venous channel as the recipient compartment for the fistula that is distinct from the compartments draining the cortical veins. Superimposition of the arterial and venous phases of the angiogram is important to illustrate the distinct compartments. Recognition of the parallel channel allows selective transvenous embolization of that part of the sinus that participates in the CVR (Fig. 11). Caragine recognized such a parallel venous channel in 10 patients, in all of whom selective transvenous embolization with preservation of the transverse/sigmoid sinus was possible (63). Piske also highlighted the concept of the parallel venous channel and reported on the selective embolization of the affected compartment in the superior sagittal sinus with preservation of the sinus itself (45). The Role of Radiosurgery

Although reported occasionally, there is a limited role for the radiosurgical treatment of cranial DAVFs, with or without CVR (64–66). The benign cranial DAVFs either need no treatment or benefit from selective embolization of the involved venous sinus. The aggressive cranial DAVFs have a poor natural history, so the delayed effects offered by radiosurgery are not acceptable (42).

SPINAL DURAL AV FISTULAS The first description of spinal AVMs dates back to 1888 by Gaupp as a cluster of hemorrhoids on the pia mater of the spinal cord (67). Medical descriptions in those days were based on postmortem dissection or surgical exposure and then related to the observed clinical symptoms. In 1900, Brasch related a severe case of myelopathy to a serpentine-like transformation of the spinal cord vasculature (68). Krause performed the first documented surgical exploration of a spinal vascular disorder (with a poor result) in 1911, which was followed by Elsberg’s first curative surgical exposure in 1916 (69,70). Following the introduction of angiography techniques, radiological descriptions of spinal pathology became available. In spite of these primitive circumstances, Wyburn-Mason published a relatively large series (110 patients) and classified angiomas into venous and arteriovenous (71). Even after the introduction of the venous congestion of the spinal cord in the early 1970s, spinal DAVFs were categorized into the group of the spinal AVMs (72). This categorization led to the unfortunate assumption that the congested perimedullary venous plexus was a part of the nidus. It needs no explanation that the stripping of the venous plexus as a part of the surgical therapy therefore often gave rise to increased postoperative neurological deficits (73).

It was not until the late 1970s that understanding the anatomy and pathophysiology of spinal DAVFs led to major breakthroughs in treatment. Spinal DAVFs were recognized as extramedullary fistulatype lesions draining into the perimedullary venous plexus, features that distinguished them from the genuine spinal AVMs (74,75). Then, after the conceptual recognition of a single shunting vein to the perimedullary venous plexus by Oldfield and Symon, the characteristic clinical syndrome, and the surgical results of simply occluding the shunting vein, DAVFs emerged as a unique entity (76,77).

Spinal Classification Spinal vascular AV shunts are a heterogeneous group of congenital malformations and acquired fistulas. Therefore, several classification schemes have been proposed, each of which tries to regroup these separate lesions on the basis of different insights in pathology. The most commonly used is the following scheme with four subdivisions: type I, dural AVF; type II, intramedullary glomus; type III, juvenile; and type IV, perimedullary (78). However, it does not take an expert to realize that these categories are too static and that classification is often difficult. In 2002, Rodesch introduced a much more advanced classification of the intradural AV shunts based on the Biceˆtre experience, in which he differentiated the lesions on the basis of quantity (single/multiple), general aspect (nidus or fistula), associated metameric disorders, and myelomere location. Notably, however, in this arrangement the spinal DAVFs are excluded (79). Spetzler introduced a classification with a wider scope that subdivides the spinal vascular lesions in neoplastic vascular lesions, spinal aneurysms, spinal AVMs, and spinal AVFs. Within the spinal AVFs, three subgroups are identified: extradural AVFs, intradural ventral AVFs, and intradural dorsal AVFs (the latter group is essentially DAVFs). Spinal DAVFs embody the vast majority of all spinal vascular lesions in the adult population and are virtually absent in the pediatric age group. The acquired AV fistula is typically located at the thoracic or lumbar level (80). Remarkably, at the cervical level, DAVFs seem not to exist. This phenomenon has been explained by the existence of numerous venous outlets at the cervical level, in contrast to the thoracolumbar level, and by the influence of gravity facilitating the venous reflux to the vena cava system. Although there are reports concerning cervical locations that presented with subarachnoid hemorrhage or without myelopathy, these cases were either cranial DAVFs or extradural AVFs draining into the perimedullary venous plexus (81).

Spinal Clinical Features The clinical features of a spinal DAVF are directly related to the venous congestion of the spinal cord caused by a direct fistula from the radiculomeningeal artery to the perimedullary venous plexus. This leads

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas

to a clinical picture of chronic progressive myelopathy of the lower thoracic cord and conus, irrespective of the location of the shunt (82). Venous congestion leading to hypoxia and ischemia is accepted as the theory of choice. The histopathological findings of an occasional biopsy of the spinal cord support this theory (83). Compression of the spinal cord by the voluptuous venous plexus has been suggested, but is highly unusual. Spinal DAVFs typically present in the fourth to sixth decade, commonly in males. In a study combining the clinical findings in 172 patients from five series in the literature, leg weakness was the initial complaint in 40%; however, at the time of diagnosis, it was present in 88% of the cases. The same phenomenon is described in bladder and bowel dysfunction (5% vs. 85%). This discrepancy is explained by the extensive delay between the initial presentation and the actual diagnosis. Symon reported that only one-third of patients were diagnosed within one year of the onset of symptoms and only two-thirds of patients had their diagnosis within three years (84). Hemorrhage practically never occurs, although it is a major presenting sign of the cranial DAVFs with reflux in the leptomeningeal veins. This phenomenon is likely explained by the slow-flow characteristics of the spinal DAVFs (85). Intracranial DAVFs with perimedullary venous reflux may present with a cervical myelopathy or rarely even with a cervical intradural hemorrhage (86). In 2002, the Toronto database contained 49 patients with a spinal DAVF. The mean age was 63.2 years, with an 80% male predominance. The fistulas were located between the nerve roots of C7 and S1, with 94% situated below the T5 root level and in majority on the left side (70%). It hypothesized that the position of the heart is related to these findings. Multiplicity was encountered in one patient (2%). At the time of diagnosis, all but one patient (who complained of radicular pain) had signs and symptoms of myelopathy. The majority of the patients (96%) suffered from spastic paraparesis or leg weakness and 90% had sensory disturbances. Bladder and bowel malfunctions as well as pain (either back or remote pain) were other significant complaints. The time interval between the initial symptoms and diagnosis was median 10.5 months (87). In the natural disease course of spinal DAVFs, the thoracic myelopathy is gradually but inexorably progressive leading eventually to a state of paraplegia and incontinence. Eventually this process ends in the Foix-Alajouanine syndrome, originally described in 1929 as a subacute necrotizing myelopathy (88). To classify the severity of the thoracic myelopathy, the classification scale for motor and bladder function as used by Aminoff is frequently applied (Table 3) (89).

Spinal Diagnostic Imaging The current mainstay of diagnosing a spinal DAVF is MRI. Typically on T2 images, there is a hyperintense signal within a slightly swollen lower thoracic spinal cord and medullary conus, outlined by a peripheral

347

Table 3 Aminoff Score of Disability Classification of gait disturbance Grade 1 Grade Grade Grade Grade

2 3 4 5

Leg weakness or abnormal gait; no restricted activity Grade 1 with restricted activity Requires one stick or similar support for walking Requires two sticks or crutches for walking Unable to stand, confined to bed or wheelchair

Classification of micturition Grade 1 Hesitance, urgency, or frequency Grade 2 Occasional urinary incontinence or retention Grade 3 Total urinary incontinence or retention

hypointensity (90). Spin-echo techniques are best to reveal the vascular nature of the congestive myelopathy. Contrast enhancement with gadolinium can be useful in differentiating the spinal cord effects of DAVF from an intramedullary tumor, although in severe congestive myelopathy due to DAVF, patchy enhancement may be noted. Enhancement of the convoluted perimedullary venous plexus supports the diagnosis of a spinal DAVF, but the abnormal flow voids are nonspecific for the actual location of the fistula. To find the level of the fistula, enhanced ATECO-MRA has been demonstrated to be very effective (Fig. 12). This technique, published by Farb, proves very useful in determining the levels of interest, after which the commonly lengthy angiography session can be focused (91). Targeted superselective angiography can thus be limited to bilateral injections of the radicular arteries at the level of the fistula and the immediate adjacent levels above and below. Usually, the fistula is located at the level of a root sleeve, however sometimes it is remote. Especially in the latter case, a multisegmental arterial supply of the DAVF can be expected. The fistulous point is recognized on the angiogram by a caliber change between the artery and the dilated vein. The intradural vein is unpredictable in its location and direction. As a rule, it follows the course of the nerve root, which implies an upward track at the lumbar level, but at the thoracic level the track may as well be horizontal. For treatment purposes, visualization of the anterior spinal artery (Adamkiewicz) is essential and warrants additional injections. The Adamkiewicz artery often arises between T8 and T11; in 50% of the cases, from the intercostal artery at T9 or T10 on the left (92). Injection of the Adamkiewicz artery, then, reveals a prolonged circulation time within the spinal cord, which is indicative of venous congestion (93). In case ATECO-MRA does not reveal a fistula site, but the MRI and the clinical picture are highly suggestive of a spinal DAVF, a more extensive, traditional angiography is necessary, including selective injections of all intercostal and lumbar arteries as well as both internal iliac arteries. If no spinal DAVF is found subsequently, the cervical cord and intracranial circulation, including the selective injections of both vertebral arteries and the costocervical and thyrocervical arteries, should be studied with angiography. A study of the intracranial circulation should include the vertebral, internal, and external carotid arteries (86).

348

van Dijk and Willinsky

Figure 12 Gadolinium-enhanced MR expedites treatment of spinal DAVF. A 63-year-old male presents with a thoracic myelopathy. The T2-weighted MR of the spine (A) shows a central T2 hyperintensity within the cord and prominent posterior perimedullary veins (arrows). The gadolinium-enhanced MR angiogram (B) identifies the fistula site (arrow) of the spinal DAVF thereby facilitating the catheter angiogram (C) that is done to embolize the fistula. Abbreviation: DAVF, dural arteriovenous fistula.

Spinal Therapeutic Considerations All cases of the ominous natural disease course of spinal DAVFs should be treated by occlusion or division of the fistulous vein. Intradural surgical ligation has proven very effective with durable results (76,94). Surgical interruption of the draining radicular vein is preferentially done using bipolar coagulation. Excision of the whole fistula site is not necessary and may cause CSF leakage or nerve sacrifice. In case of both intradural and extradural venous drainage, surgical interruption of the intradural vein will treat the congestive myelopathy; the extradural drainage can be left alone (95). Endovascular techniques using diluted liquid adhesive (e.g., N-butyl cyanoacrylate, or NBCA) have also proven effective with durable results (96,97). Particle embolization is not recommended because the incidence of symptomatic recurrence is high (98). It is essential that the embolic material reaches the proximal part of the draining vein; otherwise, collateral dural branches may keep the fistula open and recurrences may occur despite what was thought to be an ‘‘adequate’’ treatment (99). The collateral flow may not be evident at the time of embolization, and therefore patients need close follow-up. Patients may improve initially after the embolization, but will experience delayed deterioration following partial treatment. Although it is generally accepted that surgery and endovascular therapy are equally effective, a recent meta-analysis suggests that a surgical approach is preferable (100). On the other hand, advantages of endovascular embolization include the facts that it is less invasive and that it has immediate angiographic control of the treatment. Moreover, by using ATECOMRI the diagnostic angiography can even be combined with the treatment. Therefore, initial attempt at endovascular embolization with liquid adhesive

material can be justified, but should not be undertaken if catheterization of the branch to the feeding pedicle proves to be difficult on the diagnostic angiogram or in cases where there is a common segmental origin of the anterior spinal artery and the radiculomeningeal branch supplying the fistula (101). In case of an incomplete closure of the DAVF or abstention from endovascular embolization, surgical therapy should be instituted shortly after. It is worthwhile to go for a curative treatment. Using the Aminoff score, Steinmetz in his meta-analysis reports that following both microsurgery and embolization for gait disturbances, there is a 55% chance of improving and an 11% chance of being worse. Therefore, patients have an 89% chance of improving or stabilizing. The results for improvement in bladder function were less favorable. Only 33% of patients demonstrated an improvement in micturition, whereas 11% worsened (100). If aggravation of the symptoms occurs after treatment and there is angiographic proof of a cure and no multiplicity, then the clinical deterioration may be due to venous thrombosis of the perimedullary veins. Such deterioration usually occurs within days of the treatment and may warrant anticoagulation therapy. For this reason, some authors propose heparin for three days following embolization for all patients with a significantly compromised venous drainage (99).

REFERENCES 1. Newton TH, Cronqvist S. Involvement of dural arteries in intracranial arteriovenous malformations. Radiology 1969; 93(5):1071–1078. 2. Al Shahi R, Bhattacharya JJ, Currie DG, et al. Prospective, population-based detection of intracranial vascular malformations in adults: the Scottish Intracranial Vascular Malformation Study (SIVMS). Stroke 2003. 34(5):1163–1169.

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas 3. Herman JM, Spetzler RF, Bederson JB, et al. Genesis of a dural arteriovenous malformation in a rat model. J Neurosurg 1995; 83(3):539–545. 4. Houser OW, Campbell JK, Campbell RJ, et al. Arteriovenous malformation affecting the transverse dural venous sinus–an acquired lesion. Mayo Clin Proc 1979; 54(10):651–661. 5. Chaudhary MY, Sachdev VP, Cho SH, et al. Dural arteriovenous malformation of the major venous sinuses: an acquired lesion. AJNR Am J Neuroradiol 1982; 3(1):13–19. 6. Terada T, Higashida RT, Halbach VV, et al. Development of acquired arteriovenous fistulas in rats due to venous hypertension. J Neurosurg 1994; 80(5):884–889. 7. Lasjaunias P. A revised concept of the congenital nature of cerebral arteriovenous malformations. Intervent Neuroradiol 1997; 3:275–281. 8. Kerber CW, Newton TH. The macro and microvasculature of the dura mater. Neuroradiology 1973; 6(4):175–179. 9. Piton J, Guilleux MH, Guibert-Tranier F, et al. Fistulae of the lateral sinus. J Neuroradiol 1984; 11(3):143–159. 10. Roland J, Bernard C, Bracard S, et al. Microvascularization of the intracranial dura mater. Surg Radiol Anat 1987; 9(1):43–49. 11. Uranishi R, Nakase H, Sakaki T. Expression of angiogenic growth factors in dural arteriovenous fistula. J Neurosurg 1999; 91(5):781–786. 12. Rothbart D, Awad IA, Lee J, et al. Expression of angiogenic factors and structural proteins in central nervous system vascular malformations. Neurosurgery 1996; 38(5):915–924. 13. Nishijima M, Takaku A, Endo S, et al. Etiological evaluation of dural arteriovenous malformations of the lateral and sigmoid sinuses based on histopathological examinations. J Neurosurg 1992; 76(4):600–606. 14. Hamada Y, Goto K, Inoue T, et al. Histopathological aspects of dural arteriovenous fistulas in the transversesigmoid sinus region in nine patients. Neurosurgery 1997; 40(3):452–456. 15. McCutcheon IE, Doppman JL, Oldfield EH. Microvascular anatomy of dural arteriovenous abnormalities of the spine: a microangiographic study. J Neurosurg 1996; 84(2):215–220. 16. To¨nnis W. Aneurysma arteriovenosum. In: Bergstrand H, Olivecrona H, To¨nnis E, eds. Gefa¨ßmißbildungen und Gefa¨ ßgeschwu¨ lste des Gehirns. Leipzig: Thieme, 1936:88–134. 17. Fincher EF. Arteriovenous fistula between the middle meningeal artery and the greater petrosal sinus. Ann Surg 1951; 133:886–888. 18. Moniz E. L’ encephalographie arterielle, son importance dans la localization des tumeurs cerebrales. Rev Neurol (Paris) 1927; 2:72–90. 19. Newton TH, Greitz T. Arteriovenous communication between the occipital artery and transverse sinus. Radiology 1966; 87(5):824–828. 20. Houser OW, Baker HL Jr., Rhoton AL Jr., et al. Intracranial dural arteriovenous malformations. Radiology 1972. 105(1):55–64. 21. Castaigne P, Bories J, Brunet P, et al. [Meningeal arteriovenous fistulas with cortical venous drainage]. Rev Neurol (Paris) 1976; 132(3):169–181. 22. Djindjian R, Merland JJ, Theron J. Superselective arteriography of the external carotid artery. New York: SpringerVerlag, 1977. 23. Malik GM, Pearce JE, Ausman JI, et al. Dural arteriovenous malformations and intracranial hemorrhage. Neurosurgery 1984; 15(3):332–339. 24. Lasjaunias P, Chiu M, TerBrugge K, et al. Neurological manifestations of intracranial dural arteriovenous malformations. J Neurosurg 1986; 64(5):724–730.

349

25. Awad IA, Little JR, Akarawi WP, et al. Intracranial dural arteriovenous malformations: factors predisposing to an aggressive neurological course. J Neurosurg 1990; 72(6):839–850. 26. Aminoff MJ. Vascular anomalies in the intracranial dura mater. Brain 1973; 96(3):601–612. 27. Borden JA, Wu JK, Shucart WA. A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment. J Neurosurg 1995; 82(2):166–179. 28. Cognard C, Gobin YP, Pierot L, et al. Cerebral dural arteriovenous fistulas: clinical and angiographic correlation with a revised classification of venous drainage. Radiology 1995; 194(3):671–680. 29. Davies MA, TerBrugge K, Willinsky R, et al. The validity of classification for the clinical presentation of intracranial dural arteriovenous fistulas. J Neurosurg 1996; 85(5):830–837. 30. Fermand M, Reizine D, Melki JP, et al. Long term followup of 43 pure dural arteriovenous fistulae (AVF) of the lateral sinus. Neuroradiology 1987; 29(4):348–353. 31. Davies MA, Saleh J, TerBrugge K, et al. The natural history and management of intracranial dural arteriovenous fistulae. Part 1: Benign lesions. Intervent Neuroradiol 1997; 3:295–302. 32. Cognard C, Houdart E, Casasco A, et al. Long-term changes in intracranial dural arteriovenous fistulae leading to worsening in the type of venous drainage. Neuroradiology 1997; 39(1):59–66. 33. Satomi J, Van Dijk JM, TerBrugge KG, et al. Benign cranial dural arteriovenous fistulas: outcome of conservative management based on the natural history of the lesion. J Neurosurg 2002; 97(4):767–770. 34. Hurst RW, Bagley LJ, Galetta S, et al. Dementia resulting from dural arteriovenous fistulas: the pathologic findings of venous hypertensive encephalopathy. AJNR Am J Neuroradiol 1998; 19(7):1267–1273. 35. Willinsky R, TerBrugge K, Montanera W, et al. Venous congestion: an MR finding in dural arteriovenous malformations with cortical venous drainage. AJNR Am J Neuroradiol 1994; 15(8):1501–1507. 36. Willinsky R, Goyal M, TerBrugge K, et al. Tortuous, engorged pial veins in intracranial dural arteriovenous fistulas: correlations with presentation, location, and MR findings in 122 patients. AJNR Am J Neuroradiol 1999; 20(6):1031–1036. 37. Gobin YP, Rogopoulos A, Aymard A, et al. Endovascular treatment of intracranial dural arteriovenous fistulas with spinal perimedullary venous drainage. J Neurosurg 1992; 77(5):718–723. 38. Ricolfi F, Manelfe C, Meder JF, et al. Intracranial dural arteriovenous fistulae with perimedullary venous drainage. Anatomical, clinical and therapeutic considerations. Neuroradiology 1999; 41(11):803–812. 39. Brown RD Jr., Wiebers DO, Nichols DA. Intracranial dural arteriovenous fistulae: angiographic predictors of intracranial hemorrhage and clinical outcome in nonsurgical patients. J Neurosurg 1994; 81(4):531–538. 40. Davies MA, TerBrugge K, Willinsky R, et al. The natural history and management of intracranial dural arteriovenous fistulae. Part 2: Aggressive lesions. Intervent Neuroradiol 1997; 3:303–311. 41. Duffau H, Lopes M, Janosevic V, et al. Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature. J Neurosurg 1999; 90(1):78–84. 42. Van Dijk JM, TerBrugge KG, Willinsky RA, et al. Clinical course of cranial dural arteriovenous fistulas with long-term persistent cortical venous reflux. Stroke 2002; 33(5):1233–1236.

350

van Dijk and Willinsky

43. Lee SK, Willinsky RA, Montanera W, et al. MR imaging of dural arteriovenous fistulas draining into cerebellar cortical veins. AJNR Am J Neuroradiol 2003; 24(8):1602–1606. 44. Farb RI, McGregor C, Kim JK, et al. Intracranial arteriovenous malformations: real-time auto-triggered elliptic centric-ordered 3D gadolinium-enhanced MR angiography–initial assessment. Radiology 2001; 220(1):244–251. 45. Piske RL, Campos CM, Chaves JB, et al. Dural sinus compartment in dural arteriovenous shunts: a new angioarchitectural feature allowing superselective transvenous dural sinus occlusion treatment. AJNR Am J Neuroradiol 2005; 26(7):1715–1722. 46. Van Dijk JM, TerBrugge KG, Willinsky RA, et al. Multiplicity of dural arteriovenous fistulas. J Neurosurg 2002; 96(1):76–78. 47. Van Dijk JM, Willinsky RA. Venous congestive encephalopathy related to cranial dural arteriovenous fistulas. Neuroimaging Clin N Am 2003; 13(1):55–72. 48. Agid R, Willinsky RA, Haw C, et al. Targeted compartmental embolization of cavernous sinus dural arteriovenous fistulae using transfemoral medial and lateral facial vein approaches. Neuroradiology 2004; 46(2):156–160. 49. Collice M, D’Aliberti G, Talamonti G, et al. Surgical interruption of leptomeningeal drainage as treatment for intracranial dural arteriovenous fistulas without dural sinus drainage. J Neurosurg 1996; 84(5):810–817. 50. Collice M, D’Aliberti G, Arena O, et al. Surgical treatment of intracranial dural arteriovenous fistulae: role of venous drainage. Neurosurgery 2000; 47(1):56–66. 51. Goyal M, TerBrugge K, Farb RI. Endovascular retrograde cortical venous approach to disconnect retrograde leptomeningeal venous reflux in a patient with a dural AVF. Intervent Neuroradiol 1999; 5:195–198. 52. Hoh BL, Choudhri TF, Connolly ES Jr., et al. Surgical management of high-grade intracranial dural arteriovenous fistulas: leptomeningeal venous disruption without nidus excision. Neurosurgery 1998; 42(4):796–804. 53. Thompson BG, Doppman JL, Oldfield EH. Treatment of cranial dural arteriovenous fistulae by interruption of leptomeningeal venous drainage. J Neurosurg 1994; 80(4):617–623. 54. Roy D, Raymond J. The role of transvenous embolization in the treatment of intracranial dural arteriovenous fistulas. Neurosurgery 1997; 40(6):1133–1141. 55. Quinones D, Duckwiler G, Gobin PY, et al. Embolization of dural cavernous fistulas via superior ophthalmic vein approach. AJNR Am J Neuroradiol 1997; 18(5):921–928. 56. Endo S, Kuwayama N, Takaku A, et al. Direct packing of the isolated sinus in patients with dural arteriovenous fistulas of the transverse-sigmoid sinus. J Neurosurg 1998; 88(3):449–456. 57. Mullan S, Johnson DL. Combined sagittal and lateral sinus dural fistulae occlusion. J Neurosurg 1995; 82(2):159–165. 58. Mironov A. Selective transvenous embolization of dural fistulas without occlusion of the dural sinus. AJNR Am J Neuroradiol 1998; 19(2):389–391. 59. Van Dijk JM, TerBrugge KG, Willinsky RA, et al. Selective disconnection of cortical venous reflux as treatment for cranial dural arteriovenous fistulas. J Neurosurg 2004; 101(1):31–35. 60. Sundt TM Jr., Piepgras DG. The surgical approach to arteriovenous malformations of the lateral and sigmoid dural sinuses. J Neurosurg 1983; 59(1):32–39. 61. Meyers PM, Halbach VV, Dowd CF, et al. Dural carotid cavernous fistula: definitive endovascular management and long-term follow-up. Am J Ophthalmol 2002; 134(1):85–92. 62. Klisch J, Huppertz HJ, Spetzger U, et al. Transvenous treatment of carotid cavernous and dural arteriovenous

63.

64.

65.

66.

67. 68.

69. 70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

fistulae: results for 31 patients and review of the literature. Neurosurgery 2003; 53(4):836–856. Caragine LP, Halbach VV, Dowd CF, et al. Parallel venous channel as the recipient pouch in transverse/sigmoid sinus dural fistulae. Neurosurgery 2003; 53(6):1261–1266. Guo WY, Pan DH, Wu HM, et al. Radiosurgery as a treatment alternative for dural arteriovenous fistulas of the cavernous sinus. AJNR Am J Neuroradiol 1998; 19(6):1081–1087. Shin M, Kurita H, Tago M, et al. Stereotactic radiosurgery for tentorial dural arteriovenous fistulae draining into the vein of Galen: report of two cases. Neurosurgery 2000; 46(3):730–733. Pollock BE, Nichols DA, Garrity JA, et al. Stereotactic radiosurgery and particulate embolization for cavernous sinus dural arteriovenous fistulae. Neurosurgery 1999; 45(3):459–466. Gaupp J. Ha¨morrhoiden der Pia mater spinalis im Gebiet des Lendenmarks. Beitr Pathol 1888; 2:516–518. ¨ ber einen schweren spinalen SymptomencomBrasch F. U plex bedingt durch einen Aneuryma-Serpentinumartige Vera¨nderung eines Teils der Ru¨ckenmarksgefa¨sse. Berl Klin Wochenschr 1900; 37:1210–1213. Krause F. Chirurgie des Gehirns und Ru¨ckenmarks nach eigenen Erfahrungen. Berlin: Urban & Schwarzenberg, 1911. Elsberg CA. Diagnosis and treatment of surgical diseases of the spinal cord and its membranes. London: WB Saunders, 1916:194–204. Wyburn-Mason R. The vascular abnormalities and tumors of the spinal cord and its membranes. London: H. Klimpton, 1943:3–5. Aminoff MJ, Barnard RO, Logue V. The pathophysiology of spinal vascular malformations. J Neurol Sci 1974; 23(2):255–263. Yasargil MG, DeLong WB, Guarnaschelli JJ. Complete microsurgical excision of cervical extramedullary and intramedullary vascular malformations. Surg Neurol 1975; 4(2):211–224. Kendall BE, Logue V. Spinal epidural angiomatous malformations draining into intrathecal veins. Neuroradiology 1977; 13(4):181–189. Merland JJ, Riche MC, Chiras J. Intraspinal extramedullary arteriovenous fistulae draining into the medullary veins. J Neuroradiol 1980; 7(4):271–320. Oldfield EH, Di Chiro G, Quindlen EA, et al. Successful treatment of a group of spinal cord arteriovenous malformations by interruption of dural fistula. J Neurosurg 1983; 59(6):1019–1030. Symon L, Kuyama H, Kendall B. Dural arteriovenous malformations of the spine. Clinical features and surgical results in 55 cases. J Neurosurg 1984; 60(2):238–247. Anson JA, Spetzler RF. Spinal dural arteriovenous malformations. In: Awad IA, Barrow DL, eds. Dural Arteriovenous Malformations. Park Ridge, IL: American Association of Neurological Surgeons, 1993:175–191. Rodesch G, Hurth M, Alvarez H, et al. Classification of spinal cord arteriovenous shunts: proposal for a reappraisal-the Bicetre experience with 155 consecutive patients treated between 1981 and 1999. Neurosurgery 2002; 51(2):374–380. Willinsky RA, TerBrugge K, Montanera W, et al. Posttreatment MR findings in spinal dural arteriovenous malformations. AJNR Am J Neuroradiol 1995; 16(10): 2063–2071. Aviv RI, Shad A, Tomlinson G, et al. Cervical dural arteriovenous fistulae manifesting as subarachnoid hemorrhage: report of two cases and literature review. AJNR Am J Neuroradiol 2004; 25(5):854–858. Aminoff MJ, Logue V. Clinical features of spinal vascular malformations. Brain 1974; 97(1):197–210.

Chapter 19: Endovascular Management of Dural Arteriovenous Fistulas 83. Hurst RW, Kenyon LC, Lavi E, et al. Spinal dural arteriovenous fistula: the pathology of venous hypertensive myelopathy. Neurology 1995; 45(7):1309–1313. 84. Berenstein A, Lasjaunias P. Endovascular treatment of spine and spinal cord lesions. In: Surgical Neuroangiography. Berlin: Springer-Verlag, 1992:1–85. 85. Berenstein A, Lasjaunias P, TerBrugge K. Spinal dural arteriovenous fistulae. In: Surgical Neuroangiography. Berlin: Springer-Verlag, 2003:849–872. 86. Andersson T, Van Dijk JM, Willinsky RA. Venous manifestations of spinal arteriovenous fistulas. Neuroimaging Clin N Am 2003; 13(1):73–93. 87. Van Dijk JM, TerBrugge KG, Willinsky RA, et al. Multidisciplinary management of spinal dural arteriovenous fistulas: clinical presentation and long-term follow-up in 49 patients. Stroke 2002; 33(6):1578–1583. 88. Foix C, Alajouanine T. La myelite ne´crotique subaigue. Myelite centrale angio-hypertrophique a e´volution progressive. Paraple´gie amyotrophique lentement accendante, d’abord spasmodique, puis 15 flasque, s’accompagnant de dissociation albumino-cytologique. Rev Neurol (Paris) 1926; 33:1–42. 89. Aminoff MJ, Logue V. The prognosis of patients with spinal vascular malformations. Brain 1974; 97(1):211–218. 90. Hurst RW, Grossman RI. Peripheral spinal cord hypointensity on T2-weighted MR images: a reliable imaging sign of venous hypertensive myelopathy. AJNR Am J Neuroradiol 2000; 21(4):781–786. 91. Farb RI, Kim JK, Willinsky RA, et al. Spinal dural arteriovenous fistula localization with a technique of first-pass gadolinium-enhanced MR angiography: initial experience. Radiology 2002; 222(3):843–850. 92. Djindjian R. Arteriography of the spinal cord. Am J Roentgenol Radium Ther Nucl Med 1969; 107(3):461–478.

351

93. Willinsky R, Lasjaunias P, TerBrugge K, et al. Angiography in the investigation of spinal dural arteriovenous fistula. A protocol with application of the venous phase. Neuroradiology 1990; 32(2):114–116. 94. Afshar JK, Doppman JL, Oldfield EH. Surgical interruption of intradural draining vein as curative treatment of spinal dural arteriovenous fistulas. J Neurosurg 1995; 82(2):196–200. 95. Rosenblum B, Oldfield EH, Doppman JL, et al. Spinal arteriovenous malformations: a comparison of dural arteriovenous fistulas and intradural AVM’s in 81 patients. J Neurosurg 1987; 67(6):795–802. 96. Song JK, Vinuela F, Gobin YP, et al. Surgical and endovascular treatment of spinal dural arteriovenous fistulas: long-term disability assessment and prognostic factors. J Neurosurg 2001; 94(suppl 2):199–204. 97. Song JK, Gobin YP, Duckwiler GR, et al. N-Butyl 2-cyanoacrylate embolization of spinal dural arteriovenous fistulae. AJNR Am J Neuroradiol 2001; 22(1):40–47. 98. Nichols DA, Rufenacht DA, Jack CR Jr., et al. Embolization of spinal dural arteriovenous fistula with polyvinyl alcohol particles: experience in 14 patients. AJNR Am J Neuroradiol 1992; 13(3):933–940. 99. Niimi Y, Berenstein A, Setton A, et al. Embolization of spinal dural arteriovenous fistulae: results and follow-up. Neurosurgery 1997; 40(4):675–682. 100. Steinmetz MP, Chow MM, Krishnaney AA, et al. Outcome after the treatment of spinal dural arteriovenous fistulae: a contemporary single-institution series and meta-analysis. Neurosurgery 2004; 55(1):77–87. 101. Aggarwal S, Willinsky R, Montanera W, et al. Superselective angiography of a spinal dural arteriovenous fistula having a common segmental origin with the artery of Adamkiewicz. Neuroradiology 1992; 34(4):352–354.

20 Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas Nicholas J. Patronas Department of Radiology, National Institutes of Health Clinical Center, Bethesda, Maryland, U.S.A.

Donald L. Miller Department of Radiology, National Naval Medical Center and Department of Radiology, Uniformed Services University of Health Sciences, Bethesda, Maryland, U.S.A.

INTRODUCTION Inferior petrosal sinus sampling (IPSS) is used in the differential diagnosis of Cushing’s disease and when there is a strong clinical suspicion of a hyperfunctioning pituitary adenoma and noninvasive methods have failed to establish the diagnosis. These adenomas may secrete prolactin, adrenocorticotropin (ACTH), growth hormone (GH), or thyrotropin (TSH). The initial application of IPSS was the evaluation of patients with Cushing’s syndrome. This syndrome remains the main indication for its use. Cushing’s syndrome is a clinically recognizable entity. It is characterized by a variety of symptoms, the most important of which include hypertension, diabetes mellitus, weight gain with central obesity, moon faces, purple abdominal striae, hirsutism, hyperpigmentation, and osteoporosis. The common denominator of Cushing’s syndrome is hypercortisolemia. The causes of the syndrome may be classified as ACTHdependent or ACTH-independent. Simply put, patients with ACTH-dependent Cushing’s syndrome have elevated levels of ACTH and cortisol, while patients with ACTH-independent Cushing’s syndrome have elevated levels of cortisol alone. ‘‘Elevated’’ is relative; patients with hypercortisolism have low or undetectable levels of ACTH due to suppression of both corticotrophin-releasing hormone (CRH) and ACTH production by the normal feedback loops in the hypothalamic-pituitary-adrenal axis. ‘‘Normal’’ levels of ACTH are abnormal in patients with Cushing’s syndrome and indicate an ACTH-dependent cause. Overproduction of cortisol alone is typically due to an adrenal lesion. Approximately 5% of patients with Cushing’s syndrome have an adrenal lesion that demonstrates autonomous function—it does not require stimulation by ACTH to produce cortisol. These non-ACTH-dependent lesions include hyperfunctioning adrenal adenomas, adrenocortical

carcinomas, primary pigmented nodular adrenal disease (PPNAD), and macronodular hyperplasia of the adrenals. ACTH-independent Cushing’s syndrome can also occur as a result of exogenous steroid administration. There is no role for IPSS in patients with ACTH-independent Cushing’s syndrome. The remaining patients with Cushing’s syndrome have an ACTH-dependent cause. Approximately 80% of these patients have an ACTH-secreting pituitary adenoma. This etiology (and only this specific etiology) is referred to as Cushing’s disease. An additional 15% of patients with Cushing’s syndrome have an ACTHsecreting tumor at a site other than the pituitary gland. Most of these patients have an obvious primary malignancy with ectopic hormone production, typically in the lung. Some have a small, clinically occult tumor and present with what is termed the occult ectopic ACTH syndrome. Most commonly, these small tumors are found in the bronchial tree, but localization of these lesions can be extremely difficult. Both depression and alcoholism can cause elevated CRH levels and present as ‘‘pseudo-Cushing’s syndrome.’’ Very rarely, CRHsecreting tumors are responsible for ACTH-dependent Cushing’s syndrome. The endocrinologist is responsible for determining whether the patient has the ACTH-dependent or the ACTH-independent form of Cushing’s syndrome. Patients with ACTH-independent disease require adrenal imaging, but not pituitary imaging. Patients with ACTH-dependent disease require further endocrinologic evaluation to determine whether the ACTH is from a pituitary source (Cushing’s disease) or an ectopic source of ACTH. A number of biochemical tests have been developed to aid in this effort. These include suppression tests with dexamethasone and stimulation tests with CRH. The details of biochemical testing are outside the scope of this chapter. Briefly, these tests rely on differences between

354

Patronas and Miller

pituitary adenomas and ectopic tumors. In general, pituitary tumors retain some capacity to demonstrate suppression of ACTH in response to high levels of exogenous steroids and some ability to demonstrate stimulation of ACTH in response to exogenous CRH, while ectopic tumors do not respond. Unfortunately, none of these biochemical tests is 100% sensitive and 100% specific. Magnetic resonance imaging (MRI) of the pituitary gland has become a routine test for evaluating patients with ACTH-dependent Cushing’s syndrome. MRI has proven useful not only for establishing the presence of an adenoma but also for demonstrating its location within the gland. This information is of paramount importance for surgical planning, since preservation of pituitary function after successful resection of the adenoma is a cardinal objective of the operation. The sensitivity of MRI in detecting pituitary adenomas primarily depends on tumor size. The sensitivity to ACTH-secreting adenomas has been reported to be as low as 45% in postcontrast scans. In other type of adenomas the sensitivity is considerably higher, since they become clinically apparent when larger in size. One of the problems with pituitary MRI is the absence of uniformity in the imaging protocols used at various centers. One cause of false-negative results is the use of suboptimal imaging techniques. In the past we routinely used the following imaging parameters: field strength 1.5 tesla, repetition time/echo time 400/9 msec, 192  256 matrix, two excitations, 12-cm field of view in the coronal and 16 cm in the sagittal plane, and 3-mm-thick sections without gap. More recently we have used a gradient-echo T1-weighted technique before and after contrast material administration [0.01-mmol/kg gadopentetate dimeglumine (Magnevist, Berlex Laboratories, Inc., Montville, New Jersey, U.S.)], with repetition time/echo time 9.6/ 2.3 msec, a 208 flip angle, 160  256 matrix, 6 excitations, and 1.5-mm slice thickness. With this technique we can exploit the superior contrast resolution of the gradient-echo technique and acquire thinner tomographic sections. Difficulties in demonstrating ACTH-secreting pituitary adenomas have several other causes that are less amenable to correction. First, these tumors are usually very small when patients first present. The spatial resolution limitations of current MRI scanners may cause them to be obscured by averaging artifacts. Second, pituitary adenomas often enhance in a fashion similar to normal pituitary parenchyma. Finally, detection of functioning pituitary adenomas is also confounded by the fact that identical-appearing, small focal space–occupying lesions can be encountered within the pituitary parenchyma of normal subjects. Autopsy studies and MR scans of normal volunteers have demonstrated that there is a 6% to 10% incidence of nonfunctioning adenomas (incidentalomas) in the pituitary gland (1,2). Neither biochemical tests nor imaging studies can provide an accurate diagnosis in all patients with Cushing’s syndrome, and an additional diagnostic method is sometimes required. In these patients,

petrosal sinus sampling can be used to confirm or exclude the presence of a functioning pituitary adenoma. It is most useful when the results of biochemical tests and MRI are discrepant. In addition, when biochemical tests provide a firm indication of Cushing’s disease but no lesion is identified on pituitary MRI, IPSS may provide lateralization of the pituitary adenoma to one side of the pituitary gland. This lateralization permits the surgeon to perform a hemihypophysectomy and preserve pituitary function.

