Tumors of the Central Nervous System, Volume 3: Brain Tumors (Part 1)

Tumors of the Central Nervous System Tumors of the Central Nervous System Volume 3 For other titles published in thi...

Author: M.A. Hayat

44 downloads 1212 Views 12MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Tumors of the Central Nervous System

Tumors of the Central Nervous System Volume 3

For other titles published in this series, go to www.springer.com/series/8812

Tumors of the Central Nervous System Volume 3

Tumors of the Central Nervous System Brain Tumors (Part 1) Edited by

M.A. Hayat Distinguished Professor Department of Biological Sciences, Kean University, Union, NJ, USA

123

Editor M.A. Hayat Department of Biological Sciences Kean University Union, NJ, USA [email protected]

ISBN 978-94-007-1398-7 e-ISBN 978-94-007-1399-4 DOI 10.1007/978-94-007-1399-4 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2011923069 © Springer Science+Business Media B.V. 2011 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

“Although touched by technology, surgical pathology always has been, and remains, an art. Surgical pathologists, like all artists, depict in their artwork (surgical pathology reports) their interactions with nature: emotions, observations, and knowledge are all integrated. The resulting artwork is a poor record of complex phenomena.” Richard J. Reed MD

Preface

In this volume, as in volumes 1 and 2, the emphasis is on the diagnosis, therapy, and prognosis of brain tumors. In addition to describing strategies for advanced brain tumor treatment, this volume presents information on understanding the unique biology of the brain and its tumors. The information contained in this volume should aid in the development of tools for better diagnosis and effective treatment of brain malignancy. The application of various imaging techniques, including MRI, MRSI, PET, and CT, for diagnosing brain tumors including peripheral nerve sheath tumors is detailed. The use of MRS modality for classifying brain tumors is presented. This volume also contains information on the passage of malignancy to brain from tumors of other organs such as female breast and lung (tumor to tumor). The inception of both primary and secondary brain tumors is discussed. Also is included the delivery of drugs into brain tumors, considering the presence of blood brain barrier. A wide variety of treatments, such as conventional chemotherapy, electrochemotherapy, conventional resection, stereotactic radiosurgery, and magnetic resonance-guided focused ultrasound surgery in clinical practice, are explained in detail. The use of radioresponsive gene therapy for malignant brain tumors is included in this volume. The use of molecular markers as predictive and prognostic indicators in treatment decisions for individual cases are already beginning to have a significant positive effect on the clinical practice. A number of such markers are discussed in the volume. This volume also discusses pain management following craniotomy, antiepileptic drugs, and quality of life after brain tumor therapy and follow-up. By bringing together a large number of experts (oncologists, neurosurgeons, physicians, research scientists, and pathologists) in various aspects of this medical field, it is my hope that substantial progress will be made against this terrible disease. It would be difficult for a single author to discuss effectively the complexity of diagnosis, therapy, and prognosis of any type of tumor in one volume. This volume was written by 69 authors representing 12 countries. I am grateful to contributors for their promptness in accepting my suggestions. Their practical experience highlights their writings, which should build and further the endeavors of the readers in this important area of disease. I respect and appreciate the hard work and exceptional insight into the nature of cancer provided by these contributors. The contents of the volume are divided into subgroups: Introduction, Diagnosis and Biomarkers, Therapy, and Prognosis for the convenience of the readers. It is my hope that the current volume will join the preceding volumes of this series for assisting in the more complete understanding of globally relevant cancer vii

viii

Preface

syndromes. There exists a tremendous, urgent demand by the public and the scientific community to address to cancer prevention, diagnosis, treatment, and hopefully cure. I am thankful to Dr. Dawood Farahi, Dr. Kristie Reilly, and Mr. Philip Connelly for recognizing the importance of medical research and publishing in an institution of higher education, and providing the resources for completing this project. Union, New Jersey December 2010

M.A. Hayat

Contents

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M.A. Hayat 2 Brain Tumor Classification Using Magnetic Resonance Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Juan M. García-Gómez 3 Cellular Immortality in Brain Tumors: An Overview . . . . . . . . . Ruman Rahman and Richard G. Grundy Part I

1

5 21

Tumor to Tumor Passage of Malignancy . . . . . . . . . . . . .

33

4 Tumor-to-Tumor Metastasis: Extracranial Tumor Metastatic to Intracranial Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . Jian-Qiang Lu and Arthur W. Clark

35

5 Brain Metastases from Breast Cancer: Treatment and Prognosis . . Kazuhiko Ogawa, Shogo Ishiuchi, and Sadayuki Murayama

47

6 Brain Metastasis in Renal Cell Carcinoma Patients . . . . . . . . . . Aida Loudyi and Wolfram E. Samlowski

53

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain . . . . . . . . . . . . . . . . . . . . . . . . . . Naveen Sankhyan, Suvasini Sharma, and Sheffali Gulati 8 Breast Cancer and Renal Cell Cancer Metastases to the Brain . . . . Jonas M. Sheehan and Akshal S. Patel

63 75

9 Breast Cancer Brain Metastases: Genetic Profiling and Neurosurgical Therapy . . . . . . . . . . . . . . . . . . . . . . . Andreas M. Stark

85

10 Central Nervous System Tumours in Women Who Received Capecitabine and Lapatinib Therapy for Metastatic Breast Cancer . Stephanie Sutherland and Stephen Johnston

97

Part II

Biomarkers and Diagnosis . . . . . . . . . . . . . . . . . . . . .

107

11 Functional Role of the Novel NRP/B Tumor Suppressor Gene . . . . Theri Leica Degaki, Marcos Angelo Almeida Demasi, and Mari Cleide Sogayar

109

ix

x

12

13

14

15

Contents

Brain Tumors: Diagnostic Impact of PET Using Radiolabelled Amino Acids . . . . . . . . . . . . . . . . . . . . . . . Karl-Josef Langen, Matthias Weckesser, and Frank Floeth

117

Malignant Peripheral Nerve Sheath Tumors: Use of 18FDG-PET/CT . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andre A. le Roux and Abhijit Guha

127

Brain Tumors: Evaluation of Perfusion Using 3D-FSE-Pseudo-Continuous Arterial Spin Labeling . . . . . . . . . . Hanna Järnum, Linda Knutsson, and Elna-Marie Larsson

135

Cerebral Cavernous Malformations: Advanced Magnetic Resonance Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert Shenkar, Sameer A. Ansari, and Issam A. Awad

143

16

Nosologic Imaging of Brain Tumors Using MRI and MRSI . . . . . . Jan Luts, Teresa Laudadio, Albert J. Idema, Arjan W. Simonetti, Arend Heerschap, Dirk Vandermeulen, Johan A.K. Suykens, and Sabine Van Huffel

17

Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes . . . . . . . . . . . . . . . . . . . . . . . Kosuke Shimizu and Naoto Oku

155

169

18

Frozen Section Evaluation of Central Nervous System Lesions . . . . Richard Prayson

177

19

Clinical Role of MicroRNAs in Different Brain Tumors . . . . . . . . Richard Hummel, Jessica Maurer, and Joerg Haier

185

Part III Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

193

20

Electrochemotherapy for Primary and Secondary Brain Tumors . . Mette Linnert, Birgit Agerholm-Larsen, Faisal Mahmood, Helle K. Iversen, and Julie Gehl

195

21

Brain Tumors: Convection-Enhanced Delivery of Drugs (Method) . Anne-Laure Laine, Emilie Allard, Philippe Menei, and Catherine Passirani

207

22

Brain Metastases: Clinical Outcomes for Stereotactic Radiosurgery (Method) . . . . . . . . . . . . . . . . . . . . . . . . . Ameer L. Elaimy, Alexander R. MacKay, Wayne T. Lamoreaux, Robert K. Fairbanks, John J. Demakas, Barton S. Cooke, Benjamin J. Arthurs, and Christopher M. Lee

23

24

Noninvasive Treatment for Brain Tumors: Magnetic Resonance-Guided Focused Ultrasound Surgery . . . . . . . . . . . Ernst Martin and Ferenc A. Jolesz Radioguided Surgery of Brain Tumors . . . . . . . . . . . . . . . . . Laurent Menard

217

227 237

Contents

xi

25 Implications of Mutant Epidermal Growth Factor Variant III in Brain Tumor Development and Novel Targeted Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Murielle Mimeault and Surinder K. Batra 26 Endoscopic Port Surgery for Intraparenchymal Brain Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pawel G. Ochalski and Johnathan A. Engh

251

261

27 Intracranial Tumor Surgery in Elderly Patients . . . . . . . . . . . . Paul Ronning, Torstein Meling, Siril Rogne, and Eirik Helseth

269

28 Intracranial Hemangiopericytoma: Gamma Knife Surgery . . . . . Jason P. Sheehan and Edward M. Marchan

273

29 Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors . . . . . . . . . . . . . . . . . . . . . . . . Jesse J. Savage and Jason P. Sheehan

279

30 Malignant Brain Tumors: Role of Radioresponsive Gene Therapy . . Hideo Tsurushima and Akira Matsumura

287

Part IV Prognosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

293

31 Brain Tumors: Quality of Life . . . . . . . . . . . . . . . . . . . . . . Cristina D’Angelo, Antonio Mirijello, Giovanni Addolorato, and Vincenzo Antonio D’Angelo

295

32 Health-Related Quality of Life in Patients with High Grade Gliomas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eefje M. Sizoo and Martin J.B. Taphoorn

303

33 Epilepsy and Brain Tumours and Antiepileptic Drugs . . . . . . . . Sophie Dupont

313

34 Familial Caregivers of Patients with Brain Cancer . . . . . . . . . . Youngmee Kim

323

35 Pain Management Following Craniotomy . . . . . . . . . . . . . . . Doug Hughes and Scott Y. Rahimi

331

36 Air Transportation of Patients with Brain Tumours . . . . . . . . . . Peter Lindvall and Tommy Bergenheim

339

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

345

Contents of Volume 1

1 Introduction 2 Molecular Classification of Gliomas 3 Glioblastoma: Endosialin Marker for Pericytes 4 Glioma Grading Using Cerebral Blood Volume Heterogeneity 5 The Role of Ectonucleotidases in Glioma Cell Proliferation 6 Gliomas: Role of Monoamine Oxidase B in Diagnosis 7 Glioma: Role of Integrin in Pathogenesis and Therapy 8 Proton Magnetic Resonance Spectroscopy in Intracranial Gliomas 9 Infiltration Zone in Glioma: Proton Magnetic Resonance Spectroscopic Imaging 10 Malignant Gliomas: Role of E2F1 Transcription Factor 11 The Role of Glucose Transporter-1 (GLUT-1) in Malignant Gliomas 12 Malignant Gliomas: Role of Platelet-Derived Growth Factor Receptor A (PDGFRA) 13 Molecular Methods for Detection of Tumor Markers in Glioblastomas 14 Role of MGMT in Glioblastomas 15 Glioblastomas: Role of CXCL12 Chemokine 16 Cell Death Signaling in Glioblastoma Multiforme: Role of the Bcl2L12 Oncoprotein 17 Glioblastoma Multiforme: Role of Polycomb Group Proteins 18 Glioblastoma Multiforme: Role of Cell Cycle-Related Kinase Protein (Method) 19 Markers of Stem Cells in Gliomas 20 Efficient Derivation and Propagation of Glioblastoma Stem-Like Cells Under Serum-Free Conditions Using the Cambridge Protocol

xiii

xiv

Contents of Volume 1

21

Glioma Cell Lines: Role of Cancer Stem Cells

22

Glioblastoma Cancer Stem Cells: Response to Epidermal Growth Factor Receptor Kinase Inhibitors

23

Low- and High-Grade Gliomas: Extensive Surgical Resection

24

Brainstem Gangliogliomas: Total Resection and Close Follow-Up

25

Glioblastoma: Temozolomide-Based Chemotherapy

26

Drug-Resistant Glioma: Treatment with Imatinib Mesylate and Chlorimipramine

27

Glioblastoma Multiforme: Molecular Basis of Resistance to Erlotinib

28

Enhanced Glioma Chemosensitivity

29

Malignant Glioma Patients: Anti-Vascular Endothelial Growth Factor Monoclonal Antibody, Bevacizumab

30

Aggravating Endoplasmic Reticulum Stress by Combined Application of Bortezomib and Celecoxib as a Novel Therapeutic Strategy for Glioblastoma

31

Targeted Therapy for Malignant Gliomas

32

Glioblastomas: HER1/EGFR-Targeted Therapeutics

33

Epidermal Growth Factor Receptor Inhibition as a Therapeutic Strategy for Glioblastoma Multiforme

34

Role of Acyl-CoA Synthetases in Glioma Cell Survival and Its Therapeutic Implication

35

Malignant Glioma Patients: Combined Treatment with Radiation and Fotemustine

36

Malignant Glioma Immunotherapy: A Peptide Vaccine from Bench to Bedside

37

Malignant Glioma: Chemovirotherapy

38

Intracranial Glioma: Delivery of an Oncolytic Adenovirus

39

Use of Magnetic Resonance Spectroscopy Imaging (MRSI) in the Treatment Planning of Gliomas

40

Malignant Glioma Cells: Role of Trail-Induced Apoptosis

41

Long-Term Survivors of Glioblastoma

42

Glioblastoma Patients: p15 Methylation as a Prognostic Factor

Contents of Volume 2

1 Introduction 2 Gliomagenesis: Advantages and Limitations of Biomarkers 3 Molecular Subtypes of Gliomas 4 Glioblastoma: Germline Mutation of TP53 5 Familial Gliomas: Role of TP53 Gene 6 The Role of IDH1 and IDH2 Mutations in Malignant Gliomas 7 Malignant Glioma: Isocitrate Dehydrogenases 1 and 2 Mutations 8 Metabolic Differences in Different Regions of Glioma Samples 9

Glioblastoma Patients: Role of Methylated MGMT

10 Brain Tumor Angiogenesis and Glioma Grading: Role of Tumor Blood Volume and Permeability Estimates Using Perfusion CT 11 Vasculogenic Mimicry in Glioma 12 Newly Diagnosed Glioma: Diagnosis Using Positron Emission Tomography with Methionine and Fluorothymidine 13 Role of Diffusion Tensor Imaging in Differentiation of Glioblastomas from Solitary Brain Metastases 14

131 I-TM-601

SPECT imaging of Human Glioma

15 Assessment of Biological Target Volume Using Positron Emission Tomography in High-Grade Glioma Patients 16 Skin Metastases of Glioblastoma 17 Diffuse Low-Grade Gliomas: What Does “Complete Resection” Mean? 18 Quantitative Approach of the Natural Course of Diffuse Low-Grade Gliomas 19 Impact of Extent of Resection on Outcomes in Patients with High-Grade Gliomas

xv

xvi

Contents of Volume 2

20

Glioma Surgery: Intraoperative Low Field Magnetic Resonance Imaging

21

Low-Grade Gliomas: Intraoperative Electrical Stimulations

22

Malignant Gliomas: Present and Future Therapeutic Drugs

23

Recurrent Malignant Glioma Patients: Treatment with Conformal Radiotherapy and Systemic Therapy

24

Glioblastoma: Boron Neutron Capture Therapy

25

Glioblastoma: Anti-tumor Action of Cyclosporin A and Functionally Related Drugs

26

Glioblastoma Patients: Chemotherapy with Cisplatin, Temozolomide and Thalidomide

27

Glioblastoma: Role of Galectin-1 in Chemoresistance

28

Glioma-Initiating Cells: Interferon Treatment

29

Glioblastoma: Anti-tumor Action of Natural and Synthetic Cannabinoids

30

Patients with Recurrent High-Grade Glioma: Therapy with Combination of Bevacizumab and Irinotecan

31

Monitoring Gliomas In Vivo Using Diffusion-Weighted MRI During Gene Therapy-Induced Apoptosis

32

High-Grade Gliomas: Dendritic Cell Therapy

33

Glioblastoma Multiforme: Use of Adenoviral Vectors

34

Fischer/F98 Glioma Model: Methodology

35

Cellular and Molecular Characterization of Anti-VEGF and IL-6 Therapy in Experimental Glioma

36

Adult Brainstem Gliomas: Diagnosis and Treatment

37

The Use of Low Molecular Weight Heparin in the Treatment and Prevention of Thromboembolic Disease in Glioma Patients

38

Brainstem Gliomas: An Overview

39

Tumor-Associated Epilepsy in Patients with Glioma

40

Brain Tumors Arising in the Setting of Chronic Epilepsy

41

Low-Grade Gliomas: Role of Relative Cerebral Blood Volume in Malignant Transformation

42

Angiocentric Glioma-Induced Seizures: Lesionectomy

Contributors

Giovanni Addolorato Department of Internal Medicine, Catholic University of Rome, 8 – 00168 Rome, Italy, [email protected] Birgit Agerholm-Larsen Glostrup Research Institute, Copenhagen University Hospital Glostrup, 2600, Glostrup, Denmark, [email protected] Emilie Allard INSERM, U646, Universite d’Angers, Angers F-491000, France, [email protected] Sameer A. Ansari Section of Neuroradiology, Northwestern University Feinberg School of Medicine and the University of Chicago Pritzker School of Medicine, 5841 S. Maryland Ave., Chicago, IL 60637, USA, [email protected] Benjamin J. Arthurs Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Issam A. Awad Neurovascular Surgery Program, Section of Neurosurgery, University of Chicago Pritzker School of Medicine, 5841 S. Maryland Ave., Chicago, IL 60637, USA, [email protected] Surinder K. Batra Department of Biochemistry and Molecular Biology, College of Medicine, Eppley Cancer Institute, 7052 DRC, University of Nebraska Medical Center, 985870 Nebraska Medical Center, Omaha, NE 68198-5870, USA, [email protected] Tommy Bergenheim Umea University Hospital, Umea, Sweden Arthur W. Clark Department of Pathology and Laboratory Medicine, Foothills Medical Centre, Calgary, AB, Canada T2N 2T9, [email protected] Barton S. Cooke Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Cristina D’Angelo Department of Internal Medicine, Catholic University of Rome, Gemelli Hospital, l.go Gemelli, 8, 00168 Rome, Italy, [email protected] Vincenzo Antonio D’Angelo Department of Neurosurgery, IRCCS “Casa Sollievo della Sofferenza” Hospital, San Giovanni Rotondo, Italy, [email protected]

xvii

xviii

Theri Leica Degaki Department of Biochemistry, Chemistry Institute, NUCEL-Cell and Molecular Therapy Center, University of Sao Paulo, Sao Paulo 05508-900 SP, Brazil, [email protected] John J. Demakas Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Marcos Angelo Almeida Demasi Department of Biochemistry, Chemistry Institute, NUCEL-Cell and Molecular Therapy Center, University of Sao Paulo, Sao Paulo 05508-900 SP, Brazil, [email protected] Sophie Dupont Epilepsy Unit, Clinique Neurologique Paul Castaigne, Hopital de la Salpetriere, 75651 Paris cedex 13, France, [email protected] Ameer L. Elaimy Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Johnathan A. Engh Department of Neurological Surgery, University of Pittsburgh Medical Center, UPMC Presbyterian, Pittsburg, PA 15213, USA, [email protected] Robert K. Fairbanks Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Frank Floeth Department of Neurosurgery, Heinrich-Heine-University Düsseldorf, D-40225 Düsseldorf, Moorenstr. 5, [email protected] Juan M. García-Gómez Informatica Biomedica, Institudo de Aplicaciones de las Technologias de la Informacion y de las Comunicaciones Avanzadas, Universidad Politecnica de Valencia, Valencia, Spain, [email protected] Julie Gehl Department of Oncology, Copenhagen University Hospital Herlev, 2730 Herlev, Denmark, [email protected] Richard G. Grundy Children’s Brain Tumor Research Center, Medical School D Floor, School of Clinical Sciences, Queen’s Medical Centre, Nottingham, NG7 2UH, UK, [email protected] Abhijit Guha Division of Neurosurgery, Toronto Western Hospital, Toronto, ON, Canada, [email protected] Sheffali Gulati Division of Pediatric Neurology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India, [email protected] Joerg Haier Comprehensive Cancer Centre Muenster, International Patient Management, University Hospital Muenster, 48149 Muenster, Germany, [email protected] M.A. Hayat Department of Biological Sciences, Kean University, Union, NJ 07083, USA, [email protected] Arend Heerschap Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Eirik Helseth Department of Neurosurgery, OSLO University Hospital, Oslo, Norway, [email protected] Doug Hughes Department of Neurosurgery, Medical College of Georgia, Augusta, GA 30912, USA, [email protected]

Contributors

Contributors

xix

Richard Hummel Department of General and Visceral Surgery, University Hospital Muenster, 48149 Muenster, Germany, [email protected] Albert J. Idema Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Shogo Ishiuchi Department of Radiology, University of the Ryukyus, Nishihara-cho, Okinawa 903-0215, Japan, [email protected] Helle K. Iversen Glostrup Research Institute, Copenhagen University Hospital Glostrup, 2600, Glostrup, Denmark, [email protected] Hanna Järnum Department of Radiology, Aalborg Hospital, Arhus University Hospital, 9000 Aalborg, Denmark, [email protected] Stephen Johnston Breast Unit, Royal Marsden Hospital, Downs Road, Sutton, SM2 5PT, UK, [email protected] Ferenc A. Jolesz Brigham and Women’s Hospital, Harvard Medical School, Boston, USA, [email protected] Youngmee Kim Department of Psychology, University of Miami, Coral Gables, FL 33146, USA, [email protected] Linda Knutsson Department of Medical Radiation Physics, Lund University, Lund, Sweden, [email protected] Anne-Laure Laine INSERM, U646, Universite d’Angers, Angers F-491000, France, [email protected] Wayne T. Lamoreaux Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Karl-Josef Langen Institute of Neuroscience and Medicine, Forschungszentrum Jülich, D-52425 Jülich, Germany, [email protected] Elna-Marie Larsson Department of Radiology, Uppsala University, Uppsala University Hospital, SE 751 85 Uppsala, Sweden, [email protected] Teresa Laudadio Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Andre A. le Roux Division of Neurosurgery, Toronto Western Hospital, Toronto, ON, Canada, [email protected] Christopher M. Lee Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Peter Lindvall Umea University Hospital, Umea, Sweden, [email protected] Mette Linnert Department of Oncology, Copenhagen University Hospital Herlev, 2730 Herlev, Denmark, [email protected] Aida Loudyi Section of Melanoma, Renal Cancer and Immunotherapy, Nevada Cancer Institute, Las Vegas, NV 89135, USA, [email protected]

xx

Jian-Qiang Lu Department of Lab Medicine and Pathology, 5B2.24 WCM Health Sciences Centre, University of Alberta Hospital, Edmonton, AB, Canada TG6 2B7, [email protected] Jan Luts Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Alexander R. MacKay Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA, [email protected] Faisal Mahmood Department of Oncology, Copenhagen University Hospital Herlev, 2730 Herlev, Denmark, [email protected] Edward M. Marchan Department of Neurological Surgery, Health Sciences Center, Charlottesville, VA 22908, USA, [email protected] Ernst Martin University Children’s Hospital, CH-8032 Zurich, Switzerland, [email protected] Akira Matsumura Department of Neurosurgery, Clinical Medicine, Graduate School of Comprehensive Human Science, University of Tsukuba, Tsukuba, Ibaraki 305-8565, Japan, [email protected] Jessica Maurer Department of General and Visceral Surgery, University Hospital Muenster, 48149 Muenster, Germany, [email protected] Torstein Meling Department of Neurosurgery, OSLO University Hospital, Oslo, Norway, [email protected] Laurent Menard Laboratoire Imagerie et Modelisation en Neurobiologie et Cancerologie, IMNC- UMR 8165, Universita Paris Diderot 7, Paris, France, [email protected] Philippe Menei INSERM, U646, Universite d’Angers, Angers, F-491000, France; Centre Hospitalier Universitaire d’Angers, Angers cedex 9, F-49933, France, [email protected] Murielle Mimeault Department of Biochemistry and Molecular Biology, College of Medicine, Eppley Cancer Institute, 7052 DRC, University of Nebraska Medical Center, 985870 Nebraska Medical Center, Omaha, NE 68198-5870, USA, [email protected] Antonio Mirijello Department of Internal Medicine, Catholic University of Rome, 8 – 00168 Rome, Italy, [email protected] Sadayuki Murayama Department of Radiology, University of the Ryukyus, Nishihara-cho, Okinawa 903-0215, Japan, [email protected] Pawel G. Ochalski Department of Neurological Surgery, University of Pittsburgh Medical Center, UPMC Presbyterian, Pittsburg, PA 15213, USA, [email protected] Kazuhiko Ogawa Department of Radiology, University of the Ryukyus, Nishihara-cho, Okinawa 903-0215, Japan, [email protected] Naoto Oku Department of Medical Biochemistry, School of Pharmaceutical Sciences, University of Shizuoka, Shizuoka, 422-8526 Japan, [email protected]

Contributors

Contributors

xxi

Catherine Passirani INSERM, U646, Universite d’Angers, Angers F-491000, France, [email protected] Akshal S. Patel Department of Neurosurgery, Penn State Milton S. Hershey Medical Center, Hershey, PA 17033-0850, USA, [email protected] Richard Prayson Department of Anatomic Pathology, Cleveland Clinic Foundation, CCLCM, Cleveland, OH 44195, USA, [email protected] Scott Y. Rahimi Department of Neurosurgery, Medical College of Georgia, Augusta, GA 30912, USA, [email protected] Ruman Rahman Children’s Brain Tumor Research Center, Medical School D Floor, School of Clinical Sciences, Queen’s Medical Centre, Nottingham, NG7 2UH, UK, [email protected] Siril Rogne Department of Neurosurgery, OSLO University Hospital, Oslo, Norway, [email protected] Paul Ronning Department of Neurosurgery, OSLO University Hospital, Oslo, Norway, [email protected] Wolfram E. Samlowski 2435 Grassy Spring Pl, Las Vegas, NV 89135, USA, [email protected] Naveen Sankhyan Division of Pediatric Neurology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India, [email protected] Jesse J. Savage Department of Neurological Surgery, Health Sciences Center, Charlottesville, VA 22908, USA, [email protected] Suvasini Sharma Division of Pediatric Neurology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India, [email protected] Jason P. Sheehan Department of Neurological Surgery, Health Sciences Center, Charlottesville, VA 22908, USA, [email protected] Jonas M. Sheehan Department of Neurosurgery, Penn State Milton S. Hershey Medical Center, Hershey, PA 17033-0850, USA, [email protected] Robert Shenkar Neurovascular Surgery Program, Section of Neurosurgery, University of Chicago Pritzker School of Medicine, 5841 S. Maryland Ave., Chicago, IL, 60637, USA, [email protected] Kosuke Shimizu Department of Medical Biochemistry, School of Pharmaceutical Sciences, University of Shizuoka, Shizuoka, 422-8526 Japan, [email protected] Arjan W. Simonetti Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Eefje M. Sizoo Department of Neurology, VU University Medical Center, Amsterdam, The Netherlands, [email protected] Mari Cleide Sogayar Department of Biochemistry, Chemistry Institute, NUCEL-Cell and Molecular Therapy Center, University of Sao Paulo, Sao Paulo 05508-900 SP, Brazil, [email protected]

xxii

Andreas M. Stark Department of Neurosurgery, Schleswig-Holstein University Medical Center, Campus Kiel, 24105 Kiel, Germany, [email protected] Stephanie Sutherland Breast Unit, Royal Marsden Hospital, Downs Road, Sutton, SM2 5PT, UK, [email protected] Johan A.K. Suykens Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Martin J.B. Taphoorn Department of Neurology, VU University Medical Center, Amsterdam, The Netherlands; Department of Neurology, Medical Center Haaglanden, The Hague, The Netherlands, [email protected] Hideo Tsurushima Department of Neurosurgery, Clinical Medicine, Graduate School of Comprehensive Human Science, University of Tsukuba, Tsukuba, Ibaraki 305-8565, Japan, [email protected] Sabine Van Huffel Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Dirk Vandermeulen Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium, [email protected] Matthias Weckesser Department of Nuclear Medicine, Münster University, D-48149 Münster, Albert-Schweitzer-Strasse 33, Germany, [email protected]

Contributors

Chapter 1

Introduction M.A. Hayat

Keywords Tumor · CNS · Survival rate · Prognosis · Radiation · Dose Each year malignant tumors take a devastating toll on people, and among the most feared are brain tumors. Five-year survival rates per adults are disappointing, and mortality rates have not improved during the last 3 decades. Although overall 5-year survival rates have reached to 70% in children and mortality rates have declined 25% since 1970, prognosis is still poor for those inflicted with certain types of malignant tumors. There are manifold reasons, known and unknown, for lack of improved rates of survival. One of the main reasons is the difficulty encountered by drugs to cross the blood brain barrier that is a defense mechanism, protecting the brain from blood-born pathogens. Even when therapy is effective, its side-effects can cause serious disabilities. Another reason is that the diffuse infiltration of this neoplasm does not allow even the smallest surgical instruments to resect only the tumor cells bypassing the healthy neurons. In addition, these malignant cells are highly resistant to external radiation or systemic chemotherapy. Both radiation and chemotherapy can also have toxic effects not only on the tumor they are intended to treat but also on brain function. In other words, these treatments also kill normal brain cells. Functional deficits in patients after radiotherapy are probably more common than is currently reported. These deficits include mental retardation in patients and memory or cognitive deficits

M.A. Hayat () Department of Biological Sciences, Kean University, Union, NJ 07083, USA e-mail: [email protected]

in adults. Nevertheless, radiation therapy is a major component of the treatment of many primary and metastatic brain tumors. Doses higher than 60 Gy may produce vasogenic edema and necrosis in some patients. The 5-year relative survival rate following diagnosis of a primary malignant CNS tumor based on age is given below (CBTRUS): Age 0–19 years: 72.1% Age 20–44 years: 55.9% Age 45–54 years: 30.7% Age 55–64 years: 16.7% Age 65–74 years: 9.6% Age 75 or older: 5.2% From birth, males have a 0.67% lifetime risk of being diagnosed with a primary malignant CNS tumor, and 0.48% chance of dying from this cancer (excluding lymphomas, leukemias, and tumors of pituitary and pineal glands and olfactory tumors of the nasal cavity). From birth, females have a 0.54% lifetime risk of being diagnosed with this tumor, and a 0.38% chance of dying from this cancer. The 5-year relative survival rate following diagnosis of a primary malignant CNS tumor (including lymphomas and leukemias and tumors of pituitary and pineal glands, and olfactory tumors of the nasal cavity) is 33% for males and 37% for females. The estimated prevalence rate for all primary CNS tumors is 209/100.000. Approximately, more than 612,000 persons are living with this caner in the United States (malignant tumor: >124,000 and nonmalignant tumor: >488,000). The prevalence rate for all pediatric CNS tumors is estimated at 35.4/100,000, with more than 28,000 children living with this cancer in the United States.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_1, © Springer Science+Business Media B.V. 2011

1

2

The above-mentioned sobering statistics clearly indicate a considerable challenge to overcome brain tumors. To respond to this challenge, various experimental therapies have been administered, including gene therepy, antisense treatment, boron neutron capture, locoregional redioimmunotherapy, ligand-toxin conjugate administration and 5-aminolevulinic acid photodynamic therapy. Methods for sensitizing glioma cells to apoptosis induction and aiming at different targets such as the coagulation system have also been tried. These efforts have failed to significantly increase the overall survival of patients. Recently, Samnick et al. (2009) have tested the efficacy of 131 I-IPA combined with external beam photon radiotherapy as a new therapeutic approach against malignant glioma cells. This approach is based on the finding that malignant brain tumors accumulate amino acids more avidly than do healthy brains, using PET or SPECT (Hellwig et al., 2005). This finding led to the development of amino acid-based radiopharmaceuticals for detecting brain neoplasms. The use of this approach seems to merit a clinical trial to ascertain its potential in malignant glioma patients.

Causes of Developing Brain Tumors Although little is known regarding the causes of developing brain tumors, the following conditions may increase the risk of developing this neoplasm. Exposure to certain chemicals (e.g., vinyl chloride) and mutation of relevant genes are risk factors. Brain tumors can develop after medical radiation to the scalp or brain. Brain metastatic tumors can develop from cancer of other organs such as lung and breast. Certain viruses (Epstein-Barr virus and human cytomegalovirus) can also cause brain tumors. Diseased organ transplant can lead to primary CNS lymphoma. Genetic syndromes, such as neurofibromatosis types 1 or 2 and tuberous sclerosis, may increase the risk of developing brain tumors. Immune system disorders may also play a direct or indirect roe in developing these tumors. Some types of brain tumors tend to run in families. Although smoking, alcohol consumption, and certain dietary habits are associated with some types of cancer, they have not been directly linked to primary CNS tumors. Brain and spinal cord tumors are not contagious, and presently

M.A. Hayat

are not preventable. CNS tumors rarely spread outside the nervous system.

Distribution of Types of CNS Tumors There are many types of brain and spinal cord tumors (NCI): astrocytic tumors, embryonal tumors, ependymal tumors, germ cell tumors, meningeal tumors, mixed gliomas, oligodendroglial tumors, pineal parenchymal tumors, pituitary tumors, CNS lymphomas, tumors of the seller region, and other adult brain tumors. Anaplastic astrocytomas and glioblastoma account for ∼27% of brain tumors. Tumors that start in the brain are called primary brain tumors. Often tumors found in the brain are initiated somewhere else in the body and spread to one or more parts of the brain, and are called metastatic brain tumors. Brain metastases outnumber primary neoplasms by at least 10 to 1; the latter occur in 20–40% of cancer patients. The most common primary cancers metastasizing to the brain (tumor to tumor) are lung cancer (50%), breast cancer (15–20%), melanoma (10%), colon cancer (5%), and unknown primary cancers (10–15%). Approximately, 80% of brain metastases occur in the cerebral hemispheres, 15% occur in the cerebellum, and 5% occur in the brain stem. Metastases to the brain are multiple in >70% of cases, but solitary metastases also occur. Many brain tumors recur after they have been treated, and the recurrence may occur at the same cite or in other parts of the brain.

Tumor Grading Grading is based on the cellular make-up and location of the tumors. Tumors are graded in biopsy tissue or during surgery. The grade of a tumor can be used to indicate the difference between slow- and fastgrowing types of the tumor. Grade I tumors (e.g., pilocytic astrocytoma) grow slowly, do not spread into nearby tissues, and look like normal cells. It is possible to entirely remove this type of tumor by surgery. Grade II tumors (e.g., diffuse astrocytomas) also grow slowly, but may spread into nearby tissues, may recur after treatment, and may become a higher-grade tumor.

1

Introduction

Grade III tumors (e.g., annaplastic astrocytomas) grow rapidly, spread into nearby tissues, appear very different from normal cells, and may progress to a higher grade and become glioblastoma. Grade IV tumors (e.g., glioblastoma) grow and spread very quickly; the cells do not look like normal cells, and may show areas of dead cells.

Symptoms The symptoms caused by a brain tumor depend on its location in the brain, functions controlled by that part of the brain, and the size and grade of the tumor (NCI). Although the following symptoms are seen in brain tumor patients, other conditions may show the same symptoms. Headaches in the morning, which go away after vomiting. Frequent nausea and vomiting are not uncommon. Problems in normal speech, vision, and hearing are common. Trouble in walking and loss of balance may also be present. Depending on the location of the tumor in the brain, weakness on one side of the body may be found. Other symptoms include seizure, and unusual sleepiness and personal behavior.

Diagnosis Early Symptoms (mentioned elsewhere in this chapter and other chapters in this volume and in volume 1) necessitate immediate consultation with a physician. If the doctor suspects a brain tumor, a biopsy can be done to remove a sample of the tissue from the brain by removing a small part of the skull and using a needle. If a cancer is diagnosed under the microscope, the surgeon may remove as much tumor as safely possible during the same surgery or later, after detailed examination of the biopsy sample. A pathologist may check the cancer cells in the biopsy to find out the type and grade of the brain tumor and if the tumor is

3

likely to grow and spread. An imaging modality such as computed tomography (CT) or magnetic resonance imaging (MRI) can be used to find out if any cancer cells remain after surgery. These and other imaging procedures are also used to diagnose spinal tumors.

Prognosis Prognosis (chance of recovery) and treatment depend on a large number of factors, most of which are enumerated below (NCI). 1. The Type, grade, and location of the tumor in the brain. 2. Whether the tumor can be removed by surgery; if not, radiotherapy or chemotherapy, or both are alternate treatments. 3. Prognosis also depends on whether cancer cells remain after surgery. 4. Late or early diagnosis and whether the cancer has recurred. 5. The health and age of the patient. 6. The presence or absence of relevant gene mutations. 7. Whether there is a single tumor or more than one tumor in the brain. 8. Use of an imaging procedure to determine whether the tumor is responding to the treatment or is continuing to grow and spread.

References Hellwig D, Ketter R, Romieke BF, Sell N, Schaefer A, Moringlane JR, Krisch G, Samnick S (2005) Validation of brain tumor imaging with p-[123 I] iodo-L- phenylalanine and SPECT. Eur J Nucl Med Mol Imag 32:1041–1049 Samnick S, Romeike BF, Lehmann T, Israel I, Rube C, Mautes A, Reimers C, Kirsch C-M (2009) Efficacy of systemic radionuclide therapy with p-131 I-iodo- L -phenylalanine combined with external beam photon irradiation in treating malignant gliomas. J Nucl Med 50:2025–2032

Chapter 2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy Juan M. García-Gómez

Abstract The systematic compilation of Magnetic Resonance Spectroscopy (MRS) has allowed the application of statistical and signal processing techniques to analyze the contribution of metabolites and other compounds in the brain tissues. The complex nature of the MR spectra and the intrinsic difficulty of the Brain Tumor (BT) classification has led researchers towards the Machine Learning discipline, as an objective, as well as practical, methodology for discovering common patterns in the MR spectra acquired from the tumor tissues. This chapter tries to introduce the reader in the classification of brain tumor using MRS. The classification of the most prevalent types of brain tumors using MRS has been largely studied by several authors. Recently, classifiers for the childhood and for a wider range of types of tumors have been also obtained. Furthermore, incremental learning is a promising solution for the dynamism of the clinical environments. During the text we will justify the necessity of agreed acquisition protocols and prospective evaluation of the automatic classifiers to improve the predictive power of the classifiers. The aim of this chapter is to give a practical perspective of the automatic classification of brain tumors using magnetic resonance spectroscopy through the development of Clinical Decision Support Systems (CDSSs) and multicenter studies.

J.M. García-Gómez () Informatica Biomedica, Institudo de Aplicaciones de las Technologias de la Informacion y de las Comunicaciones Avanzadas, Universidad Politecnica de Valencia, Valencia, Spain e-mail: [email protected]

Keywords Magnetic resonance spectroscopy · Pattern classification · Brain tumors · Decision support systems · Multicenter evaluation study

Introduction MRS is an in-vivo noninvasive methodology requiring no ionizing radiation that allows a profile of the metabolites within a tissue to be obtained. The systematic compilation of MRS following agreed acquisition protocols has allowed the application of statistical and signal processing techniques to analyze the contribution of metabolites and other compounds in the brain tissues. Since the publication of the seminal paper by Preul et al. (1996), one major challenge during the last 2 decades has been the development of objective procedures to assist radiologists in the diagnosis of brain tumors by means of automatic classification of MRS signals from the patients. The complex nature of the MR spectra and the intrinsic difficulty of the BT classification has led researchers towards the Machine Learning discipline, as an objective, as well as practical, methodology for discovering common patterns in the MR spectra acquired from the tumor tissues. This chapter tries to introduce the reader in the classification of brain tumor using MRS. The application of the machine learning methodology will guide the exposition of the subject, illustrating the text through examples involving multicenter datasets. Along the chapter, we will try to range the next learning objectives:

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_2, © Springer Science+Business Media B.V. 2011

5

6

1. The first learning objective of the chapter will be to design of a brain tumor classification study with MRS based on the machine learning methodology. This general framework will lead us through the different steps to solve the automatic classification. 2. To enumerate the pre-processing steps needed to prepare the MR spectra for a correct classification study. 3. To summarize the feature extraction techniques applied to brain tumor diagnosis with MRS and to review relevant results in multicenter studies. 4. To summarize the classification techniques applied to brain tumor diagnosis with MRS and to review relevant results in multicenter studies and trends. 5. To justify the necessity of a correct evaluation of the classification results and to review a comparative evaluation with retrospective and prospective datasets. 6. To cite secondary outcomes of the automatic classification of brain tumors to analyze the contribution of metabolites, discover heterogeneous patterns, and detect outliers in the MRS datasets. 7. To cite Clinical Decision Support Systems (CDSSs) for brain tumor diagnosis using MRS.

MRS Classification Overview The life cycle of a Brain Tumor classification study based on MR spectroscopy mainly follows the Machine Learning methodology for solving a Pattern Recognition problem. It is composed of two main phases: the Training phase and the Recognition phase (see Fig. 2.1). During the Training phase, a set of signals following (the training corpus) a acquisition protocol is used to adapt a classification function. In this phase, a preprocessing and a features extracted from the signals are established. Afterwards, an adaptive model is fitted, selected and evaluated trying to obtain

Fig. 2.1 Design of a brain tumor classification with MR spectroscopy based on the machine learning approach

J.M. García-Gómez

the optimal generalization for predicting new cases. Once the model is ready, it can be incorporated into a CDSS to be used for the prediction of new cases, where the preprocessing and feature extraction steps will be carried out before applying the classification function. The rest of the chapter reviews the main techniques of each step of the Machine Learning methodology applied to Brain Tumor classification with MRS. Section “MRS Classification Overview” specifies the well-established pre-processing pipeline agreed in the eTUMOR project for normalizing MR spectra. In section “Preprocessing Magnetic Resonance Spectroscopy” the main pattern recognition techniques for extracting relevant features from MR spectra are studied. That section ends with a review of the effect of feature extraction from MRS in brain tumor classification. Section “Feature Extraction” studies the Machine Learning approach for classification, its techniques and its application to different problems of brain tumor diagnosis. The relevance of an accurate evaluation is studied in section “Peak Integration” by comparing retrospective and prospective evaluations of brain tumor classifiers. The use of the classification results to interpret of signal patterns, detect outliers, and perform quality control of MRS biobanks is presented in section “Stepwise Algorithm for Feature Selection in Classification”. Before conclusions, section “Relieff Feature Selection” provides an enumeration of CDSS for brain tumor diagnosis using MRS.

Preprocessing Magnetic Resonance Spectroscopy A spectrum acquired with a Time Echo (TE) <45 ms is usually considered a Short TE spectrum, and a Long TE spectrum otherwise. Different criteria have been argued in favor and against every option (e.g. Majos et al. (2004)), whereas the multicenter INTERPRET

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

and eTUMOUR projects1 defined protocols based on the acquisition of both Short TE and Long TE spectra for the same patient. Short TE (20–35 ms) 1 H MRS allows to observe several metabolites and other compounds considered useful for tumor classification: macromolecules (MM; 5.4 ppm, 2.9 ppm, 2.25 ppm, 2.05 ppm, 1.4 ppm and 0.87 ppm), myo-Inositol (mI) and Mobile Lipids (ML). However, Short TE signals are more sensible to artifacts, show a large number of overlapping peaks, and a strong MM-/ML-originated baseline. Long TE (about 135 ms) 1 H MRS is less informative than Short TE, because resonances with short T2 may be lost. However, lipid resonances (1.3 and 0.9 ppm) and MM will not be the dominating components at Long TE, making easier the analysis of the spectrum and possible the study of contributions from lactate (Lac, doublet at 1.33 ppm) and alanine (doublet at 1.47 ppm). BT classification requires the acquisition of homogeneous MRS. Therefore, acquisition protocols should be agreed at the very beginning of a classification study is started. As an example, data in the INTERPRET and eTUMOUR projects were acquired with Single Voxel (SV) 1 H MRS at 1.5 T, avoiding areas of cysts or necrosis and with minimum contamination from the surrounding non-tumoral tissue. Volume of interest size ranged between 1.5 × 1.5 × 1.5 cm3 , (3.4 mL) and 2 × 2 × 2 cm3 , (8 mL), depending on tumor dimensions. Long TE spectra were acquired with the PRESS sequence, and a recycling time (TR) between 1,500 and 2,020 ms, TE of 135 or 136 ms, spectral width of 1,000 or 2,500 Hz and 512 or 2,048 data points. Short TE were acquired using PRESS or STEAM sequences, with TR between 1,600 and 2,020 ms, TE of 20 or 30 ms, spectral width of 1,000 or 2,500 Hz and 512 or 2,048 data points. Before using the spectra for classification purposes, a quality control was carried out over both Long TE and Short TE spectra. Once a case is acquired following the agreed protocol, it is preprocessed to make it compatible, independent on the manufacturer and acquisition configuration. A worthy idea for consideration is that the preprocessing should be applied to each training and test case, hence,

1 Interpret acquisition protocols 2000. http://azizu.uab.es/ INTERPRET/mrsdata/mrsdata.html. eTUMOR acquisition protocols 2003. http://www.etumour.net

7

it is recommendable to be an automatic (or mostly automatic) procedure. The next automatic pipeline was agreed within the eTUMOUR project. It has demonstrated its usefulness in multicenter studies, such as the ones carried out by Luts et al. (2008) and García-Gómez et al. (2009a). It consists of eight steps applied consecutively to the raw data file of a SV MR spectrum (Short or Long TE) from which an unsuppressed water acquisition is also available. 1. Eddy current correction is applied to the watersuppressed Free Induction Decay (FID) of each case using the Klose algorithm. An additional manual zero-order and first-order phase correction can be performed to improve the phase correction of the MR spectra, being strongly recommended for 3 T samples. 2. The residual water resonance is removed using the Hankel-Lanczos Singular Value Decomposition (HLSVD) time-domain selective filtering using 10 singular values and a water region of [4.33, 5.07] ppm. 3. An apodization with a Lorentzian function of 1 Hz of damping is applied. 4. Before transforming the signal to the frequency domain using the Fast Fourier Transform (FFT), an interpolation is needed in order to increase the frequency resolution of the low resolution spectra to the maximum frequency resolution used in the acquisition protocols. This can be carried out with the zero-filling procedure. 5. Afterwards, the baseline offset, which can be estimated as the mean value of the region [11, 9] U [–2, –1] ppm, is subtracted from the spectrum. 6. The normalization of the spectral data vector to the L2-norm is performed based on the data-points in the region [–2.7, 4.33] U [5.07, 7.1] ppm. 7. Depending on the Signal-to-Noise Ratio (SNR) and the tumor pattern, an additional frequency alignment check of the spectrum should be performed by referencing the ppm-axis to (in order of priority) the total Creatine (Cr) at 3.03 ppm or to the Choline (Cho) containing compounds at 3.21 ppm or the Mobile Lipids (ML) at 1.29 ppm. 8. Finally, the region of interest is restricted to [0.5, 4.1] ppm. This pipeline can be adapted for the preprocessing of Magnetic Resonance Spectroscopic Imaging (MRSI) if

8

we include few steps at the beginning of the process. First, a filtering of k-space data by a Hanning filter is applied. After that, a zero filling to 32×32 and a spatial 2D Fourier transformation obtain the time domain signals for each voxel.

Feature Extraction Menze et al. (2006) defined two different approaches for feature extraction from MR spectra: (a) the quantification approach, and (b) the pattern recognition approach. The quantification approach is based on an inverse modeling of the resonance lines to infer the absolute or relative concentrations of the biochemical agents. Algorithms, such as AQSES, QUEST, AMARES, or LCModel are based on this approach. In this chapter, we will mainly discuss about the second approach, based on feature extraction from the spectra with no assumption of an underlined model. Menze et al. (2006), Luts et al. (2008), and García-Gómez et al. (2009a) have achieved highaccurate results in brain tumor classification with pattern recognition-based techniques, whereas Opstad et al. (2007) and Weis et al. (2010) have argued in favor of quantification methods based on LCModel. The feature extraction methods based on pattern recognition can be applied directly to the spectra, with no casespecific parameter to be tuned. Hence, pattern recognition approach is considered to offer a good tradeoff between usability of the technique and reproducibility of the results. Different methods have been applied for obtaining relevant features for brain tumor classification using MRS. Luts et al. (2008) examined the effect of feature extraction methods, mainly from the pattern recognition approach, prior to automated classification based on MRS for BT diagnosis. In that study, Peak Integration (PI) on selected metabolite resonance regions, peak height of typical resonances (PPM), Principal Component Analysis (PCA), Independent Component Analysis (ICA) and Wavelet (WAV) transformation, among others were compared on MRS and MRSI real data. Additionally, García-Gómez (2009a) evaluated these methods on multicenter prospective datasets. Several studies have included the simple PPM method to extract the maximum height values of the resonance peaks (Table 2.1) of the metabolites as inputs of the classification.

J.M. García-Gómez Table 2.1 Typical PPM of metabolite/molecule resonances and other molecules observed in short TE and the interval of integration used in PI Resonance Resonance frequency (ppm) Region (ppm) L2 0.92 0.15 L1 1.29 0.15 LAC 1.31 0.15 ALA 1.47 0.15 NAA 2.01 0.15 Cr 3.02 0.15 Cr2 3.92 0.15 Cho 3.21 0.15 Gly 3.55 0.15 Glx 2.04 0.15 GLx2 2.46 0.15 mI/Tau 3.26 0.15 mI2 3.53 0.15 Tau2 3.42 0.15 ALA2∗ 3.78 0.15 ∗ ALA and others alpha-CH from amino-acids (e.g. Glx)

Peak Integration Peak Integration (PI) is a simply form of estimating a quantity relative to the metabolite concentration, assuming that the amplitude of a metabolite resonance is proportional to the integral of its corresponding peaks in the spectrum. A precise estimation of the peak integrals is difficult due to several factors, including nonzero baseline, peak overlap, noise and also the discrete nature of the spectrum. In contrast, Peak Integration (PI) is a good choice when an accurate estimation of the metabolites is the objective of the study. The areas of the regions around the resonance frequencies of the metabolites can be estimated by the trapezoidal rule. Let (x(f1 ), . . . ,x(fl )) be a discretized sample of the function x(f) from which the integral with respect to f in a the range (f1 , . . . ,fl ) is wanted to be fl approximated I = x(f )df . In practice, the trapezoidal f1

integration is computed as: f = ( f2 − f1 ,..., fl − fl−1 ) x( f1 ) + x( f2 ) x( fl−1 ) + x( fl ) x = ,..., 2 2 I = f · x

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

For each selected metabolite resonance the area under the frequency peak in the magnitude spectrum is calculated. Several studies have used fifteen ranges for Short TE spectra integrated within a window of 0.15 ppm around the expected chemical shift of the main resonances of the metabolites (Table 2.1).

Stepwise Algorithm for Feature Selection in Classification StepWise (SW) is one of the most extended algorithms for solving the feature selection task. SW consists in a greedy hill climbing approach where the subset of features with the highest performance measure will be selected in each step and modified in the next step by the addition or deletion of one variable in the model. To select the next model, the algorithm compares the visited models by a measure over a validation dataset. Garczarek (2002) defined measures of the performance, such as, the Ability to Separate (AS) to check the unlike the classes are after the transformation carried out by the model, by means of the euclidean distance between the true posterior probability, p(c|x), and the posterior probability of the model p(c|x;).

9

Principal Components Analysis (PCA) PCA is a well-known projection method for studying the variability in multivariate data.Consequently, when applied to MR spectra, we will assumed the case x in frequency domain as a D-dimensional space of variables (x1 , . . . ,xD ). PCA searches in the observational D-dimensional space those p directions (called principal components) where data have the highest variability. Each principal component i = 1, . . . , p defines one direction of variability by means of a loading vector wi , and for each case x, the principal component score zi is given by zi = wTi · x, assuming zero empirical mean. PCA finds the loading vector w1 that maximizes the variance of the principal component scores zi , i.e. w1 = arg max||w1 ||=1 Var(z1 ), subject to the constraint ||w1 || = 1. The second (or a higher) principal component is defined is the same way, but also constrained to orthogonality to the first (other) component(s), w1 · w2 = 0. The transformation matrix W, composed by the loading vectors wp is usually computed by means of the eigenvalue decomposition of the covariance matrix of the original cases. Once principal components are obtained, it is usual to represent each case by means of a vector of the p-firsts z-scores.

Functional Data Analysis Relieff Feature Selection ReliefF is a feature selection method based on how well features distinguish between instances that are near to each other. In classification problems, the quality of each variable is calculated by the accumulation of the distance between randomly selected instances and their k-nearest neighbors of a different class minus the distance to the k neighbors of the same class. To make comparable the variables, the distance is normalized by the range of the feature. For neighbors of different classes, a large distance for a variable will substantially increase the quality of the variable, whereas for neighbors of the same class, a large distance for a variable will substantially decrease its quality. As a result, a discriminative variable will have a large distance with samples of other classes and a short distance with samples of the same class.

The MR spectrum can be seen as a function defined on time, x(t), or on frequency, x(f), composed by the contributions of a mixture of resonances of metabolites. Pattern Recognition (PR) methods for classification and clustering are usually applied on data of finitedimensional spaces, but they cannot be directly applied to infinite-dimensional space (such as a function, a temporal series, or a curve). Although the previous methods are based on the discretization of the frequency interval of interest, they have been successfully applied for MRS classification in previous studies. Nevertheless, the space resulting from this multidimensional approach results in highly correlated highdimensional data that should be taken into account when estimating the predictive models. Therefore, regularization techniques or dimensional reduction should be applied to avoid overfitting.

10

J.M. García-Gómez

The alternative proposed by Ramsay and Silverman (2002) consists in fitting the curve of each spectrum to la linear combination of l basis functions x(f ) = i=1 αi υi (f ), such as cubic splines, or wavelets. This can be carried out by the minimization of a least squares fitting criterion with a regularization term over the roughness of the second derivative of the fit. Then, the resulting l-vector α of coefficients can be the input of the classifiers as a finite-dimensional representation of the spectrum. In the functional version of PCA, each eigen vector wi of the multivariate data x now corresponds to a eigen function ξi (f ) and the i eigen component score is zi = ξi (f )x(f ). In fact, functional Principal Component Analysis (fPCA) finds the eigen functions ξi (f ) that maximizes the variance of the principal 2 component scores zi , subject to the constraint ξi (f )df = 1 and the orthogonality to all the previous principal components. As a result, we can represent each function xi (f ) (MR spectrum) by a vector of scores zi obtained by fPCA.

Wavelet Transform and Multi-resolution Analysis The wavelet transform consists in carrying out translations and scale transformations of a prototypical wavelet function ψ in order to adjust the shape of a signal and to successively obtain a linear expansion of it with coefficients γ(s, τ ). Mathematically this can be expressed as: x(t) =

γ (s, τ )ψs,τ (t)dτ ds,

where x(t) is the signal and the variables s and τ are referred to the translation and scale dimension. Each scaled and translated wavelet ψs,τ (t) is called child wavelet and is generated from a common mother wavelet ψ(t) by ψs,τ (t) = √1s ψ( t−τ s ). The wavelet is defined as a finite length or fast decay wave with the admissibility and the regularity conditions. These properties imply that the Fourier transform of ψ(t) tends to zero for low frequencies and therefore their behavior is similar to a band pass filter. Using this wavelet transform and taking into account

the filter behavior of the wavelet and scaling functions, the signal can be divided into different resolution levels. As a result, each MR spectrum can be represented by the parameter-space composed by the γ(s, τ ) coefficients. Afterwards it can be applied some feature selection procedure (such as SW in 3.2) to obtain an optimal input vector for classification.

Independent Component Analysis The motivation to apply ICA to a set of MR spectra is due to the presence of partial volume effects. Partial volume effects result in the fact that a signal from a specific voxel can contain components of different tissue types. The input for the ICA method is the full region of interest of the real spectrum. Given n MR spectra (x1 (t), . . . ,xn (t)), each one composed as a linear combination of n indepen dent sources, xi (t) = nj=1 aij sj (t), ∀i = 1, ..., n, ICA attempts to un-mix the sources sj (t). Let X = {xi (tk )}n1 the m×n matrix of the discretized cases xi (t), such as, X = SA, where S contains the independent sources and A the linear mixing coefficients. ICA estimates the un-mixing matrix W that makes XW = S. Due to the Central Limit Theorem, ICA assumes that the generative model X tends to be more Gaussian than the sources S. As a consequence, the optimal W is such that maximizes the non-gaussianity of the sources. Once ICA is obtained for the training dataset, coefficients in the S basis for a new case x∗ (t) = (x∗ (t1 ), . . . ,x∗ (tm )) it can be easily computed as (ST S)−1 ST x∗ (t). Most ICA algorithms start with a pre-whitening step, based on a PCA of the observations. After pre-whitening, a dimensionality reduction is obtained in the source signal subspace.

Feature Extraction for Brain Tumor Classification Based on MR Spectra Several studies have investigated the effect of feature extraction for brain tumor classification based on MR spectra. Despite of the simplicity of PI, GarcíaGómez et al. (2009a) obtained high performances in several pairwise and multi-class classification tasks

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

compared to other more complex methods. Devos et al. (2004) saw similar performances when comparing full region of interest, peak regions and PI. Simonetti et al. (2005) compared PCA, ICA, LCModel and PI for feature extraction on Short TE MRSI data and they also obtained the best results with PI. By the contrary, Opstad et al. (2007) obtained better results with LCModel quantification than with PCA for two-step LDA classification in a single-center study. Lukas et al. (2004) observed a better performance using the full region of interest rather than using PI or peak region extraction in BT classification with SV Long TE spectra. Finally, Menze et al. (2006) and Luts et al. (2008) obtained an improvement when PR approaches (e.g. ICA, PCA, binned peak region and WAV) were used in Short or Long TE instead of quantification approaches. García-Gómez et al. (2008) investigated if the combination of features from Short TE and Long TE SV at 1.5 T, improve the classification of brain tumors with respect to using only one echo time. Based on a clinically validated dataset of 50 lowgrade meningiomas (MEN), 105 aggressive tumors (AGG, glioblastoma and metastasis), and 30 low-grade glial tumors (LGG, astrocytomas grade II, oligodendrogliomas and oligoastrocytomas) predictive models based on the combination of features extracted by SW, ReliefF and PCA from Short and Long TE spectra. Significant differences in performance were found when Short TE, Long TE or both spectra combined were used as input. In their results, the combination of the two TE acquisitions produced optimal performance to discriminate MEN, whereas the use of Short TE acquisition alone was preferable to discriminate AGG from LGG. These results are consistent with those described by Majos et al. (2004). Besides, Luts et al. (2008) studied the effect of feature extraction methods in MRSI in pairwise classification. In his results, ICA, Relief-F, and full spectra in combination with LS-SVMs produced the highest overall performances. Generally speaking, the reproducibility to every case with no tuning that pattern recognition-based methods present. By contrast, quantification approach allows the researcher to obtain more accurate concentration-relative values of the metabolites of interest after manual tuning of the parameters for each case. Taken into account the comparative results of the effect of feature extraction using both

11

approach, pattern recognition methods give a reasonable usefulness-reproducibility tradeoff.

Automatic Classification The classification of Brain Tumors by MR spectroscopy involves several challenges, such as (i) complex organizational decision-making involving risk and uncertainty, (ii) difficult discrimination of tumor types, (iii) heterogeneous classes, (iv) multiple types of tumors, (v) different prevalence of the tumor types, (vi) complex signals composed by mixtures of components with biological meaning, (vii) high-/infinitedimensionality of the input space, (viii) dynamism of diagnostic techniques. The Machine Learning discipline provides the mathematical and computational mechanisms to take advance of the available biological knowledge and data gathered from the problem domain. A wide range of Machine Learning methods for automatic classification has been explored for discriminating BT using MR spectroscopy. We will review some of these methods and their applications, such as parametric discriminant analysis: Linear Discriminant Analysis (LDA); kernel-based models (Support Vector Machines (SVM) and Least Squares SVM (LS-SVM)); artificial neural networks (Multilayer Perceptron (MLP) and Bi-directional Kohonen network (BDK); ensemble models and kNearest Neighbours (KNN)).