INDICATIONS FOR PETROSAL SINUS SAMPLING Petrosal sinus sampling is performed in patients with a confirmed endocrine diagnosis of ACTH-dependent Cushing’s syndrome and one of the following: l l

l

l

l

l

Absence of a discrete pituitary lesion on MRI Equivocal biochemical tests in the presence of a discrete pituitary lesion on MRI Persistent Cushing’s syndrome after transsphenoidal surgery Clinical need to resolve other discrepancies between clinical, biochemical, and imaging tests A hyperfunctioning pituitary adenoma (acromegaly, thyrotropin-secreting hormone overproduction): Consideration of surgical resection and the results of pituitary MRI being negative

When petrosal sinus sampling is performed in patients with Cushing’s syndrome for these indications, the procedure has a sensitivity of 92% and a specificity of 90%, and lateralization provided by IPSS is correct in 70% of patients (3). Prior to sampling, 15 ‘‘lavender-top’’ tubes [Vacutainer, no.6457, with ethylenediaminetetraacetic acid (EDTA K3); Becton Dickinson, Rutherford, New Jersey, U.S.] are labeled and placed in an ice-water bath. Bilateral femoral vein puncture is performed under local anesthesia with ultrasound guidance using a micropuncture system. A sheath is placed in each femoral vein through which a 4-French (Fr) catheter is introduced. The sampling catheters are preshaped over steam to form a 758 bend for the left side and a 958 for the right. Alternatively, preshaped vertebral catheters with no side holes may be used. At least one of the femoral vein sheaths should be 1 Fr larger than the catheter used. This sheath is used to draw the peripheral vein samples that are obtained as part of the sampling procedure. A coaxial technique using a microcatheter to catheterize the inferior petrosal sinus (IPS) has become increasingly popular (3,4). This technique includes a 5-Fr or 6-Fr introducer catheter advanced into the internal jugular vein and a Target-10 or a Tracker-18 or Tracker-25 microcatheter (Target Therapeutics, Freemont, California, U.S.) and a Seeker 10 or a Seeker 16 wire (Target Therapeutics) for the selective catheterization of the IPS. Prior to introduction of any catheter into the petrosal sinus, a bolus of 3000 to 4000 IU of heparin is administered intravenously. In addition 5000 IU of heparin is added to the flush solution used to irrigate the petrosal sinus catheters

Chapter 20: Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas

and the femoral vein sheaths. The use of sedation is not recommended, but during the procedure intravenous midazolam and fentanyl may be used as needed. The right femoral vein introducer catheter is advanced into the right internal jugular vein, usually without difficulty, along a straight line through the inferior vena cava, the right atrium, and the superior vena cava. Advancement of the left femoral introducer catheter into the left internal jugular vein is usually more problematic, since the catheter must first turn by 908 from the superior vena cava into the left innominate vein and then turn by another 908 into the left internal jugular vein. In addition, a valve is located at the base of the left internal jugular vein, at its junction with the innominate vein. The valve can usually be successfully negotiated by positioning the tip of the catheter at the base of the internal jugular vein and advancing a flexible guidewire into this vessel during expiration or a reverse Valsalva maneuver. From the internal jugular vein, selective catheterization of the IPS is accomplished by rotating the introducer catheter medially and anteriorly as it is moved from the dome of the jugular bulb downward. This procedure is performed while injecting contrast to opacify the venous channels draining into this vessel. It is not uncommon to identify more than one such vessel. The tip of the introducer catheter is then anchored at the orifice of the most prominent venous channel (Fig. 1). Further advancement of the catheter into this vessel should be attempted only over a flexible guidewire coated with hydrophilic material (Glidewire, catalogue no. 46-151, Medi-tech/Boston Scientific, Natick, Massachusetts, U.S.). The guidewire should never be advanced into the cavernous sinus, and the catheter should never be advanced more than 1 to 1.5 cm into the IPS. Successful catheterization of the IPS is documented fluoroscopically in the anteriorposterior projection during gentle hand injection of contrast (Fig. 2). Digital subtraction angiography should also be obtained at this time, and opacification

Figure 1 IPS venogram in a patient with Cushing’s disease. (A) Injection in the right and (B) in the left IPS. Note asymmetry of the IPS with the left being smaller than the right. Abbreviation: IPS, inferior petrosal sinus.

355

Figure 2 Right IPS injection. There is good opacification of cavernous sinuses and retrograde flow into the left IPS. Arrow heads point to the tip of the catheters on both sides. Abbreviation: IPS, inferior petrosal sinus.

of the ipsilateral cavernous sinus as well as the opposite cavernous and IPSs documented; unless there is documentation of proper positioning of the catheter into the IPS, the results of venous sampling become questionable (Fig. 3). If one IPS is difficult to identify, it is usually advisable to abandon the attempt and switch to the other side, which may be easier to catheterize. The road map obtained from the contralateral catheterization can then be used to assist catheterization of the more difficult side (Figs. 4 and 5). Successful catheterization of the IPS requires familiarity with the anatomic variations that can be encountered at the junction of the IPS and the internal jugular vein, just inferior to the jugular bulb. There is substantial variation in the diameter, number of channels, and degree of symmetry of the IPS, and in the extent of drainage into the basilar plexus and vertebral venous plexus (Fig. 6) (5). Shiu et al. classified the spectrum of drainage patterns at the junction between the IPS and the internal jugular vein into four different variations, ranging from exclusive drainage into the internal jugular vein to exclusive drainage into the vertebral venous plexus (6). Miller et al. subsequently modified this classification (7). In type I anatomy, the IPS drains directly into the internal jugular vein as a single large channel. There may be a small communication with the vertebral venous plexus via the anterior condylar vein or other anastomotic channel. In type II anatomy, the IPS is a single channel. Drainage into the vertebral venous plexus is via a relatively large channel, greater than 1.7 mm in diameter. In type III anatomy, the IPS drains into the internal jugular vein via multiple channels. In type IV anatomy, there is no anastomosis between the IPS and the internal jugular vein. Instead, one or more veins (typically a plexus) originating from the cavernous sinus drains into the vertebral venous plexus. A variant of type II or III anatomy, incomplete type IV also

356

Patronas and Miller

Figure 3 Microcatheter technique of IPSS. (A) Subtracted AP view of injection into right IPS (inferior petrosal sinus) (white arrowhead, introducer catheter tip at junction of right jugular vein and IPSS). (B) Injection of left IPS with cross filling into right IPS (arrowheads, introducer catheter tips in jugular veins). Injection through right (C) and left (D) microcatheters in IPS (arrows, microcatheter tips in IPS). (E) Unsubtracted view showing arrangement of introducer catheters and microcatheters in position for sampling. Abbreviations: IPSS, inferior petrosal sinus sampling; IPS, inferior petrosal sinus.

occurs. In this variant, a very small connection is present between the IPS and the internal jugular vein, but the vast majority of petrosal venous drainage is into the vertebral venous plexus. IPS anatomy is symmetrical in about two-thirds of the individuals; the other one-third have one anatomic type present on the right and a different anatomic type on the left. In a venographic study of 268 IPSs, type I anatomy (the easiest to catheterize) was encountered in 20% of sinuses, type II in 46%, type III in 37%, and type IV in 0.4% of sinuses (7). Incomplete type IV anatomy was encountered in 3% of sinuses (classified above as a variant of type II or type III). Catheterization of IPSs with type II and III anatomy may occasionally be difficult, since the IPS is relatively small. The incomplete type IV variant is particularly difficult to catheterize. True type IV anatomy is fortunately rare, because this variant makes catheterization via the internal jugular vein impossible.

Figure 4 IPS venogram in a patient with Cushing’s disease. (A) Injection in the right and (B) injection in the left IPS. Note prominent vertebral plexus. The right IPS is hypoplastic. The road map from the left IPS injection was used to achieve selective catheterization of the hypoplastic right IPS. Abbreviation: IPS, inferior petrosal sinus.

Chapter 20: Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas

357

Figure 5 Contralateral injection to aid IPSS. (A) Unsubtracted and subtracted (B) AP views of introducer catheter tips (arrows) in the jugular bulbs at the junction with the IPSs. Jugular venogram in AP (B) and lateral (C) views fills left IPS into the posterior left cavernous sinus (arrowhead, C). Left sigmoid sinus also fills (black arrow, C). (D) Right jugular venogram: no filling of right IPS (arrow, introducer catheter tip). (E) Unsubtracted and (F) subtracted AP views following microcatheter placement into the left IPS. Contrast injection through microcatheter (large arrow, microcatheter tip) fills left IPS (arrowhead, impression of left cavernous carotid artery) with cross filling into right IPS, demonstrating junction with right jugular vein (small arrows), thereby aiding microcatheterization of right IPS (G). (H) Unsubtracted AP view shows microcatheters in place for IPS sampling. Abbreviations: IPSS, inferior petrosal sinus sampling; IPS, inferior petrosal sinus.

In experienced hands, bilateral petrosal sinus catheterization is possible in 93% to 99% of patients with patent internal jugular veins (3,7,8). Successful petrosal sinus sampling demands meticulous attention to detail. Petrosal venous sampling requires that simultaneous samples be obtained from two catheters and a venous sheath, and that these samples be placed immediately into correctly numbered and labeled tubes. Multiple timed samples are obtained from each catheter: a baseline set and sets at 3, 5, and 10 minutes after the intravenous administration of 1 mg/kg (maximum dose 100 mg) CRH. Prior to sampling, 2 to 3 mL of blood is withdrawn from

each catheter into a waste syringe and discarded. Each 10-mL blood sample is drawn into a plastic 10-mL syringe over 20 to 40 seconds. Each sample is then transferred into the appropriately numbered and labeled tube using a 16-gauge needle. The tube is gently tilted to mix the sample with the EDTA in the tube and is returned to the ice-water bath. In between sampling, the position of both catheters is checked fluoroscopically to confirm that neither catheter has slipped out of the petrosal sinus. At the end of the procedure, a digital subtraction venogram of each petrosal sinus is obtained separately by gentle hand injection of 5-mL nonionic contrast material. These

358

Patronas and Miller

Figure 6 Low junction of IPS with jugular vein. (A) AP and (B) lateral views of right jugular venogram show IPS junction with jugular vein (arrow) several centimeters below skull base (arrowhead, jugular bulb). (C) Lateral and (D) AP views after microcatheter placement into right IPS, injection fills IPS bilaterally (*, introducer catheter tip in left jugular vein; arrow, introducer catheter tip in right jugular vein; arrowheads, course of microcatheter in right IPS). Abbreviation: IPS, inferior petrosal sinus.

data serve as a permanent record of catheter position during sampling, if questions arise when the data are interpreted.

Interpretation of Sampling Data The physiologic basis for interpretation of sampling data is straightforward: the IPSs drain pituitary venous blood. If the patient has a functioning pituitary adenoma, ACTH will be present at higher concentration in petrosal sinus samples than in peripheral vein samples. If ACTH is coming from an ectopic source, ACTH concentrations in petrosal sinus samples will be similar to those in peripheral venous blood. In cases of Cushing’s disease, petrosal sampling can provide unequivocal evidence of ACTH-secreting adenoma by demonstrating elevated values of ACTH in the blood samples from the petrosal sinuses compared with those of the peripheral venous blood. Since ACTH concentrations in left and right petrosal sinus blood samples are usually not identical, it is essential to sample both sinuses. The side with the higher ACTH concentration in each sample set is used as the IPS value for the data analysis. In baseline samples (obtained prior to CRH administration), an IPS to peripheral (IPS/P) ACTH ratio greater than 2 is indicative of Cushing’s disease. An IPS/P ACTH ratio greater than 3 in any sample set obtained after CRH administration is also diagnostic.

When Oldfield et al. first reported the results of the method, the sensitivity and specificity for detection of a pituitary source of ACTH secretion were found to be 95% and 100%, respectively, in the baseline samples (9). After intravenous administration of CRH, both sensitivity and specificity were 100%. Subsequent investigators have found that both false-negative and, rarely, false-positive results can be encountered (3,10–12). The sensitivity and specificity after CRH administration range from 90% to 97% and from 67% to 100%, respectively (3,11,12). A negative result from petrosal sinus sampling is not conclusive proof that the patient has an ectopic ACTH source. The cause of these false-negative results is not always clear. Displacement of the catheter during sampling or incorrect catheter placement may be responsible in some cases. IPS anatomy— particularly the presence of a hypoplastic IPS—and changes in venous drainage after surgical intervention have also been implicated (13). IPSS also has a role in lateralizing the pituitary adenoma to one side of the pituitary. This lateralization is particularly important in patients with microadenomas. Successful lateralization permits the surgeon to perform a hemihypophysectomy and preserve pituitary function. A ratio of 1.4 or more between the ACTH concentrations of the two petrosal sinus samples from any sample set indicates that the adenoma is located on the side of greater ACTH

Chapter 20: Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas

Figure 7 IPS venogram in a patient with GH-secreting pituitary adenoma. The measured values of GH in pg/mL are recorded. A GH-secreting adenoma was found at surgery in the right half of the pituitary. Abbreviations: IPS, inferior petrosal sinus; GH, growth hormone.

concentration (Fig. 7). The reported sensitivity of IPSS for lateralization of pituitary adenomas in adults ranges from 57% to 90% (3,4,12,14–16). Occasionally, samples obtained before and after CRH administration provide discordant lateralization. In this situation, neither lateralization can be relied upon (17). There are several reasons for this relatively low yield. Often, the adenoma is located in the center of the pituitary and drains into both the cavernous sinuses. The same drainage pattern also occurs in larger adenomas. Anomalies in petrosal venous anatomy such as hypoplasia of one sinus can result in false lateralization to the opposite side, as can asymmetry in petrosal sinus anatomy or previous transsphenoidal surgery (4,17). In a recent study the sensitivity of lateralization in a series of 141 pediatric patients was found to be only 54%, which is more than in conventional MRI and no better than a chance value(18). In a different, smaller series of 11 patients, however, lateralization was correct in 91% of patients (19). The reasons for this discrepancy are unclear. Complications of IPSS

Various investigators have encountered neurologic and other complications during IPSS (20–24). The incidence of such complications is low and ranges from 0.2% to 1.1%. Miller et al. reported a case of hematoma in the pons associated with hemorrhage in the fourth ventricle, which resulted in right hemiplegia with partial recovery and left facial paralysis. They also reported an ischemic infarction in the medulla in a patient who underwent IPSS by other operators at a different institution. Subarachnoid hemorrhage, Raymond’s syndrome (sixth nerve palsy and hemiparesis), and brain stem infarction have been reported by other investigators (21,23,24).

359

A variety of other less severe or transient neurologic events have also been reported, including slurred speech, paresthesias, visual disturbances, transient sixth nerve palsy, vertigo, nausea, vomiting, and transient muscle weakness. These complications may not have a common cause. While the etiology is not known, it appears that most serious complications occur as a result of rupture or thrombosis of a venule in the brain stem or a bridging vein within the subarachnoid space. These complications may be due to the highly variable venous anatomy in this region. When test injections of contrast material are made to identify the orifice of the IPS, special attention should be paid to the size of the veins and to ensure that the catheter tip is not in a small vein. Catheterization of an extremely small vein or forceful hand injection of contrast material in a catheter wedged in such a vein can easily lead to either elevated venous pressure or rupture of that vein. However, the presence of adequatesized venous channels and proper position of the catheter tip do not guarantee that a serious complication will not occur, since these complications have occurred despite seemingly unremarkable petrosal sinus anatomy (20). Additionally, catheterization of extremely small petrosal sinuses has been performed without incident (7). A neurologic event may be heralded by minor and seemingly insignificant symptoms. Arterial hypertension, slurred speech, difficulty swallowing, a sensation of an enlarged tongue, a ‘‘woozy’’ feeling, and hemifacial paresthesias have all been encountered as initial manifestations of a brain stem insult. These may be subtle, and the patient may not mention them unless asked. If present, they should not be interpreted as evidence of anxiety, oversedation, or a reaction to contrast material. Brain stem injury may be preventable if the catheter is withdrawn at the earliest sign of even a minor, insignificant problem. Subtle symptoms and signs that may not appear to be neurologic may herald a clinical catastrophe if not heeded. Patients with Cushing’s syndrome are also prone to venous thrombosis, and both deep venous thrombosis and pulmonary embolus have been reported as complications of petrosal sinus sampling (25).

ALTERNATIVES TO IPSS Catheterization of, and sampling from, the cavernous sinus has been suggested as an alternative to petrosal sinus sampling, on the grounds that it is both safer and more accurate (14,26,27). Other studies indicate that sampling from the cavernous sinus is no more accurate than petrosal sinus sampling for distinguishing between Cushing’s disease and an ectopic ACTH source and is less accurate for lateralization of an adenoma within the pituitary gland (4,28). In a series of 14 cavernous sinus sampling procedures by Lefournier et al., transient sixth nerve palsies occurred in two patients (4). Doppman et al. advocated the sampling of the internal jugular veins because of the technical difficulties that can be encountered in selective catheterization

360

Patronas and Miller

of the IPS, the long learning curve for petrosal sinus sampling, and the need to abandon the procedure in patients who develop suspicious symptoms, systemic hypertension, or neurologic events during the procedure (29). In this simpler procedure, catheters are placed in both internal jugular veins at the level of the mandible. In the initial description of the procedures, the catheters were inserted through bilateral femoral vein punctures, but in a subsequent series catheters were placed via the internal jugular vein (30). In our own practice, we routinely use ultrasound guidance and a micropuncture set to access the internal jugular vein as inferiorly as possible in the neck, and advance the inner 3-Fr dilator of the micropuncture set retrogradely, so that its tip lies at the level of the mandible. The outer dilator is not used. No skin nick is necessary, and the procedure is performed with local anesthesia alone. Samples are obtained before and after CRH administration in the same fashion as for petrosal sinus sampling. In a series of 79 patients, the NIH group found a sensitivity of 83% for jugular venous sampling, using arbitrary thresholds (jugular vein/peripheral vein ACTH ratio >1.7 before CRH administration or >2.0 after CRH administration) to set specificity at 100%. In the same patients, IPSS had a sensitivity of 94% when specificity was set at 100% (30). Erickson et al. demonstrated similar results in a series of 35 patients. They suggest that the sensitivity of internal jugular vein sampling can be improved by placing the catheter near the medial rather than the lateral wall of the internal jugular vein during sampling and by using a jugular vein/peripheral vein ACTH ratio greater than 2.5 for the diagnosis of Cushing’s disease (31). Although internal jugular vein sampling is less sensitive than petrosal sinus sampling, it is simpler and avoids the risk of neurologic complications inherent in petrosal sinus sampling. It is reasonable to perform jugular venous sampling first and to reserve petrosal sinus sampling for the patients in whom jugular venous sampling does not confirm Cushing’s disease. These patients may be referred to centers where extensive experience in performing petrosal sinus sampling is available.

REFERENCES 1. Teramoto A, Hirakawa K, Senno N, et al. Incidental pituitary lesions in 1,000 unselected autopsy specimens. Radiology 1994; 193:161–164. 2. Hall WA, Luciano MG, Doppman JL, et al. Pituitary magnetic resonance imaging in normal human volunteers: occult adenomas in the general population. Ann Intern Med 1994; 120(10):817–820. 3. Bonelli FS, Huston J III, Carpenter PC, et al. Adrenocorticotropic hormone-dependent Cushing’s syndrome: sensitivity and specificity of inferior petrosal sinus sampling. AJNR Am J Neuroradiol 2000; 21(4):690–696. 4. Lefournier V, Martinie M, Vasdev A, et al. Accuracy of bilateral inferior petrosal or cavernous sinuses sampling in predicting the lateralization of Cushing’s disease pituitary microadenoma: influence of catheter position and anatomy of venous drainage. J Clin Endocrinol Metab 2003; 88(1):196–203.

5. Gebarski SS, Gebarski KS. Inferior petrosal sinus: imaging-anatomic correlation. Radiology 1995; 194:239–247. 6. Shiu PC, Hanafee WN, Wilson GC, et al. Cavernous sinus venography. AJR Am J Roentgenol 1968; 104:57–62. 7. Miller DL, Doppman JL, Chang R. Anatomy of the junction of the inferior petrosal sinus and the internal jugular vein. AJNR Am J Neuroradiol 1993; 14:1075–1083. 8. Miller DL, Doppman JL. Petrosal sinus sampling: technique and rationale. Radiology 1991; 178:37–47. 9. Oldfield EH, Doppman JL, Nieman LK, et al. Petrosal sinus sampling with and without corticotropin-releasing hormone for the differential diagnosis of Cushing’s syndrome. N Engl J Med 1991; 325:897–905. 10. Yamamoto Y, Davis DH, Nippoldt TB, et al. False-positive inferior petrosal sinus sampling in the diagnosis of Cushing’s disease. Report of two cases. J Neurosurg 1995; 83:1087–1091. 11. Swearingen B, Katznelson L, Miller K, et al. Diagnostic errors after inferior petrosal sinus sampling. J Clin Endocrinol Metab 2004; 89:3752–3763. 12. Kaltsas GA, Giannulis MG, Newell-Price JDC, et al. A critical analysis of the value of simultaneous inferior petrosal sinus sampling in Cushing’s disease and the occult ectopic adrenocorticotropin syndrome. J Clin Endocrinol Metab 1999; 84(2):487–492. 13. Doppman JL, Chang R, Oldfield EH, et al. The hypoplastic inferior petrosal sinus: a potential source of false-negative results in petrosal sampling for Cushing’s disease. J Clin Endocrinol Metab 1999; 84(2):533–540. 14. Graham KE, Samuels MH, Nesbit GM, et al. Cavernous sinus sampling is highly accurate in distinguishing Cushing’s disease from the ectopic adrenocorticotropin syndrome and in predicting intrapituitary tumor location. J Clin Endocrinol Metab 1999; 84(5):1602–1610. 15. Booth GL, Redelmeier DA, Grosman H, et al. Improved diagnostic accuracy of inferior petrosal sinus sampling over imaging for localizing pituitary pathology in patients with Cushing’s disease. J Clin Endocrinol Metab 1998. 83(7):2291–2295. 16. Oldfield EH, Chrousos GP, Schulte HM, et al. Preoperative lateralization of ACTH-secreting pituitary microadenomas by bilateral and simultaneous inferior petrosal venous sinus sampling. N Engl J Med 1985; 312:100–103. 17. Miller DL, Doppman JL, Nieman LK, et al. Petrosal sinus sampling: discordant lateralization of ACTH-secreting pituitary microadenomas before and after stimulation with corticotropin-releasing hormone. Radiology 1990; 176:429–431. 18. Batista D, Gennari M, Riar J, et al. An assessment of petrosal sinus sampling for localization of pituitary microadenomas in children with Cushing disease. J Clin Endocrinol Metab 2006; 91(1):221–224. 19. Lienhardt A, Grossman AB, Dacie JE, et al. Relative contributions of inferior petrosal sinus sampling and pituitary imaging in the investigation of children and adolescents with ACTH-dependent Cushing’s syndrome. J Clin Endocrinol Metab 2001; 86(12):5711–5714. 20. Miller DL, Doppman JL, Peterman SB, et al. Neurologic complications of petrosal sinus sampling. Radiology 1992; 185:143–147. 21. Bonelli FS, Huston J III, Meyer FB, et al. Venous subarchnoid hemorrhage after inferior petrosal sinus sampling for adrenocorticotropic hormone. AJNR Am J Neuroradiol 1999; 20:306–307. 22. Lefournier V, Gatta B, Martinie M, et al. One transient neurological complication (sixth nerve palsy) in 166 consecutive inferior petrosal sinus samplings for the etiological diagnosis of Cushing’s syndrome. J Clin Endocrinol Metab 1999; 84(9):3401–3402 (letter).

Chapter 20: Inferior Petrosal Sinus Sampling in the Diagnosis of Pituitary Adenomas 23. Seyer H, Honegger J, Schott W, et al. Raymond’s syndrome following petrosal sinus sampling. Acta Neurochir (Wien) 1994; 131(1–2):157–159. 24. Sturrock ND, Jeffcoate WJ. A neurological complication of inferior petrosal sinus sampling during investigation for Cushing’s disease: a case report. J Neurol Neurosurg Psychiatry 1997; 62(5):527–528. 25. Obuobie K, Davies JS, Ogunko A, et al. Venous thromboembolism following inferior petrosal sinus sampling in Cushing’s disease. J Endocrinol Invest 2000; 23(8): 542–544. 26. Teramoto A, Nemoto S, Takakura K, et al. Selective venous sampling directly from cavernous sinus in Cushing’s syndrome. J Clin Endocrinol Metab 1993; 76:637–641. 27. Vandorpe RA, Fox AJ, Pelz DM, et al. Direct sampling of the cavernous sinus in Cushing’s disease. Can Assoc Radiol J 1994; 45(3):234–237.

361

28. Doppman JL, Nieman LK, Chang R, et al. Selective venous sampling from the cavernous sinuses is not a more reliable technique than sampling from the inferior petrosal sinuses in Cushing’s syndrome. J Clin Endocrinol Metab 1995; 80:2485–2489. 29. Doppman JL, Oldfield EH, Nieman LK. Bilateral sampling of the internal jugular vein to distinguish between mechanisms of adrenocorticotropic hormone-dependent Cushing syndrome. Ann Intern Med 1998; 128(1):33–36. 30. Ilias I, Chang R, Pacak K, et al. Jugular venous sampling: an alternative to petrosal sinus sampling for the diagnostic evaluation of adrenocorticotropic hormone-dependent Cushing’s syndrome. J Clin Endocrinol Metab 2004; 89(8):3795–3800. 31. Erickson D, Huston J III, Young WF Jr., et al. Internal jugular vein sampling in adrenocorticotrophic hormone-dependent Cushing’s syndrome: a comparison with inferior petrosal sinus sampling. Clin Endocrinol 2004; 60:413–419.

21 Endovascular Treatment of Spinal Vascular Malformations Mayumi Oka and Kieran Murphy Department of Radiology, Division of Interventional Neuroradiology, Johns Hopkins University, Baltimore, Maryland, U.S.A.

INTRODUCTION In 1960s and 1970s, interventional neuroradiological techniques for the treatment of spinal vascular lesions were developed when understanding of these lesions deepened because of the advances in selective spinal angiography techniques and increased knowledge. Two groups of authors contributed the initial and greater part of the development of selective spinal angiography. Djindjian et al. reported their first 50 cases of transarterial embolization in 1973 (1). Di Chiro and Doppman described their own techniques and experiences in spinal angiography. Aminoff and Logue contributed to an early understanding of the pathophysiology of spinal vascular malformations (2) and established the clinical grading system (3). Later, Kendall and Logue recognized the dural arteriovenous fistula (dAVF) as a different entity from arteriovenous shunts involving the spinal cord (4). These are uncommon and complex pathologies, and the terminology has changed over the years. We will define the terminology of lesions and describe clinical manifestation, imaging findings, and management of each lesion; all are best treated by multidisciplinary approach.

CLASSIFICATION Nomenclature for spinal vascular malformations has caused confusion and controversy among clinicians, and multiple classification systems have been proposed until today (5–7). In 1978, Hurth et al. reported the first large series (8), which presented a summary of 150 cases divided into two groups: extramedullary malformations fed by the posterior spinal artery and intramedullary malformations fed by the anterior spinal artery. Their classification was aimed at a surgical approach focusing on the position of lesions relative to the spinal cord, rather than the type of shunts. As noted before, dAVFs were recognized as a distinctly different pathology from other spinal vascular malformations only in 1977 (4). Their classification likely included most of the dAVFs in the group of extramedullary malformations. In 1985, Riche et al. presented their classification, close to the modern understanding of these lesions (9), and distinguished

five different types of vascular malformations: (1) intramedullary or mixed arteriovenous malformation (IM-AVM), (2) retromedullary AVM, (3) extramedullary arteriovenous fistula (AVF) supplied by the spinal arteries, (4) extramedullary dAVF with medullary venous drainage, and (5) complex malformations (disseminated and metameric AVM). Since then, many authors have suggested modified or new classifications, and there has been drastic advancement in diagnostic modalities and our knowledge of spinal vascular malformations. However, the classification proposed by Riche et al. still retains the basic concept of spinal vascular malformations with the exception of retromedullary AVMs, which are simply included with IM-AVM in the present classification. Most authors categorize pathologies on the basis of angioarchitecture and location of lesions. Table 1 summarizes classification of spinal vascular malformations. Most spinal vascular malformations can be divided into two different types of shunts: AVFs or AVMs. Locations of the lesions are categorized as: (1) (intradural) intramedullary, (2) (intradural) perimedullary, (3) dural, and (4) extradural (epidural and paraspinal). A combination of morphological/hemodynamic and topographic information, usually provided by angiography, is used to classify these lesions. Differentiation of AVF and AVM can be difficult at times, and interpretation of angiographic images is not free from subjective judgment. Spinal vascular malformations are also a mixture of congenital and acquired lesions, etiology still needs to be elucidated and the information will be incorporated in a future classification. For now, a simplified classification with less controversy may ease communication between clinicians from different specialties. Certain genetic or hereditary syndromes/disorders are known to be associated with spinal vascular malformations. Rodesch et al. proposed the classification of intradural spinal vascular malformations based on genetics or biological features (7). They primarily distinguish AVFs and AVMs, and secondarily divide them into three categories. The first group consists of single shunts associated with genetic or hereditary disorders, mainly hereditary hemorrhagic telangiectasia (HHT) or Rendu-Osler-Weber disease. These are usually single-hole macro-AVFs and affect the pediatric

364

Oka and Murphy

Table 1 Classification of Spinal Vascular Malformations Location

Type

Feeder

Drainage

Dural Perimedullary Intramedullary Extradural

AVF AVF AVM AVF, AVM

Radiculomeningeal ASA and/or PSA ASA and/or PSA Commonly segmental

Radicular vein Perimedullary vein Medullary or perimedullary vein Epidural and/or paraspinal

Abbreviations: AVF, arteriovenous fistula; AVM, arteriovenous malformation; ASA, Anterior Spinal Artery; PSA, Posterior Spinal Artery.

population (10). The second group comprises genetic, nonhereditary, multiple AV shunts with potential metameric links. These include Cobbs syndrome, KlippelTrenaunay syndrome, and Parkes Weber syndrome (or Klippel-Trenaunay-Weber syndrome). The third group is made of single lesions consisting of a majority of spinal vascular malformations, while the first two groups comprise 16% of spinal cord AVMs in a series of 19 patients at one author’s institution (11). These are categorized as complex lesions in the classification by other authors (12,13) and as such require careful investigation of the entire pathology and determination of the lesion responsible for clinical symptoms. The aim of treatment should be symptomatic relief rather than complete cure in most cases.

CLINICAL Spinal vascular malformations are uncommon lesions. Mourier et al. studied 210 patients treated for an AVM of the spinal cord. The patients were classified into dAVF (38%), IM-AVMs (45%), and perimedullary AVFs (PM-AVFs) (17%) (14). In a series of 186 spinal vascular malformations by Biondi et al., dAVF, IM-AVM, and PM-AVF comprised 38%, 24%, and 39%, respectively (15). These authors suggest that distribution of their cases was largely influenced by a referral base and the nature of their institution being a tertiary care center. Most reports suggest a much higher rate of dAVFs, ranging from 60% to 80% of spinal vascular malformations (4). There are two main forms of presentations of spinal vascular malformations: one is progressive myelopathy of gradual onset and the other is sudden onset of neurological deficit or worsening of existing symptoms, usually secondary to hemorrhage (16). The less common form is an acute deterioration without hemorrhage, which is thought to be due to thrombosis of the draining vein of the lesion itself (8). Intradural spinal arteriovenous shunts (AVMs and AVFs) have a high rate of hemorrhage reported in the literature, ranging from 30% to 50% (8,16–18). Hemorrhages occur as a spinal subarachnoid hemorrhage (SAH) or hematomyelia. Direct destruction of neural tissue by hematoma (hematomyelia) accounts for more severe clinical signs than those secondary to SAH. Symptoms of SAH depend on the level of rupture; however, acute onset of pain, ‘‘stabbing back pain,’’ is universal with or without myelopathy or radiculopathy. When the lesion is closer to the craniocervical junction, signs and symptoms resemble those of intracranial SAH and cause a special diagnostic

dilemma. Angiography negative for intracranial aneurysms when examining a patient with SAH needs further investigation for cervical spinal vascular malformation. It should be noted that, in a series related to spinal vascular malformations, of 150 patients 55% occurred in children less than 15 years of age (8). This result is concordant with the findings observed by Rodesch et al., in which 70% of the pediatric population in their series of intradural spinal vascular malformations (excluding dAVFs) manifested hemorrhagic episodes (19). In contrast to intradural AV shunts, dAVFs are not typically associated with spinal SAH or hematomyelia (20–22). The exception to this rule being cervical dAVF (23,24). These are more typically present with complications related to venous hypertension and impaired cord venous drainage. Besides spinal SAH and hematomyelia, symptoms of spinal vascular malformations are those of nonspecific myelopathy or radiculopathy. Combination of paraparesis, sensory abnormalities, sphincter disturbances, and pain, which is often radicular in distribution, develops with highly variable speeds of progression. Stepwise progression, with incomplete recovery of symptoms between events, is common (8) and not directly correlated with the level of shunts. Progressive, slow deterioration of neurological status is a classic feature of spinal vascular malformations, often attributed to chronic venous hypertension, and eventually results in ischemic hypoxia of the spinal cord. Intermittent, transient worsening of symptoms have been documented with exercise, cough, or certain postures (16,22,25). Pregnancy is attributed to causing aggravation of the disease in a minority of cases (8,15). In 1974, Aminoff and Logue reported a series of 60 patients with spinal vascular malformations in which dural and intradural AV shunts were all mixed, as it was before dAVF was differentiated from others (3,16). Ten percent of patients presented with SAH. Severe locomotor disability occurred in 19% of patients within six months of onset and in 50% within three years. Only 9% of their patients were able to walk independently after three years. Some authors established a clinical grading system of the three major symptoms associated with spinal vascular malformations: problems with gait, micturition, and defecation (3). Gait disturbances were graded as: (1) onset of leg weakness, abnormal stance or gait, without restriction of locomotor activity; (2) diminished exercise tolerance; (3) requirement for one stick or some support for walking; (4) requirement for crutches or two sticks for walking; and (5) unable to

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

stand, confined to bed or wheelchair. Disturbances of micturition have been classified as mild—hesitancy, urgency or frequency; moderate—occasional urinary incontinence or retention; and severe—total urinary incontinence or persistent retention. Disordered control of defecation has been similarly classified as mild—constipation; moderate—occasional fecal incontinence or severe intractable constipation; severe—fecal incontinence. Delay in diagnosis is a particular problem of spinal vascular malformations, especially for those who present with nonspecific, slowly progressive radiculopathy or myelopathy, or diabetes. The duration from onset of symptoms to initial treatment averaged 2.7 years with dAVFs and 4.2 years with intradural AVMs in the series by Rosenblum et al. (18). The time from first symptoms to diagnosis was less than 1 year in 26%, 1 to 10 years in 60%, and more than 10 years in 14%. Others reported similar results.

IMAGING Early reports indicated that myelography demonstrates high rate of positive findings in patients with spinal vascular malformations. Hurth reported typical vascular filling defects (Fig. 1A) in 61%; nonspecific, abnormal findings (complete or partial obstruction of contrast column, or an enlarged cord) in 30%; and normal myelogram in only 9% of their series (8). In a series of dAVFs, dilated vessels were present in all 25 patients on supine myelogram (26). However, magnetic resonance imaging (MRI) has become the modality of choice in

365

evaluation of myelopathy, radiculopathy, or spinal SAH, since it can demonstrate other more common pathologies such as disc herniation, spinal stenosis, vertebral lesion, intra- or extramedullary neoplasm, and discitis/osteomyelitis. It can still be difficult to differentiate infectious or inflammatory myelitis or intramedullary mass (neoplasm or hematoma) from myelopathy caused by vascular malformations when not associated with significant flow voids. Even though MRI and magnetic resonance angiography (MRA) often suggest and make a diagnosis of spinal vascular malformations, spinal angiography is essential when lesions are being considered for treatment. Selective spinal angiography should focus on several points with future therapy in mind when performed: (1) First to differentiate dAVFs from intradural AV shunts, (2) to determine the exact level of shunt (by vertebral levels for surgical option), (3) identify all feeders and relationship with the radiculomedullary artery (or radiculopial artery) that is not directly feeding the fistula—continuity of spinal artery axis, and (4) presence of aneurysm and venous varix, and their relationship with symptoms (compression, rupture, etc.). Each level needs to be selected, and angiographers must be attentive to vascular blush in the hemivertebra, which implies that the dorsal spinal branch has been injected (Fig. 2A, B). The ventral and dorsal spinal branches can have separate origins from the aorta, especially when there is a common trunk for multiple levels (27). When the angiogram is negative after intercostal and lumbar artery injections, vertebral, deep cervical, ascending cervical, and internal iliac arteriogram should be performed.

Figure 1 PM-AVF, type I. A 28-year-old male presented with left lower extremity weakness. MRI of the thoracolumbar spine showed central hyperintensity of the cord on T2-weighted images and central enhancement on postgadolinium images (not shown). Myelogram (A) shows serpentine filling defect consistent with prominent draining vein at the lower thoracic levels. Left T11 intercostal artery injection (B) reveals the mildly prominent radiculomedullary artery and the anterior spinal artery. The arterial basket, connection between the anterior spinal artery and the posterior spinal artery, is outlined (long arrow). A fistula (arrow head ) is noted immediately distal to the basket, a draining vein is seen faintly on this image. Lateral projection (C) of same injection shows the anterior spinal artery (small arrows) and the fistula (large arrow). Later image (D) shows the artery (arrow) and draining veins (double arrow) posterior to the spinal cord. Abbreviations: PM-AVF, perimedullary arteriovenous fistula; MRI, magnetic resonance imaging.

366

Oka and Murphy

Figure 2 Left T9 intercostal artery injection demonstrated a normal dorsospinal artery with blush in the left hemivertebra.

SPINAL dAVFs Terms for dAVF include epidural angiomatous malformations (4), dorsal extramedullary AVM, type I, and intradural dorsal AVF (5). These are the most common types of spinal vascular malformation (28).

Pathophysiology The dAVFs are shunts between the dural branch (radiculomeningeal artery) of the dorsospinal artery and radicular vein (Fig. 3), which normally drains the perimedullary vein. The fistula is located within the dural sleeve of the exiting spinal nerve root. The fistula drains into the perimedullary venous system via radicular veins in a retrograde fashion. The pathophysiology of neurological symptoms is attributed to chronic venous hypertension caused by retrograde flow in the perimedullary vein, which normally drains the cord via the coronal venous plexus (2,4). Slow but high-pressure

retrograde venous flow into the valveless coronal venous plexus limits venous drainage of the spinal cord by the normal radial veins and results in a decreased arteriovenous gradient, eventually leading to congestive cord ischemia, which may or may not be reversible. The result of these pathological changes is irreversible necrotizing myelopathy—first described by Foix and Alajouanine (29)— also called angiodysgenetic myelomalacia or subacute necrotic myelitis, where the neural tissue may liquefy and produce a cavity (32). Spinal dAVF is an acquired disease, although the etiology is still unknown (30). Infection, trauma, syringomyelia, and surgery have been mentioned as an association or cause in the form of case reports (28,31). Venous thrombosis is the leading pathogenesis of cranial dAVFs (32) and is also considered to be a potential cause of spinal dAVFs. However, there was no association between multiple prothrombotic factors and spinal dAVFs, comparing 40 patients with dAVF and 119 control patients (33).

Figure 3 Spinal dAVF. A 39-year-old male with paraplegia. Sagittal proton density MRI (A) shows flow voids along the posterior aspect of the thoracic cord. Early arterial phase of right T5 intercostal arteriogram (B) shows a shunt (small arrow) between the radiculomeningeal branch of the dorsospinal artery and the radicular vein (large arrow). Later image shows a shunt (arrow) and venous drainage into the perimedullary vein in both cranial and caudal directions (long arrows) (C, D). Abbreviations: dAVF, dural arteriovenous fistula; MRI, magnetic resonance imaging.

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

Clinical Manifestations Spinal dAVFs commonly affect middle-aged to elderly males with 4–5:1 male to female ratio. Most patients are in their fourth to seventh decades. Patients usually present with gradually progressive myelopathy, which affects lower extremity and sphincter functions. Common initial symptoms are pain (16–39%), lower extremity weakness (29–55%), and sensory disturbance (24–47%). Sphincter dysfunction was seen in about 10% of patients. Symptoms can progress slowly and continuously, or in stepwise fashion. Most patients have a combination of motor, sensory, and sphincter symptoms by the time diagnosis is made, paraparesis in 78% to 100%, sensory disturbance in 69% to 90%, urinary incontinence in 80% to 89%, disturbed defecation in more than 80%, and disturbed sexual function in about a third of patients, though this symptom is often concealed by patients (20–22,34,35). Pain is a common but nonspecific symptom that manifests as a backache or radicular pain. This pain is often attributed to degenerative lumbar diseases or polyradiculopathy before dAVFs are suspected. Sensory disturbances start with tingling paresthesia or hyperesthesia in the feet and progress to proximal level. Flaccid and spastic paraparesis are equally common (20). Hemorrhage is uncommon and SAH is seen almost exclusively with cervical dAVFs; only one lumbar dAVF with SAH (36) and one hematomyelia in thoracic dAVF to date (37). Delay in making a diagnosis of dAVF is common, ranging from months to often several years with a median length of 10.5 to 27 months, because of the nonspecific and insidious nature of symptoms (20,21,26). Van Dijk in his report suggested that recent advances in diagnostic imaging, mainly MRI and MRA, and wide availability of the scanner may have shortened the delay in diagnosis of dAVFs. In their series, 30 out of 49 patients (61%) presented with dAVFs.