Gaussian Parametric Model Discriminant analysis is designed to find boundaries between classes. Gaussian parametric models are based on the max-likelihood estimation of the Gaussian distributions for the classes under the study. The most popular Gaussian method is LDA, which is based on the assumption of a common variance of the classes. In the Quadratic Discriminant Analysis (QDA) the covariance of the classes are independent, obtaining quadratic decision boundaries. The Fisher’s rank-reduced version of LDA (FLDA) is a reducedrank version of LDA, which projects the variables into the lower-dimensional subspace that maximizes the rate of the between-variance and the within-variance on the training corpus.

12

K-Nearest Neighbors The KNN is a non-parametric classification method on which the samples are assigned to the classes based on the distances of the test cases to the training corpus in the feature space. KNN is a type of instance-based learning where the discrimination function is approximated locally and all computation is deferred until classification.

Multilayer Perceptron The MLP is a connectionist model consisting of a network of perceptrons. One perceptron computes a non-linear function of the inner-product of the feature vector x and the weight vector w. In a MLP for solving a PR problem, the input signal propagates through the perceptrons layer-by-layer obtaining an approximation of the probability distribution of each class. During the Training phase, the output error is propagated to the previous layers to update the weight vector of the perceptrons.

Bi-Directional Kohonen Network Another type of artificial neural networks is the Kohonen network, which is a mapping technique that incorporates the topology present in the data. Each neuron in the bidimensional Kohonen map is associated with a weight vector which is adapted iteratively by some learning function, based on the properties of the objects. Individual objects are iteratively presented to the units in the network and the weight vector that is most similar to the particular object is assigned to be the winning unit. The winning unit and its neighborhood are then adjusted to become more similar to the particular object. The final Kohonen map then represents the structure of the data in an interpretable way. In the supervised Bi-directional Kohonen Networks (BDK) two separate maps are updated, namely the input map (Xmap, representing the features of the objects) and the output map (Ymap, representing e.g. the class labels of the objects). The winning unit k for a sample i is mainly determined by the similarity of the output (yi ) and the Ymapk unit. Subsequently, the

J.M. García-Gómez

Ymap winning unit is determined mainly by the similarity of the input (X) properties of an object and the Xmap units.

Ensemble Models Ensemble assumes that if a set of individual classifiers are accurate enough and diverse, an ensemble of them performs better than a single predictor in the average. This is based on the reduction of the variance component of the error. Different schemas are applied to combine the classifiers to obtain the ensemble models. For example, a typical ensemble model consists in a voting schema over classifiers trained with subsets of a dataset. Another approach for ensembling models is to perform an average of the outputs of the models of each local representation.

Automatic Classification of Brain Tumor Based on MR Spectra The comparison study performed by García-Gómez et al. (2009a) did not find significant differences were among the classification techniques for pairwise classification of BT types. In that study, 253 pairwise classifiers for Glioblastoma, Meningioma, Metastasis, and Low-Grade Glial diagnosis were inferred from 211 SV Short TE INTERPRET MR spectra obtained at 1.5 T (PRESS or STEAM, 20–32 ms), afterwards, they were tested with 97 prospective spectra compiled during eTUMOUR. They obtained similar results with FLDA than with the other methods in average; however, other methods like LS-SVM and BDK performed better for some discrimination problems (e.g. Glioblastoma (GBM) vs. Low-Grade Glial (LGG)). Devos et al. (2004) observed similar performances of their LDA and LS-SVM classifiers based on PI evaluated by the area under the ROC curves. A trend to lower results of Balance Error Rate (BER) were obtained using a BDK. Melssen et al. (2006) applied BDK to PI values to discriminate over tumor grades and other tissues in the INTERPRET multi-voxel dataset. Tate et al. (2003) based their three-class classifiers on the LDA due to the ability of this method for projecting the results in a 2-dimensional space for visualization.

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

Moreover, the developemt of multiclass classification of brain tumors using a combination of MRI and MRSI was performed by Luts et al. (2009) to generate nosologic images of the brain. Luts et al. (2008) developed a discrimination method between glioblastoma and solitary metastases based on long TE 2D-TSI and MRI based on the differencies of Cho/Cr and Cho/NAA ratios in the enhancing tumor and the perienhancing area.

Including More Classes in the Automatic Classification of Brain Tumors When designing an automatic tool for Brain Tumors classification, it would be of interest to be able to accept prospective cases from any tumor type. Most of the studies in BT classification using MRS have been focused on the prevalent brain tumor classes (i.e. Glioblastoma, Meningioma, Low Grade Glias, Metastasis). García-Gómez et al. (2009b) presented an one-class classification strategy of BT using SV Short TE MRS based on Functional Data Analysis (FDA) (see Section Functional Data Analysis) and subpattern search. The dataset used in their study was the largest multicenter database available till that moment with 967 1H-MR spectra of histopathologically diagnosed brain tumor cases compiled by ten institutions in the framework of the INTERPRET and eTUMOUR projects. The distribution of cases by pathology was: 27 Astrocytoma I (AS1), 74 Astrocytoma II (AS2), 37 Astrocytoma III (AS3), 252 Glioblastoma (GBM), 16 Lymphoma (LYM), 26 Medulloblastoma (MED), 80 Meningioma (MEN), 118 Metastasis (MET), 40 Oligodendroglioma II (OD2), 12 Oligodendroglioma III (OD3), 26 Oligoastrocytoma II (OA2), 7 Oligoastrocytoma III (OA3). There were 225 cases that belonged to other less frequent tumor types (OTH) and there were 27 Normal brain spectra (NOR). Signal processing was performed following the protocols defined in section “MRS Classification Overview”. Afterwards, each spectrum was modeled by a function of frequency defined in a base of cubic splines in the [0.5, 4.1 ppm] interval. In order to fit the spectral shape, first derivative functions were registered to their mean function. Unimodal and multi-modal linear one-class classifiers were trained based on the fPCA

13

Table 2.2 Performance estimation (ROC-AUC) by diagnosis for the one-class classifiers. Number of subpatterns (SP) is also shown. See section “Stepwise Algorithm for Feature Selection in Classification” for discussion about the subpattern discovery in heterogeneous classes Subpattern-based Diagnostic foca-Gauss foca-Gauss Class (AUC) (AUC) Subpatterns AS1 AS2 AS3 GBM LYM MED MEN MET OD2 OD3 OA2 OA3 OTH NOR

0,70 0,79 0,66 0,69 0,69 0,86 0,86 0,64 0,79 0,30 0,68 0,83 0,49 0,94

0,70 0,79 0,75 0,90 0,75 0,89 0,89 0,79 0,78 0,66 0,70 0,79 0,80 9,96

2 4 2 3

5

parameter-space for the 14 pathology diagnoses. The performances of the one-class classifiers were evaluated by means of the Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve estimated on a 5×5 Cross Validation re-sampling strategy with the results shown in Table 2.2.

Classification of Brain Tumor in Children and Adults Limited experiments have been reported in automatic classification of brain tumor in the childhood. Nevertheless, the biochemical profiles and prevalences of brain tumors give an evidence about the differences on brain tumor types in childhood and adulthood. At time, Davies et al. (2008) published one of the few studies about automatic classification in childrenhood. They were interested in the discrimination of medulloblastomas, pilocytic astrocytomas and ependymomas by linear models over LCModel quantifications of Short TE spectra and High Resolution Magic Angel Spinning (HR-MAS). Vicente et al. (2009) studied the compatibility of automatic classifiers for Brain Tumours based on MRS in the childhood and adulthood. This study was aimed

14

by the design of a filter based on the patient age to internally decide when the classifier specialized with child patients should be used. 489 (93 children, 396 adults) SV Short TE MRS with histopathological diagnosis of brain tumor were preprocessed by the pipeline defined in section “MRS Classification Overview”. Pathology distribution in children was: 60 glials (12 high grade (HG) and 48 low grade (LG)), 24 Medulloblastoma (MED), 3 Meningioma (MEN) and 6 cases belonging to less frequent tumor types (OTH). Besides, the distribution of pathologies in adults was: 233 glials (156 HG, 77 LG), 2 MED, 69 MEN, 65 Metastasis (MET), 5 Normal brain spectra (NOR) and 22 OTH. Mean age in children and adults was 11.5 and 54.1, respectively. LDA and KNN classifiers were trained with the children or adult cases to discriminate between aggressive and non-aggressive BT using PI, and PCA. The authors evaluated the models by means of the balanced accuracy rate (BAR), equivalent to 1-BER. Children classifiers were tested with two independent test sets: one containing 33 children cases and other with 396 adult cases. Analogously for the adult classifiers, an independent test set of 96 adult cases and a second with the total set of childhood cases (93) were applied. Mean performance of classifiers for adults was BAR = 0.84 tested with adulthood cases, and BAR = 0.80 for children tested with childhood cases. On the contrary, when the children test set was applied to the adult classifiers, a low mean performance was observed (BAR = 0.50), analogously to the test of children classifier with adult cases (BAR = 0.52). This showed the need of specificity on the classifiers by age. Based on the result, a “filtered classifier” was designed and tested with an independent test set of 33 children cases and 96 adult cases obtaining BAR = 0.85 in an independent test. This result reinforces the idea that the nature of childhood and adulthood BT makes necessary the design specific classifiers taking into account the age.

Incremental Learning in Brain Tumor Classification Currently, training of the classifier parameters is carried out once keeping them frozen along its time of use and failing to take benefit from future information.

J.M. García-Gómez

Tortajada et al. (2008) presented the evaluation of a incremental learning methodology for brain tumor classifiers based on SV Short TE spectra in comparison to classical methodologies of static classifiers. They chose a simple yet effective approach based on KNN to implement an incremental learning based on memory, and it was compared to the classical static classifiers. Both classical and incremental classifiers used PI as input of the classification, and the number of K neighbors was optimized in a previous step. Two classification tasks were carried out to test both approaches: (i) Meningioma (n = 73) vs. Non-Meningioma (n = 221) and (ii) Aggressive (n = 165) vs. Meningioma (n = 73) vs. Low Grade Glial (n = 56). The results showed significant improvements of the incremental classifiers compared to the classical classifiers.

Compatibility of 3 T MRS with 1.5 T-Based Classifiers Most datasets of MRS compiled for brain tumor research has been acquired in 1.5 T scanners. Nevertheless, 3 T scanners are becoming widely available in the clinical environment, complementing the more common 1.5 T scanners. Evidence that 3 T SVMRS data can be used with the currently available automatic brain tumor classiers trained on 1.5 T, may allow researchers to extract shared knowledge from existing datasets and classifiers from both MR fields. Preliminar results with brain tumor classifiers trained with 1.5 T Short TE samples to discriminate among high grade malignant tumors and common grade II glial tumors obtained an accuracy of 0.85 evaluated with subsequently-acquired 3 T samples.

Evaluation Strategies During the evaluation of automatic classifiers, we try to estimate the performance they will achieve in new cases. When enough samples are available a hold-out of the dataset in a training subset and a test subset is the most accepted strategy. Moreover, prospective test is highly desired. Nevertheless, resampling techniques have to be applied when estimating classifier performance in small-sample datasets. Several resampling

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

techniques are used for reporting the evaluation results of BT classifiers. For example, in the cross validation evaluation, the data set is divided into k subsets, and a training-test evaluation is repeated k times. In kRandom Sampling Train-Test (kRSTT), k evaluations with independent partitions with the training and test set composed by a established percentage of cases of each class. When evaluating a single classifier, the error rate ERR (or the accuracy) is usually reported. The ROC curve is a graphical technique for assessing the performance of a binary discrimination. The AUC of the ROC curve measures the discrimination capability of the binary classifiers depending on the sensitivity and specificity when varying the threshold of the latent space obtained by the models. BER is used for evaluating single classifiers when the prevalences of the classes are quite different. In a binary classifier A vs. B, BER is the average of the error rate on the A and B classes.

Prospective Evaluation of BT Automatic Classification Using MR Spectra The combined action of the INTERPRET and eTUMOR projects from 2000 to 2009 allowed the multicenter prospective evaluation of the study of

15

Section Automatic Classification of Brain Tumour Based on MR Spectra in comparison to a resampling strategy. García-Gómez et al. (2009a) evaluated 253 pairwise classifiers for Glioblastoma, Meningioma, Metastasis, and Low-Grade Glial diagnosis were inferred based on 211 SV Short TE INTERPRET MR spectra obtained at 1.5 T (PRESS or STEAM, 20–32 ms) and automatically pre-processed. Afterwards, the classifiers were tested with 97 spectra, which were subsequently compiled during eTUMOUR. In Fig. 2.2, the BER estimation obtained by the Cross Validation (CV) strategy for the INTERPRET training dataset is compared with the prospective Independent Test (IT) consisting of the new eTUMOUR cases. There was observed a general trend in the (BER(CV) < 0.2, BER(IT) < 0.3) region, except for the GBM vs. Metastasis (MET) problem, indicated by the black-dashed line. This general trend shows the underestimation of the BER by the CV evaluation, possibly produced by the overfitting of the models on the training dataset and the estimation of the error with non-fully independent samples. The scattering of the GBM vs. MET results shows the randomness for this classification problem. The noteworthy recommendation of this result is that the use of multicenter prospective tests for final evaluation of predictive models is highly recommended to ensure the independence of the training and test sets.

IT vs CV (BER)

0.5 0.45 0.4 0.35

Fig. 2.2 Scatter plot of the performance measured in BER estimated by the IT set consisting of new eTUMOUR cases and the BER estimated by the CV using the INTERPRET cases. BER(IT) = BER(CV) is represented by the solid-blue line and the trend of the (BER(CV) <0.2, BER(IT) <0.3) region is indicated by the black-dashed line

BER(IT)

0.3 0.25 0.2 0.15

MEN vs. LGG MET vs. LGG MEN vs. MET GBM vs. MEN GBM vs. LGG GBM vs. MET

0.1 0.05 0

0

0.05

0.1

0.15

0.2

0.25 BER(CV)

0.3

0.35

0.4

0.45

0.5

16

J.M. García-Gómez

Interpretation and Discussion The main result obtained from a brain tumor classification study is a classification engine to predict the diagnosis of new cases. This engine can be prepared with a user-friendly graphical interface for be used in medical environment, such as a radiological service. Nevertheless, some other outcomes are direct consequence from the classification study: • Knowledge about relevant features extracted from MRS. The classification studies usually confirms the clinical and biochemical hypotheses about the discrimination power of metabolites and their complementariness. For example, García-Gómez et al. (2008) showed how ppm with contributions of the main metabolites and molecules reported in the MRS literature for brain tumor typing and grading were frequently selected in both Short TE and Long TE spectra. Namely, ML and Lac at about 1.3 ppm, alanine at about 1.5 ppm, Glx in the [2, 2.5] ppm range, total Cr at about 3 ppm,

Cho containing compounds at about 3.2 ppm, taurine at about 3.4 ppm, mI/Gly at about 3.55 ppm and Glx/Ala at about 3.76 ppm, were frequently selected by the SW algorithm in agreement with previous data from Kinoshita et al. (1994). Comparing the frequencies of selection in Long TE ans Short TE, it was seen that the selection in Long TE was focused in the well-known metabolites, but in Short TE the selection was sparser, confirming the contributions of mixtures of metabolites and molecules useful for classification along the region of interest. • Pattern discovery. Classification problems offer also the possibility to investigate on typical and atypical patterns in the MR spectra associated to brain tumors. Due to the possible tumor heterogeneity in the acquired voxel, acquisition artifacts or molecular tumor subtypes, the in-vivo MRS pattern may be heterogeneous within each diagnostic class. In the study presented by Garcia-Gomez et al. (2009b), two subpatterns were detected in AS3, four in GBM, two in MED, three in MET and six in the OTH class. As an example, the four subpatterns detected within 1.5

1

1

a.u.

a.u.

1.5

0.5

0.5

0 4

3.5

3

2.5

2

1.5

0 4

1

3.5

3

2

1.5

1

2.5 2 ppm

1.5

1

2.5 ppm

ppm 1.5

1

1

a.u.

a.u.

1.5

0.5

0.5

0

0 4

3.5

3

2.5 2 ppm

Fig. 2.3 Four subpatterns observed for GBM

1.5

1

4

3.5

3

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

the GBM tumors are presented in Fig. 2.3. The in-vivo MR spectral pattern subtypes may arised because of small differences in the voxel positioning with respect the tumor location. • Detection of outliers and quality control of biobanks. The analysis of misclassified cases during the evaluation of the classifiers may detect abnormal cases. These potential outliers may be due to incorrect annotation, abnormal MR profiles, possible problems in voxel positioning, acquisition artifact, normal-tissue contamination, problems with the preprocessing steps or lacks in the assumed classification model. An interesting discussion was presented by García-Gómez et al. (2008) about abnormal profiles of underrepresented tumor subtypes in both short TE and long TE. In this way, even in the absence of biopsy, PR techniques can contribute to the automatic validation of cases, assisting the specialists on the detection of potential source of errors in the biomedical data acquired from patients.

Computerized Decision Support Systems for Brain Tumor Diagnosis Using MRS A CDSS is an active knowledge system, which use two or more items of patient data to generate casespecifi advice. CDSSs based on PR have been widely accepted in medical applications due to their capability to optimization, flexibility, accuracy, and interpretability. During the 2000s decades the INTERPRET and eTUMOR projects allowed the development of a research background in CDSS applied to Brain Tumor classification using MR spectroscopy. As a result, four software packages are available based on this research. The SV INTERPRET GUI, developed from 2000 to 2003, provided easy access to a database of spectra, images and clinical information from 304 validated cases of human brain tumors. It was designed to allow the display of classification plots which are useful for automating the classification of tumor spectra. The Computer Aided Diagnosis System (CADS) of eTUMOUR, developed from 2004 to 2009, to which offers advanced functionalities for patient management, as well as aided diagnosis on several questions with capability to accept SV Long TE and Short TE spectra.

17

It is designed to provide the user with a percentage of confidence associated to the answer when supporting diagnosis. The classifiers are built from the most important repository of brain tumors in Europe. The HEALTHAGENTS distributed CDSS (dDSS) was developed from 2005 to 2008. This DSS differs from the previous in its architecture definition. It offers a web accessible DSS distributed along a network of clinical and producer nodes. Conceived as a collaborative network where users requests services to knowledge suppliers, this dDSS is able to address more than 10 different questions, sort the results attending to a ranking model and show latent space projections of the LDA classifiers it contains. CURIAM-BT is a CDSS, developed from 2004 to 2010, that can work with any kind of biomedical data. It is able to handle several predictive models which can deal with categorical and numerical data or signals. For routine cases, the results of the available models for the clinical question are assembled in a common view tailored for the task domain. Besides, an advanced view is also available for research purposes on specific interesting cases where detailed information like positioning latent space projection or comparison with the prototypes of the diagnosis are displayed based on the classification results. This CDSS also offers the feature of incorporating new classifiers to the system just by simple ’drag & drop’. This functionality enables both, to improve the available classifiers and to incorporate new ones to give orientation about new clinical questions.

Conclusions The traslational research in the automatic brain tumor classification using MRS has achieved a degree of maturity to offer practical results for the clinical practice. The Machine Learning discipline has shown its suitability for the design of classifiers from brain tumor diagnosis with MRS. A multicenter dataset with a agreed acquisition protocol is essential to achieve a generalizable solution. Besides, a pre-processing pipeline may allow the extraction of relevant feature from the MR spectra. Feature extraction methods based on pattern recognition have demonstrated an optimal usefulness-reproducibility tradeoff for automatic classification of brain tumors.

18

The classification of the most prevalent types of brain tumors has been largely studied by several authors. Recent studies have also included classifiers for the childhood and a wider range of classes in their classifications, following more advanced methodologies, such as one-class classification and FDA. Furthermore, incremental learning is conceived as a solution for the dynamism of the clinical environments These advances, joined to the compatibility of the classification models based 1.5 T samples with 3 T samples, allow the use of the current multicenter datasets and classification models for research and clinical practices beyond the technological advances. A prospective evaluation of the classifiers is highly recommended, but most of the studies reported in the literature use resampling techniques due to the limitation of the available datasets. The results of the automatic classification study may be of interest to analyze the contribution of metabolites in the brain tumor types, to discover subpatterns in heterogeneous tumor types and to detect abnormal cases. At the moment, several CDSSs are available for traslational research; moreover, some of them (e.g. CURIAM-BT) are under evaluation for clinical purposes. The agreement of acquisition protocols and further studies in classification, prediction of the outcome, treatment planning, monitoring of theraphy and posttherapy evaluation, as Horska and Barker pointed out, may lead MRS to a routine clinical tool. Acknowledgements The author would like to thank the IBIME research group for contributing to results included in this chapter, particularly, Javier Vicente, Salvador Tortajada, Elies Fuster, Carlos Saez, and Montserrat Robles. The author would like to thank INTERPRET and eTUMOUR partners for providing data, particularly, Carles Majós (IDI-Bellvitge), John Griffiths and Franklyn Howe (SGUL), Arend Heerschap (RU), Witold Gajewicz (MUL), Jorge Calvar (FLENI), and Antoni Capdevila (H. de Sant Jaon de Dèu).

References Davies N, Wilson M, Harris L, Natarajan K, Lateef S, Macpherson L, Sgouros S, Grundy R, Arvanitis T, Peet A (2008) Identification and characterisation of childhood cerebellar tumours by in vivo proton MRS. NMR Biomed 21: 908–918

J.M. García-Gómez Devos A, Lukas L, Suykens JAK, Vanhamme L, Tate AR, Howe FA, Majos C, Moreno-Torres A, van der Graaf M, Arús C, Van Huffel S (2004) Classification of brain tumours using short echo time 1H MR spectra. J Magn Reson 170:164–175 García-Gómez J, Epifanio I, Julia-Sapè M, Monleón D, Vicente J, Tortajada S, Fuster E, Moreno-Torres A, Peet A, Howe F, Celda B, Arús C, Robles M (2009b) Possibilistic classification of Brain Tumors by MRS based on Functional Data Analysis and Subpattern Discovery. In Proceedings 17th Scientific Meeting, International Society for Magnetic Resonance in Medicine, Honolulu, 146 García-Gómez JM, Luts J, Julia-Sapè M, Krooshof P, Tortajada S, Robledo JV, Melssen W, Fuster-García E, Olier I, Postma G, Monleón D, Moreno-Torres A, Pujol J, Candiota AP, Martínez-Bisbal MC, Suykens J, Buydens L, Celda B, Van Huffel S, Arús C, Robles M (2009a) Multiprojectmulticenter evaluation of automatic brain tumor classification by magnetic resonance spectroscopy. Magn Reson Mater Phys 22:5–18 García-Gómez JM, Tortajada S, Vidal C, Julia-Sape M, Luts J, Moreno-Torres A, Van Huffel S, Arús C, Robles M (2008) The effect of combining two echo times in automatic brain tumor classification by MRS. NMR Biomed 21:1112–1125 Garczarek UM (2002) Classification rules in standardized partition spaces. University of Classification rules in standardized partition spaces. University of Dortmund, Dortmund Horska A, Barker PB (2010) Imaging of brain tumors: MR spectroscopy and metabolic Imaging. Neuroimaging Clin N Am 20:293–310 Kinoshita Y, Kajiwara H, Yokota A, Koga Y (1994) Proton magnetic resonance spectroscopy of brain tumors: an in vitro study. Neurosurgery 35:606–613 Lukas L, Devos A, Suykens JAK, Vanhamme L, Howe FA, Majós C, Moreno-Torres A, Graaf MVD, Tate AR, Arús C, Huffel SV (2004) Brain tumor classification based on long echo proton MRS signals. Artif Intell Med 31:73–89 Luts J, Laudadio T, Idema AJ, Simonetti AW, Heerschap A, Vandermeulen D, Suykens JAK, Van Huffel S (2009) Nosologic imaging of the brain: segmentation and classification using MRI and MRSI. NMR Biomed 22:374–390 Luts J, Poullet J-B, Garcia-Gomez JM, Heerschap A, Robles M, Suykens JAK, Van Huffel S (2008) Effect of feature extraction for brain tumor classification based on short echo time 1H MR spectra. Mag Reson Med 60:288–298 Majos C, Julia-Sape M, Alonso J, Serrallonga M, Aguilera C, Acebes JJ, Arús C, Gili J (2004) Brain tumor classification by proton MR spectroscopy: comparison of diagnostic accuracy at short and long TE. Am J Neuroradiol 25:1696–1704 Melssen W, Wehrens R, Buydens L (2006) Supervised Kohonen networks for classification problems. Chemometr Intell Lab 83:99–113 Menze BH, Lichy MP, Bachert P, Kelm BM, Schlemmer H-P, Hamprecht FA (2006) Optimal classification of long echo time in vivo magnetic resonance spectra in the detection of recurrent brain tumors. NMR Biomed 19:599–609 Opstad KS, Ladroue C, Bell BA, Griffiths JR, Howe FA (2007) Linear discriminant analysis of brain tumour 1H MR spectra: a comparison of classification using whole spectra versus metabolite quantification. NMR Biomed 20:763–770 Preul MC, Caramanos Z, Collins DL, Villemure JG, Leblanc R, Olivier A, Pokrupa R, Arnold DL (1996) Accurate,

2

Brain Tumor Classification Using Magnetic Resonance Spectroscopy

noninvasive diagnosis of human brain tumors by using proton magnetic resonance spectroscopy. Nat Med 2:323–325 Ramsay J, Silverman B (2002) Applied functional data analysis: methods and case studies, Springer series in statistics (Berlin). Springer, Heidelberg, NY Simonetti AW, Melssen WJ, Szabo de Edelenyi F, van Asten JJA, Heerschap A, Buydens LMC (2005) Combination of feature-reduced MR spectroscopic and MR imaging data for improved brain tumor classification. NMR Biomed 18: 34–43 Tate AR, Majos C, Moreno A, Howe FA, Griffiths JR, Arús C (2003) Automated classification of short echo time in in vivo 1H brain tumor spectra: a multicenter study. Magnet Reson Med 49:29–36

19

Tortajada S, García-Gómez JM, Vicente J, Robles M (2008) Dynamic learning of brain tumour classifiers based on 1HMRS. In ESMRMB 2008: 25th Annual Scientific Meeting, Valencia, Spain, 14–15 Vicente J, García-Gómez J, Tortajada S, Fuster-Garcia E, Capdevila A, Peet A, Celda B, Robles M (2009) AgeFiltered MRS Classifier to Overcome The Differences in Childhood and Adulthood Brain Tumours. In Proceedings 17th Scientific Meeting, International Society for Magnetic Resonance in Medicine, Honolulu, 981. Weis J, Ring P, Olofsson T, Ortiz-Nieto F, Wikström J (2010) Short echo time MR spectroscopy of brain tumors: grading of cerebral gliomas by correlation analysis of normalized spectral amplitudes. JMRI-J Magn Reson Imaging 31: 39–45

Chapter 3

Cellular Immortality in Brain Tumors: An Overview Ruman Rahman and Richard G. Grundy

Abstract Brain tumors are a diverse group of neoplasms that continue to present a formidable challenge in our attempt to achieve cures and a reduction in morbidity. Our conceptual framework of human brain cancer has been redrawn in the current decade. There is a growing acceptance that brain tumour formation is a phenotypic outcome of dysregulated neurogenesis, with tumors viewed as abnormally differentiated neural tissue. In relation, there is accumulating evidence that brain tumors, similar to leukaemia and many solid tumors, are organized as a developmental hierarchy which is maintained by a small fraction of cells endowed with many shared properties of tissue stem cells. Proof that neurogenesis persists throughout adult life, compliments this concept. Although the cancer cell of origin is unclear, the proliferative zones that harbour stem cells in the embryonic, post-natal and adult brain are attractive candidates within which tumour-initiation may ensue. Dysregulated, unlimited proliferation and an ability to bypass senescence are acquired capabilities of cancerous cells. These abilities in part require the establishment of a telomere maintenance mechanism for counteracting the shortening of chromosomal termini. A strategy based upon the synthesis of telomeric repeat sequences by the ribonucleoprotein telomerase, is prevalent in ∼90% of human tumors studied, including the majority of brain tumors. This review will provide a developmental perspective with respect to normal (neurogenesis) and aberrant

R. Rahman () Children’s Brain Tumor Research Center, Medical School D Floor, School of Clinical Sciences, Queen’s Medical Centre, Nottingham, NG7 2UH, UK e-mail: [email protected]

(tumourigenesis) cellular turnover, differentiation and function. Within this context our current knowledge of brain tumour telomere/telomerase biology will be discussed with respect to both its developmental and therapeutic relevance to the hierarchical model of brain tumourigenesis presented by the cancer stem cell paradigm. Keywords Neoplasms · Telomere · Telomerase · Neurogenesis · Stem cells · Tumor

Introduction Characteristics of Intracranial Neoplasms Of all solid cancers, brain tumors have the poorest survival and highest morbidity rates. Although frequently considered collectively, brain tumors represent a diverse range of tumour types affecting both children and adults. Tumors of the brain account for less than 2% of all malignancies in adult, yet now approach leukaemia as the leading cause of cancer-related death in children (30% of cancers) and the fourth leading cause in adults. The diversity of brain tumour type is a reflection of the histological complexity of the central nervous system (CNS) and can present with glial or neuronal phenotypes or a mixture of cell types. Medulloblastoma and supratentorial primitive neuroectodermal tumors (sPNET) are World Health Organisation (WHO) grade IV embryonal tumors, resulting from the transformation of primitive neuroectodermal cells. The medulloblastoma is a malignant and invasive tumour with a relatively poor prognosis and is the most common malignant brain tumour in

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_3, © Springer Science+Business Media B.V. 2011

21

22

R. Rahman and R.G. Grundy

children. It is predominately neuronal in nature and typically located in the cerebellum. In contrast the rarer CNS PNETs are more commonly found in the cerebral hemispheres and consist of poorly differentiated neuroepithelial cells which confer poor prognosis, especially in children. Despite both being classified as Grade IV tumors, CNS PNETs have a significantly worse prognosis (Packer, 2008). Also classified as a grade IV tumour by the WHO, glioblastoma multiforme (GBM) represents the most malignant and aggressive form of glioma, arising in the subcortical white matter as either a primary neoplasm, or from the malignant progression of a low grade glioma (Behin et al., 2003); this subject has been discussed in detail in volume 1 of this series of Handbooks. Typically seen in adults, prognosis is extremely poor as the diffuse infiltration makes surgical resection difficult. Ependymoma are typically classified as glial tumors, however it is currently debated as to whether ependymoma may actually reflect a separate entity. The cell of origin is uncertain; however current evidence supports a radial glial for ependymoma (Taylor et al., 2005). Composed of neoplastic ependymal cells, the ependymoma is usually located along the ventricular system of the paediatric or adolescent brain whereas spinal sites are more commonly seen in adults. Prognosis is varied but at relapse is dismal. Although an understanding of human brain cancer pathogenesis has increased in recent decades, insufficient knowledge of the molecular, genetic and cellular alterations and mechanisms is at the root of treatment failure and rudimentary capacity to determine clear prognostic markers and treatment stratification. As a result, bio-centred and target-specific brain tumour therapy has not yet been forthcoming.

in vitro. This operational definition does not however, necessarily reflect the function of NSCs in vivo and it remains uncertain whether the cell type(s) exhibiting “stemness” in vitro faithfully represents the cell exhibiting “stemness” in vivo. Mammalian developmental neuro-biology now entertains an integrated view of NSCs whereby neuroepithelial cells comprising the neural tube of the developing embryo, give rise first to radial glia, which then transform into astrocyte-like adult multipotent stem cells or directly into mature ependymal cells lining the walls of the lateral ventricles, the latter of which are post-mitotic in the adult brain. Radial glia arise from neuroepithelial cells throughout the primitive CNS at the beginning of neurogenesis and represent the predominant neuronal progenitor (Mori et al., 2005). These cells characteristically display a radial morphology and a mixed primitive cell/glial immunophenotype. In mammals, radial glial cells disappear from the brain soon after birth, with a closelyrelated astrocyte-like cell taking the role as NSC thereafter. The belief that the genesis of new CNS cells is a rare event in the adult mammalian brain was established dogma until recently. Two constitutively active germinative layers are now known to reside in the adult brain: the subventricular zone (SVZ) associated with the anterior part of the forebrain lateral ventricles and the subgranular zone (SGZ) within the hippocampus, corresponding to the inner layer of the dentate gyrus. New neurons and glia continue to be generated within these restricted regions of the adult mammalian brain (Zhao and Overstreet-Wadiche, 2008). Evidence that neurogenesis persists in the CNS of the adult mammalian brain (albeit in restricted domains) has changed our perception of the cell of origin with respect to neural neoplastic transformation.

Neuro-Developmental Biology

Telomere and Telomerase Status During Normal and Dysregulated Neurogenesis

The CNS is a highly complex structure which develops from a restricted number of extensively “plastic” cells that proliferate and acquire regional identities in space and time to produce a diverse repertoire of cell types. Collectively, these cells are defined as neural stem cells (NSCs) on the basis of their potential to differentiate into mature neurons and glia (astrocytes and oligodendrocytes) and their ability to self-renew

In virtually all types of malignant cells, unlimited replicative potential is conferred in considerable part, by the maintenance of telomere length above a critical minimum threshold. Telomeres (derived from the Greek telos, meaning end and meros, a component) are a complex of guanine-rich repeat sequences and associated proteins that form a loop structure which

3

Cellular Immortality in Brain Tumors: An Overview

caps all eukaryotic chromosome termini and prevents chromosomes from improper recombination, nuclease degradation and end fusions, thereby contributing to genomic stability. Due to incomplete replication of lagging-strand synthesis at each mitotic division (“end-replication problem”) (Hayflick and Moorhead, 1961), telomeres progressively shorten, a phenomenon that provides a molecular basis of cellular ageing and hints at a primitive tumour-suppressor mechanism. In germ cells and cancer cells, unlimited proliferative capacity is achieved through the addition of telomeric repeats by the telomerase holoenzyme. Telomerase is a ribonucleoprotein consisting of an RNA subunit (TR) which provides an intrinsic template for the synthesis of de novo telomeric repeats in a reaction catalyzed by its enzymatic component (TERT) (Greider and Blackburn, 1989). NSCs from different regions of the brain have different proliferative capacities and varying abilities to generate astrocytes, oligodendrocytes and neurons. This may be reflected in the level of telomerase activity and telomere length reserve within cells from different neuro-anatomical compartments, in particular regions of high neurogenic activity such as the SVZ and SGZ. Telomerase-mediated telomere homeostasis is therefore a likely key factor in brain plasticity throughout development and adulthood. There is a suggestion here that a fuller understanding of the replicative histories of cells within tumour hierarchical compartments may provide clues as to the cancer cell of origin for a given brain neoplasm. Indeed telomerase activity has been detected in a growing number of brain tumors in the past decade. With pre-clinical and clinical advances in telomerase-therapeutics for cancer, there is anticipation that this will impact upon paediatric and adult tumors of the CNS.

Brain Tumourigenesis and Tumour Heterogeneity A developmental relationship between tissue stem cells and cancer cells has been proposed over 50 years ago. However the pioneering demonstration of “cancer stem cells” in human cancer was first made in leukaemia (Bonnet and Dick, 1997; Lapidot et al., 1994). From this initial discovery in hematopoietic

23

tumors, evidence from most solid cancers studied defines restricted subsets of tumour cells that share characteristics of the corresponding normal tissue stem cells. This paradigm includes prospective isolation of a minority population of brain tumour cells, based on the expression of the cell surface antigen CD133, known to be highly expressed in normal neural stem cells. These putative brain cancer stem cells were isolated from medulloblastoma, pilocytic astrocytoma, glioblastoma multiforme and anaplastic ependymoma and possessed a marked capacity for self-renewal (evidenced by the formation of floating aggregates termed tumorspheres), proliferation and differentiation (Galli et al., 2004; Hemmati et al., 2003; Singh et al., 2003; Taylor et al., 2005). Crucially, only the CD133+ fraction was capable of tumour initiation upon orthotopic transplantation into non-obese diabetic severe combined immunodeficient (NOD-SCID) mouse brain (Singh et al., 2004). The CD133+ fraction produced a tumour that generated a phenocopy of the original tumour and could be serially transplanted into secondary recipients, providing evidence of self-renewal capacity in vivo. Our growing understanding of the hierarchical organization of the functional compartments in renewing brain tissues provides a theoretical and technical framework upon which to base falsifiable questions regarding brain tumour initiation and cell of origin.

A Multidisciplinary Approach to a New Synthesis Recent compelling evidence demonstrates that the brain, similar to other organs in which cancers arise, harbours a stem cell population that can continue to give rise to differentiated CNS cells. Prospective identification of neural cells with shared stem cell-like properties which are capable of initiating tumours in vivo and that are phenocopies of the original tumour, provides a new synthesis in brain tumour research. The cellular architecture of brain tumors may thus represent a distorted mimicry of normal developmental neuro-biology. This concept, coupled with an increasing understanding of the immortal status of transformed neural cells, particularly with respect to aberrant telomerase activity and the maintenance of critically short telomeres, will arguably provide an

24

experimental basis to better understand and delineate cancer changes that are causal and consequential with respect to tumour hierarchy and tumour initiation. This chapter encourages a consilient approach to investigate telomere/telomerase neurobiology and brain tumour stem cells within a neurodevelopmental paradigm.

R. Rahman and R.G. Grundy

question marks, Fig. 3.1). This concept of early neurogenesis is in dramatic contrast from the long-held view that embryonic, post-natal and adult NSCs are discrete unrelated populations.

Multipotent Neural Stem Cells and Neurogenesis in the Mature CNS Developmental Neurobiology Neuroepithelial Cells, Radial Glia and Early Embryogenesis The origin of all CNS cells can be traced to a single layer of embryological cells, which constitute the neuroepithelium. The edges of this sheet fold together to form the neural tube, which later gives rise to the ventricular system and spinal canal. Initially, neuroepithelial cells divide symmetrically in order to increase the stem cell pool; thereafter these highly plastic cells proliferate and acquire regional identities in a spatial-temporal manner. It is likely that neuroepithelial cells differentiate directly into radial glial cells in the embryonic ventricular zone (Noctor et al., 2007), although this has not been confirmed experimentally. Some recent supporting evidence is presented by the derivation of both neuroepithelial cells and radial glia from human embryonic stem cells; the course of differentiation (with respect to morphology change and lineage-specific marker expression) followed a temporal manner with neuroepithelial cells generated first, followed later by radial glial cells and finally mature neurons. The fate and function of a radial glial cell varies considerably from region to region in the CNS. It had long been accepted that radial glia are both astroglial progenitors and oligodendrocyte progenitors, but recent evidence has demonstrated that radial glia produce most of the cortical neurons (Hartfuss et al., 2001; Noctor et al., 2001; Wu et al., 2005). Radial glial cells thus serve as neural stem/progenitor cells for the majority of neurogenesis and gliogenesis in the developing brain and gives rise to the astrocyticlike neural stem cells of the adult brain. Whether radial glia transform directly into an adult multipotent neural stem cell or progenitor, or whether they first differentiate into an intermediate cell type, is unknown (see

Although widespread neurogenesis is restricted to the embryonic period, a limited number of neural stem cells persist throughout adulthood and are restricted mainly to two neurogenetic domains (Bonfanti and Peretto, 2007; Gage, 2000; Gross, 2000). The largest of these regions is the mammalian SVZ situated throughout the lateral walls of the forebrain lateral ventricles. A similar hierarchical neurogenetic system exists in the SGZ, an area corresponding to the dentate gyrus of the hippocampal granular layer. Stem cell progeny generated within these regions populate the growing CNS parenchyma, later differentiating into mature neuronal and glial fates, specified in a precise spatialtemporal manner. Cells that express the intermediate filament astroglial marker, glial fibrillary acidic protein (GFAP), have been identified as putative adult neural stem cells in this region, also known as Type B cells (Fig. 3.1). The embryonic origin of Type B cells has been confirmed using in vivo lineage tracing techniques and shows that adult subventricular cells are not related to embryonic subventricular precursors, but embryonic ventricular radial glia. Further support of the astrocytic nature of NSCs has been confirmed in experiments in which GFAP+ cells were conditionally ablated in the brain of adult mouse, resulting in near complete loss of neurogenesis. Type B cells are proposed to be relatively quiescent with a cell cycle duration of ∼24 h and constitute a small fraction of the total astrocytic population in the SVZ. Pioneering experiments retrospectively viewed, adds support to this notion; Reynolds and Weiss in 1992, isolated a small population of cells (<0.1% of total cells) from the adult striatum that could proliferate and generate multiple clones of cells in free-floating clusters in vitro called neurospheres (Reynolds and Weiss, 1992). It is worth noting that even with current refined methods to isolate neural stem cells using the neurosphere assay, the stem cell content is variable; although Type B and Type C cells can produce neurospheres,

3

Cellular Immortality in Brain Tumors: An Overview

25

Fig. 3.1 Stem cell theory for the role of telomeres and telomerase in postnatal neurogenesis and cancer. Radial glial cells (RG, green box) differentiate from embryonic neuroepithelial progenitors at the beginning of neurogenesis and are distinguished by the expression of astroglial lineage markers in the neonatal brain. Within the constitutively active germinative layers associated with the anterior part of the forebrain lateral ventricles (SVZ, shown here) and corresponding to the inner layer of the dentate gyrus within the hippocampus (SGZ, not shown), radial glia give rise to adult multipotent neural stem cells (Type B astrocytes, dark blue circle). At later developmental stages, radial glial cell division can also produce ependymal cells that constitute the epithelia lining the ventricular system of the brain and spinal cord (light blue circle). Type B neural stem cells retain markers of radial glia, have high levels of telomerase activity and may have longer telomeres relative to mature neurons or glia. During postnatal neurogenesis, neural stem cells differentiate and give rise to fast-cycling early transit amplifying progenitors (Type C putative precursors, Orange circle), which can give rise to terminally differentiated mature glia (astrocytes, oligodendrocytes) or late transit amplifying progenitors (Type A migrating neuroblasts, Yellow star) that are committed to a mature neuronal fate and that express neuronal-specific markers.

As multipotent neural stem cells differentiate into mature functional neurons and glia, the capability to leave the stem cell perivascular niche and regenerate neural tissue is decreased. This is due in considerable part to telomere shortening across time. Decreased mobilization of stem cells in aged neural tissue is likely to be consequential of senescence (replicative and stress-induced) and apoptosis. There is increasing awareness that the growth and recurrence of brain tumors may be due to a transformed CNS cell type with shared characteristics to a somatic neural stem cell (self-renewal, differentiation) and endowed with tumour-initiating ability. These “brain tumour stem cells” may arise in vivo from a transformed neural stem cell or from an early neural progenitor that acquires the ability to self-renew through mutation. Abbreviations: SC, stem cell; ET, early transit amplifying progenitor; LT, late transit amplifying progenitor; EP, ependymal cells; SVZ, subventricular zone; SGZ, subgranular zone; GFAP, glial fabrillary acidic protein; PSA-NCAM, polysialylated neural cell adhesion molecule; MAP2, microtubule-associated protein 2; O4, monoclonal antibody that recognizes the sulfatides; MBP, myelin basic protein. Blue rectangles represent stem cell/progenitor populations; Red rectangles represent terminally differentiated cell types

not all cells within the neurosphere express stem cell markers. Multipotent Type B astrocytes within the SVZ give rise to fast-cycling transiently amplifying precursor cells called Type C precursors, which in turn generate either mitotically active Type A neuroblasts or terminally differentiated mature glia (astrocytes, oligodendrocytes). Type A neuroblasts migrate towards the olfactory bulb where they integrate as mature neurons

(Fig. 3.1). Similarly, Type B astrocytes in the SGZ produce intermediate Type D precursors which give rise to Type G granule neurons. Over the course of development, NSCs not only feature altered morphology but also exhibit changes to gene expression profiles in a temporal-specific manner. This expression profile may underlie an intrinsic developmental program as progenitors grown in culture generate progeny at certain times, similar to progenitors in vivo. In addition to

26

GFAP, Type B cells express other intermediate filaments such as vimentin and nestin. As NSCs differentiate along developmental pathways, neuronal-specific (polysialylated neural cell adhesion molecule (PSANCAM), microtubule-associated protein 2 (MAP2), neurofilament protein, tyrosine hydroxylase) and glialspecific markers (O4 monoclonal antibody that recognizes the sulfatides and myelin basic protein (MBP)) are expressed in late-amplifying and mature cell compartments (Fig. 3.1).

The Role of Telomeres and Telomerase During Neurogenesis Most studies aimed at understanding the nature of NSCs with respect to self-renewal, proliferation and differentiation, focus on external cues such as secreted growth factors. Few studies have addressed the role of intrinsic mechanisms regulating NSC behaviour. Chromosome integrity is essential for cell viability, therefore highly proliferative cell types such as tissue stem cells, require active telomere maintenance strategies to proliferate indefinitely or at least long enough for the host organism to reach reproductive age. The role of telomere and telomerase dynamics in NSC compartments has scarcely been addressed. In the mouse, high levels of TERT and TERC (murine telomerase RNA subunit) mRNA are present in the developing neural tube as early as E10.5. Both telomerase subunits continue to be expressed in different regions of the developing murine CNS and correlates with the proliferation of neural progenitors (Klapper et al., 2001). This suggests a role for telomerase during specification of neural cell types. Consistent with our knowledge of neurogenetic regions of the adult brain, telomerase activity has been shown in neural precursor cells isolated from the SVZ and hippocampus in adult mice. With reference to stem cell populations in other tissues, the current working hypothesis states that telomerase activity in NSCs (and actively proliferating early neural precursors) may combat induced telomere shortening upon mitotic division. However, human progenitor cells isolated from the developing cortex which exhibit low/absent levels of telomerase activity, undergo decreased neurogenesis and eventually enter replicative senescence in vitro (Wright et al., 2006). These findings are corroborated by studies in other

R. Rahman and R.G. Grundy

tissues where low levels of telomerase activity have been found in tissue stem cells such as haematopoietic, skin, intestinal crypt and pancreas. In all these cases, telomerase activity is insufficient to completely prevent telomere loss and senescence. Rather, telomerase in tissue stem cells may slow the rate of telomere shortening and maintain a replicative reserve predetermined for the tissue stem cell in question. It is difficult to argue that this phenomenon is not a primitive tumour-suppressor mechanism and perhaps one that predates protein-mediated tumour suppression. Telomere attrition dramatically impairs the in vitro proliferation of adult NSCs isolated from the SVZ of telomerase-deficient mice, but not that of embryonic NSCs, even though the latter exhibited critically shortened telomeres. This is consistent with a recent report describing telomerase levels as high in embryonic cortical neural progenitor cells, but low in newly generated neurons and mature neurons (Cheng et al., 2007). Collectively these results hint at intrinsic differences in the protective states between adult and embryonic NSCs, with respect to chromosomal stability and response to dysfunctional telomeres. A comprehensive understanding of how the regulation of telomeres/telomerase is tightly linked to cell cycle regulation in tissue stem cells will be crucial to determining how cancer may arise when stem cell regulation is perturbed.

Cellular Immortality and Tumour Initiation Telomeres and Telomerase in Brain Tumors Tumourigenesis involves multiple oncogenic insults which collectively define the tumour phenotype. Experimental data from cultured human fibroblasts indicate that cells must bypass two checkpoints, mortality stage 1 (M1), regarded as cellular senescence and mortality stage 2 (M2), characterised by widespread apoptosis, prior to immortalization and unlimited replicative potential. Reactivation hTERT is a crucial determinant of cellular immortalization and if accompanied by inactivation of tumoursuppressor genes and/or activation of cellular oncogenes, can subsequently result in neoplastic formation.

3

Cellular Immortality in Brain Tumors: An Overview

Telomerase-mediated telomere maintenance is evident in virtually all types of malignant cells, and ∼90% of tumors show evidence of up-regulated telomerase (Shay and Bacchetti, 1997). Brain tumors in which telomerase activity has been detected include glioblastoma, anaplastic astrocytoma, oligodendroglioma, primitive neuro-ectodermal tumors, medulloblastoma and ependymoma. These data indicate that hTERT reactivation or overexpression is an essential event for tumour progression in CNS tumors. The role of hTERT may not be restricted to addition of telomere repeats as mouse TERT has been shown to promote cell survival (prevents apoptosis) in neurons, a function that does not require the telomerase RNA subunit. This “extracurricular” TERT function may contribute to cancer development and aging independently of telomere lengthening and telomerase activity. In a comprehensive analysis of telomere length, the majority of grade II astrocytomas, anaplastic astrocytomas and glioblastomas with telomerase activity, exhibited significantly shorter mean telomere lengths relative to corresponding tumors with undetectable telomerase activity and to normal brain tissue. Similarly, all primitive neuroectodermal tumors exhibited shorter telomere lengths relative to normal brain tissue. Recent findings from our own laboratory reveal that paediatric glioblastoma multiforme and ependymoma, tumors of glial origin, have significantly longer mean telomere length than those of supratentorial neuroectodermal tumors and medulloblastoma, tumors of neuroectodermal origin. This highlights the consideration required for telomere length with respect to duration/course of treatment when using anti-telomerase drugs which are predicted to exert their effects upon telomere shortening (Rahman R et al. 2010). The concurrent shortening of telomeres with telomerase reactivation/overexpression supports a hypothesis that telomerase is required to maintain both minimum length and a terminal capping structure at a subset of critically shortened telomeres. The perspective presented by the stem cell origin of brain tumors implies that the genetic alterations that lead to cancer, accumulate in NSCs rather than mature neural cells. As NSCs would already exhibit telomerase activity, one would predict the enzyme to be overexpressed rather than reactivated. Whether telomerase activity increases during later stages of carcinogenesis is unclear. Similar to self-renewing tissues where hTERT expression is insufficient to halt progressive

27

telomere shortening, telomere shortening may continue within a proliferating transformed neural stem cell. However aberrant and sustained hTERT expression likely results in telomere stabilization both with respect to telomere length and 3 telomere capping structure (Fig. 3.2). Elucidation of telomerase levels and telomere length within different neurogenetic subsets will influence strategies to target telomerase in brain cancer.

The Brain Tumour Stem Cell Paradigm An acceptance of tumour hierarchical organisation readily leads to a developmental-centred view of tumourigenesis, where brain tumors represent aberrantly differentiated tissue, in turn reflecting dysregulated neurogenesis. The stem cell paradigm for brain tumors presents tumour-initiating events as occurring within the genome of a cellular entity with neural stem cell-like properties. The identification of brain tumour stem cells was prospectively demonstrated by Peter Dirks and colleagues using cells derived from glioblastoma and medulloblastoma. As few as 100 cells sorted positively for the cell surface antigen CD133, were capable of initiating infiltrative tumors upon orthotopic transplantation into SCID mice whereas injection of 100,000 CD133- cells failed to initiate tumors (Singh et al., 2004). In a later study, tumour stem cells isolated from ependymoma manifested a radial glial phenotype with CD133+/Nestin+/RC2+/BLBP+ positivity confirmed to be tumourigenic (Taylor et al., 2005). It is important to note that a definitive demonstration of tumour stem cells in the brain is lacking as CD133 enriches for, rather than definitively identifies, cancer stem cells. This caveat has been emphasized with recent studies showing that CD133glioblastoma cells have tumour-initiating ability (Beier et al., 2007; Wang et al., 2008). In addition, it is the CD133+/Nestin+/RC2-/BLBP- phenotype that exhibited a tumourigenic phenotype in medulloblastoma in contrast to the CD133+/Nestin+/RC2+/BLBP+ phenotype in ependymoma (Taylor et al., 2005), suggesting the characteristics of these two subsets may reflect different cell of origins. Clearly stem-like properties in a tumour cell cannot be taken as direct evidence for neoplastic transformation occurring in a neural stem

28

R. Rahman and R.G. Grundy

Fig. 3.2 Proposed role for telomerase activity and telomere stabilization in brain tumour initiation and propagation. (Normal developmental neurogenesis) The postnatal brain harbours regions that retain mitotic activity throughout adulthood, in particular the subventricular zone (SVZ) and subgranular zone (SGZ). Mature neurons and glia are generated in these regions through cell cycle re-entry and differentiation of tissue-specific stem cells, giving rise to neural lineage-restricted progenitors. Faithful cell division is safeguarded in neural progenitors and mature cell types through intact functional cell cycle and senescence checkpoints. (Dysregulated developmental neurogenesis). hTERT overexpression, loss of tumour suppressors, oncogenic insults, checkpoint failure and chromosome instability, increase the tumourigenic potential of the tissue stem cell, ultimately

resulting in a transformed neural stem cell (“brain tumour stem cell”). It is unclear whether mutational events at tumourinitiation are necessary and sufficient to propagate differentiation to a brain tumour progenitor cell and “mature” malignant neural cell, or whether additional mutational events drive progression of the tumour. As the transformed neural stem cell (aberrantly) differentiates, telomere shortening is likely to occur; however due to hTERT acting on critically shortened telomeres, telomere length in the (mature) malignant neural cell would be expected to be stabilized. Although depicted here as arising in a neural stem cell, it remains plausible that the tumour cell of origin is a neural progenitor; in the latter scenario, the progenitor cell may acquire the capacity to self-renew through mutation, increasing the propensity to initiate a tumour

cell that originates in a zone of active neurogenesis. The cancer stem cell hypothesis and the cellular origin of cancer are distinct paradigms. It is conceivable that a neural progenitor or mature cell could give the appearance of “stemness” upon neoplastic transformation (Fig. 3.2). An alternative focus is that of molecular events that permit tumour initiation rather than on the cell type that initiated the tumour. The cancer stem cell hypothesis predicts mechanistic similarities between self-renewal of cancer stem cells and normal stem cells. In brain tumors there is particular interest in the Notch and Sonic-hedgehog pathways, as these have been shown to be important for normal stem cell self-renewal as well as dysregulated growth in

cancer. Overexpression of Notch receptors and their ligands Delta-like 1 (DLK1) and Jagged 1 (JAG1) correlates with the proliferative capacity of human glioma cells, implicating Notch as a positive effector of self-renewal in adult neural stem cells. Furthermore, overexpression of Notch-2 intracellular domain protein increases proliferation of medulloblastoma cell lines, whereas medulloblastoma growth is suppressed by the known Notch inhibitor, γ-secretase (Hallahan et al., 2004). Similarly Sonic Hedgehog (SHH) and its downstream effectors, glioma-associated homologue 1 (GLI1), GLI2 and GLI3 have been shown to specifically regulate self-renewal and neurogenesis within the external granular layer of the post-natal cerebellum and to control proliferation of precursors within

3

Cellular Immortality in Brain Tumors: An Overview

the adult SVZ (Palma et al., 2005). Mutations in this pathway are thus implicated in the initiation of medulloblastoma. Another emerging concept is that brain tumour stem cells may crucially be supported by a specific microenvironment which mimics the neural vascular niche. Differentiated neural cells or blood vessel cells have been implicated as niche elements. CD133+ glioblastoma cells secrete increased levels of vascular endothelial growth factor (VEGF) compared to CD133- counterparts, implicating tumour stem cells as contributing to tumour angiogenesis (Bao et al., 2006). However, whether the tumour stem cell niche differs from that of the bulk tumour population or whether it provides signals which govern aberrant self-renewal is unclear. Future work will validate the proposition that the tumour stem cell niche represents an attractive therapeutic target. The cancer stem cell field is at present contentious, in part due to differing semantic interpretations. The “cancer stem cell” term is often applied to cells identified by varying methods and criteria. The danger here is the application of the term to cells with no evidence of cancer-initiating capacity. It is important to clarify what “brain tumour stem cell” refers to in the present context: these are cells which demonstrate cancer-reinitiating ability upon orthotopic transplantation, with a capacity to generate a phenocopy of the original tumour mass (consisting both tumourigenic and non-tumourigenic cells); show extensive selfrenewal ability ex vivo and in vivo; demonstrate aberrant multipotent differentiation; harbour karyotypic and/or genetic alterations. At present, this does not exclude the possibility of more committed progenitors or even mature cells reverting to a stem-like phenotype and causing tumour-initiation in contexts that differ from those imposed by the pre-mentioned strategies. For example, it may be argued that the mouse may not represent a physiologically relevant microenvironment for engraftment and growth of human tumors, thus potentially resulting in an under-estimation of cell types capable of tumour initiation.

Towards a New Therapeutic Outlook The discovery of putative brain tumour stem cells and the growing understanding of neurogenesis and

29

cellular immortality, presents both a novel cellular target and novel molecular targets within. Much larger numbers of brain tumour samples and prospective multi-parameter cell sorting are needed to determine whether the cancer stem cell hypothesis is robust as well as to delineate patient subsets within a particular brain tumour type. Do all tumors under a common nomenclature reflect a common cell of brain cancer origin? What are the implications for therapy if they do not? In addition, evidence to date does not address whether the cell type conferring self-renewal capacity through tumorsphere-formation in vitro is the same cell type that initiates the neural tumour in vivo; these studies involve tumour-initiating populations (heterogeneous) rather than (homogenous) tumour-initiating cells. Similarly, it is important to separate tumour-initiation and tumour-propagation; this may not involve the same cell type as the tumourpropagating cell may be a much differentiated progeny of the tumour-initiating cell. It is conceivable therefore, that improved therapeutic efficacy may be achieved by targeting both the cell-type(s) which drives malignant progression as well as that which initiates – and maintains the stem cell pool of – the tumour. Since telomerase activity has been detected in almost all advanced brain tumors, the use of telomerase inhibitors may provide an attractive approach to therapy. These may be most effective in reducing the risk of relapse by targeting cancer stem cells. As normal neural stem cells would be expected to have longer telomeres than brain cancer stem cells, it follows that telomere depletion will be spared in the normal stem cell pool. There is a practical consideration here with respect to telomere length: if anti-telomerase approaches rely on telomere shortening for efficacious results, then the initial telomere length in the target cancer cell must be a consideration. This reinforces the brain tumour cell of origin, as telomere dynamics reflect the mitotic history of highly proliferating cells. Elucidation of telomere length and telomerase activity in subsets of cells within the tumour hierarchy will enable a more targeted anti-telomerase approach, perhaps in combination therapy. The identification of a greater array of surface markers will likely aid this characterisation. Key cell signalling pathways such as Notch and Sonic Hedgehog, that are involved in the regulation of self-renewal in both normal and cancer stem cells, also represent putative molecular targets. A more thorough

30

R. Rahman and R.G. Grundy

understanding of the mechanisms of signal transduction in each context, particularly with respect to gene expression levels will determine whether disrupting self-renewal pathways in brain tumour stem cells will spare normal neural stem cell pool function and result in tumour growth inhibition. In addition, better therapeutics will rely on a better understanding of the intrinsic mechanisms of resistance in brain tumour cells. The cancer stem cell hypothesis predicts that only tumourigenic cells are resistant to therapy. Recent work studying radiation resistance supports this idea as CD133+ glioma stem cells show preferential activation of the DNA damage response upon radiation treatment. However, whether the same cell type that confers radioresistance is the same cell type that confers chemoresistance is

unknown. Tumour resistance may also arise if brain tumour stem cells, similar to normal stem cells, generally reside in a quiescent state outside the cycle of division. Although direct evidence is lacking, if shown to be correct, this would affect the rationale and approach of drugs targeting telomerase and self-renewal pathways. In addition, brain tumour stem cells may plausibly have high levels of ATP-binding (ABC) drug transporters, which can protect cells from cytotoxic agents. This may result in the efflux of cytotoxic agents delivered to tumour stem cells. If ABC transporters are shown to contribute to tumour resistance, the next generation of chemotherapy regimes may involve chemosensitisers in conjunction with agents that alter drug-transporter activity. Understanding the molecular and genetic basis of the cell type(s) surviving

Fig. 3.3 Brain tumour cells exhibit intrinsic resistance to current treatment modalities. The cancer stem cell hypothesis presents an attractive framework upon which to consider a gestalt shift in treatment protocols for CNS tumors. (Top) Existing therapeutic regimes successfully achieve a substantial degree of tumour remission prior to tumour relapse fuelled by a brain tumour population(s) with intrinsic resistance to therapy. There is increasing evidence that the resistant cell type is a neural stem-like cell with dysregulated self-renewal or neural progenitor which has acquired aberrant self-renewal capabilities through mutation. Mechanisms of tumour stem cell resistance are largely based on presumptions of shared properties with neural stem cells; however evidence is beginning to emerge demonstrating that brain tumour stem/progenitor cells are intrinsically protected, relative to the bulk tumour

(yellow box). (Bottom) It is likely a new therapeutic paradigm will include chemosensitizers and cyotoxic agents that target the dysregulated pathways of neural stem/progenitor cells and mechanisms that mediate resistance, such as ABC-drug transporters. Characterisation of telomere length/structure and telomerase activation levels in brain tumour sub-compartments will reveal whether anti-telomerase therapy will be an attractive proposition for CNS tumors. To accomplish a durable response within this new therapeutic paradigm, conventional cytotoxic agents would be used to de-bulk the tumour mass in a combinatorial strategy with agents for targeted therapy. Nevertheless, a number of challenging, pertinent and often controversial biological questions surround the cancer stem cell paradigm (see section on The Brain Tumour Stem Cell Paradigm)

3

Cellular Immortality in Brain Tumors: An Overview

multi-modal therapy will be crucial to our unwavering hope of a curative response in brain cancer patients (Fig. 3.3). Is each factor contributing to resistance, both necessary and sufficient for tumour resistance? Or is coercion required involving all factors of resistance? If the latter, will targeting one mechanism of resistance suffice?