Imaging MRI should be the first imaging modality performed when any spinal vascular malformation is suspected. Although myelogram can demonstrate enlarged veins in most dAVFs (26), the myelogram must be obtained in a supine position since most of the veins are located dorsal to the spinal cord. This maneuver may not be done unless the diagnosis is already suspected. MRI findings commonly seen in patients with dAVFs, listed in order of frequency, are (1) central hyperintensity of the cord on T2-weighted images (85– 100%), (2) mild gadolinium enhancement, and (3) vascular flow voids posterior to the spinal cord and mild expansion of the cord (Fig. 4) (26,38,39). Central hyperintensity on T2-weighted images reach the tip of conus in a majority of cases (26). Peripheral hypointensity surrounding central hyperintensity has been described by Hurst et al., which is more conspicuous on true T2-weighted or gradient-echo images but subtle on Fast Spin Echo (FSE) T2-weighted images. The authors hypothesize that the finding is due to slow flow of blood containing deoxyhemoglobin

367

within distended veins (40). One must be familiar with the normal MR appearance of the spine, it can be difficult to differentiate an abnormally dilated coronal venous plexus from prominent but normal veins on today’s high-field MRI. Cerebral spinal fluid pulsation artifact should not be mistaken as abnormal flow voids. With recent advances in MRA techniques, first-pass contrast-enhanced MRA is reported to identify the level of fistula within  one level with a relatively high rate of accuracy ranging from 75% to 100% (38,41,42). Also, ever advancing multidetector row CT angiography has demonstrated precise localization of dAVFs in all eight patients (43). Conventional catheter spinal angiography, however, is indispensable for choosing treatment options and is still the gold standard to evaluate the vascular pathology of the spine. The above mentioned noninvasive imaging techniques may play a role in reducing the length of catheter angiography, thus decreasing contrast load and radiation dose, especially in those with renal insufficiency and severe atherosclerotic disease (38). Because of particular demographics of patients affected by this disease, there are few falsenegative angiograms mainly because of occlusion of the origin of feeding intercostals or lumbar arteries, severe atherosclerosis, or aortic aneurysm (44). Arterial feeders are commonly located in the midthoracic to upper lumbar level with more than 80% seen between T5 and L2, and two-thirds on the left (20–22). In one series, sacral dAVF was common (18%) (45). In case of negative spinal angiography, after selective intercostal and lumbar artery injections, a selective lateral sacral artery injection should be performed. Multiple feeders to the fistula are seen in as low as 10% to as high as 60% (20,22,45). Multiple dAVFs are an uncommon entity with a few case reports of double dAVFs in the literature, and their incidence is less than 2% of all spinal dAVFs (46–48).

Cervical dAVFs Cervical dAVFs are an uncommon subgroup of dAVFs—approximately 2.5% of all spinal dAVFs (36) (Fig. 5). Although they have the same morphology and pathophysiology as thoracolumbar dAVFs, one needs to be aware of particular characteristics of cervical dAVFs. In patients with myelopathy, motor and sensory symptoms are not always localized in the lower extremity. Hemiparesis or quadriparesis are as common as paraparesis (23). Myelopathy at a cervical level can also include brain stem signs such as cranial neuropathy or dyspnea (49). The most important difference is that they have a much higher rate of SAH when compared with their thoracolumbar counterpart. Recent literature reviews reported a 30% to 45% incidence of SAH in cervical dAVFs (24,36). The presence of a varix and superiorly directed venous drainage were significantly associated with SAH. In a review of 41 patients with cervical dAVFs, superiorly directed drainage was seen in 60% (12 out of 20) of the SAH group, which is much higher than 10% in the non-SAH group, and reaching the cranium in 50% of cases (10 of 20). Venous varix was noted in 35% and

368

Oka and Murphy

Figure 4 Spinal dAVF. A 34-year-old male with scoliosis presented with acute deterioration of bilateral lower extremity weakness and urinary incontinence, which has been present over a year. His main complaint prior to this event was back pain. T2-weighted images of the thoracolumbar spine (A, B) demonstrates flow voids dorsal to the cord and abnormal high signal in the spinal cord from T4 to the conus. Contrast enhancement is noted in the lower thoracic cord (C). The T10 intercostal angiogram (D) shows a fistula between the radiculomeningeal artery and the radicular vein (arrow). Later image (E) shows dilated perimedullary veins in both cranial and caudal directions, down to the conus (arrow). Glue embolization with NBCA was performed (F). Glue penetrates the fistula (short arrow) and occludes the proximal segment of the draining vein (large arrow). Control angiogram performed after embolization reveals residual fistula fed by right T9 (G) and T11 (H) intercostal artery branches, contribution to the fistula from these feeders were not seen prior to embolization of T10. Two feeders were embolized with glue subsequently and final angiogram showed no residual fistula. Abbreviations: dAVF, dural arteriovenous fistula; NBCA, N-butyl 2-cyanoacrylate.

Figure 5 Cervical dAVF. An 84-year-old male presented with lower extremity weakness and an unsteady gait. Sagittal T2-weighted image (A) demonstrated T2 hyperintensity in the central cord at mid- to lower thoracic region. Selective injections of all intercostal and lumbar arteries were negative. Right vertebral artery injection (B, C) demonstrates small AVF fed by the lateral spinal artery or C1 radicular artery. Venous drainage is caudal and could be followed to midthoracic level (D), which corresponds to MRI findings. Abbreviations: dAVF, dural arteriovenous fistula; AVF, arteriovenous fistula; MRI, magnetic resonance imaging.

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

5% of the SAH and non-SAH group, respectively (24). Authors also noted high prevalence of feeders from the right vertebral artery (68%). Venous drainage is via the coronal venous plexus, epidural or intracranial. Purely epidural venous drainage is associated with myelopathy due to mass effect, rather than venous congestion as seen in most dAVFs (50).

Treatment Spinal dAVF is an infrequent but potentially treatable cause of myelopathy. As all the other spinal vascular malformations, dAVF is best managed by a multidisciplinary team of neurologists, neurosurgeons, and interventional neuroradiologists. Interruption of the feeding artery only is not sufficient to eliminate the fistula and often results in recurrence—as in cranial dAVFs, the fistula recruits nearby arteries or else existing microfeeders grow. Resection of draining veins, which was once thought to be the pathology of dAVFs by means of stripping dilated coronal venous plexus, can cause a devastating outcome. It is now known that treatment should focus on the fistula and disconnecting the vein from the AVF. Meta-analysis of surgical studies demonstrated 97.9% technically successful results, 55% overall improvement, and 33% rate of improvement in micturition function (51). The same authors analyzed results of embolization and found a 46% technical success rate; however, there was not enough data on the outcome of the embolization series. Generally, approximately one-half to two-thirds of patients report improvement in motor function, one-tenth of patients experience worsening, and the remainder become stable. Function of the sphincter does not recover as much as motor strength; improvement is seen in one-third, stability of symptoms in one-third, and continuous deterioration in one-third (52). Jellema et al. noted that leg pain and muscle spasms were difficult symptoms to alleviate. In their series of 44 patients, the majority of patients who had either pain or spasms experienced worsening of the symptoms despite improvement in motor function (52).

369

Others noted correlation between the level of fistula and the outcome. Better results were seen when the lesion was in the lower thoracic region compared with those in the midthoracic or lumbar levels (35). Shorter duration of symptoms, less than a year, prior to treatment appears to correlate with better outcome, especially with sphincter dysfunction (53). The optimal treatment for spinal dAVFs is controversial, especially with ever advancing endovascular techniques. Many authors have addressed the importance of a multidisciplinary approach (20,21,52,53). They have advocated an initial attempt of endovascular therapy when possible, reserving surgery for anatomically unfavorable lesions—usually implied as the segmental artery that harbors both the feeder of dAVF and the artery of Adamkiewicz (Fig. 6A, B). Surgery can be performed immediately after embolization, as it does not interfere with any surgical technique, if embolization fails or a complex network of dural collaterals appears as a result of embolization. Those authors reported no differences in outcomes among those who were treated by surgery, embolization, or combination of both. Endovascular Technique

Those who perform endovascular treatments for spinal dAVF should know that surgery for these lesions is relatively straightforward with high success rates and low complication rates (51) (Fig. 7A, B). If the fistula does not have optimal anatomy for embolization, or when technical difficulty is encountered during the procedure, the patient should be referred for surgery. Case selection is the key for successful endovascular treatment for dAVFs. Embolization is contraindicated if the artery of Adamkiewicz, a major contributor to the anterior spinal artery, arises from the same dorsospinal artery as a feeding artery of dAVF. This contraindication occurs in approximately 10% of patients (45,54). We consider visualization of the radiculopial artery (a contribution to the posterior spinal artery) also as contraindication to embolization. Niimi et al. reported 87% technical success in 33 of

Figure 6 Spinal dAVF. A 59-year-old male presented with progressive lower extremity weakness. Selective left T6 intercostal artery injection shows a plexiform network of vessels at fistula (small arrow) and prominent perimedullary vein (double small arrows). A radiculomedullary artery arises from the same level (long arrow). Abbreviation: dAVF, dural arteriovenous fistula. Source: Courtesy of Philippe Gailloud, Division of Interventional Neuroradiology, Johns Hopkins University (unpublished material).

370

Oka and Murphy

38 cases since the introduction of the variable stiffness microcatheter, and noted that three out of five patients who had ‘‘inadequate’’ embolization had a spinal cord artery arising from the same pedicle as the feeder, which prevented more aggressive embolization (45).

Figure 7 Spinal dAVF. A 58-year-old male presented with progressive lower extremity weakness, which worsened to complete paraplegia in last 10 days with no sensation below T6. Sagittal T2-weighted MRI (A) demonstrates prominent flow voids posterior to the thoracic cord with abnormal high signal in the cord, which extends down to the conus (not shown). (B) Spinal angiogram demonstrates arteriovenous shunt (arrow) at T6 level, fed by T6 intercostal artery, which shares a common trunk with T5 and T7. The patient underwent surgery three days after the angiogram. His strength improved to three-fifth in both lower extremities, and improvement in pinprick and bilateral vibration were observed. However, he still needs self-catheterization at four months follow-up. Abbreviations: dAVF, dural arteriovenous fistula; MRI, magnetic resonance imaging.

Navigation of the microcatheter through often near-normal sized, but invariably tortuous feeder, can be difficult and stability of the guiding catheter can play a significant role in wire-catheter navigation. There are many guiding catheters of different shapes and variable stiffness suited for intercostals and lumbar arteries. We perform all endovascular spinal intervention under general anesthesia. The ability to suspend respiration at the crucial moment gives a more accurate delivery of embolic material. The goal of embolization is to have the embolic material reaching the proximal portion of the draining vein through the fistula (Fig. 8). Proximal occlusion of the feeding artery may temporarily improve symptoms by reducing arterial flow through the fistula, but will not be a cure. Angiographers need to recognize the proximal occlusion; failure to do so is the most common cause of recurrence after embolization, since postembolization angiogram shows obliteration of dAVF in either case. In case of definite proximal occlusion, surgery should be considered soon after embolization. If there is any doubt of glue staying proximal to the vein, short-term follow-up angiography should be performed. N-Butyl 2-cyanoacrylate (NBCA) is the choice of embolization material today. Use of coils or particles (mostly polyvinyl alcohol) is not acceptable because of the well-documented high recurrence rate (55,56). A mixture of NBCA and ethiodol (often 1:2) is injected slowly through the microcatheter that is optimally positioned in the feeding artery as close to the fistula as possible. Speed of injection and ratio of NBCA to ethiodol varies case by case and requires operator experience. When there is some distance between the microcatheter tip and the shunt, the D5 push technique can be useful. D5 solution infused through the guiding catheter facilitates the advancement of glue.

Figure 8 Spinal dAVF. A 73-year-old female presented with acute onset of paraplegia and numbness for 12 hours. (A) Right T6 intercostal artery injection shows a radiculomeningeal artery travels down to T7 level and forms a shunt (small arrow) with drainage into the radicular vein (long arrow). (B) Later image shows prominent perimedullary vein in caudal direction. (C) The microcatheter injection of the dorsospinal artery depict more clear images of the feeder (small arrow), the shunt (short arrow), and perimedullary vein (long arrow). (D) Glue cast (33% NBCA) follows the course of dAVF, it outlines the feeder, shunt, and proximal portion of vein. (E) Postembolization T6 intercostal angiogram shows no residual. At three months follow-up, she was able to walk with a walker. Abbreviations: dAVF, dural arteriovenous fistula; NBCA, N-butyl 2-cyanoacrylate.

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

Following embolization, it is important to document: (1) obliteration of dAVFs by injection of segmental artery at several levels above and below the treated level (Fig. 4G, H) and (2) patency of the artery of Adamkiewicz and venous drainage of normal spinal cord. The overpenetration of glue reaching beyond the proximal draining vein can be more problematic than the proximal occlusion of feeding artery, since it can worsen the venous hypertension, the results of which may be cord infarct or hemorrhage. In fact, venous thrombosis should be considered first if the patient’s symptoms deteriorate after embolization. In those cases, intravenous heparin should be immediately started with a bolus and maintained for 24 to 48 hours with possible conversion to anticoagulation. Recurrence of dAVFs can occur by collateralization or recanalization of embolized vessels. The latter is an infrequent phenomenon when NBCA is used as an embolic agent. With the use of modern devices and NBCA, initial technical success, i.e. ‘‘adequate embolization,’’ can be obtained in close to 90% in selected cases. However, even in experienced hands, 15% to 20% of dAVFs can recur following an initially successful embolization.

PERIMEDULLARY ARTERIOVENOUS FISTULAS First described by Djindjian et al. as intradural extramedullary spinal AVMs in 1977, PM-AVFs are also called type IV spinal cord AVM (57), intradural ventral AVF (5), intradural direct AVF (18), and spinal cord AVF (7). Fundamentally, PM-AVFs are abnormal direct connections between the spinal arteries and medullary veins without nidus; the fistula is on, not within, the spinal cord, as the name describes (57).

Classification Riche et al. distinguished three types of PM-AVFs (9). Type I fistulas are slow-flow, simple, single-hole fistulas fed by a single feeder, usually the anterior spinal artery, that is slightly enlarged and often flows a long distance before ending in a small shunt (Fig. 1). It is drained by a single mildly enlarged vein, often along the posterior aspect of the cord. Type II fistulas are more voluminous and are often fed by more than one spinal artery (Fig. 9). One main feeder, usually the anterior spinal artery, can be identified along with multiple smaller posterior spinal arteries. Type III fistulas are often referred as giant perimedullary fistulas (Fig. 10). They are rapid and very high-flow giant fistulas with multiple enlarged feeders. Venous drainage is markedly dilated, ecstatic, and often appears dysplastic, and a large venous aneurysm or pouch is a characteristic finding at the level of shunt. This subdivision of PM-AVF (type I, II, and III) is applied in other classifications; the most commonly used subclassification is type IV spinal cord AVM (IVa, IVb, and IVc, respectively). Type III fistulas are mostly seen in children (58) and have a high association with HHT or Rendu-Osler-

371

Weber syndrome (19). An association with Cobb syndrome has also been reported (58,59). These syndromes of vascular malformations are known to begin in early fetal life (3–5 weeks). Rodesch et al. divided PM-AVF into two subtypes: micro-AVF (mAVF) and macro-AVF (MAVF), the latter corresponds to type III PM-AVF (7). In their series, five of six MAVF in children were associated with HHT, while there was no HHT association with mAVF. Authors suggested that presence of type III AVF in children should prompt a search for HHT, and patients as well as family members should undergo screening for pulmonary AVF, which is the main cause of disability secondary to CNS disorder— ischemia due to right to left shunt, stroke, and abscess. PM-AVF and Hemorrhagic Telangiectasia

HHT is an autosomal dominant mucocutaneous and visceral vascular dysplasia with prevalence of 1 to 10 for every 100,000 cases. Diagnosis is made when at least three clinical criteria are met: epitaxis, telangiectasia, visceral vascular malformations, and an affected first-degree relative. Two mutations of endoglin (ENG) on chromosome 9 and of activin-like receptor kinase (ALK1) on chromosome 12 have been identified and represent two subtypes of HHT, i.e. HHT 1 and HHT 2, respectively (60). A higher incidence of pulmonary AVM (40%) with HHT 1, versus 3% for HHT 2, distinguishes the two types. Telangiectasias of mucous membranes and skin causes epistaxis and gastrointestinal symptoms, which are the most common presentations of the disease. Pulmonary AVM is perhaps the most important abnormality to detect because of the relatively high incidence (10–20%) of serious consequences involving the brain (stroke, abscess). Ten to twenty percent have cerebral AVMs, which are commonly described as mAVM (nidus less than 10 mm) or small AVM (1–3 cm). Spinal cord vascular malformations are seen in 8% of HHT with neurological manifestations (61). Treatment of PMAVF in patients with HHT will be the same for those without HHT; however, the presence of other visceral organ AVMs and right to left shunt may differ treatment priority and perioperative management, including general anesthesia and anticoagulation.

Clinical A large series of spinal AVMs reported various incidences of PM AVM ranging from 13.5% (11 out of 81) to as high as 34% (27 out of 80) (6,14,18) of all spinal AVMs. This variability in many series can be explained largely by different referral patterns. In one series, dAVFs comprised only 10% of all spinal AVMs because dAVFs were treated by physicians in smaller centers. In a series of 157 intradural spinal AVMs without including dAVFs, 32 patients were found to have PM-AVFs (20%) (19). Patients are younger than those affected by dAVFs, most present in their second to fourth decades. About two-thirds are younger than 25 years and one-third less than 15 years (19). Many patients have months to years of radiculomedullary symptoms,

372

Oka and Murphy

Figure 9 PM-AVF, type II. A 31-year-old male presented with progressive urinary and bowel incontinence, erectile dysfunction, and spastic paraplegia over several months. Sagittal T2-weighted images (A, B) show expansion and extensive signal abnormality involving the high- to midthoracic cord and prominent flow voids. Right T5 intercostal artery injection (C) shows a moderate-flow fistula fed by the posterior spinal artery with reflux into the intrinsic veins. There is no apparent nidus. Later image (D) shows prominent draining veins, venous drainage through the radicular vein is also noted (arrow). Selective injection of the right T7 intercostal artery (E) demonstrates the fistula fed by a small feeder. Glue embolization of the right T5 intercostal artery (F) followed by embolization of the T7 was performed. The patient has shown slow improvement in strength and experiences no more incontinence; however, no change is noted with erectile dysfunction. Abbreviation: PM-AVF, perimedullary arteriovenous fistula. Source: Courtesy of Philippe Gailloud, Division of Interventional Neuroradiology, Johns Hopkins University (unpublished material).

which are progressive with or without episodes of acute deterioration. If untreated, complete spinal transection develops in seven to nine years (62). As in spinal dAVFs, venous ischemia due to venous congestion is likely the main cause of progressive symptoms in type I and some type II patients, and venous thrombosis may play a role in nonhemorrhagic episodic deterioration. Symptoms seen in some type II and III patients are multifactorial, i.e. not only venous hypertension but steal phenomenon and direct compression are also responsible for development of disease. A large varix at the level of shunt in type III patients causes direct compression of the spinal cord or nerve roots, which may explain asymmetric

distribution of signs and symptoms in some patients (59). Approximately one-third of patients present with spinal SAH (18,59,62). Hemorrhage likely occurs on the venous side of malformations—venous drainage is commonly seen in the posterior aspect of the cord and posterior spinal veins are located exclusively in the subarachnoid space. On the other hand, hematomyelia is likely a result of rupture of the anterior spinal vein, which is subpial in location (17). A much higher incidence of hemorrhage (SAH or hematomyelia) has been observed in pediatric populations. In one series, 70% of patients under 15 years of age presented with some type of hemorrhage (19), they also tend to present with acute symptoms rather than progressive

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

373

Figure 10 PM-AVF, type III. A 23-year-old female with known PM-AVF developed acute deterioration of spastic paraplegia with new bladder and bowel incontinence. At age 18, the diagnosis was made by angiogram (A, B) after an episode of spinal SAH. The treatment was offered but declined. T12 intercostals artery injection (A) shows a large anterior spinal artery supplying a high-flow fistula (arrow) and a large venous aneurysm (double arrow). Later image (B) shows opacification of contiguous venous pouches. At the time of treatment, the venous pouch appears more dysplastic (C). Also noted was an interval development of posterior spinal artery contribution to the fistula (arrow). (D) A detachable coil was placed first at the site of fistula (arrow), and postcoil angiogram shows a reduction of flow through the fistula. Glue embolization (E) was performed (90% NBCA) through the coil (large arrow), glue (small arrow) stays at the fistula without escaping into the venous side. Postembolization angiogram of T12 (F) shows almost complete obliteration, the posterior spinal artery is visualized. T9 intercostal artery injection (G) demonstrates contiguity of the anterior spinal artery. Abbreviations: PM-AVF, perimedullary arteriovenous fistula; SAH, subarachnoid hemorrhage; NBCA, N-butyl 2-cyanoacrylate. Source: Courtesy of Philippe Gailloud, Division of Interventional Neuroradiology, Johns Hopkins University (unpublished material).

symptoms of chronic nature. Delay in diagnosis is unfortunately very common, more than 20 years of delay has been reported (14,62).

Imaging Type II and III lesions are easily detected as prominent serpentine filling defects on myelography or perimedullary flow voids, often accompanied by signal abnormality within the cord on MRI (Fig. 9A, B). In type III lesions, integrity and architecture of the spinal cord can be very difficult to assess because of the large size of venous outflow and resulting distortion of the cord (Fig. 11). Though early reports questioned the ability of MRI to diagnose type I lesions (14,59), recent case series suggested that MRI with MRA is a reliable modality, particularly because myelotomography is not available anymore (63). Contrast administration

increases the visualization of perimedullary vessels and demonstrates intramedullary enhancement at the level of signal abnormality. For all three types of PM-AVFs, only selective spinal angiography can provide the information necessary to achieve the subclassification of a lesion and to choose its treatment. The number and size of feeders and the size and location of the fistula dictate treatment. Oblique or lateral views are often necessary. The feeders arise from various levels; however, the fistula itself is commonly located at the level of conus medullaris ranging from 64% to 75% in reported series, followed by filum terminale (64). Type III lesions can be seen in the cervical region. Some lesions, mostly type II lesions, may be mistaken as IM-AVM on angiography because of a ‘‘pseudonidus’’ appearance caused by reflux of venous flow into the intrinsic network of congested veins immediately distal to the shunt (19).

374

Oka and Murphy

Figure 11 PM-AVF, type III (same patient as in Fig. 10). Sagittal T2-weighted images of the lumbar spine (A, B) at the time of presentation demonstrate prominent tortuous flow voids around the distal spinal cord. There is a large flow void suggesting a venous aneurysm. The cord is distorted by dilated vessels, and it is difficult to evaluate the cord parenchyma (C). Abbreviation: PM-AVF, perimedullary arteriovenous fistula.

Treatment Subclassification of PM-AVF does not hold significant implications for clinical symptoms or treatment outcomes, rather indicates therapeutic approaches. The size of fistula and the size and number of feeders included in classification are critical information, as well as the location of the fistula relative to the spinal cord. Emergent intervention is not necessary in most of hemorrhagic cases as a high rate of spontaneous recovery is reported (72% by Rodesch et al.) (17). Most authors agree that surgery should be considered first in type I lesions because of the small size of the feeder and long distance to the shunt. Transarterial embolization has been attempted in a small number of patients, with reported success, when a lesion was located anterior to the conus medullaris or spinal cord (63,65). It is reasonable to try an endovascular approach first, in selected patients, as long as operators recognize proximal occlusion, which results in the same angiographic appearance as true obliteration of the shunt. Those with proximal occlusion need to undergo surgical resection without a long interval to avoid growth of more complex, recanalized fistula fed by collaterals. For type II lesions, some think that embolization is rarely effective because of multiple feeders and fistulas. Some feeders are transmedullary branches, and catheterization of those branches may not be technically possible and may not be safe. Surgery is indicated in most posterolateral AVFs, and embolization can be performed in conjunction with surgery in anterior lesions. Resection or clipping of PM-AVF that is interposed, often hidden deep behind markedly enlarged veins, is technically difficult (14). Although surgery may be the treatment of choice in type II PM-AVFs, reports on the surgical outcome of these lesions remain scarce. Type III lesions have multiple large feeders and giant venous ectasia, which represent high operative risk. Thus, embolization is the first line of treatment. Successful obliteration of giant PM-AVF has been reported with a detachable balloon (14,58); however, the balloon is not currently available in the U.S. market. Acrylic glue (NBCA) should be used

whenever transarterial access to the shunt is achievable. In a series of 22 patients with type III PM-AVF, 15 patients, whose angiogram showed complete disappearance of the lesion at the time of embolization, had recovered completely (14). Endovascular Techniques

Embolization is performed under general anesthesia. Ability to control respiration and any patient’s motion is critical when visualization of fine vasculature is critical. The patient is given 5000 units of heparin at the beginning and the dose is adjusted to maintain Activated Clotting Time (ACT) above 250 seconds. Usually, an hourly bolus of 1000 second is effective. A nonglide guiding catheter is placed at the origin of the feeding artery. A 6-Fr system is used whenever possible to acquire a better road map during navigation of the microcatheter as well as to add stability to the system. A braided microcatheter enables us to use either coils or liquid adhesive. Some newer small-diameter microcatheters (e.g., Echelon 10, Micro Therapeutics Inc., Irvine, California, U.S.) have a large inner diameter equipped for detachable coil placement, but still come with an advantage of small outer diameter that makes navigation easier and allows contrast injection through a 5-Fr guiding catheter. It is imperative to reach the site of the fistula, as proximal closure of the feeder results in development of a more complex, inaccessible lesion. Embolization should be performed with liquid adhesive (NBCA mixed with iodized oil). For a type II lesion with medium-flow, primary glue embolization is performed with various concentrations of NBCA (Fig. 9F). For very high-flow type III lesions, a glue injection following placement of coils at the site of fistula prevents glue migration through a high-flow shunt into the venous side. Placement of coils also assures accurate deposition of fast glue (Fig. 10D, E). Coil placement at the fistula may not be possible when the feeder continuously enlarges as it gets closer to the fistula, in those cases, few coil loops can be positioned in the proximal venous pouch. Communication with the anesthesiologist is important during the procedure, especially, immediately before injection of liquid

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

adhesive. The anesthesiologist must know the nature of glue embolization and that any patient motion can cause disaster or failure of embolization. Since liquid adhesive is a permanent agent, and failure of embolization usually means losing an access to the lesion in the best scenario. Some advocate a provocative test before embolization, including a balloon occlusion test of spinal arteries and injection of barbiturate or lidocaine. We do not use neurophysiological monitoring or provocative tests, rather detailed analysis of a pre-embolization microcatheter angiogram provides crucial and adequate information to decide where to deposit and when to stop injecting glue. Transient worsening of symptoms after intervention is common but most return to baseline (59). Worsening of symptoms can occur during the immediate postembolization period or in the subacute phase (4–6 weeks) after embolization. The former is likely a result of progressive retrograde thrombosis of the draining vein of fistula and veins of adjacent cord parenchyma due to sudden hemodynamic changes. The latter is secondary to mass effect and inflammatory changes of thrombosis of a large venous pouch, which peaks weeks after thrombosis happens. For the immediate postoperative period, especially in high-flow fistula, patients will be kept on intravenous heparization for 24 to 48 hours to keep the normal draining vein patent. A thrombosis can also occur on the arterial side (57). When a large varix is obliterated, intravenous

375

steroid is used during the hospitalization, which likely reduces the inflammatory effects of acute thrombus. A tapering dose of oral steroid may be added in selective cases. For patients who develop symptoms in the subacute phase, again a steroid is often used to reduce edema and inflammatory changes associated with thrombus formation.

INTRAMEDULLARY ARTERIOVENOUS MALFORMATION IM-AVM is also called type II malformation, glomus AVM, and angioma arteriovenosum. It consists of feeding arteries, nidus, and draining veins, as in cerebral AVM. The nidus of AVM can be compact, called glomus type (Fig. 12), or more diffuse in appearance involving a longer segment, called juvenile type (Fig. 13). Differentiation of the two types, although it is widely used, is very loosely defined in the literatures. Though it was a classification based on angiographic findings, it has been used for surgical lesions that lack a clear plane between the nidus and normal cord. In our opinion, the juvenile type should be used specifically to describe IM-AVM with involvement of neighboring nonneural structures, such as dura, bone, muscle, subcutaneous tissue, or skin, to avoid confusion, although initial descriptions of the juvenile type suggests this finding as common and not essential (18). In a new modified classification by Spetzler, the

Figure 12 IM-AVM. A 38-year-old female presented with progressive right lower extremity weakness associated with right-sided hip and back pain over a year. Right T9 intercostal artery injection (A) shows the anterior spinal artery (small arrows) supplying the intramedullary nidus (long arrow) at the T10–T11 level. Later image (B) shows early venous drainage through the perimedullary veins (double arrows) in both cranial and caudal directions. Lateral projection of same artery shows the anterior spinal artery and fistula in early phase (C) and drainage veins ventral to the cord in cranial direction (large arrows) and dorsal to the cord in caudal direction(small arrows) in late phase (D). Right T11 intercostal artery injection (E) reveals small contribution from the posterior spinal artery. There appears to be a small component of AVF from the right L2 lumbar artery (F). Abbreviations: IM-AVM, intramedullary arteriovenous malformations; AVF, arteriovenous fistula.

376

Oka and Murphy

Figure 13 A 14-year-old male with port wine stain in midline back over upper thoracic level, severe headache, nuchal rigidity, and new onset left upper extremity weakness with paraparesis. (A, B) Sagittal and axial T2-weighted images demonstrate flow voids throughout cord and in adjacent soft tissue and bony structures (including T1 vertebral body on sagittal images). (C, D) Left vertebral artery injection demonstrated intramedullary spinal cord AVM with ASA supply originating from intradural vertebral artery. (E) Injection of left T4 intercostal artery shows soft tissue component as well as supply to intramedullary component of AVM, feeding artery aneurysm on pedicle. (F) Selective injection of intramedullary pedicle followed by glue embolization. (G) Plain film demonstrating radiopaque glue in pedicle and nidus. (H) Postembolization—no filling of inferior intramedullary supply or aneurysm. (I) Injection of adjacent intercostals fills soft tissue AVM component. Abbreviation: AVM, arteriovenous malformation.

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

377

juvenile type is called extradural-intradural AVM (5). Other names applied to this type are metameric angiomatosis (66) and type III malformation (67). In this section, this subgroup of IM-AVM is described under metameric angiomatosis.

Clinical Manifestations Although IM-AVM is uncommon, it is two to three times more common than PM-AVFs (14,15). Patients usually present in the second and third decade of life. There is slight male predominance in many series. In a series of 90 patients with intradural AVM/AVFs, 57% were male (8). In the series by Rodesch et al., 57% of 155 patients with intradural AVM/AVFs were male but no significant gender difference was seen in the adult population (17). Approximately, one-half to two-thirds of patients experience hemorrhage (SAH and/or hematomyelia) by the time diagnosis is made (15,17,68). Hemorrhage leads to acute onset of neurological deficits in two-thirds, the rest only have back pain (characteristic severe spinal pain with or without root pain) without deficits (69). Recurrent SAH is seen in 18% to 30% (15,17). Rodesch et al. found a significantly higher risk of hemorrhage in cervical lesions (65%) compared with thoracolumbar lesions (40%) (17), concordant with other observations (8), although their series does not differentiate IM-AVM from PM-AVF. Other symptoms include back pain, motor deficits, sensory disturbance, bladder and bowel incontinence, and impotence. These symptoms usually develop in a slow progressive fashion. In the Hurth et al. series, approximately 40% of patients with untreated or partially treated intradural AVM/AVFs had acceptable neurological conditions after 15 years, the number is somewhat better in cervical lesions (8).

Imaging Diagnosis of IM-AVM is easily made by MRI. It demonstrates intramedullary nidus and dilated draining veins along the spinal cord. It provides important information of associated abnormalities such as SAH, myelomalacia, gliosis, cord edema, venous ectasia, and aneurysm. IM-AVM distributes uniformly along the longitudinal axis of the spinal cord without thoracolumbar predominance, unlike dAVFs. IM-AVM has a nidus fed by anterior spinal (radiculomedullary) or posterior spinal (radiculopial) arteries. It may be supplied directly by spinal arteries or their branches, such as sulcocommissural arteries and pial branches. Multiple arterial feeders are often present (Fig. 14A, B), which can be extensive, especially in the cervical IM-AVM due to the presence of an embryological connection of vertebral, ascending cervical, deep cervical arteries, and external carotid artery branches. Venous drainage can be ventral or dorsal to the cord and cranial or caudal in the longitudinal course. IM-AVM nidus can be confined to the cord parenchyma, or on the pial surface, or both. Intraoperatively, most AVMs have a

Figure 14 IM-AVM. A 19-year-old female presented with mild left lower extremity weakness and numbness in abdomen and left thigh. Right T9 angiogram (A) shows the right posterior spinal artery feeding a nidus at T8 level. Left T8 injection (B) shows the anterior spinal artery supply to the nidus. Embolization was performed via right T9 feeder using glue. Abbreviation: IM-AVM, intramedullary arteriovenous malformation.

varying degree of extramedullary/subpial component that is accessible to surgeons (68). Biondi et al. reported a 20% (14 out of 70) incidence of aneurysm in IM-AVMs (Fig. 13D) (15). Earlier reports presented a much lower frequency of spinal aneurysm; however, this is due to mixing of all vascular malformations, including dAVFs and PMAVFs, into one group. Many reports also misinterpreted the venous pouch in PM-AVFs as an aneurysm. There were no aneurysms associated with dAVFs (44 patients) or PM-AVFs (72 patients) in his series. SAH was present in 100% of patients with IM-AVMs and associated spinal aneurysm; of those 43% of cases had recurrent SAH. In patients with IM-AVMs and no spinal aneurysms on the angiogram, SAH was present in 70% (39 out of 56) with recurrent SAH in 13 patients (15). Following embolization of IM-AVM without obliteration of aneurysm, size of 8 out of 11 aneurysms followed the size of an AVM, suggesting that flow change is an important factor in the formation and growth of aneurysms (70). It is important to carefully study the angioarchitecture of IM-AVM, especially in the early arterial phase, to differentiate an aneurysm from a venous pouch. These aneurysms are thought to be flow-related aneurysms as a result of hemodynamic changes, though underlying dysplasia or abnormalities of local vessels are likely to contribute to aneurysm formation since an aneurysm is rarely seen in high-flow PM-AVF. Pseudoaneurysms related to rupture or dysplastic change in or near a nidus can be seen; however, it can be difficult to differentiate the true aneurysm from the pseudoaneurysm, and the Biondi et al. article did not differentiate two abnormalities. In the series of intradural AVM/AVFs by Rodesch et al., true arterial aneurysms were seen in 49 out of 155 patients (31.6%) and pseudoaneurysms in

378

Oka and Murphy

26 patients (17%). Interestingly, there was no increased risk of hemorrhage in patients with true aneurysms. Pseudoaneurysms were associated with hemorrhage in all cases (17).

Metameric Angiomatosis Metameric angiomatosis are complex vascular abnormalities involving the spinal cord in various fashions, extending into or separately affecting the vertebrae, muscle, subcutaneous tissue, and skin along the dermatome. Juvenile AVM, also referred to as type III vascular malformation, and complex spinal cord AVM are diffuse lesions that do not respect tissue boundaries and are typically seen in young patients. Wellknown associated syndromes are Cobb syndrome, Klippel-Trenaunay syndrome, and Parkes Weber syndrome. Cobb syndrome is characterized by intradural AVM/AVFs and is associated with vertebral, cutaneous, or paraspinal lesions in the same or adjacent segment (Fig. 13B, C). A cutaneous angioma can be small and subtle, but it is the hallmark of this syndrome (71). Klippel-Trenaunay syndrome manifests with cutaneous angioma and limb venolymphatic lesions without AV shunts (72–75), whereas Parkes Weber syndrome has limb lesions with high-flow shunts (76,77). The limb lesions also follow the dermatome distribution. Matsumaru et al. found 16% of metameric vascular malformations in their series of 119 spinal cord AVMs (11). They include nine Cobb, two Klippel-Trenaunay, and one Parkes Weber syndrome. There were seven cases of nonsyndromic association with bifocal intradural metameric lesions. Incidence of metameric angiomatosis in patients with intradural AVM/AVFs is 19% in the series of Rosenblum et al., 38% in a series of thoracic IM-AVM by Biondi et al., and 26% in the Hurth et al. series (excluding HHT). Obviously, these are extremely rare lesions; therefore, optimal specific treatment has not been established. Prognosis is generally poor and a complete cure of lesions should not be a goal of treatment since it is likely associated with high morbidity, given the complexity of the abnormalities. It should be targeted to the lesion or site responsible for clinical symptoms, and it may be the best to leave them alone if patients are stable and not symptomatic.

Treatment Because of the known progressive course and poor prognosis, treatment is recommended to prevent the onset or progression of symptoms. Clinically, patients with a history of hemorrhage or progressive neurological deficits should be treated more aggressively to prevent recurrent hemorrhage or progression of disease process. Treatment by any methods should aim at obliteration of AVM; however, some lesions are not curable by means of surgery or embolization. In these scenarios, the goal of treatment can be tailored to improve clinical symptoms or to target specific angioarchitecture such as aneurysm.

Embolization has become the first line of treatment in many centers, either as a primary treatment or as a preoperative adjunct (66,78). A liquid adhesive, NBCA, should be used as an embolic material, whenever feasible, especially when embolization is performed as a sole treatment. In the largest series of glue embolization of intradural AVM/AVFs by Rodesch et al., which consisted of 114 patients excluding patients who only consulted, 69 patients (60%) underwent embolization, 21 (18%) received surgery, and 24 (21%) were considered untreatable. Among 69 patients, 68 lesions were treated with acrylic glue, and a good clinical outcome was achieved (15% asymptomatic, 43% improved, 25% unchanged) at mean follow-up of 5.6 years. More than 50% reduction in size of AVM/AVF was seen in 86% of cases. Complications were transient deficits in 10 patients and permanent deficits in 9 (13%)—three of which (4%) suffered a mild disabling deficit, whereas other six deficits did not limit normal activity. All complications occurred during the embolization through the anterior spinal artery. No recanalization was noted (79). Even in one of the most experienced centers for treatment of spinal vascular malformations, a relatively large number of patients (21%) do not receive any treatment. All treating physicians need to know the limits of endovascular treatment. Arterial aneurysms on feeders of IM-AVM should be treated, especially in patients with a history of hemorrhage. Even partial treatment may be acceptable to eradicate the aneurysm and reduce risk of future hemorrhage (80). Preoperative embolization can be useful to facilitate surgical resection of certain lesions (81,82). Particle embolization was used prior to the days of liquid adhesive. Biondi et al. followed 35 patients with thoracic IM-AVM who underwent particle embolization (69). Multiple embolizations were usually required because of recanalization, which range from 1 to 17 sessions. Clinical outcome was generally good despite a high recanalization rate of 80% at the last angiographic follow-up. Clinical status was unchanged in 54% of patients, improved in 26%, and worsened in 20% as compared with status after the first embolization. Surgical resection of IM-AVM is a challenging procedure for surgeons, and several different approaches have been advocated (68,82,83). In the series by Rosenblum et al, 43 out of 54 patients with intradural AVM/AVF (43, IM-AVM; 11, PM-AVM) underwent surgical resection—neurological status improved in 33% of patients, worsened in 14%, and was stable in 51%. An acceptable outcome is seen in 84%, though persistent AVM was detected in 13 out of 32 (41%) on available postoperative angiograms (18). In another surgical series by Connolly consisting of 15 IM-AVMs, subjective improvement was noted in 80% of patients, objective neurological improvement in 40%, stability in 53%, but worsening of symptoms in only 1 patient. Recurrence was noted in 3 out of 13 patients who had follow-up angiography (68). Radiosurgery may be an alternative treatment for symptomatic lesions that are not amenable to surgery or embolization (84,85).

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

379

Figure 15 IM-AVM. An 83-year-old female with severe neck pain and quadriplegia. (A, B) Right vertebral artery injection shows PSA feeding pedicle to Spinal Cord AVM originating from right C4 level. ASA originating from right intradural vertebral artery gave small component of AVM supply. Note soft tissue AVM at right cervico-occipital level (drainage laterally into neck veins). (C–F) Plain and subtracted images of AP and lateral views of selective injection of feeding pedicle pre-embolization. (G) Lateral plain film after embolization demonstrating glue in nidus of AVM. (H, I) Postembolization RVA injection, minimal component from ASA remaining. Abbreviation: IM-AVM, intramedullary arteriovenous malformation; RVA, right vertebral artery.

Endovascular Techniques

We perform all IM-AVM embolization under general anesthesia. The patient is fully heparinized, and ACT is monitored during the procedure. A nonglide guiding catheter will be placed at the origin of the feeding artery. When an AVM is high flow, a flow-guided microcatheter can be used; however, in many instances, a braided microcatheter with support from a guidewire is needed to navigate through tortuous spinal arteries. Whenever a posterior spinal axis, including a radiculopial artery, is one of the prominent feeders, it should be chosen first as a target since embolization is safer with less risk of morbidity. The posterior spinal artery supplies the posterolateral aspect of the cord in radial fashion and multiple circumferential and longitudinal anastomoses exist between them. As described before, in patients with a history of hemorrhage, a feeder with aneurysm should be targeted first. When the anterior spinal artery (radiculomedullary artery) needs to be embolized, superselective catheterization in or near the nidus beyond the longitudinal axis of the artery is essential to keep any normal branch out of danger

(Fig. 15), which usually means catheterization of the sulcocommissural artery. Once the catheter tip is in position, a superselective angiogram should be done to confirm the absence of normal branches. Embolization is performed with NBCA mixed with iodized oil, and dilution of NBCA depends on the angioarchitecture and flow of the lesion and location of the catheter tip. We commonly use 30% to 50% of NBCA, unless high-flow shunt is noted within the nidus, for which higher concentration of NBCA is used. After embolization, patients are monitored in the neurocritical care unit for one day, then in the floor for another day or two before discharge. IV heparinization is continued for 24 hours in most of the patients to prevent progressive thrombosis of normal veins.

EXTRADURAL ARTERIOVENOUS MALFORMATION/FISTULAS Extradural AVMs/AVFs are rare lesions that can cause neurological symptoms. In this group, paraspinal or paravertebral AVM/AVF and epidural

380

Oka and Murphy

AVF are included, most common form of which is vertebral-vertebral AVF.

Clinical Several case report and case series are available (86–90), and they provide clinical, angiographic, and treatment information of 22 cases. One case was reported twice and was excluded. Extradural AVMs/AVFs are common in pediatric patients; among 22 cases, 13 patients were less than 17 years of age, including 6 infants. However, there is no significant gender prediction (10 male, 12 female). It appears to occur equally throughout the spinal axis with seven cervical, ten thoracic, one thoracolumbar, and four lumbar lesions. Several forms of clinical presentations were reported but progressive neurological deficits are the most common presentation. This deficit is caused by two mechanisms: the first is venous hypertension or congestion secondary to venous reflux into the perimedullary vein via epidural venous plexus, which is comparable to the pathophysiology of dAVFs. The second is direct compression of the spinal cord or spinal nerves by dilated arterialized veins. Other symptoms include mass effect on other organs, one infant presented with difficulty swallowing secondary to compression of esophagus (89), and another patient with Klippel-Trenaunay syndrome was found to have asymptomatic erosion of cervical vertebrae caused by high-pressure venous masses (90). High-flow shunts caused congestive heart failure because of volume and pressure overload in one neonate (89). Only one SAH has been reported, angiogram of the patient showed high-flow shunt with pseudoaneurysm on venous side (88). There are three patients with metameric angiomatosis and two patients with metameric angiomatosis and spinal vascular malformation. One patient with neurofibromatosis presented with pain and CT finding of epidural mass through neural foramen, which led to surgery with assumption of the lesion being neurofibroma (87). This lesion was later treated by endovascular method. The association of neurofibromatosis and AVF has been reported (91). On the other hand, vertebral-vertebral AVF is often asymptomatic and rarely cause neurological deficits. Most patients are found to have a bruit or experience tinnitus. The largest series of vertebral-vertebral AVFs by Beaujeux et al. reported that out of 46 patients of this condition, only 3 had neurological deficits, whereas 21 had tinnitus (92). In their series, majority were spontaneous (59%), whereas in other series 100% cases were spontaneous (93). Trauma was thought to be the cause in 41% of patients. In half of the cases vertebral-vertebral AVFs are located in C1 to C2 levels, and C5 is the other common location.

Imaging MRI often provides useful information regarding geography of lesion, presence of dilated perimedullary veins, cord compression, and coexisting abnormality (Fig. 16). In the series by Goyal et al., 3 of 10 patients

had increased signal within the cord on T2-weighted images and prominent perimedullary vessels (87). Two of them had paravertebral AVF with reflux into perimedullary veins on angiogram. The author emphasized importance of this finding, since simple clipping of the radicular vein can alter venous flow and eliminate symptoms. Extradural AVM/AVF can be located in the epidural space, bone, or adjacent soft tissue, therefore any neighboring artery can become feeder (Fig. 17A–D). The spinal artery (including radiculomedullary and radiculopial arteries) does not contribute to shunts. Venous drainage involves epidural venous plexus at some point, which is drained by paraspinal veins such as azygos or hemiazygos veins or reflux into the perimedullary veins (Fig. 16E–H).

Treatment Extradural AVF/AVM with neurological symptoms should be treated if feasible. Because of variable angioarchitecture and involved vessels, treatment needs to be individualized depending on clinical symptoms and pathophysiology. If venous congestive myelopathy is caused by perimedullary venous reflux, treating arterial side by endovascular approach will not resolve spinal venous hypertension unless complete obliteration of shunts is achieved. However, alteration of venous drainage by clipping of radicular vein will likely eliminate symptoms. The lesion can be left alone unless systemic symptoms develop. When symptoms are caused by large high-pressure veins in the epidural space, lesions usually have very high flow and feeders are multiple and dilated. Transarterial embolization using liquid adhesives is useful with high rate of success (88). One can use similar method as treating type III PM-AVF. Placement of coil at the shunt or proximal venous pouch can facilitate precise deposition of liquid adhesive without its migration to systemic veins (Fig. 16I–H). When arterial route is exhausted or there is single venous drainage with multiple arterial feeders, transvenous route can be used (87,90), coils in the epidural space do not appear to cause symptomatic compression of spinal cord. Vertebral-vertebral AVF is often neurologically asymptomatic and indication of treatment should be thought out as in benign intracranial dAVF. When a high-flow single-hole fistula is found, often caused by trauma, placement of detachable coils or balloon at the shunt or proximal venous pouch is effective in closing the shunt (Fig. 18). This procedure can be performed via transarterial or transvenous route (94,95). If it results in incomplete obliteration of the fistula, sacrifice of the vertebral artery must be considered. The vertebral artery harboring the fistula needs to be occluded distal and proximal to the fistula.

ISOLATED SPINAL ARTERY ANEURYSMS Isolated spinal artery aneurysms without associated AVM are exceedingly rare and few have been reported (96–99). Rengachary et al. reviewed literatures

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

381

Figure 16 Paraspinal AVF. A 28-year-old male presented with progressive lower extremity weakness with vascular malformation identified in the lumbar region on MRI (A–D). Further workup prior to embolization showed asymptomatic pulmonary hypertension. Pelvic angiogram (E, F) demonstrates high-flow fistula fed by anterior division of the internal iliac artery bilaterally with robust venous drainage into the internal iliac veins. Left L1 angiogram (G) demonstrates a large lumbar artery feeding the AVM nidus lateral to the spinal canal. Venous drainage is via the large epidural vein with radicular veins (H, arrows) seen at multiple levels. Nearly all intercostals and lumbar arteries from T10 to L4 contribute to AVM or AVF. Following first embolization of a few prominent feeders, he was asymptomatic for a year. Second embolization was done when he presented with sensory symptoms in the lower extremities. Left L2 lumbar angiogram (I) shows a fast-flow AVM lateral to the spinal canal. Glue embolization was performed in oblique projection that best demonstrates the proximal segment to prevent reflux (J). Unsubtracted image (K) shows the coil mass, glue cast from current and prior embolizations. Abbreviations: AVF, arteriovenous fistula; MRI, magnetic resonance imaging; AVM, arteriovenous malformation. Source: Courtesy of Philippe Gailloud, Division of Interventional Neuroradiology, Johns Hopkins University (unpublished material).

and found 57 spinal artery aneurysms (97) of which association with spinal AVM was noted in 33 cases and coarctation of aorta in 4 additional patients. Etiology of spinal artery aneurysms in other 20 patients is thought to be idiopathic in eight, congenital/vascular anomaly in four, arteritis in two, association with syphilis, fibromuscular dysplasia, and pseudoxanthoma elasticum is seen in one patient each. No information regarding etiology was available in three patients. Among eight patients with idiopathic aneurysms, five presented with SAH and two with large aneurysms (>20 mm in

diameter) with compressive symptoms. One ‘‘pea’’ size aneurysm in cervical region presented with leg paralysis without SAH. Recently, Massand et al. presented four patients, all with SAH, and three of four were thought to have dissecting aneurysms (98). Berlis et al. had three patients with spinal artery aneurysms and associated SAH. One was mycotic (Candida) and was treated with antifungal medication, which resulted in occlusion of aneurysm as well as parent spinal artery. Other two were thought to have dissecting aneurysms (99). Both authors postulated that dissection is likely the etiology

382

Oka and Murphy

Figure 17 Paraspinal AVF. An 11-year-old boy status post chest wall lymphatic malformation resection at age six and known vascular lesion in the spine suddenly developed near complete paralysis and urinary and bowel incontinence. Left suprascapular artery angiogram (A) shows the feeder of paraspinal AVF (approximate midline location is indicated by an arrow pointing an endotracheal tube). Late arterial phase (B) demonstrates a large venous aneurysm (large arrow) at the level of T5 and T6, which drains into the epidural vein on the left (small arrow). Glue embolization was performed with a microcatheter tip (C, arrow) just proximal to two main feeding branches of the AVF using D5 solution push technique. Postembolization angiogram via the left subclavian artery (D) shows no contribution from the suprascapular artery. Faint opacification of the feeder fed by small collaterals via the superior intercostals artery. Glue cast is outlined by small arrows. He underwent second embolization via the T6 intercostal artery (not shown) and recovered completely. Abbreviation: AVF, arteriovenous fistula. Source: Courtesy of Philippe Gailloud, Division of Interventional Neuroradiology, Johns Hopkins University (unpublished material).

Figure 18 Vertebral-vertebral fistula. A 23-year-old male presented with proptosis of the right eye two months after a motor vehicle accident. AP (A) and lateral (B) angiogram of the left vertebral artery show high-flow direct fistula between distal cervical segment of vertebral artery and vertebral vein. Using transarterial approach, the fistula was embolized with detachable coils (C). Postembolization AP (D) and lateral (E) angiogram show minimal flow through the coil mass, which continuously decreased over several minutes documented on repeated angiograms. Also noted is a high-flow carotid cavernous fistula (type I) (not shown).

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

in many of the idiopathic spinal artery aneurysms. Location of aneurysms varies in the artery but usually seen along the course of artery, thus fusiform shape is more common. This difference from intracranial saccular aneurysms, which are typically seen at branching point, makes treatment of spinal artery aneurysms more challenging. Trapping or wrapping of diseased segment have been performed with success since clipping and coiling are associated with higher risk of parent artery damage or occlusion (99,100). If dissection is indeed the most common etiology, wait-and-see strategy with hope for spontaneous healing might play a role when surgical exploration is not an option or is associated with greater risk than benefit.

REFERENCES 1. Djindjian R, Cophignon J, Rey A, et al. Superselective arteriographic embolization by the femoral route in neuroradiology: study of 50 cases. II. Embolization in vertebromedullary pathology. Neuroradiology 1973; 6(3):132–142. 2. Aminoff MJ, Barnard RO, Logue V. The pathophysiology of spinal vascular malformations. J Neurol Sci 1974. 23(2):255–263. 3. Aminoff MJ, Logue V. The prognosis of patients with spinal vascular malformations. Brain 1974; 97(1):211–218. 4. Kendall BE, Logue V. Spinal epidural angiomatous malformations draining into intrathecal veins. Neuroradiology 1977; 13(4):181–189. 5. Spetzler RF, Detwiler PW, Riina HA, et al. Modified classification of spinal cord vascular lesions. J Neurosurg 2002; 96(suppl 2):145–156. 6. Bao YH, Ling F. Classification and therapeutic modalities of spinal vascular malformations in 80 patients. Neurosurgery 1997; 40(1):75–81. 7. Rodesch G, Hurth M, Alvarez H, et al. Classification of spinal cord arteriovenous shunts: proposal for a reappraisal—the Bicetre experience with 155 consecutive patients treated between 1981 and 1999. Neurosurgery 2002; 51(2):374–379; discussion 379–380. 8. Hurth M, Houdart R, Djindjian R. Arteriovenous malformations of the spinal cord. clinical, anatomical, and therapeutic considerations: a series of 150 cases. Prog Neurol Surg 1978; 9:166–238. 9. Riche MC, Reizine D, Melki JP, et al. Classification of spinal cord vascular malformations. Radiat Med 1985. 3(1):17–24. 10. Krings T, Ozanne A, Chng SM, et al. Neurovascular phenotypes in hereditary haemorrhagic telangiectasia patients according to age. Review of 50 consecutive patients aged 1 day–60 years. Neuroradiology 2005; 47(10):711–720. 11. Matsumaru Y, Pongpech S, Laothomas J, et al. Multifocal and metameric spinal cord arteriovenous malformations. Review of 19 cases. Intervent Neuroradiol 1999; 5:27–34. 12. Caragine LP Jr., Halvach VV, Ng PP, et al. Vascular myelopathies-vascular malformations of the spinal cord: presentation and endovascular surgical management. Semin Neurol 2002; 22(2):123–132. 13. Merland JJR, Laurent D, Khayata A, et al. Embolization of spinal cord vascular lesions. s: Vinuela F, ed. Interventional Neuroradiology: Endovascular Therapy of the Central Nervous System. New York: Raven Press, 1992:153–165. 14. Mourier KL, Gobin YP, George B, et al. Intradural perimedullary arteriovenous fistulae: results of surgical and endovascular treatment in a series of 35 cases. Neurosurgery 1993; 32(6):885–891; discussion 891.

383

15. Biondi A, Merland JJ, Hodes JE, et al. Aneurysms of spinal arteries associated with intramedullary arteriovenous malformations. I. Angiographic and clinical aspects. AJNR Am J Neuroradiol 1992; 13(3):913–922. 16. Aminoff MJ, Logue V. Clinical features of spinal vascular malformations. Brain 1974; 97(1):197–210. 17. Rodesch G, Hurth M, Alvarez H, et al. Angio-architecture of spinal cord arteriovenous shunts at presentation. Clinical correlations in adults and children: the Bicetre experience on 155 consecutive patients seen between 1981–1999. Acta Neurochir (Wien) 2004; 146(3):217–226; discussion 226–227. 18. Rosenblum B, Oldfield EH, Doppman JL, et al. Spinal arteriovenous malformations: a comparison of dural arteriovenous fistulas and intradural AVM’s in 81 patients. J Neurosurg 1987; 67(6):795–802. 19. Rodesch G, Hurth M, Alvarez H, et al. Spinal cord intradural arteriovenous fistulae: anatomic, clinical, and therapeutic considerations in a series of 32 consecutive patients seen between 1981 and 2000 with emphasis on endovascular therapy. Neurosurgery 2005; 57(5):973–983; discussion 973–983. 20. Westphal M, Koch C. Management of spinal dural arteriovenous fistulae using an interdisciplinary neuroradiological/ neurosurgical approach: experience with 47 cases. Neurosurgery 1999; 45(3):451–457; discussion 457–458. 21. Van Dijk JM, TerBrugge KG, Willinsky RA, et al. Multidisciplinary management of spinal dural arteriovenous fistulas: clinical presentation and long-term follow-up in 49 patients. Stroke 2002; 33(6):1578–1583. 22. Symon L, Kuyama H, Kendall B. Dural arteriovenous malformations of the spine. Clinical features and surgical results in 55 cases. J Neurosurg 1984; 60(2):238–247. 23. Kinouchi H, Mizoi K, Takahashi A, et al. Dural arteriovenous shunts at the craniocervical junction. J Neurosurg 1998; 89(5):755–761. 24. Aviv RI, Shad A, Tomlinson G, et al. Cervical dural arteriovenous fistulae manifesting as subarachnoid hemorrhage: report of two cases and literature review. AJNR Am J Neuroradiol 2004; 25(5):854–858. 25. Wyburn-Mason R. The vascular abnormalities and tumors of the spinal cord and its membranes. London: Henry Kimpton, 1943. 26. Gilbertson JR, Miller GM, Goldman MS, et al. Spinal dural arteriovenous fistulas: MR and myelographic findings. AJNR Am J Neuroradiol 1995; 16(10):2049–2057. 27. Siclari F, Fasel JH, Gailloud P. Direct emergence of the dorsospinal artery from the aorta and spinal cord blood supply. Case reports and literature review. Neuroradiology 2006; 48(6):412–414. 28. Koch C. Spinal dural arteriovenous fistula. Curr Opin Neurol 2006; 19(1):69–75. 29. Foix C, Alajouanine T. Subacute necrotic myelitis, slowly progressive central myelitis with vascular hyperplasia, and slowly ascending, increasingly flaccid amyotrophic paraplegia accompanied by albuminocytologic dissiciation. Rev Neurol 1926; 33:1–42. 30. Benhaiem N, Poirier J, Hurth M. Arteriovenous fistulae of the meninges draining into the spinal veins. A histological study of 28 cases. Acta Neuropathol (Berl) 1983; 62(1–2): 103–111. 31. Finsterer J, Bavinzski G, Ungersbock K. Spinal dural arteriovenous fistula associated with syringomyelia. J Neuroradiol 2000; 27(3):211–214. 32. Mironov A. Classification of spontaneous dural arteriovenous fistulas with regard to their pathogenesis. Acta Radiol 1995; 36(6):582–592. 33. Jellema K, Tijssen CC, Fijnheer R, et al. Spinal dural arteriovenous fistulas are not associated with prothrombotic factors. Stroke 2004; 35(9):2069–2071.

384

Oka and Murphy

34. Jellema K, Canta LR, Tijssen CC, et al. Spinal dural arteriovenous fistulas: clinical features in 80 patients. J Neurol Neurosurg Psychiatry 2003; 74(10):1438–1440. 35. Cenzato M, Versari P, Righi C, et al. Spinal dural arteriovenous fistulae: analysis of outcome in relation to pretreatment indicators. Neurosurgery 2004; 55(4):815–822; discussion 822–823. 36. Koch C, Gottschalk S, Giese A. Dural arteriovenous fistula of the lumbar spine presenting with subarachnoid hemorrhage. Case report and review of the literature. J Neurosurg 2004; 100(suppl 4 Spine):385–391. 37. Mascalchi M, Mangiafico S, Marin E. Hematomyelia complicating a spinal dural arteriovenous fistula. Report of a case. J Neuroradiol 1998; 25(2):140–143. 38. Luetmer PH, Lane JI, Gilbertson JR, et al. Preangiographic evaluation of spinal dural arteriovenous fistulas with elliptic centric contrast-enhanced MR Angiography and effect on radiation dose and volume of iodinated contrast material. AJNR Am J Neuroradiol 2005; 26(4):711–718. 39. Bowen BC, Fraser K, Kochan JP, et al. Spinal dural arteriovenous fistulas: evaluation with MR angiography. AJNR Am J Neuroradiol 1995; 16(10):2029–2043. 40. Hurst RW, Grossman RI. Peripheral spinal cord hypointensity on T2-weighted MR images: a reliable imaging sign of venous hypertensive myelopathy. AJNR Am J Neuroradiol 2000; 21(4):781–786. 41. Farb RI, Kim JK, Willinsky RA, et al. Spinal dural arteriovenous fistula localization with a technique of first-pass gadolinium-enhanced MR angiography: initial experience. Radiology 2002; 222(3):843–850. 42. Saraf-Lavi E, Bowen BC, Quencer RM, et al. Detection of spinal dural arteriovenous fistulae with MR imaging and contrast-enhanced MR angiography: sensitivity, specificity, and prediction of vertebral level. AJNR Am J Neuroradiol 2002; 23(5):858–867. 43. Lai PH, Pan HB, Yang CF, et al. Multi-detector row computed tomography angiography in diagnosing spinal dural arteriovenous fistula: initial experience. Stroke 2005; 36(7):1562–1564. 44. Oldfield EH, Bennett A III, Chen MY, et al. Successful management of spinal dural arteriovenous fistulas undetected by arteriography. Report of three cases. J Neurosurg 2002; 96(suppl 2):220–229. 45. Niimi Y, Berenstein A, Setton A, et al. Embolization of spinal dural arteriovenous fistulae: results and follow-up. Neurosurgery 1997; 40(4):675–682; discussion 682–683. 46. Krings T, Mull M, Reinges MH, et al. Double spinal dural arteriovenous fistulas: case report and review of the literature. Neuroradiology 2004; 46(3):238–242. 47. Van Dijk JM, TerBrugge K, Willinsky RA, et al. Multiplicity of dural arteriovenous fistulas. J Neurosurg 2002; 96(1): 76–78. 48. Pierot L, Vlachopoulos T, Attal N, et al. Double spinal dural arteriovenous fistulas: report of two cases. AJNR Am J Neuroradiol 1993; 14(5):1109–1112. 49. Willinsky R, TerBrugge K, Lasjaunias P, et al. The variable presentations of craniocervical and cervical dural arteriovenous malformations. Surg Neurol 1990; 34(2):118–123. 50. Kohno M, Takahashi H, Ide K, et al. A cervical dural arteriovenous fistula in a patient presenting with radiculopathy. Case report. J Neurosurg 1996; 84(1):119–123. 51. Steinmetz MP, Chow MM, Krishnaney AA, et al. Outcome after the treatment of spinal dural arteriovenous fistulae: a contemporary single-institution series and meta-analysis. Neurosurgery 2004; 55(1):77–87; discussion 87–88. 52. Jellema K, Tijssen CC, van Rooij WJ, et al. Spinal dural arteriovenous fistulas: long-term follow-up of 44 treated patients. Neurology 2004; 62(10):1839–1841.

53. Song JK, Vinuela F, Gobin YP, et al. Surgical and endovascular treatment of spinal dural arteriovenous fistulas: long-term disability assessment and prognostic factors. J Neurosurg 2001; 94(suppl 2):199–204. 54. Eskandar EN, Borges LF, Budzik RF Jr., et al. Spinal dural arteriovenous fistulas: experience with endovascular and surgical therapy. J Neurosurg 2002; 96(suppl 2):162–167. 55. Nichols DA, Rufenacht DA, Jack CR Jr., et al. Embolization of spinal dural arteriovenous fistula with polyvinyl alcohol particles: experience in 14 patients. AJNR Am J Neuroradiol 1992; 13(3):933–940. 56. Morgan MK, Marsh WR. Management of spinal dural arteriovenous malformations. J Neurosurg 1989; 70(6):832–836. 57. Heros RC, Debrun GM, Ojemann RG, et al. Direct spinal arteriovenous fistula: a new type of spinal AVM. Case report. J Neurosurg 1986; 64(1):134–139. 58. Ricolfi F, Gobin PY, Aymard A, et al. Giant perimedullary arteriovenous fistulas of the spine: clinical and radiologic features and endovascular treatment. AJNR Am J Neuroradiol 1997; 18(4):677–687. 59. Halbach VV, Higashida RT, Dowd CF, et al. Treatment of giant intradural (perimedullary) arteriovenous fistulas. Neurosurgery 1993; 33(6):972–979; discussion 979–980. 60. Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). Am J Med Genet 2000; 91(1):66–67. 61. Roman G, Fisher M, Perl DP, et al. Neurological manifestations of hereditary hemorrhagic telangiectasia (RenduOsler-Weber disease): report of 2 cases and review of the literature. Ann Neurol 1978; 4(2):130–144. 62. Gueguen B, Merland JJ, Riche MC, et al. Vascular malformations of the spinal cord: intrathecal perimedullary arteriovenous fistulas fed by medullary arteries. Neurology 1987; 37(6):969–979. 63. Oran I, Parildar M, Derbent A. Treatment of slow-flow (type I) perimedullary spinal arteriovenous fistulas with special reference to embolization. AJNR Am J Neuroradiol 2005; 26(10):2582–2586. 64. Hurst RW, Bagley LJ, Marcotte P, et al. Spinal cord arteriovenous fistulas involving the conus medullaris: presentation, management, and embryologic considerations. Surg Neurol 1999; 52(1):95–99. 65. Cho KT, Lee DY, Chung CK, et al. Treatment of spinal cord perimedullary arteriovenous fistula: embolization versus surgery. Neurosurgery 2005; 56(2):232–241; discussion 232–241. 66. Merland JJ. Embolization of spinal cord vascular lesions. In: Vinuela F, ed. Interventional Neuroradiology: Endovascular Therapy of the Central Nervous System. New York: Raven Press, 1999. 67. Nelson PK, Bose A. Spinal vascular malformations: angiographic diagnosis and neuroendovascular treatment. In: Jafar JJ, Awad IA, Rosenwasser RH, eds. Vascular Malformations of the Central Nervous System. Philadelphia: Lippincott Williams & Wilkins, 1999:413–435. 68. Connolly ES Jr., Zubay IA, McCormick PC, et al. The posterior approach to a series of glomus (Type II) intramedullary spinal cord arteriovenous malformations. Neurosurgery 1998; 42(4):774–785; discussion 785–786. 69. Biondi A, Merland JJ, Reizine D, et al. Embolization with particles in thoracic intramedullary arteriovenous malformations: long-term angiographic and clinical results. Radiology 1990; 177(3):651–658. 70. Biondi A, Merland JJ, Hodes JE, et al. Aneurysms of spinal arteries associated with intramedullary arteriovenous malformations. II. Results of AVM endovascular treatment

Chapter 21: Endovascular Treatment of Spinal Vascular Malformations

71. 72.

73. 74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

and hemodynamic considerations. AJNR Am J Neuroradiol 1992; 13(3):923–931. Cobb S. Haemangioma of the spinal cord: associated with skin naevi of the same metamere. Ann Surg 1915; 62:641–649. Nakstad PH, Hald JK, Bakke SJ. Multiple spinal arteriovenous fistulas in Klippel-Trenaunay-Weber syndrome treated with platinum fibre coils. Neuroradiology 1993; 35(2):163–165. Klippel M, Trenaunay C. Du Naevus variqueux osteohypertrophique. Arch Gen Med 1900; 77:641–672. Kojima Y, Kuwana N, Sato M, et al. Klippel-TrenaunayWeber syndrome with spinal arteriovenous malformation– case report. Neurol Med Chir (Tokyo) 1989; 29(3):235–240. Gourie-Devi M, Prakash B. Vertebral and epidural hemangioma with paraplegia in Klippel-TrenaunayWeber syndrome. Case report. J Neurosurg 1978; 48(5):814–817. Ziyeh S, Spreer J, Rossler J, et al. Parkes Weber or KlippelTrenaunay syndrome? Non-invasive diagnosis with MR projection angiography. Eur Radiol 2004; 14(11): 2025–2029. Parkes-Weber F. Haemangiectatic hypertrophy of limbs congenital phlebarteriectasis and so-called congenital vericose. Br J Child Dis 1918; 15:13–23. Rodesch G, Lasjaunias P. Spinal cord arteriovenous shunts: from imaging to management. Eur J Radiol 2003; 46(3):221–232. Rodesch G, Hurth M, Alvarez H, et al. Embolization of spinal cord arteriovenous shunts: morphological and clinical follow-up and results–review of 69 consecutive cases. Neurosurgery 2003; 53(1):40–49; discussion 49–50. Konan AV, Raymond J, Roy D. Transarterial embolization of aneurysms associated with spinal cord arteriovenous malformations. Report of four cases. J Neurosurg 1999. 90(suppl 1):148–154. Barrow DL, Alleyne CH. Surgical approaches: spinal lesions. In: Jafar JJ, Awad IA, Rossenwasser RH, eds. Vascular Malformations of the Central Nervous System. Philadelphia: Lippincott-Raven, 1999:339–345. Martin NA, Khanna RK, Batzdorf U. Posterolateral cervical or thoracic approach with spinal cord rotation for vascular malformations or tumors of the ventrolateral spinal cord. J Neurosurg 1995; 83(2):254–261. Menku A, Akdemir H, Durak AC, et al. Successful surgical excision of juvenile-type spinal arteriovenous malformation in two stages following partial embolization. Minim Invasive Neurosurg 2005; 48(1):57–62. Sinclair J, Chang SD, Gibbs IC, et al. Multisession CyberKnife radiosurgery for intramedullary spinal cord arteriovenous malformations. Neurosurgery 2006; 58(6): 1081–1089; discussion 1081–1089. Hida K, Shirato H, Isu T, et al. Focal fractionated radiotherapy for intramedullary spinal arteriovenous

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

385

malformations: 10-year experience. J Neurosurg 2003; 99(suppl 1):34–38. Cognard C, Semann H, Bakchine S, et al. Paraspinal arteriovenous fistula with perimedullary venous drainage. AJNR Am J Neuroradiol 1995; 16(10):2044–2048. Goyal M, Willinsky R, Montanera W, et al. Paravertebral arteriovenous malformations with epidural drainage: clinical spectrum, imaging features, and results of treatment. AJNR Am J Neuroradiol 1999; 20(5):749–755. Rodesch G, Alvarez H, Chaskis C, et al. Clinical manifestations in paraspinal arteriovenous malformations. Int J Neuroradiol 1996; 2(5):430–436. Hui F, Trosselo MP, Meisel HJ, et al. Paraspinal arteriovenous shunts in children. Neuroradiology 1994; 36(1): 69–73. Szajner M, Weill A, Piotin M, et al. Endovascular treatment of a cervical paraspinal arteriovenous malformation via arterial and venous approaches. AJNR Am J Neuroradiol 1999; 20(6):1097–1099. Deans WR, Bloch S, Leibrock L, et al. Arteriovenous fistula in patients with neurofibromatosis. Radiology 1982; 144(1):103–107. Beaujeux RL, Reizine DC, Casasco A, et al. Endovascular treatment of vertebral arteriovenous fistula. Radiology 1992; 183(2):361–367. Goyal M. Spontaneous vertebrovertebral arteriovenous fistulae clinical features, angioarchitecture and management of twelve patients. Intervent Neuroradiol 1999; 5:219–224. Guglielmi G, Vinuela F, Duckwiler G, et al. High-flow, smallhole arteriovenous fistulas: treatment with electrodetachable coils. AJNR Am J Neuroradiol 1995; 16(2):325–328. De Keukeleire K, Defreyne L, Vanlangenhove P, et al. Asymptomatic congenital extradural vertebral arteriovenous fistula: treatment with electrolytically detachable coils. Cardiovasc Intervent Radiol 2000; 29(2):311–314. Caglar YS, Torun F, Pait G, et al. Ruptured aneurysm of the posterior spinal artery of the conus medullaris. J Clin Neurosci 2005; 12(5):603–605. Rengachary SS, Duke DA, Tsai FY, et al. Spinal arterial aneurysm: case report. Neurosurgery 1993; 33(1):125–129; discussion 129–130. Massand MG, Wallace RC, Gonzalez LF, et al. Subarachnoid hemorrhage due to isolated spinal artery aneurysm in four patients. AJNR Am J Neuroradiol 2005; 26(9):2415–2419. Berlis A, Scheufler KM, Schmahl C, et al. Solitary spinal artery aneurysms as a rare source of spinal subarachnoid hemorrhage: potential etiology and treatment strategy. AJNR Am J Neuroradiol 2005; 26(2):405–410. Gonzalez LF, Zabramski JM, Tabrizi P, et al. Spontaneous spinal subarachnoid hemorrhage secondary to spinal aneurysms: diagnosis and treatment paradigm. Neurosurgery 2005; 57(6):1127–1131; discussion 1127–1131.

22 Percutaneous Vertebroplasty Mary E. Jensen Departments of Radiology and Neurosurgery, University of Virginia Health Systems, Charlottesville, Virginia, U.S.A.

INTRODUCTION Vertebral body augmentation encompasses a group of minimally invasive therapies that fortify vertebral body insufficiency fractures due to osteoporosis, osteolytic conditions, or minor trauma. Percutaneous vertebroplasty is the best-known augmentation procedure, with a proven track record of safety and efficacy. A similar procedure called kyphoplasty, also known as ‘‘balloon-assisted vertebroplasty,’’ has recently gained popularity. Evolving techniques such as lordoplasty and new devices such as deployable grafting systems (OptiMesh, Spineology, St. Paul, Minnesota, U.S.) and permanent structural implants (StaXx Fracture Repair System, Spine Wave, Inc., Shelton, Connecticut, U.S.) are being promoted. The fundamental goal of all these procedures is to provide improved compressive strength to the vertebral body and prevent its further collapse through the introduction of a stabilizing material. The major clinical benefit of vertebral augmentation is pain relief, the mechanism of which is unclear. Other clinical benefits such as restoration of the vertebral body height and reduction of kyphotic angulation have been suggested, although not proven. Although early vertebroplasty reports focused mainly on its use in the treatment of painful vertebral hemangiomas and bony metastases, the majority of the current literature addresses augmentation procedures in osteoporotic crush fractures, since these fractures respond particularly well. This chapter will focus primarily on the clinical and technical aspects of vertebroplasty in the treatment of osteoporotic and malignant vertebral fractures.

THE BIOMECHANICS OF VERTEBRAL BODY AUGMENTATION The loss of substantive bone tissue from primary or secondary osteoporosis, tumor erosion, or osteonecrosis may lead to vertebral collapse when the axial load is more than the involved vertebral body can withstand. Polymethylmethacrylate (PMMA), an acrylic polymer noted for its excellent compressive strength (but poor shear strength), has long been used

by spine surgeons for vertebral packing following tumor debulking (1–3). During the PMMA preparation phase, liquid and powdered acrylic components are mixed together to create a ‘‘dough,’’ which is then used to fill the surgically created void. The material cures in a matter of minutes to form dense ‘‘cement.’’ The curing process is an exothermic reaction, thus generating significant heat. Extensive research on PMMA as a suitable material for vertebroplasty has been published. Biomechanical testing of PMMA injected into osteoporotic vertebral bodies demonstrates an increase in the force by almost 200% to compress treated vertebrae when compared with an untreated control group (4). Even when altered by the addition of opacification agents or antibiotic powders (5), or by changing the monomer to polymer ratio (6), the compressive strength easily surpasses that of an unadulterated osteoporotic vertebral body. When vertebrae are compressed past the point of initial failure, injected specimens are more likely to resist continued deformation than native vertebrae (7), thereby maintaining spinal axis alignment. When PMMA is applied directly to tumor tissue, the acrylic causes necrosis at the PMMA/tumor interface, probably from direct cytotoxic effects and tissue coagulation caused by the exothermic polymerization process (8). Despite the extensive biomechanical research that has been published, the mechanism of pain relief associated with vertebroplasty remains unclear. Many theories have been proposed. The mechanical, vascular, chemical, and/or thermal effects of PMMA may cause destruction of nerve endings in adjacent sensitive tissues or incite tumor necrosis. Stabilization of microfractures and decreased mechanical stresses applied to the affected vertebrae may also play a role (9). However, if this vertebral strengthening effect is the cause of the therapeutic response, one would expect to find the degree of pain relief to be proportional to the total amount of injected acrylic and the extent of vertebral filling. To date, there has been no correlation between pain relief and the volume of PMMA used (9), and the physiological explanation of the analgesic effect associated with vertebroplasty remains obscure.

388

Jensen

HISTORY OF PERCUTANEOUS VERTEBROPLASTY IN OSTEOPOROTIC VERTEBRAL COMPRESSION FRACTURES In 1987, Galibert et al. (10) described the percutaneous application of acrylic polymer (PMMA) to vertebral body defects associated with painful hemangiomas, with resultant good control of pain. Other small series followed with emphasis on the treatment of hemangiomas or metastases (11–13). In 1991, the first report of vertebroplasty in the osteoporotic spine was published. Debussche-Depriester (14) reported five patients suffering from painful osteoporotic vertebral compression fractures (VCFs), all of whom showed complete, immediate relief of pain after vertebroplasty with no or minimal residual discomfort. Vertebroplasty was virtually unknown in North America until the early 1990s; Dion and Jensen successfully treated the first patient in the United States at the University of Virginia in 1993. The first article focusing on the technical aspects of vertebroplasty was published in 1997 by the authors (15). In this small clinical trial, 29 patients with 47 osteoporotic VCFs, who had failed conservative medical therapy, underwent vertebroplasty. Ninety percent of this cohort experienced significant pain relief as evidenced by patients’ verbal expression of perceived pain and analgesic use. In 1998, Deramond et al. (16) reported the results of vertebroplasty on 80 patients with osteoporotic fractures, with rapid and complete relief of pain in greater than 90% of cases. Follow-up of 1 month to 10 years showed prolonged analgesic effect, and only a single complication was reported. The first open prospective study was not published until 1999 (17); no control group was used and the follow-up period ended at six months. Since then, several retrospective (18,19) and prospective nonrandomized studies (20–22) have shown statistically significant improvement in pain and function, particularly ambulation. These results have been confirmed in a prospective study using a control group (23) and a prospective randomized control study (24). Vertebroplasty was enthusiastically accepted by interventional radiologists and embraced by the elderly population. On the basis of the positive outcomes seen with vertebroplasty, kyphoplasty was introduced in 2001 (25) as an alternative approach, and rocketed to popularity, primarily in the surgical community.

VERTEBROPLASTY IN OSTEOPOROTIC COMPRESSION FRACTURES In 1995, an estimated 700,000 vertebral fractures occurred in elderly individuals as a sequela of osteoporosis (26). The lifetime risk of a clinically detected VCF is 15.6% for white females and 5% for white males (27). Clearly, osteoporosis of the spine and its clinical consequences are important health care and public health issues that deserve attention. Osteoporotic VCFs most likely occur in postmenopausal Caucasian and Asian females. Although the

majority of fractures result from age-related bone loss, underlying factors that may contribute to osteoporosis include steroid therapy, early oophorectomy, hypogonadism in males, hyperthyroidism, chronic obstructive pulmonary disease, immobility, anticonvulsant use, smoking, and alcohol consumption. Twenty percent of females and 40% of males presenting with vertebral or hip fractures have one of these associated conditions (28). Both low bone mass and a history of previous fracture independently predict the risk of subsequent fracture, with a sevenfold increased risk in females with low bone mass and a 25-fold risk in females with low bone mass and a single fracture (29). A vertebral fracture may be defined as reduction in vertebral height by 15% or greater, or classified by degree and type of deformity (wedge, biconcavity, or compression) (30). The most common locations for the development of a compression fracture are the T8, T12, L1, and L4 levels (31,32). The physiologic thoracic kyphosis places the greatest axial load at T8, and the thoracolumbar spinal junction is frequently affected because of the change in mobility between the relatively restricted thoracic spine and the more freely moving lumbar vertebrae (31). Although many fractures are asymptomatic, clinically detected VCFs are associated with some degree of pain in 84% of patients (33). Most fractures occur spontaneously (59%) (31) or are associated with trivial strain or exertion (32). Pain is often described as intense and deep, localized to the level of the involved vertebra, and exacerbated by palpation over the affected site (32,33). Pain is often position dependent with reduction or relief when supine, while weight bearing or bending causes the most discomfort. In some cases, pain may be referred to adjacent levels of the vertebrae as far removed as four levels, or radiate to the flank or along the ribs (32). Frank radicular pain involving the legs is uncommon (32) and may be caused by foraminal stenosis due to a retropulsed bone fragment or severe vertebral collapse. Pain associated with VCFs is usually self-limiting, lasting from two weeks to three months. For this reason, treatment of acute fractures has been largely conservative, with current medical therapy emphasizing pain control using narcotic and/or anti-inflammatory medications and strict bed rest (34). However, extended bed rest and narcotic use in the elderly is not without risk, and the decision to treat sooner rather than later should be on the basis of the patient’s overall medical condition, degree of infirmity, and rapidity of improvement. Surgery is rarely indicated, and internal fixation is reserved for patients with gross deformity, instability, or neurological deficits (35). Local application of heating pads and ice packs, massage therapy, or trigger point injections may be useful. Other treatments, such as back bracing, physical therapy, and exercise, are introduced once the patient is capable of bearing weight. Patients should be evaluated by their primary care physician, geriontologist, or endocrinologist for initiation of preventive medical therapy [bisphosphonates, calcitonin, or hormonal replacement therapy (HRT)] to prevent new fractures.