Concluding Remarks The brain cancer stem cell paradigm without reference to neurodevelopment and dysregulated cellular proliferation is perhaps reductionist and oversimplified. The emerging concept of a close relationship between developmental neurobiology, telomere/ telomerase biology and brain cancer stem cells, encourages a consilient approach to develop the next generation of brain cancer therapeutic interventions. At one time disparate, these areas of investigative biology should help determine the cellular culprit of particular brain cancer types, both with respect to the cell type where neoplastic transformation occurs and to the molecular events that permit it. If these tumour-initiating cells are intrinsically resistant to conventional therapeutic strategies, then a comprehensive characterisation of this mechanism will allow more effective combination therapy to be developed in the future.

References Bao S, Wu Q, Sathornsumetee S, Hao Y, Li Z, Hjelmeland AB, Shi Q, McLendon RE, Bigner DD, Rich JN (2006) Stem cell-like glioma cells promote tumour angiogenesis through vascular endothelial growth factor. Cancer Res 66:7843– 7848 Behin A, Hoang-Xuan K, Carpentier AF, Delattre JY (2003) Primary brain tumors in adults. Lancet 361:323–331 Beier D, Hau P, Proescholdt M, Lohmeier A, Wischhusen J, Oefner PJ, Aigner L, Brawanski A, Bogdahn U, Beier CP (2007) CD133(+) and CD133(−) glioblastoma-derived cancer stem cells show differential growth characteristics and molecular profiles. Cancer Res 67:4010–4015 Bonfanti L, Peretto P (2007) Radial glial origin of the adult neural stem cells in the subventricular zone. Prog Neurobiol 83:24–36 Bonnet D, Dick JE (1997) Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3:7307

31 Cheng A, Shin-ya K, Wan R, Tang SC, Miura T, Tang H, Khatri R, Gleichman M, Ouyang X, Liu D (2007) Telomere protection mechanisms change during neurogenesis and neuronal maturation: newly generated neurons are hypersensitive to telomere and DNA damage. J Neurosci 27:3722–3733 Gage FH (2000) Mammalian neural stem cells. Science 287:1433–1438 Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C, Dimeco F, Vescovi A (2004) Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res 64:7011–7021 Greider CW, Blackburn EH (1989) A telomeric sequence in the RNA of Tetrahymena telomerase required for telomere repeat synthesis. Nature 337:331–337 Gross CG (2000) Neurogenesis in the adult brain: death of a dogma. Nat Rev Neurosci 1:67–73 Hallahan AR, Pritchard JI, Hansen S, Benson M, Stoeck J, Hatton BA, Russell TL, Ellenbogen RG, Bernstein ID, Beachy PA (2004) The SmoA1 mouse model reveals that notch signaling is critical for the growth and survival of sonic hedgehog-induced medulloblastomas. Cancer Res 64: 7794–7800 Hartfuss E, Galli R, Heins N, Gotz M (2001) Characterization of CNS precursor subtypes and radial glia. Dev Biol 229:15–30 Hayflick L, Moorhead PS (1961) The serial cultivation of human diploid cell strains. Exp Cell Res 25:585–621 Hemmati HD, Nakano I, Lazareff JA, Masterman-Smith M, Geschwind DH, Bronner Fraser M, Kornblum HI (2003) Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci USA 100:15178–15183 Klapper W, Shin T, Mattson MP (2001) Differential regulation of telomerase activity and TERT expression during brain development in mice. J Neurosci Res 64:252–260 Lapidot T, Sirard C, Vormoor J, Murdoch B, Hoang T, CaceresCortes J, Minden M, Paterson B, Caligiuri MA, Dick JE (1994) A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature 367:645–648 Mori, T, Buffo, A, and Gotz, M. (2005) The novel roles of glial cells revisited: the contribution of radial glia and astrocytes to neurogenesis. Curr Top Dev Biol 69:67–99 Noctor SC, Flint AC, Weissman TA, Dammerman RS, Kriegstein AR (2001) Neurons derived from radial glial cells establish radial units in neocortex. Nature 409:714–720 Noctor SC, Martinez-Cerdeno V, Kriegstein AR (2007) Neural stem and progenitor cells in cortical development. Novartis Found Symp 288:59–73; discussion 73–78, 96–98 Packer RJ (2008) Childhood brain tumors: accomplishments and ongoing challenges. J Child Neurol 23:1122–1127 Palma V, Lim DA, Dahmane N, Sanchez P, Brionne TC, Herzberg CD, Gitton Y, Carleton A, Alvarez-Buylla A, Ruiz i Altaba A (2005) Sonic hedgehog controls stem cell behavior in the postnatal and adult brain. Development 132: 335–344 Rahman R, Osteso-Ibanez T, Hirst R, Levesley J, Quinn S, O’Callaghan C, Coyle B, Grundy R (2010) Histone deacetylase inhibition attenuates cell growth with associated telomerase inhibition in high grade paediatric brain tumour cells. Mol Canc Therapeut 9:2568–2581 Reynolds BA, Weiss S (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255:1707–1710

32 Shay JW, Bacchetti S (1997) A survey of telomerase activity in human cancer. Eur J Cancer 33:787–791 Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB (2003) Identification of a cancer stem cell in human brain tumors. Cancer Res 63:5821–5828 Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB (2004) Identification of human brain tumour initiating cells. Nature 432:396–401 Taylor MD, Poppleton H, Fuller C, Su X, Liu Y, Jensen P, Magdaleno S, Dalton J, Calabrese C, Board J (2005) Radial glia cells are candidate stem cells of ependymoma. Cancer Cell 8:323–335 Wang J, Sakariassen PO, Tsinkalovsky O, Immervoll H, Boe SO, Svendsen A, Prestegarden L, Rosland G, Thorsen F, Stuhr L (2008) CD133 negative glioma cells form tumors in

R. Rahman and R.G. Grundy nude rats and give rise to CD133 positive cells. Int J Cancer 122:761–768 Wright LS, Prowse KR, Wallace K, Linskens MH, Svendsen CN (2006) Human progenitor cells isolated from the developing cortex undergo decreased neurogenesis and eventual senescence following expansion in vitro. Exp Cell Res 312:2107–2120 Wu SX, Goebbels S, Nakamura K, Kometani K, Minato N, Kaneko T, Nave KA, Tamamaki N (2005) Pyramidal neurons of upper cortical layers generated by NEX positive progenitor cells in the subventricular zone. Proc Natl Acad Sci USA 102:17172–17177 Zhao CS, Overstreet-Wadiche L (2008) Integration of adult generated neurons during epileptogenesis. Epilepsia 49 (Suppl 5):3–12

Part I

Tumor to Tumor Passage of Malignancy

Chapter 4

Tumor-to-Tumor Metastasis: Extracranial Tumor Metastatic to Intracranial Tumors Jian-Qiang Lu and Arthur W. Clark

Abstract Tumor-to-tumor metastasis (TTM) is a relatively rare but well-documented phenomenon. When malignant tumor metastasizes to an intracranial tumor, meningiomas are most often the recipient; and breast or lung carcinoma the most common donor (primary tumor). The diagnosis of TTM can be made only by histopathological examination. Awareness of TTM is essential in clinical practice for timely diagnosis and early detection of malignant disease. The pathogenesis of TTM is related to various factors including vascularity and indolent growth of the recipient tumor; and simultaneous occurrence of the particular donor and recipient tumors (most notably breast carcinoma and meningioma). Most essential for understanding pathogenesis, however, are current concepts in the pathogenesis of metastatic cancer, an edifice rising on the concept of “seed and soil” expressed by Paget in 1889. The recently developed metastatic niche model is built on “seed and soil” theory, and describes the evolution of a conducive microenvironment in which disseminated tumor cells engraft and proliferate at the secondary sites. Early interventions that target both the disseminating seed and the metastatic soil may enable improvements in prognosis of malignant tumors. Keywords Metastasis · TTM · Extracranial · Intracranial · Pathogenesis · Gliomas

J.-Q. Lu () Department of Lab Medicine and Pathology, 5B2.24 WCM Health Sciences Centre, University of Alberta Hospital, Edmonton, AB, Canada TG6 2B7 e-mail: [email protected]

Introduction and Definitions Tumor metastases to distant sites represent a major cause of mortality, being responsible for 90% of all malignant tumor-related deaths (Fokas et al., 2007). Tumor-to-tumor metastasis (TTM) is a rare but welldocumented entity. It was first reported by Berent in 1902 (Campbell et al., 1968); about two hundred cases have since been reported (Petraki et al., 2003; Lu et al., 2009). TTM are metastatic deposits from a systemic neoplasm into a pre-existing (intracranial or extracranial) tumor. The diagnostic criteria of TTM as discussed by Campbell et al. (1968) pertain to tumors in general, and can be summarized as follows: (1) more than one primary tumor must exist; (2) the recipient tumor is a true neoplasm; (3) the metastatic neoplasm is a true metastasis with established growth in the host tumor, not the result of contiguous growth or embolization of tumor cells; and (4) tumors that have metastasized to the lymphatic system where lymphoreticular malignant tumors already exist are excluded. Authors reporting extracranial to intracranial TTM have referred to the criteria proposed by Chambers and colleagues (Chambers et al., 1980), adopted by Pamphlett (1984), and elaborated by Jarrell and colleagues (Jarrell et al., 2006): (1) the metastatic nidus must be at least partially enclosed by a rim of histologically distinct primary tumor tissue; (2) the existence of the primary tumor must be proven; and (3) the metastatic tumor must be demonstrated to be compatible with the primary tumor by morphological or immunohistochemical means. Once a TTM has been recognized, it is obviously important to prove the existence of the metastasizing primary tumor. However, the

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_4, © Springer Science+Business Media B.V. 2011

35

36

first evidence that a primary tumor exists may come either before or only after the diagnosis of metastases within the recipient tumors (Caroli et al., 2006; Lu et al., 2009). TTM should be distinguished from a “collision tumor” as for example when leptomeningeal spread of carcinoma extends to or even invades a pre-existing meningioma (see discussion in Takei and Powell, 2009); or the contiguous occurrence of a meningioma and breast carcinoma with brain metastasis (the case reported by Seckin et al., 2006, and referred to as a collision by Binello et al., 2010). However TTM and “collision tumor” have occasionally been used interchangeably in the literature because the two phenomena can be difficult or impossible to distinguish in some cases. The term “collision tumor” refers to two neoplasms occurring together in the same location. In passing, it should be noted that a patient may have multiple primary brain tumors of different cell types; and in some cases the two primary brain tumors may be contiguous (Deen and Laws, 1981), qualifying as a “collision tumor.” Several other terms are widely used in literature on primary and metastatic tumors. The terms “synchronous” and “metachronous” are used to refer to the metastases discovered (respectively) close to, or much later than, the discovery of the primary tumor. The definition of the time boundary between “synchronous” and “metachronous” varies from study to study, and may be as short as 2 months or as long as a year. The terms “single” and “solitary” are used to refer to the intracranial metastases which are (respectively) associated, or not associated, with metastases elsewhere. The tumors frequently seen as donors in TTM (to include both extracranial and intracranial recipients) include lung, breast, prostate, and malignant melanoma. The common recipient tumors are meningioma, renal cell carcinoma, sarcoma, thyroid neoplasm, and pituitary adenoma (Bret et al., 2001; Campbell et al., 1968; Lu et al., 2009; Petraki et al., 2003). Although TTM is typically seen as an aggressive epithelial neoplasm (carcinoma) metastatic to a benign or low-grade malignant neoplasm, the recipient tumors may be benign or malignant. Of the malignant recipients, renal cell carcinoma is the most common; of the benign recipients, meningioma is overwhelmingly the most common (Petraki et al., 2003; Takei and Powell, 2009).

J.-Q. Lu and A.W. Clark

The focus of this review is on the metastases of extracranial to intracranial tumors, hereinafter referred to as intracranial TTM. In 1930, Fried first described a metastasis from lung carcinoma to a meningioma (Pamphlett, 1984). Since then the reported cases of intracranial TTM have been steadily accumulating (Lu et al., 2009). Breast and lung carcinomas are common donor tumors both for the metastases to intracranial tumors and for the metastases to the brain and leptomeninges. However melanoma, which is relatively frequent among malignant tumors metastasizing to the brain, has rarely been the donor for intracranial TTM. Intracranial tumors that have been reported as recipients include meningiomas (by far the most common), pituitary adenomas, hemangioblastomas, gliomas, schwannomas, craniopharyngioma, and cavernous hemangioma (Table 4.1). Table 4.1 summarizes the reported cases of tumor metastases to intracranial tumors. For the purposes of tabulation, we have excluded the metastases to tumors in the spinal region, where hemangioblastomas have been the most frequent recipient (Jarrell et al., 2006) but TTM involving spinal meningiomas have also been reported (Caroli et al., 2006). The relative frequency of meningiomas and rarity of gliomas as recipients in intracranial TTM immediately raise questions about pathogenesis. For insights into pathogenesis we will refer to the literature on TTM in general; and to the more basic concepts of progression and metastasis in malignant tumors.

Pathogenesis Preferential targeting of specific organs for the metastases from specific malignant tumors is well known. Examples include the tendency of prostate cancer to metastasize to bone but not to the brain; the ease with which melanoma metastasizes to either the brain or bone; and the remarkable predilection of metastases from uveal melanoma for the liver. It has been acceptable that colonization of metastatic tumor cells in particular target organs is obviously dependent on the biology of the target organ (i.e., organ-specific metastases). The sites, temporal course, and severity of metastases to the brain and other particular organs are determined by several factors such as the cellular origin, intrinsic properties of the tumor,

4

Tumor-to-Tumor Metastasis

37

Table 4.1 Reported cases of metastases to intracranial tumorsa Recipient tumor

Donor tumor

Approx. no. of reported cases

Meningiomas

Breast ca Lung ca Renal cell ca Prostate ca Melanoma Pancreatic ca Gastric ca Colon ca Carcinoid (lung) Cervical squamous cell ca Endometrial adenoca Lymphoma Ca, unknown primary origin

28 27 3 3 2 2 1 1 1 1 1 1 3

Benedetto et al. (2007), Caroli et al. (2006), Petraki et al. (2003), Tally et al. (1988), Takei and Powell (2009)

Pituitary adenomas

Breast ca Lung ca Renal cell ca Gastric ca Colorectal ca Carcinoid (mediastinum) Pancreatic ca Prostate ca Ca, unknown primary origin

4 3 2 2 1 1 1 1 3

Bret et al. (2001), Petraki et al. (2003), Tally et al. (1988), Weber et al. (2003)

Hemangioblastomas

Renal cell ca Pancreatic ca

6 1

Jarrell et al. (2006), Martin et al. (2010)

Schwannomas (CPA)

Lung ca Breast ca

4 2

Caroli et al. (2006), Tally et al. (1988)

Oligodendrogliomas

Breast ca Colon ca Melanoma

2 1 1

Caroli et al. (2006), Tally et al. (1988)

Gliomas, not specified

Melanoma Ca, unknown primary origin

1 2

Caroli et al. (2006), Tally et al. (1988)

Glioblastomas

Thyroid adenoca Renal cell ca

1 1

Caroli et al. (2006), Petraki et al. (2003)

Acoustic neurinomas

Breast ca Lung ca

1 1

Caroli et al. (2006)

Selected references

Ependymoma Lung ca 1 Caroli et al. (2006) Craniopharyngioma Lung adenoca 1 Fraggetta et al. (2000) Cavernous hemangioma Lung ca 1 Tally et al. (1988) Lymphoma 1 Tally et al. (1988) Choristoma of infundibulum Paraganglioma (sellar) Esophageal ca 1 Lu et al. (2009) a Cases of metastases to the spinal cord are not included. Ca, carcinoma; CPA, cerebellopontine angle

tissue affinities and circulation patterns (Nguyen et al., 2009). Preferential targeting of a specific (benign) tumor for the metastases from a specific malignant tumor has also been identified. Jarrell and colleagues (2006) studied several patients with von Hippel Lindau disease in whom a renal cell carcinoma had metastasized to a hemangioblastoma but not to other tissues. They

concluded that hemangioblastomas are an early and preferred site for metastasis in von Hippel-Lindau disease. The reasons that hemangioblastomas are attractive targets for the metastases might be their biological features such as slow growth, high vascularity, as well as high lipid content (Jarrell et al., 2006; Martin et al., 2010). In addition, for hemangioblastomas associated with von Hippel-Lindau disease, the

38

J.-Q. Lu and A.W. Clark

impaired VHL protein function disrupts fibronectin matrix assembly, potentially making these tumors more vulnerable to metastatic seeding (Jarrell et al., 2006). It is not clear whether tumors are generally more or less hospitable to metastases as compared to nonneoplastic (“normal”) organ tissue. TTM has been considered rare, and several factors, such as tumorproducing substances locally antagonistic to other new growth, antigenic difference in membranous structures, and nutritional competition with other new growth (see discussion in Petraki et al., 2003), might suppress the growth of cells from one tumor inside a second tumor. A contrasting view emerges from the work of Greene and Strauss (1949), who concluded from their 17-year-long study of spontaneous neoplasia in a colony of rabbits that, when the rabbits developed metastases from more than one tumor, there was a higher than expected tendency for the cells of one tumor to lodge and grow in the metastases of another. That the internal environment of a tumor is an excellent nidus for growth is self-evident, and the absence of any factor of incompatibility between different tumor types is readily demonstrated by the simple procedure of transplanting from one tumor to another. In such an experiment, in which a carcinoma was transplanted to the normal subcutaneous tissues as well as to a previously transplanted sarcoma in the same animal, a take not only occurred in the latter site but growth was actually much more rapid. – Greene HSN and Strauss JS. (1949)

Meningiomas, the most common benign recipient tumor in TTM, have several biological features that might make them fertile “soil” for seeding of metastases: (1) rich vascularity that may enhance the chances of meningiomas to receive hematogenous metastases, (2) slow growth rate that may allow a greater period of time for metastasis to develop, (3) particular hormonal factors that may be related to breast carcinomas (a very common donor to meningiomas), and (4) low metabolic rate that may provide a noncompetitive metabolic environment conducive to the growth of metastatic tumors (Caroli et al., 2006; Takei and Powell, 2009). Other evidence that may be relevant to the frequency of breast carcinoma metastases to meningioma include a higher-than-expected simultaneous occurrence of those two tumors; and common expression in both tumors of E-cahedrin (Watanabe et al., 2002), and the c-myc oncogene (Caroli et al., 2006). Meningiomas seem to have

particular propensity of receiving breast and lung carcinomas (Table 4.1) but not melanomas, despite the propensity for melanoma to metastasize to the brain. In contrast to meningiomas and hemangioblastomas, pituitary adenomas lack high collagen or lipid content (Bret et al., 2001; Weber et al., 2003). They have occasionally been the recipient tumor in metastases. The disruption of either the normal pituitary vascular system by an adenoma (Weber et al., 2003) or the adenoma’s microvasculature by a previous surgery or radiation (Bret et al., 2001) may predispose to infiltration by blood-born metastases. Of course the nonadenomatous pituitary gland more frequently receives a metastasis than does a pituitary adenoma; and in those cases, it is the posterior lobe (neurohypophysis) rather than the anterior lobe (adenohypophysis) that is more often the target – possibly because of the direct systemic arterial blood supply to the neurohypophysis (Weber et al., 2003). A serious and comprehensive approach to the pathogenesis in TTM must be based in the larger efforts to understand progression and metastasis in malignant tumors. This has been the subject of intense research for over a century (Fidler, 2003; Fokas et al., 2007). More than a century ago, tumor dissemination was thought to be purely determined by mechanical factors that caused tumor cell emboli to lodge in the vasculature (reviewed in Psaila and Lyden, 2009). However, Stephen Paget (1889) made an extremely important observation that certain organs (e.g., liver) appeared to be particularly receptive to metastases (e.g., from breast cancer). Since his finding was not explicable by blood flow alone, he concluded that the “soil” or local microenvironment of these organs must be more conducive for disseminating tumor cells to “seed” than that of other organs such as the spleen. The “seed and soil” theory proposed by Paget (1889) is a milestone in the study of malignant tumors and their metastases. Forty years later, Paget’s theory was challenged by James Ewing’s theory that re-affirmed the idea that metastasis was determined by the anatomy of the vascular and lymphatic channels (Psaila and Lyden, 2009). Ewing’s theory prevailed until seminal studies by Fidler and Kripke (1977) who conclusively demonstrated that, although tumor cells reached the vasculature of all organs, metastases are organ-specific (Nguyen et al., 2009). The “seed” and “soil” theory was then revived and became the most popular theory for the metastasis of malignant tumors (Fidler, 2003; Fokas et al., 2007).

4

Tumor-to-Tumor Metastasis

The metastatic process was envisioned as several steps: invasion, entry into systemic circulation (intravasation); movement from the circulatory system into a new host tissue (extravasation); and proliferation and growth of the secondary tumor (Fidler, 2003). Metastasis has been traditionally thought of as an event that occurred at a late stage in malignant progression of the primary tumor. Recent progress, however, strongly challenges the latter assumption. Disseminated tumor cells (DTCs) may spread throughout the body at a much earlier stage than had been supposed. The term DTC is used for any tumor cell that has left the primary lesion and travelled to ectopic environment. Klein (2009) has provided an excellent review of two fundamental models of metastasis: the linear progression model in which tumor ontogeny proceeds to full malignancy within the primary tumor microenvironment, followed by the tumor cell dissemination to found a metastasis; and the parallel progression model in which tumor cells depart the primary lesion before the acquisition of fully malignant phenotypes to undergo somatic progression and metastatic growth at distant sites. The data from disease courses, tumor growth rates, autopsy studies, clinical trials and molecular genetic analyses of the primary tumor cells and DTCs are in support of the parallel progression model, although direct and incontrovertible evidence for either model is lacking. Once the primary tumors are removed, metachronous metastases must arise from DTCs. However, the majority of DTCs will not grow into a metastasis. In terms of Paget’s seed and soil conceptual framework, the molecular traits of the seeds (DTCs) as well as the anatomic, cellular, and molecular features of the host environment (soil) will determine whether metastasis evolves (Nguyen et al., 2009). The “metastatic niche model,” built on Paget’s seed and soil concepts, emphasizes the interactions of malignant cells with their microenvironments at the metastatic sites (Psaila and Lyden, 2009). The microenvironment or soil comprises supportive (non-malignant) stromal cells, soluble factors, vascular networks, nutrients and metabolic components and the structural extracellular matrix architecture. DTCs must promote these adaptations of the foreign soil in order to make the transition from micro- to macrometastases. Thus the evolution of a metastatic niche includes (1) a pre-metastatic stage: mechanical forces of the vascular channels govern the initial delivery of cells from the primary tumor to distant tissues, (2) a micro-metastatic stage:

39

metastatic tumors cells engraft the niche to populate micro-metastases, and (3) a macro-metastatic stage: recruitment of endothelial progenitor cells to the early metastatic niche mediates the angiogenic switch and enables progression to macro-metastases (Psaila and Lyden, 2009). This model of the metastatic process has been supported by substantial data, but it is somewhat speculative and how it applies to brain metastases remains an open question (Ellison et al., 2008; Nguyen et al., 2009). The process of understanding the pathogenesis of metastases to the brain is likely to accelerate in the next decade. The spread of extracranial malignant tumors to the brain must be viewed in the context of both “seed” and “soil”, with recognition of the “soil” of the central nervous system (CNS) may be quite distinct from that of other organs (Ellison et al., 2008). Approximately 20–30% of patients with primary tumors develop metastases to the brain (Fokas et al., 2007) and several factors influencing the process have been identified. First, angiogenesis is necessary for the development and growth of tumors. Clinical and experimental evidence has suggested that spreading the tumor cells and formation of metastases are directly related to the number of micro-vessels in the primary tumor. Brain metastases selected from the breast carcinoma variants MDA231-BR1, -BR2 and BR3 with increased metastatic ability are characterised by richer vascularity in comparison to the variants with lower spreading potential. The treatment with VEGFA inhibitor PTK787/Z 222584 resulted in the shrinkage of brain metastatic lesions and diminished vascularisation, implicating VEGF as a contributor of brain metastases from primary breast carcinomas (Fokas et al., 2007). The patients with high microvascular densities in brain tumors have shorter postoperative survival than do the patients with low microvascular densities, suggesting that brain tumor vasculature is important to its growth (Kim and Lee, 2009). Secondly, astrocytes produce various cytokines and growth factors such as IL-1, IL-6, TNG-α, TGF-β, IFN-γ, and PDGF, which modulate brain microenvironment. MDA-MB-435 BR1 breast carcinoma cells derived from MDA-MB-435-induced brain metastases showed increased adhesion to astrocytes, which was reversed by antibodies against IL-6, TGF-β, and IGF-1 (Fokas et al., 2007). The release of these chemokines and growth factors favors specific malignant tumors (for example, breast carcinomas) seeding to the brain parenchyma (Sierra et al., 1997).

40

Thirdly, various signal neurotransmitters in the brain can influence seeding and development of metastases to the CNS (Entschladen et al., 2004). It has been suggested that met-enkephalin, substance P, bombesin, dopamine, and norepinephrine may enhance while gamma-aminobutyric acid (GABA) may inhibit chemotactic cell migration (Fokas et al., 2007). Finally, the identification of metastasis suppressor genes (MSGs) and proteins that specifically inhibit the ability of cells to form metastases provides new insights into molecular mechanisms underlying the regulation of this complex process (Rinker-Schaeffer et al., 2006). The MSGs that have been identified to date include Non metastatic (Nm23) gene, differentiation related gene (Drg-1), Src-suppressed C kinase substrates (SSeCKs), Vitamin D3 upregulated protein 1 (VDUP), mitogen activated protein kinase 4 (MKK4), MKK6, MMK7, and Kiss-1 gene. These genes often affect fundamental signalling pathways of the cell. MSGs do not affect the spread of tumor cells to distant tissue, but inhibit the colonization of the micro-metastases. In brain metastases from ductal invasive breast carcinomas, prominent reduction in transcriptional and protein expression of the MSGs such as Kiss-1 and MMK4 was found in comparison to the primary malignant tumors (Stark et al., 2005). In addition, the GTP-binding proteins of the Rho superfamily (Rho, Rac, Cdc42) also operate as metastasis suppressor proteins. Several other proteins including signal transducer and activator of transcription 3 (Stat3), neurotrophins, and the receptor tyrosine kinase HER-2/neu, may also play a role in metastasizing malignant tumor cells to the brain (Fokas et al., 2007). The different classes of metastasis genes, such as metastasis initiation genes, metastasis progression genes, and metastasis virulence genes, have been thought to mediate metastatic progression in certain tumor types (Nguyen et al., 2009). However, their role in tumor metastases to the brain needs further investigation.

Clinical and Epidemiologic Correlates Metastatic tumors to the brain are approximately 10 times more common than primary intracranial tumors (Ellison et al., 2008). The incidence rates found in

J.-Q. Lu and A.W. Clark

the literature for brain metastases (up to 11 per 100, 1000 population per year) probably underestimate the true incidence, due to underdiagnosis and inaccurate reporting (Wesseling et al., 2007). Similarly, intracranial TTM may be significantly underdiagnosed and underreported. The incidence of brain metastases varies according to the histological type of the primary tumors, but increases with age. Up to 30% of adults and 6–10% of children with malignant tumors develop brain metastases. In contrast to the enormous number of brain metastases from extracranial tumors, the reported cases of intracranial TTM number is about 117 (Table 4.1, based on English language literature to June 2010). A gender bias is not significant for metastases to the brain (Wesseling et al., 2007), whereas females are more often affected by intracranial TTM (Tally et al., 1988). This female bias is demonstrable in the most common recipients including meningiomas (Caroli et al., 2006), pituitary adenomas (Bret et al., 2001), and hemangioblastomas (Jarrell et al., 2006; Martin et al., 2010). The female predominance in the metastases to meningiomas and pituitary adenomas (Bret et al., 2001) is chiefly because of the high prevalence of breast carcinoma among the donor tumors. In up to 10% of brain metastases, no primary tumor is found at presentation (Wesseling et al., 2007). Similarly, the primary extracranial tumor metastasizing to an intracranial tumor is often unknown at the time of intracranial surgery. The metastases to other sites may be known or identified at the time of diagnosis of the intracranial TTM, and this is probably true for the vast majority of intracranial TTM that are first identified at autopsy (Caroli et al., 2006; Chambers et al., 1980). Clinical manifestations of intracranial metastases vary with the location, size, and tumor type of metastases. Brain metastases may be, but not always symptomatic; about 6 percent of patients with lung carcinomas may have asymptomatic brain metastases (Ellison et al., 2008). The neurological signs and symptoms are generally caused by increased intracranial pressure or local effect of the tumors on the adjacent brain tissue. The most common signs and symptoms of intracranial metastases include headache, altered mental status, paresis, ataxia, visual complaints, nausea or sensory disturbances. Some patients present acutely with seizure, infarct or hemorrhage (Wesseling et al., 2007).

4

Tumor-to-Tumor Metastasis

The fact that a large percentage of intracranial TTM are first diagnosed at autopsy (Bhargava et al., 1999; Caroli et al., 2006; Chambers et al., 1980) suggests that many are either asymptomatic or obscured by other clinical problems such as metastatic disease at other sites. Clinical manifestations of intracranial TTM generally correlate with the site of the recipient tumor, but the evolution of signs and symptoms may be more rapid than expected for the recipient tumor. Meningiomas, for example, are generally indolent tumors for which surgical removal alone is sufficient for cure; but when meningiomas are the recipient of a metastasis, the clinical course is sometimes stormy (Bhargava et al., 1999). Our case (Lu et al., 2009) of a suprasellar TTM, in which the recipient tumor was a paraganglioma, illustrates these principles. The patient was an 81-year-old male, previously healthy, who presented with a 3-week history of headache and blurred vision. Examination revealed bitemporal hemianopsia. Laboratory tests including hormone levels were unremarkable. Brain MRI revealed a well-circumscribed sellar and suprasellar tumor with heterogeneous T1-, T2-weighed images (Fig. 4.1, left), and gadolinium enhancement (Fig. 4.1, right). Microscopic examination revealed a paraganglioma harbouring a metastatic carcinoma. His primary esophageal carcinoma was not found until 2

41

months later and he died 4 months after the neurosurgical procedure.

Diagnosis The correct diagnosis of TTM cannot be established before histopathological studies (Bret et al., 2001). Pre-operative imaging has sometimes shown the features suggestive of, but never specific for, the diagnosis. In cases of the metastases to meningioma, some authors have reported hypodense areas in a hyperdense, enhancing mass on CT scan; or on MRI an area of marked enhancement surrounded by a moderately enhancing (recipient) tumor (Caroli et al., 2006; Pamphlett, 1984). Some reports indicate severe mass effort of the TTM even when a low grade meningioma is the recipient; and peritumoral edema may be prominent (Takei and Powell, 2009). Watanabe and colleagues (2002) found high choline/creatine ratio and lactate/lipid peak on proton magnetic resonance spectroscopy, suggesting the malignant component in the meningioma harbouring breast carcinoma metastasis. Bret et al. (2001) reported a case of breast carcinoma metastatic to pituitary adenoma, in which a gadolinium-enhanced MRI revealed a distinct nodular area with sharp demarcation from the surrounding

Fig. 4.1 Preoperative coronal T2- (left) and post-gadolinium T1-weighted (right) images showing the sellar and suprasellar tumor (arrow) with intense heterogeneous enhancement. Reproduced with permission from Lu et al. (2009)

42

tumor mass. The clinical context might lower the threshold for consideration of TTM. However, to our knowledge, the diagnosis has not been made prior to histopathologic examination, even in cases of von Hippel Lindau with renal cell carcinoma metastases to hemangioblastoma (Jarrell et al., 2006; Martin et al., 2010). Intraoperatively, most metastases growing within intracranial recipient tumors cannot be recognized grossly (Jarrell et al., 2006). Occasionally, the metastases may be distinguishable from their recipient tumors by a focally different color and/or texture. The tissue being resected has sometimes been described as “hard,” “hemorrhagic,” or “composed of both soft and solid elements” (Bret et al., 2001). In one of the 6 cases of metastases to hemangioblastomas reported by Jarrell et al. (2006) the metastatic lesion was tan, well circumscribed, and visible through a thin layer of the hemangioblastoma during resection. However, the other cases of that report gave no visible or palpable evidence to the surgeon that a TTM was being removed. In our case study of the esophageal carcinoma metastatic to an intracranial paraganglioma (Lu et al., 2009), during the surgery, the tumor had a variable consistency, both soft and firm. The softer component was removed using gentle suction, while the firmer component required the use of angled curettes and biopsy forceps. The diagnosis of TTM is sometimes achievable at the intraoperative pathology consultation (with frozen sections and smear preparations), because the histologic distinction of recipient and metastatic tumors can usually be made with routine staining. Various diagnostic difficulties may be encountered, however, even after examination of the paraffin sections. For example, it is difficult to identify the metastasis from renal cell carcinoma to a hemangioblastoma, in which the donor and recipient tumors may have similar histopathologic patterns. In such cases, immunohistochemistry can be particularly important in confirming the presence of TTM (Jarrell et al., 2006; Martin et al., 2010). Immunohistochemistry is also important more generally, in characterizing the donor and recipient tumors, and in gathering evidence to suggest where the primary malignant tumor is located (or, if that is known already, in confirming the immunohistochemical compatibility of the metastasis and the primary tumor). In our case of an esophageal carcinoma metastatic to an intracranial paraganglioma (Lu et al., 2009)

J.-Q. Lu and A.W. Clark

there were two distinct components admixed in some areas (Fig. 4.2a): one component showed the classic Zellballen pattern, characteristic of a paraganglioma (Fig. 4.2b–c); the other component demonstrated malignant features with focal necrosis and frequent mitoses (Fig. 4.2d–f), which exhibited immunoreactivity for CAM5.2, pancytokeratin, but not for thyroid transcription factor-1. The proliferation rate, assessed by Ki-67 immunostaining, was very low in the benign component, but extremely high (more than 90%) in the malignant component (Fig. 4.2f). These histopathological features indicated the diagnosis of poorly differentiated carcinoma metastatic to a paraganglioma of the sellar and suprasellar region. It was only two months later, however, that the primary site of the malignant tumor became known, when the patient presented with upper gastrointestinal bleeding. Endoscopy with multiple biopsies of the esophagus and stomach were performed. Histopathological examination of the biopsy specimens revealed Barrett’s esophagus and an invasive poorly differentiated carcinoma, adjacent to an adenoma, arising in the lower esophagus (Fig. 4.3), consistent with a primary carcinoma. The proliferation rate of this carcinoma, assessed by Ki-67 immunostaining, was approximately 60%. The immunohistochemical profile of this esophageal carcinoma was identical to that of the intracranial metastasis within the sellar/suprasellar paraganglioma, indicative of the nature of TTM.

Treatment and Prognosis Treatment and prognosis of patients with intracranial TTM should be similar to that for patients with resectable brain metastases, except for the rare cases in which the intracranial recipient tumor is a malignant glioma that imposes its own survival limitations and management exigencies. Thus treatment and prognosis for the patients with intracranial TTM largely depend on the factors other than the intracranial TTM itself. Metastatic tumors of the CNS are a major cause of morbidity and mortality for the patients with malignant tumors. The prognosis of patients with brain metastasis is poor, with median survival typically less than 1 year. A variety of factors improve prognosis, including aggressive therapy, better performance status, less extensive systemic lesions and fewer brain metastases.

4

Tumor-to-Tumor Metastasis

Fig. 4.2 (a) Photomicrograph of a sellar and suprasellar tumor demonstrating two distinct components: benign (right lower) and malignant (upper) (H & E, original magnification X 200). (b) The benign component showing arrangement of chief cells in nests surrounded by a single layer of sustentacular cells, characteristic of paraganglioma (H & E, original magnification X 400). (c) The sustentacular cells partially highlighted by S100 immunostaining (original magnification X 400). (d) The malignant component exhibiting patternless

Fig. 4.3 Photomicrograph of the esophageal carcinoma showing poorly differentiated features adjacent to an esophageal adenoma (left upper) (H & E, original magnification X 200). Reproduced with permission from Lu et al. (2009)

43

arrangement of malignant cells with frequent mitoses (H & E, original magnification X 400). (e) Immunoreactivity for cytokeratin CAM5.2 (original magnification X 400) identified in the malignant component (upper), but not in the benign component. (f) Proliferation rate, labeled by Ki67 immunostaining (original magnification X 200), being extremely high in the malignant component (upper and left), but very low in the benign component. Reproduced with permission from Lu et al. (2009)

44

In particular, combined neurosurgical removal of isolated metastases and whole-brain radiation therapy appear to decrease neurological morbidity and may prolong survival (Ellison et al., 2008; Wesseling et al., 2007). The patients with “solitary” brain metastases by definition have no other known metastatic disease, while the patients with “single” brain metastases may have metastatic disease outside the brain (Schiff, 2001). In either case, the primary tumor may have been located or not. In the patients with “synchronous” brain metastases, the primary tumor and the brain metastasis have been discovered at about the same time; whereas in the patients with “metachronous” brain metastases, the primary tumor has been discovered at a comparatively long interval prior to discovery of the brain metastasis. Compared to the patients with brain metastases in general, those with solitary and synchronous brain metastases have a significantly better prognosis; and, in exceptional cases, survivals of ten years or more have been reported (Shahidi and Kvale, 1996). A median survival of 24 months and a 5-year survival of 21.4% have been reported for the patients with non-small cell lung cancer and synchronous brain metastases (many of them solitary) in whom surgical resection for both the brain metastasis and the lung primary tumor were carried out (Billing et al., 2001). Other groups have also reported significant improvements in outcomes in managing the patients with single or solitary brain metastases. It is within the foregoing context that the treatment and prognosis for intracranial TTM should be understood. A majority of intracranial TTM are associated with the metastases to extracranial sites (Campbell et al., 1968; Caroli et al., 2006). Survival times may be very short, even when the TTM is the presenting manifestation of the malignancy. This was true of our patient, who died 4 months after resection of the intracranial tumor (Lu et al., 2009). By contrast, although the reports of intracranial TTM usually do not provide very long term followup information, significant disease-free intervals can occur after the resection. In two cerebellar hemangioblastomas that were the recipients of metastases from renal cell carcinoma, follow-ups of 12 and 20 months revealed no evidence of primary disease progression or metachronous metastases and no adjuvant therapy had been required (Jarrell et al., 2006). In another case reported by these authors, the 28-year-old

J.-Q. Lu and A.W. Clark

female patient had pancreatic neuroendocrine tumor metastases to a cerebellar hemangioblastoma. She had undergone partial resection of a 10-cm pancreatic neuroendocrine tumor 66 months before the resection of the cerebellar hemangioblastoma. At the time of surgery for the cerebellar hemangioblastoma, CT imaging revealed a 2.5-cm primary pancreatic tumor and multiple metastases in the liver. These metastases were successfully treated. At the last follow up, 12 months after the resection of the cerebellar hemangioblastoma containing metastasis, the patient showed no evidence of new metastases. In the case studies of metastases to intracranial meningiomas, the postoperative follow-up revealed survival at 12 months in a 69-year-old female patient (Takei and Powell, 2009), and at 7 months in a 59-yearold female patient (Caroli et al., 2006), but death at 15 months in a 65-year-old female patient (Caroli et al., 2006). Weber et al. (2003) reported the metastasis of renal cell carcinoma to pituitary adenoma in a 62-yearold female patient who died 8 months after the subtotal resection of the intracranial tumor. McCormick et al. (1989) described a 35-year-old woman with a pituitary adenoma that was occupied by a metastatic renal cell carcinoma 12 months after the decompression of the adenoma. This patient expired at 22 months after the last decompression of the metastatic tumor. Current progress in malignant tumor treatment has included a critical review of the treatment of metastatic disease. The traditional assumption was that metastases were essentially similar to the primary tumors and that control of tumor growth applicable to the primary tumor should effectively suppress metastatic growth. This assumption has been challenged by accumulating molecular evidence and pre-clinical modelling (Sleeman and Steeg, 2010). Tumor growth control in itself can actually promote rather than suppress the formation and growth of metastases, and there are increasing numbers of examples where chemotherapy, radiotherapy and biological/targeted therapies have this effect. For example, adjuvant radiotherapy used for local growth control can promote metastasis via the socalled tumor bed effect, in which tumor recurrence in the irradiated field is associated with higher metastasis and poor prognosis. Such observations indicate that the traditional approaches to the treatment of metastases may be not only inadequate but also misdirected. New therapies built on a solid understanding of the

4

Tumor-to-Tumor Metastasis

process of metastatic disease are needed (Sleeman and Steeg, 2010). In conclusion, TTM is uncommon, but deserves more recognition. The metastases of extracranial malignant tumors to intracranial tumors are encountered in clinical practice from time to time. Awareness of TTM is essential in clinical practice so as to make accurate diagnoses and sometimes prompt an additional workup for the patients with unknown primary malignancy (Lu et al., 2009). The treatment and prognosis of intracranial TTM is one part of the larger challenge: prevention and treatment of metastatic tumors. Understanding of the cellular and molecular pathogenesis for metastatic disease is rapidly expanding on the “seed and soil” foundation articulated by Paget in 1889. Early intervention that targets both the disseminating seed and metastatic soil may be essential to improving prognosis in this devastating form of disease (Psaila and Lyden, 2009).

References Benedetto N, Perrini P, Scollato A, Buccoliero AM, Di Lorenzo N (2007) Intracranial meningioma containing metastatic colon carcinoma. Acta Neurochir (Wien) 149:799–803 Bhargava P, McGrail KM, Manz HJ, Baidas S (1999) Lung carcinoma presenting as metastasis to intracranial meningioma: case report and review of the literature. Am J Clin Oncol 22:199–202 Billing PS, Miller DL, Allen MS, Deschamps C, Trastek VF, Pairolero PC (2001) Surgical treatment of primary lung cancer with synchronous brain metastases. J Thorac Cardiovasc Surg 122:548–553 Binello E, Bederson JB, Kleinman GM (2010) Hemangiopericytoma: collision with meningioma and recurrence. Neurol Sci 31:625–630 Bret P, Jouvet A, Madarassy G, Guyotat J, Trouillas J (2001) Visceral cancer metastasis to pituitary adenoma: report of two cases. Surg Neurol 55:284–290 Campbell LV Jr, Gilbert E, Chamberlain CR Jr, Watne AL (1968) Metastases of cancer to cancer. Cancer 22:635–643 Caroli E, Salvati M, Giangaspero F, Ferrante L, Santoro A (2006) Intrameningioma metastasis as first clinical manifestation of occult primary breast carcinoma. Neurosurg Rev 29:49–54 Chambers PW, Davis RL, Blanding JD, Buck FS (1980) Metastases to primary intracranial meningiomas and neurilemomas. Arch Pathol Lab Med 104:350–354 Deen HG Jr, Laws ER Jr (1981) Multiple primary brain tumors of different cell types. Neurosurgery 8:20–25 Ellison DW, Perry A, Rosenblum M, Asa S, Reid R, Louis DN (2008) Tumors: non-neuroepithelial tumors and secondary effects. In: Love S, Louis DN, Ellison DW (eds)

45 Greenfield’s Neuropathol, 8th edn. Hodder Arnold, London, pp 2002–2182 Entschladen F, Drell TL 4th, Lang K, Joseph J, Zaenker KS (2004) Tumour-cell migration, invasion, and metastasis: navigation by neurotransmitters. Lancet Oncol 5: 254–258 Fidler IJ (2003) The pathogenesis of cancer metastasis: the ‘seed and soil’ hypothesis revisited. Nat Rev Cancer 3:453–458 Fidler IJ, Kripke ML (1977) Metastasis results from preexisting variant cells within a malignant tumor. Science 197: 893–895 Fokas E, Engenhart-Cabillic R, Daniilidis K, Rose F, An HX (2007) Metastasis: the seed and soil theory gains identity. Cancer Metastasis Rev 26:705–715 Fraggetta F, Galia A, Grasso G, D’Arrigo C, Cristaudo C, Giangaspero F (2000) Pulmonary adenocarcinoma metastatic to pituitary craniopharyngioma. J Clin Pathol 53:946–947 Greene HS, Strauss JS (1949) Multiple primary tumors in the rabbit. Cancer 2:673–691 Jarrell ST, Vortmeyer AO, Linehan WM, Oldfield EH, Lonser RR (2006) Metastases to hemangioblastomas in von HippelLindau disease. J Neurosurg 105:256–263 Kim WY, Lee HY (2009) Brain angiogenesis in developmental and pathological processes: mechanism and therapeutic intervention in brain tumors. FEBS J 276:4653–4664 Klein CA (2009) Parallel progression of primary tumours and metastases. Nat Rev Cancer 9:302–312 Lu JQ, Khalil M, Hu W, Sutherland GR, Clark AW (2009) Tumor-to-tumor metastasis: esophageal carcinoma metastatic to an intracranial paraganglioma. J Neurosurg 110:744–748 Martin SE, Al-Khatib SM, Turner MS, Douglas-Akinwande AC, Hattab EM (2010) A 41-year-old woman with von Hippellindau and a cerebellar lesion. Brain Pathol 20:511–514 McCormick PC, Post KD, Kandji AD, Hays AP (1989) Metastatic carcinoma to the pituitary gland. Br J Neurosurg 3:71–79 Nguyen DX, Bos PD, Massagué J (2009) Metastasis: from dissemination to organ-specific colonization. Nat Rev Cancer 9:274–284 Paget S (The Lancet 1889, reprinted in) (1989) The distribution of secondary growths in cancer of the breast. Cancer Metastasis Rev 8:98–101 Pamphlett R (1984) Carcinoma metastasis to meningioma. J Neurol Neurosurg Psychiatr 47:561–563 Petraki C, Vaslamatzis M, Argyrakos T, Petraki K, Strataki M, Alexopoulos C, Sotsiou F (2003) Tumor to tumor metastasis: report of two cases and review of the literature. Int J Surg Pathol 11:127–135 Psaila B, Lyden D (2009) The metastatic niche: adapting the foreign soil. Nat Rev Cancer 9:285–293 Rinker-Schaeffer CW, O’Keefe JP, Welch DR, Theodorescu D (2006) Metastasis suppressor proteins: discovery, molecular mechanisms, and clinical application. Clin Cancer Res 12:3882–3889 Schiff D (2001) Single brain metastasis. Curr Treat Options Neurol 3:89–99 Seckin H, Yigitkanli K, Ilhan O, Han U, Bavbek M (2006) Breast carcinoma metastasis and meningioma. A case report. Surg Neurol 66:324–327

46 Shahidi H, Kvale PA (1996) Long-term survival following surgical treatment of solitary brain metastasis in non-small cell lung cancer. Chest 109:271–276 Sierra A, Price JE, García-Ramirez M, Méndez O, López L, Fabra A (1997) Astrocyte-derived cytokines contribute to the metastatic brain specificity of breast cancer cells. Lab Invest 77:357–368 Sleeman J, Steeg PS (2010) Cancer metastasis as a therapeutic target. Eur J Cancer 46:1177–1180 Stark AM, Tongers K, Maass N, Mehdorn HM, Held-Feindt J (2005) Reduced metastasis-suppressor gene mRNAexpression in breast cancer brain metastases. J Cancer Res Clin Oncol 131:191–198 Takei H, Powell SZ (2009) Tumor-to-tumor metastasis to the central nervous system. Neuropathology 29:303–308

J.-Q. Lu and A.W. Clark Tally PW, Laws ER Jr., Scheithauer BW (1988) Metastases of central nervous system neoplasms. Case report. J Neurosurg 68:811–816 Watanabe T, Fujisawa H, Hasegawa M, Arakawa Y, Yamashita J, Ueda F, Suzuki M (2002) Metastasis of breast cancer to intracranial meningioma: case report. Am J Clin Oncol 25:414–417 Weber J, Gassel AM, Hoch A, Spring A (2003) Concomitant renal cell carcinoma with pituitary adenoma. Acta Neurochir (Wien) 145:227–231 Wesseling P, von Deimling A, Aldape KD (2007) Metastatic tumours of the CNS. In: Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (eds) World Health Organization Classification of Tumours of the Central Nervous System. IARC Press, Lyon, pp 248–251

Chapter 5

Brain Metastases from Breast Cancer: Treatment and Prognosis Kazuhiko Ogawa, Shogo Ishiuchi, and Sadayuki Murayama

Abstract The incidences of brain metastases from breast cancer have recently increased due to prolonged survival of the patients and the development of imaging techniques. However, there is little information regarding the optimal treatments and the prognosis for these patients. Identification of subgroups of patients with substantially different outcomes is, thus, mandatory to enable tailoring of optimal therapy and to influence the stratification of future clinical trials. This article provides a brief overview of the treatment results of brain metastases from breast cancer, and identifies the factors that influence the prognosis of these patients. Keywords Brain metastases · Breast cancer · Corticosteroids · Radiosurgery · Chemotherapy · Trastuzumab

Introduction Brain metastases represent an important cause of morbidity and mortality, and are the most common intracranial tumors in adults (Loeffler et al., 1997). Despite numerous studies designed to improve treatment outcome, a median survival of 3–6 months has been reported (Borgelt et al., 1980). The risk of developing brain metastases varies according to primary tumor type, with breast cancer accounting for approximately 10–20% of all brain metastases, making breast cancer the second most common source of brain metastases (Chang and Lo, 2003).

Recently, advances in neuroimaging, such as computed tomography (CT) scans and magnetic resonance imaging (MRI) have allowed careful monitoring of cancer patients. This fact, along with the increased survival of patients, has led to more frequent and earlier detection of brain metastases, and number of published clinical reports on treatment of brain metastases has gradually increased. However, most of the large studies have been dominated by cases of brain metastases from lung cancer (Borgelt et al., 1980), and the optimal management for brain metastases from breast cancer has not yet been fully investigated. Identification of subgroups of patients with substantially different outcomes is, thus, mandatory to enable tailoring of optimal therapy and to influence the stratification of future clinical trials. This article provides a brief overview of brain metastases from breast cancer, and identifies the factors that influence the prognosis of these patients.

Treatments Corticosteroids The initial therapy should promptly start with corticosteroids, which effectively improve edema and neurologic deficits (approximately 60% of patients). In patients treated with corticosteroids alone, the median survival time has been approximately 1–2 months.

Whole Brain Radiotherapy (WBRT) K. Ogawa () Department of Radiology, University of the Ryukyus, Nishihara-cho, Okinawa, 903-0215, Japan e-mail: [email protected]

WBRT continues to be the standard of care in patients with brain metastases. The current standard external

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_5, © Springer Science+Business Media B.V. 2011

47

48

beam radiotherapy is 30Gy given in 10 fractions (Cox and Ang, 2010), however, the optimal dose and fractionation schedule have yet to be decided. In the majority of patients, the treatment of brain metastases is a palliative measure, because the primary disease is sometimes advanced, and the general condition of the patients often is poor. Concerning patients with brain metastases from breast cancer treated with WBRT, a median survival of 3–6 months has been reported (DiStefano et al., 1979; Ogawa et al., 2008). Therefore, the prognoses of patients with brain metastases from breast cancer were generally poor. It has been well-established that WBRT is effective in the palliation of symptoms resulting from intracranial metastases (Kurtz et al., 1981). The results of the first two Radiation Therapy Oncology Group (RTOG) metastatic brain studies, which mainly incorporated patients with metatatic lung and breast cancer, suggested that the administration of WBRT could improve neurologic function in 50% of patients; 70–80% of patients spent their remaining lifetime in an improved or stable neurologic state (Borgelt et al., 1980). Concerning brain metastases from breast cancer, Ogawa et al. (2008) indicated that for patients with unfavorable prognoses, palliative WBRT was effective in improving the quality of the remaining lifetime as well as those frequently seen from other primaries. For the 43 patients treated with WBRT without systemic chemotherapy, the improvements of neurological symptoms were observed in 34 of patients (79%) with the median duration of improvement was 3.7 months (range, 0.4–11.8 months). These results indicated that for breast cancer patients with unfavorable prognosis who developed brain metastases, palliative WBRT was effective in improving the quality of the remaining lifetime as well as those frequently seen from other primaries.

Surgical Resection Surgical resection followed by WBRT is considered the best option for patients with solitary and/or resectable brain metastases. Two randomized trials that excluded patients with multiple brain metastases showed that surgical resection plus radiotherapy was significantly better than radiotherapy alone (Patchell et al., 1990; Vecht et al., 1993). In approximately 10% of patients, surgical resection followed by WBRT prolongs median survival up to 1–2 years.

K. Ogawa et al.

Radiosurgery Radiosurgery provides a substitute or alternative to conventional surgery. Recently, stereotactic radiosurgery has provided local control rates equivalent to those from surgical series and has also made it possible to treat patients with surgically inaccessible or multiple lesions. Several reports have indicated that median survival of patients treated with radiosurgery has been 10–16 months (Kasad et al., 2009). In selected patients who might profit from effective local tumor control, such multimodal treatments might provide better results. The relative roles of stereotactic radiotherapy vs. WBRT in the treatment of patients with brain metastases from breast cancer remain to be undefined.

Chemotherapy Use Recently, the role of chemotherapy after WBRT has been investigated because breast cancer is a rather chemosensitive disease and not all breast cancer patients with brain metastases die due to brain metastases. Previously, it was generally considered that systemic chemotherapy could not be effective against central nervous system tumors, because the bloodbrain barrier would preclude significant penetration of antineoplastic drugs into brain tumor tissue. However, based on a number of profound studies, it now is generally agreed that the integrity of the blood-brain barrier varies widely in different regions of malignant brain tumors, resulting in an impaired or absent barrier function in large areas of the tumor vasculature (Greig et al., 1989). This suggests that the effects of chemotherapy on brain tumors may be more dependent on the chemosensitivity of the tumor than on the ability of the drugs to cross an intact blood-brain barrier. Supporting this hypothesis are the results of a clinical study in patients with breast cancer with brain metastases showing an effective response of systemic and neurologic disease to systemic chemotherapy, with a neurologic response in 50% of the patients and an overall median survival time of 5.5 months (Rosner et al., 1986). Table 5.1 indicates the summary of reports indicating the efficacy of systemic chemotherapy for brain metastases form breast cancer. Several reports indicated that systemic chemotherapy prolonged

20

14

13

6

15

Park et al.

Lee et al.

Bartsch et al.

Ogawa et al.

Park et al.

77

65

174

198

125

222

Yes 119/125 Yes 186/198 pts. Yes 174/174 pts. Yes 65/65 pts. Yes 44/77 pts.

Yes 212/222 pts. Yes 63/125 pts Yes 89/198 pts. Yes 115/174 pts Yes 11/65 pts. Yes 65/77 pts.

Yes 160/222 pts.

Yes 42/77 pts

–

–

Yes 9/125 pts –

Yes 59/222 pts.

TZ

CT

CT

CT

CT/TZ

CT/TZ

OS = overall survival; MS = median survival; CI = confidence intervals; Mo = months; CT/TZ = CT and TZ.

16

Niwinska et al.

With CT/TZ vs. Without CT/TZ Hazard ratio (95%CI) 0.58 (0.41–0.81) With CT/TZ: MS = 11.5 Mo Without CT/TZ: MS = 3.6 Mo With CT: MS = 7.8 Mo Without CT: MS = 3.6 Mo With CT: MS = 10 Mo Without CT: MS = 5 Mo With CT: 1-year OS: 82% Without CT: 1-year OS: 17% With TZ: MS = 15 Mo Without TZ: MS = 10 Mo

Table 5.1 Summary of reports indicating the efficacy of systemic chemotherapy and/or Trastuzumatib for brain metastases from breast cancer Systemic chemotherapy Trastuzumab Factors Author Reference no. No. of pts. Radiotherapy (CT) (TZ) affecting OS Prognosis

0.035

0.02

< 0.001

0.001

< 0.001

0.002

p value

5 Brain Metastases from Breast Cancer: Treatment and Prognosis 49

50

survival in these patients (Ogawa et al., 2008; Bartsch et al., 2006; Lee et al., 2008; Niwinska et al., 2010). Ogawa et al. (2008) indicated that the administration of systemic chemotherapy after radiotherapy was an independent prognostic factor for overall survival. Bartsh et al. also indicated that systemic chemotherapy significantly prolonged not only cerebral time to progression but also overall survival (Bartsch et al., 2006). A possible explanation of the favorable results seen in the patients treated with systemic chemotherapy may be that systemic chemotherapy was active against to not only the brain metastases from breast cancer but also to active extracranial diseases. Several reports indicated that most of the patients with an unfavorable prognosis died of extracranial disease, whereas a substantial subgroup with a favorable prognosis usually died of brain metastases, which determined the length of survival (Boogerd et al., 1993; Snee et al., 1985). Brain metastases from breast cancer are usually associated with active extracranial diseases, and the fact that approximately 50% of patients with brain metastases die of active extracranial disease illustrates the need for improved therapy for all metastatic diseases, as well as for brain metastases (DiStefano et al., 1979). However, at present, the role of chemotherapy in the management of brain metastases from breast cancer has been explored in a very limited number of controlled comparative trials, and therefore the class of evidence and the level of recommendations have limited applicability. Further studies are required to determine whether systematic chemotherapy could be effective in prolonging the time to progression of not only brain metastases but also active extracranial diseases.

Trastuzumab Use In approximately 25% of patients, human epithelial growth factor receptor (HER)-2 has been overexpressed, and patients with HER-2 overexpressing metastatic breast cancer are at a higher risk for brain metastases (Stemmler and Heinemann, 2008). Trastuzumab is a recombinant humanized monoclonal antibody targeted against the extracellular domain of HER-2. Recently, several reports have indicated the utility of trastuzumab for patients with brain metastases from HER-2 positive breast cancer (Park et al.,

K. Ogawa et al.

2009b; Niwinska et al., 2010; Park et al., 2009a). Table 5.1 indicates the summary of reports indicating the efficacy of trastuzumab for brain metastases from breast cancer. Park et al. indicated that the administration of trastuzumab in the treatment of HER-2 positive breast cancer resulted in prolonged time to death from the appearance of brain metastases (Park et al., 2009a). They also indicated that extracranial systemic disease control at the time of brain metastases and durable prolongation of extracranial systemic disease are the main independent prognostic factors for survival. Because of the limited information regarding the efficacy of trastuzumab to brain metastases, further studies are required to investigate the role of trastuzumab in the treatment of brain metastases from HER-2 positive breast cancer.

Prognostic Factors The prognoses of patients brain metastases are considered to correlate with combined several prognostic factors, including patient-related factors and treatment-related factors. Concerning patient-related factors, the prognostic factors that affect survival outcomes include age, KPS, number of brain metastases, tumor size, performance status, HER-2 status, number of brain metastases and the presence of extracranial metastases (Bartsch et al., 2006; Niwinska et al., 2010; Ogura et al., 2003; Boogerd 1996; Saito et al., 2006). Concerning treatment-related prognostic factors, several reports have indicated that treatment modality, such as the addition of surgery, systemic chemotherapy, endocrine therapy and targeted therapy, was an independent prognostic factor for survival (Ogawa et al., 2008; Niwinska et al., 2010; Saito et al., 2006). It has now been established clearly that achieving local control in the brain improves the survival of selected patients (Wronski et al., 1997). Ogawa et al. indicated that the prognosis of the patients treated with surgical resection and radiotherapy was significantly better than those treated with radiotherapy alone (Ogawa et al., 2008). Saito et al. also indicated that surgical resection followed by WBRT yielded better survival rates than WBRT alone (Saito et al., 2006). Recent reports have also indicated that the addition of systemic treatment can improve survival in patients with

5

Brain Metastases from Breast Cancer: Treatment and Prognosis

brain metastases. Niwinska et al. indicated that survival from brain metastases depended on the use of systemic treatment (Niwinska et al., 2010). Several authors attempt to adapt the prognostic scores to predict survival in patients with brain metastases from breast cancer, such as Recursive Partitioning Analysis (Le Scodan et al., 2007). Further studies are needed to investigate the optimal prognostic scores that can predict the prognosis of these patients. In conclusions, the prognoses for patients with brain metastases from breast cancer were generally poor, although selected patients may survive longer with intensive brain tumor treatment and/or systemic chemotherapy. Recent reports have also indicated that the addition of trasutuzumab may prolong survival in selected patients with brain metastases. For patients with unfavorable prognoses, palliative radiotherapy was effective in improving the quality of the remaining lifetime. Further prospective studies are necessary to investigate the optimal treatments for brain metastases from breast cancer.

References Bartsch R, Fromm S, Rudas M, Wenzel C, Harbauer S, Roessler K, Kitz K, Steger GG, Weitmann HD, Poetter R, Zielinski CC, Dieckmann K (2006) Intensified local treatment and systemic therapy significantly increase survival in patients with brain metastases from advanced breast cancer- a retrospective analysis. Radiother Oncol 80:313–317 Boogerd W (1996) Central nervous system metastases in breast cancer. Radiother Oncol 40:5–22 Boogerd W, Vos VW, Hart AA, Baris J (1993) Brain metastases in breast cancer; natural history, prognostic factors and outcome. J Neurooncol 15:165–174 Borgelt B, Gelber R, Kramer S, Brady LW, Chang CH, Davis LW, Perez CA, Hendrickson FR (1980) The palliation of brain metastases: final results of the first two studies by the Radiation Therapy Oncology Group. Int J Radiat Oncol Biol Phys 6:1–9 Chang EL, Lo S (2003) Diagnosis and management of central nervous system metastases from breast cancer. Oncologist 8:398–410 Cox JD, Ang KK (2010) Radiation oncology (9th edn). Rationale, Technique, Results, Philadelphia, Mosby, Elsevier, pp P859–P860 DiStefano A, Yong Yap Y, Hortobagyi GN, Blumenschein GR (1979) The natural history of breast cancer patients with brain metastases. Cancer 44:1913–1918 Greig NH (1989) Brain tumors and the blood-brain barrier. In: Neuwelt EA (ed) Implications of the Blood-Brain Barrier and its manipulation, vol 2. Plenum, New York, pp P77–P106

51

Kased N, Binder DK, McDermott MW, Nakamura JL, Huang K, Berger MS, Wara WM, Sneed PK (2009) Gamma knife radiosurgery for brain metastases from primary breast cancer. Int J Radiat Oncol Biol Phys 75:1132–1140 Kurtz JM, Gelber R, Brady LW, Carella RJ, Cooper JS (1981) The palliation of brain metastases in a favorable patient population: a randomized clinical trial by the Radiation Therapy Oncology Group. Int J Radiat Oncol Biol Phys 7: 891–895 Le Scodan R, Massard C, Mouret-Fourme E, Guinebretierre JM, Cohen-Solal C, De Lalande B, Moisson P, Breton-Callu C, Gardner M, Goupil A, Renody N, Floiras JL, Labib A (2007) Brain metastases from breast carcinoma: validation of the radiation therapy oncology group recursive partitioning analysis classification and proposition of a new prognostic score. Int J Radiat Oncol Biol Phys 69:839–845 Lee SS, Ahn JH, Kim MK, Sym SJ, Gong G, Ahn SD, Kim SB, Kim WK (2008) Brain metastases in breast cancer: prognostic factors and management. Breast Cancer Res Treat 111:523–530 Loeffler JS, Patchell RA, Sawata R (1997) Metastatic brain tumor. In Devita VT Jr., Hellman S, Rosenberg SA (eds) Cancer: principles and practice of oncology, 5th edn. Lippincott-Raven, Philadelphia, PA, pp P2523–P2536 Niwinska A, Murawska M, Pogoda K (2010) Breast cancer brain metastases: differences in survival depending on biological subtype, RPA RTOG prognostic class and systemic treatment after whole-brain radiotherapy (WBRT). Ann Oncol 21: 942–948 Ogawa K, Yoshii Y, Nishimaki T, Tamaki N, Miyaguni T, Tsuchida Y, Kamada Y, Toita T, Kakinohana Y, Tamaki W, Iraha S, Adachi G, Hyodo A, Murayama S (2008) Treatment and prognosis of brain metastases from breast caner. J Neurooncol 86:231–238 Ogura M, Mitsumori M, Okumura S, Yamauchi C, Kawamura S, Oya N, Nagata Y, Hiraoka M (2003) Radiation therapy for brain metastases from breast cancer. Breast Cancer 10: 349–355 Park BB, Uhm JE, Cho EY, Choi YL, Ji SH, Nam do H, Il Lee J, Park W, Huh SJ, Park YH, Ahn JS, Im YH (2009b) Prognostic factor analysis in patients with brain metastases from breast cancer: how can we improve the treatment outcome? Cancer Chemother Pharmacol 63: 627–633 Park YH, Park MJ, Ji SH, Yi SY, Lim DH, Nam DH, Lee JI, Park W, Choi DH, Huh SJ, Ahn JS, Kang WK, Park K, Im YH (2009a) Trastuzumab treatment improves brain metastasis outcomes through control and durable prolongation of systemic extracranial disease in HER2-overexpressing breast cancer patients. Br J Cancer 100:894–900 Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ, Maruyama Y, Kryscio RJ, Markesbery WR, Macdonald JS, Young B (1990) A randomized trial of surgery in the treatment of single metastases to the brain. N Engl J Med 322: 494–500 Rosner D, Nemoto T, Lane WW (1986) Chemotherapy induces regression of brain metastases in breast carcinoma. Cancer 58:832–839 Saito EY, Viani GA, Ferrigno R, Nakamura RA, Novaes PE, Pellizzon CA, Fogaroli RC, Conte MA, Salvajoli JV (2006) Whole brain radiation therapy in management of brain

52 metastasis: results and prognostic factors. Radiat Oncol 1:20 Snee MP, Rodger A, Kerr GR (1985) Brain metastases from carcinoma of breast: a review of 90 cases. Clin Radiol 36:365–367 Stemmler HJ, Heinemann V (2008) Central nervous system metastases in HER-2 overexpressing metastatic breast cancer: a treatment challenge. Onclogist 13:739–750

K. Ogawa et al. Vecht CJ, Haaxma-Reiche H, Noordijk EM, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR (1993) Treatment of single brain metastasis: radiotherapy alone or combined with neurosurgery? Ann Neurol 33:583–590 Wronski M, Arbit E, McCormick B (1997) Surgical treatment of 70 patients with brain metastases from breast carcinoma. Cancer 80:1746–1754

Chapter 6

Brain Metastasis in Renal Cell Carcinoma Patients Aida Loudyi and Wolfram E. Samlowski

Abstract The advent of “targeted” therapy directed against the VEGF or mTOR pathways has markedly increased the survival of patients with metastatic renal cell carcinoma (RCC) over the last 5 years. Since RCC ranks 3rd in incidence-proportion percentage of brain metastases, it is not surprising, that management of this complication has become an increasingly important clinical challenge. Some form of systematic CNS screening (e.g. brain MRI scans) appears reasonable, especially at the time of disease progression or in patients with long-standing metastatic disease. Historically, surgical resection was the preferred treatment of patients with one or two superficial brain metastases. Whole brain radiotherapy (WBRT) was used for palliation of multiple or deeper metastases. Unfortunately, these approaches achieved long-term survival only in a small fraction of patients. This is because the vast majority of RCC patients do not have isolated brain recurrences, but rather progress in the brain at the time of systemic cancer progression. Advances in gamma-knife (GK) or by linear accelerator based stereotactic radiosurgery (SRS) have resulted in apparent improvement in CNS control in patients with up to 5 metastases. Long-term control of treated lesions is achieved in the majority of patients with only a single GK/SRS treatment. Whether WBRT should be immediately added is controversial. It is currently reasonable to treat with primary GK/SRS and follow patients closely for progression with the intention of using salvage therapy when appropriate. This

A. Loudyi () Section of Melanoma, Renal Cancer and Immunotherapy, Nevada Cancer Institute, Las Vegas, NV 89135, USA e-mail: [email protected]

approach shortens the delay in starting systemic therapy. The use of “targeted” therapy following GK/SRS appears to be feasible without an increased risk of CNS complications. The use of systemic immunotherapy (e.g. IL-2) following GK/SRS to induce remissions is intriguing, as about 15% of patients achieve long term or longer survival. A two-compartment approach in patients with metastatic RCC and brain involvement employing GK/SRS treatment of the brain, followed by systemic therapy is likely to result in improved outcomes. Keywords Renal cell carcinoma · WBRT · CNS · Glucocorticosteroid · Immunotherapy · Radiotherapy

Introduction Renal cell carcinoma (RCC) represents over 90% of all tumors arising from the kidney. Jemal et al. (2009) estimated that 57,800 people in the United States were newly diagnosed with RCC in 2008. Approximately 25% of RCC patients present with metastatic disease at the time of original diagnosis and another 25–30% eventually develop metastatic disease after initial surgical resection. These metastatic disease patients account for approximately 13,000 deaths/year in the United States. The prevalence of RCC brain metastases across all stages of RCC is approximately 2–17% antemortem. This incidence-proportion percentage ranks 3rd among all tumor types, exceeded only by lung cancer and melanoma in a population-based study by Barnholtz-Sloan et al. (2004). Harada et al. (1999) found that brain metastases generally occur late in the course of RCC, at a median

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_6, © Springer Science+Business Media B.V. 2011

53

54

of 3.4 years from initial diagnosis. This was confirmed by Shuch et al. (2008), who found that patients most frequently had CNS involvement after a substantial interval (10–11 months) with metastatic disease (57% of brain metastases patients in their series). However, synchronous CNS involvement at the time of diagnosis of metastatic RCC was observed in 37 patients (27%). Isolated CNS involvement was uncommon, and occurred in only 7 patients (5.1%). These investigators noted that brain metastases most commonly were identified in patients with clear cell histology (92.7% of affected patients). The remainder of this review will be on the treatment of brain metastases in patients with clear cell RCC. Loeffler (2004) reported that a short interval between diagnosis of the primary tumor and the development of a brain metastasis is generally associated with a poorer prognosis. Current studies suggest that the median survival of RCC patients following development of brain metastasis is short, ranging from 3–7 months.

Clinical Presentation of RCC Brain Metastases In the past, brain lesions derived from clear cell RCC frequently presented with symptoms, because these metastases are highly vascular, with surrounding edema. As a result, brain metastases from RCC appear to have an increased propensity for spontaneous hemorrhage. Presenting symptoms can range from subtle personality changes, headaches, and deteriorating vision to confusion and seizures. Headaches are present in almost half of the patients and are more frequent with multiple lesions or posterior fossa localization. They are often associated with signs of increased intracranial pressure such as nausea or vomiting. Early morning headache, although specific, is not common. Therefore, any positional or pattern change of previous migraines, any symptoms suggestive of raised intracranial pressure or neurological deficit should act as warning signs to the presence of intracranial process. While cognitive impairment historically has been observed in over 30% of patients, it should be noted that cancer patients may develop confusion and altered mental status for other reasons, and metabolic encephalopathy (e.g. hypercalcemia) should

A. Loudyi and W.E. Samlowski

be excluded. Motor deficits or hemiparesis occur in 20–40% of the patients. Stroke, CNS hemorrhage, ischemic events or vascular invasion occur in another 5–10% of the cases. Seizures occur in approximately 15% of patients with brain metastases. Prophylactic therapy to prevent seizures is generally not employed, due to a high frequency of adverse effects related to seizure medications. It should be noted that neurologic signs and symptoms frequently persist following treatment, altering subsequent quality-of-life. This makes early detection of brain metastasis an important consideration. In contemporary series, such as published by Shuch et al. (2008), the more frequent use of brain scans has resulted in detection of an increasing number of small, asymptomatic CNS metastases.

Screening To date, there are no specific screening guidelines for early detection of brain metastases. Due to the low overall incidence of brain metastases in the surgically treated primary renal cancer patients, recommendations by Lam et al. (2005) and Levy et al. (1998), advise against systematic screening, unless neurologic symptoms are noted. It does seem reasonable, though, to screen patients with newly diagnosed metastatic disease for synchronous brain metastases, and to rescreen metastatic disease patients at the time of progression on systemic therapy, when a treatment change is contemplated. The incidence of brain metastases in this group is relatively high. If metastatic disease treatment with immunotherapy or anti-angiogenic drugs is anticipated, a brain CT or MRI should definitely be performed before initiation of therapy. Immunotherapy with interleukin-2 (IL-2) causes vascular leak that increases peritumoral brain edema and may exacerbate neurologic symptoms. This agent may also lower the seizure threshold. Anti-angiogenic drugs such as sorafenib and sunitinib appear to increase the risk of spontaneous hemorrhage in untreated CNS metastases in a retrospective review by Pouessel and Culine (2007). The risk of intracerebral hemorrhage following the VEGF antibody bevacizumab treatment is less clear. Besse et al. (2010), performed a safety analysis of 13,000 patients, and found a relatively small risk of intracerebral bleeding. There is additional concern that immunotherapy

6 Brain Metastasis in Renal Cell Carcinoma Patients

and “targeted therapy” agents do not penetrate the blood brain barrier well, necessitating use of other treatment modalities.

Palliative Treatment Approaches For many decades, there were no effective drugs available to treat metastatic renal cancer. From the 1980s to 2005, only alpha-interferon and IL-2 were available for treatment of metastatic renal cancer. Since 2005, the survival of patients with metastatic renal cancer has doubled or tripled due to the advent of “targeted” therapy, such as VEGF pathway or mTOR inhibitors. The treatment of brain metastasis in RCC is problematic because of the blood-brain barrier that limits penetration of immunotherapy, chemotherapy and “targeted therapy” into the CNS. The brain is considered an immunologically privileged site, limiting the effectiveness of cancer immunotherapy. Furthermore, systemic glucocorticosteroids are often used in the treatment of brain edema related to metastases. Steroids are believed to decrease the effectiveness of IL-2 based immunotherapy. Steroids also induce liver p450 enzymes and thereby increase clearance of most “targeted” therapy agents. It should be noted that palliative administration of steroids (e.g. dexamethasone) may provide transient improvement of neurologic symptoms by decreasing edema. This strategy has not been studied prospectively, and do not appear to prolong survival.

Surgical Treatment In patients with RCC brain metastases, neurosurgery may be a useful clinical option. Unfortunately, only, a minority with RCC brain metastasis are appropriate candidates. To use of neurosurgery is strongly influenced by performance status, extent of the disease in extracranial organs, the number of brain metastasis and their anatomic location. It is also desirable to consider post-operative morbidity and loss-of-function related to a planned resection. Generally patients with 1–2 superficial metastases are the best candidates. Stereotactic surgical approaches may be employed to more precisely localize lesions for resection, and the use of intraoperative functional mapping techniques

55

may help minimize post-operative sequelae. Sawaya et al. (1998), have suggested that use of modern surgical approaches may diminish risks associated with infectious risks, cerebral hemorrhage and infarct and possible neurological sequelae in 7–13% of patients. Surgery is the treatment of choice for large solitary lesions (>3 cm). Adequate control of lesions of this size is generally not achieved by radiotherapy, without considerable damage to surrounding normal brain. Shiau et al. (1997), found that the rate of local failure after treatment with SRS and neurotoxicity increases proportionally with increasing lesion size. Another potential indication for neurosurgical intervention is in patients presenting with need for immediate decompression, due to herniation. Neurosurgical resection for brain metastases is generally followed up by radiotherapy, as Patchell et al. (1998) found that this may improve CNS control.

Whole Brain Radiotherapy In the early 1970s whole brain radiation therapy (WBRT) was employed in an attempt to control unresectable CNS metastases from RCC. This approach resulted in short median survival. Predictive factors for post-WBRT survival of brain metastases were evaluated by Gaspar et al. (1997). These authors performed a recursive partitioning analysis (RPA) of 1,200 patients with brain metastases from a number of different cancers. In this analysis, three prognostic groups were characterized, based on their Karnofsky performance score (KPS), age, presence of extracranial metastasis, and control at the primary site. Class I patients had a KPS of ≥70, age ≤ 65, no extracranial metastasis, and without recurrence at the resected primary tumor site. Class II is similar, with active systemic metastases or age > 65. Patients who had systemic metastatic disease and a KPS < 70 were defined as class III. Survival was 7.1 months in RPA Class I patients, 4.2 months in Class II, and 2.3 months in III. We have not found this classification system very helpful in managing renal cancer patients with brain metastases, as the vast majority of these patients also have active systemic metastases (Class II or III), which are not well separated in this classification scheme. In a RCC specific series, Wronski et al. (1997) evaluated 119 patients treated with WBRT. These investigators found

56

a median survival of 4.4 months following treatment of single lesion and 3 months for those with multiple metastases. Over two-thirds of patients died from CNS progression. In selected patients it may be advantageous to combine surgical resection and WBRT. Patchell et al. (1990) demonstrated improved survival and local control with surgery followed by WBRT versus WBRT alone in 48 patients. Survival was 40 weeks in the surgery + WBRT group, compared to 15 weeks with WBRT alone. The rate of local recurrence was 20% in the combination group versus 52%, respectively. Vecht et al. (1993), validated the preceding results in 63 patients, describing an overall survival of 10 months in the surgery + WBRT group as compared to 6 months with WBRT alone. The major acute toxicity associated with WBRT is worsening cerebral edema. Corticosteroids are frequently administered before and during therapy to prevent this complication. Late toxicities include radiation leukoencephalopathy with cognitive deterioration and dementia. These complications are difficult to separate from tumor related changes in mental function, but are a significant concern due to the increasing survival duration in renal cell cancer patients. Other complications include normal pressure hydrocephalus, radiation necrosis and neuroendocrine dysfunction such as hypothyroidism.

A. Loudyi and W.E. Samlowski

small case series have suggested that SRS and GK may be more effective than WBRT for treatment of “radioresistant” tumors such as renal cell carcinoma and melanoma. In a recent multi-institute prospective Japanese study, Serizawa et al. (2010), obtained excellent results using GK in patients with up to 10 lesions without prophylactic WBRT. Thus, a number of patients who are not candidates for neurosurgery are able to benefit from this minimally invasive procedure. An important technical limitation of GK/SRS appears to be the amount of time patients must be immobilized during the treatment of multiple lesions. It is also possible that the biology of renal cancer in patients with oligometastatic disease (e.g., ≤3 brain lesions) may differ substantially from that in patients with many (>10) metastases. A third technology in development employs heavy particles (protons and neutrons) as a radiation source. These approaches may have the theoretical advantage of a sharp drop-off in delivered dose at a precise depth in the brain. This technology is expensive and has not been widely disseminated. Therefore the usefulness of these approaches compared to WBRT, SRS or GK is less well established, although some dramatic responses have been reported by Maor et al. (1988), in patients with renal cancer brain metastases.

Comparison of SRS and Surgery Radiosurgery Focal radiotherapy allows delivery of higher doses of radiotherapy to precisely defined volumes of brain tissue. Two delivery systems are in common use. One is the Gamma-Knife (GK) system pioneered by Leksell (1987). The other is stereotactic radiosurgery (SRS) employing a linear accelerator, as described by Alexander et al. (1995), more recently employing a tomotherapy gantry (e.g., Novalis). Sophisticated computerized dose mapping allows planning of treatment fields that minimize the dose of radiation delivered to sensitive structures (e.g. the optic nerves). While SRS and GK have not been compared head-to-head in a randomized trial, current results published in small case series appear comparable. Additionally, these techniques also allow treatment of surgically inaccessible lesions, as well as treatment of multiple lesions at the same time. A number of

Multiple studies have showed similar effectiveness between surgery and SRS for small (<2 cm) brain metastases. Alexander et al. (1995) described a 9.4 months median survival in a study of 248 patients with 421 metastatic lesions. The 1-, 2-, and 3-year local control rates were 85, 65, and 65%, respectively. In a series of 122 patients, Mehta et al. (1997), reported a local control rate of 86% with a median survival of 56 weeks. O’Neill et al. (2003), found no difference in the CNS recurrence rate following surgery or SRS. Mori et al. (1998), found that SRS for renal cell brain metastases results in CNS control in the majority of cases and was associated with minimal complications. A very recent study by Shuto et al. (2010), described the results of GK treatment versus craniotomy in 280 metastatic brain tumors, GK was found to be quite effective for growth control of brain metastases from RCC. They also stated that in selected patients,

6 Brain Metastasis in Renal Cell Carcinoma Patients

GK also helped control non-symptomatic peritumoral edema. While these results are encouraging, in the absence of prospective randomized trials directly comparing SRS to surgery, superiority of one approach over the other cannot be definitely established. Therefore, the treatment of brain metastasis in RCC patient should be multidisciplinary and individualized to each case. For lesions that are surgically difficult to access or when marked functional morbidity would be predicted, such as in the mid brain, basal ganglia and proximity to cranial nerves, SRS is the treatment of choice. This is especially true if the size of the lesion is ≤2 cm. Samlowski et al. (2008) showed that SRS alone is also an excellent option for multiple brain metastases for up to five lesions from metastatic renal cancer. Control of the CNS lesion(s) without subsequent CNS progression can be achieved in 87% of patients with a single treatment. Subsequent salvage in patients with CNS progression can also be accomplished, using surgery, additional SRS and WBRT. These results support Wowra et al. (2002) findings that radiosurgery is a minimally invasive and effective outpatient treatment for multiple brain metastases from renal cell cancer. These investigators recommended SRS as being the method of choice to control intracranial disease, especially in selected patients with limited extracranial disease. More recently some centers have begun to use SRS to the tumor bed as adjuvant therapy to surgery, rather than WBRT. This approach, described by Mathieu et al. (2007), is justified by the “radioresistant” nature of clear cell renal carcinoma to standard radiation fractions used in WBRT, and provides the advantage of deferring WBRT. Hwang et al. (2009), also showed a trend towards longer median survival with adjuvant GK to surgical resection cavities in the brain of RCC patients. A trend towards improvement in overall survival was observed with adjuvant GK, compared to WBRT.

SRS Plus Immediate WBRT Versus Deferred WBRT There is currently substantial disagreement whether GK/SRS should be followed by immediate WBRT for treatment of patients with brain metastases from

57

RCC. Historically, the reverse sequence was employed. SRS boost was generally added to larger brain metastases following WBRT. In a single institutional study, Kondziolka et al. (1999), suggested that WBRT followed by immediate radiosurgery improved CNS control in patients with two to four brain metastases. Andrews et al. (2004) randomized 333 patients with 1–3 new brain metastases from a variety of cancers to whole brain radiation therapy or WBRT followed by immediate stereotactic radiosurgery boost (RTOG 9508). Of note, only 3% of these patients in this series had brain metastases from renal cancer. There was a significant survival advantage in the group receiving WBRT plus SRS boost in patients with only a single brain metastasis (median survival time 6.5 versus 4.9 months, p = 0.04), but not in patients with 2 or 3 metastases. These authors therefore concluded that WBRT plus SRS boost should be standard treatment for patients with a single unresectable brain metastasis and that this treatment should be considered for patients with two or three brain metastases. Aoyama et al. (2006) randomized 132 patients with 1–4 brain metastases (<3 cm in size) to receive SRS alone or in combination with immediate WBRT. This study included patients with breast, lung, renal (8%) and colorectal cancer patients. The 1-year actuarial survival was 38.5% in the SRS + WBRT group versus 28.4% in the SRS group, respectively (p = 0.42). The probability of developing new CNS lesions by 12 months was 46.8% in the WBRT + SRS group and 76.4% for SRS alone group (p < 0.001). Salvage brain treatment was less frequently required in the WBRT + SRS group (n = 10) than with SRS alone (n = 29) (p < 0.001). These authors therefore recommended SRS with immediate WBRT. It should be noted that salvage SRS was able to be employed in 19 patients in SRS only group and 9 patients in the SRS + WBRT group, and that death attributed to neurologic causes occurred in a similar number of patients (19.3% versus 22.8%) (p = 0.64). Median survival in the WBRT + SRS group was not different from the group treated with SRS alone (7.5 months versus 8.0 months) (p = 0.42). In a phase II study (ECOG 6397), Manon et al. (2005), evaluated SRS alone without immediate WBRT in 31 patients with 1–3 “radioresistant” brain metastases from melanoma (45% of patients), renal cell carcinoma (45%), or sarcoma (10%). With a median potential follow-up of 32.7 months, median

58

survival was 8.3 months (95% CI, 7.4 to 12.2). Threeand 6-month intracranial failure with SRS alone was 25.8 and 48.3%, respectively. Failure within the SRS treated volume was 19.3% at 3 months and 32.2% at 6 months. Approximately 38% of patients experienced death attributable to neurologic cause. There were 3 grade 3 toxicities related to SRS. These authors concluded that delaying WBRT may be appropriate for some subgroups of patients with radioresistant tumors, but routine avoidance of WBRT should be approached judiciously. This has become an important issue, as Chang et al. (2009), found that patients who are treated with the combination of WBRT + SRS may be at greater risk of neurologic sequelae. This may prove to be an increasing obstacle, in an era where median survival of metastatic renal cancer is steadily increasing.

Multimodality Treatment of Brain Metastases The vast majority of patients with RCC brain metastases also have concurrent extracranial disease. It makes little sense, therefore to focus on CNS treatment as a stand-alone entity. Treatment of brain metastases needs to be incorporated into a multi-disciplinary treatment plan. We have hypothesized that improvement in survival can be obtained by employing aggressive GK/SRS treatment of brain metastases and subsequent systemic therapy for extracranial disease. Samlowski et al. (2008) employed a systematic approach to the detection and treatment of RCC brain metastases. All stage IV renal cell carcinoma patients were screened with a brain MRI at the time of initial evaluation or at the time of disease progression (when a change in therapy is contemplated). Follow-up brain imaging was also performed if a patient develops new neurologic symptoms. The goal was to aggressively treat any brain metastases identified, employing GK/SRS as the primary treatment modality if there were ≤5 brain metastases. If >5 metastases were present WBRT was utilized as the primary modality, with stereotactic boost to large lesions >1 cm. Patients treated with either GK/SRS or WBRT were followed with brain imaging studies at least every 2–3 months, with the goal of employing salvage treatment,

A. Loudyi and W.E. Samlowski

whenever possible. If ≤5 new CNS lesions were found, additional SRS was used. If there were >5 new lesions, WBRT was added. Palliative surgery was performed if there was a surgically accessible dominant or symptomatic lesion. The GK/SRS treatment dose was prescribed to the isodose line covering 95% of the target volume (range, 80–97%), with a planned dose based on the maximal diameter of each metastatic lesion: <2 cm, 20–24 Gy; 2–3 cm, 18 Gy; 3–4 cm, 15 Gy. This dose selection was based on RTOG protocol 9508 Andrews et al. (2004). Lesions ≥4 cm in diameter were excluded from treatment with SRS. These large lesions are currently rare, and generally reflect a failure of the screening paradigm. SRS treatment was well tolerated. Transient adverse effects were seen in a few patients, including facial edema and ecchymoses from head frame application. Two patients developed symptomatic radiation necrosis of a brain metastasis requiring palliative neurosurgical resection. Both patients had been treated with sequential SRS and WBRT. A ventriculoperitoneal shunt was required in third patient due to ventricular entrapment. This probably represented a tumor-related complication. Only a single SRS treatment was required to achieve adequate CNS control in 27/32 patients (84%), while 5 patients received 2 SRS treatments (16%). Failure was generally due to the development of new metastases that could be treated for salvage with additional SRS or WBRT. The overall CNS control rate was 60% at 1 year and 32% at 2 years. Median survival from the time of the diagnosis of brain metastases was 10.1 months (CI 6.4–14.8). One-year and 3-year survival from the onset of brain metastases was 43 and 16% respectively. Six patients remain alive without CNS progression with follow-up of 9.6–55.7 months from diagnosis of brain metastases. There may be a trend for improved survival in patients with fewer brain metastases. Fourteen patients with single brain metastases had a median survival of 8.9 months compared to 5.4 months for the 18 patients treated with multiple metastases (p = 0.19). This observation will require confirmation in larger patient samples. Local control data was available for 22 patients and for 42 of 71 treated brain metastases. There were only 6/42 failures at the site of an SRS-treated metastasis. The probability of local control in individual treated lesions was 86, 74 and 59% at 1, 2, and 3 years

6 Brain Metastasis in Renal Cell Carcinoma Patients

following SRS, respectively. Durable CNS control was achieved in 91% of evaluable lesions with a single SRS treatment. Only 8 of 22 patients eventually developed new brain metastases. This was always in the context of extracranial tumor progression. Median time to CNS progression of any treated lesion following SRS was 15 months (confidence interval, CI = 8.7 months –∞). Once the patient recovers from GK/SRS and is tapered from any glucocorticosteroids (which are not always necessary), systemic therapy should be strongly considered. We found that the majority of RCC brain metastasis patients could tolerate treatment for extracranial disease. Patients at our institution have been treated with high dose IL-2, or a variety of investigational antiangiogenic agents as part of clinical trials. It should be noted that concurrent treatment with GK/SRS and anti-angiogenic agents has not yet been adequately tested and could result in increased toxicity. Bevacizumab, 10 mg/kg i.v. every 2 weeks, was administered to 10 patients following SRS. Sunitinib (SU11248), given 50 mg daily p.o. for 4 weeks on/2 weeks off, was administered to 4 patients. Sorafenib, 400 mg p.o. twice daily, was given to 3 patients. Thalidomide was administered to two patients following SRS treatment. These agents were tolerated without apparent increase in toxicity, or additional neurologic symptoms. There were no episodes of intracerebral bleeding. In univariate analysis, surgical resection, addition of WBRT to SRS, antiangiogenic therapy (bevacizumab, sunitinib, sorafenib, or thalidomide), or the number of CNS lesions (one vs ≥2 metastases) did not appear to affect survival very much, although the sample size was small. Administration of immunotherapy following SRS appeared to significantly prolong survival (p = 0.0007). A significant component of this effect appeared to be due to IL-2 treatment following SRS (p = 0.04). A few of these patients currently remain in long-term remissions of RCC. It is also interesting to note that known prognostic factors published by Motzer et al. (1999), and Bukowski et al. (2004), had an overriding effect on survival even in the presence of RCC brain metastasis. The patients with the best outcome had “good risk” disease, one brain metastasis and were treated with immunotherapy (high-dose IL-2). Due to the small and retrospective nature of our study, these findings will need to be confirmed in a larger prospective clinical trial.

59

Conclusions and Recommendations Due to the substantial risk of brain metastases over time in patients with metastatic renal cancer, periodic screening seems important. This practice has led to more early detection of asymptomatic brain metastasis, which are more easily controlled. Our impression is that this leads to less neurologic morbidity and a higher level of function. Diagnosis of brain metastases presents a therapeutic challenge in many aspects. Treatment of renal cancer patients with brain metastases may be considered a bicompartmental problem (brain and systemic metastases), as current VEGF and mTOR directed agents have poor activity in the CNS. Treatment of these lesions with WBRT can delay systemic therapy, which is needed in the vast majority of patients. The use of single dose-fraction GK/SRS approaches facilitates the ability to move more rapidly to systemic treatment. Surgery followed by GK/SRS to treatment bed or alternatively WBRT may play a role in controlling superficial, solitary brain metastases. While there is a concern about drug-induced morbidity in patients with RCC brain metastases, our experience shows that antiangiogenic and immunologic therapies can be safely added after these lesions are treated with radiotherapy. Close post-treatment surveillance (e.g. every 2–3 months) is critical, as patients who progress in the CNS may be candidates for additional salvage therapy (surgery, GK/SRS or WBRT). Management of these lesions in a multidisciplinary fashion should be strongly encouraged, with input from neurosurgery, oncology, and radiation therapists. At present the longevity of patients with metastatic renal cancer overall, and the high local control achievable with a single GK/SRS session has encouraged deferral of WBRT as a salvage treatment. This approach is predicated on the ability to perform close follow-up of patients with brain MRI scans. Further improvements in CNS treatment approaches are needed, as a percentage of patients eventually progress at treated CNS sites (40% by 3 years), or less commonly, with new brain metastases (about 32%). More effective systemic disease control, including agents that produce complete remissions of RCC, is also needed, as this represents another important site of eventual mortality. Nevertheless, we are encouraged

60

that long-term survivorship (>3 years) of renal cancer patients with brain metastases may currently be achievable in ∼15% of patients by use of GK/SRS, followed by systemic immunotherapy. It is likely that broader application of combined modality treatment will result in further improvements in the outcome of RCC metastatic to the brain.

References Alexander E 3rd, Moriarty TM, Davis RB, Wen PY, Fine HA, Black PM, Kooy HM, Loeffler JS (1995) Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases. J Natl Cancer Inst 87:34–40 Andrews DW, Scott CB, Sperduto PW, Flanders AE, Gaspar LE, Schell MC, Werner-Wasik M, Demas W, Ryu J, Bahary JP, Souhami L, Rotman M, Mehta MP, Curran WJ Jr. (2004) Whole brain radiation therapy with or without stereotactic radiosurgery boost for patients with one to three brain metastases: phase III results of the RTOG 9508 randomised trial. Lancet 363:1665–1672 Aoyama H, Shirato H, Tago M, Nakagawa K, Toyoda T, Hatano K, Kenjyo M, Oya N, Hirota S, Shioura H, Kunieda E, Inomata T, Hayakawa K, Katoh N, Kobashi G (2006) Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial. J Am Med Assoc 295:2483–2491 Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, Sawaya RE (2004) Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. J Clin Oncol 22: 2865–2872 Besse B, Lasserre SF, Compton P, Huang J, Augustus S, Rohr UP (2010) Bevacizumab safety in patients with central nervous system metastases. Clin Cancer Res 16:269–278 Bukowski RM, Negrier S, Elson P (2004) Prognostic factors in patients with advanced renal cell carcinoma: development of an international kidney cancer working group. Clin Cancer Res 10:6310S–6314S Chang EL, Wefel JS, Hess KR, Allen PK, Lang FF, Kornguth DG, Arbuckle RB, Swint JM, Shiu AS, Maor MH, Meyers CA (2009) Neurocognition in patients with brain metastases treated with radiosurgery or radiosurgery plus wholebrain irradiation: a randomised controlled trial. Lancet Oncol 10:1037–1044 Gaspar L, Scott C, Rotman M, Asbell S, Phillips T, Wasserman T, McKenna WG, Byhardt R (1997) Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys 37:745–751 Harada Y, Nonomura N, Kondo M, Nishimura K, Takahara S, Miki T, Okuyama A (1999) Clinical study of brain metastasis of renal cell carcinoma. Eur Urol 36:230–235 Hwang SW, Abozed MM, Hale A, Eisenberg RL, Dvorak T, Yao K, Pfannl R, Mignano J, Zhu JJ, Price LL, Strauss GM, Wu JK (2009) Adjuvant Gamma Knife radiosurgery following

A. Loudyi and W.E. Samlowski surgical resection of brain metastases: a 9-year retrospective cohort study. J Neurooncol 98:77–82 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ (2009) Cancer statistics, 2009 CA Cancer J Clin 59:225–249 Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC (1999) Stereotactic radiosurgery plus whole brain radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 45:427–434 Lam JS, Shvarts O, Leppert JT, Pantuck AJ, Figlin RA, Belldegrun AS (2005) Postoperative surveillance protocol for patients with localized and locally advanced renal cell carcinoma based on a validated prognostic nomogram and risk group stratification system. J Urol 174:466–472; discussion 472; quiz 801 Leksell DG (1987) Stereotactic radiosurgery. Present status and future trends. Neurol Res 9:60–68 Levy DA, Slaton JW, Swanson DA, Dinney CP (1998) Stage specific guidelines for surveillance after radical nephrectomy for local renal cell carcinoma. J Urol 159:1163–1167 Loeffler JS (2004) Can combined whole brain radiation therapy and radiosurgery improve the treatment of single brain metastases? Nat Clin Pract Oncol 1:12–13 Manon R, O‘Neill A, Knisely J, Werner-Wasik M, Lazarus HM, Wagner H, Gilbert M, Mehta M (2005) Phase II trial of radiosurgery for one to three newly diagnosed brain metastases from renal cell carcinoma, melanoma, and sarcoma: an Eastern Cooperative Oncology Group study (E6397). J Clin Oncol 23:8870–8876 Maor MH, Frias AE, Oswald MJ (1988) Palliative radiotherapy for brain metastases in renal carcinoma. Cancer 62:1912– 1917 Mathieu D, Kondziolka D, Cooper PB, Flickinger JC, Niranjan A, Agarwala S, Kirkwood J, Lunsford LD (2007) Gamma knife radiosurgery in the management of malignant melanoma brain metastases. Neurosurgery 60:471–481; discussion 481–472 Mehta M, Noyes W, Craig B, Lamond J, Auchter R, French M, Johnson M, Levin A, Badie B, Robbins I, Kinsella T (1997) A cost-effectiveness and cost-utility analysis of radiosurgery vs. resection for single-brain metastases. Int J Radiat Oncol Biol Phys 39:445–454 Mori Y, Kondziolka D, Flickinger JC, Logan T, Lunsford LD (1998) Stereotactic radiosurgery for brain metastasis from renal cell carcinoma. Cancer 83:344–353 Motzer RJ, Mazumdar M, Bacik J, Berg W, Amsterdam A, Ferrara J (1999) Survival and prognostic stratification of 670 patients with advanced renal cell carcinoma. J Clin Oncol 17:2530–2540 O‘Neill BP, Iturria NJ, Link MJ, Pollock BE, Ballman KV, O‘Fallon JR (2003) A comparison of surgical resection and stereotactic radiosurgery in the treatment of solitary brain metastases. Int J Radiat Oncol Biol Phys 55:1169–1176 Patchell RA, Tibbs PA, Regine WF, Dempsey RJ, Mohiuddin M, Kryscio RJ, Markesbery WR, Foon KA, Young B (1998) Postoperative radiotherapy in the treatment of single metastases to the brain: a randomized trial. J Am Med Assoc 280:1485–1489 Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ, Maruyama Y, Kryscio RJ, Markesbery WR, Macdonald JS, Young B (1990) A randomized trial of surgery in the treatment of single metastases to the brain. N Engl J Med 322:494–500

6 Brain Metastasis in Renal Cell Carcinoma Patients Pouessel D, Culine S 2007. High Frequency of Intracerebral Hemorrhage in Metastatic Renal Carcinoma Patients with Brain Metastases Treated with Tyrosine Kinase Inhibitors Targeting the Vascular Endothelial Growth Factor Receptor. Eur Urol Samlowski WE, Majer M, Boucher KM, Shrieve AF, Dechet C, Jensen RL, Shrieve DC (2008) Multidisciplinary treatment of brain metastases derived from clear cell renal cancer incorporating stereotactic radiosurgery. Cancer 113: 2539–2548 Sawaya R, Hammoud M, Schoppa D, Hess KR, Wu SZ, Shi WM, Wildrick DM (1998) Neurosurgical outcomes in a modern series of 400 craniotomies for treatment of parenchymal tumors. Neurosurgery 42:1044–1055; discussion 1055–1046 Serizawa T, Hirai T, Nagano O, Higuchi Y, Matsuda S, Ono J, Saeki N (2010) Gamma knife surgery for 1–10 brain metastases without prophylactic whole-brain radiation therapy: analysis of cases meeting the Japanese prospective multiinstitute study (JLGK0901) inclusion criteria. J Neurooncol 93:163–167 Shiau CY, Sneed PK, Shu HK, Lamborn KR, McDermott MW, Chang S, Nowak P, Petti PL, Smith V, Verhey LJ, Ho M, Park E, Wara WM, Gutin PH, Larson DA (1997) Radiosurgery for brain metastases: relationship of dose and pattern of

61 enhancement to local control. Int J Radiat Oncol Biol Phys 37:375–383 Shuch B, La Rochelle JC, Klatte T, Riggs SB, Liu W, Kabbinavar FF, Pantuck AJ, Belldegrun AS (2008) Brain metastasis from renal cell carcinoma: presentation, recurrence, and survival. Cancer 113:1641–1648 Shuto T, Matsunaga S, Suenaga J, Inomori S, Fujino H (2010) Treatment strategy for metastatic brain tumors from renal cell carcinoma: selection of gamma knife surgery or craniotomy for control of growth and peritumoral edema. J Neurooncol Vecht CJ, Haaxma-Reiche H, Noordijk EM, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR et al. (1993) Treatment of single brain metastasis: radiotherapy alone or combined with neurosurgery? Ann Neurol 33:583–590 Wowra B, Siebels M, Muacevic A, Kreth FW, Mack A, Hofstetter A (2002) Repeated gamma knife surgery for multiple brain metastases from renal cell carcinoma. J Neurosurg 97:785–793 Wronski M, Maor MH, Davis BJ, Sawaya R, Levin VA (1997) External radiation of brain metastases from renal carcinoma: a retrospective study of 119 patients from the M. D. Anderson Cancer Center. Int J Radiat Oncol Biol Phys 37:753–759

Chapter 7

Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain Naveen Sankhyan, Suvasini Sharma, and Sheffali Gulati

Abstract Inflammatory myofibroblastic tumors (IMT) are rare tumors of unknown etiology, composed of proliferating myofibroblasts and accompanying lymphoplasmacytic infiltration. They are most commonly seen in the lung, but may rarely occur in other organs. The authors review the current literature of coexisting inflammatory myofibroblastic tumors of lung and brain. Keywords Plasma cell granuloma · Inflammatory granuloma · Anaplastic lymphoma kinase · Pseudotumor · ALK

Introduction Inflammatory myofibroblastic tumor is a quasineoplastic lesion consisting of inflammatory cells and myofibroblastic spindle cells (Scott et al., 1988). The myofibroblast is ubiquitous in soft tissues and its precise role in any given lesion may vary considerably, ranging from “innocent” bystander through being a reactive stromal component, to representing the primarily proliferating cell type (Petridis et al., 2004). Myofibroblastic tumors can be broadly classified into 4 main groups: reactive lesions, benign tumors (neoplastic, reactive, hamartomatous), the locally aggressive fibromatosis and sarcomas showing myofibroblastic differentiation (Fletcher, 1998). Inflammatory

N. Sankhyan () Division of Pediatric Neurology, Department of Pediatrics, All India Institute of Medical Sciences, New Delhi, India e-mail: [email protected]

myofibroblastic tumor was first described in the lungs by Brunn in 1939. It mainly affects children and young adults. Lungs or the upper respiratory tract are the most common sites of these tumors (Bahadori and Liebow, 1973; Lawson et al., 2010). However they have also been reported to occur in other areas like the skin (Vadmal and Pellegrini, 1999), spleen (Herman et al., 1994) thyroid (Kojima et al., 2009), breast (Chetty and Govender, 1997), uterus (Rabban et al., 2005), kidney (Dogra and Bhatt, 2009), heart (Hartyánszky et al., 2000), liver (Tang et al., 2010), pancreas (Dagash et al., 2009) retroperitoneum (Koirala et al., 2010), gastrointestinal tract (Sanders et al., 2001), mediastinum (Sugiyama and Nakajima, 2008) and central nervous system (Trojan et al., 2001). Other terms which have been used to describe this entity include plasma cell granuloma, fibroxanthoma, xanthogranuloma, pseudolymphoma, inflammatory pseudotumor, and inflammatory myofibrohistiocytic proliferation. The term inflammatory myofibrohistiocytic proliferation was suggested by Tang et al. (1990) to overcome the inadequacy and inaccuracy of the conventional designations of plasma cell granuloma and inflammatory pseudotumor; especially as plasma cells are not always a major feature of this lesion.

Histopathology It is a benign solid tumor composed mainly of spindleshaped cells and has a chronic inflammatory component consisting of plasma cells, lymphocytes and occasional histiocytes. Absence of anaplasia, intermixture of lymphocytes and plasma cells among spindle cells, and paucity of mitotic cells leads to the diagnosis

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_7, © Springer Science+Business Media B.V. 2011

63

64

of inflammatory myofibroblastic tumor (Karnak et al., 2001). Major subgroups have been identified among inflammatory tumors not affecting the CNS, including a predominant myxoid/vascular pattern resembling granulation tissue, a compact spindle cellular and a hypocellular fibrotic pattern in nonpulmonary, and an organizing pneumonia pattern with central hyalinization, a fibrous histiocytoma-like pattern, and a lymphoplasmacytic pattern in pulmonary inflammatory tumors (Coffin et al., 1995; Matsubara et al., 1988). Inflammatory myofibroblastic tumor in the CNS is a neoplasm similar to its soft tissue counterpart and should be distinguished from the histologically similar, nonneoplastic inflammatory pseudotumors. Inflammatory pseudotumor, a term used in the past interchangeably with IMT is now best considered a separate entity. Inflammatory pseudotumor, a term synonymous with plasma cell granuloma and lymphoid hyperplasia, is a chronic inflammatory lesion of uncertain etiology that more often affects the central than the peripheral nervous system. Recent studies with stringent diagnostic criteria indicate that inflammatory pseudotumors lack anaplastic lymphoma kinase expression, with such expression being a feature of inflammatory myofibroblastic tumors (Swain et al., 2008).

Etiopathogenesis The myofibroblasts, fibroblasts, and histiocytes in inflammatory myofibroblastic tumor are probably derivatives of primitive mesenchymal cells which are widely distributed in the body and thus may contribute to the ubiquitous occurrence of inflammatory myofibroblastic tumor (Tang et al., 1990). The true etiology of inflammatory myofibroblastic tumors remains unknown. Some authors believed this tumor was a low grade fibrosarcoma with inflammatory cells. The propensity of these tumors to be locally aggressive, sometimes multifocal, and to progress occasionally to a true malignant tumor supported the idea (Narla et al., 2003). Others hypothesized that inflammatory myofibroblastic tumor represented an immunological response to an infectious or non-infectious agent. This idea was supported by the presence of clinical and laboratory signs of systemic inflammation in 15–30% of previous cases (Coffin et al., 1998). In part the

N. Sankhyan et al.

confusion arose because; historically both neoplastic and nonneoplastic processes were combined as inflammatory pseudotumors. Recent reports suggest that inflammatory myofibroblastic tumors are neoplastic (Swain et al., 2008). Evidence favouring the neoplastic nature was presented by demonstrating a clonal population harboring the abnormal Anaplastic Lymphoma Kinase (ALK) receptor tyrosine kinase expression due to aberrations on chromosomal locus 2p23 (Clarke et al., 2005; Su et al., 1998; Griffin et al., 1999). Hence, it may be reasonable to classify a lesion as inflammatory myofibroblastic tumor in the presence of typical histopathological features and ALK expression. But the nosology of lesions with similar morphology but absent ALK expression or those with unclear morphology but ALK expression remains unsettled. Human herpesvirus 8 (HHV-8) DNA sequences have been found in adult pulmonary inflammatory myofibroblastic tumors and presence of Ebstein-Barr Virus (EBV) has been reported in splenic and hepatic inflammatory myofibroblastic tumors, suggesting the role of these viruses in inflammatory myofibroblastic tumor pathophysiology (Gomez et al., 2000; Arber et al., 1998). However, Tavora et al. (2007), in a study of 20 patients with pulmonary inflammatory myofibroblastic tumors did not find HHV-8 related transcripts. Similar was the experience of Swain et al., in six patients with central nervous system inflammatory myofibroblastic tumors (Swain et al., 2008). It is clear that a lot remains to be known about the pathogenesis of this tumor.

Clinical Features The clinical presentation depends on the site, extent and spread of the tumor. Virtually any organ may be involved and exceptionally, even bone marrow spread has been described (Hagenstad et al., 2003). Extrapulmonary inflammatory myofibroblastic tumor affects a younger population of patients, with a predilection for the first and second decades; this is in contrast to a peak incidence in mid-adulthood for the pulmonary form (Coffin et al., 1995). Cough, dyspnea and hemoptysis are the usual presenting features of pulmonary inflammatory myofibroblastic tumors (Kim et al., 2002). Patients with inflammatory myofibroblastic tumors may also have systemic features

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain

including fever, growth impairment, iron deficiency anemia, thrombocytosis, elevated erythrocyte sedimentation rate and hypergammaglobulinemia (Coffin et al., 1998).

Intracranial Inflammatory Myofibroblastic Tumors Intracranial inflammatory myofibroblastic tumors may be either isolated or associated with a primary inflammatory myofibroblastic tumor in another organ, usually lung, mesentery or mediastinum. Intracranial inflammatory myofibroblastic tumor may present with seizures, focal neurological deficits, features of raised intracranial pressure (headache, vomiting, papilledema), cranial nerve deficits and in case of sellar/suprasellar location, with endocrine dysfunction (Makino et al., 1995). Spinal cord and meningeal involvement, suggesting hematogenous spread have also been described (Narla et al., 2003). Greiner et al. reviewed 38 published cases (16 female, 22 male, mean age 34 years) of intracranial inflammatory myofibroblastic tumor. The major complaints were headache (52.6%), seizures (26.3%), visual disturbance (36.8%), ataxia (13.2%), paresis (15.8%) and diabetes insipidus (7.9%). Intracranial inflammatory myofibroblastic tumor was found incidentally on cranial imaging in 5.3% of the patients (Greiner et al., 2003).

Co-existence of Pulmonary and Brain Inflammatory Myofibroblastic Tumor On reviewing the literature we found only eight case reports of combined pulmonary and intracranial inflammatory myofibroblastic tumors (Tang et al., 1990; Chan et al., 1994; Le Marc’hadour et al., 1995; Malhotra et al., 1991; Greiner et al., 2003; Petridis et al., 2004; Jeba et al., 2008; Sharma et al., 2009) (Table 7.1). The most frequent pulmonary symptoms in these patients included cough and hemoptysis. The intracranial symptoms included headache, seizures and vomiting (Table 7.1). Four of the reported cases had multifocal brain lesions (Chan et al., 1994; Sharma et al., 2009; Malhotra et al., 1991; Jeba et al., 2008). Five of these eight patients were children. Interestingly, though the lung lesions in these cases

65

were solitary large masses, the brain lesions were small, multifocal and discrete, suggesting the possibility of hematogenous dissemination. In all but one (Petridis et al., 2004), the discovery of the lung lesion either preceded the brain lesion (duration ranging from 5 months to 5 years) or was simultaneous; suggesting that the primary pathology originated in the lung. The case described by Petridis et al. (2004) was unusual for two other aspects as well: the brain lesions were hemorrhagic, resembling cavernous hemangiomas; and secondly unlike all the other patients who had a benign course, this patient had a fulminant course and died. It is unclear whether these co-occurring lesions represent a metastasis or a multifocal, exaggerated inflammatory response to some unidentified etiological agent(s). The metastasis hypothesis would also have to account for selective predilection for brain involvement, and the slow growth and overall indolent nature of these tumors. Also, no nuclear atypia or abnormal mitoses have been seen. Demonstration of ALK expression at both sites would favor a metastatic etiology; however this has not been performed in any of the reported cases to date.

Cranial Imaging Findings in Inflammatory Myofibroblastic Tumor Both CT and MRI demonstrate well-circumscribed, solid, homogenous masses with mild-to-moderate enhancement (Makino et al., 1995). In the brain, both solitary and multifocal brain lesions have been reported. MRI reveals iso-to-hypointense T2 lesions which have occasionally demonstrated hemorrhage within the tumors. Calcification and necrosis are rare. A lack of mobile protons, due to the densely fibrotic background of these lesions, may account for the T2 hypointensity and weaker enhancement on MR images (Makino et al., 1995). T2 weighted MRI sometimes reveal interdigitation with the adjacent cortex which is in line with observation of pathological changes in the neighboring cerebral tissue (Greiner et al., 2003). Although inflammatory myofibroblastic tumors are well circumscribed lesions, lymphoplasmacytic inflammation, neuronal loss, and reactive gliosis can be found within the adjacent cortex (Tekkök et al., 2000).

8

30

13

Le Marc’hadour et al. (1995)

Greiner et al. (2003)

20

M

M

M

M

Age (year) Sex 13 M

Chan et al. (1994)

Malhotra et al. (1991)

Author, year Tang et al. (1990)

Cough

Asymptomatic

Cough

Pulmonary symptoms/ signs Asymptomatic, Examinationdull percussion note Hemoptysis

Seizure

Recurrent headaches

Seizure

Seizures

Intracranial symptoms Intermittent throbbing headache

4 years

Simultaneous

2 years

5 years

Interval between detection of lung & brain lesion Simultaneous

Solid lesion in Lt lower lobe

3 masses in Rt lower lobe

Night sweats, malaise

Nil

Developed lesion in opposite lung, and in pneumonectomy space NA

No progression on 9 year follow up

Remarks No regression, but no progession on 3 year follow up.

Both lung & brain No recurrence on lesions 4.5 year follow successfully up resected

NA

Lung: lesion resection Brain: conservative Lung: pneumonectomy Brain: conservative

Systemic features Treatment Anemia, Hyper- Radiotherapy gamma globulinemia

Mass extending NA from lateral aspect of the left cavernous sinus to the tentorium cerebelli and the infratemporal fossa Round lesion in Rt Fever frontal lobe

Chest imaging Brain imaging findings findings large mass with Single 1.5 cm dense enhancing calcification in lesion in Lt Rt lung parietooccipital region Circumscribed 3 lesions in Lt mass in Rt temporal, Rt & upper lobe Lt parietal regions Large calcified 3 round lesions in mass in Rt Rt frontal, Lt lower lobe parietal, & Lt frontal

Table 7.1 Summary of cases reported with co-existing inflammatory myofibroblastic tumors of lung and brain

66 N. Sankhyan et al.

M

Sharma et al. (2009)

Headache, vomiting

Intracranial symptoms Headache, vomiting, diplopia, seizures

Cough, dyspnea Seizure

Cough

Pulmonary symptoms/ signs Hemoptysis, cough

M-Male, NA-Not available, Lt-Left, Rt-Right

10

M

Age (year) Sex 29 M

Jeba et al. (2008) 12

Author, year Petridis et al. (2004)

Table 7.1 (continued)

1 year

5 months

Interval between detection of lung & brain lesion Brain lesion preceded ling lesion by 8 months

Large, lobulated, calcified mass in Lt Lung

Fever

Lung lesionpneumonectomy Brain lesion-surgery, radiotherapy, single agent chemotherapy (doxorubicin) Lung lesioninoperable Brain lesionchemotherapy (Methotrexate, 6-Mercaptopurine)

Systemic features Treatment Fever Lung & brain lesions resected

2 round lesions in Anemia Rt occipital lobe

Brain imaging findings Hemorrhagic lesions mimicking cavernomas in Lt frontal, Rt occipital, Lt parietal region Well defined mass 2 solid enhancing lesion with irregular mass central lesions in Rt & calcification in Lt frontal lobes Lt lower lobe

Chest imaging findings Cystic lesion with central calcification in Lt lung

Brain lesions reduced in size, Lung lesion static at 2 year follow up

NA

Remarks Had metastasis in spinal cord. Died due to pulmonary complications

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain 67

68

Differential Diagnosis The differential diagnosis is broad because of the contrast-enhancing and sometimes multifocal pattern of CNS involvement by inflammatory myofibroblastic tumor. It includes infections as well as neoplasms. Because of the similar appearance on brain imaging infective granulomas particularly tuberculomas are important differentials. This is particularly true in asian countries where tuberculomas are common causes of enhancing ring lesions in the brain. CNS tumors, like plasmacytoma and meningioma containing plasma cells and lymphocytic infiltration may mimic CNS inflammatory myofibroblastic tumor (Le Marc’hadour et al., 1995). Both these tumors occur in older individuals. The presence of Bence-Jones proteins in the urine and the M spike on serum electrophoresis confirms the diagnosis of plasmacytoma. Histopathologically, meningiomas may be differentiated from inflammatory myofibroblastic tumor by the presence of meningeal whorls and positive epithelial membrane antigen (EMA) staining in the former (Greiner et al., 2003). Rarely, Histiocytosis X and Wegener’s granulomatosis may cause central nervous system involvement mimicking inflammatory myofibroblastic tumor. Multifocal central nervous system inflammatory myofibroblastic tumor may also mimic metastasis or CNS lymphoma.

Treatment and Prognosis The biological potential of inflammatory myofibroblastic tumors is highly variable. Complete surgical resection, if possible, is the treatment of choice for most inflammatory myofibroblastic tumors. Radiation therapy has been tried in unresectable cases but it is associated with significant morbidity. Response to steroids is unpredictable, some patients have shown improvement while some have even shown tumor progression (Narla et al., 2003). The other treatment modalities include immunomodulation (Cyclosporin A) and combination chemotherapy. Chemotherapeutic agents like methotrxate, azathioprine, chlorambucil, cyclophosphamide, ifosfamide, vincristine and dactinomycin have been tried in these patients without much success (Karnak et al., 2001). Anti-inflammatory agents including non-steroidal

N. Sankhyan et al.

anti-inflammatory drugs and infliximab (anti-TNF alpha binding antibody) have also been tried with variable success (Su et al., 2000; Germanidis et al., 2005). Shah and McClain described a 14-year old girl with recurrence of intracranial inflammatory myofibroblastic tumor after radiotherapy and steroids. This patient was treated successfully with a combination of methotrexate and 6-mercaptopurine, given for 2 years (Shah and McClain, 2005). This therpay was also successful in the case reported by Sharma et al. (2009). Till further specific therapy emerges this possibly beneficial simple anti-metabolite regimen can be used in children with in-operable IMT affecting the brain (Shah and McClain, 2005).