Chapter 22: Percutaneous Vertebroplasty

Quality of life and functional status are severely affected by vertebral osteoporosis. Elderly females with symptomatic fractures demonstrate significant performance impairments in physical, functional, and psychosocial testing when compared with a control group with no fractures (36). A late consequence of the disease is the development of progressive kyphosis, which may lead to chronic pain and disability, decreased exercise tolerance, fear of falling, early satiety, weight loss, and depression (33).

CONSEQUENCES OF CONSERVATIVE THERAPY Before vertebroplasty, VCFs were essentially the only fracture not treated orthopedically. As noted above, initial treatment usually is ‘‘conservative,’’ consisting of immobilization and narcotic analgesia. Although ‘‘conservative therapy’’ implied ‘‘safe,’’ it is neither benign nor risk-free, and its complications are well documented (37–39). Narcotic analgesia, commonly used in conjunction with bed rest in the treatment of acute and chronic nonmalignant musculoskeletal pain (37,40), may lead to adverse drug reactions (ADRs) in over 70% of individuals, with the elderly more likely to suffer severe ADR such as confusion. ‘‘Immobilization’’ encompasses enforced bed rest, use of braces or corsets, and pain causing protective limitations of motion. During bed rest, virtually every organ system is adversely affected, and these effects tend to be more pronounced in older patients who have less reserve than younger ones. Bone density declines approximately 2% per week (37), with the most dramatic changes seen in the first 12 weeks of immobilization. Muscle strength declines 10% to 15% per week and the rate of recovery from disuse is slower than the rate of loss (38). Decreased endurance is seen with a sense of fatigue and reduced patient motivation, setting up a vicious circle of greater inactivity. Muscle and ligament complexes are affected, resulting in muscular shortening and contracture formation (38). There is abundant evidence that early active mobilization after initial stabilization—a benefit of both vertebroplasty and kyphoplasty—is the key to contracture prevention. Early mobilization also leads to the prevention of pressure sores, the prevalence of which tends to increase significantly with age. Cardiovascular effects include increased heart rate, shorter diastolic times, and reduced coronary blood flow. Overall cardiac output, stroke volume, and left ventricular function decline as well as cerebral perfusion (38). Depending on the length of bed rest, it may take 20 to 72 days to restore pre-bed rest cardiac function. The lungs suffer from decreased ciliary clearance, less effective coughing, atelectasis, and a predilection for pneumonia. Respiratory capacity decreases by 25% to 50% from deconditioning of the respiratory muscles and restrictive impairment (39). In one study of patients immobilized by pelvic fracture, the incidence of deep vein thrombosis (DVT) was 61%, with

389

proximal DVT occurring in 29%. Pulmonary embolism is seen in 2% to 12% of patients and is fatal in 0.5% to 10% (41). Patients are at increased risk of genitourinary calculus formation, incontinence, urinary tract infections, and urosepsis. Gastrointestinal effects include reduced appetite, constipation, and fecal impaction, all exacerbated by the administration of narcotics. Even the central nervous system is not immune; patients at bed rest exhibit higher levels of anxiety, depression, insomnia, and pain intolerance (39). In a recent prospective study of 498 hospitalized patients (70 years or older), low mobility (defined as bed rest or ability to transfer to chair) and intermediate mobility (defined as ambulation one to two times with total assistance) were independent predictors of several poor hospital outcomes at discharge (42). When compared with high mobility (defined as ambulation two or more times with partial or no assistance) patients, the low and intermediate mobility patients showed decline in activities of daily living, new institutionalization, and death. The contribution of low mobility to these outcomes remained statistically significant in multivariate analyses even after controlling for age, sex, severity of illness, and comorbidities. In a recent study of vertebroplasty in an affected inpatient population (43), treatment facilitated a rapid discharge as well as long-term improvement in refractory pain. In addition, vertebroplasty leads to greater decreases in analgesic requirement when administered earlier in hospitalization. In short, conservative treatment leads to adverse outcomes associated with low mobility and bed rest, which may be viewed as iatrogenic events leading to complications such as functional decline. Restoring mobility quickly and minimizing narcotic use should be major goals of compression fracture therapy, and vertebral augmentation has proven effective for both.

PATIENT SELECTION CRITERIA The primary goal of vertebroplasty is to alleviate pain and improve mobility; vertebral body stabilization for prevention of further collapse is a secondary goal. Treatment is directed toward affected patients who have failed a reasonable course of medical therapy. Selection criteria are outlined in detail in the American College of Radiology (ACR) Standards’ ‘‘Guidelines for the Performance of Percutaneous Vertebroplasty’’ (44). All practitioners should be familiar with this document’s content. In short, appropriate candidates include patients with painful VCF refractory to medical therapy, with failure defined as no or minimal pain relief following the administration of prescription analgesics for an unspecified time period; patients who are unable to ambulate because of the pain; painful VCF associated with osteonecrosis (Ku¨mmell’s disease) (45); and unstable VCF that demonstrates movement at the wedge deformity. Patients with multiple compression deformities, who are at risk for pulmonary compromise, gastrointestinal dysfunction, or altered center of gravity if further collapse occurs,

390

Jensen

are also specified in the ACR document, although no data to support this position is available. Absolute contraindications are few. Patients with asymptomatic stable fractures or who are clearly improving with conservative treatment are not candidates. There is no evidence to support prophylactic vertebroplasty in osteopenic patients with no acute fracture. Systemic infection, osteomyelitis, uncorrectable coagulopathies, and allergic sensitivity to any of the required components are other contraindicated conditions. Although traumatic compression fracture of nonosteoporotic vertebra is considered an absolute contraindication in the ACR guidelines, a recent study has shown a positive clinical outcome from vertebroplasty in patients suffering from thoracolumbar burst fractures (46). Relative contraindications are not defined and are often operator specific. Patients with significant spinal canal compromise from retropulsed fragments, vertebra plana, or chronic fractures may be candidates, but relief is variable. Radicular pain or radiculopathy involving the lower extremities is an infrequent finding with VCFs, and an appropriate search for other compressive pathology unrelated to the collapse should be performed prior to vertebroplasty.

PATIENT SCREENING AND EVALUATION A clinical coordinator, such as a nurse, nurse practitioner, or experienced assistant, is invaluable for the smooth operation of a busy vertebroplasty service. The coordinator can collect pertinent information such as a ‘‘pain’’ history, other relevant medical conditions or previous surgeries, current analgesic use, and radiological studies, prior to scheduling an appointment. In many cases, non-candidates are discovered early on and can be redirected. Requiring a referral from an individual’s primary care physician also helps to eliminate inappropriate patients who are self-referred. Potential candidates for treatment should fulfill relevant clinical and radiological criteria, and the information should be appropriately documented in the patient’s chart. Inevitably, the practitioner will be faced with the previously treated patient who experiences a new VCF and demands an immediate vertebroplasty. Following the same screening methods used for evaluation of the initial fracture will ensure that the practitioner considers all potential pathologic processes prior to performing a second procedure.

History of Present Illness A detailed history concentrating on the patient’s back pain, mobility, relevant medication use (including analgesics, steroids, bisphosphonates, calcitonin, HRT), and general medical condition is obtained. Presenting symptoms, indications for the procedure, pertinent medical and surgical history, a list of all current medications, history of allergies, and detailed documentation of failed medical therapy are recorded. Use of visual analog scales for determining pain level,

dermatome drawings for pain localization, or questionnaires are useful for collecting data. Patients with atypical back pain should be evaluated for a concomitant disease process. Any condition that results in bacteremia, e.g., urinary tract infection, may seed the spinal column resulting in discitis or epidural abscess.

Neurological and Physical Examination A focused physical and neurological examination to identify painful vertebral levels and evaluate for possible radicular symptoms or neurological deficits is mandatory. Sites of point tenderness to percussion or palpation and positional ‘‘trigger points’’ are identified. In patients with multiple acute or subacute compression fractures, the site of point tenderness often correlates with the pain generator that should be targeted at the initial treatment. A lack of preoperative spinous process tenderness does not preclude clinical success of vertebroplasty (47). Patients with diffuse or nonfocal pain, low back pain that radiates to the hip or iliac crest, or lumbar radiculopathy may have other pathology such as facet or disc disease, which should first be excluded. Evaluation of the patient’s ability to lie prone without pulmonary compromise is recommended, particularly in individuals with known chronic obstructive pulmonary disease. A detailed physical examination is indicated when significant concurrent illnesses are suspected.

Radiological Evaluation Osteoporotic postmenopausal females with a documented new or subacute fracture on conventional radiographs and who meet the clinical criteria may proceed to vertebroplasty without other imaging. Occasionally, plain films will show intravertebral gas-filled clefts indicating the presence of avascular necrosis (45,48). Kyphotic movement at the fracture site on flexion/extension films also may be associated with a cleft (49). Bony sclerosis and osteophyte formation are indicative of healed chronic fractures. Adjunctive imaging is indicated in patients with single or multiple fractures of uncertain age, when serial conventional radiographs are unavailable or when a marrow-replacement disease process, such as multiple myeloma, is suspected. For all practical purposes, most patients have had MRI as part of their diagnostic evaluation prior to referral. MRI and/or bone scan imaging are very useful for identifying active fractures (50,51) and predicting outcome (48,52–55). Uncomplicated VCFs typically exhibit decreased signal on T1-weighted sequences (Fig. 1A) and increased or inhomogeneous signal on T2-weighted sequences (Fig. 1B) (56). Edema may involve the entire vertebral body or may be limited to the area adjacent to an endplate. Fluid-filled clefts are readily identified but their presence is underestimated when compared to vertebroplasty findings, with only 50% of clefts seen on MRI (48). Subacute or chronic painful fractures may demonstrate normal (fatty) marrow signal intensity on T1- and T2-weighted images.

Chapter 22: Percutaneous Vertebroplasty

391

Figure 1 T1-weighted sagittal image (A) shows low signal intensity involving the L3 vertebral body in addition to the inferior endplate of L5 and the superior endplate of L2. The corresponding areas on the T2-weighted image (B) show mild hyperintensity, which is inhomogeneous. STIR sequence (C) clearly identifies edema at all three levels. Abbreviation: STIR, short-tau inversion recovery.

A limited MR study consisting of T1 (Fig. 1A) and short-tau inversion recovery (STIR) (Fig. 1C) sagittal images may be the only study needed to spot vertebral body edema. Although MRI is sensitive for the detection of acute compression fractures, the duration of vertebral body edema with respect to the presence of pain is unknown. Three recent studies (53–55) have correlated clinical outcomes with preprocedural MR findings. Patients whose fractures showed extensive bone marrow edema were more likely to exhibit a positive clinical response to vertebroplasty than those patients whose fractures did not display edema. However, the lack of edema did not preclude a positive response, and these individuals should not be automatically ineligible for vertebroplasty. Bone scans (Fig. 2) and MRIs are usually positive in the first three to four months, but bone scintigraphy has been shown to be more accurate than MRI in the detection of older fractures (51). In patients suspected of having active VCFs with no obvious acute fracture on MRI, bone scintigraphy is often the next study performed. In evaluating the use of scintigraphy in preprocedural evaluation of patients being considered for vertebroplasty, Maynard et al. (52) found that a high percentage of patients (94%) achieved nearly complete pain relief after treatment of those levels that showed increased uptake of tracer, even in patients with multiple fractures of uncertain age. One pitfall of bone scanning is that activity in chronic facet disease may be confused with activity in a partially collapsed vertebral body on a routine scan (Fig. 3). SPECT scanning can localize the tracer uptake within the vertebral body as opposed to the adjacent facet joints. In patients with complex or severe fractures, computed tomography (CT) prior to vertebroplasty may be used to evaluate the integrity of the posterior wall of the vertebral body, to locate fracture lines involving the vertebral body and pedicles, to detect intravertebral gas-filled clefts, and to assess posterior displacement of fragments (Fig. 4). Canal compromise

from retropulsed bone is not considered an absolute contraindication provided there is no cord or nerve root compression resulting in neurological symptoms or dysfunction. In ambiguous cases, fluoroscopic examination of the painful sites may reveal an alternative explanation for back pain. Most common are patients with low back pain radiating to the hip who demonstrate facet hypertrophy and point tenderness over the joint. Diagnostic facet injection can be performed first as part of the screening process.

Figure 2 Anterior and posterior whole body bone scan show intense focal uptake of tracer at the L4 level consistent with a compression fracture.

392

Jensen

Figure 3 Anterior (A) and posterior (B) bone scan images show increased tracer activity throughout the whole vertebral body at L3, but focal uptake most notable overlying the posterior and lateral aspects of the vertebral hemispheres at L4. This patient was found to have a new compression fracture at L3 and significant degenerative joint disease at L4 and L5.

the surrounding epidural or paravertebral veins resulting in worsening pain or paralysis, pulmonary compromise, and death. The potential need for immediate surgical intervention should be discussed, and surgical backup at the time of the procedure must be available. Elderly patients often have chronic conditions that require special consideration. When indicated, preprocedure laboratory testing is done and often includes tests for hemoglobin, hematocrit, electrolytes and renal function, coagulation parameters, and complete blood count with differential and sedimentation rate. On an outpatient basis, individuals taking Coumadin can be given enoxaparin (Lovenox) subcutaneously once or twice a day. Lovenox can be reversed with protamine sulfate at the time of the procedure and immediately reinstituted on its completion, followed by resumption of Coumadin therapy. This process eliminates the need for a lengthy hospitalization but requires coordination with the patient’s primary care physician. Vertebroplasty should be avoided in patients with known infections, fevers, or elevated white blood count (unless due to steroid use). Patients with chronic obstructive pulmonary disease or asthma may have difficulty in breathing when lying prone, and anesthesia-managed conscious sedation may be required. General anesthesia is usually not indicated, except in the uncooperative or unstable patient.

Preprocedure Preparation and Counseling

TECHNICAL ASPECTS OF VERTEBROPLASTY

Vertebroplasty is usually performed on an outpatient basis. Important preprocedure instructions should be given at the time of the evaluation or the night prior to the procedure. Patients are asked to receive nothing by mouth after midnight and to avoid taking their morning analgesics. Transdermal narcotic patches need not be removed. A responsible adult must be available to transport the patient home after completion of the observation period. Informed consent is obtained in all cases. Risks cited should include infection, bleeding, fracture, extravasation of acrylic into

Different techniques have evolved on the basis of the predominant European (16,57,58) and North American (15,59–61) experiences. Descriptions of the procedure abound primarily in the radiology literature; variations in technique are mostly minor and related to the availability of the products and equipment utilized, and the operators’ training and personal style. However, there is no substitute for ‘‘hands-on’’ experience, and interested operators are strongly encouraged to attend one of the many educational courses currently available.

Figure 4 Coronal (A) and sagittal (B) reconstructions of a lumbar CT showing an intraosseous air-filled cleft with mild sclerosis along the inferior border. Note the vertical fracture through the posterior third of the vertebral body with gas in the disc spaces and retropulsion of the posterior fracture fragment into the spinal canal. Abbreviation: CT, computed tomography.

Chapter 22: Percutaneous Vertebroplasty

Equipment Requirements and Operator Skills Needle placement within the vertebral body has been described using standard fluoroscopy (15,16,57), CT guidance (58,62), or CT fluoroscopy (63). Regardless of the modality used to position the needle, acrylic injection into the trabecular space is, in essence, a venous embolization and should always be performed under continuous fluoroscopic observation. Operators should strive to use the highest quality fluoroscopy available, with multiple levels of magnification and small focal spot sizes. Use of a biplane digital angiography unit is ideal; biplane monitoring of fluoroscopic images decreases procedural time and enables orthogonal visualization of the acrylic injection. However, a high-quality single-plane unit that can rapidly move from the lateral to the anteroposterior (AP) positions will suffice. Low-quality analog fluoroscopy portable units are to be avoided as the image quality is usually too poor for adequate visualization of bony landmarks and acrylic flow. In addition to a high-quality imaging chain, the operator should possess appropriate cognitive and technical skills to ensure quality and safety of the study. These skills include but are not limited to knowledge of the radiographic anatomy of the spine and associated structures on both CT and fluoroscopy; formal training in radiation physics, equipment, and techniques to minimize exposure to self and patient; skill in CT or fluoroscopic-guided biopsy procedures of the spine, including radiographic triangulation; and knowledge of proper embolization technique.

Patient Preparation and Monitoring From start to finish, a dedicated nurse or other trained professional, whose primary responsibility is to establish and maintain venous access, administer conscious sedation, monitor the patient’s physiologic status, and maintain the medical record, must be present. Particularly, patients with decreased respiratory excursion when in the prone position are problematic because of unsatisfactory oxygenation. Patients with respiratory compromise may require supplemental oxygen or anesthesia support. Performing the procedure in the decubitus position or in a myelogram suite with the table tilted and the patient semierect are other options. Equipment and medications for emergency resuscitation should always be immediately available. Many patients are anxious about rolling into the prone position, and intravenous administration of 25 to 50 mg of fentanyl (Sublimaze, Abbott Laboratories, North Chicago, Illinois, U.S.) five minutes prior to positioning is useful. The patient is placed prone on the angiography table, and physiological monitors, including EKG leads, pulse oximeter, and blood pressure cuff, are attached, in addition to oxygen via nasal cannula. Additional conscious sedation may be given in the form of fentanyl and midazolam (Versed, Roche Pharma, Manati, Puerto Rico, U.S.) in small increments. To minimize infection risk, the procedure is performed under strict sterile conditions. All personnel in the room don surgical caps and masks, and the

393

operators and their assistants wear sterile gowns and gloves. The level to be treated is identified under fluoroscopy and marked, and the overlying skin surface is sterilely prepped and draped. If the drape fenestration does not expose all levels to be treated, sterile Tegaderm is applied over the area that will be covered by the drape to keep it sterile. The image intensifiers are covered with sterile bags, as they will be in close position to the sterile filed devices as well as vertebroplasty devices. Prophylactic antibiotic therapy, either given intravenously and/or mixed with the acrylic polymer, has been advocated (15,16,59–61).

Pedicle Targeting The pedicle to be punctured is isolated under AP fluoroscopy. In the simple ‘‘bulls-eye’’ approach to the pedicle, the fluoroscopic tube is either in a straight AP position or obliqued slightly. In this approach, the largest surface area of the pedicle is presented for targeting and its entire cortical circumference is easily seen. This approach is most likely to be used in the upper and midthoracic vertebral bodies (Fig. 5) as the pedicles jut posteriorly from the vertebral body at a 908 angle in the axial plane. The needle is advanced until its tip is positioned in the midportion of the ipsilateral vertebral hemisphere. If holovertebral filling is desired, a contralateral puncture may be necessary. Puncture of the pedicle using the more oblique (*208 of ipsilateral angulation) ‘‘scotty-dog’’ view will result in a steeper lateral-to-medial needle track with the final needle position near the midline of the vertebral body (Fig. 6). From this location, it is more likely that a single transpediculate injection will fill the central portion of the vertebra between the pedicles, minimizing the need for a contralateral puncture. This approach is more technically challenging since the pediculate cortex is not as well seen as it is in the ‘‘bulls-eye’’ view, and the surface area is smaller, particularly in the thoracic spine. If the needle is positioned too laterally, it may traverse the transverse process or the thoracic cavity with subsequent fracture or pneumothorax. However, the unipediculate approach results in a shorter procedure time, diminished risk as only one needle is placed, and better visualization during injection since only a single injected bolus of acrylic is observed (64). With either approach, the puncture site should avoid the medial and inferior borders of the pedicles. Tracks in these locations can result in a breach of the cortical wall and entry into the spinal canal or neural foramen. Once the angle of approach is determined, the skin, subcutaneous soft tissues, and pediculate periosteum are anesthetized with 7 to 10 cc of bupivacaine hydrochloride (0.25%) (Abbott Laboratories), using a 2-inch, 25-G spinal needle. Prior to removing this needle, AP and lateral fluoroscopy should show the tip of the needle approximating the same location on the pedicle in the superior-inferior plane. If there is a discrepancy between the two and the patient is in the true lateral position, then the AP tube needs to be adjusted in either the cranial or caudal direction until

394

Jensen

Figure 5 The pedicle to be traversed is first anesthetized using a 25-G spinal needle (A, B). In this illustration, the anesthesia needle tip is slightly inferior, and the actual puncture site will be made higher on the pedicle to parallel the superior endplate fracture (C, D). Note the difficulty in visualizing the inferior aspect of the pedicles on the lateral view (D) due to overlapping bony edges from the ribs, and burnout from the lung fields. After vertebroplasty (E, F), the PMMA is noted to fill a cavity superior to the inferior endplate. Abbreviation: PMMA, polymethylmethacrylate.

Figure 6 In the ‘‘scotty-dog’’ approach, the AP tube is obliqued approximately 208 (A). The needle is advanced through the pedicle (B) and is positioned in the midline (C), in the anterior one-third of the vertebral body (D). In this patient, a large cavity was filled within the central portion of the L4 vertebral body (E, F). A small amount of PMMA has decompressed into the needle track (arrow) but remains within the vertebra. Abbreviations: AP, anterioposterior; PMMA, polymethylmethacrylate.

Chapter 22: Percutaneous Vertebroplasty

the needle tip approximates the same location on the pedicle as on the lateral view. A small skin incision is made with a number 11 scalpel blade to allow easy passage of the vertebroplasty needle.

Positioning of the Needle A variety of disposable vertebroplasty needles or trocars are available for use, and there are no performance comparison studies among the different products that might guide selection. These devices are generally listed as ‘‘bone biopsy’’ needles and range in size from 11 to 13 G; injection of acrylic is difficult through smaller gauge needles, although 15-G needles have been used particularly in the cervical region. Important features for consideration include the availability of different stylet tip shapes and cannula sizes and lengths, radiolucency of the handle, ‘‘locking’’ of the stylet within the cannula, and compatibility of the cannula Luer lock hub with various injection devices and methacrylates. Specialty needles are also available with beveled cannulas to direct acrylic flow or with curved tips to reach specific locations in the vertebral body. (Fig. 16A, B). The needle is advanced until the stylet tip abuts the cortical surface in the superior to midpoint portion of the pedicle. Depending on the shape of the pedicle, the needle should enter at the widest point, away from the medial and inferior borders. With hourglassshaped pedicles, the operator may need to choose the extreme superior or inferior aspect for entrance. The angle of approach on the lateral view is determined by the degree of endplate compression or anterior wedging. Often the course of the needle will parallel that of the superior endplate (Fig. 5D), in which case the stylet tip position will begin more superiorly on the pedicle. On the AP view, the needle should traverse the pedicle and vertebral body from lateral to medial (Fig. 5C); otherwise it may abut or exit the lateral wall of the vertebral body. The stylet tip of the needle should be positioned precisely before a cortical break is made. Positioning is best made with a diamond-point stylet, as beveled stylets have a tendency to slip off the pedicle. Once the track is started, repositioning becomes difficult as the stylet has a tendency to slide into the initial divot. In this situation, changing the angle of approach by rotating the AP tube slightly may present a better entrance point, or the contralateral pedicle can be used instead. A slight back-and-forth twisting motion is used to advance the tip through the cortex, with frequent fluoroscopic checks in both the AP and lateral planes as the needle traverses the pedicle. Alternatively, a small sterile orthopedic hammer can be used to tap gently on the needle handle, advancing the tip in small increments. Once within the trabecular bone, less pressure is required to advance the needle and care must be taken not to pierce the endplates or vertebral wall. Use of the single-bevel stylet often will deflect the needle tip in the direction opposite to the bevel, allowing minor adjustments in either plane. The needle is advanced using continuous or intermittent

395

lateral fluoroscopy until the stylet tip is placed in the anterior one-third to one-quarter of the vertebral body. The closer the tip is to the midline on the AP view, the further anterior it may be positioned on the lateral view. Because the stylet tip projects beyond the end of the cannula, the final cannula tip position will be slightly more posterior.

Placement of a Contralateral Needle Many experienced practitioners position a single needle in the midportion of the vertebral body and perform only a single injection of acrylic, filling the midportion of the body (Fig. 6C). If the initial needle placement is within the lateral aspect of the hemivertebra, the acrylic will more than likely remain in the ipsilateral hemivertebra. Some operators prefer to fill the entire vertebra at a single sitting and will place a second needle if the initial fill pattern is deemed unsatisfactory or incomplete. Whether this procedure is necessary for a good clinical result is a matter of debate. An in vitro study by Tohmeh et al. (65), evaluating PMMA augmentation of osteoporotic vertebrae from a single or bipedicular approach, showed no significant difference in height changes between either augmented group; specifically, preferential deformation of the single-side augmented group was not noted. In a retrospective clinical study by Kim et al. (64), use of a unipediculate approach resulted in filling of both vertebral halves from a single puncture site with no statistically significant difference in clinical outcome from that of bipediculate vertebroplasty. The bipediculate approach presents unique challenges. One problem is the obscuration of the basivertebral plexus during injection by overlapping needles. Changing the lateral obliquity makes the visualization around the single needle easy, but the presence of a second trocar makes observation of this critical area difficult. Waiting to place the second needle after completion of the first injection is one solution, and if acrylic fills the contralateral hemisphere, the second injection is not needed. If both needles are placed at the same time, the contralateral stylet remains in place during the initial ipsilateral acrylic injection; otherwise, the material will track through the trabecular space and egress out the contralateral needle. The first needle can be removed prior to injection of the second hemivertebra. Another technical difficulty is observing acrylic flow during contralateral injection because of the presence of PMMA in the ipsilateral hemisphere. Potential solutions include adding extra barium sulfate to the acrylic mixture used during the contralateral injection so that it is seen through the ipsilateral acrylic cast; using final images of the ipsilateral injection displayed on an adjacent monitor as a guide by looking for acrylic extending outside of the existing cast; or injecting under a combination of lateral and AP oblique views. Use of road-mapping technique is not advised as respiratory and bowel gas movement makes precise visualization impossible.

396

Jensen

Vertebrography The initial technical description of vertebroplasty (15) advocated the use of vertebrography prior to acrylic injection as a safety feature. Injection of small amounts of contrast into the vertebral body confirms the cannula location within the trabecular space, evaluates potential routes of acrylic extravasation, and clearly defines the location of the basivertebral plexus, which channels much of the vertebral venous outflow into the anterior internal epidural venous plexus. On the lateral view, the egress point of this plexus is seen as a bony depression located anterior to the posterior vertebral body margin between the pedicles, which may not be easily visualized in osteoporotic bone. The location of this vascular junction is critically important, as extravasation of acrylic into the epidural veins is the major cause of neurological complications in vertebroplasty. Controversy exists over the need for vertebrography, particularly in the hands of experienced practitioners (66,67). Gaughen et al. (68) retrospectively evaluated the safety and efficacy of vertebroplasty performed in two patient populations, one in which venography was performed and the other without venography. No significant differences in frequency or amount of venous extravasation, or in clinical outcome between the two groups were found. However, this study was done at a major medical center by senior interventionalists with extensive experience, and its conclusions may not be valid for all operators. In short, some operators may find the vertebrogram helpful, as it easily identifies the location of the needle tip, visualizes the exact point where the basivertebral plexus exits the vertebral body, outlines the epidural and paraspinal venous system, and may predict PMMA flow characteristics and possible sites of egress (66). However, contrast that extravasates through fracture lines into the paravertebral spaces may obscure visualization and make injection of PMMA difficult to see (Fig. 8B). The decision to perform vertebrography either consistently, on a case-by-case basis, or not at all is left to the individual operator.

Acrylic Preparation Although a variety of bone filler substances have been used in the treatment of vertebral body disease, currently the only biomaterial approved for use in vertebroplasty in the United States is PMMA. There are several commercially available PMMA products that are used for vertebroplasty, all with different handling characteristics. PMMA consists of two components—a fine-grained powdered polymer and a volatile liquid monomer. When the two substances are combined, an exothermic chemical reaction begins that leads to progressive polymerization of the mixture to its solid state. Users should be familiar with PMMA prior to starting a vertebroplasty service. Bench testing is the recommended way to evaluate the material to ensure that the resultant mixture can be injected effectively through a needle and visualized fluoroscopically. This testing is best done at a formal course in which acrylic preparation and injection is performed on

models or cadavers and a variety of materials and mixtures can be tried. The major parameters of PMMA that impact its use in vertebroplasty are polymerization time and opacification. The polymerization time, or curing rate, varies among the different products, and the slurry may be suitable for injection from as little as 5 minutes to close to 20 minutes. The polymerization time of any PMMA can be prolonged by refrigerating the kit prior to its use, cooling the procedure room, or by chilling the prepared acrylic in an ice bath. For acrylics with longer curing times, the powdered polymer component needs to dissolve completely in the liquid monomer before injection. If adequate ‘‘solvation time’’ (1–2 minutes after mixing) is not allowed, the pressure from injection may cause the monomer to leach out of the mixture, leaving a powder plug in the cannula. The second parameter of great significance is opacification. As most clinically relevant complications are due to the migration of acrylic into the extraosseous spaces, fluoroscopic visualization of the material during injection is of paramount importance. Visualization is influenced by the amount of barium sulfate within the product, size of the patient, location of the treated vertebral body, and quality of the imaging chain. The percent of barium sulfate by weight or volume varies between products, and operators must be knowledgeable about their chosen material’s opacification characteristics. Sterile barium sulfate for use in vertebroplasty is commercially available and users should be prepared to supplement their mix with extra barium sulfate if necessary. Antibiotic powders for infection prophylaxis, such as tobramycin or vancomycin, also may be added to the powdered polymer. Added substances should be thoroughly mixed with the powdered polymer first to guarantee homogeneity of the slurry. Certain caveats apply—any alteration of the manufacturers’ product or mixing instructions, either by adding substances or changing the powder to liquid ratio, may change the consistency and/or polymerization time of the material. Readers are advised that, strictly speaking, any altered material is no longer FDA-approved.

Acrylic Injection Injection of the acrylic slurry is performed using either 1-mL Luer lock syringes or commercially available cement delivery systems. The 1-mL syringes are inexpensive, require minimal storage space, and allow exquisite tactile feedback during injection, which improves acrylic flow control; however, their use places the operator’s hands close to or within the radiation field. Commercially available injection devices are self-contained systems, with a reservoir into which the PMMA is loaded and a twist-type or triggeractivated plunger that advances the material into the tubing. Some newer systems also incorporate a mixing chamber attached to the reservoir, making the device entirely self-contained. The system is attached to the cannula hub via high-pressure tubing. Each turn

Chapter 22: Percutaneous Vertebroplasty

of the plunger or pull on the trigger delivers a consistent amount of acrylic into the cannula. Injection devices increase the distance between the operator and the X-ray tube, thus minimizing the dose to the hands, especially in the AP plane (69). With a delivery system, only a single connection of the tubing to the cannula hub is necessary, resulting in less exposure of the acrylic to the atmosphere and of the hub’s Luer lock threads to the acrylic. Unfortunately, the tactile feedback with delivery systems is diminished and the operator has to rely more on visual cues, such as crowding of the barium particles in the cannula, to detect compromised acrylic flow. In addition, pressure buildup in the system resulting in sudden expulsion of acrylic from the cannula tip is more likely with injection devices than 1-mL syringes. Regardless of the system used, operators should practice first on models or cadavers to become familiar with the tactile feedback and visual cues used during acrylic injection. The application of PMMA to the trabecular space is an embolization procedure, and all injections are visualized under continuous fluoroscopic monitoring. Some authors have advocated injection of small aliquots (0.1–0.2 cc) using intermittent fluoroscopy (70,71). However, as with any vascular embolization, direction of flow can shift suddenly as the embolization progresses. Small amounts of material can move quickly into unintended vascular spaces without recognition (69). Embolization of acrylic to the pulmonary system is particularly problematic because the material does not remain in the field of view, and deposition into the lungs may not be suspected until the patient becomes symptomatic. Furthermore, ‘‘small’’ aliquots are difficult to measure with injection devices and the amount perceived to be delivered may be different from what is actually delivered because of compliance in the system. Lateral imaging is used primarily to ensure that epidural extravasation of cement does not occur; intermittent AP fluoroscopy monitors any lateral paravertebral extravasation. As the acrylic exits the cannula, it permeates the trabecular space, giving the appearance of a concentrically expanding cloud (Fig. 7). Alternatively, it may seep along intraosseous cracks, leak through endplate fractures, or fill an internal cavity (Fig. 6). In some instances, vertebral body expansion with reduction of kyphotic and wedge angulation will occur (Fig. 8) (72–74). The cannula is withdrawn slightly whenever injection becomes difficult, creating a space for acrylic flow. When using an injector, forward pressure is removed prior to needle withdrawal to avoid sudden PMMA deposition into a new space. Typically, the injection is terminated when the acrylic reaches the posterior one-quarter of the vertebral body to avoid embolization of the basivertebral plexus (Fig. 7). Good pain relief occurs with filling of two-third of the vertebral body (64), and overzealous attempts at complete vertebral filling risks complication for little clinical gain. Failure of the acrylic to egress from the cannula tip may be due to obstruction from bony trabeculae, or from a blockage within the 1-mL syringe, injector tubing, or cannula. Acrylic compaction occurs when continued injection against a relative obstruction forces the

397

Figure 7 Multiple adjacent compression fractures were treated at one sitting. All three needles were placed followed by sequential PMMA injection. Note the cloud-like, wispy pattern of trabecular filling. The embolization is terminated when the PMMA reaches the posterior quarter of the vertebral body. Abbreviation: PMMA, polymethylmethacrylate.

Figure 8 This 70-year-old male presented with an acute wedge compression fracture of the L3 vertebral body (A). Following filling of an intraosseous cavity with PMMA, the fracture is reduced and the height is restored (B). The density surrounding the anterior inferior border of the vertebral body (arrows) is from contrast extravasation into the surrounding tissues during vertebrography. Abbreviation: PMMA, polymethylmethacrylate.

liquid monomer out of the slurry. The resultant plug will obstruct the cannula lumen, necessitating its removal. Compaction is best identified by the lack of movement of PMMA into the vertebra, with crowding of the constrained barium particles within the cannula. If repositioning of the cannula tip slightly posteriorly does not result in acrylic flow, then the syringe or delivery system is disconnected and evaluated for

398

Jensen

plug formation. If no obstruction is present, the cannula is cleared with the stylet under fluoroscopic observation and injection resumes. Small acrylic leaks through endplate fractures are acceptable, but large amounts of PMMA within the disc space may act as a wedge causing fracture of the adjacent vertebra (75). If the acrylic preferentially flows to a paravertebral or epidural vein, the needle is repositioned more posteriorly and the material is allowed to thicken. Injection is terminated if continued venous filling occurs. Occasionally, partially solidified PMMA will disconnect from the needle tip during withdrawal and deposit in the subcutaneous soft tissues. This retained fragment may become a source of pain or infection. Decompression of PMMA along the needle track has been seen with needle removal after filling large intraosseous cavities. Several techniques can be employed to prevent subcutaneous deposition. Prior to needle removal, a 360( twisting motion is performed to separate any stream of acrylic that may be attached to the material within the cannula dead space. The needle is withdrawn slightly and the needle track is observed for retrograde acrylic movement. It may be left in place until the acrylic hardens as long as connection between the intraosseous bolus and the needle has been broken. If necessary, redirection and advancement of the needle across the acrylic core will separate the two (76). If inadequate filling of the vertebral body requires a contralateral puncture, then the procedure is repeated on the opposite side. Otherwise, the skin incision is cleaned and dressed with small adhesive bandages, and the patient is transferred to the recovery room for further observation and care.

POSTPROCEDURAL CARE In the outpatient setting, most postvertebroplasty patients are observed for two hours prior to discharge. Individuals remain supine for one hour and are gradually allowed to sit up and/or stand over the next hour under direct nursing or physician supervision. Ambulatory patients are discharged to the care of a responsible adult after recovery. In many instances, patients experience some immediate pain relief either from the residual effects of the local anesthetic or from the procedure, or a combination of both. Patients are advised that focal pain at the puncture sites is common and may last up to 48 hours. Pain medication may be taken as needed; however, they are encouraged to limit narcotic use so that efficacy can be determined. Nonsteroidal anti-inflammatory agents such as ibuprofen may be substituted. Prior to discharge, patients are evaluated for chest or back pain, new neurological dysfunction, dyspnea, or other potential complications of the procedure. Most significant complications are due to extraosseous acrylic deposition and patients quickly become symptomatic. Early recognition is vital so that appropriate treatment can be instituted, and suspected complications should be treated as emergencies. For

this reason, immediate access to CT scanning and surgical backup is an absolute requirement for any vertebroplasty service. Small waterproof bandages placed over the puncture sites may be removed the next day, and skin incisions are kept clean and dry. Follow-up either by direct contact or telephone interview is done within 48 hours and at 7 days following the procedure. Patients are to notify the physician of redness or discharge at the operative site, recurrent or new back pain, chest pain, or shortness of breath, or unexplained fever or neurological symptoms. Any new symptom requires clinical evaluation and possibly imaging. New back pain may indicate recurrent or new fracture, unrecognized facet pain, or epidural abscess. Chest pain may be the result of rib fractures or unsuspected pulmonary embolization of acrylic. All neurological symptoms require immediate CT scanning to search for misplaced PMMA, and suspected osteomyelitis or abscess is best investigated with MRI. For people who have been immobilized for a long period of time, a gradual increase of activity is recommended. Some individuals who feel better immediately try to return to full activity only to develop new vertebral fractures or fall and break a hip or a wrist. A short course of physical therapy with continued use of a brace may be helpful. Patients who are not receiving preventative medical therapy are referred to endocrinology or geriatrics for further evaluation and implementation of appropriate treatment.

CLINICAL OUTCOMES Over 450 articles concerning vertebroplasty have been published in the last 20 years. Among these papers, about 100 studies address the clinical outcomes of patients treated with percutaneous vertebroplasty. Without exception, these reports describe vertebroplasty as a successful therapy for the relief of the pain associated with VCFs caused by either osteoporosis or tumor involvement. The earliest literature consisted of small, retrospective, uncontrolled case series introducing the technique, and claiming excellent results for the patients involved (10–14). Since that time, larger case series have been published (9,15–22, 47,52,57,58,77–79). Vertebroplasty has consistently shown immediate and considerable improvement in pain and patient mobility following treatment (15–24). In a recent prospective, nonrandomized study of 79 consecutive patients with osteoporotic compression fractures, of whom 55 (70%) were treated with vertebroplasty and 24 (30%) with conservative therapy, the vertebroplasty group showed statistically significant reduction in pain and improvement in physical functioning at 24 hours over the conservative treatment group (23). In addition, 24% of the treated patients were able to cease all analgesia after 24 hours compared to none in the conservative treatment group. These markedly different clinical outcomes at 24 hours to 1 week represent the enormous benefit of vertebroplasty over conservative therapy in terms of

Chapter 22: Percutaneous Vertebroplasty

early mobilization, even though at 6 weeks, 6 months, and 12 months the clinical outcomes were the same in the two groups. In a prospective trial of vertebroplasty versus best medical therapy (24), 40 patients with acute osteoporotic compression fractures were randomized to vertebroplasty or conservative therapy, with crossover for the medically treated group allowed at six weeks. The vertebroplasty group showed statistically significant improvement in pain and mobility, and reduction in medication use immediately after vertebroplasty. None of the patients randomized to medical therapy showed significant improvement, and 16 of the 19 patients were offered vetebroplasty. This postmedical therapy vetebroplasty group also showed statistically significant improvement in all three parameters immediately following the procedure. At 12 weeks, both groups showed statistically significant durability of the therapeutic response. It is well documented that the natural history of healing compression fractures is comprised of gradual improvement in pain over 2 to 12 weeks with variable return of function. What is not described as ‘‘natural history’’ is sudden improvement in pain and return in function—the hallmark picture of a positive therapeutic response with vertebroplasty. Most of the patients enrolled in the initial vertebroplasty studies did not undergo treatment until all noninvasive therapeutic options had been exhausted. These patients acted as their own internal controls, as vertebroplasty was performed at a point in their clinical course in which if improvement associated with healing were to occur, it should have happened. It is therefore unlikely that the rapid marked improvement in clinical findings following vertebroplasty was associated with the natural course of the disease. It may also be argued that patients treated medically are just as likely to have a long-term positive outcome similar to that of the treated population, a finding noted in the Diamond study. However, equality in long-term outcomes does not negate the early positive effects of a successful vertebroplasty. The potential complications associated with conservative therapy are more likely to happen early in the course of a patient’s immobilization, leading to physiological losses from which the patient may not recover. Another consideration is that the positive outcomes seen in vertebroplasty are due to the placebo effect. Vertebroplasty reports have consistently shown positive responses in the range of 80% to 90% for osteoporotic fractures, regardless of cohort demographics, etiology of osteoporosis, geographic location, or type of institution where the procedure was performed. The question will be laid to rest with the completion of a sham trial. A feasibility study reported in an abstract by Kallmes et al. (80) showed that patients could be successfully randomized to vertebroplasty or a sham procedure, but no meaningful clinical information was obtained. This small study was used to obtain NIH funding for a multicenter ‘‘vertebroplasty versus sham procedure’’ trial that is currently enrolling patients.

399

In summary, despite the general endorsement in the literature of the procedure to date, published reports suffer from at least one of three primary methodological limitations: (1) retrospective assessment of patient status (pain and functional ability); (2) assessment of patient status using either nonvalidated instruments, e.g., self-developed, or validated instruments nonspecific to VCFs; and (3) lack of control subjects. Though it is true that surgical therapies are rarely subjected to the same prospective randomized evaluation as medical therapies, it must be admitted that at this time, there is actually little Level 1 scientific evidence that vertebroplasty is an effective therapy.

TREATMENT OF VCFs ASSOCIATED WITH NEOPLASTIC DISEASE Malignancies involving the spinal axis are not uncommon in the cancer population. Each year, approximately 5% of cancer patients will develop spinal metastases, although not all will become clinically relevant (81). Malignant cells may access the spinal column by direct extension from a paravertebral tumor or via hematogenous, lymphatic, or perineural spread (82). Pain arises from tumor impingement on nerve roots or the spinal cord, or from periosteal irritation caused by cortical erosion and vertebral collapse. Treatment options include medical therapy, systemic chemotherapy, radiation therapy, chemoembolization, transarterial embolization, surgery, or a combination of modalities. Radiation therapy is highly effective in many patients, but relief may be delayed by two weeks, and minimal if any bone strengthening is not recognized for up to two to four months (83). Extensive multifocal disease is most likely to be treated by radiation therapy, or conventional medical therapy consisting of immobilization and corticosteroid medication or narcotic analgesia. Conservative treatment is associated with multisystem complications as discussed earlier in this chapter. Unlike osteoporotic VCFs, surgery may be indicated in selected individuals. Patients with compressive neurological symptoms from single or adjacent level disease with a life expectancy of six months or greater are often considered for surgical intervention. Most techniques consist of anterior decompression and stabilization by replacement of the affected vertebral bodies with PMMA, with the addition of pedicle screws, cages, or distraction rods as needed (1–3). However, there is a significantly increased perioperative morbidity and mortality in surgical patients who have received previous chemotherapy, radiation therapy, or steroids (2,3). A variety of minimally invasive techniques have been effective as primary or adjunctive treatment of vascular metastases. Percutaneous transarterial embolization may be used for preoperative devascularization (84,85), delivery of chemotherapeutic agents (86), and pain palliation (87,88) through shrinkage of tumor bulk. Direct percutaneous puncture with instillation of a cytotoxic material such as absolute ethanol (89)

400

Jensen

has also been described. However, with the exception of surgery, none of these therapies restore strength and stability to the spinal column, and patients remain at risk for vertebral collapse. Vertebroplasty of metastatic lesions presumably palliates pain and provides structural stability to the spine through the solidification of the osteolytic lesion. In addition, there may be some cytotoxic effect of PMMA that results in tumor necrosis and shrinkage. In some patients, vertebroplasty may facilitate surgery by providing stabilization of the anterior column in an otherwise nonsurgical candidate.

relative contraindication. Severely compressed vertebrae are technically difficult to treat, and compression of the cord or nerve roots by displaced fracture fragments and/or significant epidural tumor bulk may be made worse with the injection of PMMA. Disease involving the posterior elements cannot be treated with PMMA; instead, vascular tumors may be injected in a manner similar to that described for hemangiomas (93), with N-butyl cyanoacrylate prior to surgical removal.

Patient Selection

Spinal MRI is the most common study obtained in patients with malignant disease. MRI is extremely useful for identifying the extent of the disease, including the location and number of involved vertebrae, the presence of canal compromise, and the compression of neural structures (Fig. 9A). However, MRI does not adequately visualize the bony cortex, and CT is better at defining the degree of osseous destruction and cortical involvement (Fig. 9B) as well as identifying displaced fragments that may encroach upon the canal. The addition of intrathecal contrast in conjunction with preoperative CT scanning may provide valuable information about the presence of epidural disease. In patients with severe vertebral destruction or lesions in difficult locations such as the cervical spine, CT can be utilized for treatment planning, needle placement, and in some cases, acrylic injection. Preprocedure plain films also play an important role in lesion evaluation as it allows the operator to assess the visibility of the bony cortex, particularly the appearance of the pedicles. Since most vertebral bodies are entered via the transpedicular route under fluoroscopic guidance, poor definition of the pedicles may alter the approach or necessitate the need for needle placement under CT.

Because of the variety of therapeutic choices available to this patient population, the decision to treat is made by a multidisciplinary team and takes into consideration the local and general extent of the disease, the patient’s medical condition, response to previous therapy, general state of health, and life expectancy. In general, patients suffering from significant focal, mechanical pain unresponsive to analgesia are potential candidates for vertebroplasty (16,57,90,91). Blastic metastases with an osteolytic component may be considered for vertebroplasty. Patients with minimal or no pain but impending or frank vertebral collapse due to extensive vertebral involvement may be treated for spinal stabilization. Absolute contraindications for vertebroplasty in malignancies are the same as for osteoporotic VCFs. Relative contraindications unique to neoplastic disease also exist. In many of the early studies, patients with cortical osteolysis were excluded from treatment for fear of causing canal compromise (92); however, further studies show that these lesions can be successfully treated provided there is no significant epidural involvement (9,16). Extensive destruction and significant collapse leading to height loss of 70% or more is a

Imaging Evaluation

Figure 9 This 65-year-old male with renal cell carcinoma was evaluated for acute back pain and was found to have an L1 metastatic deposit on MRI (A). Note the marked tumor involvement of the vertebral body, and the left pedicle and lamina with epidural extension into the left lateral recess. (B) CT confirms the osseous destruction and better demonstrates the cortical erosion at the lateral recess. CT after vertebroplasty (C) shows PMMA within the tumor located in the vertebral body with extension into the lateral recess. The needle track through the right pedicle was sealed with PMMA. Lateral plane film (D) shows the PMMA bulging into the spinal canal (arrows). The PMMA within the needle track can also be seen (open arrow). The patient developed a radiculopathy that responded completely to a short course of oral steroids. Abbreviations: MRI, magnetic resonance imaging; CT, computed tomography; PMMA, polymethylmethacrylate.

Chapter 22: Percutaneous Vertebroplasty

Patients with multiple myeloma or lymphoma may have plain film findings identical to osteoporotic patients (Fig. 14A). Individuals who present with an ‘‘osteoporotic’’ fracture but who do not fit the clinical picture of primary or secondary osteoporosis (e.g., most males, African-American females) should be evaluated with MRI and biopsy, if necessary.

Special Considerations in the Performance of Vertebroplasty for Malignant Disease The basic principles that guide the injection of acrylic into osteoporotic VCFs also apply to neoplastic lesions. However, there are certain technical issues that are unique to the treatment of patients with tumor involvement. Cancer-ridden individuals are often more uncomfortable in the prone position than osteoporotic patients and may require large amounts of narcotic analgesia or general anesthesia. In addition, these people are more prone to infection as a side effect of chemotherapy or radiation treatments, and prophylactic intravenous injection of a broad spectrum antibiotic prior to vertebroplasty coupled with the antibiotic powder mixed with the PMMA is recommended. As noted above, the traditional approach for thoracic and lumbar vertebrae is via the transpedicular route. However, with tumor involvement the bony landmarks often are not seen, making the transpedicular approach problematic. One technique is to estimate the location of the affected pedicle using the medial and lateral borders of the adjacent pedicles as markers. As long as the needle tip remains lateral to the medial border of the adjacent pedicles, the needle will not pass through the spinal canal. If the transpedicular approach is not viable, the posterolateral approach for lumbar or lower thoracic vertebrae or the intercostopedicular approach for upper thoracic vertebrae may be used. Unlike osteoporosis, the cervical spine is often involved with malignant processes. The anterolateral approach for cervical lesions may be done under fluoroscopic guidance alone (94) or with the use of CT. Fluoroscopic visualization of the lower cervical vertebrae may be difficult because of the overlying shoulder density. Placement of the needle using CT guidance allows the operator to observe and avoid major vascular structures. Injection of PMMA under CT is problematic, as the 3D movement of the acrylic cannot be adequately monitored without excessive table movement. There have been recent reports of injection of minute aliquots of PMMA using CTfluoroscopy but extensive use of this technique has not been described. PMMA injection should be done under continuous fluoroscopic monitoring whenever possible, and should not be performed unless the posterior wall of the vertebral body can be seen. Vertebrography is not routinely performed in neoplastic vertebrae. Unlike osteoporosis, the significant cortical destruction and bony erosion present with tumors lead to substantial leakage of contrast into the canal and paravertebral spaces. Contrast cannot be readily removed from these spaces and its

401

presence may obscure visualization, impeding or preventing cement injection. The larger vertebroplasty trocars (10–11 G) readily accept smaller (15 G or less) biopsy systems for sampling of tumor, or a specialized vertebroplasty trocar with a biopsy cannula can be used (Fig. 10). Some authors advocate using small 15-G trocars for acrylic injection of cervical lesions, or in vertebrae with extensive destruction or extreme vascularity. Use of small needles may decrease the risk of hematoma formation or leakage of acrylic out of the needle track. Because of the variable density of tumor tissue relative to osteoporotic bone, the acrylic pattern may appear spotty and discontinuous. One exception is multiple myeloma, where the affected vertebrae may fill in a similar manner as osteoporotic fractures. Hypervascular tumors may demonstrate free arterial flow through the cannula once the stylet is removed, and the operator should be prepared to deposit acrylic material within the track during needle withdrawal to prevent excess bleeding (Fig. 9C). There has been one reported case of tumor displacement into the canal as a result of vertebroplasty (95). The introduction of intrathecal contrast immediately prior to treatment may allow the operator to identify tumor displacement during acrylic injection. Routine postprocedure CT scanning is recommended in malignant lesions. Studies provide information about the distribution of the acrylic (Fig. 9C) and may show changes in position of the tumor mass, cement leakage, intratumoral epidural cement, iatrogenic fracture or hematoma, or other unsuspected

Figure 10 This patient underwent a biopsy prior to vertebroplasty. Note the hollow cannula (arrow) extending outside of the trocar. The biopsy cannula is inserted in place of the stylet when the trocar is located at the posterior wall of the vertebral body. The trocar is then advanced to the anterior third of the vertebral body, the biopsy cannula is removed and the tissue core is retrieved.

402

Jensen

complications. It also provides a permanent record for comparison to future studies.

Vertebroplasty and Adjunctive Therapies Radiation therapy is typically used concomitantly with vertebroplasty (9,90,96,97). The timing of the radiation therapy does not affect the vertebroplasty response (96) and the radiation does not affect the integrity of the acrylic. Most investigators advocate initial treatment with vertebroplasty because of its immediate analgesic effect and improved spinal stability (9,57,90,96), followed by radiation therapy. Transarterial or direct puncture embolization of a particulate or liquid agent may be useful in reducing frank epidural disease or posterior element involvement. Recent reports of radiofrequency ablation of tumor followed by vertebroplasty have shown positive results in a small number of patients (98,99).

Clinical Outcomes As noted above, percutaneous vertebroplasty was initially described in 1987 as a treatment for painful vertebral hemangiomas (10). In 1989, Kaemmerlen et al. reported the first series limited to twenty patients with malignant disease (11). Eighty-five percent had substantial pain relief in less than 48 hours, and one complication was reported. No recurrent local pain or secondary vertebral body collapse was observed in an average follow-up period of 2.8 months. In 1996, Weill et al. (77) reported their experience with vertebroplasty in 37 patients. Seventy-three percent of patients had complete or marked relief in pain, which remained stable at six months, and in 65% of patients at one year. Recurrence of pain in five patients was attributed to the development of adjacent metastatic vertebral lesions. Five (13.5%) local complications were noted—three patients developed sciatica and two patients with treated cervical vertebrae experienced dysphagia. Three patients were successfully treated with steroid therapy, while one patient required surgical removal of epidural cement. The same year, Cotten et al. (9) published a prospective study focusing on patient outcomes in 37 patients with 40 involved vertebrae. The efficacy of the treatment was evaluated by a different physician from the one who performed the procedure, and the efficacy criterion was the change in the pain score on the McGill-Melzack verbal scale (0–5 scale). Patients were evaluated at 48 hours, 3 months, and 6 months following vertebroplasty, with a mean follow-up period of 4.2 months. All patients received radiation therapy 12 to 22 days following the procedure. Thirty-six patients (97.3%) reported some pain relief within the first 48 hours; 13.5% were pain-free, 55% showed substantial improvement, and 30% were moderately improved. The extent of vertebral body filling was compared to the degree of pain relief and no correlation was found. Of 40 vertebral bodies, 29 (72.5%) demonstrated leakage of PMMA outside of the vertebral body, but the majority of leaks were

clinically silent. Three patients (8.1%) developed radicular pain from leaks into the neural foramen, and two of these patients required decompressive surgery for relief. The beneficial effects of vertebroplasty appeared to be durable. All patients who were available for reevaluation had sustained or increased pain relief at one month, with 88.9% at three months and 75% at six months. In addition, no new or progressive vertebral collapse was seen in the injected vertebrae. Deramond et al. (16) reported their experience in 101 patients treated for spinal malignancies. Eighty percent of patients had moderate to complete pain relief following vertebroplasty, with a 10% reported complication rate. Fifty percent of treated vertebrae had osteolysis of the posterior wall, yet radicular pain accounted for only 4% of the complications. The remaining complications were associated with the patient’s primary disease. The North American literature on vertebroplasty outcomes in metastatic lesions is not as extensive as the European literature. In the first report by Barr et al. (91), 8 of 48 patients suffered from metastatic disease, of which 50% described substantial pain relief following vertebroplasty. Seven of eight patients (87.5%) demonstrated no further collapse of the treated vertebrae at follow-up. In 2003, Fourney et al. (100) described a larger series of 56 patients. Eighty-four percent of patients noted marked or complete pain relief following vertebroplasty or kyphoplasty, and no patient had worsening symptoms or suffered a complication. Asymptomatic acrylic leakage occurred in 9% of vertebroplasty patients and none in the kyphoplasty patients. Median follow-up was 4.5 months, and reductions in visual analog pain scores remained significant up to one year. All of these series point out the difficulty in determining the efficacy and durability of vertebroplasty in a patient population suffering from malignancies. Many patients are too ill to attend follow-up sessions or have died. Also, the addition of radiation therapy or other concomitant therapies cloud the picture as to the long-term benefits derived from vertebroplasty. The initial pain relief easily could be ascribed to the effects of vertebroplasty, but its role in the durability of the pain relief versus that of the additional therapy remains uncertain.

COMPLICATIONS The number of percutaneous augmentation procedures being performed around the world is steadily increasing. Neophyte operators are most likely to experience complications during their learning phase. Complications are best avoided by awareness of the factors that contribute to their occurrence. Often, it is the overzealous quest for complete vertebral body filling that results in complications, and practitioners new to the procedure must realize that ‘‘more’’ definitely is not ‘‘better’’ where augmentation is concerned. The primary cause of a symptomatic vertebroplasty complication is the passage of acrylic into

Chapter 22: Percutaneous Vertebroplasty

adjacent spaces via fracture lines or cortical destruction (Fig. 9D), along the needle track, or into the epidural and paravertebral venous complexes (9,16,57,90,96). Acrylic material located within the epidural venous plexus or foraminal veins may cause spinal cord or nerve root compression, with resultant radiculopathy or myelopathy (Fig. 11). Migration of small amounts of PMMA through the epidural or paravertebral venous system to the pulmonary vasculature (Fig. 12) is usually without clinical significance, but symptomatic pulmonary embolus and death have been reported (101). Perivertebral acrylic is usually asymptomatic, although dysphagia from esophageal compression after a cervical vertebroplasty for malignancy has occurred (96). Acrylic within the disc space may decrease its cushioning ability, leading to focal fractures at adjacent endplates (102). More often than not, PMMA leakage is asymptomatic, even in malignant lesions. Cotten et al. (9) demonstrated acrylic leaks by CT, both venous and cortical, in 29 of 40 patients with osteolytic metastases or myeloma. Most of these leaks were asymptomatic, but two of eight foraminal leaks produced nerve root compression that required decompressive surgery. In a later series, Cotten et al. (57) reported that 1 patient, of 258 treated, experienced spinal cord compression that required surgery. Of 13 patients with radicular pain, only 3 required surgical decompressions, while 10 responded to local anesthetic infiltration or medical therapy. Deramond et al. (16) noted a single transient neurological complication in 80 patients with osteoporotic fractures. Review of all major vertebroplasty

Figure 11 This patient complained of leg pain following vertebroplasty done at an outside institution. CT showed PMMA within the L2–L3 neural foramina bilaterally, causing nerve root compression. The patient responded to steroid therapy. Note the PMMA located within the paravertebral veins draining into the IVC (open arrow). Abbreviations: CT, computed tomography; PMMA, polymethylmethacrylate; IVC, inferior vena cava.

403

Figure 12 The history is unknown of this woman who presented for a chest film. A midthoracic level vertebroplasty is seen through the cardiac silhouette in addition to extensive pulmonary PMMA in the apices and the right hilum. Lateral film (not shown) demonstrated six other vertebroplasty sites. Abbreviation: PMMA, polymethylmethacrylate. Source: Courtesy of Jonas Goldstein, M.D., Asheville, North Carolina, U.S.A.

series shows that the complication rate ranges from 1% to 10%; Murphy and Deramond (90) divide it further into 1.3% for osteoporosis, 2.5% for hemangiomas, and 10% for neoplastic disease. Fortunately, most patients with radicular symptoms respond to antiinflammatory or narcotic analgesics or local anesthetic infiltration. But if significant neurological compromise were to occur, surgical colleagues must be available for immediate consultation or intervention. Vertebroplasty should only be performed at sites where surgical backup is available. Complications also may occur from inappropriate needle positioning. Improper placement of the cannula tip within or near the basivertebral plexus places the patient at risk for deposition of PMMA into the epidural venous plexus. Advancement of the needle through the anterior vertebral body wall could damage the aorta or inferior vena cava. Use of the paravertebral approach may injure the intercostal or lumbar artery. Also, leakage of PMMA into the paravertebral space through the needle track is more common with vertebrae that are accessed through the vertebral body rather than the pedicle. Transgression of the dura may lead to a symptomatic cerebrospinal fluid (CSF) leak or decompression of PMMA into the thecal sac after cannula removal (Fig. 13). Pneumothorax is a potential complication of thoracic vertebroplasty. Other complications that have occurred, as reported in the literature or through personal knowledge, include fracture of the transverse process or pedicle, paravertebral hematoma, epidural abscess, pneumothorax, rib fractures, CSF leak, seizure or respiratory arrest from oversedation, and acute disc extrusion. Severely osteoporotic patients may sustain rib fractures (15) or sternal fractures from lying prone on the procedure table. Padding the table, performing the puncture with the patient in the decubitus position, or advancing the needle through the bone with

404

Jensen

Figure 13 This 80-year-old female with a T11 compression fracture underwent vertebroplasty at an outside institution. Following the procedure, she complained of incontinence and leg weakness. Spiral CT with axial reconstruction shows decompression of PMMA along the needle track through the subarachnoid space. MRI (not shown) demonstrated application of the PMMA along the lateral aspect of the conus. Abbreviations: CT, computed tomography; PMMA, polymethylmethacrylate; MRI, magnetic resonance imaging.

the use of a hammer may help to decrease the chance of a rib fracture. Hemodynamic compromise has been associated with packing of the acetabulum with PMMA during hip replacement surgery. Transient systemic hypotension during acrylic injection in vertebroplasty has been reported (103), but a large retrospective study of the cardiovascular effects of PMMA in vertebroplasty found no generalized association between acrylic injection and systemic cardiovascular derangement (104). One theoretical complication is thermal injury to adjacent neurological structures during acrylic polymerization. There have been no clinical reports of this phenomenon and its possibility appears unlikely on the basis of in vitro tests, which showed no significant temperature rise in the spinal canal with vertebroplasty (105), and in vivo animal experiments, which showed no spinal cord damage from PMMA located adjacent to the dural sac in dogs (106). Exposure to methylmethacrylate and the radiation used for its application are potential occupational hazards to physicians, staff, and patients. Cloft et al. studied the exposure of medical personnel to methymethacrylate vapor during vertebroplasty (107). A vertebroplasty exposes the physician to less than 5 ppm (parts per million) of methacrylate vapor, well below the allowable limit of 100 ppm per eight hours set by Occupational Safety and Health Administration (OSHA). Exposure to the patient and other personnel is presumed to be even lower as they are usually further away from the acrylic mixing area. However, even though vapor exposure is negligible, some

people may experience an idiosyncratic reaction or asthma exacerbation in response to the pungent smell of the material. Radiation doses to the operator have already been described in the technical aspects portion of this chapter. When stringent radiation safety practices are followed, operators can perform up to 500 vertebroplasties in a year before reaching the extremity dose limit (69). Patient peak skin doses in vertebroplasty have also been studied (108). In 61 treated patients, the mean peak skin dose was 684 mGy, and in only 1 patient did peak skin dose exceed 2 Gy. At no time did the peak dose ever exceed 3 Gy even though more than one level was treated in many of these individuals. New fractures following vertebroplasty often occur and may represent the natural history of osteoporosis rather than a complication of the procedure. However, the issue of increased risk for fracture at an adjacent level has been raised in the literature. Grados et al. (19) found a slight but statistically significantly increased risk of vertebral fracture in the vicinity of an augmented vertebra when compared to a vertebral fracture in the vicinity of an untreated fracture. Lin et al. (102) evaluated a small group of patients who developed adjacent endplate fractures following vertebroplasty and found a higher proportion than those who did not have an acrylic leak into the adjacent disc space. These results must be considered with caution as association does not necessarily mean causation, and avoiding treatment of fractures that involve the endplate may change the clinical response (109). It has been suggested that the location and degree of height restoration of the treated level (110), or filling of an intraosseous cavity (111) may play a role in the development of adjacent fractures. In summary, complications are most commonly associated with (1) poor visualization owing to inadequate fluoroscopic equipment, poor patient cooperation (the moving target), or unsatisfactory acrylic opacification; (2) operator error, such as inappropriate patient selection; lack of knowledge of the radiographic spinal anatomy, particularly bony and venous; poor fluoroscopic-triangulation skills; unfamiliarity with equipment, devices, and PMMA; and poor embolization technique; (3) lack of patient monitoring; and (4) improper aseptic technique. By recognizing and avoiding these potential pitfalls, and thoroughly educating oneself prior to performing vertebroplasty, operators will markedly decrease their chances of causing a significant complication.

KYPHOPLASTY In 2001, a new technique for percutaneous augmentation of osteoporotic VCFs called ‘‘kyphoplasty’’ was described in the literature (25). In kyphoplasty, the vertebral body is accessed in a similar manner as vertebroplasty, but a balloon catheter (KyphX Inflatable Bone Tamp, Kyphon, Inc., Sunnyvale, California, U.S.) is used to create a space within the hemivertebra prior to acrylic injection (Fig. 14). Kyphoplasty is quite similar to vertebroplasty, differing only in the use of the

Chapter 22: Percutaneous Vertebroplasty

405

Figure 14 Kyphoplasty performed on a patient with a T12 compression fracture from biopsy-proven multiple myeloma. With the exception of the superior endplate compression, the vertebral body looks normal. After positioning of the balloons (A, B) they are symmetrically inflated (C, D), resulting in lifting of the superior endplate (arrows). After the created cavities are filled with PMMA and the cannulas are removed (E, F), the superior endplate returns to its original position. Abbreviation: PMMA, polymethylmethacrylate.

balloon, and has been referred to as ‘‘balloon-assisted kyphoplasty.’’ Theoretical advantages of kyphoplasty over vertebroplasty include the potential for vertebral body height restoration, reduction of kyphotic angulation of the spine, and lower rate of acrylic leakage into adjacent soft tissues. The realization of these benefits, however, has not been established in clinical studies. Although several papers have been published (112–121), the clinical outcome data are not as extensive as for vertebroplasty. The majority of the kyphoplasty literature addresses outcomes in the treatment of acute or subacute fractures, usually defined as vertebral levels that are edematous on MRI or show tracer uptake on nuclear medicine studies. However, similar positive outcomes in pain relief and improved mobility regardless of fracture age have been described. In one study (119), patients with fractures over four months old had nearly as favorable clinical

outcomes (87% response) as patients with fractures less than ten weeks old (90% response). However, most series describe a positive clinical response rate that is very similar to what is seen in vertebroplastytreated populations. All these reports have been hampered by the same methodologic flaws that were described for the vertebroplasty literature. Complications associated with kyphoplasty are similar to those seen in vertebroplasty. Six major complications in 531 patients (1.1%) treated with kyphoplasty were reported in a multicenter collection of patients, four of which were neurological complication. Nussbaum et al. (122) evaluated vertebroplasty and kyphoplasty complications reported to the FDA and found a number of kyphoplasty complications not reported elsewhere. Eighteen cases involving spinal canal intrusion were reported, of which thirteen cases required surgical decompression and five cases resulted

406

Jensen

Sacroplasty

Figure 15 This elderly lady underwent an L4 kyphoplasty (A) and showed clinical improvement. Six weeks later, she developed new back pain. Lateral plain film (B) shows new compression fractures from T12 through L3.

in permanent neurological deficit. Kyphoplasty has also been reported to be associated with adjacent fracture formation, occurring primarily within a two-month window following the treatment (Fig. 15) (123).

CONTROVERSIES IN VERTEBRAL BODY AUGMENTATION PROCEDURES The major controversy between kyphoplasty and vertebroplasty groups relates to height restoration, kyphosis correction, and safety. Currently, there are multiple studies that show improved height and reduced kyphosis can occur with both techniques. There is no data to support that these changes provide any additional clinical benefit. Kyphoplasty proponents state that the risk of PMMA extravasation is diminished due to the creation of a cavity, thus making kyphoplasty safer than vertebroplasty. However, studies have shown that the rate of leakage is similar, and in both instances the rate of clinical relevant complication due to leakage remains small. Because of additional equipment, anesthesia, and hospital costs, kyphoplasty has been estimated to cost 10 to 20 times more than vertebroplasty (124). It is possible that certain subgroups of patients may drive more benefit from one particular procedure. Features that might affect choice of procedure include degree of compression deformity, age of the fracture, and the presence of neoplastic involvement, but the benefits of kyphoplasty relative to vertebroplasty in such subgroups currently remain totally undefined. With the considerable added financial expense of kyphoplasty, a significant clinical benefit over vertebroplasty would have to be proven to justify this cost.

Sacral insufficiency fractures are another cause of lower back, hip, and leg pain in the elderly. These fractures are usually difficult to diagnose as the symptoms are nonspecific or radicular in nature, and the imaging findings are less obvious than those found with VCFs. Plain radiographs are often nondiagnostic or difficult to interpret because of the curvature of the sacrum and overlying bowel gas. MRI may show marrow edema, and cortical disruption and bony sclerosis may be detected on CT. The H- or butterfly-pattern of uptake in the sacrum on bone scintigraphy is typical of fractures of the body of the sacrum that involve the sacral alae (125). Like vertebroplasty, medical treatment is usually limited to immobilization and analgesics. In 2002, Garant (126) first described sacroplasty. A natural extension of vertebroplasty, sacroplasty, involves the injection of acrylic into the sacral alar fracture site under fluoroscopic control (127). The operator is faced with unique technical challenges. The sacrum is a thin, curved bone and localization of the needle tip with respect to the fracture line, the sacral foramina, and the inner cortex on the pelvic side is often problematic. Poor visualization of the sacral foramina during acrylic injection is a significant hazard. When performed under fluoroscopy, the needle is placed either through the dorsal cortex along the short axis of the sacrum (126) or in the plane of the long axis of the upper sacrum between the foramina and the sacroiliac joint (128). Contrast injection to confirm trocar placement and positioning of 20-G Chiba needles at the sacral foramina for localization may be useful safeguards (126). Placement of the trocar under CT guidance with acrylic injection under direct fluoroscopic visualization (127,129) or performance of the entire procedure under CT fluoroscopy (130) has also been described. Acrylic injected along the short axis has a tendency to pool as a round collection at the needle tip with only a short length of PMMA deposited along the fracture (128). Placement of needles at each sacral segment may be required to adequately fill the fracture (126,127). As the injection is viewed in a frontal oblique view, extravasation into the pelvis may not be detected. The long-axis approach places a vertical column of acrylic along the fracture line. With this approach, only one needle per sacral ala is needed. Since the ventral border of the sacrum is visualized during the lateral injection, extraosseous extravasation is more readily identified (128). With either approach, acrylic is deposited along the track from anterior to posterior in the short-axis approach and from superior to inferior in the long-axis approach. In the case studies and small series reported to date (126,127,129,131), almost all patients reported marked or complete relief of pain following sacroplasty. This technique has also been used in sacral fractures associated with hemangiomas (132) and metastatic lesions (131,133) (Fig. 16). Although technically more challenging than vertebroplasty, sacroplasty is another tool in the radiologist’s kit of advanced augmentation procedures.

Chapter 22: Percutaneous Vertebroplasty

407

Figure 16 Sacroplasty was performed for pain control on this patient with widespread multiple myeloma lesions. Both sacral ala were injected. Various areas of involvement were injected from a single puncture site with the use of a (A, B) curved cannula (Cardinal Medical, Inc., Chicago, Illinois, U.S.). Coronal CT reconstructions following sacroplasty (C) shows PMMA throughout the sacral ala without compromise of the neural foramina. Abbreviation: PMMA, polymethylmethacrylate.

SUMMARY Percutaneous vertebroplasty has advanced from an obscure technique reserved for a few special cases to a highly utilized procedure that consistently benefits a significant number of patients worldwide. The practice of vertebroplasty continues to grow in size and scope, fostering new developments, research, and products. Hopefully in the near future, conclusive data through randomized controlled trials will confirm what seems apparent—that vertebroplasty is safe and efficacious in the vast majority of treated individuals. It is vital to the growth of minimally invasive spine procedures that all interventional radiologists take an active interest in bringing this exciting technology to their patients and their practices.

REFERENCES 1. Harrington KD. Anterior decompression and stabilization of the spine as a treatment for vertebral collapse and spinal cord compression from metastatic malignancy. Clin Orthop 1988; 233:177–197. 2. Cybulski GR. Methods of surgical stabilization for metastatic disease of the spine. Neurosurgery 1989; 25:240–252. 3. Sundaresan N, Galicich JH, Lane JM, et al. Treatment of neoplastic epidural cord compression by vertebral body resection and stabilization. J Neurosurg 1985; 63:676–684. 4. Heini PF, Berlemann U, Kaufmann M, et al. Augmentation of mechanical properties in osteoporotic vertebral bones–a biomechnaical investigation of vertebroplasty efficacy with different bone cements. Eur Spine J 2001; 10:164–171. 5. Belkoff SM, Maroney M, Fenton DC, et al. An in vitro biomechanical evaluation of bone cements used in percutaneous vertebroplasty. Bone 1999; 25(suppl 2):23S–26S. 6. Jasper LE, Deramond H, Mathis JM, et al. The effect of monomer_to_powder ratio on the material properties of cranioplastic. Bone 1999; 25(suppl 2):27S–29S. 7. Dean JR, Ison KT, Gishen P. The strengthening effect of percutaneous vertebroplasty. Clin Radiol 2000; 55:471–476. 8. San Millan Ruiz D, Burkhardt K, Jean B, et al. Pathology findings with acrylic implants. Bone 1999; 25(suppl 2): 85S–90S. 9. Cotten A, Dewatre F, Cortet B, et al. Percutaneous vertebroplasty for osteolytic metastases and myeloma: effects of the percentage of lesion filling and the leakage of

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

methyl methacrylate at clinical follow-up. Radiology 1996; 200:525–530. Galibert P, Deramond H, Rosat P, et al. Note pre´liminaire sur le traitement des angiomes verte´braux par vertebroplastie percutane´e. Neurochirurgie 1987; 33:166–168. Kaemmerlen P, Thiesse P, Bouvard H, et al. Verte´broplastie percutane´e dans le traitement des me´tastases. Technique et re´sultats. J Radiol 1989; 70:557–562. Nguyen JP, Djindjian M, Pavlovitch JM, et al. Vertebral hemangioma with neurologic signs. Therapeutic results. Survey of the French Society of Neurosurgery. Neurochirurgie 1989; 35:299–303, 305–308. Deramond H, Darrason R, Galibert P. Percutaneous vertebroplasty with acrylic cement in the treatment of aggressive spinal angiomas. Rachis 1989; 1:143–153. Debussche-Depriester C, Deramond H, Fardellone P, et al. Percutaneous vertebroplasty with acrylic cement in the treatment of osteoporotic vertebral crush fracture syndrome. Neuroradiology 1991; 33(suppl):149–152. Jensen ME, Evans AE, Mathis JM, et al. Percutaneous polymethylmethacrylate vertebroplasty in the treatment of osteoporotic vertebral body compression fractures: technical aspects. AJNR Am J Neuroradiol 1997; 18:1897–1904. Deramond H, Depriester C, Galibert P, et al. Percutaneous vertebroplasty with polymethylmethacrylate. Technique, indications and results. Radiol Clin North Am 1998; 36:533–546. Cortet B, Cotten A, Boutry N, et al. Percutaneous vertebroplasty in the treatment of osteoporotic vertebral compression fractures: an open prospective study. J Rheumatol 1999; 26:2222–2228. Evans AJ, Jensen ME, Kip KE, et al. Vertebral compression fractures: pain reduction and improvement in functional mobility after percutaneous polymethylmethacrylate vertebroplasty retrospective report of 245 cases. Radiology 2003; 226:366–372. Grados F, Depriester C, Cayrolle G, et al. Long-term observations of vertebral osteoporotic fractures treated by percutaneous vertebroplasty. Rheumatology (Oxford) 2000; 39:1410–1414. Heini PF, Walchli B, Berlemann U. Percutaneous transpedicular vertebroplasty with PMMA: operative technique and early results. A prospective study for the treatment of osteoporotic compression fractures. Eur Spine J 2000; 9:445–450. McGraw JK, Lippert JA, Minkus KD, et al. Prospective evaluation of pain relief in 100 patients undergoing percutaneous vertebroplasty: results and follow-up. J Vasc Interv Radiol 2002; 13(9 pt 1):883–886.

408

Jensen

22. Zoarski GH, Snow P, Olan WJ, et al. Percutaneous vertebroplasty for osteoporotic compression fractures: quantitative prospective evaluation of long-term outcomes. J Vasc Interv Radiol 2002; 13(2 pt 1):139–148. 23. Diamond TH, Champion B, Clark WA. Management of acute osteoporotic vertebral fractures: a non-randomized trial comparing percutaneous vertebroplasty with conservative therapy. Am J Med 2003; 114:257–265. 24. Do HM, Marcellus ML, Weir RU, et al. Percutaneous vertebroplasty versus medical therapy for treatment of acute vertebral body compression fractures: a prospective randomized study. Proceedings of the ASNR, 2002, and Proceedings of the AUR, 2002. 25. Lieberman IH, Dudeney S, Reinhardt MK, et al. Initial outcome and efficacy of ‘‘kyphoplasty’’ in the treatment of painful osteoporotic vertebral compression fractures. Spine 2001; 26:1631–1638. 26. Riggs BL, Melton LJ. The worldwide problem of osteoporosis: insights afforded by epidemiology. Bone 1995; 17 (suppl):505S–511S. 27. Melton LJ, Chrischilles EA, Cooper C, et al. Perspective. How many women has osteoporosis?. J Bone Miner Res 1992; 7:1005–1010. 28. Riggs BL, Melton LJ. Involutional osteoporosis. N Engl J Med 1986; 314:1676–1686. 29. Ross PD, Davis SW, Epstein RS, et al. Preexisting fractures and bone mass predict vertebral fracture incidence in women. Ann Intern Med 1991; 114:919–923. 30. Eastell R, Cedel SL, Wahner HW, et al. Classification of vertebral fractures. J Bone Miner Res 1991; 6:207–215. 31. Cooper C, Atkinson EJ, O’Fallon WM, et al. Incidence of clinically diagnosed vertebral fractures: a populationbased study in Rochester, Minnesota, 1985-1989. J Bone Miner Res 1992; 7:221–227. 32. Patel U, Skingle S, Campbell GA, et al. Clinical profile of acute vertebral compression fractures in osteoporosis. Br J Rheumatol 1991; 30:418–421. 33. Silverman SL. The clinical consequences of vertebral compression fracture. Bone 1992; 13:S27–S31. 34. Rapado A. General management of vertebral fractures. Bone 1996; 18(suppl 3):191S–196S. 35. Cohen LD. Fractures of the osteoporotic spine. Orthop Clin North Am 1990; 21:143–150. 36. Lyles KW, Gold DT, Shipp KM, et al. Association of osteoporotic vertebral compression fractures with impaired functional status. Am J Med 1993; 94:595–601. 37. Cherasse A, Muller G, Ornetti P, et al. Tolerability of opioids in patients with acute pain due to non-malignant musculoskeletal disease. A hospital-based observational study. Joint Bone Spine 2004; 71:572–576. 38. Dittmer DK, Teasell R. Complications of immobilization and bedrest. Part 1: Musculoskeletal and cardiovascular complications. Can Fam Physician 1993; 39:1428–1432. 39. Teasell R, Dittmer DK. Complications of immobilization and bedrest. Part 2: Other complications. Can Fam Physician 1993; 39:1440–1442. 40. Turk DC, Brody MC, Okifuji EA. Physicians’ attitudes and practices regarding the long-term prescribing of opioids for non-cancer pain. Pain 1994; 59:201–208. 41. Babayev M, Lachmann E, Nagler W. The controversy surrounding sacral insufficiency fractures: to ambulate or not to ambulate?. Am J Phys Med Rehabil 2000; 79:404–409. 42. Brown CJ, Friedkin RJ, Inouye SK. Prevalence and outcomes of low mobility in hospitalized older patients. J Am Geriatr Soc 2004; 53:1263–1270. 43. Trout AT, Gray LA, Kallmes DF. Vertebroplasty in the inpatient population. AJNR Am J Neuroradiol 2005; 26:1629–1633.