Illustrative Case A 10-year-old boy was symptomatic since the age of five. He was first seen at age of seven with history of recurrent episodes of cough and breathlessness. These episodes were occasionally associated with fever and wheeze. He had received multiple courses of oral antibiotics, bronchodilators and steroids; on which he would show improvement for a few days. He had received antitubercular treatment without any symptomatic relief for 1 year prior to presentation. Examination revealed a thin built, afebrile child with mild pallor. Chest examination revealed tracheal shift to left, vesicular breath sounds with reduced intensity and dull percussion noted on left side. Chest X-ray revealed left opaque hemithorax with ipsilateral mediastinal shift. Pleural tap was dry. Contrast enhanced CT of the chest showed a large (6×6×7 cm) lobulated mass with large chunks of calcification closely abutting diaphragm, left heart border, and left hilum with narrowing of left lower lobe bronchus (Fig. 7.1). A few enlarged pretracheal and precarinal lymph nodes were also noted. The patient underwent left thoracotomy which revealed opaque appearance of left lung. Only the upper lobe and upper part of lower lobe were inflatable. A large calcified mass involving hilum of left lung, adherent to pulmonary artery, pericardium and left dome of diaphragm was noted. The mass was non-resectable and a wedge biopsy was taken. Biopsy showed fascicles of spindle cells with variable collagenized and myxoid regions and variable density of chronic inflammatory cells. There were abundant

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain

69

thin walled curvilinear blood vessels. Proliferation of fibroblastic and myofibroblastic cells with collagen production and interspersed infiltrate of lymphocytes and plasma cells was evident. The spindle cells were immunopositive for vimentin and smooth muscle actin

and were negative for CD 34 immunostain (Fig. 7.2). Findings were consistent with inflammatory myofibroblastic tumor. A trial of oral prednisolone was given for 2 months but the patient did not show any improvement. Parents refused the options of radiotherapy and chemotherapy. The child continued to be in follow up, receiving conservative treatment including antibiotics for chest infections and bronchodilators and inhaled steroids for wheeze. One year after diagnosis, he developed one episode of afebrile generalized tonic clonic seizure. There was no history of headache, vomiting, limb weakness, cranial nerve deficits, personality or behavior changes. CECT and MR scan showed two well-defined rounded enhancing lesion in right occipital region with minimal perifocal edema (Fig. 7.3). He was given antitubercular treatment for 1 year and phenytoin. He remained well for the next 1 year then he had an episode of breakthrough seizure. A follow up MRI of brain was obtained which revealed persistence of lesions (Fig. 7.4 A & B). CT abdomen and bone marrow aspirate did not show further systemic spread. The chest pathology and the nature of the brain imaging findings suggested a diagnosis of multifocal

Fig. 7.2 Low power photomicrographs showing fascicles of spindle cells with variable collagenized and myxoid regions and variable density of chronic inflammatory cells. Abundant thin walled curvilinear blood vessels are identified (a & b, H & E × 40). There is proliferation of fibroblastic and myofibroblastic

cells with collagen production and interspersed infiltrate of lymphocytes and plasma cells (c, H & E × 400). The spindle cells are immunopositive for vimentin (d, IHC; DAKO × 100) and smooth muscle actin (e, IHC; DAKO × 100) and are negative for CD 34 immunostain (f, IHC; DAKO × 40)

Fig. 7.1 CECT thorax – Axial section at mid-thorax level shows a mixed-density mass lesion with area of calcification in left hemithorax

70

N. Sankhyan et al.

Fig. 7.3 CECT scan (a) show well-defined enhancing lesion in right occipital region with minimal perifocal edema. On MR scan, the lesion is cortical-subcortical in location, isointense on T1-weighted (b) and T2-weighted images (c) with mild perilesional edema and shows homogenous enhancement

following gadolinium administration (e, f). Another smaller T2-isointense lesion (d) with moderate perifocal edema is seen in right postcentral gyrus, which also shows homogenous enhancement (e)

inflammatory myofibroblastic tumor. The child was started on oral methotrexate and 6-mercaptopurine. At 2 years follow up, he was asymptomatic and tolerated the therapy well. A magnetic resonance imaging of brain obtained at the end of 2 years shows significant reduction in the size of the lesions (Fig. 7.4 C & D). Comment: Successful treatment of recurrent of intracranial IMT with a combination of methotrexate and 6-mercaptopurine has been reported in the past. This treatment was based on the premise that IMT’s form a histopathological spectrum with Rosai-Dorfman disease (inflammatory sinus histiocytosis with massive lymphadenopathy) (Govender and

Chetty, 1997). Horneff et al. (1996) recommended this regimen for Rosai-Dorfman disease. Our case emphasizes the possible beneficial therapeutic effect of this simple anti-metabolite regimen in children with inoperable IMT affecting the brain. Spontaneous reduction in size cannot be excluded, however, the fact the lesions persisted the same for 2 years before starting treatment makes this unlikely. [The definitive version of this case was published in J. Child Neurol.; 24(10): Oct/2009: 1302–1306. SAGE Publications Ltd. SAGE Publications, Inc. 2010, All rights reserved. ©] In conclusion, the co-occurrence of intracranial and pulmonary inflammatory fibroblastic tumors is

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain

71

Fig. 7.4 T1 weighted post contrast axial MRI images of the brain at the time of initial presentation (a and b) shows focal enhancing lesions representing metastases in the right temporooccipital region (a) and in right frontal and paritel lobes (b).

Follow up T1 weighted post contrast MRI images at the same levels after 2 years of antimetabolite therapy (c and d) shows marked regression in the size and enhancement of the lesions with minimal residual lesions

a very rare phenomenon. It is unclear whether these lesions represent a metastasis or a multifocal, exaggerated inflammatory response to some unidentified etiological agent(s). Diagnosis is suggested by magnetic resonance imaging and confirmed by

histopathology. ALK expression needs to be studied in these patients to understand the tumor biology better. Treatment of choice remains surgery. When this is not feasible, simple anti-metabolite regimens may benefit.

72

References Arber DA, Weiss LM, Chang KL (1998) Detection of EbsteinBarr virus in inflammatory pseudotumor. Semin Diagn Pathol 15:155–160 Bahadori M, Liebow AA (1973) Plasma cell granulomas of the lung. Cancer 31:191–208 Brunn H (1939) Two interesting benign lung tumors of contradictory histopathology: remarks on the necessity of maintaining chest tumor registry. J Thorac Surg 9: 119–131 Chan YF, White J, Brash H (1994) Metachronous pulmonary and cerebral inflammatory pseudotumor in a child. Pediatr Pathol 14:805–815 Chetty R, Govender D (1997) Inflammatory pseudotumor of the breast. Pathology 29:270–271 Clarke AJ, Jacques TS, Galloway MJ, Thom M, Kitchen ND, Plant GT (2005) ALK positive inflammatory myofibroblastic tumour of the pineal region. J Clin Pathol 58:981–983 Coffin CM, Humphrey PA, Dehner LP (1998) Extrapulmonary inflammatory myofibroblastic tumor: a clinical and pathological survey. Semin Diagn Pathol 15:85–101 Coffin CM, Watterson J, Priest JR, Dehner LP (1995) Extrapulmonary inflammatory myofibroblastic tumor: a clinicopathologic and immunohistochemical study of 84 cases. Am J Surg Pathol 19:859–872 Dagash H, Koh C, Cohen M, Sprigg A, Walker J (2009) Inflammatory myofibroblastic tumor of the pancreas: a case report of 2 pediatric cases–steroids or surgery? J Pediatr Surg 44:1839–1841 Dogra VS, Bhatt S (2009) Inflammatory pseudotumor of the kidney. Ultrasound Q 25:69–70 Fletcher CD (1998) Myofibroblastic tumors: an update. Verh Dtsch Ges Pathol 82:75–82 Germanidis G, Xanthakis I, Tsitouridis I, Zaramboukas T, Kiskinis D, Konstantaras C, Miliaras S, Sirakos T, Pagkalos E (2005) Regression of myofibroblastic tumor of the gastrointestinal tract under infliximab treatment. Dig Dis Sci 2:262–265 Gomez-Roman JJ, Ocejo-Vinyals G, Sánchez-Velasco P, Nieto EH, Leyva-Cobián F, Val-Bernal JF (2000) Presence of human herpesvirus-8 DNA sequences and overexpression of human IL-6 and cyclin D1 in inflammatory myofibroblastic tumor. Lab Invest 80:1121–1126 Govender D, Chetty R (1997) Inflammatory pseudotumor and Rosai-Dorfman disease of soft tissue: a histological continuum? J Clin Pathol 50:79–81 Greiner C, Rickert CH, Möllman FT, Rieger B, Semik M, Heindel W, Wassmann H (2003) Plasma cell granuloma involving lung and brain. Acta Neurochir 145:1127–1131 Griffin CA, Hawkins AL, Dvorak C, Henkle C, Ellingham T, Perlman EJ (1999) Recurrent involvement of 2p23 in inflammatory myofibroblastic tumors. Cancer Res 59:2776–2780 Hagenstad CT, Kilpatrick SE, Pettenati MJ, Savage PD (2003) Inflammatory myofibroblastic tumor with bone marrow involvement. A case report and review of the literature. Arch Pathol Lab Med 127:865–867 Hartyánszky IL, Kádár K, Hubay M (2000) Rapid recurrence of an inflammatory myofibroblastic tumor in the right ventricular outflow tract. Cardiol Young 10:271–274

N. Sankhyan et al. Herman TE, Shackelford GD, Ternberg JL, Dehner LP (1994) Inflammatory myofibroblastic tumor of the spleen: report of a case in an adolescent. Pediatr Radiol 24:280–282 Horneff G, Jürgens H, Hort W, Karitzky D, Göbel U (1996) Sinus histiocytosis with massive lymphadenopathy (RosaiDorfman disease): response to methotrexate and mercaptopurine. Med Pediatr Oncol 27:187–192 Jeba J, John S, Backiyanathan S, Christopher DJ, Kurian S (2008) Inflammatory pseudotumour of the lung with sarcomatous brain metastasis. Eur J Cancer Care 17: 412–414 Karnak I, Senocak ME, Ciftci AO, Ca˘glar M, Bingöl-Kolo˘glu M, Tanyel FC, Büyükpamukçu N (2001) Inflammatory myofibroblastic tumor in children: diagnosis and treatment. J Pediatr Surg 36:908–912 Kim JH, Cho JH, Park MS, Chung JH, Lee JG, Kim YS, Kim SK, Shin DH, Choi BW, Choe KO, Chang J (2002) Pulmonary inflammatory pseudotumor – a report of 28 cases. Korean J Intern Med 17:252–258 Koirala R, Shakya VC, Agrawal CS, Khaniya S, Pandey SR, Adhikary S, Pathania OP (2010) Retroperitoneal inflammatory myofibroblastic tumor. Am J Surg 199:e17–e19 Kojima M, Suzuki M, Shimizu K, Masawa N (2009) Inflammatory pseudotumor of the thyroid gland showing prominent fibrohistiocytic proliferation. A case report. Endocr Pathol 20:186–190 Lawson SL, Azoumah DK, Lawson-Evi K, N’Timon B, Savi de Tove HM, Yehouessi-Vignikin B, Kpemissi E (2010) Inflammatory myofibroblastic tumour of nose and paranasal sinuses in a little girl of 7-year-old. Arch Pediatr 17:34–37 Le Marc’hadour F, Lavielle JP, Guilcher C, Brambilla E, Brichon PY, Lebas JF, Charachon R, Pasquier B (1995) Coexistence of plasma cell granulomas of lung and central nervous system. Pathol Res Pract 191:1038–1045 Makino K, Murukami M, Kitano Y, Ushio Y (1995) Primary intracranial plasma-cell granuloma. A case report and review of the literature. Surg Neurol 43:374–378 Malhotra V, Tatke M, Malik R, Gondal R, Beohar PC, Kumar S, Puri V (1991) An unusual case of Plasma cell granuloma involving lung and brain. Ind J Cancer 28:223–227 Matsubara O, Tan-Liu NS, Kenney RM, Mark EJ (1988) Inflammatory pseudotumors of the lung: progression from organizing pneumonia to fibrous histiocytoma or to plasma cell granuloma in 32 cases. Hum Pathol 19:807–814 Narla LD, Newman B, Spottswood SS, Narla S, Kolli R (2003) Inflammatory pseudotumor. Radiographics 3:719–729 Petridis AK, Hempelmann RG, Hugo HH, Eichmann T, Mehdorn HM (2004) Metastatic low-grade inflammatory myofibroblastic tumor in the central nervous system of a 29-year old male patient. Clin Neuropathol 23:158–166 Rabban JT, Zaloudek CJ, Shekitka KM, Tavassoli FA (2005) Inflammatory myofibroblastic tumor of the uterus: a clinicopathologic study of 6 cases emphasizing distinction from aggressive mesenchymal tumors. Surg Pathol 29:1348–1355 Sanders BM, West KW, Gingalewski C, Engum S, Davis M, Grosfeld JL (2001) Inflammatory pseudotumor of the alimentary tract: clinical and surgical experience. J Pediatr Surg 36:169–173 Scott L, Blair G, Taylor G, Dimmick J, Fraser G (1988) Inflammatory pseudotumors in children. J Pediatr Surg 23:755–758

7 Coexsistence of Inflammatory Myofibroblastic Tumor in the Lung and Brain Shah MD, McClain KL (2005) Intracranial plasma cell granuloma: case report and treatment of recurrence with methotrexate and 6-mercaptopurine. J Pediatr Hematol Oncol 27:599–603 Sharma S, Sankhyan N, Kalra V, Garg A, Gupta SD, Agarwala S, Das P (2009) Inflammatory myofibroblastic tumor involving lung and brain in a 10-year-old boy: a case report. J Child Neurol 24:1302–1306 Su LD, Atayde-Perez A, Sheldon S, Fletcher JA, Weiss SW (1998) Inflammatory myofibroblastic tumor: cytogenetic evidence supporting clonal origin. Mod Pathol 11:364–368 Su W, Ko A, O’Connell TX, Applebaum H (2000) Treatment of pseudotumors with nonsteroidal anti-inflammatory drugs. J Pediatr Surg 35:1635–1637 Sugiyama K, Nakajima Y (2008) Inflammatory myofibroblastic tumor in the mediastinum mimicking a malignant tumor. Diagn Interv Radiol 14:197–199 Swain RS, Tihan T, Horvai AE, Di Vizio D, Loda M, Burger PC, Scheithauer BW, Kim GE (2008) Inflammatory myofibroblastic tumor of the central nervous system and its relationship to inflammatory pseudotumor. Hum Pathol 39:410–419

73

Tang L, Lai EC, Cong WM, Li AJ, Fu SY, Pan ZY, Zhou WP, Lau WY, Wu MC (2010) Inflammatory myofibroblastic tumor of the liver: a cohort study. World J Surg 34:309–313 Tang TT, Segura AD, Oechler HW, Harb JM, Adair SE, Gregg DC, Camitta BM, Franciosi RA (1990) Inflammatory myofibrohistiocytic proliferation simulating sarcoma in children. Cancer 65:1626–1634 Tavora F, Shilo K, Ozbudak IH, Przybocki JM, Wang G, Travis WD, Frank TJ (2007) Absence of human herpesvirus-8 in pulmonary inflammatory myofibroblastic tumor: immunohistochemical and molecular analysis of 20 cases. Mod Pathol 20:995–999 Tekkök IH, Ventureyra EC, Jimenez CL (2000) Intracranial plasma cell granuloma. Brain Tumor Pathol 17:97–103 Trojan A, Stallmach T, Kollias S, Pestalozzi BC (2001) Inflammatory myofibroblastic tumor with CNS involvement. Onkologie 24:368–372 Vadmal MS, Pellegrini AE (1999) Inflammatory myofibroblastic tumor of the skin. Am J Dermatopathol 21:449–453

Chapter 8

Breast Cancer and Renal Cell Cancer Metastases to the Brain Jonas M. Sheehan and Akshal S. Patel

Abstract Metastatic tumors from systemic cancers comprise the majority of brain tumors. These tumors most commonly originate from the lung and breast skin or kidney cancers. There are numerous similarities regarding the pathophysiology of these entities. Primary tumors spread to the central nervous system in a stepwise and highly concerted fashion. Tumor particles must breach the containment organ and subsequently travel via the blood stream to lodge within the brain. Trans-endothelial migration allows cells to penetrate the blood brain barrier. Tumor emboli must then survive and grow within the brain microenvironment with local nutrient supply and glial support. Breast and renal tumors utilize an array of similar molecular signals to accomplish these tasks: such as the Ras/MEK/MAPK and PI-3K/Akt pathways. This may explain why these primaries have a preponderance to metastasize to the brain. As antioncogenic therapies become more effective and patients with systemic cancers are afforded longer survival, cerebral invasion becomes more common and more important to overall management. Based on recent scientific data, emerging therapeutic targets for brain metastasis include vascular endothelial growth factor (VEGF), the epidermal growth factor receptor (EGFR) family, chemokines and mTOR. Keywords Tumor · Metastasis · BBB · Chemokines · VEGF · EGFR · mTOR

J.M. Sheehan () Department of Neurosurgery, Penn State Milton S. Hershey Medical Center, Hershey, PA 17033-0850, USA e-mail: [email protected]

Introduction Brain metastases are common entities, comprising more than half of all brain tumors. (Ranasinghe and Sheehan, 2007). As systemic therapies improve, patients with cancer are living longer, and the number of cancer patients with metastatic tumors to brain is increasing. By the time of death, nearly 40% of all patients with cancer will have involvement of the central nervous system with metastases (Schouten et al., 2002). Metastases are either discovered on a staging evaluation of a patient with a known primary disease or less commonly during investigation of new neurologic deficits with occult tumor. The incidence of metastatic disease is not accurately known due to underreporting or inaccurate diagnoses. Epidemiological studies estimate the incidence of brain metastases up to 11 per 100,000 of the population (Sawaya, 2004). However autopsy data places the incidence even higher. The largest and most comprehensive study from the Memorial Sloan-Kettering Cancer Center found an intradural involvement rate of 20% (Gavrilovic and Posner, 2005). After age 60, the incidence appears to climb greater than 30 per 100,000 population (Sawaya, 2004). Though breast cancer has an obvious sex predilection in regards to primary tumor, renal cell cancer does not. Gender does not seem to affect independently the occurrence of cerebral metastases. The cumulative incidence of cerebral metastases from breast cancer at 1 and 5 years after initial diagnosis of primary is 1 and 5% respectively. For renal cell cancer at 1 and 5 years the cumulative incidence is 5.2 and 9.8%, respectively (Schouten et al., 2002). The cerebral hemispheres are involved in 80–85% of cases of metastases, and the cerebellum or

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_8, © Springer Science+Business Media B.V. 2011

75

76

posterior-fossa in 15% of the time. Metastases to the brain stem comprise only 1–5% of cases of all cancers. The location of metastatic disease in the brain seems to follow brain mass and blood distribution. However, renal cell cancer metastasized disproportionately to the posterior fossa. In more than half of the cases, multiple metastatic lesions are present on initial imaging. When single metastases are found, they are rarely “solitary,” i.e. the only detectable cancer in the body. There has been renewed focus on the treatment of brain metastases in recent years. Systemic therapies have increased patient longevity and thus increased the likelihood of late tumor occurrence and growth in the brain. Advances in neuroimaging have permitted more accurate surveys of the CNS so that smaller lesions earlier in their growth can be detected and addressed. Accounting for 10–15% of metastatic brain tumors, breast cancer is the second most common tumor likely to metastasize to brain, following lung cancer. Renal cell cancer is the fourth most common and comprises 7–10% of CNS metastases (Eichler and Plotkin, 2008). Why these entities possess an affinity for the central nervous system is a topic of intense investigation. Although simplistic, if one were to think of tumor cells as parasites, the brain provides a perfect nesting ground. The blood-brain-barrier (BBB) excludes most of our current systemic therapies and thus affords a “sanctuary” to treatment. Although not “tight” enough to exclude invasion of tumor cells, the BBB manages to successfully block entry of potential therapeutic agents into the CNS. There have been many allegories to falsely suggest that invasion of brain tissue occurs via penetration of the blood brain barrier by tumor cells. Recently the paradigm has shifted and it appears that the CNS may preferentially sustain tumor cells within this sanctuary. We will discuss the pathobiology of breast cancer and renal cell cancer in this regard.

Presenting Signs and Symptoms Cerebral metastases present with either focal or diffuse signs. Cancer staging, with neuroimaging workup, will detect asymptomatic lesions in up to 50% of patients with locally metastatic disease. The two major determinants of clinical presentation are size and anatomic location of the tumor. It is often telling whether there is torpid progression or rapid neurologic decline.

J.M. Sheehan and A.S. Patel

Although many brain tumors may hemorrhage, renal cell metastases have a greater tendency to hemorrhage and rapidly compress and affect the surrounding parenchyma. The hemorrhage event often produces sharp neurologic deterioration or a severe sudden headache. Of the common general complaints, headache and focal neurological deficit (such as weakness) are the foremost, each being present in 50% of patients. Seizures are the initial presenting event in 15% of patients. More subtly, metastases may present with cognitive or behavioral changes. Localizing symptoms such as hemiparesis, aphasias and visual disturbances will initially broaden a clinician’s differential to include stroke. If we combine three studies on this topic and collect information from over one thousand patients: 31% presented with headaches, 24% with weakness, 19% with seizures, 5% with visual changes and 9% were asymptomatic (Nussbaum et al., 1996; Zimm et al., 1981). Stephen Paget (1855–1926), an English surgeon, presented a paper analyzing 735 autopsy cases of breast cancer, with the addition of additional cases from the literature, and argued that the distribution of metastases was not due to chance, but rather suggested “the best work in pathology of cancer is done by those who. . . are studying the nature of the seed. . .” (micrometastases), and the “observations of the properties of the soilmay also be useful”. Thus the “seed-soil” paradigm of metastases was born. In apparent contrast to this came a slightly later theory proposed by James Ewing (1846–1943). This American pathologist believed that cancer cells migrated across the body based on the fluid mechanics of blood circulation. Breast cancer, for example, can burst into the arterial circulation and be filtered by the lung were it can flourish before making an onward journey to other organs. Renal cell cancer, on the other hand, can slip into Batson’s venous plexus and make its way directly to the brain. A persistently patent foramen ovale between the right and left atrium of the heart would change the dynamics of spread.

Imaging This century has seen tremendous progress in imaging technology. From an inauspicious beginning in skull

8

Breast Cancer and Renal Cell Cancer Metastases to the Brain

plain X-ray films and pneumoencephalograms, modern neuroimaging modalities such as Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), and MR Spectroscopyhave become useful adjuncts the clinical history and neurologic exam. The subtle and insipid growths of micro-metastases are impossible to detect with physical assessment alone. Current state-of-the-art magnetic resonance technology, along with its sub-modalities (such as spectroscopy, perfusion, and diffusion techniques), is the neuroimaging modality of choice for metastatic CNS disease. The classic imaging findings for cerebral metastases include: a well-circumscribed round lesion in the region of the gray-white junction and significant vasogenic edema (Fig. 8.1). If lesions are multiple this may also hint at metastatic disease versus a primary neoplasm. Renal cell cancer metastases are highly cellular and will tend to bleed, or in other words present as hemorrhagic metastases. In differentiating metastases versus primary neoplasms of the brain, imaging of the peritumoral tissue provides the most useful diagnostic information. Magnetic resonance spectroscopy (MR SPECT) patterns for metastatic lesions usually display elevated signals for lipid (products of brain destruction), choline (cell membrane marker) and lactate (product of anaerobic glycolysis), with a decreased N-aceyl-aspartate or NAA (neuronal marker) peak. Primary brain tumors are more likely to infiltrate surrounding brain compared to metastatic lesions. Voxel calculations of the peritumoral region can thus aid in differentiating primary glioma versus metastases (Law et al., 2002). More recently, radiologists have begun to examine the relative cerebral blood volume of the edema that surrounds tumor and use this calculated property to again differentiate between high-grade glioma and metastatic disease (Hakyemez et al., 2010).

General Tumor Biology The etymology of metastases derives from the Greek word for displacement. There are two theories, as mentioned above, that explain how this displacement or spillage of tumor particles results in cerebral involvement. Both breast and renal cell cancer utilize similar

77

Fig. 8.1 Gadolinium enhanced T1 weighted MRI axial images. Above: Metastatic Breast Carcinoma. Bottom: Metastatic Renal Carcinoma

mechanisms to leave the primary site, spread and settle within the brain in a metachronous fashion. Metastatic emboli often wedge near the temporoparieto-occiptal junction, in the distribution of the middle cerebral artery. This distribution correlates with blood flow and mirrors the location of other embolic events in the CNS such as ischemic strokes. They rarely involve the cortical surface but rather embed at the junction where arterioles transition into capillaries.

78

This wedging concept is rather simplistic given that we now know that there is a complex interaction between cancer cells and the endothelium. Trans-endothelial migration of tumor particles involves cell adhesion molecules (CAMs) such as integrin and destructive cytokines. Regulators of the cell cytoskeleton such as Rho GTPases facilitate integrin based trans-endothelial migration of cancer cells. Therapies that inhibit this extravasation process may slow the development of metastases (Miles et al., 2008).

Blood Brain Barrier Physiology The blood brain barrier (BBB) is perhaps the most complex biologic interface in nature. It invests 99% of the CNS and acts like a gatekeeper to ensure highly specific flow in and out of the CNS cellular environment. The “Rule of Five” highlights favorable characteristics for permeability through the BBB: a molecular weight <500, an octanol-to-water partition coefficient less than five, hydrogen bond donors less than five and hydrogen bond acceptors less than ten (Palmieri et al., 2007). At the choroid plexus, where cerebrospinal fluid is produced, the floating milieu of CSF is protected by choroid plexus epithelial cells which form the bloodCSF barrier (BCSFB). Both the BBB and the BCSFB are the architecture that tumor emboli must over-come to embed themselves within the brain parenchyma. The critical buttresses of the BBB are the interendothelial tight junctions. They comprise of transmembrane claudins, occludin and junctional adhesion molecules; as well as cytoplasmic anchors such as zonula occludens, cingulin, AF-6 and 7H6. All are linked to the actin cytoskeleton like chandeliers in a cellular cathedral. The endothelial layer has no fenestrae and a very low pinocytic rate, both properties unique to the CNS. Communication and molecular exchange between endothelial cells is vital to ensure the physical and metabolic permeability barrier. Beneath the endothelium lies the basement membrane, studded with pericytes and ensheathed by the pseudo-podia of astrocytes (Engelhardt and Sorokin, 2009). Claudins play an important role in human cancer as evidenced in recent gene and protein expression profiling experiments. In addition to their role for sealing cellular sheets, they also maintain cell polarity

J.M. Sheehan and A.S. Patel

and signal transduction. Interestingly, there is overlap in the claudin family subtype expression in breast and renal cell carcinomas. Claudins 3 and 4 appear to be differentially over-expressed in these tumors. The significance of this is unknown. Current evidence points to an over-expression rather than dysfunction of claudin proteins. One can hypothesize that a derangement of claudin function may contribute to loss of primary tumor cohesiveness and potential for metastasis. How primary tumor claudin expression affects the likelihood of metastasis is unknown but future investigations regarding tumor cell mobility and transendothelial leakage may be productive (Ouban and Ahmed, 2010).

Glial Microenvironment “Beware the Greeks, even when they bear gifts. . .” A line from Virgil’s Aeneid reminds us that tumor cells have a way to trick our physiology into letting them cohabit the brain. Fitzgerald et al. (2008) examined a model of metastatic breast cancer and noted that numerous glial cells reside within the center of the tumor mass. In vitro cultures of these reactive glia appeared to induce a fivefold increase in tumor cell proliferation. The authors concluded that reactive glia may support tumor cells in the same manner as they do normal neuronal tissue; or more colorfully put, tumor cells seem to hijack the brain’s infrastructure in order to provide a permissive microenvironment for growth and invasion.

Chemokines Chemokines are a type of cytokine that mediate chemotaxis. Their receptors are found in varying degrees within the tumor microenvironment. Tumor cells secrete these molecules, as do leukocytes, epithelial cells, stromal cells and endothelial cells. 48 different chemokines have been characterized (Ali and Lazennec, 2007). Systemic dissemination of cancer cells is a highly orchestrated process. Some tumors, such as breast and renal cell cancers, prefer the brain as a sanctuary. Who is the conductor? Chemokines appear to play an organizing role to the process of

8

Breast Cancer and Renal Cell Cancer Metastases to the Brain

tumor proliferation, metastases and angiogenesis. One essential fulcrum of chemokine function is chemotactic gradient establishment. These gradients determine the metastastic potential for breast and renal cell tumors (Balkwill, 2004). The CXCL12 chemokine and its receptor CXCR4 appear to be key mediators of metastases in both breast and renal cell cancers (Lai et al., 2004). We will discuss later the signaling cascades for both breast and renal cell cancer metastases but suffice to say that CXCR4 expression is regulated by and related to both disease processes (Ali and Lazennec, 2007; Muller et al., 2001; Reckamp et al., 2008).

Vascular Endothelial Growth Factors For tumor cells to grow beyond 2 mm in size they must create or co-opt new vasculature. The role of nutrient delivery cannot be understated as blood vessel density is correlated with higher incidence of metastasis (Kato et al., 2003). Angiogenesis is the process of blood vessel growth and appropriation. It is under both positive and negative control in order that neovascularization occurs during the appropriate circumstance. The prime growth factor controlling angiogenesis is vascular endothelial growth factor (VEGF). This molecule has a number of roles but the most important are: stimulation of endothelial cell mitogenesis, promotion of endothelial cell survival, degradation of extracellular matrix and increasing vascular permeability. The VEGF gene is upregulated in both breast and renal cell cancer; and tumor cells themselves appear to produce this factor rather than endothelial cells. Indeed, tumor growth is angiogenesis dependent and VEGF is key to this process (Folkman, 1995). Hence a number of neutralizing molecules have been created to target the VEGF or angiogenic system to retard the proliferation of cancer at primary and metastatic sites. There are a host of anti-angiogenic agents currently under study for the treatment of breast and renal cell cancer. These include direct inhibitors of VEGF, tyrosine kinase inhibitors and epidermal growth factor (EDGF) inhibitors. Most are now at phase II or III trials to compare new versus standard of care or older treatment regimens (Heldwein et al., 2009; Palmieri et al., 2006).

79

For the treatment of metastatic breast and renal cell cancer, Lapatinib (Glaxo Smith Kline) has shown early promise. Lapatinib is an oral selective and reversible inhibitor of both EDGF and Her-2 tyrosine kinase. For patients who have Her-2 (see below) positive breast cancer brain metastases, two recent phase II studies have shown that Lapatinib has some modest effect in reduction of brain tumor volume (Lin et al., 2008; Lin et al., 2009). For metastatic renal cell cancer, a phase III trial using Lapatinib as second-line versus hormone therapy demonstrated longer progression free survival in those patients with EGFR super-expression (Ravaud et al., 2008). Multitargeted tyrosine kinase inhibitors such as Sorafenib (Nexavar; Bayer Pharmaceuticals Corporation) and Sunitinib (Sutant; Pfizer, Inc.) have also shown promise for the treatment of advanced renal cell and breast cancer (Grandinetti and Goldspiel, 2007; Lee et al., 2008). Interestingly, brain MR images after treatment with these kinase inhibitors may show an apparent worsening of disease progression. It is thought that this reversible and contradictory change is due to chemotherapy related effects on vascular permeability and impairment of the blood brain barrier (Hill et al., 2009).

Breast Cancer Over 6 million life years (Disability adjusted Life Years or DALY) are lost per annum to this cancer worldwide. Nearly one third of breast cancer patients will experience brain involvement during the course their disease. This estimate is based on the incidence of clinically overt metastases coupled with the incidence of asymptomatic lesions (Barnholtz-Sloan et al., 2004; Lin et al., 2004; Miller et al., 2003; Patanaphan et al., 1988; Tsukada et al., 1983). Metastasis is a late event, often after the primary focus has already spread to lung, liver and bone (Weil et al., 2005). A large retrospective study found that the median age of patients at diagnosis is 45 years (range 22–77 years) and with improvements in systemic therapy there is a median lag time of 30.9 months, between initial diagnosis and the discovery of brain pathology (Altundag et al., 2007). Breast cancer is the foremost cause of cancerrelated death in women in the USA, and out of all solid tumors has the highest affinity for the leptomeninges.

80

A number of risk factors have been studied in terms of patient characteristics. It appears that youth and the absence of the estrogen receptor (ER-negative) on the primary tumor are the most consistent risk factors for the development of brain metastases (Evans et al., 2004; Hicks et al., 2006). The median age of breast cancer patients with brain metastases is about 5 years younger than those without (Cheng and Hung, 2007). ER-negative tumors have almost twice the selective tropism for brain compared to ER-positive counterparts (Chang et al., 2003). The epidermal growth factor receptor Erb-2 (or Her2/neu) is pivotal in the pathogenesis of breast cancer seeding to the CNS. Its major role in this circumstance is receptor signaling protein kinase activity. The Her-2/neu gene is a proto-oncogene on chromosome 17. Its product is a transmembrane receptor kinase. A fifth of breast cancers express elevated levels of Her-2/neu, as the gene is amplified and the transcriptional promoter is upregulated. (Meric-Bernstam and Hung, 2006). The receptor system activates myriad intracellular signals. The major and most studied are: the mitogen-activated protein kinase pathway (Ras/MEK/MAPK), the extracellular signal-related pathway (Ras/MEK/Erk), the phosphatidylinositol-3OH kinase pathway (PI-3K/Akt) and vascular endothelial growth factor (VEGF) pathway (Fig. 8.2). These pathways play an orchestral role to induce growth, spread and survival of breast cancer metastases via modulation in gene expression. The introduction of trastuzumab (Herceptin; Hoffman-La Roche, Basel, Switzerland), a humanized monoclonal antibody receptor inhibitor of Her 2-neu, heralded a new era in breast cancer treatment. It afforded a way to inhibit tumor growth for a select population of breast cancer patients (Slamon et al., 2001). The pharmacology of trastuzumab is largely unknown within the cerebrospinal fluid compartment and as a monoclonal antibody there is no evidence that this drug can penetrate the blood brain barrier.

Fig. 8.2 A schema to highlight the interplay of factors involved in breast cancer metastases

J.M. Sheehan and A.S. Patel

That being said, recent literature appears conflicted whether Her-2 positive tumors pose a greater risk of brain metastases (Eichler et al., 2008; Stemmler and Heinemann, 2008; Tham et al., 2006). The reason for this apparent discrepancy is not known. The International Breast Cancer Study Group performed a retrospective analysis of over 3,000 patients and concluded that the cumulative 10-year risk of developing cranial metastases is doubled for Her-2 positive tumor primaries (Pestalozzi et al., 2006). Nevertheless, another retrospective study performed by Lai et al. showed no difference in rate of brain metastases between patients receiving trastuzumab and those that did not. One has to differentiate between trastuzumab treatment studies and Her-2 status studies. It has been known that tumor biology changes during the course of disease and we should not compare a biologic property with its proposed therapy. This is especially the case when examining our contradicting retrospective data on this subject. There are three culpable explanations for the association of Her-2 overexpression and cerebral metastases: 1: Trastuzumab has poor blood brain barrier penetration and thus in an ineffective treatment (Grossi et al., 2003), 2: With greater systemic control, patients receiving therapy will unmask cerebral disease as they live longer, 3: It is possible that Her-2 gene products target cellular pathways that facilitate cerebral invasion. The unmasking phenomenon is vital to consider given that patients are likely to succumb to brain metastases related death. Indeed, CNS screening of asymptomatic patients is becoming increasingly popular.

Renal Cell Cancer Renal cell cancer is the third leading cause of uroepithelial cancer related deaths worldwide. It tends to metastasize to the brain in up to 10% of patients with primary disease (Sheehan et al., 2003). It has been documented that the incidence of renal cell carcinoma has increased by 130% in the last 50 years (Jemal et al., 2008). The majority of studies find the clear cell subtype as the most common histopathologic entity (70–80%), while the rest consist of papillary, chromophobe, collecting duct and medullary subtypes. In the last couple of decades, interferon alpha (IFN-a) and interleukin 2(IL2) therapies have become

8

Breast Cancer and Renal Cell Cancer Metastases to the Brain

the mainstay of metastatic renal cell cancer treatment. Nevertheless, there have been many theoretical risks proposed with the use of IL2 therapy for the specific treatment of brain metastases. These include: increases in peritumoral edema, increased risk of cerebral hemorrhage, treatment-related confusion, and reduction in seizure threshold (Shuch et al., 2008). Newer targeted medications are being developed with tumor pathogenesis in mind to provide better outcomes for these patients. During the early part of the last century, giants of their respective fields, Eugene Von Hippel, Arvid Lindau, Harvey Cushing and Percival Bailey recognized the syndrome now named Von-Hippel-Lindau (VHL). This syndrome affects 1 in 35,000 individuals and has autosomal dominant transmission, with a defect localized on chromosome 3. The VHL gene is defective in 100% of familial cases of renal cell carcinoma and in 60–70% of sporadic cases; thus most of our insights concerning the molecular pathogenesis of renal cell cancer arise from our study of patients with VHL disease. The current molecular schema of VHL pathogenesis involves dis-regulation of VHL gene product. In the absence of VHL protein, hypoxic conditions will induce an accumulation of large amounts of hypoxia inducible factor (HIF). This in turn promotes oncogenesis through a similar cascade involved in breast cancer. (Heldwein et al., 2009). Vascular endothelial growth factor (VEGF), platlet-derived growth factor (PDGF), and epidermal growth factor (EGF) activate genetic expression via Ras/MEK/MAPK and PI-3K/Akt. HIF activates the transcription of these growth factors. The mammalian target of rapamycin (mTOR) has also been shown to be involved in conjunction with the growth factor pathways noted above. The mTOR kinase protein controls a variety of growth related cellular functions. Activation of protein kinase B (Akt) results in phosphorylation of mTOR and there is crosstalk with a myriad of signaling proteins involved in oncogenesis. These affect actin cytoskeleton remodeling and transcription. mTOR may also play a role in HIF expression though association with proteins involved in its translation. (Kapoor, 2009). Although the PTEN tumor suppressor gene mutations have never been implicated in renal cell cancer, the overlap in trafficking signals with mTOR suggests that inhibition of mTOR could halt tumor

81

Fig. 8.3 A schema to highlight the interplay of factors involved in renal cell cancer metastases

progression on multiple levels (Fig. 8.3). Knowledge of mTOR and research in VHL patients has led to the development mTOR inhibitors for use in renal cell caricinoma: Sirolimus (Rapamune; Wyeth, Madison, NJ, USA), Everolimus (RAD001; Novartis Pharmaceuticals, St. Louis, MO, USA) Temsirolimus (analogue of rapamysin). Everolimus is an oral drug that acts as a direct inhibitor of mTOR. A recent phase III trial of 410 patients, whose disease was classified as advanced due to progression while on VEGF inhibitors, found that the median progression free survival was significantly prolonged compared to placebo (4 versus 1.9 months). (Motzer et al., 2008). Temsirolimus, on the other hand, functions as a competitive inhibitor of mTOR kinase. The effectiveness of this drug was tested as a lone agent for untreated poor-prognosis patients and compared to interferon alpha (IFN-a) treatment. As monotherapy, temsirolimus significantly prolonged median overall survival compared to IFN-a (10.9 versus 7.3 months) (Hudes et al., 2007). In conclusion, breast and renal cell metastases to the brain remain a treatment challenge. Systemic therapies are improving, and early detection of CNS involvement is facilitated by advanced neuro-imaging capabilities widely available today. Both diseases use similar molecular and cellular pathways to metastasize, settle and proliferate within the brain. The process is highly organized and allows numerous avenues for investigation and intervention. Analysis of the genetic expression of various factors and receptors highlighted in this chapter will provide the keys to unlocking which patients will ultimately benefit our treatment and how the disease will flourish within the CNS.

82

References Ali S, Lazennec G (2007) Chemokines: novel targets for breast cancer metastasis. Canc Metastasis Rev 26:401–420 Altundag K, Bondy ML, Mirza NQ, Kau SW, Broglio K, Hortobagyi GN, Rivera E (2007) Clinicopathologic characteristics and prognostic factors in 420 metastatic breast cancer patients with central nervous system metastasis. Cancer 110:2640–2647 Balkwill F (2004) Cancer and the chemokine network. Nat Rev Canc 4:540–550 Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, Sawaya RE (2004) Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. J Clin Oncol 22: 2865–2872 Chang J, Clark GM, Allred DC, Mohsin S, Chamness G, Elledge RM (2003) Survival of patients with metastatic breast carcinoma: importance of prognostic markers of the primary tumor. Cancer 97:545–553 Cheng X„ Hung MC (2007) Breast cancer brain metastases. Cancer Metastasis Rev 26:635–643 Eichler AF, Kuter I, Ryan P, Schapira L, Younger J, Henson JW (2008) Survival in patients with brain metastases from breast cancer: the importance of HER-2 status. Cancer 112: 2359–2367 Eichler AF, Plotkin SR (2008) Brain metastases. Curr Treat Options Neurol 10:308–314 Engelhardt B, Sorokin L (2009) The blood-brain and the bloodcerebrospinal fluid barriers: function and dysfunction. Semin Immunopathol 31:497–511 Evans AJ, James JJ, Cornford EJ, Chan SY, Burrell HC, Pinder SE, Gutteridge E, Robertson JF, Hornbuckle J, Cheung KL (2004) Brain metastases from breast cancer: identification of a high-risk group. Clin Oncol (R Coll Radiol) 16: 345–349 Fitzgerald DP, Palmieri D, Hua E, Hargrave E, Herring JM, Qian Y, Vega-Valle E, Weil RJ, Stark AM, Vortmeyer AO, Steeg PS (2008) Reactive glia are recruited by highly proliferative brain metastases of breast cancer and promote tumor cell colonization. Clin Exp Metastasis 25:799–810 Folkman J (1995) Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med 1:27–31 Gavrilovic IT, Posner JB (2005) Brain metastases: epidemiology and pathophysiology. J Neurooncol 75:5–14 Grandinetti CA, Goldspiel BR (2007) Sorafenib and sunitinib: novel targeted therapies for renal cell cancer. Pharmacotherapy 27:1125–1144 Grossi PM, Ochiai H, Archer GE, McLendon RE, Zalutsky MR, Friedman AH, Friedman HS, Bigner DD, Sampson JH (2003) Efficacy of intracerebral microinfusion of trastuzumab in an athymic rat model of intracerebral metastatic breast cancer. Clin Cancer Res 9:5514–5520 Hakyemez B, Erdogan C, Gokalp G, Dusak A, Parlak M (2010) Solitary metastases and high-grade gliomas: radiological differentiation by morphometric analysis and perfusionweighted MRI. Clin Radiol 65:15–20 Heldwein FL, Escudier B, Smyth G, Souto CA, Vallancien G (2009) Metastatic renal cell carcinoma management. Int Braz J Urol 35:256–270

J.M. Sheehan and A.S. Patel Hicks DG, Short SM, Prescott NL, Tarr SM, Coleman KA, Yoder BJ, Crowe JP, Choueiri TK, Dawson AE, Budd GT, et al (2006) Breast cancers with brain metastases are more likely to be estrogen receptor negative, express the basal cytokeratin CK5/6, and overexpress HER2 or EGFR. Am J Surg Pathol 30:1097–1104 Hill KL Jr., Lipson AC, Sheehan JM (2009) Brain magnetic resonance imaging changes after sorafenib and sunitinib chemotherapy in patients with advanced renal cell and breast carcinoma. J Neurosurg 111:497–503 Hudes G, Carducci M, Tomczak P, Dutcher J, Figlin R, Kapoor A, Staroslawska E, Sosman J, McDermott D, Bodrogi I, et al (2007) Temsirolimus, interferon alfa, or both for advanced renal-cell carcinoma. N Engl J Med 356:2271–2281 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Murray T, Thun MJ (2008) Cancer statistics, 2008. CA Cancer J Clin 58:71–96 Kapoor A (2009) Inhibition of mTOR in kidney cancer. Curr Oncol 16(Suppl 1):S33–S39 Kato T, Kameoka S, Kimura T, Nishikawa T, Kobayashi M (2003) The combination of angiogenesis and blood vessel invasion as a prognostic indicator in primary breast cancer. Br J Cancer 88:1900–1908 Lai R, Dang CT, Malkin MG, Abrey LE (2004) The risk of central nervous system metastases after trastuzumab therapy in patients with breast carcinoma. Cancer 101:810–816 Law M, Cha S, Knopp EA, Johnson G, Arnett J, Litt AW (2002) High-grade gliomas and solitary metastases: differentiation by using perfusion and proton spectroscopic MR imaging. Radiology 222:715–721 Lee SS, Ahn JH, Kim MK, Sym SJ, Gong G, Ahn SD, Kim SB, Kim WK (2008) Brain metastases in breast cancer: prognostic factors and management. Breast Cancer Res Treat 111:523–530 Lin NU, Bellon JR, Winer EP (2004) CNS metastases in breast cancer. J Clin Oncol 22:3608–3617 Lin NU, Carey LA, Liu MC, Younger J, Come SE, Ewend M, Harris GJ, Bullitt E, Van den Abbeele AD, Henson JW et al (2008) Phase II trial of lapatinib for brain metastases in patients with human epidermal growth factor receptor 2-positive breast cancer. J Clin Oncol 26: 1993–1999 Lin NU, Dieras V, Paul D, Lossignol D, Christodoulou C, Stemmler HJ, Roche H, Liu MC, Greil R, Ciruelos E et al. (2009) Multicenter phase II study of lapatinib in patients with brain metastases from HER2-positive breast cancer. Clin Cancer Res 15:1452–1459 Meric-Bernstam F, Hung MC (2006) Advances in targeting human epidermal growth factor receptor-2 signaling for cancer therapy. Clin Canc Res 12:6326–6330 Miles FL, Pruitt FL, van Golen KL, Cooper CR (2008) Stepping out of the flow: capillary extravasation in cancer metastasis. Clin Exp Metastasis 25:305–324 Miller KD, Weathers T, Haney LG, Timmerman R, Dickler M, Shen J, Sledge GW Jr. (2003) Occult central nervous system involvement in patients with metastatic breast cancer: prevalence, predictive factors and impact on overall survival. Ann Oncol 14:1072–1077 Motzer RJ, Escudier B, Oudard S, Hutson TE, Porta C, Bracarda S, Grunwald V, Thompson JA, Figlin RA, Hollaender N, Urbanowitz G, Berg WJ, Kay A, Lebwohl D, Ravaud A (2008) Efficacy of everolimus in advanced renal cell

8

Breast Cancer and Renal Cell Cancer Metastases to the Brain

carcinoma: a double-blind, randomised, placebo-controlled phase III trial. Lancet 372:449–456 Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, McClanahan T, Murphy E, Yuan W, Wagner SN, Barrera JL, Mohar A, Verástegui E, Zlotnik A (2001) Involvement of chemokine receptors in breast cancer metastasis. Nature 410:50–56 Nussbaum RR, Werner U, Pichlmayr R (1996) Liver metastases in breast carcinoma. Results of partial liver resection. Chirurg 67(3):234–247 Ouban A, Ahmed AA (2010) Claudins in human cancer: a review. Histol Histopathol 25:83–90 Palmieri D, Chambers AF, Felding-Habermann B, Huang S, Steeg PS (2007) The biology of metastasis to a sanctuary site. Clin Cancer Res 13:1656–1662 Palmieri D, Smith QR, Lockman PR, Bronder J, Gril B, Chambers AF, Weil RJ, Steeg PS (2006) Brain metastases of breast cancer. Breast Dis 26:139–147 Patanaphan V, Salazar OM, Risco R (1988) Breast cancer: metastatic patterns and their prognosis. South Med J 81:1109–1112 Pestalozzi BC, Zahrieh D, Price KN, Holmberg SB, Lindtner J, Collins J, Crivellari D, Fey MF, Murray E, Pagani O, Simoncini E, Castiglione-Gertsch M, Gelber RD, Coates AS, Goldhirsch A (2006) Identifying breast cancer patients at risk for Central Nervous System (CNS) metastases in trials of the International Breast Cancer Study Group (IBCSG). Ann Oncol 17:935–944 Ranasinghe MG, Sheehan JM (2007) Surgical management of brain metastases. Neurosurg Focus 22:E2 Ravaud A, Hawkins R, Gardner JP, von der Maase H, Zantl N, Harper P, Rolland F, Audhuy B, Machiels JP, Petavy F, Gore M, Schoffski P, El-Hariry I (2008) Lapatinib versus hormone therapy in patients with advanced renal cell carcinoma: a randomized phase III clinical trial. J Clin Oncol 26:2285–2291 Reckamp KL, Strieter RM, Figlin RA (2008) Chemokines as therapeutic targets in renal cell carcinoma. Expert Rev Anticancer Ther 8:887–893

83 Sawaya R (2004) Chairman’s reflection on the past, present and future of neurosurgical oncology. J Neuro-Oncol 69: 19–23 Schouten LJ, Rutten J, Huveneers HA, Twijnstra A (2002) Incidence of brain metastases in a cohort of patients with carcinoma of the breast, colon, kidney, and lung and melanoma. Cancer 94(10):2698–2705 Sheehan JP, Sun MH, Kondziolka D, Flickinger J, Lunsford LD (2003) Radiosurgery in patients with renal cell carcinoma metastasis to the brain: long-term outcomes and prognostic factors influencing survival and local tumor control. J Neurosurg 98:342–349 Shuch B, La Rochelle JC, Klatte T, Riggs SB, Liu W, Kabbinavar FF, Pantuck AJ, Belldegrun AS (2008) Brain metastasis from renal cell carcinoma: presentation, recurrence, and survival. Cancer 113:1641–1648 Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, Fleming T, Eiermann W, Wolter J, Pegram M, et al (2001) Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 344:783–792 Stemmler HJ, Heinemann V (2008) Central nervous system metastases in HER-2-overexpressing metastatic breast cancer: a treatment challenge. Oncologist 13:739–750 Tham YL, Sexton K, Kramer R, Hilsenbeck S, Elledge R (2006) Primary breast cancer phenotypes associated with propensity for central nervous system metastases. Cancer 107: 696–704 Tsukada Y, Fouad A, Pickren JW, Lane WW (1983) Central nervous system metastasis from breast carcinoma. Autopsy study. Cancer 52:2349–2354 Weil RJ, Palmieri DC, Bronder JL, Stark AM, Steeg PS (2005) Breast cancer metastasis to the central nervous system. Am J Pathol 167:913–920 Zimm S, Wampler GL, Stablein D, Hazra T, Young HF (1981) Intracerebral metastases in solid-tumor patients: Natural history and results of treatment. Cancer 48(2):384–394

Chapter 9

Breast Cancer Brain Metastases: Genetic Profiling and Neurosurgical Therapy Andreas M. Stark

Abstract Brain metastases occur in 10–15% of patients with metastastic breast cancer. The annual incidence in the USA and Europe is in the range of 5–10 per 100,000 per year. The mean latency period between diagnosis of the primary tumor and the detection of brain metastases is 2–3 years. Most often, brain metastases arise at the white/grey matter border in the supratentorial space. The standard imaging is contrast-enhanced magnetic resonance imaging (MRI). Treatment consists in open surgical resection and/or stereotactic radiosurgery. It is appropriate to resect 1–3 accessible lesions within 1–2 operations. Adjuvant radiotherapy is applied on a routine basis. Chemotherapy might be applied according to the primary tumor. Overall median survival is in the range of 5.5–14.5 months. Research activities in this field include molecular profiling of breast cancer brain metastases. We have attempted genetic cDNA array profiling using a 15 K microarray of brain-selective MDA-MB-231 cells. These have been acquired from multiple passages of MDA-MB-231 invasive and metastastic breast cancer cells in mice. Profiling of selective metastatic cells showed accumulation of gene expression characteristic for brain metastasis. Several factors were identified which are potentially important for this process: MetastasisSuppressor Gene KiSS-1, Matrix-Metallo-Proteinase(MMP)-1 and vascular factor endoglin. Further structural and functional studies in selective metastastic cells and human specimens were undertaken. These results underline the importance of these factors for the molecular process of breast cancer brain metastasis.

Keywords Neurocranium · MSG · MDA-MB-231 · Imaging · Neuronavigation · Brain

A.M. Stark () Department of Neurosurgery, Schleswig-Holstein University Medical Center, Campus Kiel, 24105 Kiel, Germany e-mail: [email protected]

Approximately 10–15% of patients with metastatic breast cancer develop brain metastases. The annual incidence in the USA and Europe is in the range of

Introduction Breast Cancer Metastases to the Neurocranium Hematogenous breast cancer metastases to the neurocranium may arise in the bone, the dura or the brain. Affected patients are often in an advanced stage of disease (Weil et al., 2005). Bone metastases may be located within the calvaria or the skull base. Surgical resection and radiotherapy are the mainstays of therapy (Constans and Donzelli, 1981; Stark et al., 2003). Dural metastases are rare, and they may occur in one of two morphologies: (1) as a solid mass they are a differential diagnosis of meningioma, (2) as a subdural effusion they must be suspected in patients with subdural hematoma and underlying malignant disease. Cytological examination of subdural fluid is advised in these cases (Stark and Mehdorn, 2004). The most common presentation of breast cancer brain metastasis to the neurocranium is intraparenchymal “brain” metastases.

Breast Cancer Brain Metastases: Epidemiology

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_9, © Springer Science+Business Media B.V. 2011

85

86

5–10/100,000 per year. The median age at the time of diagnosis of breast cancer and brain metastases is 46–48 and 50–59 years, respectively. Brain metastases are widely regarded as a late complication in breast cancer. Typically, they grow at the white/gray matter border, more often in the cerebrum than in the cerebellum or the brain stem. The mean latency between diagnosis of the primary tumor and the detection of brain metastases is 2–3 years. However, breast cancer brain metastases may occur even 20 years after diagnosis of the initial breast tumor (Stark et al., 2005a; Weil et al., 2005; Wronski et al., 1997). Beside metastasis to the brain parenchyma, breast cancer has an obvious tendency towards leptomeningeal dissemination (meningeosis carcinomatosa). Leptomeningeal dissemination mostly occurs in patients with late stage metastatic breast cancer who have also solid brain metastases. Meningeosis carcinomatosa is associated with a poor prognosis.

A.M. Stark

individual characteristics of the host tissue influencing metastases formation (Puduvalli, 2001; Liotta and Kohn, 2001).

The Brain as a Special Microenvironment Characteristic sites of breast cancer metastases are bone (pelvis, vertebral body, calvaria), lung, liver, and brain. The brain constitutes a special microenvironment. It is separated from the blood stream by the blood-brain barrier (BBB), it has an extensive capillary bed, it lacks significant extracellular space, it lacks large diffusion channels, and it has no lymph drainage. Consequently, the brain tissue is characterized by restriction of unwanted movement of cells. Thus, metastatic cells which are able to invade the brain must have specific, unique properties to overcome these restrictions (Puduvalli, 2001). In order to answer this question we conducted an experimental design using modern molecular biology techniques.

The Seed and Soil Hypothesis Selective Metastatic Breast Cancer Cells The process of tumor metastasis includes certain steps which metastatic cells are forced to overcome in order to form metastatic lesions at distant organs. They need to escape from the primary tumor site, invade the blood vessels, escape from the vasculature, and migrate into the host tissue. Finally, they need to grow (Welch et al., 2000). It is well known that metastasis formation is not random. In other words, it is not sufficiently explained by the proportion of the vasculature and the endothelial surface. As early as in 1889, Paget reported the results of post-mortem examinations of 650 patients with breast cancer. He answered the question “What is it that decides what organs shall suffer in a case of disseminated cancer?” by the statement that “distribution of the secondary growths is not a matter of choice”. As a consequence, Paget is regarded as the founder of the “seed and soil” hypothesis of malignant tumor dissemination. Notably, in this series the cranium was affected in 36 individuals (5.5%) (Paget, 1889). It is now accepted that metastasis is defined by the type and individual properties of the tumor (the “seed”) and by the molecular background of the affected organ (the “soil”). Two molecular mechanisms are quite interesting in this model. First, the interaction between the metastatic cell and the involved organ. Second,

Yoneda et al. (2001) established a brain- and bone-seeking clone of MDA-MB-231 invasive and metastatic breast cancer cells. They injected parental cells (231-parental) in mice and observed multiple metastases to several organs, and resected only the brain and bone metastases and injected again one of these newly formed cell lines (231-brain and 231bone). After several passages 231-brain-cells exclusively formed brain metastases and 231-bone-cells formed exclusively bone metastases. It can be assumed that these selective metastatic cells accumulated characteristics which are essential for metastasis to the brain and bone.

The 231 Array Experiment Gene Profiling of MDA-MB-231 Selective Metastatic Breast Cancer Cells In cooperation with the National Cancer Institute and the National Human Genome Research Institute, NIH, Bethesda, USA, we have conducted a 15 K gene

9

Breast Cancer Brain Metastases

array of these selective metastatic cells. The cDNA microarray included 15,360 human genes printed on glass and hybridized with fluorescently labelled cDNA (Khan et al., 1999). In one array the genetic profile of 231-brain was compared to 231-parental, and in another array 231-bone was compared to 231parental. A threshold of 2.5-fold under- or overexpression was defined as a significant change. Data analysis revealed 113 genes which were differentially expressed in brain- and/or bone-selective clones versus parental cells. Of these, 79 genes had a known function. Nine genes were selected for further analysis based on the expression differences and the literature background (Stark et al., 2010). Data validation was performed using quantitative real-time polymerase chain reaction (RT-PCR). The following genes were found to be differentially expressed in brain- and/or bone-seeking clones (Fig. 9.1): • The Metastasis-Suppressor Gene Kiss1 was strongly reduced in brain- and in bone-selective cells. Kiss1 has initially been found to suppress the metastatic capability of malignant melanoma in vitro and in vivo (Harms et al., 2003; Shirasaki et al., 2001). • The tyrosine-protein kinase receptor Tie1 is involved in angiogenetic processes. Tie-1 mRNA expression was significantly increased in brain- and bone-seeking clones.

Fig. 9.1 Expression of selected genes in MDA-MB-231 Parental cells (P) versus MDA-MB-231 Brain-selective clones (Br) and MDA-MB-231 Bone-selective clones (Bo). ∗ = significant changes. Gene names are explained in the text

87

• Matrix Metalloproteinases (MMPs) are zincdependent endopeptidases which hydrolyse components of the extracellular matrix (Visse and Nagase, 2003). MMP-1 was significantly increased in the brain-seeking clone and even higher expressed in the bone-seeking variant. • Endoglin (CD105) is a cell membrane glycoproteine that is strongly expressed on endothelial tumor cells. It functions as an adhesion molecule for integrins and other receptors (Fonsatti et al., 2003). Endoglin was significantly higher expressed in both selective MDA-clones. The following genes were also tested by quantative real-time RT-PCR but were found not to be differentially expressed: MMP14 (matrix metalloproteinase 14), UBA2 (Sumo-1 activating enzyme subunit 2), DEK oncogene, CTGF (connective tissue growth factor), WFS1 (Wolframin syndrome 1).

Examination of Metastasis Suppressor Genes (MSG) in Selective Metastatic Breast Cancer Cells and Human Breast Cancer Brain Metastases Metastatic cells constitute a specialized subset of tumor cells that have acquired the ability to complete

88

the metastatic cascade: (1) migration from the primary tumor, (2) dissemination into the blood stream/survival in blood vessels, (3) extravasation into the host tissue, and (4) proliferation in the host tissue. During the past 2 decades, several factors have been identified, which suppress metastasis in vitro and in vivo. These metastasis suppressor genes (MSGs) are assigned a pivotal role in the metastatic process of human cancers (Welch et al., 2000; Debies and Welch, 2002). These genes are defined by their ability to inhibit metastasis to secondary organs without affecting the original tumor. Meanwhile, over 20 MSGs have been identified and tested in vivo (Horak et al., 2008). The first cloned metastasis suppressor gene, identified in a murine melanoma cell line, was Nm23 (Steeg et al., 1998). Transfection of Nm23 into cell lines affects invasion, motility, colonization, and differentiation. Nm23 is decreased in late-stage human cancers. It is located on the long arm of chromosome 17 (17q21). Kai1 (Kangai 1, CD 82, C33, 11p11.2) is a member of the TM4SF family of adhesion molecules. Its expression is inversely correlated with aggressive behavior in several human cell lines. Kai1 has been well described in prostate cancer and breast cancer. In breast cancer, Kai1 expression is correlated with prolonged patient survival. The exact mechanism of action of Kai1 is not known. It is discussed here if Kai1 may be a secondary target of the tumor suppressor protein p53. As mentioned above, Kiss1 was discovered as a melanoma metastasis suppressor: metastasis of human melanoma cells lines is inhibited after Kiss1 is introduced. Interestingly, Kiss1 expression is lost, when melanoma cells convert from the benign to the malignant phenotype. In breast cancer cell lines, introduction of Kiss1 suppresses metastasis. However, the mechanism of action of Kiss1 has not yet been identified. Kiss1 maps to 1q32 (Harms et al., 2003; Shirasaki et al., 2001). The introduction of Mkk4 (mitogenactivated protein kinase kinase 4) has been shown to suppress the metastatic ability of prostate cancer cells in a rat model. Mkk4-containing cells can escape from the primary tumor site but are growthsinhibited until Mkk4 is lost (Robinson et al., 2003). When BRMS1 (breast cancer metastasis-suppressor 1) is introduced into metastatic breast cancer cells, metastasis is suppressed. BRMS1 is located on chromosome 11 (11q13.1-13.2), a highly conserved domain. It may be involved in intercellular communication. However,

A.M. Stark

the exact mechanism of action is not known (Welch et al., 2000; Debies and Welch, 2002). Specific MSGs can be assigned to specific steps of the metastatic cascade. Most important in the light of future treatment options, Nm23, Kiss1, Kai1, and Mkk4 promote growths arrest of metastatic cells after they have invaded the host organ (Horak et al., 2008). Based on the 231 array experiment results, we have examined the expression of several MSGs in human samples of breast cancer brain metastases. The results have been compared to primary tumors and normal breast tissue (Stark et al., 2005a; 2010). In a first attempt, we performed quantitative realtime RT-PCR for the expression of metastasis suppressor genes Nm23, Kiss1, Kai1, BRMS1 and Mkk4 in breast cancer brain metastases versus primary tumors. We found significantly reduced expression of Kiss1, Kai1, BRMS1, and Mkk4 in metastatic lesions. These changes could also be observed at the protein level using immunohistochemistry. The mRNA expression of Nm23 was also reduced. However, this result was not statistically significant. In general, the Kiss1 mRNA content was much lower than the mRNA content of the other MSGs tested. Nm23 mRNA content was at the highest. According to the reported possible function of the examined genes, Kai1, Kiss1 and Mkk4 are promising candidates for future treatment strategies. Because these factors inhibit metastatic growths at the secondary site, clinically significant metastases could be possibly avoided by introduction of these MSGs into tumor cells. This approach is quite interesting. However, as outlined before, metastasis is a very complex process which is not yet fully understood. It can be assumed that reams of tumor cells invade the blood stream and are spread all over the vasculature. Few cells possibly attach to the endothelium, even fewer cells invade the surrounding stroma and possibly form clinically significant metastases. In many patients micrometastases are present when the primary tumor is detected and treated. At this time a hypothetic MSG therapy could be introduced. One would have to deliver the MSGs are delivered into the disseminated cells. Future research will hopefully make this scenario possible. In a second attempt, we examined the mRNA and protein expression of MSG maspin in breast cancer brain metastases versus primary tumors and

9

Breast Cancer Brain Metastases

normal breast tissue. Maspin (mammary serine protease inhibitor, Serpin B5) is a member of the serpin family of serine protease inhibitors. It was first described in 1994. The gene encodes for a 42 kDa protein that is expressed in myoepithelial cells in normal breast tissue. The gene maps to the long arm of chromosome 18 (18q21.3–q23). Maspin is involved in many physiological and pathological processes. It provokes apoptosis and inhibits tumor growth. It inhibits cell migration and angiogenesis. It has been described that maspin expression is lost during the progression of breast cancer (Bieche et al., 2003; Debies and Welch, 2002). It is thought to be down regulated in breast cancer and lost in distant metastases. However, some breast cancer tumors seem to show maspin upregulation (Bieche et al., 2003; Maas et al., 2001). There is evidence that the estrogen receptor (ER)-α controls maspin expression. The estrogen receptor alpha may itself be induced by tamoxifen. Maspin expression has prognostic value in ER-α positive postmenopausal breast cancer. We examined the expression of maspin in human tissue samples of normal breast tissue, primary breast cancer and brain metastases by RT-PCR and immunohistochemistry. We found decreased maspin mRNA expression in breast cancer versus normal tissue and, again, in brain metastases versus primary tumors. Expression values were normalized against normal breast tissue (=1). Herein, mean expression of primary tumors was 0.3 and maspin mRNA expression, in brain metastases was 0.13 (Stark et al., 2010). The same tendencies were observed for the protein level. All normal tissue probes showed maspin expression whereas only 19% of primary tumors were positive. Interestingly, none of the brain metastasis expressed maspin protein. The expression of ER-α was also reduced in brain metastases supporting the hypothesis that maspin might be regulated by ER-α. Maspin mRNA expression was also detected in breast cancer cell lines. In poorly invasive and nonmetastatic breast cancer cells (MCF-7, T47-D), maspin mRNA was much higher expressed than in highly metastatic cells (MDA-MB-231 parental). In conclusion, maspin is a potential therapeutic target for the treatment of breast cancer brain metastases. Selective ER modulators such as tamoxifen might lead to increased expression of maspin in patients with breast cancer.

89

Examination of Matrix-MetalloProteinases in Selective Metastatic MDA-MB-231 Breast Cancer Cells Matrix Metalloproteinases (MMPs) are an interesting family of zinc-dependent endopeptidases which hydrolyze components of the extracellular matrix. MMPs play a significant role in normal tissue remodeling (embryogenic development, angiogenesis, ovulation, wound healing) and migration, invasion and metastasis of many human cancers. They are produced by tumor cells as well as by neighbouring stromal cells. A loss of MMP activity control can lead to arthritis, atherosclerosis, aneurysm formation, and fibrosis. Four groups of MMPs have been described: (1) gelatinases (MMP-2 and -9), (2) interstitial collagenases (MMP-1, -8 and -13), (3) stromyelysins (MMP-3, -7, -10 and -11), and (4) membrane MMPs (MMP-14 to -17) (Visse and Nagase, 2003). Certain MMPs are involved in breast cancer migration, invasion, and metastasis. Breast cancer cells express high levels of MMPs 1, 3, 9, 13, and 14. The role of MMPs for tumor genesis has been well described. However, only few data exist concerning the role of MMPs for selective metastatic properties. We have examined the mRNA and protein expression of MMPs 1–3, 8, 9, 13 and 14 in selective metastatic MDA-MB 231 parental, brain-, and boneseeking breast cancer cells. Using quantitative RTPCR, we found MMP-1 and -9 overexpression in brain- and bone-selective cells versus parental cells. Expression for both MMPs was highest in brainselective clones, MMP-1 showed the highest overall expression values. In contrast to MMP-1 and -9, MMP14 was highly expressed in all three cell lines, and MMP-13 showed reduced mRNA expression in both selective clones. MMP-2 and -8 were not expressed in any of the MDA-MB 231 cell line. Examination of protein expression using enzymelinked immunosorbant essay (ELISA) also revealed overexpression of MMP-1 and -9 in both selective cell lines. Again, expression was the highest in brainselective cells. It means that the amount of total as well as the amount of active protein was higher in 231-brain and – bone than in 231-parental cells. Functional studies demonstrated that MMP-1 and 9 mRNA and protein could be stimulated by TPA in

90

231-parental cells much more than in selective clones. This is a strong evidence that MMP-1 and -9 are already upregulated in 231-brain and – bone. Next, it was tested whether MMP-1 and -9 contribute to migration and invasion properties of the examined cell lines. We observed that migration of the brain-seeking clone could be inhibited by the administration of MMP-9 inhibitor. The same effect could be observed if MMP-9 inhibitor was combined with MMP-1 inhibitor but not with MMP-1 inhibitor alone. In contrast, migration properties of selective metastatic clones was not significantly different when treated with MMP-inhibitor (Stark et al., 2007). In conclusion, MMP-1 and -9 were identified as potential therapeutic targets in breast cancer brain metastases. MMP-1 belongs to the interstitial collagenases. It cleaves different collagen types and can be induced by epidermal growth factor receptor stimulation or by the application of tumor necrosis factor alpha. MMP-1 is overexpressed in breast cancer cells with high invasion potential only. MMP-9 (gelatinase B) degrades elements of the extracellular matrix. It has been associated with disruption of the basement membrane that opens the BBB after tissue injury. MMP-9 expression has been associated with aggressive behavior of many human cancers. Przybylowska et al. (2005) reported that MMP-1 is responsible for local breast cancer invasion and MMP-9 for tumor growths. As a result of clinical practice, MMPs can be therapeutically influenced by synthetic small molecule inhibitors. Mendes et al. (2005) have demonstrated that the development of breast cancer brain metastases can be suppressed by MMP inhibitors in a rat model.

Examination of Endoglin in Selective Metastatic MDA-MB-231 Breast Cancer Cells Endoglin (CD105) is a disulfide-linked homodimeric glycoprotein expressed on the cell membrane, in particular on endothelial tumor cells. It is an adhesion molecule for integrins and acts as an accessory component of the transforming growths factor ß (TGF-ß) (Fonsatti et al., 2003). Endoglin gene defects cause the Osler-Rendu-Weber syndrome 1 (hereditary hemorrhagic teleangiectasia type 1). Therefore, endoglin is assigned a crucial role for normal vasculature development. Moreover, endoglin is expressed in many human

A.M. Stark

cancer types. However, its exact biological function in cancer genesis is unclear. Oxman et al. (2008) have examined the mRNA and protein expression of endoglin in MDA-MB 231 parental cells as well as in brain- and boneselective clones. They found highly overexpressed endoglin in both selective clones, endoglin expression was highest in brain-seeking clones. Moreover, endoglin expression was associated with the secretion of matrix-metallo-proteinases –1 and –19. Using functional experiments in parental and brain-seeking cells, endoglin expression was furthermore associated with higher migration and higher invasion capacity of the clones. In conclusion, endoglin can be associated with the metastatic phenotype of brain-selective MDAMB-231 clones. Further research is needed to answer the question whether endoglin might be used as a therapeutic target.

Neurosurgical Therapy of Breast Cancer Brain Metastases Neurosurgical treatment for breast cancer brain metastases does not differ substantially from the treatment of brain metastases from other types of cancers. Open microsurgical resection is warranted whenever clinically possible. For small lesions (size ≤ 3 cm diameter, n ≥3) stereotactic radiosurgery should be considered as an alternative or additional treatment to surgery. The advantage of surgery is complete erasure of the lesion. If the brain metastasis is surrounded by significant edema, the edema usually decreases after removal of the metastasis. Due to advances in operative techniques and neuroanesthesia, the removal of 1–3 (maybe 4) accessible lesions in 1–2 operations is possible. If two operations are required they may be performed in one operative session.

Epidemiology It is estimated that in the USA 1.3 million patients are diagnosed with breast cancer each year. At least 100,000 of these individuals will develop brain metastases which would result in an incidence of at least 8%. According to autopsy data, the rate of brain metastases in patients with metastatic breast cancer is in the range of 20–40% (Weil et al., 2005). Breast

9

Breast Cancer Brain Metastases

cancer is the second most common primary tumor forming brain metastases. The most common tumor is lung cancer. There are some clinical differences concerning the occurrence of brain metastases of different tumor types.

Interval Between Diagnosis of the Primary Tumor and the Detection of Brain Metastases Breast cancer brain metastases occur usually years after diagnosis of the primary tumor. According to Wronski et al. (1997), the median interval of the 70 patients included in the study was 28 months. There was no significant association with prognosis when patients with an interval <28 months were compared to those > 28 months. In our own series, the median interval was 39 months with a range between 8 months and 18 years. Most other cancer types forming brain metastases show a smaller median interval from diagnosis to brain metastases than breast cancer. In lung cancer, we have observed a median interval of 9 months in small cell lung cancer and 6 months in non small cell lung cancer. In non small cell lung cancer as well as in small cell lung cancer, synchronous diagnosis of brain metastases and the primary tumor is common. This condition is rare in breast cancer. Further tumor types with relatively long interval between diagnosis of the primary tumor and detection of brain metastasis is colorectal cancer (44 months according to our data), larynx cancer (59 months) and ovarian cancer (35 months) (Stark et al., 2005b).

Location of Brain Metastases Most often, brain metastases arise at the white/grey matter border in the supratentorial space. Metastasis to the infratentorial space is less common. Location may be superficial (=accessible) or deep in the basal ganglia or brain stem.

Solitary, Singular and Multiple Brain Metastases Solitary brain metastasis means that one single brain metastasis is present without detection of any other

91

metastases in the body. In contrast, the term singular brain metastasis describes one single brain metastasis with further extracranial metastatic lesions. The term “multiple brain metastases” is usually used if more than three brain lesions are detected. According to our data, solitary brain metastasis in breast cancer patients is in the range of 20% of patients undergoing neurosurgical treatment. If the lesion is accessible, solitary brain metastases is usually an indication for surgery. Surgical treatment in singular breast cancer brain metastasis is indicated if the primary tumor and extracranial metastases are under control. In general, a life expectancy of at least 6 months should be expected. Notable, treatment decisions should be made on an individual basis.

Symptoms and Findings Brain metastases lead to symptoms primary referring to their location. In general, focal neurological deficits (hemiparesis, aphasia, sensation deficits, visual disturbances) and signs of raised intracranial pressure (headache, nausea, vomiting) can be distinguished. Raised intracranial pressure may result from large brain metastases or from significant surrounding edema or both. Edema can be extreme in brain metastases and can be treated with steroids until surgery and/or radiotherapy can be applied. According to our observations in 47 female patients, the most common leading symptom was ataxia (23%), followed by headache (21%), visual disturbances (15%), hemiparesis and vertigo (11%), aphasia, neuropsychological deficits and cranial nerve affection (4%), and hyposensibility (2%). Brain metastases (4%) were asymptomatic findings in routine follow-up magnetic resonance imaging.

Imaging The standard imaging tool for brain metastases, as well as for intracranial tumors in general, is magnetic resonance imaging (MRI). It is highly superior to computed tomography (CT) in showing anatomical and pathological details. CT is useful in scanning patients who present in an emergency situation. Herein, intracranial bleeding can be ruled out and metastasis might be suspected due to isodense or

92

A.M. Stark

hypodense lesions surrounded by edema. If contrast material is given, brain metastases will show homogenous or ring-like enhancement. Notably, CT contrast administration is only advisable if MRI is not possible. Brain metastases might present with intratumoral bleeding. This is common in renal cell cancer and melanoma brain metastases, but might also occur in colorectal cancer and breast cancer lesions. Often, there is significant edema surrounding the bleeding. The bleeding typically has a sharp margin which can be distinguished from spontaneous (non tumorigenic) intracerebral bleeding. If brain metastasis is suspected in CT, MRI should be followed before and after administration of contrast material. Using MRI, the number of brain metastases can be determined to lesions of 1 mm in diameter and even below. The exact formation can be visualized: solid, cystic, or partially cystic. Breast cancer brain metastases are often at least partially cystic. The exact anatomical location can be noticed and the operative access can be planned. In cases with tumor location inside or close to eloquent structures (motor cortex, speech area, visual cortex) functional MRI imaging can help to plan treatment. For example, fiber tracking can visualize motor fibers and functional speech or motor MRI can identify brain regions of motor or speech action. These are helpful tools for operation planning.

a

In cases where bone infiltration is suspected, additional CT with 1–2 mm slices is necessary. This is especially indicated for skull base lesions or superficial lesions with contact to the bone. Intraoperatively, the adjacent bone might be replaced. Imaging examples are demonstrated in Fig. 9.2.

Treatment Decision/Indication for Surgery In general, surgery for brain metastases is indicated when life expectancy is at least 6 months. The primary tumor and possibly existing extracranial metastases should be under control. However, in clinical practice surgery might also be applied to patients presenting in an emergency situation with large tumors and/or large surrounding edema in risk for brain herniation. This might especially occur in patients with infratentorial brain metastases. The number of brain metastases should generally not exceed 3 lesions (maybe 4). Lesions planned for surgery should be accessible. This means that they should be superficial and not located within eloquent brain areas. Patients with lesions located very close to eloquent areas may be candidates for awake craniotomies. The neurological function can be measured by patient’s active movement or speech under local anesthesia.

b

Fig. 9.2 a Contrast enhanced T1-weighted MRI of an 59-year old female with cystic breast cancer brain metastases (arrows). b Contrast enhanced T1-weighted MRI of an 38-year old female with solid breast cancer brain metastases (arrow)

9

Breast Cancer Brain Metastases

Prognostic Factors and Survival Overall median survival in patients after surgical resection for breast cancer brain metastasis is in the range of 5.5–14.5 months (Wronski et al., 1997; Stark et al., 2005b). Generally, accepted independent prognostic factors are only patient’s age and performance. The age threshold separating patients with favourable and unfavourable diagnosis is usually set between 55 and 70 years. Analysis of our data has revealed that the age of 65 years is statistically most useful in determining this threshold. Patient performance is generally specified using the Karnofsky Performance Score (KPS). This score was established by Karnofsky and Burchenal in 1948 rating cancer patients for the amount of independent patient activity. The, activity is measured in 10 point steps from 10 to 100 with 100 meaning full independency. A KPS of 60–70 is usually required to enter clinical studies. Further prognostic factors have been described. However, their prognostic value is debatable: the presence of extracranial metastases, risk factors, and the hormone status. Undoubtfully, leptomeningeal carcinomatosis is associated with a worse prognosis. This condition describes diffuse metastatic growth along the leptomeninges. In MRI, carcinomatosis is visualized by leptomeningeal contrast enhancement. Hydrocephalus is often present and may require diversion of cerebrospinal fluid. To prevent malignant cell carry-over into the peritoneum or the blood-stream, an Ommaya reservoir is placed. This reservoir can be punctuated with a needle and CSF can be aspirated repeatedly. Intrathecal chemotherapy may be applied via the reservoir.

Planning Surgical Treatment: Neuronavigation Surgical preparation for removal of brain metastases includes choosing the approach and, in most circumstances, neuronavigation. Neuronavigation refers to an intraoperatively usable 3D-computer model based on a preoperative CT or MRI scan that helps the surgeon to navigate inside the skull. As a first step, patient’s head is covered with several fiducia markers. Then, a special thin-sliced CT or, better, MRI scan is performed after contrast administration. Imaging

93

data are transferred to a workstation where a 3D model of the patient’s head is established. Immediately before the operation, the patient’s head is fixed in a Mayfield-clamp, preventing movement during the time of surgery. A reference star is connected to the clamp and patient’s head is registered to the computer model. Now, the surgeon can check for the exact location of the tumor at any time of the operation. This is especially valuable for creating an exact access to the tumor. In glioma surgery, neuronavigation is also used for resection control. It should be noted that after opening the skull, the brain will considerably move which leads to inaccurateness of the neuronavigation system. This incident is called brainshift.

Surgical Treatment After opening the skull, the operation is continued under the microscope. Small vessels can be visualized much easier and surgery is much more gentle. The dura is opened and CSF may be diversed in order to gain space for surgical intenvention. The brain might have to be devided to gain access to the lesion. Self-holding retractors may be used for deeper-seated tumors. Cystic lesions may be punctured to facilitate removal of the solid part or the cyst wall. Then, the margin around the lesion is dissected with bipolar forceps and cottonoids. Usually, breast cancer brain metastases are quite firm and have a sharp margin. Small and/or superficially-located lesions might be erased in toto, larger lesions might be reduced in size before they can be completely removed. The tumor bed is searched for bleeding and finally tapered with cellulose stripes. It should be the goal of surgery to remove the lesion(s) completely. It is currently under discussion if a margin of brain tissue should also be resected. After closure of the dura, the bone flap is replaced. Only in few cases with infiltration of the dura and bone grafts need to be placed. The dura can be replaced by autologous galea. If more material is needed, fascia lata may be removed from the leg or xenograft material may be used. The bone can easily be replaced by bone cement.

Surgical Treatment: Optional Tools The integration of MRI into the operating room enables the surgeon to control the resection status

94

intraoperativly with high resolution. Mostly this technique is used for gliomas. However, it may also be used in selected cases of brain metastasis. It may be useful in a few cases of recurrent brain metastases when more than one small lesion is operated. Neuronavigation can be updated after removal of the first lesion and before resection of the second one. This practice will overcome brainshift problems. Intraoperative ultrasound is another technique which enables an intra-operative imaging. It is less expansive and easy to use. The depths is limited and the imaging quality depends on the nature of the lesion. Awake-craniotomy may be performed when the metastasis is located close to an eloquent brain area. Usually, this technique is used when gliomas infiltrate the motor cortex or speech area. However, it might be occasionally used for brain metastases. The patient is given local anesthesia for rigid head fixation and skin incision. After skull opening, the dura is covered with cottonoids soaked with local anesthetic. Then the operation is performed. Because the brain itself does not express pain sensors, the patient can contribute to monitor his/her neurological function. The operation is always assisted by a neuropsychology team member.

Complications of Surgical Treatment Complications during the perioperative period may or may not be related to the surgical procedure. Complications related to the surgical procedure include bleeding and ischemia due to injury to the brain vessels. Bleeding and, in very few cases, ischemia might occur after the operation is finished. As a consequence, neurological deficits may occur. Secondary bleeding is common in patients with altered blood clotting function. CSF fistula may result from insufficient dural closure. Superficial or deep infections might occur. Hydrocephalus is an uncommon complication, the cause of which is often debatable. It may occur from bleeding into the CSF space, meningitis or meningeosis carcinomatosa. Patients with systemic cancer are more prone to develop nonsurgery-related complications. The most frequent complications are deep vein thrombosis, lung arterial embolism, and myocardial infarction. Patients who undergo craniotomy for removal of brain metastases should be watched at least overnight on an intensive care ward. A CT scan must be available at

A.M. Stark

any time. Contrast-enhanced MRI should be used for resection control.

Adjuvant Treatment After the nature of the lesion has been proven histologically, patients with brain metastases should undergo radiotherapy. Most regimens include fractionated external beam radiation with a total dose of 30 Gy. In patients with multiple brain metastases where one single lesion has been removed and another 1–2 small lesions are present radiosurgery may be used in addition to surgery. Systemic chemotherapy for breast cancer brain metastases is only applied for the primary tumor and/or extracranial metastases. There is currently no specific chemotherapy available for breast cancer brain metastases. As outlined above, intrathecal chemotherapy might be applied in patients with meningeosis carcinomatosa.

Post-operative Follow-up There is no overall standard procedure for postoperative follow-up visits for patients with brain metastases in the neuro-oncological community. In our department, after craniotomy and radiotherapy have been applied, patients report every three months to our outpatient department. Herein, clinical examination is combined with contrast enhanced MRI. Close observation of the patients enables the possibility to re-operate on recurrent metastases before they become symptomatic. It is well known that as much as 50% of patients with breast cancer brain metastases develop recurrent metastases (either at the site of the initial lesion or somewhere else in the brain). Furthermore, 50% of these patients will die from brain lesions.

References Bieche I, Girault I, Sabourin JC, Tozlu S, Driouch K, Vidaud M, Lidereau R (2003) Prognostic value of maspin mRNA expression in ER alpha-positive postmenopausal breast carcinomas. Br J Cancer 88:863–870 Constans JP, Donzelli R (1981) Surgical features of cranial metastases. Surg Neurol 15:35–38

9

Breast Cancer Brain Metastases

Debies MT, Welch DR (2002) Genetic basis of human breast cancer metastasis. J Mamm Gliand Biol Neopl 6:441–451 Fonsatti E, Altomonte M, Nicotra MR, Natali PG, Maio M (2003) Endoglin (CD105): a powerful therapeutic target on tumor-associated angiogenetic blood vessels. Oncogene 22:6557–6563 Harms JF, Welch DR, Miele ME (2003) Kiss1 metastasis suppression and emergent pathways. Clin Exp Metastasis 20:11–18 Horak CE, Lee JH, Marshall JC, Shreeve SM, Steeg PS (2008) The role of metastasis suppressor genes in metastatic dormancy. APMIS 116:586–601 Karnofsky, DA, Burchenal, JH (1949) The clinical evaluation of chemotherapy agents. Macleod CM (ed) Evaluation of chemotherapy agents. Columbia Press, New York, NY, pp 191–205 Khan J, Saal LH, Bittner ML, Chen Y, Trent JM, Meltzer PS (1999) Expression profiling in cancer using cDNA microarrays. Electrophoresis 20:223–229 Liotta LA, Kohn EC (2001) The microenvironment of the tumour-host interface. Nature 411:375–379 Maas, N, Hojo, T, Rosel, F, Ikeda, T, Jonat, W, and Nagasaki, K (2001) Down regulation of the tumor suppressor gene maspin in breast carcinoma is associated with a higher risk of distant metastasis. Clin Biochem 34:303–307 Mendes O, Kim HT, Stoica G (2005) Expression of MMP2, MMP9 and MMP3 in breast cancer brain metastasis in a rat model. Clin Exp Metastasis 22:237–246 Oxmann D, Held-Feindt J, Stark AM, Hattermann K, Yoneda T, Mentlein R (2008) Endoglin expression in metastatic breast cancer cells enhances their invasive phenotype. Oncogene 27:3567–3575 Paget S (1889) The distribution of secondary growths in cancer of the breast. Lancet 1:571–573 Przybylowska K, Kluczna A, Zadrozny M, Krawczyk T, Kulig A, Rykala J, Kolacinska A, Morawiec Z, Drzewoski J, Blasiak J (2005) Polymorphisms of the promoter regions of matrix metalloproteinases genes MMP-1 and MMP-9 in breast cancer. Breast Cancer Res 95:65–72 Puduvalli VK (2001) Brain metastases: biology and the role of the brain microenvironment. Curr Oncol Rep 3:467–475 Robinson VL, Hickson JA, Vander Griend DJ, Dubauskas Z, Rinker-Schaeffer CW (2003) MKK4 and metastasis suppression: a marriage of signal transduction and metastasis research. Clin Exp Metastasis 20:25–30 Shirasaki F, Takata M, Hatta N, Takehara K (2001) Loss of expression of the metastasis suppressor gene Kiss1 during melanoma progression and its association with LOH of chromosome 6q16.3-q23. Cancer Res 61:7422–7425

95 Stark AM, Anuszkiewicz B, Mentlein R, Yoneda T, Mehdorn HM, Held-Feindt J (2007) Differential expression of matrix metalloproteinases in brain- and bone-seeking clones of metastatic MDA-MB-231 breast cancer cells. J Neurooncol 81:39–41 Stark AM, Eichmann T, Mehdorn HM (2003) Skull metastases – clinical features, differential diagnosis and review of the literature. Surg Neurol 60:219–225 Stark AM, Mehdorn HM (2004) Dural metastases. J NeuroOncol 68:10 Stark AM, Mentlein R, Mehdorn HM, Held-Feindt J (2010) Genetisches Profil von Hirn- und Knochen-selektiven MDAMB-231 Klonen. Chirurg Forum 39:19–21 Stark AM, Schem C, Maass N, Hugo HH, Mehdorn HM, Jonat W, Held-Feindt J (2010) Reduced Expression of Metastasis Suppressor Gene Maspin in Breast Cancer Brain Metastases. Neurol Res 32:303–308 Stark AM, Tongers K, Maass N, Mehdorn HM, HeldFeindt J (2005a) Reduced Metastasis-Suppressor Gene mRNA-Expression in Breast Cancer Brain Metastases. J Cancer Res Clin Oncol 131:179–183 Stark AM, Tscheslog H, Buhl R, Held-Feindt J, Mehdorn HM (2005b) Surgical treatment for brain metastases: prognostic factors and survival in 177 patients. Neurosurg Rev 28: 115–119 Steeg PS, Bevilacqua G, Kopper L, Thorgeirsson UP, Talmadge JE, Liotta LA, Sobel ME (1998) Evidence for a novel gene associated with low tumor metastatic potential. J Natl Cancer Inst 80:200–204 Visse R, Nagase H (2003) Matrix metalloproteinases and tissue inhibitors of metalloproteinases, structure, function and biochemistry. Circ Res 92:827–839 Weil RJ, Palmieri DC, Bronder JL, Stark AM, Steeg PS (2005) Breast cancer metastasis to the central nervous system. Am J Pathol 167:913–920 Welch DR, Steeg PS, Rinker-Schaeffer CW (2000) Molecular biology of breast cancer metastasis. Genetic regulation of human breast carcinoma metastasis. Breast Cancer Res 2:408–416 Wronski M, Arbit E, McCormick B (1997) Surgical treatment of 70 patients with brain metastases from breast carcinoma. Cancer 80:1746–1754 Yoneda T, Williams PJ, Hiraga T, Niewolna M, Nishimura R (2001) A bone-seeking clone exhibits different biological properties from the MDA-MB-231 parental human breast cancer cells and a brain-seeking clone in vivo and in vitro. J Bone Miner Res 16:1486–1495

Chapter 10

Central Nervous System Tumours in Women Who Received Capecitabine and Lapatinib Therapy for Metastatic Breast Cancer Stephanie Sutherland and Stephen Johnston

Abstract Breast cancer is a common disease affecting many women worldwide. Approximately 20–25% of breast cancers belong to a sub-group that overexpress the HER2 oncogene which is associated with a poorer prognosis and a higher incidence of CNS metastases. The use of adjuvant Trastuzumab has halved the risk of relapse and recurrence in this population but sadly a significant number of women still develop systemic disease including CNS metastases. This poses a challenge for treatment: there is a limit to the amount of radiation that the brain can tolerate and cytotoxics are not generally effective. Trastuzumab has been shown to improve survival in CNS disease but this is due to better systemic disease control as it does not cross the BBB in any meaningful concentration. Lapatinib a small molecule TKI has been shown to be effective in pre-clinical models and this was validated in further phase II studies. The LEAP study was a global open access programme for women with HER2+ MBC that had disease progression following treatment with anthracycline, taxane and Trastuzumab. Amongst the sub-group of patients with CNS metastases overall response rates were 18–21% with approximately half achieving clinical benefit. Clinical studies are ongoing looking at the role of lapatinib in treating or delaying the onset of CNS disease in the metastatic setting (LANTERN & CEREBRAL) and to see whether using it adjuvantly can prevent or reduce the incidence of CNS metastases (ALTTO). Future developments include development of novel targeted

S. Sutherland () Breast Unit, Royal Marsden Hospital, Downs Road, Sutton, SM2 5PT, UK e-mail: [email protected]

agents acting further along the intra-cellular portion of the HER2 signalling pathway, new routes of administration of existing therapies and consideration of preventive strategies such as PCI. Survival of patients with HER2+ CNS metastases has improved significantly and it is important that alternative treatment strategies are found once patients have progressed after local treatment (WBRT, surgery or SRS) and Trastuzumab. Capecitabine in combination with Lapatinib has been shown to be effective in this setting and Lapatinib may have a role to play in prevention of CNS metastases when used as adjuvant treatment. Keywords Central nervous system · Tumour · Capecitabine · Lapatinib · MBC · LEAP study

Introduction Breast cancer is a common disease affecting 1 in 9 women in the developed world and despite advances in treatment, remains one of the leading causes of cancer related deaths in women. Approximately 10–30% (depending on the stage of their disease at diagnosis) of patients treated for early breast cancer will go on to develop systemic disease and in those with metastatic breast cancer (MBC) 10–16% will include Central Nervous System (CNS) metastases (Tsukada et al., 1983; Barnholtz-Sloan et al., 2004). Treatment of CNS metastases is a challenge in that options are relatively limited. Cytotoxic drugs have not proved to be particularly effective and radiotherapy has hitherto been the mainstay of treatment. For selected patients of good performance status with

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_10, © Springer Science+Business Media B.V. 2011

97

98

a solitary lesion, and in whom systemic disease is absent or well controlled, surgery is the treatment of choice. Stereotactic RT (SRS) can be used in similar patients with 1–3 lesions providing the tumours are of a suitable size (<3 cm diameter) and position but otherwise whole brain radiotherapy (WBRT) is the standard of care (Richards et al., 2007; Leyland-Jones, 2009). The role of WBRT in addition to surgery or SRS is still under debate with some centres using it in addition to improve CNS specific PFS and other centres preferring to keep it in reserve for use on progression due to concerns over side effects such as lethargy and cognitive impairment. Once patients develop disease progression after local CNS treatment, options are limited and until recently were mostly directed towards symptom control and palliative care. Previously this was not so much of a clinical issue as brain metastases tended to occur late in the course of the disease and were often superseded by extracranial disease progression. However as treatments and survival improves, symptomatic CNS metastases are becoming more of a clinical problem, causing significant morbidity to patients and control of CNS disease is becoming a critical part of maintaining quality of life. Approximately 25–30% of patients with breast cancer belong to a subgroup who over express the human epidermal growth factor receptor (HER2) ErbB2 oncogene. This oncogene encodes the trans membrane tyrosine kinase receptors that activate intra-cellular signaling pathways responsible for proliferation, cell survival and angiogenesis and is associated with a higher risk of disease progression and death (Paik et al., 1990; Meric et al., 2002). Historically patients with HER2+ breast cancer had a significantly worse outcome than those with HER2disease but the advent of Trastuzumab, a humanised monoclonal antibody targeting the extra-cellular domain of the HER2 receptor, has radically changed the outlook for these patients. Results from the adjuvant HERA study showed that 12 months of adjuvant treatment with Trastuzumab reduced the risk of recurrence by approximately 50% for these patients (Smith et al., 2007). Despite this improvement and reduction in risk of recurrence, unfortunately a significant number of HER2+ patients still relapse with MBC of whom 28–43% will go on to develop central nervous system (CNS) disease either as a first site of relapse, or in addition despite other sites of systemic disease

S. Sutherland and S. Johnston

being well controlled and responding to Trastuzumab– based therapy (Bendell et al., 2003; Clayton et al., 2004). The incidence of Lepto-meningeal disease in HER2+ patients is quoted as 2–5% by De Angelis in Diseases of the Breast (2000) and these are probably under-diagnosed in clinical practice. Certain factors are known to increase risk of CNS metastases in the HER2+ population, namely young age, premenopausal status at diagnosis and ER negativity as well as visceral disease as a first site of metastatic relapse (Gori et al., 2007). The documented incidence of CNS metastases appears to be increasing in HER2+ MBC but this may be due to better control of systemic disease resulting in improved survival which allows patients to live long enough to develop CNS metastases (Stemmler et al., 2006; Park et al., 2009). Advances in imaging modalities and techniques have also led to earlier detection of CNS metastases which may have contributed to the apparent rise in incidence. Approximately 10% of HER2+ patients will develop isolated CNS disease as a first site of progression (Burstein et al., 2005) suggesting that HER2 + disease may have a particular affinity for the brain; the other possible reason being that the large monoclonal antibody Trastuzumab (molecular weight 145 Kda) does not readily cross the blood brain barrier (BBB) allowing the brain to behave as a sanctuary site whilst disease remains well controlled elsewhere. Interestingly survival is better for HER2+ patients after a diagnosis of CNS disease than those who are HER2 negative (Kirsch and Hochberg, 2003) and it is believed that this is due to better control of extra-cranial disease. Trastuzumab has not been shown to be effective in treating CNS disease per se although it is well recognized that giving or continuing Trastuzumab after a diagnosis of CNS disease improves overall survival despite the fact that it does not penetrate the BBB in any meaningful concentration (Eichler et al., 2008; Park et al., 2009). The recent development of Lapatinib, an orally active small molecule reversible dual tyrosine kinase inhibitor (TKI) targeting the intracellular cytoplasmic adenosine tri-phosphate (ATP) binding sites of both HER2 and ErbB1 (EGFR) pathways, has led to improvements in treatments for HER2+ metastatic patients that have progressed either in the CNS or systemically following local CNS treatment,. It has been shown to be effective both in laboratory and phase 1 studies and more recently in clinical trials such as

10 Central Nervous System Tumours in Women

EGF100151 and the LEAP study where it was used in combination with Capecitabine.

Pre-clinical Data Pre-clinical data appears to show that combining lapatinib with Trastuzumab can overcome resistance to Trastuzumab. Lapatinib demonstrated growth inhibitory activity in a population of ErbB2+ (HER2+) cells selected for long term growth in media treated with a therapeutic level of Trastuzumab suggesting that there is little overlapping of resistance mechanisms (Konecny et al., 2006). Studies using ErbB2 positive cell lines demonstrated a synergistic action between lapatinib and Trastuzumab which suggested that dual blockade was more effective than single agent treatment, resulting in increased apoptosis and may be useful in overcoming Trastuzumab refractory or resistant cancers. Lapatinib, being a small molecule TKI, molecular weight <1 KDa, in theory may penetrate more readily into the CNS which would make it particularly useful in the treatment of CNS disease. A pre-clinical study by Gril et al. showed in-vitro inhibition of phosphorylation of EGFR, HER2, downstream signalling proteins and cell proliferation in brain seeking breast cancer cell lines MDA-MB-231-BR (with and without HER2), and a corresponding 50–53% reduction in the number of brain metastases developed by nude mice treated with lapatinib following an intra-cardiac injection of these cells when compared with a control group (Gril et al., 2008). However, it has also been shown in pre-clinical models that Lapatinib does not cross the intact BBB to any significant degree (Lin et al., 2008; Polli et al., 2008) and this may indicate that some of the method of action in apparent reduction of risk for development of CNS disease is related to systemic reduction of circulating tumour cells (CTCs). It is probable that disruption of the BBB by the tumour, surgery, SRS or WBRT may also contribute by allowing better CNS drug penetration.

Clinical Data It is well documented that patients with HER2 overexpressing tumours seem to have a higher incidence

99

of intracranial metastases (up to 30%) (Clayton et al., 2004; Bria et al., 2008), and the brain is often described as a “sanctuary site” due to difficulties with cytotoxic agents or monoclonal antibodies such as Trastuzumab crossing the “blood-brain barrier”. Studies have shown that HER2 status is maintained between the primary breast tumour and CNS metastases (Fuchs et al., 2002), which has increased interest in developing HER2 targeted strategies for treatment of CNS disease. In a study by Dawood et al. analyzing the hazard of death in patients with CNS disease it was found that the hazard of death was increased (HR, 1.66; 95% CI, 1.31–2.12; p < 0.0001) in those who had HER2-ve disease or HER2+ disease and had never received Trastuzumab compared with those who had HER2+ disease and had received Trastuzumab at anytime (Dawood et al., 2008). It is well documented that there is little CSF penetration of Trastuzumab across an intact BBB (serum to CSF ratio 420:1) but Stemmler et al. have shown increased CSF concentrations in the presence of lepto-meningeal disease (ratio 49:1) and following WBRT (ratio 76:1) suggesting that there may be a locally therapeutic benefit in continuing Trastuzumab in the presence of CNS disease as well as the survival benefit gained from increased systemic disease control (Stemmler et al., 2007). Responses have been documented to Capecitabine monotherapy in both treated and untreated CNS metastases from breast cancer (Wang et al., 2001; Ekenel et al., 2007) and there are anecdotal reports of responses to a variety of other chemotherapeutic agents but numbers are small. Poor responses of HER2+ CNS disease to agents such as Temozolomide suggest that therapies most likely to obtain CNS responses in HER2+ breast cancer are those that have documented activity against systemic disease. The evidence cited above showing activity for lapatinib in pre-clinical models of CNS disease (Gril et al., 2008) has been followed by similar results from clinical trials of women with HER2+ MBC with CNS metastases. There are several reports of CNS responses to lapatinib based therapy. Following an initial single centre phase II study of lapatinib monotherapy in patients with HER2-positive breast cancer who developed progressive CNS metastases after prior loco-regional therapy, (Lin et al., 2008) a larger multi-centre phase II study involving 242 patients reported a 6% CNS ORR to lapatinib alone (Lin et al., 2009). In an exploratory analysis a further

100

21% patients experienced a ≥20% volumetric reduction in CNS lesions, with many having significant improvement in neurological symptoms. An association was observed between volumetric reduction and improvement in PFS and neurological signs and symptoms. Subsequently within this study, a cohort of 50 patients whose CNS disease progressed on lapatinib monotherapy entered an extension phase involving treatment with both Capecitabine and lapatinib, with an ORR of 20%, with 40% patients having a volumetric reduction of ≥20% in their CNS lesions. In EGF100151, a pivotal phase III trial comparing Capecitabine monotherapy with the combination of Capecitabine and Lapatinib, objective tumour response rates of 22% were reported for lapatinib and Capecitabine compared with 14% in the Capecitabine alone group, with a CBR of 27 and 18% respectively (Geyer et al., 2006). Although numbers were small, fewer patients in the combination arm developed CNS metastases compared with those on Capecitabine monotherapy (4 versus 13), suggesting that lapatinib may also have a role to play in prevention (Cameron et al., 2008). The global LEAP study (Capri et al., 2010) was a single arm open label clinical trial that allowed access to lapatinib in combination with Capecitabine for patients with HER2 over-expressing breast cancers, who had previously received Anthracycline, Taxane and Trastuzumab, with the aim of providing clinical benefit whilst awaiting regulatory approval in individual countries, and further validating safety data observed in EGF100151 in a larger population. 4283 patients from 45 countries worldwide were enrolled in the study. Entry criteria were similar to those in the pivotal EGF100151 trial; in contrast however, prior Capecitabine use was permitted and patients with cerebral metastatic disease could participate providing they were on a steroid dose ≤2 mg/day of dexamethasone (or equivalent) and were performance status (PS) 0–2. The median treatment duration was 24.7 weeks and the most common drug-related serious adverse events (SAEs) were diarrhea (9.7%), vomiting (4.3%), and nausea (2.4%) 0.5% of patients experienced a decrease in left ventricular ejection fraction consistent with the known safety profile of lapatinib. The median PFS and OS were 21.1 [95% confidence interval (CI) = 20.1–22.3] and 39.6 (95% CI = 37.7–40.7) weeks, respectively (n = 4,006), and subgroup analysis

S. Sutherland and S. Johnston

showed longer PFS and OS in patients who had not received prior Capecitabine. In a separate report of 138 patients with brain metastases treated with Lapatinib and Capecitabine within the LEAP study and the French ATU program, Boccardo et al. (2008) reported an ORR of 18% (CR and PR) with a further 47% of patients achieving stable disease, although the duration of SD was not recorded. In addition they noted an improvement in neurological symptoms in 25% of patients. In a further cohort of patients from the 5 largest participating UK sites of the LEAP study, efficacy was looked at in more detail, both systemically and in the CNS. Tumour response rate, time to disease progression (TTP) and treatment related toxicities were assessed and compared with those reported in EGF100151. A separate analysis of the cohort of patients with CNS metastases was performed to assess response in this group. The UK subgroup of 34 patients with CNS metastases who had progressed following prior WBRT compared favourably with these previous studies, with an investigator assessed response rate of 21% and evidence of clinical benefit in half of the treated patients (Sutherland et al., 2010). For patients with symptomatic CNS disease progressing after prior WBRT, prognosis is normally very poor (weeks to a few months at most), so this level of benefit in a hitherto “difficult-to-treat” scenario represents meaningful clinical efficacy. The MRI scans (Fig. 10.1) from a 61 year old patient treated on the LEAP study, illustrate a partial response to Capecitabine in combination with Lapatinib. The scans on the left were taken at baseline and those on the right after 2 cycles of treatment. The patient also had an improvement in her neurological symptoms, which consisted of right sided facial twitching and a feeling of loss of awareness and abnormal sensation in her left hand. These had resolved completely after 3 cycles.

Current Trials For patients with CNS metastatic disease clinical trials are often limited and specific studies are required to address how best to use HER2 targeted therapies such as lapatinib in this situation. The “LANTERN” Study (Fig. 10.2) is a randomised, phase II, multi-centre, prospective, controlled, open label trial in patients with

10 Central Nervous System Tumours in Women

101

Fig. 10.1 Scans showing a partial response to treatment with Lapatinib and Capecitabine in the LEAP study

Before

Before advanced HER2+ MBC pre-treated with Trastuzumab who have newly diagnosed CNS metastases, with equal randomisation comparing lapatinib in combination with Capecitabine versus continued Trastuzumab in combination with Capecitabine after local therapy (WBRT/stereotactic radio surgery (SRS)). This will attempt to assess whether there is any difference in time to progression of CNS metastases between the 2 arms and also assess PFS, OS, CNS response rates, clinical benefit, neurological symptom improvement and quality of life as secondary endpoints. There is also

After

After a functional imaging sub-study which will investigate the relationship between permeability of the BBB and response to treatment. In addition, the “CEREBRAL” trial (Fig. 10.3) has been established in HER2+ MBC to see whether lapatinib can prevent the development of CNS metastases. This trial is a phase III Randomized Multi-centre Open label study comparing Lapatinib Plus Capecitabine Versus Trastuzumab Plus Capecitabine. It is designed specifically to look at the effect of Lapatinib on the incidence of brain metastases as a site of first relapse in

102 Fig. 10.2 Lantern Trial flow diagram

S. Sutherland and S. Johnston

Patient has early or metastatic HER2+ breast cancer and is receiving Trastuzumab

Patient diagnosed with brain metastases, +/– extra cranial disease progression and receives local therapy (WBRT or SRS)

Randomisation

Lapatinib (1250mg od) plus Capecitabine (1000mg/m2 bd) d1–14 21 day cycles until PD

Trastuzumab (6mg/kg 3weekly) day 1 plus Capecitabine (1000mg/m2 bd) d1–14 21 day cycles until PD

Disease progression (CNS or non-CNS) as assessed by RECIST LANTERN TRIAL TREATMENT STOPPED

Collection of any subsequent treatment and survival data every 12 weeks until 6 months post – randomisation or death whichever is earlier

patients with MBC who have a negative screening MRI brain scan, but are at risk of developing CNS disease due to their recurrence of HER2+ MBC. The study is currently open for recruitment in Europe and the USA and is aiming for a total of 650 patients. Patients are excluded from the study if they have had prior lapatinib or an ErbB2 inhibitor other than Trastuzumab and are not allowed to have pre-existing CNS disease. (Previous Capecitabine use is allowed) The study aims to report a final analysis once all patients have been followed for a minimum of 12 months. This has the potential to be a pivotal study as it will provide evidence to whether or not the theoretical and pre-clinical advantages of lapatinib over Trastuzumab for HER2+ CNS disease translate into genuine clinical benefit for patients. A further study, “COMPLETE” is a first line trial of lapatinib in combination with a Taxane versus Trastuzumab and a Taxane for patients with HER2+ MBC. While designed to compare overall

efficacy of Lapatinib versus Trastuzumab in the first line metastatic setting, a pre-specified secondary endpoint is to compare the frequency of CNS metastases between the 2 arms. This will provide additional evidence for differentiating any treatment benefit between lapatinib and Trastuzumab with respect to CNS metastases. Other trials in the adjuvant setting such as ALTTO, a 4 arm phase III adjuvant study comparing the activity of monotherapy with either lapatinib or Trastuzumab vs the combination of concurrent lapatinib and Trastuzumab vs sequential Trastuzumab and lapatinb, may help to provide more insight into the role of lapatinib as regards the incidence and time of onset of HER2+ CNS metastases. Again, like COMPLETE, comparison of the incidence of CNS metastases between the 4 arms is a pre-specified secondary endpoint to determine whether Lapatinib has a role to play in delaying or preventing the onset of CNS disease in HER2+ metastatic breast cancer.

10 Central Nervous System Tumours in Women

103

Fig. 10.3 Flow diagram for “CEREBRAL” Lapatinib Plus Capecitabine Versus Trastuzumab Plus Capecitabine in ErbB2 (HER2) Positive Metastatic Breast Cancer

Patients with HER2+ metastatic breast cancer, who have received prior anthracycline or Taxane chemotherapy

Screening Brain MRI shows no evidence of CNS metastases

Randomisation

Capecitabine 1000mg/m2 po bd days1–14 Plus Lapatinib 1250 mg po od 21 day cycles until PD

Capecitabine 1000mg/m2 po bd days1−14 Plus Trastuzumab (6mg/kg 3 weekly) day 1 21 day cycles until PD

Follow up for minimum of 1 year or until death or with drawn from the study

Primary outcome Incidence of CNS metastases as first site of relapse

The Future As knowledge about the intracellular part of the HER2 signaling pathways increases, so does our understanding of the mechanisms of resistance to therapy which will allow development of drugs to overcome this. Many such compounds are being explored in the laboratory and are showing activity in Phase I trials at the present time including PI3-Kinase, AKT, mTOR and HSP90 Inhibitors. Newer monoclonal antibody drugs such as pertuzumab which prevents HER2 dimerization and T-DM1 an antibody drug conjugate consisting of Trastuzumab with an attached cytotoxic anti-microtubule agent (DM1) have also shown promise in the treatment of HER2+ MBC, but like Trastuzumab are large antibody molecules and are therefore unlikely to have a direct impact on CNS disease. They may however contribute to improved overall survival of patients with HER2+ MBC and CNS disease by providing further gains in systemic disease control after progression on Trastuzumab.

New routes of administration of existing drugs may also be developed and it is possible that there will be intra-thecal administration of HER2 targeted agents for those with CNS disease in an attempt to overcome both the BBB and the effluxing of agents by the P-glycoprotein pump which is so prevalent in neurological tissue. There have been isolated case reports of intra-thecal administration of Trastuzumab (Laufman and Forsthoefel, 2001; Platini et al., 2006; Stemmler et al., 2006) but as yet this has not been widely investigated. However as numbers of patients with CNS metastases from HER2+ MBC are increasing and preclinical data (Grossi et al., 2003) suggests there may be a benefit, this is likely to be another avenue for exploration. In addition, oral TKIs are being looked at as radio-sensitizers in combination with radiotherapy and there is an ongoing phase I trial, at the USA’s Dana Farber Institute, involving the use of Lapatinib and WBRT (NCT00470847), as well as a planned phase II study using Capecitabine in

104

combination with WBRT followed by a combination of Capecitabine and Sunitinib. Other trials involving Lapatinib in combination with cytotoxics e.g. Temozolomide (NCT00614978) or Capecitabine or Topotecan (EGF107671,) are currently recruiting. Prophylactic Cranial Irradiation (PCI) such as is used in small cell lung cancer (SCLC) has also been considered as a potential preventive treatment option for those with HER2+ disease due to the high risk of CNS metastases. A recent retrospective review by Graesslin et al. (2010) looked at the ability to predict development of CNS metastases in patients with MBC and developed a nomogram which they subsequently validated. They found the risk of CNS disease to be independently related to younger age, higher histological grade, short interval between diagnosis and development of metastatic disease, number of sites of metastatic disease and Estrogen and Progesterone Receptor (ER/PR) expression and HER2 status. This would potentially allow for more accurate selection of those most at risk of CNS disease and perhaps would facilitate the development of trials aimed at prevention. This has not previously been done due to the associated morbidity and long term neurological sequelae which can result from WBRT.

Conclusions The outlook for patients with HER2+ MBC changed dramatically with the advent of Trastuzumab and this in turn has prolonged the survival of patients who develop CNS disease. It is important that these patients receive adequate treatment both in terms of initial local therapy (Surgery, SRS, WBRT) and ongoing systemic HER2 targeted therapy such as lapatinib or Trastuzumab in order to maintain disease control. Lapatinib in combination with Capecitabine has been shown to be an effective and reasonably well tolerated treatment for patients with disease progression following prior Trastuzumab-containing regimens, particularly in those who are Capecitabine naive and in those who have progressive and symptomatic CNS metastases following prior WBRT (Capri et al., 2010; Sutherland et al., 2010) Strategies aimed at prevention are important and Lapatinib may well have a role to play either as part of primary treatment in high

S. Sutherland and S. Johnston

risk HER2+ early breast cancer or as an adjunct to first line MBC in an attempt to reduce or delay the onset of CNS disease.