44. Barr JD, Mathis JM, Barr MS, et al. Standard for the performance of percutaneous vertebroplasty. In: American College of Radiology Standards 2000–2001. Reston, VA: American College of Radiology, 2000:441–448. 45. Do HM, Jensen ME, Marx WF, et al. Percutaneous vertebroplasty in vertebral osteonecrosis (Ku¨mmell’s spondylitis). Neurosurg Focus 1999; 7(1):e2. 46. Chen JF, Lee ST. Percutaneous vertebroplasty for treatment of thoracolumbar spine bursting fracture. Surg Neurol 2004; 62:494–500. 47. Gaughen JR Jr., Jensen ME, Schweickert PA, et al. Lack of preoperative spinous process tenderness does not affect clinical success of percutaneous vertebroplasty. J Vasc Interv Radiol 2002; 13:1135–1138. 48. Lane JI, Maus TP, Wald JT, et al. Intravertebral clefts opacified during vertebroplasty: pathogenesis, technical implications, and prognostic significance. AJNR Am J Neuroradiol 2002; 23:1642–1646. 49. Kim DY, Lee SH, Jang JS, et al. Intravertebral vacuum phenomenon in osteoporotic compression fracture: report of 67 cases with quantitative evaluation of intravertebral instability. J Neurosurg 2004; 100(suppl 1 Spine):24–31. 50. Do HM. Magnetic resonance imaging in the evaluation of patients for percutaneous vertebroplasty. Top Magn Reson Imaging 2000; 11:235–244. 51. Masala S, Shillaci O, Massari F, et al. MRI and bone scan imaging in the preoperative evaluation of painful vertebral fractures treated with vertebroplasty and kyphoplasty. In Vivo 2005; 19:1055–1060. 52. Maynard AS, Jensen ME, Schweickert PA, et al. Value of bone scan imaging in predicting pain relief from percutaneous vertebroplasty in osteoporotic vertebral fractures. AJNR Am J Neuroradiol 2000; 21:1807–1812. 53. Brown DB, Glaiberman CB, Gilula LA, et al. Correlation between preprocedural MRI findings and clinical outcomes in the treatment of chronic symptomatic vertebra compression fractures with percutaneous vertebroplasty. AJR Am J Roentgenol 2005; 184:1951–1955. 54. Tanigawa N, Komemushi A, Kariya S, et al. Percutaneous vertebroplasty: relationship between vertebral body bone marrow edema pattern on MR images and initial clinical response. Radiology 2006; 239:195–200. 55. Alvarez L, Pirez-Hiqueras A, Grazino JJ, et al. Predictors of outcomes of percutaneous vertebroplasty for osteoporotic vertebral fractures. Spine 2005; 30:87–92. 56. Baker LL, Goodman SB, Perkash I, et al. Benign versus pathological compression fractures of vertebral bodies: assessment with conventional spin-echo, chemical-shift, and STIR MR imaging. Radiology 1990; 174:495–502. 57. Cotten A, Boutry N, Cortet B, et al. Percutaneous vertebroplasty: state of the art. Radiographics 1998; 18:311–320. 58. Gangi A, Guth S, Imbert JP, et al. Percutaneous vertebroplasty: indications, technique, and results. Radiographics 2003; 23:e10. 59. Mathis JM, Barr JD, Belkoff SM, et al. Percutaneous vertebroplasty: a developing standard of care for vertebral compression fractures. A review. AJNR Am J Neuroradiol 2001; 22:373–381. 60. Jensen ME, Dion JE. Percutaneous vertebroplasty in osteoporotic compression fractures. Neuroimaging Clin N Am 2000; 10:547–568. 61. Kallmes DF, Jensen ME. Percutaneous vertebroplasty. Radiology 2003; 229:27–36. 62. Barr MS, Barr JD. Invited commentary. Radiographics 1998; 18:320–321. 63. Kim JH, Park KS, Yi S, et al. Real-time CT fluoroscopy (CTF)-guided vertebroplasty in osteoporotic spine fractures. Yonsei Med J 2005; 46:635–642.

Chapter 22: Percutaneous Vertebroplasty 64. Kim AK, Jensen ME, Dion JE, et al. Unilateral transpedicular percutaneous vertebroplasty: initial experience. Radiology 2002; 222:737–741. 65. Tohmeh AG, Mathis JM, Fenton DC, et al. Biomechanical efficacy of unipedicular versus bipedicular vertebroplasty for the management of osteoporotic compression fractures. Spine 1999; 24:1772–1776. 66. McGraw JK, Heatwole EV, Strnad BT, et al. Predictive value of intraosseous venography before percutaneous vertebroplasty. J Vasc Interv Radiol 2002; 13:149–153. 67. Vasconcelos C, Gailloud P, Beauchamp NJ, et al. Is percutaneous vertebroplasty without pretreatment venography safe? Evaluation of 205 consecutives procedures. AJNR Am J Neuroradiol 2002; 23:913–917. 68. Gaughen JR Jr., JensenME, Schweickert PA, et al. Relevance of antecedent venography in percutaneous vertebroplasty for the treatment of osteoporotic compression fractures. AJNR Am J Neuroradiol 2002; 23:594–600. 69. Kallmes DF, O E, Roy SS, et al. Radiation dose to the operator during vertebroplasty: prospective comparison of the use of 1-cc syringes versus an injection device. AJNR Am J Neuroradiol 2003; 24:1257–1260. 70. Ortiz AO, Natarajan V, Gregorius DR, et al. Significantly reduced radiation exposure to operators during kyphoplasty and vertebroplasty procedures: methods and techniques. AJNR Am J Neuroradiol 2006; 27:989–994. 71. Mathis JM, Wong W. Percutaneous vertebroplasty: technical considerations. J Vasc Interv Radiol 2003; 14:953–960. 72. Teng MM, Wei CJ, Wei LC, et al. Kyphosis correction and height restoration effects of percutaneous vertebroplasty. AJNR Am J Neuroradiol 2003; 24:1893–9000. 73. Hiwatashi A, Moritani T, Numaguchi Y, et al. Increase in vertebral body height after vertebroplasty. AJNR Am J Neuroradiol 2003; 24:185–189. 74. Dublin AB, Hartman J, Latchaw RE, et al. The vertebral body fracture in osteoporosis: restoration of height using percutaneous vertebroplasty. AJNR Am J Neuroradiol 2005; 26:489–492. 75. Lin EP, Ekholm S, Hiwatashi A, et al. Vertebroplasty: cement leakage into the disc increases the risk of new fracture of adjacent vertebral body. AJNR Am J Neuroradiol 2004; 25:175–180. 76. Kauffman TJ, Wald JT, Kallmes DF. A technique to circumvent subcutaneous cement tracks during percutaneous vertebroplasty. AJNR Am J Neuroradiol 2004; 25:1595–1596. 77. Anselmetti GC, Corrao G, Monica PD, et al. Pain relief following percutaneous vertebroplasty: results of a series of 283 consecutive patients treated in a single institution. Cardiovasc Intervent Radiol 2007 Jan 2; [Epub]. 78. Martin JB, Jean B, Sugiu K, et al. Vertebroplasty: clinical experience and follow-up results. Bone 1999; 25(suppl 2): 11S–15S. 79. Kaufmann TJ, Jensen ME, Schweickert PA, et al. Age of fracture and clinical outcomes of percutaneous vertebroplasty. AJNR Am J Neuroradiol 2001; 22:1860–1863. 80. Kallmes DF, Jensen ME, Marx WF, et al. A pilot study for a sham-controlled, randomized, prospective, crossover trial of percutaneous vertebroplasty. ASNR Annual Meeting, Vancouver, BC, May 13–17, 2002. 81. Ries LAG, Kosary CL, Hankey BF, et al, (eds.) SEER Cancer Statistics Review, 1973-1996. NIH Publication No. 99-2789. Bethesda, MD: National Cancer Institute, 1999. 82. Black P. Spinal metastasis: current status and recommended guidelines for management. Neurosurgery 1979; 5:726–746. 83. Gilbert HA, Kagan AR, Nussbaum H, et al. Evaluation of radiation therapy for bone metastases: pain relief and quality of life. AJR Am J Roentgenol 1977; 129:1095–1096.

409

84. Sundaresan N, Choi IS, Hughes JE, et al. Treatment of spinal metastases from kidney cancer by presurgical embolization and resection. J Neurosurg 1990; 73:548–554. 85. Gellad FE, Sadato N, Numaguchi Y, et al. Vascular metastatic lesions of the spine: preoperative embolization. Radiology 1990; 76:683–686. 86. Kato T, Nemoto R, Mori H, et al. Arterial chemoembolization with mitomycin C microcapsules in the treatment of primary or secondary carcinoma of the kidney, liver, bone and intrapelvic organs. Cancer J 1981; 48:674–680. 87. O’Reilly GV, Kleefield J, Klein LA, et al. Embolization of solitary spinal metastases from renal cell carcinoma: alternative therapy for spinal cord or nerve root compression. Surg Neurol 1989; 31:268–271. 88. Soo CS, Wallace S, Chuang VP, et al. Lumbar artery embolization in cancer patients. Radiology 1982; 145:655–659. 89. Chiras J, Cognard C, Rose M, et al. Percutaneous injection of an alcoholic embolizing emulsion as an alternative preoperative embolization for spine tumor. AJNR Am J Neuroradiol 1993; 14:1113–1117. 90. Murphy KJ, Deramond H. Percutaneous vertebroplasty in benign and malignant disease. Neuroimaging Clin North Am 2000; 10:535–545. 91. Barr JD, Barr MS, Lemley TJ, et al. Percutaneous vertebroplasty for pain relief and spinal stabilization. Spine 2000; 25:923–928. 92. Kaemmerlen P, Thiesse P, Jonas P, et al. Percutaneous injection of orthopedic cement in metastatic vertebral lesions. N Engl J Med 1989; 321(2):121 (letter). 93. Cotten A, Deramond H, Cortet B, et al. Preoperative percutaneous injection of methyl methacrylate and N-butyl cyanoacrylate in vertebral hemangiomas. AJNR Am J Neuroradiol 1996; 17:137–142. 94. Mont’Alverne F, Vallee JN, Cormier E, et al. Percutaneous vertebroplasty for metastatic involvement of the axis. AJNR Am J Neuroradiol 2005; 26:1641–1645. 95. Jensen ME, Kallmes DF. Percutaneous vertebroplasty in the treatment of malignant spine disease. Cancer J 2002; 8:194–206. 96. Weill A, Chiras J, Simon J, et al. Spinal metastases: indications for and results of percutaneous injection of acrylic cement. Radiology 1996; 199:241–247. 97. Jang JS, Lee SH. Efficacy of percutaneous vertebroplasty combined with radiotherapy in osteolytic metastatic spinal tumors. J Neurosurg Spine 2005; 2:243–248. 98. Gronemeyer DH, Schirp S, Gevargez A. Image-guided radiofrequency ablation of spinal tumors: preliminary experience with an expandable array electrode. Cancer J 2002; 8:33–39. 99. Buy X, Basile A, Bierry G, et al. Saline-infused bipolar radiofrequency ablation of high-risk spinal and paraspinal neoplasms. AJR Am J Roentgenol 2006; 186(suppl 5): S322–S326. 100. Fourney DR, Schomer DR, Nader R, et al. Percutaneous vertebroplasty and kyphoplasty for painful vertebral body fractures in cancer patients. J Neurosurg 2003; 98 (suppl 1):21–30. 101. Padovani B, Kasriel O, Brunner P, et al. Pulmonary embolism caused by acrylic cement: a rare complication of percutaneous vertebroplasty. AJNR Am J Neuroradiol 1999; 20:375–377. 102. Lin EP, Ekholm S, Hiwatashi A, et al. Vertebroplasty: cement leakage into the disc increases the risk of new fracture of adjacent vertebral body. AJNR Am J Neuroradiol 2004; 25:175–180. 103. Vasconcelos C, Gailloud P, Martin JB, et al. Transient arterial hypotension induced by polymethylmethacrylate injection during percutaneous vertebroplasty. J Vasc Interv Radiol 2001; 12:1001–1002.

410

Jensen

104. Kaufmann TJ, Jensen ME, Ford G, et al. Cardiovascular effects of polymethylmethacrylate use in percutaneous vertebroplasty. AJNR Am J Neuroradiol 2002; 23:601–604. 105. Deramond H, Wright NT, Belkoff SM. Temperature elevation caused by bone cement polymerization during vertebroplasty. Bone 1999; 25(suppl 2):17S–21S. 106. Wang GW, Wilson CS, Hubbard SL, et al. Safety of anterior cement fixation in the cervical spine: in vivo study of dog spine. South Med J 1984; 77:178–179. 107. Cloft HJ, Easton DN, Jensen ME, et al. Exposure of medical personnel to methylmethacrylate vapor during percutaneous vertebroplasty. AJNR Am J Neuroradiol 1999; 20:352–353. 108. Miller DL, Balter S, Cole PE, et al. Radiation dose in interventional radiology procedures: the RAD-IR study. Part II: Skin dose. J Vasc Intervent Radiol 2003; 14:977–990. 109. Jensen ME, Kallmes DF. Does filling the crack break more of the back?. Am J Neuroradiol 2004; 25:166–167. 110. Kim SH, Kang HS, Choi JA, et al. Risk factors of new compression fractures in adjacent vertebrae after percutaneous vertebroplasty. Acta Radiol 2004; 45:440–445. 111. Trout AT, Kallmes DF, Layton KF, et al. Vertebral endplate fractures: an indicator of the abnormal forces generated in the spin after vertebroplasty. J Bone Miner Res 2006; 21:1797–1802. 112. Garfin SR, Yuan HA, Reiley MA. New technologies in spine: kyphoplasty and vertebroplasty for the treatment of painful osteoporotic compression fractures. Spine 2001; 26:1511–1515. 113. Coumans JV, Reinhardt MK, Lieberman IH. Kyphoplasty for vertebral compression fractures: 1-year clinical outcomes from a prospective study. J Neurosurg Spine 2003; 99(1):44–50. 114. Feltes C, Fountas KN, Machinis T, et al. Immediate and early postoperative pain relief after kyphoplasty without significant restoration of vertebral body height in acute osteoporotic vertebral fractures. Neurosurg Focus 2005; 18(3):e5. 115. Kasperk C, Hillmeier J, Noldge G, et al. Treatment of painful vertebral fractures by kyphoplasty in patients with primary osteoporosis: a prospective nonrandomized controlled study. J Bone Miner Res 2005; 20(4):604–612. 116. Ledlie JT, Renfro M. Balloon kyphoplasty: one-year outcomes in vertebral body height restoration, chronic pain, and activity levels. J Neurosurg 2003; 98:36–42. 117. Gaitanis IN, Hadjipavlou AG, Katonis PG, et al. Balloon kyphoplasty for the treatment of pathological vertebral compressive fractures. Eur Spine J 2005; 14(3):250–260.

118. Berlemann U, Franz T, Orler R, et al. Kyphoplasty for treatment of osteoporotic vertebral fractures: a prospective non-randomized study. Eur Spine J 2004; 13(6):496–501. 119. Crandall D, Slaughter D, Hankins PJ, et al. Acute versus chronic vertebral compression fractures treated with kyphoplasty: early results. Spine J 2004; 4(4):418–424. 120. Rhyne A III, Banit D, Laxer E, et al. Kyphoplasty: report of eighty-two thoracolumbar osteoporotic vertebral fractures. J Orthop Trauma 2004; 18(5):294–299. 121. Phillips FM, Ho E, Campbell-Hupp M, et al. Early radiographic and clinical results of balloon kyphoplasty for the treatment of osteoporotic vertebral compression fractures. Spine 2003; 28(19):2260–2265; discussion 2265–2267. 122. Nussbaum DA, Gailloud P, Murphy K. A review of complications associated with vertebroplasty and kyphoplasty as reported to the Food and Drug Administration medical device related web site. J Vasc Interv Radiol 2004; 15:1185–1192. 123. Fribourg D, Tang C, Sra P, et al. Incidence of subsequent vertebral fracture after kyphoplasty. Spine 2004; 29:2270–2276. 124. Mathis JM, Ortiz AO, Zoarski GH. Vertebroplasty versus kyphoplasty: a comparison and contrast. AJNR Am J Neuroradiol 2004; 25:840–845. 125. Balserio J, Brower AC, Zeissman HA. Scintigraphic diagnosis of sacral fractures AJR Am J Roentgenol 1987; 148:111–113. 126. Garant M. Sacroplasty: a new treatment for sacral insufficiency fractures. J Vasc Interv Radiol 2002; 13:1265–1267. 127. Pommersheim W, Huang-Hellinger F, Baker M, et al. Sacroplasty: a treatment for sacral insufficiency fractures. AJNR Am J Neuroradiol 2003; 24:1003–1007. 128. Smith DK, Dix JA. Percutaneous sacroplasty: long-axis injection technique. AJR Am J Roentgenol 2006; 186:1252–1255. 129. Brook AL, Mirsky DM, Bello JA. Computerized tomography guided sacroplasty: a practical treatment for sacral insufficiency fracture: case report. Spine 2005; 30:E450–E454. 130. Layton KF, Thielen KR, Wald JT. Percutaneous sacroplasty using CT fluoroscopy. AJNR Am J Neuroradiol 2006; 27:356–358. 131. Butler CL, Given CA II, Michel SJ, et al. Percutaneous sacroplasty for the treatment of sacral insufficiency fractures. AJR Am J Roentgenol 2005; 184:1956–1959. 132. Atalay B, Caner H, Yilmaz C, et al. Sacral kyphoplasty for relieving pain caused by sacral hemangioma. Spinal Cord 2006; 44:196–199. 133. Uemura A, Matsusako M, Numaguchi Y, et al. Percutaneous sacroplasty for hemorrhagic metastases from hepatocellular carcinoma. AJNR Am J Neuroradiol 2005; 26:493–495.

Index Abciximab, 168 Absolute ethyl alcohol, 178 Acetabulum, 404 Acrylic compaction, 396, 397 ACTH. See adrenocorticotropin (ACTH) hormone Activated clotting time (ACT), 164 Activated coagulation time (ACT), 247 Acute ischemic infarct, 101, 103 Acute ischemic stroke (AIS), 168 Adamkiewicz artery, 41, 49 Adjunctive imaging, 390 Adrenal lesion, overproduction of cortisol, 353. See also Cushing’s syndrome Adrenals, macronodular hyperplasia of, 353 Adrenocorticotropin (ACTH) hormone dependent disease, 353 independent Cushing’s syndrome, 353 and adrenal imaging, 353 secreting pituitary adenomas, 353 difficulties in demonstration of, 354 Adult polycystic kidney disease, 126 Adverse drug reactions (ADRs), 389 African-Americans and intracranial atherosclerosis, 325 Aggressive Borden type 2 DAVF conversions, 344–345 Allen’s test, 162 Allergic reaction, 161 Alteplase thrombolysis for acute noninterventional therapy in ischemic stroke (ATLANTIS) trial, 305 Amaurosis, 215 American Society of Interventional and Therapeutic Neuroradiology (ASITN), 246 Amplatz Goose Neck Snare, 179 Amyloid angiopathy, 92, 99, 105 Anastomoses, 29, 32. See also intracranial collateral routes and anastomoses; intraparenchymal venous anastomoses in aneurysm, 81–82 arterial blood supply and, 39–40, 44–45 cerebellar, 59, 75–81 interventional neuroradiology and, 57–87 intracranial atherosclerosis disorders and, 70–72 intraspinal extradural, arterial blood supply in, 39–40 ischemic stroke and, 65–71, 82 lateral geniculate, 82 leptomeningeal, 57–61, 66, 81 pericallosal, 82 pericallosal-pericallosal, 29 persistent caroticobasilar, 29–31 transmedullary midline, 47–48 venous midline, 49 Aneurysm, 27, 92–96, 109 associated with brain AVM. See brain arteriovenous malformations (AVMs) clinical applications of MRA in intracranial, 124–127

clips, 90 coiled intracranial, surveillance, 143–144 coil protocol, 246–248 CT techniques clinical applications in, 92–96 decision matrix for treatment of, 239–241 disorders, 81–82 endovascular therapy of ruptured unsecured, 270–271 endovascular treatment of associated, 223 fusiform thrombosed, 82 intracranial. See intracranial aneurysms intracranial collateral routes and anastomoses in, 81–82 isolated spinal artery, 380–383 nonaneurysmal subarachnoid hemorrhage, 96 pericallosal, 109 radiographic follow-up of treated, 248 reconstructive repair of, 246 recurrence, following coiling, 249 ruptured cavernous, 231 saccular aneurysmal rupture, 92–96 subarachnoid hemorrhage, 92–96 AngiJet Rheolytic Thrombectomy System, 179 Angiogenesis, 60–61 Angiographic diagnosis extradural ICA dissection, 216–218 extradural VA dissection, 216–218 Angiography 3D, 325 catheter, 141 conventional, 94, 124, 131–132 CT. See CT angiography (CTA) digital subtraction angiography (DSA). See digital subtraction angiography (DSA) high-resolution, 325 iliac artery, 181 spinal, catheter for, 164 Angioplasty, 169 balloon, 266–270 carotid, 146–147 coronary, 146 percutaneous, in posterior arteries, 147 of symptomatic MCA stenosis, 327 transluminal, concept of, 325 Angioplasty and stenting case illustrations, 319–320 complications and preventive measures, 319 early trials, 313–314 evolution, 312–317 postintervention follow-up, 318 risk factors, 318 stenting procedure, 317–318 trials of angioplasty and stent placement with dep vs. cea in high-risk patients, 314–317 trials of angioplasty and stent placement with dep vs. cea in low-risk patients, 317

Angio-Seal, 181 Angular artery (AngA), 27 Anterior cerebral artery (ACA), 25–29, 31, 57–59, 66, 69, 72–73, 82, 94, 109, 138–139, 141–143, 202 dissection, 213, 215 Anterior choroidal artery (AchA), 25 Anterior communicating artery (ACoA), 57–58, 62, 74–75, 81, 90, 92–93, 138, 143–144 Anterior ethmoidal artery (AeA), 25 Anterior inferior cerebellar artery (AICA), 4, 30, 32–33, 59 Anterior internal frontal artery (AntIFA), 25 Anterior radiculomedullary artery, 41 Anterior spinal artery (ASA), 4, 32–34, 41–45, 51 Anterior temporal artery (AntTemp), 27 Anticoagulation, 164 therapy, 180 Aortic arch atheroma, 146 and branches, 1–2 Arterial access, 162 Arterial and venous occlusion tests, 183–187, 190 brain stem–evoked potentials monitoring, 186 cerebral blood flow measurements, 185 complications, 186 CT perfusion, 185 electroencephalography, 185–186 induced hypotension, 185 measurement of stump pressures, 185 MR perfusion, 185 neurophysiological monitoring, 186 perfusion imaging, 185 single-photon emission tomography, 185 somatosensory-evoked potentials monitoring, 185–186 techniques, 183–187, 190 transcranial Doppler examinations, 185–186 xenon CT, 185–186 Arterial blood supply in cauda equina, 43 in cervical region, 42–43 extra- and intraspinal extradural anastomoses, 39–40 intrinsic spinal cord arteries and, 44 radicular supply, 40–42 by region of spinal cord, 42–45 sources, 39–40 to spine and spinal cord, 39–45 superficial and intrinsic arterio-arterial anastomoses, 44–45 superficial spinal cord arteries and, 40–42 in thoracic region, 43 in thoracolumbar region, 43 Arterial collateral anatomy, 57 Arterial dissection, 103–104 in duplex US sonography, 150 Arterial feeders, classification of, 279–280 Arterial spin-labeled perfusion MRI, 89

412

Index

Arterioarterial thromboembolic stroke, medical therapy for, 326 Arteriogenesis, 60–61 Arteriovenous fistula (AVF), 92, 96–97, 106, 108 vertebral-vertebral, 380 Arteriovenous malformation (AVM), 20, 23, 51, 92, 140, 144, 146, 178, 189–190. See also specific AVMs anatomical evaluation, 52–54 associated with intracranial arterial aneurysms. See intracranial arterial aneurysms CT techniques clinical applications in, 96–99 embolization. See embolization, of AVMs endovascular management of, 205–210 facial, 162, 206 mandibular, 207 nasomaxillary, 206 nontraumatic hemorrhage, 96–99 vascular malformations, 205–208 venous occlusive disorders, 97–99 Ascending aortic arch (AOA), 1 Ascending cervical artery, 1, 4, 39–40 Ascending pharyngeal artery (APA), 7–10, 31–32 ASITN. See American Society of Interventional and Therapeutic Neuroradiology (ASITN) Aspirin, 172 therapy, 180 Associated aneurysms, endovascular treatment, 223 Asymptomatic carotid artery surgery (ACAS) trial, 148 Asymptomatic carotid atherosclerosis study (ACAS), 311 Atherosclerosis, 83, 101–103 Atherosclerosis, location and severity of in internal carotid artery (ICA) origin, 325 in middle cerebral artery (MCA), 325 in posterior cerebral artery (PCA), 325 proximal and midportion of basilar artery, 325 spared areas anterior cerebral artery (ACA), 325 posterior inferior cerebellar arteries (PICAs), 325 superior cerebellar artery (SCA), 325 Atherosclerotic CVD, 146 endovascular treatment of, 328 Atherosclerotic plaques, 331 Atretic carotid arteries, 102 Atrial fibrillation, 146 Autotriggered elliptic centric-ordered 3D gadolinium enhanced MRA (ATECO-MRA), 339, 347. See also magnetic resonance angiography (MRA) AVF. See arteriovenous fistula Aviator balloon catheter, 168 AVM. See arteriovenous malformation Balloon(s), 168 endovascular, 168 high-pressure, 168–169 hyperform, 171 low-pressure, 169–171 occlusion, 183, 185–186 occlusive devices, 171 sentry, 269 Balloon angioplasty, 266–270 first success, 325 and stent placement, comparison, 328 for vasospasm, 170

Balloon-assisted kyphoplasty. See kyphoplasty Balloon-assisted remodeling, 170 Balloon-assisted vertebroplasty, 387 Balloon-expandable coronary stents, 328 drawbacks, 329 Balloon overdilatation, in basilar artery and MCA, 327 Balloon test occlusion (BTO), 170–171 technique, 241–242, 251–252 of the ICA, 244 Balt Magic, 165 Barium sulfate, 396 Basilar artery (BA), 30, 33–34, 138–139, 141, 143 Basilar plexus, 59 Bench testing, 396 Bernasconi tentorial artery, 18, 23–24 Bioeffects and safety, of US imaging, 135–136 Bioimplants, 325 Bipediculate approach, 395 Biplane road mapping, 325 Blastic metastases, 400 Bleeding diathesis, 161 Blood-brain barrier (BBB), 153–154, 186 Blood urea nitrogen, 161 Blood velocity measurements, use of US imaging, 136 Blood vessel anatomy, clinical application of, 52–54 Blunt carotid injury (BCI), 108 Blunt carotid vertebral injury (BCVI), 108–109 Blunt injury, CT techniques clinical applications in, 107–109 Blunt neck trauma, 107 Bolus contrast CT perfusion, 89–92 Bone scintigraphy, 390, 406 Bony sclerosis, 390 Boston Scientific, 329 Bovine arch, 2 Brachiocephalic artery (BCA), 1–2 Brain applied neurovascular anatomy of, 23–37 coronal view of, 27 MRI through rostral midbrain, 24 Brain arteriovenous malformations (AVMs). See also arteriovenous malformation (AVM) aneurysms associated with classification, 281–282 clinical implications, 284–286 epidemiology, 282–283 pathogenesis of, 283–284 treatment approaches, 286–287 angioarchitecture of, 279–287 angiography of, 277–279 classification and pathogenesis, 275–276 embolization of background, 288–289 complications, 297–299 cure rates, 297 indications for, 289–290 palliative, 297 postprocedural care, 296 preradiosurgical result, 296–297 presurgical result, 296 tools and technique, 290–296 epidemiology, 276 grading systems for, 287–288 natural history of, 276–277 nidus, 280–281 Brain stem–evoked potentials monitoring, 186

Bruit, 103 Bupivacaine hydrochloride, 393 Calcarine artery (CalcA), 34 Calcification, in duplex US sonography, 151 Callosomarginal artery (Call-MargA), 25–26 Capillary hemangiomas, 195–196 Capillary malformations, 205, 208 Cardiopulmonary arrest, 99 Cardiovascular effects, 389 Carney triad, 199 Carotid and vertebral arteries dissections of, 213–227 2D TOF for imaging, 117 US imaging of, 147–151 Carotid angioplasty, 146–147 Carotid angioplasty and stenting (CAS), 146–147, 150, 163–164. See also stenting Carotid artery stenosis contrast-enhanced MRA for, 119 MRA clinical applications in, 127–129 Carotid-basilar anastomosis, 188 Carotid body tumors, 200–201 Carotid cavernous fistula (CCF) anatomy and pathophysiology, 232–233 classification, 233 clinical features, 231–232 etiology, 231 follow-up, 237 imaging studies, 234–235 indications for treatment, 233–234 symptoms and signs associated with, 233–234 treatment options and considerations, 235–237 endovascular therapy, 235–237 medical therapy, 235 surgical therapy, 235 transarterial approach, 235–237 transorbital approach, 237 transvenous approach, 237 Carotid duplex, clinical utility of, 149–150 Carotid endarterectomy (CEA), 146–148, 150. See also angioplasty and stenting Carotid ligation surgery, 241 Carotid occlusion in duplex US sonography, 150–151 testing, 183 Carotid occlusion surgery study (COSS), 75 Carotid revascularization of stenosis, 74–75 Cassinari tentorial artery, 18, 23–24 Catheter angiography, 141 Catheter-based thrombolysis, 105 Catheterization intra-arterial (IA), 163 transfemoral, 163 Catheters. See specific catheter Caucasians, proneness to extracranial disease, 325 Cauda equina, arterial blood supply in, 43 Cavernous hemangiomas, 195 Cavernous segment, 18 Cavernous sinus sampling, 359 Central artery (CenA), 27 Central retinal artery (CrA), 25 occlusion, 215 Cephalosporin, 181 Cerebellar anastomoses, 59, 78–81 Cerebral angiography, 146 Cerebral arteriovenous malformation (CAVMs), 64. See also arteriovenous malformation (AVM) intracranial collateral routes and anastomoses in, 78–81

Index Cerebral artery revascularization, by using balloon dilatation, 329 Cerebral blood flow (CBF), 89, 91–92, 100–101, 105, 132, 140, 143, 185–186 imaging, 171 Cerebral blood volume (CBV), 91, 101, 132, 185 Cerebral hypoperfusion, 179 Cerebral ischemia, 83, 146 Cerebral perfusion imaging, 104 Cerebral perfusion pressure (CPP), 105–106 Cerebral vascular disease, 146 Cerebral vascular malformations. See brain arteriovenous malformations (AVMs) Cerebral vasospasm, 264 diagnosis and monitoring, 141–143 Cerebral veins, 35–37 deep venous system, 37 superficial venous system, 35–37 Cerebral venous thrombosis (CVT), 60, 63 intracranial collateral routes and anastomoses in, 75–78 Cerebrospinal fluid (CSF) leak, 403 Cerebrovascular CTA, 101–104 arterial dissection, 103–104 atherosclerosis, 101–103 stenotic-occlusive disease, 101–103 Cerebrovascular disease (CVD), 135 atherosclerotic, 146 endovascular treatment of, 328 Cervical angiogram, 169 Cervical region arterial blood supply in, 42–43 venous drainage in, 48 Cervical segment, 18 Cervical veins, 19 Chiba needles, 406 Circle of Willis, 45, 57–58, 61, 63, 74–75, 79, 81, 88, 92, 119, 139, 143, 183, 231 Cisternography, 87–88 Clinical applications of CT techniques, 92–109 in ischemia, 99–106 in nontraumatic hemorrhage, 92–99 in trauma, 105–109 in traumatic intracranial aneurysms, 109 Clopidogrel, 172, 180 CLOTBUST study, 153–154 Clot formation, 306 Coagulopathy, 92, 96, 99, 105, 209 Cobra 1 and 2, 164 Coiled intracranial aneurysms surveillance, 143–144 Coil occlusion technique, 244 Collagen deficiency diseases, 231 Collagen-vascular disease, 126 Common carotid arteries (CCAs), 32, 151, 198, 200–204, 207, 216–217, 220 cervical carotid variations, 6 external carotid artery, 5–18 ascending pharyngeal artery, 7–10 external carotid anastomotic network, 14–17 facial artery, 10–11 internal maxillary artery, 12, 14–15 lingual artery, 10–11 occipital artery, 10–12 posterior auricular artery, 11 restoration of flow, 17 superficial temporal artery, 12–14 superior thyroid artery, 7–8 terminal branching, 7 internal carotid artery, 15–18 Common femoral artery (CFA), 162 Computed tomographic angiography (CTA), 62–64, 88–90, 92–109, 138, 140–141, 216

Computed tomographic venography (CTV), 77 Computed tomography (CT), 391, 400. See also CT techniques conventional, 87–88 fluoroscopy, 393 guidance, 401, 406 myelography, 87–88 Computed tomography perfusion (CTP), 89–92, 100–101, 185 Computer-assisted tomography. See CT techniques Contrast-enhanced CT (CECT), 88 Contrast-enhanced MRA, 119–121 Conventional angiography, 94, 124, 131–132 Coronary angioplasty, 146 Coronary artery bypass grafting (CABG), 312 Coronary catheterization, 146 Corpus callosum, 34 Cortical venous reflux (CVR) and DAVFs, 336–339, 342 Corticosteroid therapy, 196 Corticotrophin- releasing hormone (CRH), 353 Coumadin therapy, 392 Cranial DAVFs. See dural arteriovenous fistula (DAVF) Cranial nerve defects, 103 Craniocerebral angiographic anatomy, 23 Creatinine, 161 CT techniques. See also computed tomography angiography, 88–90, 92–109, 138, 140–141, 216 bolus contrast CT perfusion, 89–92 cisternography, 87–88 clinical applications, 92–109 in ischemia, 99–106 in nontraumatic hemorrhage, 92–99 stroke-like conditions, 104–106 subacute and chronic setting, 101 in trauma, 105–109 conventional CT, 87–88 development of, 87 myelography CT, 87–88 perfusion CT, 89–92 stable xenon, 92 CT venogram (CTV), 88, 98–99 Curative embolization, 290 Curing, 387 Cushing’s syndrome, 354 ACTH-dependent or ACTH-independent, 353 symptoms of, 353 Cystic medial necrosis, 214 Decision matrix for treatment, of aneurysms, 239–241 Deconstructive approach, to intracranial arterial aneurysms safety and efficacy, 241 therapeutic vessel occlusion techniques, 241–246 Deconvolution techniques, 91 Deep cervical artery, 39–40 Deep middle cerebral vein, 59 Deep venous system, in cerebral veins, 37 Dexamethasone, 177, 235 Diagnostic catheters, 164 Diagnostic US imaging carotid angioplasty, 146–147 cerebral vasospasm diagnosis and monitoring, 141–143

413

coiled intracranial aneurysms surveillance, 143–144 in interventional neuroradiology, 136–147 intracerebral venous system, 145 intracranial aneurysms endovascular treatment monitoring, 147 arterial occlusion detection, 137–140 atherosclerotic stenosis, 140–141 microemboli detection, 145–147 in neurointerventional procedures monitoring, 146–147 percutaneous angioplasty in posterior arteries, 147 reference values, 137 stenting, 146–147 TCD imaging, 136–137 vascular malformations diagnosis, 144 Diamox, 235 Diffusion-tensor imaging-based MR tractography, 183 Diffusion-weighted imaging (DWI), 99–100, 132–133 Digital subtraction angiography (DSA), 88–89, 92–94, 96–99, 101–102, 106, 108, 125–127, 130–132, 141–144, 148–149, 189, 355. See also angiography of SDAVF, 53–54 Dimethyl sulfoxide (DMSO), 178 Diphenhydramine, 161 Direct CCF. See carotid cavernous fistula (CCF) Direct current cardioversion, 146 Direct puncture, 162 embolization, 402 Dissections, 231. See also dissection of carotid and vertebral arteries distal cervical carotid, 103 intracranial, 96 MRA clinical applications in, 129 Dissection of carotid and vertebral arteries, 213–227 angiographic diagnosis, 216–218 extradural ICA dissection, 216–218 extradural VA dissection, 216–218 clinical manifestations extradural ICA dissection, 215 extradural VA dissection, 215 intradural ICA dissection, 215–216 intradural VA dissection, 216 endovascular treatment associated aneurysms, 223 extradural dissection, 220–223 intradural dissection, 224–226 epidemiology, 213 medical treatment extradural dissection, 219 intradural dissection, 220 pathogenesis, 213–215 pathophysiological process of, 221 prognosis and outcome, 226–227 surgical treatment extradural dissection, 219–220 intradural dissection, 223 treatment, 219–226 endovascular treatment, 220–226 medical treatment, 219, 223 surgical treatment, 219–220, 224 Distal internal carotid artery (dICA), 31–32 Distal middle cerebral artery, 29–30 Distal posterior cerebral artery, in vertebrobasilar system, 35 Distal vasospasm, 264 Doppler display modes, in US imaging, 136 Dorsal meningeal artery, 18, 23 Dotter, Charles T., 325

414

Index

Duplex US sonography, 147–151 arterial dissection, 150 calcification, 151 carotid occlusion, 150–151 clinical utility of carotid duplex, 149–150 degree of stenosis, 148–149 extracranial VA, 151 high bifurcation, 151 monitoring after revascularization procedures, 149–150 pitfalls, 150–151 reference values, 149 tortuosity, 151 vascular pathology sonographic assessment, 147–148 Dural arteriovenous fistula (DAVF), 77–78, 131–132. See also dural arteriovenous shunts disorders; Spinal vascular malformations adult type, 335 aggressive forms, 338–339 anatomical setting of the fistula, 335 benign and aggressive clinical features of, 337 classification schemes Borden classification, 337 Cognard classification, 337 cranial classification, 336 clinical features, 336–339 diagnostic imaging, 339–340 therapeutic considerations, 340–346 pathophysiology, 335–336 spinal classification, 346 clinical features, 346–347 diagnostic imaging, 347 therapeutic considerations, 348 transarterial approach, to venous packing of a traumatic, 343 transvenous facial approach, to bilateral cavernous, 341 Dural arteriovenous shunts disorders, intracranial collateral routes and anastomoses in, 77–78 Dural sinus drainage, in DAVF management, 342–346 Dyspnea, 398 Echo time, in TOF techniques, 119 Eclampsia, 92, 99, 105, 143 Ectopic tumors and pituitary adenomas, 354 Edema, 105, 390 Ehlers-Danlos syndrome, 103, 214, 231 EKOS ultrasound-emitting infusion catheter, 309 Electroencephalography (EEG), 185–186, 188 Embolectomy, 180 mechanical, 178–179 Embolic agents, 292 Embolic material microcoils, 175–177 Embolic protection devices, 171–172 Embolization of a cavernous sinus DAVF with liquid adhesives, 340 endovascular, 196 IM-AVM, 378–379 percutaneous transarterial, 399 PM-AVF, 374–375 preoperative meningioma, 177 preradiosurgical, 289–290 presurgical, 289 for spinal dAVF, 369–371 targeted transvenous, of a cavernous DAVF, 342

transarterial, 402 tumors, 177 Embolization, of AVMs, 167, 183. See also AVMs background, 288–289 complications, 297–299 cure rates, 297 indications for, 289–290 palliative, 297 postprocedural care, 296 preradiosurgical result, 296–297 presurgical result, 296 tools and technique, 290–296 Embosphere microspheres, 177 Embospheres, 177 EMI Mark I scanner, 87 Endovascular embolization, 196 Endovascular infusions vasodilator infusions, 167–168 Wada testing, 167 Endovascular management, of tumors and vascular malformations of head and neck, 195–211 Endovascular occlusion, of the ICA, 241 Endovascular strategies acute ischemic stroke. See intra-arterial thrombolysis for atherosclerotic disease, 328 for AVMs. See embolization, of AVMs for DAVFs, 348 intracranial atherosclerotic occlusive disease (ICD), 332 for post-subarachnoid hemorrhage vasospam balloon angioplasty, 266–270 intra-arterial antispasmodics, 264–266 of a ruptured unsecured aneurysm, 270–271 treatment outcomes, 271 in the treatment of aneurysms conditions favouring, 250–251 deconstructive approach, 241–246 reconstructive approach, 246, 248–249 Endovascular treatment, 70 associated aneurysms, 223 extradural dissection, 220–223 intradural dissection, 224–226 Endovascular US thrombolysis, 153 Enoxaparin (Lovenox), 392 Entry slice phenomenon, 113 Envoy catheter, 165 Envoy guide catheter, 166 Epilepsy, 183, 186 Epistaxis, 177, 208–210 Ethanol, 178 Ethiodol, 178, 293 Ethmoidal artery, 24 European carotid surgery trial (ECST), 148, 311 European cooperative acute stroke study (ECASS), 305 External carotid artery (ECA), 3, 5–18, 31–32, 198, 200, 202–204, 206, 209–210 anatomic considerations, 31–32 ascending pharyngeal artery, 7–10, 31–32 external carotid anastomotic network, 14–17 facial artery, 10–11 internal maxillary artery, 12, 14–15 lingual artery, 10–11 occipital artery, 10–12 posterior auricular artery, 11 restoration of flow, 17 superficial temporal artery, 12–14 superior thyroid artery, 7–8 terminal branching, 7