References Barnholtz-Sloan JS, Sloan AE, Davis FG, Vigneau FD, Lai P, Sawaya RE (2004) Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System. J Clin Oncol 22(14):2865–2872 Bendell JC, Domchek SM, Burstein HJ, Younger J, Kuter I, Bunnell C, Rue M, Gelman R, Winer E (2003) Central nervous system metastases in women who receive trastuzumabbased therapy for metastatic breast carcinoma. Cancer 97(12):2972–2977 Boccardo F, Kaufman B, Baselga J, Dieras V, Link J, Casey MA, Fittipaldo A, Oliva C, Zembryki D, Rubin SD (2008) Evaluation of lapatinib (Lap) plus capecitabine (Cap) in patients with brain metastases (BM) from HER2+ breast cancer (BC) enrolled in the Lapatinib Expanded Access Program (LEAP) and French Authorisation Temporaire d’Utilisation (ATU). J Clin Oncol 26(May 20 suppl):Abstract 1094 Bria E, Cuppone F, Fornier M, Nistico C, Carlini P, Milella M, Sperduti I, Terzoli E, Cognetti F, Giannarelli D (2008) Cardiotoxicity and incidence of brain metastases after adjuvant trastuzumab for early breast cancer: the dark side of the moon? A meta-analysis of the randomized trials. Breast Cancer Res Treat 109(2):231–239 Burstein HJ, Lieberman G, Slamon DJ, Winer EP, Klein P (2005) Isolated central nervous system metastases in patients with HER2-overexpressing advanced breast cancer treated with first-line trastuzumab-based therapy. Ann Oncol 16(11):1772–1777 Cameron D, Casey M, Press M, Lindquist D, Pienkowski T, Romieu CG, Chan S, Jagiello-Gruszfeld A, Kaufman B, Crown J, Chan A, Campone M, Viens P, Davidson N, Gorbounova V, Raats JI, Skarlos D, Newstat B, Roychowdhury D, Paoletti P, Oliva C, Rubin S, Stein S, Geyer C (2008) A phase III randomized comparison of lapatinib plus capecitabine versus capecitabine alone in women with advanced breast cancer that has progressed on trastuzumab: updated efficacy and biomarker analyses. Breast Cancer Res Treat 112(3):533–543 Capri G, Chang J, Chen SC, Conte P, Cwiertka K, Jerusalem G, Jiang Z, Johnston S, Kaufman B, Link J, Ro J, Schutte J, Oliva C, Parikh R, Preston A, Rosenlund J, Selzer M, Zembryki D, De Placido S (2010) An open-label expanded access study of lapatinib and capecitabine in patients with HER2-overexpressing locally advanced or metastatic breast cancer. Ann Oncol 21(3):474–480 Clayton AJ, Danson S, Jolly S, Ryder WD, Burt PA, Stewart AL, Wilkinson PM, Welch RS, Magee B, Wilson G, Howell A, Wardley AM (2004) Incidence of cerebral metastases in patients treated with trastuzumab for metastatic breast cancer. Br J Cancer 91(4):639–643

10 Central Nervous System Tumours in Women Dawood S, Broglio K, Esteva FJ, Ibrahim NK, Kau SW, Islam R, Aldape KD, Yu TK, Hortobagyi GN, Gonzalez-Angulo AM (2008) Defining prognosis for women with breast cancer and CNS metastases by HER2 status. Ann Oncol 19(7): 1242–1248 Eichler AF, Kuter I, Ryan P, Schapira L, Younger J, Henson JW (2008) Survival in patients with brain metastases from breast cancer: the importance of HER-2 status. Cancer 112(11):2359–2367 Ekenel M, Hormigo AM, Peak S, Deangelis LM, Abrey LE (2007) Capecitabine therapy of central nervous system metastases from breast cancer. J Neurooncol 85(2):223–227 Fuchs IB, Loebbecke M, Buhler H, Stoltenburg-Didinger G, Heine B, Lichtenegger W, Schaller G (2002) HER2 in brain metastases: issues of concordance, survival, and treatment. J Clin Oncol 20(19):4130–4133 Geyer CE, Forster J, Lindquist D, Chan S, Romieu CG, Pienkowski T, Jagiello-Gruszfeld A, Crown J, Chan A, Kaufman B, Skarlos D, Campone M, Davidson N, Berger M, Oliva C, Rubin S, Stein S, Cameron D (2006) Lapatinib plus capecitabine for HER2-positive advanced breast cancer. N Engl J Med 355(26):2733–2743 Gori S, Rimondini S, De Angelis V, Colozza M, Bisagni G, Moretti G, Sidoni A, Basurto C, Aristei C, Anastasi P, Crino L (2007) Central nervous system metastases in HER2 positive metastatic breast cancer patients treated with trastuzumab: incidence, survival, and risk factors. Oncologist 12(7):766–773 Graesslin O, Abdulkarim BS, Coutant C, Huguet F, Gabos Z, Hsu L, Marpeau O, Uzan S, Pusztai L, Strom EA, Hortobagyi GN, Rouzier R, Ibrahim NK (2010) Nomogram to predict subsequent brain metastasis in patients with metastatic breast cancer. J Clin Oncol 28(12):2032–2037 Gril B, Palmieri D, Bronder JL, Herring JM, Vega-Valle E, Feigenbaum L, Liewehr DJ, Steinberg SM, Merino MJ, Rubin SD, Steeg PS (2008) Effect of lapatinib on the outgrowth of metastatic breast cancer cells to the brain. J Natl Cancer Inst 100(15):1092–1103 Grossi PM, Ochiai H, Archer GE, McLendon RE, Zalutsky MR, Friedman AH, Friedman HS, Bigner DS, Sampson JH (2003) Efficacy of intracerebral microinfusion of trastuzumab in an athymic rat model of intracerebral metastatic breast cancer. Clin Cancer Res 9(15):5514–5520 Kirsch DG, Hochberg FH (2003) Targeting HER2 in brain metastases from breast cancer. Clin Cancer Res 9(15): 5435–5436 Konecny GE, Pegram MD, Venkatesan N, Finn R, Yang G, Rahmeh M, Untch M, Rusnak DW, Spehar G, Mullin RJ, Keith BR, Gilmer TM, Berger M, Podratz KC, Slamon DJ (2006) Activity of the dual kinase inhibitor lapatinib (GW572016) against HER-2-overexpressing and trastuzumab-treated breast cancer cells. Cancer Res 66(3):1630–1639 Laufman LR, Forsthoefel KF (2001) Use of intrathecal trastuzumab in a patient with carcinomatous meningitis. Clin Breast Cancer 2(3):235 Leyland-Jones B (2009) Human epidermal growth factor receptor 2-positive breast cancer and central nervous system metastases. J Clin Oncol 27(31):5278–5286

105 Lin NU, Carey LA, Liu MC, Younger J, Come SE, Ewend M, Harris GJ, Bullitt E, Van den Abbeele AD, Henson JW, Li X, Gelman R, Burstein HJ, Kasparian E, Kirsch DG, Crawford A, Hochberg F, Winer EP (2008) Phase II trial of lapatinib for brain metastases in patients with human epidermal growth factor receptor 2-positive breast cancer. J Clin Oncol 26(12):1993–1999 Lin NU, Dieras V, Paul D, Lossignol D, Christodoulou C, Stemmler HJ, Roche H, Liu MC, Greil R, Ciruelos E, Loibl S, Gori S, Wardley A, Yardley D, Brufsky A, Blum JL, Rubin SD, Dharan B, Steplewski K, Zembryski D, Oliva C, Roychowdhury D, Peoletti P, Winer E (2009) Multicenter phase II study of lapatinib in patients with brain metastases from HER2-positive breast cancer. Clin Cancer Res 15(4):1452–1459 Meric F, Hung MC, Hortobagyi GN, Hunt KK (2002) HER2/neu in the management of invasive breast cancer. J Am Coll Surg 194(4):488–501 Paik S, Hazan R, Fisher ER, Saas RE, Redmond C, Schlessinger J, Lippman ME, King CR (1990) Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project: prognostic significance of erbB-2 protein overexpression in primary breast cancer. J Clin Oncol 8(1):103–112 Park YH, Park MJ, Ji SH, Yi SY, Lim DH, Nam DH, Lee JI, Park W, Choi DH, Huh SJ, Ahn JS, Kang WK, Park K, Im YH (2009) Trastuzumab treatment improves brain metastasis outcomes through control and durable prolongation of systemic extracranial disease in HER2-overexpressing breast cancer patients. Br J Cancer 100(6):894–900 Platini C, Long J, Walter S (2006) Meningeal carcinomatosis from breast cancer treated with intrathecal trastuzumab. Lancet Oncol 7(9):778–780 Polli JW, Humphreys JE, Harmon KA, Castellino S, O’Mara MJ, Olson KL, John-Williams LS, Koch KM, Serabjit-Singh CJ (2008) The role of efflux and uptake transporters in N-{3chloro-4-[(3-fluorobenzyl)oxy]phenyl}-6-[5-({[2-(methylsulfonyl)ethyl]amino}methyl)-2-furyl]-4-quinazolinamine (GW572016, lapatinib) disposition and drug interactions. Drug Metab Dispos 36(4):695–701 Richards GM, Khuntia D, Mehta MP (2007) Therapeutic management of metastatic brain tumors. Crit Rev Oncol Hematol 61(1):70–78 Smith I, Procter M, Gelber RD, Guillaume S, Feyereislova A, Dowsett M, Goldhirsch A, Untch M, Mariani G, Baselga J, Kaufmann M, Cameron D, Bell R, Coleman R, Wardley A, Harbeck N, Lopez RI, Mallmann P, Gelmon K, Wilcken N, Wist E, Sanchez Rovira P, Piccart-Gebhart MJ (2007) 2-year follow-up of trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer: a randomised controlled trial. Lancet 369(9555):29–36 Stemmler HJ, Kahlert S, Siekiera W, Untch M, Heinrich B, Heinemann V (2006) Characteristics of patients with brain metastases receiving trastuzumab for HER2 overexpressing metastatic breast cancer. Breast 15(2):219–225 Stemmler HJ, Schmitt M, Willems A, Bernhard H, Harbeck N, Heinemann V (2007) Ratio of trastuzumab levels in serum and cerebrospinal fluid is altered in HER2-positive breast cancer patients with brain metastases and impairment of blood-brain barrier. Anticancer Drugs 18(1):23–28

106 Sutherland S, Ashley S, Miles D, Chan S, Wardley A, Davidson N, Bhatti R, Shehata M, Nouras H, Camburn T, Johnston SR (2010) Treatment of HER2-positive metastatic breast cancer with lapatinib and capecitabine in the lapatinib expanded access programme, including efficacy in brain metastases – the UK experience. Br J Cancer 102(6): 995–1002

S. Sutherland and S. Johnston Tsukada Y, Fouad A, Pickren JW, Lane WW (1983) Central nervous system metastasis from breast carcinoma. Autopsy study. Cancer 52(12):2349–2354 Wang ML, Yung WK, Royce ME, Schomer DF, Theriault RL (2001) Capecitabine for 5-fluorouracil-resistant brain metastases from breast cancer. Am J Clin Oncol 24(4): 421–424

Part II

Biomarkers and Diagnosis

Chapter 11

Functional Role of the Novel NRP/B Tumor Suppressor Gene Theri Leica Degaki, Marcos Angelo Almeida Demasi, and Mari Cleide Sogayar

Abstract Elucidation of the mechanisms underlying the aggressive nature of GBM (glioblastoma multiforme) is important to improve gene-, radio- and chemo- therapy. The rat C6 glioma cell line has been used as an experimental model system for GBM, providing several promising results. Upon searching for glucocorticoid-regulated cDNAs sequences associated with the transformed to normal phenotypic reversion of C6/ST1 rat glioma cells (variant cell line, which is hyper-responsive to glucocorticoid treatment and was isolated from C6 glioma cells), we identified the nuclear restrict protein in brain (NRP/B) as a novel rat gene. Nuclear matrix proteins have been involved in malignant transformation, such as alterations in nuclear shape, DNA content and proliferative state. The nuclear matrix protein NRP/B is identified as a novel gene expressed in the nervous system, including brain cell lines and human brain tumors and, also, in other tissues (adipose, breast and hairy cell leukemia). Sequence analysis indicates that the rat NRP/B displays a high level of homology with the equivalent orthologs genes from other organisms. Over expression of the NRP/B cDNA in C6/ST1 cells suppresses anchorage independence in vitro and tumorigenicity in vivo, altering their malignant nature towards a more benign phenotype. Some mechanisms are discussed here, based on previous studies from other groups: the interaction of NRP/B with RB1 and p53. Therefore,

M.C. Sogayar () Department of Biochemistry, Chemistry Institute, NUCEL-Cell and Molecular Therapy Center, University of Sao Paulo, Sao Paulo 05508-900 SP, Brazil e-mail: [email protected]

NRP/B may be postulated as a novel tumor suppressor gene, with possible relevance for glioblastoma therapy. Keywords NRP/B · Glioblastoma multiforme · Glucocorticoid · C6/ST1 rat glioma cell · Functional analysis · Tumor suppressor gene

Introduction Gliomas Gliomas have been defined as tumors which display histopathological, immunohistochemical and ultrastructural evidence of glial differentiation. Four malignancy grades are recognized by the WHO (World Health Organization), with grade I tumor being the biologically least aggressive and grade IV, the most aggressive tumors (Louis et al., 2007). This classification is based on presumed histogenesis together with grading of individual tumor classes, reflecting the anticipated biological behaviour. Thus, higher grade tumors (grades III and IV), in the absence of treatment, are expected to follow a more aggressive clinical course than their lower grade counterparts (grades I and II). At the molecular level, many genetic alterations targeting cell cycle regulatory genes are involved in malignant progression from grade II to grade III and GBM, namely: p16, cyclin-dependent kinases (Cdks) 4 and 6, cyclin D1 and the retinoblastoma (p110RB ) protein. This occurs due to loss of heterozygosity (LOH) in some chromosomes or to specific mutations, namely: chromosome 9p loss occurs in 50% of GBM and grade III, and affects the region

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_11, © Springer Science+Business Media B.V. 2011

109

110

of CDKN2A gene which encodes the p16 and ARF proteins; chromosome 13q loss occurs in one third to one half of high-grade astrocytomas, affecting the p110RB gene; chromosome 10 loss is found in 60– 95% of GBMs, affecting the PTEN tumor suppressor gene. Amplification of the cdk4 gene, located on chromosome 12q, facilitates progression to GBM in 15% of malignant gliomas. The p53 tumor suppressor networks are also disrupted in GBM and amplification of EGFR (epidermal growth factor receptor) gene are found in 40% of GBMs (Huse and Holland, 2010; Louis, 2006). The carcinogenesis of gliomas, particularly the molecular defects which underlie initiation and progression of these tumors, is ill-defined, hampering the development of targeted, individualized and successful intervention for gliomas. Despite the difficulties to cross the blood-brain barrier, some strategies have been designed to overcome this problem, such as: increasing the drugs dosage, conjugation of therapeutic agents to lipophilic moieties or to other vectors, co-administration with inhibitors of the blood-brain barrier transporters or using the invasive approaches, in which the anticancer agent is infused directly into the tumor (Huse and Holland, 2010).

Nuclear Restrict Protein in Brain The nuclear matrix plays a pivotal role in processing of the genetic information, acting on transcription, RNA splicing/transport, DNA replication and repair. Transcription factors, including tumor suppressors (e.g. RB1, TP53) and hormone receptors (estrogen receptor), dynamically associate with specific nuclear matrix sites that support their assembly into functional macromolecular complexes (Davie et al., 1998). Irregular nuclear appearance, chromatin rearrangements and changes in the composition of nuclear matrix proteins in malignant cells have recently been mechanistically linked to functional changes that promote tumor development, because the nuclear matrix coordinates processes occurring at chromatin sites, which are under stringent cell cycle control (He et al., 2008). Nuclear matrix protein NRP/B (nuclear restricted protein/brain) (Kim et al., 1998), also termed ENC1 (Hernandez et al., 1997) or PIG10 (Polyak et al.,

T.L. Degaki et al.

1997), a member of the BTB/Kelch repeat family, was shown to be expressed in the nucleus of neuronal cells, as a component of the nuclear matrix. NRP/B was first implicated as a target for p53-transcriptional activation in the DLD1 colorectal cancer cell line, using the SAGE (Serial Analysis of Gene Expression) technology (Polyak et al., 1997). Later on in 1997, researchers isolated a cDNA fragment, corresponding to murine Enc-1, from mRNA of mouse brain at different stages of development. ENC-1 co-localized with the actin cytoskeleton with possible physical association (Hernandez et al., 1997). Only in the following year (1998) researchers isolated the human NRP/B gene, using the single pass sequencing technology, which identified proteins of potential importance in brain development (Kim et al., 1998). The NRP/B protein is predominantly expressed in primary neurons, interacts with RB1 during neuronal differentiation and participates in the regulation of neuronal process formation (Hernandez et al., 1998; Kim et al., 1998). The NRP/B mRNA expression was detected in astrocytes and glial cells derived from brain tumors, suggesting its involvement in brain tumorigenesis (Kim et al., 2000). The structural analysis of the NRP/B protein revealed that the N-terminal region is α-helical, while the C-terminal region is a β-sheet, consisting of a six times repeated motif which forms a superbarrel structure. The BTB domain, localized in the Nterminal region of NRP/B, has been proposed to mediate protein-protein interactions, with 5–10% of the zinc finger proteins in humans being estimated to contain these domains (Albagli et al., 1995). Recently, the BTB-domain of NRP/B was shown to be essential for nuclear localization and interaction with RB1 had been described as an important regulator of neuronal differentiation. Based on threedimensional modeling of NRP/B, a mutant (NRP/BBTB) was generated and shown to inhibit cell cycle progression and the interaction with RB1 and neurite outgrowth (Kim et al., 2005). The six copies of Kelch repeats in the C-terminal region of NRP/B are important for cytoskeletal organization and function. Mutations in this region are found both in brain tumor cell lines and in primary GBM tissues (Liang et al., 2004). More importantly, in this same study, NRP/B mutations in the kelch domain conferred cell growth advantage by elevating ERK activation and suppressing cellular apoptosis in response to stressful

11 Functional Role of the Novel NRP/B Tumor Suppressor Gene

stimulus, by inhibiting TP53-mediated caspase activation and reducing its binding affinity to actin. Abundant expression of NRP/B was observed in the cytoplasm of several brain cell lines and in human brain tumors, including glioblastomas and astrocytomas (Kim et al., 2000), but alterations and gene mutations were found in these samples. These alterations may contribute to brain tumorigenesis by promoting cell proliferation, affecting the cellular apoptosis machinery through a TP53-mediated caspase pathway and inducing a shift of NRP/B localization from the nucleus to the cytoplasm (Liang et al., 2004). More recently, a novel role for NRP/B was discovered, related to oxidative stress responses in breast cancer cells via the Nrf2 (NF-E2-related factor 2) pathwaymediated NQO1 (NAD(P)H:quinine oxidoreductase 1) transcriptional activity (Seng et al., 2007). A more recent study describing the mechanism by which the NRP/B modulates Nrf2-dependent NQO1 induction in cellular protection against ROS (reactive oxygen species), provides insights into the biological function of NRP/B as a modulator of cellular protection in brain tumors (Seng et al., 2009). Additionally, another work describes the physical association of NRP/B and F-actin, implicating a role for NRP/B in cytoskletal reorganization in the differentiation of granulose cells, which occurs during the ovulatory process in the rat (Kim et al., 2009). In cancer, NRP/B has been associated not only with the human nervous system and with breast cancer cells (Seng et al., 2007), but, also, with colorectal carcinomas (Fujita et al., 2001) and hairy cell leukemia (Hammarsund et al., 2004).

111

of the C6 rat glioma cell line is inhibited by glucocorticoid hormones (GC), which constitute a class of corticosteroids produced by the adrenal gland, being routinely prescribed for patients with brain tumors to decrease edema and, also, due to their anti-neoplastic properties (Nahaczewski et al., 2004). GC enters the cells through passive diffusion forming a complex with its specific receptor protein GR (glucocorticoid receptor). The GC-GR complex then undergoes activation and enters the cell nucleus, where it binds to the DNA, leading to a number of biological responses induced by these hormones (Beato et al., 1996). To study the molecular mechanisms of GC as an anti-tumor agent in glioma cells, we have previously isolated the C6/ST1 variant (Armelin et al., 1983) from the C6 rat cell line. This variant is hyper-responsive to GC, going from a fully transformed and tumorigenic phenotype to a normal one, both in vitro and in vivo (Armelin and Armelin, 1983). Using genomics and proteomics approaches, we found several GCregulated genes which are associated with the transformed to normal phenotypic reversion that occurs in the C6/ST1 variant (Armelin et al., 1996; Demasi et al., 2007; Vedoy and Sogayar, 2002). Among these, we identified the rat Nrp/b (r-Nrp/b) gene (Vedoy and Sogayar, 2002), with a single transcript of approximately 5.5 kb, that is induced upon a 5 h treatment with hydrocortisone.

Cloning and Functional Characterization of the Rat Full Length cDNA Corresponding to the Nuclear Restrict Protein in Brain Gene Associated with Rat Brain Tumor Model the Transformed to Normal Phenotypic Model systems are important in the study of gliomas, Reversion of RAT C6/ST1 Glioma Cells as well in testing targeted therapeutics. The widely employed rat brain tumor models, have provided a wealth of information on the biology, biochemistry and experimental therapeutics of brain tumors in experimental neuro-oncology and allowed for precision testing of some candidate tumor suppressor and oncogenes in appropriate cellular context (Barth and Kaur, 2009). The rat C6 glioma cell is a widely used experimental model system for glioblastoma (GBM) growth, invasion and metastasis and for the design and evaluation of anticancer therapies (Miura et al., 2008). Growth

Based on the cDNA sequence inferred by searching the NCBI/GenBankTM databases, we amplified the coding region of the NRP/B rat ortholog, whose sequence was deposited in GenBank database under the accession number AY669396. In order to assess the functional role of r-NRP/B in the phenotypic reversion of the rat glioma C6/ST1 cell line, cells were transduced with the pLPCX-NRP/B retroviral construct. The overexpression of r-NRP/B was confirmed at the RNA (Northern blot) and protein (Western

112

blot and Immunofluoresce) levels. The immunofluorescence pattern of C6/ST1 and C6/ST1-NRP/B cells labeled with a specific monoclonal antibody to NRP/BENC-1 showed that r-NRP/B is distributed throughout the cytoplasm in positive cells, with the level of expression being higher in C6/ST1 cells transduced with retrovirus containing NRP/B gene (C6/ST1-NRP/B), when compared to the parental C6/ST1cells. Our results indicated that r-NRP/B overexpression did not directly affect C6/ST1 cell proliferation in monolayer cultures. However, C6/ST1-NRP/B cells displayed a significantly decreased ability to form colonies in the soft agar colony formation assay when compared with the parental C6/ST1 cell line, suggesting a positive correlation between NRP/B overexpression and decreased survival/growth advantage of C6/ST1 cells under anchorage-independent conditions. Anchorage-independent growth is considered to be the in vitro equivalent of tumorigenesis in vivo. The tumorigenesis results shown in Fig. 11.1a, obtained in two independent experiments, indicated that both the frequency of tumor formation and the average tumor size were decreased at 6 weeks in mice injected with the C6/ST1-NRP/B cells, when compared with those induced by C6/ST1 parental cells, or by C6/ST1-EGFP or C6/ST1-pLPCX (empty vector) cells. Among the eight animals injected with the C6/ST1-NRP/B cells, only one mouse displayed a small tumor (Fig. 11.1b). To determine whether the lower tumorigenicity was, in fact, induced by overexpression of the NRP/B transgene, we examined the expression of NRP/B by immunohistochemistry in the single C6/ST1-NRP/B tumor formed in comparison with other tumors induced by C6/ST1-pLPCX cells. A more pronounced immunohistochemical staining to NRP/B was observed in tumors induced by C6/ST1NRP/B cells than by the vector alone control cells (Fig.11.1c).

Discussion – Role of Nuclear Restrict Protein in Brain as a Tumor Suppressor Gene Our study elucidated the general genomic structure of the r-Nrp/b. Alignment of the product of rat Nrp/b with its orthologs indicated that this gene is highly conserved among different species. In particular, the

T.L. Degaki et al.

amino acid sequence identity between rat, human and mouse NRP/B is 99% (Degaki et al., 2009). C6/ST1 cell fractionation followed by Western blotting, indicated that abundant NRP/B expression is evident in the cytoplasmic fraction, with a weaker signal found in the nuclear matrix and in cytoskeletonassociated and nuclear proteins fractions (Demasi et al., 2007). In brain tumors, NRP/B expression is described mainly in the cytoplasm (Liang et al., 2004), in agreement with the immunocytochemistry results we obtained with C6/ST1 cells transduced with the NRP/B retrovirus (Degaki et al., 2009). GCs have been used for more than 40 years in the treatment of intracranial gliomas, however, the exact molecular mechanisms by which glucocorticoids affect tumor growth, reduce tumor-associated edema and improve the patients clinical status are still largely unknown (Piette et al., 2006). Recent results obtained in our laboratory demonstrate enhancement of p53 activity with inhibition of cell proliferation upon glucocorticoid receptor activation in C6/ST1 cells (Macedo, 2007). Dexamethasone treatment of rat hippocampus results in elevated levels of p53, suggesting p53 to be a player in GR-mediated growth arrest (Almeida et al., 2000). Kinetic analysis of C6/ST1 cells Hydrocortisone (Hy)-responsiveness showed elevated expression of the NRP/B gene upon a 5 h of treatment with this GC hormone (Degaki et al., 2009; Vedoy and Sogayar, 2002). NRP/B was first detected as one of the “p53-induced genes” or PIGs in a colorectal cancer cell line transfected with p53 (Polyak et al., 1997) and other p53 gene-targets were also induced by Hy in ST1 cells, such as trombospondin-1 and cyclin-G (Vedoy and Sogayar, 2002), suggesting growth arrest mediated by Hy via p53 activation. Overexpression of wt-NRP/B in the p53–/– colorectal carcinoma HCT 116 cells restored caspase-3 activation and cell sensitivity to apoptosis in response to cisplatin treatment and mutated NRP/B overexpression in HCT 116 p53+/+ cells attenuated caspase-3 activation (Liang et al., 2004), suggesting that NRP/B mutations affect the cellular apoptosis machinery through a p53-mediated caspase pathway, contributing to brain tumorigenesis. NRP/B is involved in the process of p53 stabilization by modulating Nrf-2-dependent NQO1 in tumors, because in response to oxidative stress stimuli, Nrf2 interacts with NRP/B, promoting the induction of NQO1 which interacts with p53, preventing p53 degradation (Seng et al., 2009).

11 Functional Role of the Novel NRP/B Tumor Suppressor Gene

In our study, the r-NRP/B cDNA was transfected into the C6/ST1 glioma cell line, and stable cell lines expressing the NRP/B protein were obtained, but no direct effect of NRP/B was observed in cell proliferation in monolayer cultures. However, C6/ST1-NRP/B cells displayed a significantly decreased ability to grow in soft-agar suspension, when compared with parental C6/ST1 cells or with cells overexpressing the EGFP protein or the empty pLPCX vector, suggesting a diminished tumorigenic capacity of these cells, which

A

113

was further supported by the tumor formation assays in nude mice. When subcutaneously injected into nude mice, the tumorigenicity potential of the cells obtained from the unique tumor formed from the NRP/B overexpressing cell line was only 25–30% of that displayed by both the C6/ST1 parental cells and by cells transformed with the empty retroviral pLPCX vector or with the vector carrying only the EGFP gene (Fig. 11.1b). Differently from normal cells, which require adherence to a substratum for growth (Stupack et al., 2001),

B

ST1 EGFP

ST1

Tumor volume (cm3)

3

ST1 ST1 EGFP ST1 pLPCX ST1 NRP/B

2

1

0 0

ST1 NRP/B

1

2

3 4 Weeks

5

6

7

ST1 pLPCX

C1

Fig. 11.1 Effect of retroviral-mediated transfer of Nrp/b in C6/ST1 cells in tumor growth in vivo (tumorigenesis). (a) Representative athymic mice showing tumor growth induced by C6/ST1 cells. Mice were individually injected subcutaneously with C6/ST1 cells (2×106 in 100 ml DMEM) and monitored for tumor growth during 6 weeks. Mice injected with C6/ST1, C6/ST1-EGFP and C6/ST1-pLPCX (empty vector) display a large tumor and the mouse injected with C6/ST1-NRP/B displays a very small tumor. (b) Graph showing the volume (cm3 ) of tumors induced by C6/ST1, C6/ST1-NRP/B, C6/ST1-EGFP and C6/ST1-pLPCX cells throughout the 6 weeks period. ∗ P < 0.05: comparing C6/ST1-NRP/B tumor volume with the others at the

C2

5th week. (c) Immunohistochemistry staining of the NRP/B protein in tumors originating from the C6/ST1 cells transduced with the empty pLPCX vector or with the pLPCX-NRP/B construct. (c1) C6/ST1-pLPCX cells. (c2) C6/ST1-NRP/B cells showing NRP/B positive-staining mainly localized in the cytoplasm. Original magnification: ×400. (Reprinted from J Steroid Biochem Mol Biol 117, Degaki TL, Demasi MAA, Sogayar MC, Overexpression of Nrp/b (nuclear restrict protein in brain) suppresses the malignant phenotype in the C6/ST1 glioma cell line., Pages 113 and 114, Copyright 2009, with permission from Elsevier)

114

transformed cells usually grow in agarose suspension (anchorage independence), an ability which strongly correlates with tumorigenesis in vivo (Cifone, 1982; Freedman and Shin, 1974). Anchorage dependence for growth is confined to the G1 phase of the cell cycle (Assoian, 1997). Once after cells pass the restriction point (Blagosklonny and Pardee, 2002), they no longer require adhesion to a substratum to complete the cycle (Bohmer et al., 1996). Restoration of anchorage dependence to the C6/ST1-NRP/B cell line may be due to the effect of NRP/B on limitation of cells passing beyond the G1 phase, an effect which is similar to that of C6/ST1 cells upon treatment with Hydrocortisone (Armelin et al., 1983). Kim and collaborators suggest that the growth suppressive effect of NRP/B may occur through interaction of its BTB domain with p110RB or, possibly, through other key cell cycle regulatory proteins, such as E2F or the CDKs (Kim et al., 2005). Previous results show that in vivo phosphorylation of NRP/B is regulated during cell cycle progression and that overexpression of NRP/B induces hypophosphorylation of p110RB during neuronal differentiation (Kim et al., 1998). Upon this induction of p110RB hypophosphorylation, associated with NRP/B overexpression, cell cycle reentering could not be promoted by E2F, which is released when p110RB is phosphorylated, an important transcription factor which induces the expression of genes whose products are required for cell cycle progression. It is widely known that abnormalities in molecular regulators of the G1 checkpoint (p16/pRb/E2F pathway) are commonly present in most of human gliomas (Louis et al., 2007), therefore, it is important to develop more effective therapies which function by specifically targeting these signaling pathways (e.g. the excess of E2F1) in malignant gliomas (Alonso et al., 2008). Re-location of NRP/B from the nucleus to the cytoplasm and gene mutations in NRP/B may affect interaction with p110RB , therefore, could result in deregulation of cell proliferation (Liang et al., 2004). Taken together, our results imply that r-NRP/B plays a tumor suppressive role and may partly reverse the malignant growth potential of glioma cells both in vitro and in vivo (Degaki et al., 2009). Our study underscores the functional importance of NRP/B over-expression in limiting the tumorigenic potential of C6/ST1 cells, but further studies are required to characterize the signaling mechanisms underlying this effect and the exact role played by this gene

T.L. Degaki et al.

during development of the Central Nervous System. We believe that model systems and organisms may be very useful, reflecting the genes role in human diseases, thus contributing to better understand the involvement of nuclear matrix proteins in cancer.

(Nuclear Restrict Protein in Brain) Tumor Suppressor Expression in Human Cell Lines and Clinical Samples of Astrocytic Tumors The relative NRP/B expression was analyzed by qRTPCR in several human cell lines available in our lab from brain (T98G, A172, U87, HOG, U343 and U373), uterus (C33A, Siha, Hela, Caski), lung (DU145, IMR90, NCIH23, H1155), prostate (NL, PC3, DU145), and breast (MCF-7, ZR751, MB231, MB435, HS5778T). Figure 11.2 shows the relative expression of NRP/B in a linear scale. The results, shown in Fig. 11.2a, indicate higher levels of the NRP/B transcript in some samples from brain and breast and lower expression in all samples from uterus, lung and prostate. The expression in brain and breast tumor cells is not homogeneous, a pattern which may due to the difference in metastatic or invasive potential. Using cDNAs obtained from samples of astrocytic tumors of different grades, we identified the expression pattern of the NRP/B transcript (Fig. 11.2b). Normal CNS tissue (11 samples) and grade I to IV astrocytomas (8 Pilocytic, 8 Diffuse, 12 Anaplastic and 19 glioblastomas) were analyzed for the expression of the NRP/B transcript, relative to the HPRT mRNA, in each of the samples, by qRT-PCR. The mean values of NRP/B transcript levels were higher in all grade tumors relative to normal samples, except for the pilocytic astrocytomas group; results which were statistically significant (P < 0.01). Cancer is characterized by uncontrolled division of cells and by a lack of features typical of the differentiated cell corresponding to the lineage in which they arise. Using cDNAs from human tumor cell lines as template, we found high NRP/B expression in brain and breast, and lower expression in uterus, prostate and lung. Expression of NRP/B was already detected in the following mammary cell lines: MCF-10A, MCF-7, and T-47D, but in MDA-MB-231, the translational levels

8

115 B

7 6 5 4 3 2 1 0 Brain

Uterus Prostate Lung

Breast

NRP/B relative expression

A

NRP/B relative expression

11 Functional Role of the Novel NRP/B Tumor Suppressor Gene 60 50 40 30 20 10 0 Normal Grade I Grade IIGrade III GBM

Clinical Samples Fig. 11.2 NRP/B mRNA expression in human tissues analysed by qRT-PCR. (a) Cells lines derived from brain, uterus, lung, prostate and breast. Relative expression obtained with respect to HPRT and MCF-7 cells expression levels. (b) Different grade astrocytomas (grade I, pilocytic astrocytoma; grade II, low grade

of NRP/B were abundant (Seng et al., 2007), similar results as those obtained in our study, which detected abundant relative expression in MDA-MB-231, followed by ZR751 and HS578T. In others glioblastomas cell lines, such as: SNB-19, SK-MG12, U-251MG U118-MG (Hernandez et al., 1998; Kim et al., 2000), the NRP/B expression was detected by Northern blot analysis. Mutations in the Kelch domain of NRP/B were detected in U87-MG cell line and mutations in both the Kelch and in the Intervening sequence (IVS) domains were found in the A172 cell line (Liang et al., 2004); whereas in our study we detect higher relative expression of NRP/B in U373MG, followed by A172 and U87MG. Using cDNAs obtained from samples of astrocytic tumors of different grades, quantitative real-time PCR analysis for the NRP/B gene revealed higher expression when compared to that of non-tumoral CNS tissue. For the first time, we describe this correlation of higher expression of NRP/B in higher grades astrocytoma tumors. The lower relative expression of NRP/B in pilocytic astrocytomas is similar to that of normal tissue samples, which could be explained due to the distinct benign characteristics of these tumors. Mutated forms of NRP/B were detect in brain tumor cell lines and in primary human malignant GBM tissues (Liang et al., 2004). These alterations within the NRP/B gene contribute to brain tumor tumorigenesis, triggering a higher expression of this mutated form. Acknowledgements We thank the excellent technical assistance of Zizi de Mendonça, Sandra Regina de Souza, Debora

astrocytoma; grade III, anaplastic astrocytomas; GBM, glioblastoma) and normal white (glial) nervous system tissue analyzed by qRT-PCR. Relative expression obtained with respect to HPRT gene expression levels in normal brain

Cristina da Costa, Ricardo Krett de Oliveira for the work carried out in our laboratories: Cellular and Molecular Biology Lab and Human Pancreatic Islet Unit. The financial support of FAPESP, CAPES, CNPq is greatly appreciated.

References Albagli O, Dhordain P, Deweindt C, Lecocq G, Leprince D (1995) The BTB/POZ domain: a new protein-protein interaction motif common to DNA- and actin-binding proteins. Cell Growth Differ 6:1193–1198 Almeida OF, Conde GL, Crochemore C, Demeneix BA, Fischer D, Hassan AH, Meyer M, Holsboer F, Michaelidis TM (2000) Subtle shifts in the ratio between pro- and antiapoptotic molecules after activation of corticosteroid receptors decide neuronal fate. FASEB J 14:779–790 Alonso MM, Alemany R, Fueyo J, Gomez-Manzano C (2008) E2F1 in gliomas: a paradigm of oncogene addiction. Cancer Lett 263:157–163 Armelin MC, Armelin HA (1983) Glucocorticoid hormone modulation of both cell surface and cytoskeleton related to growth control of rat glioma cells. J Cell Biol 97:459–465 Armelin MC, Oliveira ML, Mercado JM, Sasahara RM, Valentini SR, Carvalho LH (1996) Molecular genetic approach to cell proliferation control and neoplasia. Braz J Med Biol Res 29:911–919 Armelin MC, Stocco RC, Armelin HA (1983) Control of rat C6 glioma cell proliferation: uncoupling of the inhibitory effects of hydrocortisone hormone in suspension and monolayer cultures. J Cell Biol 97:455–458 Assoian RK (1997) Anchorage-dependent cell cycle progression. J Cell Biol 136:1–4 Barth RF, Kaur B (2009) Rat brain tumor models in experimental neuro-oncology: the C6, 9L, T9, RG2, F98, BT4C, RT-2 and CNS-1 gliomas. J Neurooncol 94: 299–312

116 Beato M, Truss M, Chavez S (1996) Control of transcription by steroid hormones. Ann N Y Acad Sci 784:93–123 Blagosklonny MV, Pardee AB (2002) The restriction point of the cell cycle. Cell Cycle 1:103–110 Bohmer RM, Scharf E, Assoian RK (1996) Cytoskeletal integrity is required throughout the mitogen stimulation phase of the cell cycle and mediates the anchorage-dependent expression of cyclin D1. Mol Biol Cell 7:101–111 Cifone MA (1982) In vitro growth characteristics associated with benign and metastatic variants of tumor cells. Cancer Metastasis Rev 1:335–347 Davie JR, Samuel S, Spencer V, Bajno L, Sun J, Chen HY, Holth LT (1998) Nuclear matrix: application to diagnosis of cancer and role in transcription and modulation of chromatin structure. Gene Ther Mol Biol 1:509–528 Degaki TL, Demasi MA, Sogayar MC (2009) Overexpression of Nrp/b (nuclear restrict protein in brain) suppresses the malignant phenotype in the C6/ST1 glioma cell line. J Steroid Biochem Mol Biol 117:107–116 Demasi MA, Montor WR, Ferreira GB, Pimenta DC, Labriola L, Sogayar MC (2007) Differential proteomic analysis of the anti-proliferative effect of glucocorticoid hormones in ST1 rat glioma cells. J Steroid Biochem Mol Biol 103:137–148 Freedman VH, Shin SI (1974) Cellular tumorigenicity in nude mice: correlation with cell growth in semi-solid medium. Cell 3:355–359 Fujita M, Furukawa Y, Tsunoda T, Tanaka T, Ogawa M, Nakamura Y (2001) Up-regulation of the ectodermal-neural cortex 1 (ENC1) gene, a downstream target of the betacatenin/T-cell factor complex, in colorectal carcinomas. Cancer Res 61:7722–7726 Hammarsund M, Lerner M, Zhu C, Merup M, Jansson M, Gahrton G, Kluin-Nelemans H, Einhorn S, Grander D, Sangfelt O, Corcoran M (2004) Disruption of a novel ectodermal neural cortex 1 antisense gene, ENC-1AS and identification of ENC-1 overexpression in hairy cell leukemia. Hum Mol Genet 13:2925–2936 He S, Dunn KL, Espino PS, Drobic B, Li L, Yu J, Sun JM, Chen HY, Pritchard S, Davie JR (2008) Chromatin organization and nuclear microenvironments in cancer cells. J Cell Biochem 104:2004–2015 Hernandez MC, Andres-Barquin PJ, Holt I, Israel MA (1998) Cloning of human ENC-1 and evaluation of its expression and regulation in nervous system tumors. Exp Cell Res 242:470–477 Hernandez MC, Andres-Barquin PJ, Martinez S, Bulfone A, Rubenstein JL, Israel MA (1997) ENC-1: a novel mammalian kelch-related gene specifically expressed in the nervous system encodes an actin-binding protein. J Neurosci 17:3038–3051 Huse JT, Holland EC (2010) Targeting brain cancer: advances in the molecular pathology of malignant glioma and medulloblastoma. Nat Rev Cancer 10:319–331 Kim SG, Jang SJ, Soh J, Lee K, Park JK, Chang WK, Park EW, Chun SY (2009) Expression of ectodermal neural cortex 1

T.L. Degaki et al. and its association with actin during the ovulatory process in the rat. Endocrinology 150:3800–3806 Kim TA, Jiang S, Seng S, Cha K, Avraham HK, Avraham S (2005) The BTB domain of the nuclear matrix protein NRP/B is required for neurite outgrowth. J Cell Sci 118:5537–5548 Kim TA, Lim J, Ota S, Raja S, Rogers R, Rivnay B, Avraham H, Avraham S (1998) NRP/B, a novel nuclear matrix protein, associates with p110(RB) and is involved in neuronal differentiation. J Cell Biol 141:553–566 Kim TA, Ota S, Jiang S, Pasztor LM, White RA, Avraham S (2000) Genomic organization, chromosomal localization and regulation of expression of the neuronal nuclear matrix protein NRP/B in human brain tumors. Gene 255:105–116 Liang XQ, Avraham HK, Jiang S, Avraham S (2004) Genetic alterations of the NRP/B gene are associated with human brain tumors. Oncogene 23:5890–5900 Louis DN (2006) Molecular pathology of malignant gliomas. Annu Rev Pathol 1:97–117 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A, Scheithauer BW, Kleihues P (2007) The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 114:97–109 Macedo AFA (2007) Functional link between glucocorticoid hormones and the TP53 tumor suppressor gene in a rat glioma cell model. PhD thesis, University of Sao Paulo, Sao Paulo Miura FK, Alves MJ, Rocha MC, Silva RS, Oba-Shinjo SM, Uno M, Colin C, Sogayar MC, Marie SK (2008) Experimental nodel of C6 brain tumors in athymic rats. Arq Neuropsiquiatr 66:238–241 Nahaczewski AE, Fowler SB, Hariharan S (2004) Dexamethasone therapy in patients with brain tumors a focus on tapering. J Neurosci Nurs 36:340–343 Piette C, Munaut C, Foidart JM, Deprez M (2006) Treating gliomas with glucocorticoids: from bedside to bench. Acta Neuropathol 112:651–664 Polyak K, Xia Y, Zweier JL, Kinzler KW, Vogelstein B (1997) A model for p53-induced apoptosis. Nature 389:300–305 Seng S, Avraham HK, Birrane G, Jiang S, Li H, Katz G, Bass CE, Zagozdzon R, Avraham S (2009) NRP/B mutations impair Nrf2-dependent NQO1 induction in human primary brain tumors. Oncogene 28:378–389 Seng S, Avraham HK, Jiang S, Yang S, Sekine M, Kimelman N, Li H, Avraham S (2007) The nuclear matrix protein, NRP/B, enhances Nrf2-mediated oxidative stress responses in breast cancer cells. Cancer Res 67:8596–8604 Stupack DG, Puente XS, Boutsaboualoy S, Storgard CM, Cheresh DA (2001) Apoptosis of adherent cells by recruitment of caspase-8 to unligated integrins. J Cell Biol 155:459–470 Vedoy CG, Sogayar MC (2002) Isolation and characterization of genes associated with the anti-tumor activity of glucocorticoids. Brain Res Mol Brain Res 106:57–69

Chapter 12

Brain Tumors: Diagnostic Impact of PET Using Radiolabelled Amino Acids Karl-Josef Langen, Matthias Weckesser, and Frank Floeth

Abstract Magnetic resonance tomography (MRT) allows an exact morphologic visualisation of the brain and is the investigation of choice for diagnosing cerebral glioma, but its capacity to differentiate tumor tissue from non-specific tissue changes is limited. Positron emission tomography (PET) using radiolabeled amino acids provides additional metabolic information which helps increase diagnostic accuracy. The use of radiolabeled amino acids allows better delineation of tumor margins and improves targeting of biopsy and radiotherapy, and planning surgery. In addition, amino acid imaging appears useful in distinguishing tumor recurrence from nonspecific post-therapeutic edema and scar tissue, in predicting prognosis in low grade gliomas, and in monitoring metabolic response during treatment. Most PET studies of cerebral gliomas have been performed with the amino acid 11 C-methyl-L-methionine (MET), although the short half-life of 11 C (20 min) limits the use of this technique to the few centers that are equipped with an in-house cyclotron facility. In recent years 18 F-labelled amino acids (half-life 109 min) such as O-(2-[18 F]fluoroethyl)-L-tyrosine (FET) have been developed that allow a more widespread use of amino acid imaging. The logistical prerequisites for amino acid imaging of brain tumors have become markedly less difficult with the introduction of 18 F-labelled amino acids like FET. The scientifically documented utility of amino acid imaging of cerebral gliomas

K.-J. Langen () Institute of Neuroscience and Medicine, Forschungszentrum Jülich, D-52425 Jülich, Germany e-mail: [email protected]

justifies its use as a routine diagnostic technique for certain indications. Keywords Tumor · PET · Amino acids · MRT · Gliomas · MET · FET

Introduction Malignant glial neoplasms of the nervous system arise with an incidence of 5–6 new cases per 100.000 persons per year (Ohgaki and Kleihues, 2005) and these diffuse gliomas are the most common type of primary brain tumors. The treatment of cerebral glioma consists of biopsy or surgical resection and depending on grade of malignancy local radiotherapy and chemotherapy. Despite all efforts, the results of treatment, measured in terms of survival time and quality of life, remain unsatisfactory over the decades to this day with a median survival time of 5.6 years for low-grade glioma WHO grade II, 1.6 years for anaplastic glioma WHO grade III and 0.4 years for glioblastoma WHO grade IV (Ohgaki and Kleihues, 2005). Magnetic resonance imaging (MRI) with its excellent soft tissue contrast, the high spatial resolution of 1–2 mm and its multiplanar reconstruction capabilities is currently the method of first choice for the diagnosis and differential diagnosis of primary brain tumors. With MRI, the tumor can be reliably detected and, usually, well characterized in terms of its internal structure and the extent to which it disrupts the blood-brain barrier. Imaging with standard T1-weighted sequences serves to demonstrate the extent and the anatomical relationship of the tumor to the adjacent normal structures; when contrast medium is given, the tumor

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_12, © Springer Science+Business Media B.V. 2011

117

118

K.-J. Langen et al.

per se can be better delineated from the accompanying reaction in the surrounding brain tissue (edema) (Ernst, 1997). The differentiation of glioma tissue with its pronounced regional heterogeneity from surrounding edema is unreliable, however, particularly when the tumor is not sharply demarcated from normal brain tissue due to diffuse tumor cell infiltration, and when the blood-brain barrier remains intact (Jansen et al., 2000). Radiogenic changes in peritumoral brain tissue after treatment may result in pathological uptake of contrast medium that cannot be reliably distinguished from vital tumor tissue of recurrent glioma (Lustig et al., 2007). T2-weighted sequences, supplemented with a proton-density or FLAIR (fluid-attenuated inversion recovery) sequence, can reveal the maximal extent of structural changes, but, like T1-weighted images, they do not permit reliable differentiation of tumor tissue from the surrounding edema (Ernst, 1997). Metabolic studies, such as the measurement of glucose metabolism with 18 F-fluorodeoxyglucose (FDG) and PET, have been used for many years as supplementary

methods beyond morphological imaging. In cerebral gliomas, FDG uptake is correlated with the degree of malignancy of the tumor (WHO grading) and with the patient’s outcome (Padma et al., 2003; Chen et al., 2007). Because of the high rate of glucose metabolism in normal brain tissue, however, it is often difficult to distinguish glioma tissue from normal brain tissue by FDG-PET. While most high-grade glioma WHO grade III and nearly all grade IV glioblastoma show a high glucose uptake, low-grade glioma WHO grade II usually exhibit a normal or even a decreased glucose uptake (Fig. 12.1). Aside from FDG, radiolabeled amino acids have also been used successfully for many years in PET to measure tumor metabolism (Jager et al., 2001). Because the uptake of amino acids by normal brain tissue is relatively low, cerebral gliomas can be distinguished from the surrounding normal tissue with high contrast. Most PET studies of cerebral gliomas are performed with the amino acid MET (Singhal et al., 2008), although the short half-life of 11 C (20 min) limits the use of this technique to the few

Fig. 12.1 Astrocytoma WHO Grade II: T1-weighted MRI after application of Gd-DTPA and T2 weighted MRI (upper row) show widespread abnormalities. FDG PET shows

hypometabolism in the tumor area and is not helpful to guide biopsy (lower left). FET PET identifies a hot spot within the tumour and detects an optimal biopsy site (lower right)

12 Diagnostic Impact of PET Using Radiolabelled Amino Acids

centers that are equipped with an in-house cyclotron facility. In recent years some 18 F-labelled amino acids (half-life 109 min) have been introduced which, like FDG, can be transported from a cyclotron to multiple external PET centers. This enables a wider application of amino acid PET in clinical diagnosis. One of the best established tracers is FET that can be produced in large amounts for clinical purposes like the widely used FDG (Langen et al., 2006). The purpose of this chapter is to provide an overview of the current state of development of amino acid diagnosis for cerebral glioma. The report is focussed on the clinical experiences with MET and FET which are at present the best validated amino acid tracers for PET. The clinical results obtained with these tracers, however, are in principle similar to those obtained with other large neutral amino acid tracers like 11 C-tyrosine or 18 F-DOPA.

The General Properties of Radiolabeled Amino Acids The increased uptake of MET and FET by cerebral glioma tissue is due almost entirely to increased transport via specific amino acid transporters (Langen et al., 2006; Singhal et al., 2008). Since large neutral amino acids also enter normal brain tissue, a disruption of blood brain barrier, i.e., enhancement of contrast media in CT or MRI scans, is not a prerequisite for intratumoral (BBB) accumulation of MET or FET. Consequently, uptake of the tracer has been reported in many low grade gliomas without BBB leakage (Singhal et al., 2008; Floeth et al., 2007). Comparative studies of these substances have shown that both are taken up to roughly the same extent by cerebral glioma tissue (Weber et al., 2000; Langen et al., 2006), so that the clinical experience gained with them can be considered together. Animal experiments have shown that FET, in contrast to MET, exhibits no uptake in inflammatory cells and in inflammatory lymph nodes, promising a higher specificity for the detection of tumor cells (Langen et al., 2006). Nevertheless, false positive uptake has been observed for both tracers MET and FET in brain abscesses, demyelinating processes, near cerebral ischemia and hematomas. Therefore, increased uptake of the tracers is not specific for cerebral gliomas although high

119

amino acid uptake has a high positive predictive value for cerebral gliomas. For both tracers, the radiation exposure to the patient remains within the same order of magnitude as that of conventional radiological studies (Langen et al., 2008). No side effects have been reported to date with the use of these tracers after several thousand studies have been performed worldwide. The duration of tracer uptake and image acquisition with any of these radiopharmaceuticals is about 30–50 min. The cost of FET-PET is comparable to that of FDG-PET while MET-PET is substantially more expensive since no more than 2–3 patients can be studied with a single radiosynthesis.

Clincial Applications of Amino Acid PET in Brain Tumors Showing the Extent of Tumor for Biopsy and Treatment Planning An important aspect of the diagnostic assessment of cerebral glioma is the measurement of the extent of the tumor and the detection of the areas within the tumor in which the rate of cellular proliferation is highest. Because the tumor biology is dominated by the most malignant parts of the heterogeneous glioma, representative tissue samples are vitally important for histological tumor diagnosis, prognostication, and treatment planning. The ability of MRI to show the most rapidly proliferating portions of the usually inhomogeneous glioma is, unfortunately, limited, particularly when the tumor does not take up contrast medium. Multiple studies in which the radiological findings were compared with the histological findings in tissue samples obtained by biopsy or open surgery have provided clear evidence that MET-PET detects rapidly proliferating glioma tissue more reliably than either CT or MRI (Singhal et al., 2008). MET-PET can also be used to detect the regions with the most pronounced anaplastic changes inside a heterogeneous glioma (Goldmann et al., 1997; Pirotte et al., 2004). Similar results have been reported for FET-PET. In a study involving 31 patients, tumor tissue was found in 94% of biopsies taken at sites where FET-PET had pointed to probable tumor tissue, but only in 53% of the suspicious areas identified by MRI (Pauleit et al., 2005). These figures impressively document the decidedly better

120

diagnostic reliability of neuroimaging with radiolabeled amino acids. Amino acid imaging thus optimizes the targeting biopsies and helps prevent the problem of non-diagnostic biopsies from non-specifically altered tissue (Fig. 12.1). Amino acid imaging can, therefore, be expected to improve patient care, particularly in cases of glioma without any disruption of the blood-brain barrier (i.e., without contrast enhancement on MRI), and when the tumor lies adjacent to eloquent structures. Furthermore, amino acid imaging also improves the planning of treatment for glioma, both resective surgery and local radiotherapy. A restriction of the target volume to the actual tumor tissue can substantially lessen the adverse effects of radiation therapy, while permitting treatment of the tumor itself with a higher dose than otherwise possible (Fig. 12.2). Grosu et al. (2005) have reported that, in 29% of cerebral gliomas, pathological amino acid uptake is found outside the abnormal brain tissue revealed by MRI while 90% of the hyperdense areas identified by T2-weighted MRI showed no pathological amino acid uptake. The authors recommend integrating amino acid imaging into CT- and MRI-based radiotherapy planning, particularly when high-precision

Fig. 12.2 Anaplastic glioma, WHO grade III: this large tumor shows only a mass effect in the upper hemisphere but the solid tumor mass cannot be identified neither in the T1-weighted image before and after application of Gd-DTPA (upper row) nor in the T2-weighted MR image (lower left). FET-PET (lower right) clearly depicts the area of the solid tumor (red-yellow area)

K.-J. Langen et al.

radiotherapy is to be given or in the setting of dose escalation studies or for the re-irradiation of recurrent tumors. A preliminary trial of amino acid imaging for radiotherapy planning in recurrent glioma has shown a significantly longer survival time than when the planning is based on CT and MRI alone (median, 9 months versus 5 months) (Grosu et al., 2005).

The Differential Diagnosis of Intracranial Tumors The differential diagnosis of a space occupying or a diffuse lesion in the brain includes primary and metastatic brain tumors, hemorrhage, infarction, infections like abscess, virus encephalitis, proliferative multifocal leukencephalopathy, and inflammatory pathologies like multiple sclerosis and post infectious encephalomyelitis. The overall sensitivity of METPET for gliomas, including both high and low grade gliomas, has been estimated to be around 76 and 95% (Singhal et al., 2008). Herholz et al. (1998) determined a specificity of 87% for distinguishing non-tumoral brain lesions using a tumor/normal tissue ratio of 1.47

12 Diagnostic Impact of PET Using Radiolabelled Amino Acids

as the diagnostic cut-off in a sample of 28 nontumoral cases. There have been reports of perifocal MET uptake around hematomas and areas of ischemia, as well as of rare cases of MET and FET uptake in or around ring enhancing lesion like brain abscesses and acute inflammatory demyelination (Singhal et al., 2008; Floeth et al., 2006). The occasional uptake of radiolabeled amino acids by non-neoplastic lesions reduces the specificity of this method to roughly 90%. Although the uptake of radioactive amino acids is considered relatively specific for neoplastic cells, the predictive value is limited and the possibility of nonspecific enhancement in inflammatory cells or reactive glial tissue must still be borne in mind. Therefore, a histological evaluation of tumor tissue by biopsy remains necessary especially in contrast enhancing lesions. A recent study by Floeth et al. (2005) indicated the utility of combining FET-PET with proton magnetic resonance spectroscopy (1 H-MRS) in the diagnostic evaluation of 50 newly diagnosed brain lesions with suspicion of glioma. In a biopsy-controlled study using a tumor/normal tissue ratio of 1.6 as the diagnostic cutoff, a neoplastic lesion was found by biopsy in 97% of cases where both FET-PET and MRS had indicated the presence of a tumor. On the other hand, a neoplastic lesion was not found in any case in which both techniques had yielded a negative finding. It follows that, when a patient has a brain lesion of unknown type, the combination of two methods of metabolic imaging with amino acid PET and 1 H-MRS can be used to detect or rule out a tumor with a very high degree of diagnostic accuracy. If the findings of both FET-PET and 1 H-MRS are normal, it would seem reasonable to dispense with a biopsy and follow the lesion further with MRI at regular intervals. Very similar findings of successful differentiation of tumorous and non-tumorous tissue were achieved in an actual study of patients with small nonspecific incidental brain lesions with a combination of the standard morphologic imaging by MRI and metabolic FET-PET imaging (Floeth et al., 2008). Incidental findings are defined as previously undetected abnormalities that are unexpectedly discovered and unrelated to the purpose of the examination and their management is a critical ethical problem. Due to an increasing number of cerebral MRI investigations in patients with relatively nonspecific symptoms, such as headaches or dizziness, in asymptomatic patients as a “brain check-up” MRI,

121

or in healthy volunteers as research scans, a rapidly increasing number of small white matter incidentalomas of the brain with the differential diagnosis of a glioma in statu nascendi is visualized. In a prospective study 21 aymptomatic patients with small nonenhancing white matter lesions suspective of a diffuse low grade glioma were screened with MRI and FET-PET using a tumor/normal tissue ratio of 1.6 as the diagnostic cut-off at the time of first diagnosis. No biopsy was obtained and only follow-up with MRI and clinical investigation was done in 4- to 6-month intervals. All 17 lesions with a circumscribed growth pattern on MRI and normal or low FET uptake on PET had a benign course: The patients remained clinically asymptomatic and most lesions vanished completely, regressed or persisted with eventual development of a low-grade glioma WHO grade II in 2 patients after several years. In contrast, all 4 lesions with a diffuse growth pattern on MRI and increased FET uptake underwent rapid clinical deterioration and died due to the development of a high grade glioma WHO grade III or IV usually within one year. These data suggest that the assessment with the combination of MRI and FETPET provides a good prediction of course and outcome in unspecific white matter lesions with the differential diagnosis of a glioma. While a circumscribed growth pattern on MRI and a normal or low FETmetabolism is a strong predictor of a benign course, an early biopsy for histologic assessment does not appear to be mandatory and follow-up in regular 6 months intervals is sufficient. Diffuse lesions on MRI with increased FET uptake on PET are highly at risk of a malignant course with fatal outcome and early biopsy, short follow-up intervals and early aggressive therapy seems to be necessary for these devastating lesions.

Glioma Grading and Prognosis While FDG-PET is considered as a relative accurate predictor of the World Health Organization (WHO) grading and prognosis of cerebral gliomas (Padma et al., 2003), most studies employing amino acid imaging have shown that gliomas of different WHO grades overlap in their degree of amino acid uptake, so that the tumor grade cannot be reliably predicted with this technique (Singhal et al., 2008; Pauleit et al., 2005).

122

More recently, differences in the time course of FET uptake depending on the WHO grade have been found (Weckesser et al., 2005). High-grade gliomas (HGG) of grade III (anaplastic glioma) or IV (glioblastoma) are characterized by an early peak around 10–15 min after injection followed by a decrease of FET uptake. In contrast, time–activity curves slightly and steadily increase in low-grade gliomas (LGG) of WHO grade II. Using dynamic evaluation of selected regions of the tumor, HGG and LGG could be distinguished with an accuracy >90% in primary tumors as well as in recurrent tumors (Pöpperl et al., 2006a, b, 2007). Thus, the sensitivity and specificity of dynamic FET-PET for tumor grading appears to be comparable to that achieved by FDG-PET. The prognostic significance of the findings of amino acid imaging is currently debated. Some studies seem to show that lower amino acid uptake especially in astrocytic glioma is associated with a better prognosis but there may be high uptake in oligodendrogliomas in spite of their apparently better prognosis (Kaschten et al., 1998; Singhal et al., 2008). In low-grade glioma the majority of tumors undergoes astrocytic differenciation with some oligodendrocytic and some mixed oligoastrocytic tumors. There is a general observation that astrocytoma WHO grade II show a higher FET uptake than the surrounding brain in two third of the tumors while one third exhibits a normal or lower FET-metabolism. In contrast all oligodendroglioma and mixed glioma WHO grade II do have increased FET uptake and this is usually a high to very high FET-metabolism (Floeth et al., 2006). Obviously the oligodendrocytic glioma are a different entity with different metabolic features. A prognostic aspect of amino acid PET concerns also its ability to reflect the tumor extent for planning surgery of gliomas. Recently, Pirotte et al. (2009) demonstrated that complete resection of the tumor area with increased amino acid uptake prolongs the survival of patients with high grade gliomas. Conversely, whether or not MRI enhancement was present on the postoperative scan did not have an impact on the survival of those patients. These data indicate that resection of malignant gliomas guided by amino acid PET may increase the amount of anaplastic tissue removed and thus patient’s survival because amino acid PET represents a more specific marker of anaplastic tumor tissue than contrast enhancement in MRI.

K.-J. Langen et al.

Furthermore, there appears to be an important clinical role of amino acid imaging in prognostication for patients with low-grade gliomas (LGG) that account for ∼15% of the total population of glioma patients. Some of these patients will enjoy a stable course with an excellent quality of life for decades even without treatment, while others experience rapid tumor progression with malignant transformation to a high-grade glioma (HGG) and a poor prognosis. A small number of prognostic factors have been identified, but the individual course remains unpredictable, and the optimal treatment strategy is controversial. A better identification of individuals with either a poor or a favourable prognosis is highly desirable to optimize treatment. Delayed or insufficient treatment of progressive lesions may shorten survival time, whereas too early aggressive treatment may cause unnecessary hospitaliszation and treatment-related morbidity. A study with METPET showed that these patients benefit from a surgical procedure only when increased amino acid uptake can be demonstrated (Ribom et al., 2001). FET-PET combined with MR morphology has also been found to be a statistically significant prognostic predictor for patients with newly diagnosed low-grade gliomas (LGG). Within a 7 year period a group of 33 consecutive patients with previously untreated non enhancing grade II glioma were included in a prospective study (Floeth et al., 2006). A baseline MRI and FET-PET was performed before histology was established in all patients on tissue samples by biopsy and a “wait and see” strategy without further treatment was started. During the follow-up it turned out, that baseline FET uptake and a circumscribed versus a diffuse growth pattern on MRI were highly significant predictors for patients’ course and outcome: Those LGG that were well demarcated in the MRI and did not take up FET had an excellent prognosis with long progression free intervals, good clinical condition and without malignant transformation. In contrast the LGG with diffuse tumor margins and FET uptake had a poor outcome with early progression and malignant transformation to a HGG, rapid clinical deterioration and death within a few years. Thus, combined assessment with FET-PET and MRI can identify a subgroup of patients with “stable” LGG who may be best “treated” with observation alone and another subgroup of patients with “instable” LGG who should receive early and aggressive treatment before they transform into a HGG.

12 Diagnostic Impact of PET Using Radiolabelled Amino Acids

The Diagnostic Assessment of Recurrent Tumors It is difficult to distinguish recurrent glioma from nonspecific post-therapeutic changes with conventional MRI alone, because pathological enhancement with contrast medium may reflect either new growth of tumor or tissue necrosis after radio- or chemotherapy (Brandsma et al., 2008). The role of FDG-PET in such cases is limited because of the frequency of nonspecific uptake (Langen et al., 2008). Multiple studies have shown that MET-PET is highly sensitive to detect recurrent tumor but the specificity for the differentiation of vital tumor tissue from non-neoplastic changes is not optimal and in the range of 70% (Singhal et al., 2008). The specificity of FET-PET for the differentiation of recurrent tumor from non-neoplastic changes appears to be higher than of MET-PET. In a study involving 46 patients, the sensitivity and specificity of FET-PET for the detection of recurrent gliomas were 100 and 93%, respectively, compared with 93 and 50% for MRI (Rachinger et al., 2005). Additional use of dynamic FET-PET allowed a differentiation of highgrade and low-grade recurrences with a sensitivity and specificity of 92% (Pöpperl et al., 2006a, b). Thus, especially FET-PET is considered as a valuable tool in differentiating recurrent tumor from non-neoplastic changes.

Fig. 12.3 Glioblastoma (WHO grade IV) in the left temporal lobe. Brain imaging after surgery (upper row) and after completion of radiochemotherapy (RCX). There is an enlargement of the contrast-enhancing area in the T1-weighted image after RCX suggesting tumor progression while FET-PET indicates good response due to decreasing amino acid uptake. The patient had a favourable outcome

123

Treatment Monitoring Changes in apparent tumor size or contrast enhancement that are seen in MRI and CT are taken as indicators of the response to therapy but this approach is limited by the difficulty in distinguishing vital tumor tissue and unspecific treatment effects. Early detection of treatment response in MRI may be misleading because of the known reactive transient blood-brain barrier (BBB) alterations with consecutive contrast enhancement mimicking tumor progression. This phenomenon, so called “pseudoprogression” is seen in 20–47% of cases and can lead to an unnecessary overtreatment (Lustig et al., 2007; Piroth et al., 2010). The feasibility and usefulness of MET-PET for therapy assessment and follow-up in conjunction with anatomical imaging after surgery, chemotherapy, and radiotherapy have been demonstrated in several studies. The currently available data suggest that a reduction of MET uptake by a glioma is a sign of a response to treatment (Singhal et al., 2008). Initial longitudinal studies employing FET-PET after locoregional chemo- and radioimmunotherapy of gliomas have also shown a good correlation between FET uptake and treatment response (Pöpperl et al., 2006). Recently, a first prospective study evaluated the prognostic value of early changes of FET uptake after postoperative

124

radiochemotherapy in glioblastomas (RCX) (Piroth et al., 2010) (Fig. 12.3). It could be demonstrated that PET responders with a decrease of the tumor/brain ratio of more than 10% had a significantly longer disease free survival and overall survival than patients with stable or increasing tracer uptake after RCX. Thus, PET using amino acids like FET appears at present one of the most promising imaging methods to predict treatment response of malignant gliomas at an early stage after RCX. The identification of non-responders may help to optimize individual treatment strategy using alternative methods and to reduce the negative impact of ineffective therapy approaches on the patient’s quality of life. Furthermore, individually optimized treatment may also have economical impact since the costs of PET imaging are relatively small in relation to the expenses of local or systemic treatment approaches and consecutively the management of adverse effects. The information provided by amino acid PET may assist to optimize the individual treatment strategy, to define the individual need for dose escalation in radiotherapy and to minimize negative impact of treatment approaches on quality of life.