External jugular vein (EJV), 19 Extracranial arterial stenosis disorders, intracranial collateral routes and anastomoses in, 74–76 Extracranial tumors, 82 Extracranial vertebral arteries, 151 Extradural arteriovenous malformation/ fistulas clinical presentation, 380 imaging, 380 treatment, 380 Extradural dissection, 213 angiographic findings in, 216–217 associated aneurysms, 223 endovascular treatment, 220–223 medical treatment, 219 surgical treatment, 219–220 Extradural ICA dissection, 213, 215 Extradural VA dissection, 213, 215 Extradural venous spaces, venous drainage in, 52 Extraspinal extradural anastomoses, arterial blood supply in, 39–40 Facial artery, 31 Facial AVM, 162, 206 Facial veins, 19 False lumen (FL), 2 Falx cerebri, 26 Fat embolization syndrome, 146 Feeding arteries, 280 Fentanyl, 393 Fibromuscular dysplasia (FMD), 103, 105, 126, 214, 216, 219, 223, 231 Fisher four point grading scale, 263 FLAIR MRI vascular hyperintensity (FVH), 64, 67–68 Flow compensation, in TOF techniques, 118 Flow-guided microcatheters, 165–166, 178 Fluid-attenuated inversion recovery (FLAIR), 64, 67–69, 73, 75, 204 Fluoroscopic visualization, 401 Frontopolar artery (FpA), 25 Functional magnetic resonance imaging (fMRI), 183, 188, 190. See also magnetic resonance imaging (MRI) Fusiform thrombosed aneurysm, 82 Gadolinium-enhanced MR, in treatment of spinal DAVF, 348 Gadolinium-enhanced MRA, 119–121, 127 Gastrointestinal effects, 389 GDC system. See Guglielmi detachable coil (GDC) system General anesthesia, 178 Gerstmann’s syndrome, 27 Glomus jugulare tumors, 199–201 Glomus tumors. See paragangliomas Glomus tympanicum tumors, 199 Glycoprotein (GP) IA, 168 IIb/IIIa inhibitors, 306 Gradient echo imaging sequences, 100 Gradient-recalled echo (GRE) sequences, 64–65 Granulomatous angiitis, 105 Gray-scale imaging, in US imaging, 136 Groin closure, 181 Groin fibrosis, 163 Guglielmi detachable coils (GDC), 175, 245–246, 249 Guide catheters, 164–165 Guidewire-directed microcatheters, 166 Guidewires, 166–167 Gyral AVMs, 278–279. See also arteriovenous malformation (AVM)

Index Headache, 103 Head veins, 18–20 HEAL trial, 250 Hemangiomas, 195–196 Hemangioma thrombocytopenia syndrome, 196 Hematuria, 207 Hemorrhagic and stroke-like conditions, T techniques clinical applications, 104–106 Hemorrhagic complications, 180 Heparin, 162, 164 Heparin-coated stents, 173 Hereditary hemorrhagic telangiectasia (HHT), 363 diagnostic criteria for, 209–210 Herniation, 105 Herpes virus, 105 Heubner artery, 25 High-resolution angiography, 325 Homocysteinuria, 103 Horner’s syndrome, 103 ipsilateral, 215 Humanitarian device exemption (HDE), stent as, 328 Hydrogel coils, 175 Hyperacute and acute setting CT techniques clinical applications, 99–101 in ischemia, 99–101 Hypercortisolemia. See Cushing’s syndrome Hyperform balloon, 171 Hyperperfusion syndrome, 105 Hypertension, 92 Hypertensive encephalopathy, 105 Hypertensive hemorrhage, 99, 105 Hypertensive therapy, 179–180 Hypervascular tumors, 401 Hypervolemic therapy, 180 Hypoglossal artery, 31 Hypoxic-ischemic injury, 99 Ibuprofen, 398 Iliac artery angiography, 181 Induced hypotension test, 185 Inferior hypophyseal artery, 18, 23 Inferior petrosal sinus (IPS), 59, 232, 237 Inferior petrosal sinus sampling (IPSS), 353 alternatives to, 359–360 coaxial technique using microcatheter to catheterize, 354 complications of IPSS, 359 contralateral injection for, 357 for differential diagnosis of Cushing’s disease, 353–354 indications for, 354 interpretation of sampling data, 358–359 ‘‘lavender-top’’ tubes in, 354 microcatheter technique of, 356 negative result from, 358 role in lateralizing pituitary adenoma, 358–359 successful catheterization of IPS, 355 use of sedation in, 354 Inferolateral trunk, 23 Innominate artery, 1 Insufflator device, 169 Insula, arterial anatomy of, 28 Intelligel, 175 Intercostopedicular approach, 401 Interferon therapy, 196 Interhemispheric fissure, 96 Intermediate lenticulostriate artery (IL), 28 Internal carotid artery (ICA), 5, 60, 62, 70, 72, 74–75, 92–95, 97, 102–103, 105, 108–109, 138–140, 144–145, 148,

150–151, 198, 200–203, 209–210, 214–221, 223, 226, 231–238 anatomic considerations, 27–30 anterior cerebral artery, 27–29 distal middle cerebral artery, 29–30 angiogram of, 30 anterior cerebral artery, 25–29, 31 anterior choroidal artery, 25 bifurcation, 28 diagnostic parameters for, 149 lateral angiographic projection of right, 24 middle cerebral artery, 26–30 ophthalmic artery, 23–25 posterior communicating artery, 24–25, 28 stenosis, 146, 149 Internal iliac arteries, 39–40 Internal jugular vein sampling, 360. See also inferior petrosal sinus sampling (IPSS) Internal mammary (INM) artery, 1 Internal maxillary artery (IMA), 6–7, 31–32 Internal parietal artery (IpA), 34 International Study of Unruptured Intracranial Aneurysms (ISUIA), 239–240 International subarachnoid aneurysm trial (ISAT), 246 Interventional neuroradiology CT imaging and physiologic techniques in, 87–109 intracranial collateral routes and anastomoses in, 57–87 MR angiography applications in, 113–133 ultrasonographic imaging and physiological techniques in, 135–154 Interwoven microfilaments, 175 Intra-arterial antispasmodics, 264–266 Intra-arterial thrombolysis, 168, 221 combined intravenous and intra-arterial approach, 306 combined with glycoprotein IIb/IIIa inhibitors, 306 multimodal, 306–309 as sole treatment, 305–306 and thrombolysis in myocardial infarction (TIMI) grading scale, 309 Intracerebral venous system, 145 Intracranial aneurysms endovascular treatment monitoring, 147 MRA clinical applications in, 124–127 Intracranial angioplasty. See stent techniques Intracranial arterial aneurysms AVMs associated with classification, 281–282 clinical implications, 284–286 epidemiology, 282–283 pathogenesis of, 283–284 treatment approaches, 286–287 clinical presentation, 239 definitions and epidemiology, 239 incidence rate, 239 treatment, 241, 249–251 aneurysm coil protocol, 246–248 balloon test occlusion (BTO), 251–252 decision matrix, 239–241 deconstructive approach, 241–246 follow-up, 248 improved techniques, 252–253 reconstructive approach, 246, 248–249 stent techniques, 253–256 surgery following partial embolization, 252–253 thromboembolic complications, 254 using neuroform devices, 254

415

Intracranial arterial occlusion detection, 137–140 Intracranial arteries, morphology and hemodynamics of, 328 Intracranial arteriovenous malformations (AVMs), 129. See also arteriovenous malformation (AVM) Intracranial atherosclerosis disorders, intracranial collateral routes and anastomoses in, 70–72 Intracranial atherosclerotic occlusive disease (ICD) aspects of endovascular revascularization for, 325 demographic characteristics, 325 endoluminal revascularization of, 326 endovascular treatment of (by angioplasty and stenting), 332 natural history of, 326 stenting and angioplasty for, 325 treatment indications, 326 warfarin-aspirin recurrent stroke study (WARSS) trial, 326 warfarin versus aspirin study, 326 Intracranial atherosclerotic stenosis, 140–141 Intracranial collateral routes and anastomoses anatomy of, 57–59 cerebral hemodynamics and, 62 clinical correlates, 62 disorders, 65–83 aneurysms, 81–82 cerebral arteriovenous malformations, 78–81 cerebral venous thrombosis, 75–77 dural arteriovenous shunts, 77–78 extracranial arterial stenosis, 74–76 intracranial atherosclerosis, 70–72 ischemic stroke, 65–71 moyamoya syndrome, 72–74 occlusion, 74–76 tumors, 82–83 epidemiology of, 60 imaging, 62, 65 in interventional neuroradiology, 57–87 pathophysiology of, 60–62 angiogenesis, 60–61 arteriogenesis, 60–61 Intracranial dissection, 96 Intracranial embolization coils, 90 Intracranial neoplasms, 92, 99, 105 Intracranial pressure monitoring, 162 Intracranial stenotic-occlusive disease, 102 Intracranial tumors, 82 Intracranial vascular malformations, MRA clinical applications in, 129–131 Intradural dissection, 213 angiographic findings in, 217, 218 clinical manifestations, 216 endovascular treatment, 224–226 medical treatment, 223 surgical treatment, 224 Intradural ICA dissection, 213, 215–216 Intradural VA dissection, 213 angiographic features in, 219 angiographic findings in, 217, 218 clinical manifestations, 216 Intramedullary arteriovenous malformation (IM-AVMs). See also arteriovenous malformation (AVM) clinical manifestations, 377 imaging, 377–378 metameric angiomatosis, 378 treatment, 378–379 Intraparenchymal venous anastomoses, venous drainage in, 47–48

416

Index

Intraspinal extradural anastomoses, arterial blood supply in, 39–40 Intravascular deoxygenation, 61 Intraventricular hemorrhage, 96 Intrinsic arterio-arterial anastomoses, arterial blood supply in, 44–45 Intrinsic spinal cord arteries, 44, 46–47 Intrinsic veins, venous drainage in, 46–47 Introducer sheaths, standard, 163–164 IPSS. See inferior petrosal sinus sampling (IPSS) IPS venogram, patient with Cushing’s disease, 355 ISAT. See international subarachnoid aneurysm trial (ISAT) Ischemia, 132 cerebrovascular CTA, 101–104 CT techniques clinical applications in, 99–106 hemorrhagic and stroke-like conditions, 104–106 hyperacute and acute setting, 99–101 subacute and chronic setting, 101 Ischemic complications, 179–180 Ischemic event, risk of, 104 Ischemic optic neuropathy, 215 Ischemic stroke intracranial collateral routes and anastomoses in, 65–71, 82 MRA clinical applications in, 132–133 Isobutyl-2-cyanoacrylate (IBCA), 289 Isolated spinal artery aneurysms, 380–383 ISUIA. See International Study of Unruptured Intracranial Aneurysms (ISUIA) Joint Study of Extracranial Arterial Occlusion, 325 Jugular venous sampling, 360 Jugulotympanic paragangliomas, 199–200 Juvenile angiofibroma (JAF), 196–199, 209 Kaposiform hemangioendothelioma, 196 Kasabach-Merritt syndrome, 196 ’’Kissing balloon’’ technique, 331 Klippel-Trenaunay syndrome, 275 Kommerell’s diverticulum, 1 Ku¨mmell’s disease, 390 Kyphoplasty, 387, 402, 404 advantages over vertebroplasty, 405 controversies with vertebroplasty, 406 Kyphosis, 389 Lacrimal artery (LA), 24–25 LAO arch configuration, 2 Laser thrombectomy, 179 Lateral geniculate anastomoses, 82 Lateral lenticulostriate artery (LL), 28 Lateral posterior choroidal artery (LpcA), 34 Left common carotid artery (LCCA), 1 Left internal carotid artery (LICA), 222, 224 Left subclavian artery (LSUB), 1 Lenticulostriate arteries (LentStrA), 25–27 Lenticulostriate collaterals, 70, 72, 81 Leptomeningeal anastomoses, 57–61, 66, 81 Leptomeningeal collaterals, 62–64, 70, 82 Lidocaine, 162 cardiac, 162, 167 Lingual artery, 10–11, 31 Liquid coils, 176 Luer lock syringes, 396–397 Lumbar arteries, 4 Lumbar region, venous drainage in, 50 Luminal stenosis, assessment with MRA, 127

Lymphatic malformations, 205, 208 Lymphoma, 401 Magnetic resonance and recanalization of stroke clots using embolectomy (MR RESCUE) trial, 308 Magnetic resonance angiography (MRA), 39, 57, 62–64, 67–69, 71, 88, 97, 99, 101–102, 216, 248, 365 application in interventional neuroradiology, 113–133 autotriggered elliptic centric-ordered 3D gadolinium enhanced (ATECO-MRA), 339, 347 clinical applications, 124–133 carotid stenosis, 127–129 dissections, 129 intracranial aneurysms, 124–127 intracranial vascular malformations, 129–131 ischemic stroke, 132–133 spinal vascular malformations, 131–132 contrast-enhanced MRA, 119–121 gadolinium-enhanced, 119–121, 127 multiplanar reconstructions (MPRs) in, 126 phase contrast MRA, 121–124 TOF techniques, 113–119 Magnetic resonance digital subtraction angiography (MR DSA), 97, 130 Magnetic resonance imaging (MRI), 39, 97, 99–101, 216, 354, 365. See also functional magnetic resonance imaging (fMRI) arterial spin-labeled perfusion, 89 pituitary. See pituitary MRI through rostral midbrain, 24 Magnetic resonance venography (MRV), 64, 76–77, 80 Magnetization transfer imaging (MTI), 118 Magnetoencephalography (MEG), 183 Marfan’s syndrome, 103, 214 Massive oral bleeding, 2 MATRIX-ACTIVE trial, 250 Maximum intensity projection (MIP), 90, 93, 97, 99, 105 MCA. See middle cerebral artery McConnel’s capsular artery, 18, 23 Mechanical embolus removal in cerebral ischemia (MERCI) trial, 308 Medial posterior choroidal artery (MpcA), 34 Medical treatment extradural dissection, 219 intradural dissection, 220 MEG test, 188–189 Meningiomas, 82, 202–204 Meningohypophyseal trunk, 23 Merci Retrieval System, 308 Merci Retriever, 163, 178–179 Metameric angiomatosis, 378 Methylmethacrylate, 404 Microbubble-augmented US thrombolysis, 154 Microcatheters, 179 flow-guided, 165–166 guidewire-directed, 166 shapes, 166 Microembolic signals (MES), 145–146 Microemboli detection, by diagnostic US imaging, 145–147 Microguidewires, 167, 179 MicroLysUS infusion catheter, 179 Microvena snare, 308–309 Midazolam, 393

Middle cerebral artery (MCA), 25–30, 58–59, 62–64, 66–70, 72–73, 79, 82, 92, 109, 138–143, 151, 153–154, 222 aneurysms, 93–94 balloon overdilation, 327 dissection, 213, 215 stenosis, 70 Middle internal frontal artery (MidIFA), 25 Middle lenticulostriate artery (ML), 28 Middle meningeal artery (MMA), 31, 200, 202–204 Middle temporal artery (MidTemp), 27 Milrinone, 266 Mirage wire, 167 Mixed (sulcogyral) type AVMs, 279. See also arteriovenous malformation (AVM) MOTSA technique. See multiple overlapping thin slab acquisition (MOTSA) technique Moyamoya syndrome, 96, 105 intracranial collateral routes and anastomoses in, 72–74, 83 transient ischemic attacks (TIAs) of, 62 MRA. See magnetic resonance angiography MR DSA methods. See magnetic resonance digital subtraction angiography (MR DSA) MRI, 406 spinal, 400 MR perfusion, 185 Multidetector CT (MDCT), 87–88, 101, 107, 109 Multiholed pigtail catheters, 164 Multimodal thrombolysis, 306–309 Multiplanar reconstructions (MPRs), in MRA, 126 Multiple overlapping thin slab acquisition (MOTSA) technique, 117 Murray’s law, 61 Myelography CT, 87–88 Myeloma, 401 Myocardial infarction, 146 N-acetylcysteine, 161 Narcotic analgesia, 389 Nasal telangiectasias, 210 NASCET trial, 127 Nasomaxillary AVMs, 206 National institute of neurological diseases and stroke (NINDS) study, 305 National Institutes of Health Stroke Scale (NIHSS), 306 N-butyl cyanoacrylate (NBCA), 235–236, 292, 400 in the nidus, 293 polymerization of, 292–293 technique, 294–295 N-butyl 2-cyanoacrylate (NBCA), 370–371 N-butyl cyanoacrylic acid (NBCA), 177–178 Neck pain, 103 veins, 18–20 Neo-endothelialization, 172 Neointimal hyperplasia, 173 Neurofibromatosis type 1, 199 Neuroform, 254 Neuroform Microdelivery System, 173 Neuroform stent, 173–175 Neuronet Endovascular Snare, 179 Neurophysiological monitoring (NPM) test, 186 Nicardipine, 168, 266 Nitinol (nickel-titanium alloy), 172, 179 cerebrovascular stents, 329 wire, 178 Nitroglycerin, 162

Index Nonaneurysmal subarachnoid hemorrhage, 96 Nonbifurcated common carotid artery, 5 Nondeconvolution methods, 91 Nonenhanced CT (NECT), 88, 90–95, 97–101, 104, 106 Nonsteroidal anti-inflammatory agents, 398 Nontraumatic hemorrhage AVM, 96–99 causes of, 99 CT techniques, clinical applications in, 92–99 venous occlusive disorders, 97–99 North-American symptomatic carotid endarterectomy trial (NASCET), 148, 311 Occipital artery (Occ), 31 Occipital auricular arteries, 7 Occlusion, 92, 105, 138, 150 disorders, 74–76 Occlusion test, measures, 242 Occlusive vascular disease distribution, 325 Occupational Safety and Health Administration (OSHA), 404 Ocular ischemic syndromes, 215 Onyx, 178, 253, 292, 294 technique, 295–296 Opacification, 396 Ophthalmic artery, 23–25 Orbitofrontal artery (Ofr), 27 Osler-Weber-Rendu disease, 363 Osteogenesis imperfecta, 214, 231 Osteolysis, 400, 402 Osteonecrosis, 390 Osteophyte formation, 390 Osteoporotic fractures, 403 Pain, 399 Palliative embolization, 290 Papaverine, 168, 264–266 Paragangliomas, 199–202 Paravertebral approach, 403 Parenchyma, 96 Parent artery occlusion (PAO), 28–29 Parietooccipital artery (PoA), 34 Parks-Weber syndrome, 275 Patient screening and evaluation history of present illness, 390 neurological and physical examination, 390 preprocedure preparation and counseling, 392 radiological evaluation, 390–392 Patient selection criteria, 389–390 Penetrating injury, CT techniques clinical applications in, 107–108 Percutaneous angioplasty in posterior arteries, 147 Percutaneous transarterial embolization, 399 Percutaneous transluminal angioplasty (PTA), 147, 325–327 balloon advantage of, 328 disadvantage of, 329 for femoral artery stenosis, 326 for intracranial atherosclerotic stenosis, 327 problems of intracranial balloon angioplasty, 326 success rate for PTA (92%), 326 Perfused blood volume (PBV), 92 Perfusion CT, 82 Perfusion harmonic imaging (PHI), 140 Perfusion imaging, 104, 185

Perfusion-weighted imaging (PWI), 100–101, 132 Pericallosal anastomoses, 82 Pericallosal aneurysms, 109 Pericallosal artery (PeriA), 25 Pericallosal-pericallosal anastomosis, 29 Perimedullary arteriovenous fistulas (PM-AVFs) classification, 371 clinical presentation, 371–373 imaging, 373 treatment, 374–375 Perinatal asphyxic injury, 99 Perivertebral acrylic, 403 Persistent caroticobasilar anastomoses, 29–31 Petrous segment, 18 Pharmacological testing, 189–190 noninvasive alternatives, 190 technique, 190 Phase contrast MRA, 121–124 Pituitary adenomas and inferior petrosal sinus sampling (IPSS), 353 Pituitary MRI. See also magnetic resonance imaging (MRI) absence of uniformity in the imaging protocols, 354 for evaluating ACTH-dependent Cushing’s syndrome, 354 Pneumothorax, 403 Polycystic kidney disease, 239 Polyglycolic acid– lactide copolymer (PGLA)-coated coils, 250 Polymethylmethacrylate (PMMA), 387, 396, 397, 398, 399, 400, 401, 403 acetabulum with, 404 extravasation, 406 Polyurethane filter, 171 Polyvinyl alcohol particles, 177, 292 Positron emission tomography (PET), 68, 73, 75–76, 189 Possis AngioJet thrombectomy catheter, 309 Posterior auricular arteries, 7 Posterior cerebral artery (PCA), 138–139, 141, 143, 188 in vertebrobasilar system, 27, 32–35 Posterior communicating artery (PCoA), 24–25, 28, 57–58, 73–75, 81, 92, 138 Posterior ethmoidal artery (PeA), 25 Posterior inferior cerebellar artery (PICA), 4, 31–33, 43, 59, 69, 92–94, 127, 226 Posterior intercostal artery, 4, 39–40 Posterior internal frontal artery (PostIF), 25 Posterior reversible encephalopathy syndromes (PRES), 105 Posterior temporal artery, 27, 33 Posterior venous plexus, 50 Post-revascularization injury, 180 Post-subarachnoid hemorrhage vasospasm. See also vasopasm diagnosis and medical management, 264 endovascular therapy balloon angioplasty, 266–270 intra-arterial antispasmodics, 264–266 of a ruptured unsecured aneurysm, 270–271 treatment outcomes, 271 Fisher four point grading scale, 263 historical aspects, 264 pathophysiology, 263–264 Postvertebroplasty care, 398 Precentral artery (PcA), 25 Prednisone, 161 Preoperative meningioma embolization, 177 Preprocedure plain films, 400 Preradiosurgical embolization, 289–290

417

Presurgical embolization, 289 Primary pigmented nodular adrenal disease (PPNAD), 353 PROACT trial, 306, 308 Prophylaxis, 396 Prosthetic cardiac valves, 146 Protamine sulfate, 164 Proximal posterior cerebral artery (PCA), 57–59, 66, 68–69, 73, 82 Pseudoaneurysm, 109. See also aneurysm Pseudo-Cushing’s syndrome, 353. See also Cushing’s syndrome Pseudoxanthoma elasticum, 214 Pterygoid canal artery, 18 Pterygoid plexus, 59 Pterygopalatine venous plexus (PVP), 19–20 Pulmonary arteriovenous, 210 Pulmonary embolism, 389 Pulsatile tinnitus, 103 Pulse oximetry, 162 Radial artery, 162 Radiation therapy, 196, 399, 402 Radicular pain, 390 Radicular supply arteries, and arterial blood supply, 40–42 Radiculomedullary artery, 39–40 Radiculomedullary veins, venous drainage in, 50–51 Radiculopathy, 390 Radiographic follow-up, of treated aneurysms, 248 Radionucleotide, 171 Radiopacity, 178 Radiosurgical treatment, of cranial DAVFs, 346 Rated burst pressure (RBP), 168 Raymond’s syndrome, 359 Recombinant tissue plasminogen activator (r-tPA), 305 Reconstructive repair, of aneurysms, 246 with aneurysm coils, 248–249 Rendu-Osler-Weber syndrome, 209, 275 Reperfusion injury, 99 Reperfusion syndrome, 105 Restenosis, 173 after stenting, 330 Restorative neurology, 61 Retroperitoneal hematoma, 162 Revascularization procedures monitoring, 149–150 Right aortic arch, 2 Right common carotid artery (RCCA), 1 Right subclavian artery (RSUB), 1 Right vertebral artery, 2 Ringer’s solution, 176 Robust leptomeningeal collaterals, 67 Ruptured cavernous aneurysms, 231 RX Acculink Carotid Stent System, 168 Saccular aneurysmal rupture, 92–96 Sacral arteries, 39–40 Sacral foramina, 406 Sacroplasty, 406 Sacrum, 406 SAH. See subarachnoid hemorrhage (SAH) Sclerotherapy, 162 Seldinger technique, 163 Selective angiographic evaluation, of brain AVMs, 278 Selective arterial spin-labeled (SASL) perfusion, 65 Self-expanding stent implantation, 329 SENSE technique, 119, 130 Sentry balloon, 269

418

Index

Shaded surface display (SSD) methods, 89 Short-tau inversion recovery (STIR), 390 Shuttle Select Tuohy-Borst Introducer, 164 Sickle cell disease, 96, 105 Simmons II catheter, 164 Single-photon emission computerized tomography (SPECT), 68, 185, 263 Sinus access, 168 Sinus anatomy, 36 Sirolimus (rapamycin), 330 Skull applied neurovascular anatomy of, 23–37 base veins, 18–20 SMASH technique, 119 Snow plow’ effect, of atherosclerotic plaques, 331 Sodium bicarbonate, 161 Somatosensory-evoked potentials (SSEP) monitoring, 185–186 Specialty needles, 395 SPECT scanning, 390 Spetzler-Martin grading system, for AVMs, 288. See also arteriovenous malformation (AVM) Spinal angiography, catheter for, 164 Spinal DAVFs. See dural arteriovenous fistula (DAVF) Spinal dural arteriovenous fistulas (SDAVFs), 52–54 Spinal radicular arteries, 40 Spinal vascular malformations classification, 363–364 clinical presentation, 364–365 dAVFs cervical, 367–369 clinical manifestations, 367 imaging, 367 pathophysiology, 366 treatments, 369–371 extradural arteriovenous malformation/ fistulas clinical presentation, 380 imaging, 380 treatment, 380 imaging, 365 intramedullary arteriovenous malformation (IM-AVFs) clinical manifestations, 377 imaging, 377–378 metameric angiomatosis, 378 treatment, 378–379 isolated spinal artery aneurysms, 380–383 perimedullary arteriovenous fistulas (PM-AVFs) classification, 371 clinical presentation, 371–373 imaging, 373 treatment, 374–375 Spinal vascular malformations, MRA clinical applications in, 131–132 Spine and spinal cord arterial blood supply, 39–45 in cauda equina, 43 in cervical region, 42–43 extra- and intraspinal extradural anastomoses, 39–40 intrinsic spinal cord arteries and, 44 radicular supply, 40–42 by region, 42–45 sources, 39–40 superficial and intrinsic arterio-arterial anastomoses, 44–45 superficial spinal cord arteries and, 40–42 in thoracic region, 43 in thoracolumbar region, 43 AVMs anatomical evaluation, 52–54

vascular anatomy of, 39–54 venous drainage, 45–52 in cervical region, 48 in extradural venous spaces, 52 intraparenchymal venous anastomoses, 47–48 intrinsic veins, 46 in lumbar region, 50 in radiculomedullary veins, 50–51 superficial veins, 46–47 in thoracic region, 48–50 in transdural course, 50–51 Spinnaker Elite, 165 Splenial artery, 34 Stable xenon CT perfusion, 89, 92, 101, 105, 185–186 Stenosis, degree of, 148–149 Stenotic-occlusive disease, 101–103 Stent(s), 172–175 balloon-expandable coronary, 328 coated, 173 drugeluting, 173 evaluation, 89 grafts, 175 as humanitarian device exemption (HDE), 328 intracranial, 172 Neuroform, 173–175 notion of, 327 placement, and success rates, 328 self-expanding, 172 sirolimus-eluting, 330 smaller, 172–173 Wingspan, 175 Stenting, 146–147. See also carotid angioplasty and stenting (CAS) and angioplasty for intracranial atherosclerotic occlusive disease (ICD), 325 endovascular treatment of, 332 restenosis after, 330 Stenting and angioplasty case illustrations, 319–320 complications and preventive measures, 319 early trials, 313–314 evolution, 312–317 postintervention follow-up, 318 risk factors, 318 stenting procedure, 317–318 trials of angioplasty and stent placement with dep vs. cea in high-risk patients, 314–317 trials of angioplasty and stent placement with dep vs. cea in low-risk patients, 317 Stenting of symptomatic atherosclerotic lesions in the vertebral or intracranial arteries (SSYLVIA), 328 Stent techniques, 306–308 advanced, 254 assisted angioplasty, 325 advantages over simple angioplasty, 328 as replacement to angioplasty, 327 balloon-expandable coronary stents, 328 drawbacks, 329 complications, 254–256 effects of heparin-coated, 330 exclusion criteria for, 330 humanitarian device exemption (HDE), 328 inclusion criteria for, 330 for intracranial atherosclerotic occlusive disease (ICD), 325 nitinol (nickel-titanium alloy) cerebrovascular stents, 329

procedure and possible complications, 330–331 rationale for, 253–254 restenosis, after stenting, 330 sirolimus-eluting stents, 330 stent, notion of, 327 stent placement, and success rates, 328 technology, 325 Sternal fractures, 403 Steroid therapy, 402 Stroke-like conditions, CT techniques clinical applications, 104–106 Stroke outcomes and neuroimaging of intracranial atherosclerosis (SONIA) study, 140 Stump pressures, measurement of, 185 Sturge-Weber disease, 275 Subacute and chronic setting CT techniques clinical applications, 101 in ischemia, 101 Subarachnoid hemorrhage (SAH), 92–96, 98, 104, 124–125, 141–143, 214–216, 218, 224–226, 239–240 Subclavian steal syndrome, 151 Subdural hematoma (SDH), 94 Sulcal AVMs, 278. See also arteriovenous malformation (AVM) Superficial arterial anastomoses, arterial blood supply in, 44–45 Superficial spinal cord artery, 39–40 arterial blood supply in, 40–42 Superficial spinal cord veins, 48, 52 Superficial temporal artery (STA), 31 Superficial veins, venous drainage in, 46–47 Superficial venous system, in cerebral veins, 35–37 Superior cerebellar artery (SCA), 32–34, 59, 69 Superior ophthalmic vein (SOV), 19 Superior petrosal sinus, 59 Superior thyroid artery (SUT), 7–8 Superior thyroid (ST), 31 Superselective angiographic evaluation, of brain AVMs, 278 Super Stiff wire, 167 Supraclinoid segment, 18 Surgical treatment extradural dissection, 219–220 intradural dissection, 223 Swan-Ganz catheter, 247 Sylvian fissures, 96 Symptomatic vertebroplasty, complications in, 402–404 Syphilis, 105 Takayasu arteritis, 105 Tantalum powder, 178 Targeted transvenous embolization, of a cavernous DAVF, 342 TCD ultrasonography, 67 Tegaderm, 393 Temporal artery (Tp), 27 Temporooccipital artery (OccTemp), 27 TempoVert, 164 Terumo Glidecath Angled Taper, 164 Terumo glidewire, 163 Therapeutic vessel occlusion techniques, 241–246 Thoracic region arterial blood supply in, 43 venous drainage in, 48–50 Thoracolumbar region, arterial blood supply in, 43 3D angiography, 325 3D gadolinium-enhanced sequence, 117 Three-dimensional coil shapes, 252

Index 3D road map, 325 3D SPGR technique, 119 3D TOF techniques, 117–118, 125–127, 130–131 Thrombolysis, 138, 179 catheter-based, 105 endovascular ultrasound, 153 intra-arterial. See intra-arterial thrombolysis mechanism of US-accelerated, 152 microbubble-augmented US, 154 multimodal, 306–309 transcutaneous US-enhanced, 153–154 transvenous, 168 Thrombolysis in myocardial infarction (TIMI) grading scale, 309 Thrombolytic infusions, 168 Thrombosis dural venous sinus, 168, 179 prevention, 162, 164, 173, 175, 178, 180, 181 transverse sinus, 99 Tilted optimized nonsaturating excitation (TONE), 118 Time-of-flight magnetic resonance angiograph (TOF MRA), 73–74 Time-of-flight (TOF) techniques, 113–122, 125–127, 130–131 2D TOF, 115–117, 119, 121, 127 3D TOF, 117–118, 125–127, 130–131 echo time, 119 flow compensation, 118 limitations, 118 Titanium-nitrous-oxide (TiNOX), 173 Tobramycin, 396 Tortuosity, 3–4, 151 Tortuous vasculature, 118 Tractography, 190 Transarterial approach, to venous packing of a traumatic DAVF, 343 Transarterial embolization, 402 Transbrachial approach, 162 Transcranial color-coded duplex sonography (TCCS), 136–146 Transcranial Doppler (TCD), 64, 71, 264 imaging, 96, 135–147, 152–154, 185–186 Transcranial low-frequency US mediated thrombolysis in brain ischemia (TRUMBI) trial, 153 Transcutaneous US-enhanced thrombolysis, 153–154 Transdural course, venous drainage in, 50–51 Transfemoral approach, 162–163 complications in, 162 Transient ischemic attacks (TIAs), 62, 212 Transluminal angioplasty, concept of, 325 Transmedullary midline anastomoses, 47–48 Transradial approach, 162 Transvenous disconnection, of a Borden type 3, 343 Transvenous facial approach, to bilateral cavernous DAVFs, 341 Transverse sinus thrombosis, 99 Trauma blunt injury, 108–109 CT techniques clinical applications in, 105–109 penetrating injury, 107–108 Traumatic intracranial aneurysms, CT techniques clinical applications in, 109 Triple-H therapy, 270 Trocars, 395, 406 15-G, 401 vertebroplasty, 401

Tufted angioma, 196 Tumors embolization, 177 endovascular management of, 195–211 intracranial collateral routes and anastomoses in, 82–83 Two 2D TOF techniques, 115–117, 119, 121, 127 Ulcerations, 102 Ultrasonographic imaging. See US imaging University of Oregon in Portland, 325 Unruptured intracranial aneurysm. See intracranial arterial aneurysms US imaging accelerated thrombolysis mechanism, 152 bioeffects and safety, 135–136 contrast agents effects, 136 nonthermal effects, 135–136 thermal effects, 135 blood velocity measurements, 136 of carotid and vertebral arteries, 147–151 diagnostic imaging in interventional neuroradiology, 136–147 carotid angioplasty, 146–147 cerebral vasospasm diagnosis and monitoring, 141–143 coiled intracranial aneurysms surveillance, 143–144 intracerebral venous system, 145 intracranial aneurysms endovascular treatment monitoring, 147 intracranial arterial occlusion detection, 137–140 intracranial atherosclerotic stenosis, 140–141 microemboli detection, 145–147 neurointerventional procedures monitoring, 146–147 percutaneous angioplasty in posterior arteries, 147 reference values, 137 stenting, 146–147 TCD imaging, 136–137 vascular malformations diagnosis, 144 diagnostic techniques, 136 Doppler display modes, 136 duplex sonography, 147–151 arterial dissection, 150 calcification, 151 carotid occlusion, 150–151 clinical utility of carotid duplex, 149–150 degree of stenosis, 148–149 extracranial VA, 151 high bifurcation, 151 monitoring after revascularization procedures, 149–150 pitfalls, 150–151 reference values, 149 tortuosity, 151 vascular pathology sonographic assessment, 147–148 endovascular US thrombolysis, 153 gray-scale imaging, 136 in interventional neuroradiology, 135–154 microbubble-augmented US thrombolysis, 154 technical aspects, 135–147 therapeutic use in acute stroke, 151–154 transcutaneous US-enhanced thrombolysis, 153–154 Valsalva maneuver, 354 Vancomycin, 396

419

Varicose convolutions, 50 Vascular access, 162 Vascular access sites common femoral artery, 162 direct puncture, 162 radial artery, 162 Vascular access technique micro access approach, 163 standard approach, 162–163 Vascular malformations, 205–211 AVM, 205–208 diagnosis, 144 endovascular management of, 195–211 epistaxis, 208–210 Vascular pathology sonographic assessment, 147–148 Vascular smooth muscle cells, 263 Vascular tumors, 195–204 capillary hemangioma, 195–196 juvenile angiofibroma (JAF), 196–199 meningiomas, 202–204 paragangliomas, 199–202 Vasculitis, 92, 99, 105 Vasculopathies, 96, 105 Vasoconstriction, 263–264 Vasodilation, 263–264 Vasospasm (VSP), 97, 104 balloon angioplasty for, 170 cerebral, diagnosis and monitoring, 141–143 distal, 264 post-subarachnoid hemorrhage. See post-subarachnoid hemorrhage vasospasm Vein of Labbe, 59, 76–77 Vein of Trolard, 59 Venogenesis, 60 Venography, 396 Venous collateral anatomy, 59 Venous drainage, 59 in cervical region, 48 in extradural venous spaces, 52 intraparenchymal venous anastomoses, 47–48 intrinsic veins, 46 in lumbar region, 50 in radiculomedullary veins, 50–51 in spine and spinal cord, 45–52 superficial veins, 46–47 in thoracic region, 48–50 in transdural course, 50–51 Venous hypertension, 92, 105 Venous midline anastomoses, 49 Venous occlusion testing. See arterial and venous occlusion tests Venous occlusive disorders, 97–99 Verapamil, 162, 167–168, 266 infusion, 180 Vertebral arteries (VA), 32, 39–40, 43, 138–139, 141, 143, 150–151, 200 cervical branches, 3–5 dissections, 213, 216–218, 223–225 distal variations, 5 extracranial, 151 scope on head turning, 4 tortuosity, 3–4 Vertebral compression fractures (VCF), 388 associated with neoplastic diseases, treatment of, 399 clinical outcomes, 402 imaging evaluation, 400–401 patient selection, 400 vertebroplasty and adjunctive therapies, 402 vertebroplasty for malignant disease, 401–402 Vertebral fracture, 388, 404

420

Index

Vertebral-vertebral AVF, 380 Vertebrobasilar system, 32–35 anatomic considerations, 35 distal posterior cerebral artery, 35 posterior cerebral artery, 32–35 Vertebrography, 401 Vertebroplasty, 388, 392 balloon-assisted, 387 clinical outcomes, 398–399 controversies with kyphoplasty, 406 symptomatic, complications in, 402–404 technical aspects, 392 acrylic injection, 396–398 acrylic preparation, 396 equipment requirements and operator skills, 393 patient preparation and monitoring, 393 pedicle targeting, 393–395 placement of a contralateral needle, 395 positioning of needle, 395 vertebrography, 396 trocars, 401

Viatrac 14 Plus balloon catherer, 168 Vidian artery, 18 Vinblastine therapy, 196 Vincristine therapy, 196 Vista Brite Tip, 164 Von Hippel-Lindau disease, 199 Wada test, 186–190 alternative agents, 188 epilepsy, 186 MEG, 188–189 noninvasive testing, 188 predictive value, 188 technique, 187–188 Warfarin-aspirin recurrent stroke study (WARSS) trial, 326 Warfarin aspirin symptomatic intracranial disease (WASID) trial, 71 Warfarin versus aspirin for symptomatic intracranial disease (WASID) study, 326 Warfarin versus aspirin study, 326 Wheal, 162

Willisian collaterals, 57–61, 63–64, 67, 69–70, 74–75 WingSpan Multicenter European Study, 329 WingSpan self-expanding stents, 330 WingSpan system (Smart Therapeutics), 329 for treatment of symptomatic intracranial stenosis, 329 and treatment with Gateway PTA balloon, 329 Wyburn-Mason syndrome, 275 XeCT perfusion. See stable xenon CT perfusion Y-stents, 254, 256 Zoom digital fluoroscopy, 325 Zygomatic-orbital artery, 24