Imaging Brain Tumors in Children The histological subtypes of brain tumours in children differ considerably from that in adults. Only few mainly retrospective studies have been performed in children with brain tumours. It is however evident that the assessment of glucose metabolism with FDG is less suitable for the evaluation of tumour malignancy than it is the case in adults (Weckesser et al., 2001). The main reason for this is the high glucose metabolism in pilocytic astrocytomas. These low grade tumours may exhibit metabolic rates with the intensity of grey matter, an association of the metabolic activity of the tumour and clinical presentation or outcome is not evident. In children, the determination of tumour grade with amino acids seems to be even less reliable than in adults. A broad overlap of amino acid uptake is observed in low grade and high grade tumors (Utriainen et al., 2002). Similar to glucose metabolism amino acid uptake may be high in low grade tumors like pilocytic astrocytomas and gangliogliomas, uptake may be relatively low in the highly

K.-J. Langen et al.

aggressive medulloblastomas (WHO IV), a common diagnosis in infratentorial brain tumors. The dynamic acquisition protocol in FET has not yet been systematically evaluated in children with brain tumours. The potential of amino acids to determine the site of stereotactic biopsy or for imageguided surgical resection of infiltrative low-grade brain tumours in children has been reported (Pirotte et al., 2007). The presence of amino acid uptake after surgery is indicative of residual tumor in case of ambiguous findings in early postoperative MRI (Pirotte et al., 2005).

Perspectives for Amino Acid PET Diagnostic assessment in nuclear medicine by imaging with radiolabeled amino acids permits a more specific representation of the spatial extent of solid and diffuse glioma tissue than is possible by conventional MRI alone. This is very advantageous for the planning of biopsies, resections, and radiotherapy. Furthermore, recurrent tumors can be differentiated from posttherapeutic changes with a high degree of specificity, valuable prognostic information can be obtained for low-grade gliomas, and the treatment response can probably be judged early on in the course of treatment. Unfortunately, only a small number of studies of amino acid imaging for brain tumors in children have been performed to date, but here, too, the technique seems to be comparably useful (Singhal et al., 2008). Other imaging methods like 1 H-MRS may also yield metabolic information that is markedly more specific than that obtainable by conventional MRI for the differentiation of tumor tissue from non-specific changes (Herholz et al., 2007). Unlike PET, however, MRS can only be used to analyze selected small volumes or partial areas in single planes and the quality of the studies is often impaired by susceptibility artefacts. Other techniques like perfusion-weighted MRI (pwMRI or PWI) are more easily available than PET and may yield information that is correlated with the degree of malignancy of gliomas (Herholz et al., 2007). The diagnostic accuracy of this technique in comparison with amino acid PET, however, remains to be investigated. The scientifically documented utility of amino acid imaging of cerebral gliomas seems to justify its use as a routine diagnostic technique for certain indications,

12 Diagnostic Impact of PET Using Radiolabelled Amino Acids

but it remains to be confirmed that this will improve the overall quality of care. The logistical prerequisites for amino acid imaging have become markedly less difficult to achieve in recent years with the introduction of FET-PET. The costs of these diagnostic techniques would appear to be well justified by their clinical utility, not least because their timely application in a larger number of patients can be expected to save the costs incurred today by the use of other, less diagnostically reliable techniques.

References Brandsma D, Stalpers L, Taal W, Sminia. P, van den Bent MJ (2008) Clinical features, mechanisms, and management of pseudoprogression in malignant gliomas. Lancet Oncol 9:453–461 Chen W (2007) Clinical applications of PET in brain tumors. J Nucl Med 48:1468–1481 Ernst S (1997) Hirntumoren. In Heindel W, Kugel H, Lackner K (eds) Rationelle MR-Untersuchungstechniken, pp 3–7. Thieme Verlag, Stuttgart Floeth FW, Pauleit D, Langen KJ, Reifenberger G, Stoffels G, Stummer W, Rommel F, Hamacher K, Sabel M (2006) Differentiation of tumorous and non-tumorous ring enhancing lesions (REL) with FET PET. J Nucl Med 47:776–782 Floeth FW, Pauleit D, Sabel M, Stoffels G, Reifenberger G, Riemenschneider MJ, Jansen P, Hamacher K, Coenen HH, Stummer W, Steiger HJ, Langen KJ (2007) Prediction of prognosis in low grade glioma patients with magnetic resonance imaging and positron emission tomography using F-18-fluoroethyltyrosine. J Nucl Med 48:519–527 Floeth F, Pauleit D, Wittsack HJ, Langen KJ, Reifenberger G, Hamacher K, Müller HW, Zilles K, Messing-Jünger M, Weber F, Stummer W, Steiger HJ, Coenen HH, Sabel M (2005) Multimodal metabolic imaging of cerebral gliomas using positron emission tomography with [18 F]-fluoroethyll-tyrosine and magnetic resonance spectroscopy. J Neurosurg 102:318–327 Floeth FW, Sabel M, Stoffels G, Pauleit D, Hamacher K, Steiger HJ, Langen KJ (2008) Prognostic value of 18F-fluoroethylL-tyrosine PET and MRI in small nonspecific incidental brain lesions. J Nucl Med 49:730–737 Goldman S, Levivier M, Pirotte B, Brucher JM, Wikler D, Damhaut P, Dethy S, Brotchi J, Hildebrand J (1997) Regional methionine and glucose uptake in high-grade gliomas: A comparative study on PET-guided stereotactic biopsy. J Nucl Med 38:1459–1462 Grosu AL, Weber WA, Franz M„ Stark S, Piert M, Thamm R, Gumprecht H, Schwaiger M, Molls M, Nieder C (2005) Reirradiation of recurrent high-grade gliomas using amino acid PET(SPECT)/CT/MRI image fusion to determine gross tumor volume for stereotactic fractionated radiotherapy. Int J Radiat Oncol Biol Phys 63:511–519 Herholz K, Coope D, Jackson A (2007) Metabolic and molecular imaging in neuro-oncology. Lancet Neurol 6:711–724

125 Herholz K, Hölzer T, Bauer B, Schröder R, Voges J, Ernestus RI, Mendoza G, Weber-Luxemburger G, Löttgen J, Thiel A, Wienhard K, Heiss WD (1998) 11 C-methionine PET for differential diagnosis of low-grade-gliomas. Neurology 50:1316–1322 Jager PL, Vaalburg W, Pruim J, de Vries EG, Langen KJ, Piers DA (2001) Radiolabeled amino acids: basic aspects and clinical applications in oncology. J Nucl Med 42:432–445 Jansen EP, Dewit LG, van Herk M, Bartelink H (2000) Target volumes in radiotherapy for high-grade malignant gliomas of the brain. Radiother Oncol 56:151–156 Kaschten B, Stevenaert A, Sadzot B, Deprez M, Degueldre C, Del Fiore G, Luxen A, Reznik M (1998) Preoperative evaluation of 54 gliomas by PET with fluorine-18fluorodeoxyglucose and/or carbon-11-methionine. J Nucl Med 39:778–785 Langen KJ, Hamacher K, Weckesser M, Floeth F, Stoffels G, Bauer D, Coenen HH, Pauleit D (2006) O-(2[18 F]fluoroethyl)-L-tyrosine: uptake mechanisms and clinical applications. Nucl Med Biol 33:287–294 Langen KJ, Tatsch K, Grosu AL, Jacobs AH, Weckesser M, Sabri O (2008) Diagnostics of cerebral gliomas with radiolabeled amino acids. Dtsch Arztebl Int 105:55–61 Lustig RA, Seiferheld W, Berkey B, Yung AW, Scarantino C, Movsas B, Jones CU, Simpson JR, Fishbach J, Curran WJ Jr (2007) Imaging response in malignant glioma, RTOG 90-06. Am J Clin Oncol 30:32–37 Ohgaki H, Kleihues P (2005) Population-based studies on incidence, survival rates, and genetic alterations in astrocytic and oligodendroglial gliomas. J Neuropathol Exp Neurol 64:479–489 Padma MV, Said S, Jacobs M, Hwang DR, Dunigan K, Satter M, Christian B, Ruppert J, Bernstein T, Kraus G, Mantil JC (2003) Prediction of pathology and survival by FDG PET in gliomas. J Neurooncol 64:227–237 Pauleit D, Floeth F, Hamacher K, Riemenschneider MJ, Reifenberger G, Müller HW, Zilles K, Coenen HH, Langen KJ (2005) O-(2-[18 F]fluoroethyl)-L-tyrosine PET combined with Magnetic Resonance Imaging Improves the Diagnostic Assessment of Cerebral Gliomas. Brain 128:678–687 Piroth MD, Pinkawa M, Holy R, Klotz J, Nussen S, Stoffels G, Coenen HH, Kaiser JH, Langen KJ, Eble MJ (2010) Prognostic Value of Early 18 F-Fluoroethyltyrosine PET after Radiochemotherapy in Glioblastoma multiforme. Int J Radiat Oncol Biol Phys 2010, Epub 2010 Jun 21, DOI: 10.1016/j.ijrobp.2010.01.055 Pirotte B, Goldman S, Massager N, David P, Wikler D, Vandesteene A, Salmon I, Brotchi J, Levivier M (2004) Comparison of 18 F-FDG and 11 C-methionine for PETguided stereotactic brain biopsy of gliomas. J Nucl Med 45:1293–1298 Pirotte BJ, Levivier M, Goldman S, Massager N, Wikler D, Dewitte O, Bruneau M, Rorive S, David P, Brotchi J (2009) Positron emission tomography-guided volumetric resection of supratentorial high-grade gliomas: a survival analysis in 66 consecutive patients. Neurosurgery 64:471–481 Pirotte B, Levivier M, Morelli D, Van Bogaert P, Detemmerman D, David P, Baleriaux D, Brotchi J, Goldman S, (2005) Positron emission tomography for the early postsurgical evaluation of pediatric brain tumors. Childs Nerv Syst 21: 294–300

126 Pirotte BJ, Lubansu A, Massager N, Wikler D, Goldman S, Levivier M (2007) Results of positron emission tomography guidance and reassessment of the utility of and indications for stereotactic biopsy in children with infiltrative brainstem tumors. J Neurosurg 107(Suppl):392–399 Pöpperl G, Gotz C, Rachinger W, Schnell O, Gildehaus FJ, Tonn JC, Tatsch K (2006a) Serial O-(2-[(18)F]fluoroethyl)L-tyrosine PET for monitoring the effects of intracavitary radioimmunotherapy in patients with malignant glioma. Eur J Nucl Med Mol Imaging 33:792–800 Pöpperl G, Kreth FW, Herms J, Koch W, Mehrkens JH, Gildehaus FJ, Kretzschmar HA, Tonn JC, Tatsch K (2006b) Analysis of 18 F-FET PET for grading of recurrent gliomas: is evaluation of uptake kinetics superior to standard methods? J Nucl Med 47:393–403 Pöpperl G, Kreth FW, Mehrkens JH, Herms J, Seelos K, Koch W, Gildehaus FJ, Kretzschmar HA, Tonn JC, Tatsch K (2007) FET PET for the evaluation of untreated gliomas: correlation of FET uptake and uptake kinetics with tumour grading. Eur J Nucl Med Mol Imaging 34:1933–1942 Rachinger W, Goetz C, Pöpperl G, Gildehaus FJ, Kreth FW, Holtmannspotter M, Herms J, Koch W, Tatsch K, Tonn JC (2005) Positron emission tomography with O-(2-[18 F]fluoroethyl)-l-tyrosine versus magnetic resonance imaging in the diagnosis of recurrent gliomas. Neurosurgery 57:505–511 Ribom D, Eriksson A, Hartman M, Engler H, Nilsson A, Langstrom B, Bolander H, Bergstrom M, Smits A (2001)

K.-J. Langen et al. Positron emission tomography (11)C-methionine and survival in patients with low-grade gliomas. Cancer 92: 1541–1549 Singhal T, Narayanan TK, Jain V, Mukherjee J, Mantil J (2008) 11C-L-methionine positron emission tomography in the clinical management of cerebral gliomas. Mol Imaging Biol 10:1–18 Utriainen M, Metsähonkala L, Salmi TT, Utriainen T, Kalimo H, Pihko H, Mäkipernaa A, Harila-Saari A, Jyrkkiö S, Laine J, Någren K, Minn H (2002) Metabolic characterization of childhood brain tumors: comparison of 18Ffluorodeoxyglucose and 11C-methionine positron emission tomography. Cancer 95:1376–1386 Weber WA, Wester HJ, Grosu AL, Herz M, Dzewas B, Feldmann HJ, Molls M, Stöcklin G, Schwaiger M (2000) O-(2[18F]Fluoroethyl)-L-tyrosine and L-[methyl11 C]methionine uptake in brain tumours: initial results of a comparative study. Eur J Nucl Med 27:542–549 Weckesser M, Langen KJ, Rickert CH, Kloska S, Straeter R, Hamacher K, Kurlemann G, Wassmann H, Coenen HH, Schober O (2005) Initial experiences with O-(2[18 F]fluorethyl)-L-tyrosine PET in the evaluation of primary brain tumors. Eur J Nucl Med 32:422–429 Weckesser M, Matheja P, Rickert CH, Sträter R, Palkovic S, Löttgen J, Kurlemann G, Schober O (2001) High uptake of L-3-[123I]iodo-alpha-methyl tyrosine in pilocytic astrocytomas. Eur J Nucl Med 28:273–281

Chapter 13

Malignant Peripheral Nerve Sheath Tumors: Use of 18FDG-PET/CT Andre A. le Roux and Abhijit Guha

Abstract Malignant Peripheral Nerve Sheath Tumors (MPNST) are rare but lethal lesions. They are more often found in patients with the cancer pre-disposition syndrome, Neurofibromatosis 1 (NF1). NF1 patients harbor many peripheral nerve tumors, the majority of which are benign. Early detection of these benign plexiform neurofibromas (PNf), which may have transformed to a MPNST, is of paramount importance in the long term management and outcome of these patients. However, surveillance of these PNfs is cumbersome as the clinical and radiological features suggestive of malignant transformation are often inaccurate and surgery can lead to significant morbidity. One imaging strategy that is of promise is to take advantage of the fact that malignant tumors inherently are actively proliferating with accompanying increased and altered glucose metabolism, compared to benign lesions. This is the theoretical premise for using 18-Fluorodeoxyglucose (18-FDG)-Positron Emission Tomography (PET) imaging to help diagnose and guide management of MPNSTs. Keywords Neurofibromatosis 1 · 18-FDG PET · MPNST · Plexiform neurofibromas · SUV · CT-PET

Introduction Malignant Peripheral Nerve Sheath Tumors (MPNST) are defined as any malignant lesion developing from or differentiating towards cells from the peripheral A. Guha () Division of Neurosurgery, Toronto Western Hospital, Toronto, ON, Canada e-mail: [email protected]

nerve sheath (Wong et al., 1998). Due to confusion as to cell of origin in MPNSTs which have significant intra- and inter-tumoral heterogeneity, a large number of nomenclatures have been used including malignant schwannoma, neurogenic sarcoma, neurofibrosarcoma and malignant neurilemmoma. They are rare with an incidence of 0.001% in the general population (Karabatsou et al., 2009). They comprise 5–10% of all soft tissue tumors or sarcomas (Frank et al., 2003; Grobmyer et al., 2008), which remains a more common and key differential diagnosis. The gender distribution of MPNST’s is equal in the general population, with the majority of patients being between 20 and 50 years of age. However, up to 10% of patients diagnosed with MPNST are <20 years of age. Half of all MPNST occur in the setting of patients with Neurofibromatosis type 1 (NF1), the most common of all human Cancer pre-disposing genetic syndromes (Benz et al., 2010). In patients with NF1 MPNSTs develop some 10 years earlier than their non-NF1 counterparts (Grobmyer et al., 2008). Various risk factors have been identified for the development of MPNST. In a recent review by Grobmyer et al. (2008), the presence of NF1 was the most important, especially NF1 patients with benign plexiform neurofibromas (PNf). Previous radiation exposure was documented in up to 10% of patients with MPNST, but only 4% of patients with radiation induced sarcomas have MPNST. This radiation exposure may have occurred between 4 and 41 years earlier. The estimated lifetime risk of malignant transformation of a benign PNf in the general population is 3–5% versus 15–20% in patients with NF1 (Karabatsou et al., 2009). In fact, MPNSTs are the most common malignant tumor in patients with NF1 (Karabatsou et al., 2009) and ∼50% of patients with MPNST have NF1.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_13, © Springer Science+Business Media B.V. 2011

127

128

This increased incidence of malignant transformation of PNf is in contrast to schwannomas, although both arise from transformation of schwann cells. Malignant schwannomas are extremely rare and are not reported in the setting of other much rarer schwannoma predisposition syndromes NF2 or NF3, unless stimulated by additional carcinogens such as radiation. It should be noted that although schwann cells are the targets of transformation, the genetic alterations are totally different between the three NF syndromes identified to date, hence representing distinct diseases.

Neurofibromatosis Type 1 (NF1) NF1 is the most common human genetic syndrome with an incidence of 1:3000 live births. Half of the time it is inherited from one of the parents in an autosomal dominant manner, while the rest are due to de-novo mutations in the germline. The responsible NF1 gene, which is altered mainly due to point mutations and to a lesser degree larger deletions, has been cloned and localized to chromosome #17q (Anghileri, 2006). The defective gene encodes a large protein called neurofibromin and is a typical tumor suppressor gene (TSG), with loss of the normal copy in schwann cells leading to formation of benign PNf and similarly other associated cancers in other cell types. NF1 patients harbor a large number of dermal subcutaneous neurofibromas, which rarely if ever undergo malignant transformation. However, a significant percentage of NF1 patients harboring deeper PNf, do have a risk of malignant transformation to a MPNST. It is estimated that the risk of malignant transformation to a MPNST is 20fold higher in PNf compared to dermal neurofibromas (Ratner and Miller, 2006). The exact lifetime risk to a MPNST in NF1 patients varies widely in the literature and ranges from 2 to 29% (D’Agostino et al., 1963; Sorensen et al., 1986; King et al., 2000). However, there is agreement that this risk is more common than a PNf occurring in the general population. Anghileri (2006) estimates that 25–50% of all MPNST present in NF1 patents, while most clinicians site a much higher incidence of 50–60%. Specific mutations in the NF1 gene, which is hard to analyze due to its enormous size, is not able to predict this transformation, except for the observation that it occurs at an higher incidence in NF1

A.A. le Roux and A. Guha

patients with larger chromosomal deletions and not just point mutations in the NF1 gene (Anghileri, 2006). However, most NF1 patients with a MPNST have a small truncating point mutation of the NF1 gene locus resulting in a 8–13% incidence of MPNST. Of interest 5–10% of NF1 patients, who have a larger chromosomal micro-deletion that affects a larger portion of the NF1 gene, results in an increased risk of developing MPNST between16–26% (Birindelli et al., 2001; De Raedt et al., 2003; Perrone et al., 2003; Anghileri, 2006). Although not fully verified by others, Frank et al. (2003) reported a male predominance in patients with NF1 who progress to MPNST and a significant poorer 5-year survival.

MPNST: Presentation, Treatment and Prognosis The commonest presenting features of MPNST include an enlarging mass (new or change in a pre-existing mass) or the development of new symptoms associated with a neurofibroma. These may include pain (highly significant) (King et al., 2000), paresthesias, mass effect from the lesion and the development of new or worsening neurological deficit. MPNST are most commonly found in the proximal upper and lower extremities and pelvis. However, they may occur in uncommon locations and this does not seem to stratify between NF1 and non-NF1 patients (Grobmyer et al., 2008). The primary modality of treatment is radical surgical excision of the nerve and adjacent soft-tissue structures, with the objective of obtaining tumor free margins. Usually this leads to some form of functional deficit if it involves a major nerve and hence the importance of insuring the diagnosis, which as detailed below can be problematic and proper discussion with the patient and family. In addition to the origin of the MPNST, the location is also a determinant of the ability and associated morbidity to undertake radical removal as evidenced by MPNSTs that occur near the spinal column. Radiotherapy regimes include options of preoperative external beam radiation, post-operative external beam radiation, intra-operative high-energy electron irradiation (IOERT) and brachytherapy with

13 Malignant Peripheral Nerve Sheath Tumors

iridium-192. Overall, the benefit of one modality over another is not clear, each having their own theoretical advantages and risks. Chemotherapy agents overall have not been very useful, due to inherent resistance of MPNSTs. Chemotherapies that have been utilized include bleomycin, cisplatin, cyclophosphamide, dacarbazine, dactinomycin, doxorubicin, etoposide, ifosphamide, mitomycin and vincristine. Various regimes of these chemotherapies have been used, including pre-operative, post-operative and pre and postoperative chemotherapy. Overall none have been proven to be effective, with some centers such as ours reserving chemotherapy and its associated systemic toxicities for metastatic MPNSTs. The lack of effectiveness with non-targeted chemotherapy has fuelled the need for biologically targeted therapies against MPNSTs. Increased molecular understanding of MPNSTs has facilitated development of these agents which target aberrant activated signaling pathways such as those mediated by Ras, PI3K/Akt/mTOR as well as receptors such as Epidermal Growth Factor Receptor and Platelet Derived Growth Factor Receptor. Some of these agents are in clinical trials, though likely a multimodal approach with radical surgery, radiation and multiple biologicals will likely be required. The rarity of MPNSTs is a significant hurdle in evaluating these novel therapies, underscoring the need of cooperative clinical trials. Wong et al. (1998) documented a 5- and 10year survival rate of 52 and 34% respectively for all MPNSTs, with a median survival of 61 months. Anghileri (2006) in their large series had a disease specific 10-year mortality rate of 43%. Significant prognostic factors for overall survival included: tumor size, grade and histological subtype; tumor location; positive diagnosis of NF1 and surgical margin clearance (Wong et al., 1998). Stage 1 & 2 patient’s fared significantly better than those with disease that had already metastasized. Similarly lower grade MPNST with lower mitotic rate (<20 per 10 High Power Field: Hruban et al., 1990) and lack of necrosis fared better than higher grade tumors as per the grading schema used for soft-tissue sarcomas in general, patients with abdominal and central disease have the worst outcome compared to peripheral MPNSTs, which were likely detected earlier and could undergo radical tumor-free marginal resection. In support, another study reported a threefold increase in disease recurrence in abdominal

129

or central MPNSTs, with patients harboring head and neck tumors doing the worse (Anghileri 2006). Patients with NF1 tended to have larger tumors and a higher mitotic count, than found in patients without this predisposition. NF1 patients also tended to present at a later stage of disease, perhaps due to their overall tumor burden and cognitive deficiencies. However, if stratified by stage outcome was not different between NF1 and non-NF1 MPNST patients (Anghileri 2006). Doorn et al. (1995) also looked at the difference in outcome between NF1 and nonNF1 patients in their series of 22 patients where half were in the setting of NF1. There was a uniformly poor outcome in NF1 patients with MPNST, with a high incidence of local and distant treatment failure of 63.6%. The overall outcome between the two groups was the same but the NF1 patients had a high risk for developing a second MPNST. Doorn et al. (1995) also documented a median survival time of 20 months for patients with local treatment failure and only 3 months for patients with distant failure. Patients with NF1 were diagnosed at a younger age as compared to patients without NF1 and their tumors were usually larger in size. Doorn et al. (1995) also found that survival was worse if symptoms were present for more than 6 months. Recurrence of disease, either local and certainly distant, is associated with a poorer outcome, with a 2.5-fold increase in mortality Anghileri (2006). Local recurrence ranged from 20 to 40% (Grobmyer et al., 2008), with most occurring within 24 months of surgery. Distant metastatic disease is common and the most important cause of mortality. Like other sarcomas metastasis is mainly by the bloodstream, involving the lung, liver, brain and lymph nodes. Clear surgical margins is the singular determinant of increased survival along with grade of the MPNST, whereas irradiation did not show a dramatic difference on outcome. Anghileri (2006) reported a 2.4-fold increase in disease recurrence with positive surgical margins and an associated increased mortality rate in these patients. Complete tumor excision with negative surgical margins was also the single most significant prognostic factor for survival and local disease control in another study (Wong et al., 1998). Of interest was that this increased risk (in patients with positive surgical margins) was not related to tumor location or size of tumor. However, patients with tumors smaller than 10 cm had better outcomes likely reflecting the ability to obtain clear surgical margins in smaller

130

tumors rather than an inherent difference in the biology (Ducatman et al., 1986; Hruban et al., 1990). In some series brachytherapy and IOERT was found to be beneficial, but this has not been universally demonstrated and not utilized at most centers. Of interest, radiation induced MPNST patients tended to fare worse compared to patients with a de novo mode of origin (Sordillo et al., 1981; Wong et al., 1998). The risk for malignant transformation of a benign PNf to a MPNST is highest in NF1 patients due to their cancer pre-disposition and hence radiation is not recommended to be used unless the diagnosis of MPNST is pathologically established.

MPNST: Macroscopic, Microscopic and Cytogenetic Features Neurofibromas are intermixed with functional nerve fascicles, unlike schwannomas, hence associated with a higher morbidity when surgically manipulated. One of the main dilemmas of these patients and their caring physicians is to determine when a PNf has transformed to a MPNST. Patients with PNf need to be continually followed as transformation may occur at any given stage. This surveillance may prove to be problematic as the clinical and radiological features of malignant transformation may be obscure. Some benign lesions may also de-differentiate without any initial change in clinical symptoms or radiological appearance. The normal anatomy of a nerve sheath shows a variety of cell types. These include perineural cells, mesenchymal type cells (fibroblasts, pericytes, endothelial cells and epineurial lipocytes) and Schwann cells. Some MPNST shows some degree of Schwannian differentiation (Wong et al., 1998). Most of these lesions do arise de-novo from normal peripheral nerves or from pre-existing neurofibromas and as mentioned not schwannomas which occur sporadically or associated with NF2 and NF3 tumor pre-disposition syndromes (Frank et al., 2003; Grobmyer et al., 2008). Features of schwann cell differentiation with positive staining for the S-100 schwann cell marker may be seen in MPNST, but in high-grade MPNST, this can be absent in more than 50% of the cases (Wong et al., 1998). Macroscopically MPNST are usually fusiform in shape, enlarging the nerve from which they originate, much like the gross and radiological features of benign

A.A. le Roux and A. Guha

PNf. In a series by Wong et al. (1998), the largest tumors were found in the abdominal compartment where more than half exceeded 10 cm in greatest diameter. Size is an important prognosticator as it impacts on the surgeons ability to undertake total removal including the surrounding soft-tissue structures to minimize risk of metastasis. The most common nerves of origin were the brachial plexus and the affected nerves could be identified in about a third of cases at time of surgery. A positive history of NF1 was present in 24% of their patients. The cut surface has a varying appearance, ranging from a fleshy tan look to one with hemorrhage and necrosis, while some may be fibrous and firm. Although MPNST usually have a pseudo capsule while benign lesions tend to be more encapsulated (Grobmyer et al., 2008), it is not possible to distinguish between the two on macroscopic features only. The microscopic histological features of MPNST include spindled cells arranged in a fascicular growth pattern (low magnification) with hyperchromatic nuclei and an increase in mitotic figures (>4/HPF). Grobmyer et al. (2008) described these lesions as “marbled” with light and dark areas alternating one another. The spindle cells are large and uniform with the nuclei having a buckled appearance. Necrosis with or without pseudopalisading is also commonly seen. Low-grade lesions tend to have less mitotic figures and little or no necrosis. Various histological subtypes are described, with de-differentiation present in the highgrade MPNST, but the scope of this description falls outside this chapter and the interested reader is referred to the appropriate text. The karyotypes of patients with MPNST show complex numerical and structural aberrations (Schmidt et al., 1999), in addition to the primary alterations in the NF1 gene locus on chromosome #17q. In these additional genetic alterations likely resides the molecular explanation of the transformation from a benign PNf to a MPNST. Many gain and loss of chromosomal regions have been identified in MPNST with high resolution array CGH (aCGH) analysis, with the main unresolved issue being which are causal for the transformation and progression. Chromosomal alterations include regions on 1p, 9p, 11, 12p, 14q, 17q, 22q, X and Y. Gain of material has been limited to chromosomes 2 and 7 (Plaat et al., 1999). Frank et al. (2003) documented a series in which all cases showed aberrant near triploid karyotypes with several structural and numerical aberrations. A partial deletion and loss of

13 Malignant Peripheral Nerve Sheath Tumors

heterozygosity of 1p has been described in MPNST, suggesting that loss of 1p may be involved in the tumorigenesis of MPNST and other solid tumors (neuroblastomas and pheochromocytomas) (Frank et al., 2003). Complete or partial loss of chromosome# 9 is also a common feature of sporadic and NF1 associated MPNST. The NF genes (1 & 2) and their gene products act as tumor suppressor genes with decreased expression leading to an increased incidence of tumor development. Much is yet to be discovered and hopefully the further development of this field will increase our understanding of this highly complex disease.

131

has been made, as there are no criteria to differentiate MPNST from benign lesions (Grobmyer et al., 2008). CT is the imaging modality of choice for metastatic disease, specifically disease that involves the liver and/ or lung. Lesions of the retro-peritoneum can be evaluated in great detail with CT. Gallium-67 scan has also been used to identify malignant transformation of neurofibromas in patients with NF1. Gallium uptake was associated with malignant transformation, but this has not borne out to be reliable modality, as there may be necrosis and associated inflammatory cells which uptake Gallium in large plexiform neurofibromas, hence is not routinely used.

MPNST: Imaging Modalities

18-Fluorodeoxyglucose (18-FDG)-Positron Emission Tomography As mentioned this differentiation between a benign PNf which perhaps is growing but not biologically (PET) transformed to a MPNST is the main dilemma for the patient and physician, especially in the context of NF1 where these lesions are often multiple. The imaging modality of choice for soft tissue tumors is magnetic resonance imaging (MRI). Common features of most benign or malignant neurofibromas include a fusiform shape with tapered ends, the lesion longitudinally orientated in the anatomical plane of a peripheral nerve, fascicular appearance, atrophy of surrounding muscles and the split-fat sign (Grobmyer et al., 2008). However, MRI has a limited ability to distinguish between benign and malignant lesions. Some classical features of benign lesions are proposed and include a central hypo-intense area (called the “target sign”) on T2 images. This “target sign” is often absent in MPNST (Grobmyer et al., 2008). However, other radiological features, including that of contrast enhancement, irregular margins, location and in-homogenous signal intensity are non-specific for MPNST and hence offer little help in the early identification of malignant transformation. Rapid size increase is a significant finding, but the availability of previous imaging to serve for comparison is not always possible. Lesions that invade surrounding tissues, fat planes, have poorly defined margins and peri-lesional edema are all suggestive, but not pathognomic, of a malignant process or transformation. In only about 10% of MPNST may these lesions found to be multiple (Ozturk et al., 2007). Other imaging modalities include CT scans, which may be of benefit once the diagnosis of a MPNST

The diagnostic dilemma to distinguish benign plexiform neurofibroma from MPNST, with its importance in early detection to improve overall survival and minimize morbidity has led to the interest in 18-Fluorodeoxyglucose (18-FDG)-Positron Emission Tomography (PET) (Cordona et al., 2003). PET imaging is based on the metabolic characteristics of the tissue/lesion and thus differs from normal imaging in that highly metabolic active tissue demonstrates a “hot spot”. Malignant lesions show an increased volume of proliferating cells with increased glucose transport and utilization and up-regulation of hexokinase activity as compared to normal tissue (Beyer et al., 2000). Glucose can be labeled and thus imaged by 18-FDG PET. Imaging with 18-FDG-PET has been used for identification of other human tumors including head and neck cancers, colon cancer, lung cancers, soft tissue tumors and gliomas. For example, the identification of primary sarcomas and even identification of pulmonary metastasis from bone or soft tissue sarcomas has been enhanced by PET imaging as reported by Iagaru et al. (2006). PET imaging has the potential to distinguish intra-tumoral pathological heterogeneity, as found commonly in MPNST, where within the same tumor there may be regions which are similar to a benign plexiform neurofibroma while other regions may represent low and high grade changes. The worse grade dictates the biological and clinical behavior of

132

the entire MPNST. Biopsy can theoretically be directed at PET “hot spots” within an MPNST to help diagnosis the most aggressive biological component of the tumor. In addition to cancer, PET imaging has also been used to differentiate between inflammatory tissue and tumor tissue. This non-cancer specificity is of course the limitation of PET imaging as the confidence of its result relies on the difference in metabolism between normal and abnormal tissue. Hence, some low-grade tumors will be missed while non-cancerous inflammatory process secondary to the tumor or treatments given such as surgery and radiation may lead to false positivity. Recently the value of PET to distinguish between schwannomas and neurofibromas was documented by Benz et al. (2010). Hence, PET can be used to follow non-invasively benign PNfs and also potential early detection of non-invasive in both sporadic and more importantly NF1 patients who may harbour multiple such lesions. In addition, this study also demonstrated the usefulness of PET to distinguish regions of MPNST with intra-tumoral heterogeneity, such as MPNST’s with both benign and malignant components. The use of 18-FDG-PET to indicate metabolic activity is beyond debate, however, this imaging modality does not provide precise anatomical imaging. Currently, the MRI anatomical imaging is matched side by side as closely as possible by the clinician, however, the development of CT-PET where imaging from both these modalities can be fused has helped with this obstacle. Antoch et al. (2004) found that using both CT-PET in conjunction with MRI imaging provided more accurate diagnostic capability, compared to the imaging modalities in isolate. In addition, CT-PET was helpful in staging and able to detect asymptomatic metastatic disease or a asymptomatic MPNST in another neurofibroma. It is hoped that with our increasing familiarity with CT-PET in MPNST, in the future one may be able to undertake stereotactic guided biopsy of “hot spots” to minimize morbidity and optimize diagnostic ability. Its usefulness will likely not be in those patients where other clinical and radiological characteristics are highly suggestive of MPNST’s as these patients will require radical compartmental surgery to minimize recurrence as previously described. However, in known benign PNf, especially in NF1 patients, where there is some increase in pain or radiological size, a percutaneous biopsy of the CET-PET hot spot may very well alleviate the requirement of open surgery and associated morbidity.

A.A. le Roux and A. Guha

Although these benefits of PET and CT-PET imaging to augment the anatomical MRI images are of potential high importance, sensitivity and specificity parameters in MPNST still remain to be proven. Much of the problem is intrinsic to the rarity of MPNST’s, where no one singular institute sees enough of these tumors to evaluate the value of PET in a rigorous manner. Second, there is variation in uptake of FDG in MPNST’s, speculated to be a result of differences in hypoxia-inducible factor 1α (HIF1α) (Shintani et al., 2006). There is also variation in isotope type used, the actual PET or CT-PET imaging machine, variation in interpretation and cut-off of standardized uptake values (SUV) and their pathological meaning etc. Additional limitations include availability of the hardware (PET or CT-PET) with its large initial set up costs, isotope lability and requirement for a local cyclotron manufacturing and the need of interested nuclear medicine and radiology specialist for interpretation of the collected data. Although the role of PET and CT-PET is currently uncertain, further experience, especially in a multi-institutional standardized manner will likely provide us a better idea of the utility of this promising imaging modality for MPNST’s. In the future, a management protocol as schematized below in Fig. 13.1 may evolve for optimal screening and diagnosis of MPNST’s. Towards this we have recently proposed such a scheme at our institute, though we still have not evaluated enough patients to make firm comments as to the role of CT-PET in MPNST’s. Based on our evaluation of 10 NF1 suspected MPNST patients (Karabatsou et al., 2009), those tumors with a SUV <4 and a low index of clinical suspicion had benign PNf which did not change to a MPNST over a follow up of many years. Most of these patients can safely be followed and repeat imaging done on an interval determined by the clinicians index of suspicion. It is our practice to follow such NF1 patients annually. The other group of patients where there was little debate are those tumors with a SUV >7. These patients all had high-grade MPNST’s and went on to open biopsy with sampling of the “hot spot”, since we do not have the set up for percutaneous biopsies for MPNST’s yet. With a positive biopsy of MPNST confirmed, these patients went on to radical compartmental resection which includes the nerve of origin and adjacent softtissue structures to obtain tumor-free margins. Pre- or post-operative radiation is given with chemotherapy

13 Malignant Peripheral Nerve Sheath Tumors

133

Fig. 13.1 Proposed management strategy based on limited experience with CT-PET at University of Toronto

Clinical & Radiological Suspicion of MPNST

MRI scan & 18-FDG-PET/CT scan

Measurement of Standard Uptake Value (SUV) 4 hours post injection

SUV < 4

SUV 4 –7

SUV > 7

Close follow-up

Multiple open biopsies (?Low grade MPNST?)

Likely high grade MPNST: Biopsy of “hot spot”

Established Diagnosis of MPNST

Radical Compartmental Resection with Tumor-Free Margins

reserved for those who develop metastatic disease. The middle group of patients with SUV 4–7 is the diagnostic conundrum. These patients have lesions that are of suspect of harboring likely low grade MPNST’s, but not enough numbers to date keeps this a open question. The options in management include either careful follow-up if low clinical suspect or currently open multiple biopsies and perhaps in the future CT-PET guided biopsy of the suspicious area. In summary, what is the threshold of SUV values to distinguish benign neurofibroma from a lowgrade MPNST is not known and likely differs from one imaging facility to another, for reasons discussed below. As in all matters in medicine, there are controversies with the role of PET imaging in MPNST’s. The threshold of SUV for quantitative analysis and provide enough sensitivity and specificity towards the diagnosis of MPNST remains unresolved. In contrast to our findings, Ferner et al. (2000) indicated that a SUV of 2.5 should reliably detect malignancy although an overlap may exist in the range between 2.7 and 3.3, likely a result of different machines and interpretation by our radiology colleagues. In addition to the actual cut-off SUV to be used, there are technical issues that also stir controversy. The optimal time after 18-FDG

injection at which the SUV value should be read, has not yet determined. Most units use between 1 and 4 h, but there is no standard. The actual method of SUV calculation also differs and depends on the imaging analysis software used, leading to different ways of reporting the SUV. For example, one can use the average or the maximal SUV. In our limited experience we found that the maximal SUV is more accurate to report on heterogenous tumors such as MPNSTs (Karabatsou et al., 2009). In conclusion, MPNST are lesions with poor prognosis, which although rare in the total population is relatively common in the setting of NF1. Obstacles with screening, diagnosis and optimal management do exist, leading to lethality of these tumors. Early detection followed by radical surgery and obtaining tumor free margins is critical to minimize local and distant recurrence and improve overall survival. 18FDG PET/CT may play an important role in screening and diagnosis and augment current MRI based imaging. However, although individual centers have reported some promise, they have been on limited patient numbers. Larger scale multi-institutional studies are required not only to determine the role of CTPET, but also optimize current and emerging biological therapies.

134

References Anghileri M (2006) Malignant peripheral nerve sheath tumors – prognostic factors and survival in a series of patients treated at a single institution. Cancer 107:1065–1074 Antoch G, Saoudi N, Kuehl H, Dahmen G, Mueller SP, Beyer T, Bockisch A, Debatin JF, Freudenberg LS (2004) Accuracy of whole body dual-modality fluorine-18-2-fluoro-2-D-glucose positron emission tomography and computed tomography (FDG-PET/CT) for tumor staging in solid tumors: comparison with CT and PET. J Clin Oncol 22:4357– 4368 Benz MR, Czernin J, Dry SM, Tap WD, Allen-Auerbach MS, Elashoff D, Phelps ME, Weber WA, Eiber F (2010) Quantitative F18-fluorodeoxyglucose positron emission tomography accurately characterizes peripheral nerve sheath tumors as malignant or benign. Cancer 116:451–458 Beyer T, Townsend DW, Brun T, Kinahan PE, Charron M, Roddy R, Jerin J, Young J, Byars L, Nutt R (2000) A combined PET/CT scanner for clinical oncology. J Nucl Med 41(8):1369–1379 Birindelli A, Perrone F, Oggionni M, Lavarino C, Pasini B, Vergani B, Ranzani GN, Pierotti MA, Pilotti S (2001) Pathway alterations in sporadic and NF1-related malignant peripheral nerve sheath tumours. Lab Invest 81:833–844 Cardona S, Schwarzbach M, Hinz U, Dimitrakopoulou-Strauss A, Attigah N, Mechtersheimer G, Lehnert T (2003) Evaluation of F18-deoxyglucose positron emission tomography (FDG PET) to assess the nature of neurogenic tumours. Eur J Surg Oncol (EJSO) 29:536–541 D’Agostino AN, Soule EH, Miller RH (1963) Sarcomas of the peripheral nerves and somatic soft tissue associated with multiple neurofibromatosis (von Recklinghausen’s disease). Cancer 16:1015–1027 De Raedt T, Brems H, Wolkenstein P, Vidaud D, Pilotti S, Perrone F, Mautner V, Frahm S, Sciot R, Legius E (2003) Elevated risk for MPNST in NF1 microdeletion patients. Am J Hum Genet 72:1288–1292 Doorn PF, Molenaar WM, Buter J, Hoekstra HJ (1995) Malignant peripheral nerve sheath tumors in patients with and without neurofibromatosis. Eur J Surg Oncol 21:78–82 Ducatman BS, Scheithauer BW, Piepgras DG (1986) Malignant peripheral nerve sheath tumors: a clinicopathologic study 120 cases. Cancer 57:2006–2021 Ferner RE, Lucas JD, O’Doherty MJ, Hughes RA, Smith MA, Cronin BF, Bingham J (2000) Evaluation of (18) fluorodeoxyglucose positron emission tomography (18 FDG PET) in the detection of malignant peripheral nerve sheath tumors arising from within plexiform neurofibromas in neurofibromatosis 1. J Neurol Neurosurg Psychiatry 68:353–357 Frank D, Gunawan B, Holtrup M, Füzesi L (2003) Cytogenetic characterization of three malignant peripheral nerve sheath tumors. Cancer Genet Cytogen 144:18–22

A.A. le Roux and A. Guha Grobmyer SR, Reith JD, Shahlaee A, Bush CH, Hochwald SN (2008) Malignant peripheral nerve sheath tumor: molecular pathogenesis and current management considerations. J Surg Oncol 97:340–349 Hruban RH, Shiu MH, Senie RT (1990) Malignant peripheral nerve sheath tumors of the buttock and lower extremity: a study of 43 cases. Cancer 66:1253–1265 Iagaru A, Quon A, McDougall IR, Gambhir SS (2006) F-18 FDG PET/CT evaluation of osseous and soft tissue sarcomas. Clin Nucl Med 31:754–760 Karabatsou K, Kiehl T-R, Wilson DM, Hendler A, Guha A (2009) Potential role of 18 fluorodeoxyglucose positron emission tomography/computed tomography in differentiating benign neurofibroma from malignant peripheral nerve sheath tumor associated with Neurofibromatosis 1. Neurosurgery 65:A160–A170 King AA, DeBaun MR, Riccardi VM, Gutmann DH (2000) Malignant peripheral nerve sheath tumors in Neurofibromatosis 1. Am J Med Gene 93:388–392 Ozturk E, Erdem I, Sonmez G, Haholu A, Sildiroglu HO, Mutlu H, Basekim CC, Kizilkaya E (2007) Multicentric malignant peripheral nerve sheath tumor. Clin Imaging 31: 363–366 Perrone F, Tabano S, Colombo F, Dagrada G, Birindelli S, Gronchi A, Colecchia M, Pierotti MA, Pilotti S (2003) p15INK4b, p14ARF and p16INK4a inactivation in sporadic and neurofibromatosis type 1-related malignant peripheral nerve sheath tumours. Clin Cancer Res 9:4132–4138 Plaat BE, Molenaar WM, Mastik MF, Hoekstra HJ, te Meerman GJ, van den Berg E (1999) Computer assisted cytogenetic analysis of 51 malignant peripheral nerve sheath tumors: sporadic vs. neurofibromatosis type 1 associated malignant schwannomas. Int J Cancer 83:171–178 Ratner N, Miller SJ (2006) Model systems for neurofibroma and malignant nerve sheath tumor drug discovery today: disease models. Cancer 3(2):175–182 Schmidt H, Würl P, Taubert H, Meye A, Bache M, Holzhausen H-J, Hinze R (1999) Genomic imbalances of 7p and 17 q in malignant peripheral nerve sheath tumors are clinically relevant. Gene Chromos Cancer 25:205–211 Shintani K, Matsumine A, Kusuzaki K (2006) Expression of hypoxia inducible factor (HIF) 1-alpha as a biomarker of outcome in soft tissue sarcomas. Virch Arch 449:673–681 Sordillo PP, Helson L, Hajdu SI (1981) Malignant schwannoma – clinical characteristics, survival and response to therapy. Cancer 47:2503–2509 Sorensen SA, Mulvihill JJ, Nielsen A (1986) Long-term follow up of von Recklinghausen neurofibromatosis. N Engl J Med 314:1010–1015 Wong WW, Hirose T, Scheithauer BW, Schild SE, Gunderson LL (1998) Malignant peripheral nerve sheath tumor: analysis of treatment and outcome. Int J Rad Oncol Biol Phys 42(2):351–360

Chapter 14

Brain Tumors: Evaluation of Perfusion Using 3D-FSE-Pseudo-Continuous Arterial Spin Labeling Hanna Järnum, Linda Knutsson, and Elna-Marie Larsson

Abstract In this chapter, the advantages and disadvantages of 3D fast spin echo (FSE) pCASL as a measure of brain tumor perfusion are reviewed. In addition, we compare pCASL with other perfusion techniques and discuss future ASL approaches. A prospective study of 28 patients with contrast-enhancing brain tumors was performed at 3 T using dynamic susceptibility contrast (DSC) MRI and pCASL. The visual qualitative evaluation of signal enhancement in tumor was scored from 0 to 3 (0 = no signal enhancement compared with white matter (WM), 3 = pronounced signal enhancement with similar or higher signal intensity than in gray matter (GM)/basal ganglia). The extent of susceptibility artifacts in the tumor was scored from 0 to 2 (0 = no susceptibility artifacts, 2 = extensive susceptibility artifacts (maximum diameter >2 cm). Absolute ASL cerebral blood flow (CBF) values in tumor, GM, WM, and cerebellum were also measured. Using normalized tumor blood flow values (ASL nTBF, DSC nTBF), tumor-to-healthy tissue perfusion ratios were calculated by dividing the mean value of tumor ROI by the mean value in ROIs in the two cerebellar hemispheres. ASL had in comparison with DSC-MRI both a lower signal enhancement score (p = 0.02) and a lower susceptibility artifact score (p < 0.01). The highest absolute ASL CBF values were measured in tumor tissue and the lowest in WM. There was a good correlation between DSC nTBF and ASL nTBF values, with a correlation coefficient of 0.82.

H. Järnum () Department of Radiology, Aalborg Hospital, Arhus University Hospital, 9000 Aalborg, Denmark e-mail: [email protected]

Keywords FSE · Tumor · Arterial spin labeling · Perfusion analysis · Meningiomas · pCASL

Introduction Tumors of the central nervous system are prevalent and often lethal or chronically disabling. Brain tumors are a leading cause of cancer-related deaths, being diagnosed in 10–15 per 100,000 persons per year in both Europe and the United States (Ohgaki and Kleihues, 2005). Gliomas constitute more than 90% of primary brain tumors diagnosed after the second decade of life (Kleihues et al., 1995). Glioblastoma multiforme is the most common glioma in adults; approximately 4,800 new cases are seen each year in the United Kingdom and about 17,500 in the United States (Rahman et al., 2010). Despite optimal treatment with chemotherapy, radiation, or radiosurgery, the long-term survival is poor, only 8% of patients surviving for 4 years (Rahman et al., 2010). Glioblastoma multiforme together with other highgrade tumors, e.g. metastases, lymphoma, and several benign intracranial tumors (i.e. meningiomas), all exhibit increased vascularity (Jarnum et al., 2010). Much time and effort are spent on generating new and effective antiangiogenic drugs against malignant gliomas. Monitoring of such therapies requires a reliable tool for perfusion measurements. Cerebral perfusion defines the process involved in the delivery of nutrients and oxygen from arterial blood to the capillary bed within the brain tissue. Neovascularization and hypervascularity may be assessed by several techniques. One method is to employ the nuclear medicine technique, positron emission tomography

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_14, © Springer Science+Business Media B.V. 2011

135

136

(PET), which involves injection of a labeled gammaemitting radioisotope into the bloodstream of the patient. PET allows for quantification of cerebral blood flow (CBF). Cerebral blood flow is defined as the volume of blood delivered to a given mass of brain tissue in a given time, stated in ml/min/100g. By computed tomography (CT) using x-rays and intravenous injection of iodinated contrast medium, perfusion parameters such as CBF and cerebral blood volume (CBV) can be calculated. Cerebral blood volume describes the fraction of a voxel that contains blood vessels and is usually expressed in ml/100g or as a percentage. Although many perfusion parameters have been proposed, CBF and CBV are the two most commonly used perfusion parameters. In contrast to PET and CT, magnetic resonance imaging (MRI) offers a non-ionizing approach. Dynamic susceptibility contrast MRI (DSC-MRI) together with arterial spin labeling (ASL) constitute the two main MRI perfusion techniques. T1-weighted dynamic contrast enhanced (DCE) MRI is a third technique that can be used to assess brain perfusion as well as vascular permeability (Larsson et al., 2009). Dynamic susceptibility contrast MRI utilizes a paramagnetic gadolinium-chelated contrast agent to visualize perfusion. The passage of the contrast agent induces a signal intensity decrease due to induced local magnetic field gradients that extend from the vascular compartment into the surrounding tissue. The decrease in signal intensity can be converted into a concentrationtime curve from which perfusion parameters can be derived. Gadolinium-based contrast agents are contraindicated in patients with renal failure because of the risk of developing nephrogenic systemic fibrosis (NSF). Unlike DSC-MRI, ASL is a promising tool for absolute quantification of CBF. It uses arterial water as an endogenous tracer by magnetically labeling blood–water protons in a cerebral feeding artery. In addition, ASL facilitates longitudinal and serial studies, because it allows for multiple repetitive follow-ups without the risks that are associated with ionizing radiation. The following is a summary of the diagnostic performance of 3D fast spin echo (FSE) pseudocontinuous arterial spin labeling (pCASL) for evaluation of brain tumor perfusion. The methodology section meticulously delineates our set-up of the sequence (Jarnum et al., 2010). Significant outcomes from our study (Jarnum et al., 2010) are briefly reported in the

H. Järnum et al.

results section and thoroughly reviewed in the discussion together with interesting findings from other perfusion studies.

Methods: Arterial Spin Labeling The general principle for obtaining CBF is common for all ASL techniques: to produce a label image and a control image. During the label experiment, an inversion or saturation RF pulse is applied below the imaging slab. After a transit delay, the labeled spins reach the imaging plane, where a blood-tissue-water exchange takes place, rendering a signal decrease. This MRI signal is lower than the one produced during the control experiment, during which the spins are not affected by a RF pulse. Cerebral blood flow is calculated by subtracting label images from control images because the difference in arterial magnetization between control and label is directly related to the amount of arterial blood delivered. The signal difference is only about 1–3% of the total signal in the image, thus multiple repetitions are needed. Depending on how the labeling pulse is applied, two branches of ASL exist: pulsed ASL (PASL) and continuous ASL (CASL). In CASL, the supplying blood is continuously labeled below the imaging slab until the tissue magnetization reaches a steady state, whereas the PASL approach labels a thick slab of arterial blood at a single instance in time, and the imaging is performed after a time long enough for the labeled blood to reach the tissue and exchange at the region of interest (Petersen et al., 2006). Both methods have their pros and cons, which are described in detail elsewhere (Petersen et al., 2006). Pseudo-continuous arterial spin labeling (pCASL) is a form of CASL in which the labeling RF pulses are not played out continuously. However the advantages of continuous labeling are maintained, i.e. high signal to noise ratios.

3D-FSE-Pseudo-Continuous Arterial Spin Labeling In our study, MR scanning was executed on a 3 T whole-body MRI system (Signa HDx, R14M5, GE Healthcare) by using an eight-channel receive-only head array coil. Perfusion imaging was performed

14 Brain Tumors

137

Fig. 14.1 Combined labeling and background suppression preparation sequence: The preparaion begins with repeated selective saturation and a single, selective adiabatic inversion pulse for background suppression. Thereafter, labeling is applied for 1.5 s. After the labeling, four inversion pulses are applied

for optimal background suppression. In between these inversions, pulses are inferior saturation pulses to minimize signal from inflowing blood arriving after the end of labeling. With permission from Neuroradiology

using pCASL, background suppression, and a stack-ofspiral 3D fast spin echo imaging sequence (Fig. 14.1). Pseudo-continuous ASL followed the method of Dai et al. (2008). It employed Hanning RF pulses of 500 μs duration spaced 1,500 μs apart, an average RF amplitude of 1.8 μT, an average gradient of 0.9 mT/m, and a gradient amplitude during the RF pulses of 9 mT/m. Labeling was performed from 3,000 to 1,500 ms before image acquisition. Background suppression was achieved by selective saturation at 4,100 ms before imaging, selective inversion just before the labeling begins at 3,000 ms before imaging, and then weakly selective inversion pulses applied at 1,500, 680, 248, and 57 ms before imaging. The selective inversion and saturation pulses were applied to a slab containing the imaged region and ending at the labeling plane. All inversion pulses were adiabatic pulses. In addition to the background suppression pulses, inferior saturation pulses were applied at 1,037, 392, and 116 ms to suppress inflowing arterial blood spins after completion of labeling. A reference image volume for quantification was also acquired using a simple saturation recovery preparation with saturation applied 2,000 ms before image acquisition. After ASL preparation, images were acquired with an interleaved 3D stack-of-spiral FSE sequence with the following parameters: 512 sampling points on eight spirals, reconstructed matrix 128 × 128, TR = 9.2 ms, TE = 1.9 ms, NEX = 3, slice thickness = 5 mm, and acquisition time = 5:36 min. The eight axial spiral interleaves were acquired to encode images with 3.7mm resolution on a 24-cm field of view. Each FSE

train acquired all slices encoded for a particular spiral interleave, and subsequent interleaves were acquired after additional preparation and acquisition. The total spiral duration of 4.1 ms combined with the FSE refocusing helped to minimize sensitivity to field nonuniformity. Three averages of the label and control pairs were obtained. Quantification was performed using the model of Alsop and Detre (1996) with the inclusion of a term for the finite labeling duration as in Wang et al. (2005) and correction for the incomplete recovery of the tissue signal in the reference image due the saturation performed tsat (2,000 ms) before imaging. Flow was calculated with the equation − tsat (Sctrl −Slbl ) 1−e T1g w λ T1b f= e τ − Sref 2αT1b 1 − e T1b where f is the flow, S is the signal on the control, label, or reference image, T1b is the T1 of blood, T1g is the T1 of gray matter, α is the labeling efficiency, λ is the brain-blood partition coefficient, τ is the labeling duration (1,500 ms), and w (1,500 ms) is the postlabeling delay time. This equation assumes that the label is primarily in the microvasculature rather than in the tissue, so that the T1 of blood rather than that of tissue is used throughout. We used an estimate of gray matter (GM) T1, 1,200 ms, for the correction of incomplete recovery, but the effect was relatively small, and therefore differences in tissue T1 should not have a major effect on quantification. We used an assumed T1 for blood of 1,600 ms (Lu et al., 2004) and an average

138

H. Järnum et al.

brain value for λ of 0.9 (Herscovitch and Raichle, 1985). The labeling efficiency was assumed to be the product of 0.8 for pCASL efficiency and 0.75 for the additional attenuation from the background suppressed pulses (Garcia et al., 2005; Dai et al., 2008; Jarnum et al., 2010).

Perfusion Analysis Three dimensional FSE pCASL was used in our analysis to assess CBF in 28 contrast-enhancing brain tumors: 12 grade IV glioblastomas, 2 grade III anaplastic astrocytomas, 1 grade III anaplastic oligodendroglioma, 3 grade II oligodendrogliomas, 1 grade II astrocytoma, 4 meningiomas, 3 metastases, 1 lymphoma, and 1 primitive neuroectodermal tumor (PNET). Tumor regions-of-interest (ROIs), range 0.28–2.0 cm2 , were manually drawn in an area

T1Gd

ASL CBF

T1Gd

ASL CBF

Fig. 14.2 Upper row: Forty-nine-year-old woman with glioblastoma multiforme in the frontal lobes. The region within the tumour with pronounced signal enhancement is marked with circle on ASL CBF, DSC rCBF (color-coded)

with maximum signal enhancement on a gray-scale DSC relative CBF (rCBF) map and copied to the corresponding DSC relative CBV (rCBV) map using in-house developed software. Effort was made to localize the same ROI position on ASL CBF as on the DSC-MRI maps (rCBF and rCBV). The visual evaluation of signal enhancement in tumor tissue was scored from 0 to 3, where 0 = no signal enhancement compared with white matter (WM), 1 = slightly higher signal enhancement than WM, 2 = moderate signal enhancement (i.e. slightly lower signal enhancement than cortical GM/basal ganglia), and 3 = pronounced signal enhancement with similar or higher signal intensity than in GM/basal ganglia (Fig. 14.2). Qualitatively, we also approximated the number of susceptibility artifacts (signal loss/displacement) and scored them from 0 to 2, where 0 = no susceptibility artifacts, 1 = small/moderate susceptibility artifacts (maximum diameter <2 cm),

DSC rCBF

DSC rCBF

DSC rCBV

DSC rCBV

and DSC rCBV (gray scale). Lower row: A 54-year-old woman with an olfactory groove meningioma. Susceptibility artifacts are prominent on DSC-MRI maps (rCBF and rCBV) but not on ASL CBF. With permission from Neuroradiology

14 Brain Tumors

139

which did not affect tumor evaluation, and 2 = extensive susceptibility artifacts (maximum diameter >2 cm), which worsened tumor evaluation (Fig. 14.2). Quantitatively, absolute perfusion values were measured with ASL CBF in the above-mentioned ROIs in tumor tissue, in the central parts of the bilateral cerebellar hemispheres, and in contralateral normal appearing GM and WM. Normalized perfusion values in tumor tissue were used when pCASL was compared with DSC-MRI because in clinical practice only a relative perfusion measurement is usually possible with DSC-MRI (Petersen et al., 2006). Normalized tumor blood flow (nTBF) values were thus evaluated on ASL CBF (ASL nTBF), DSC rCBF maps (DSC nTBF) and DSC rCBV maps (DSC nTBV). Tumor-to-healthy tissue perfusion ratios were calculated by dividing the mean value of tumor ROI by the mean value in ROIs in the two cerebellar hemispheres.

signal enhancement was lower in ASL CBF than in DSC-MRI maps: 70 versus 80. There was a statistically significant difference in total visual score for signal enhancement between the two sequences (p = 0.02). Nineteen out of 28 patients had the same visual score with ASL and DSC-MRI.

Statistical Analysis

Quantitative Perfusion Analysis

Comparison between 3D FSE pCASL and DSC-MRI was performed with the Wilcoxon signed rank test (significance level set at p < 0.05) for both the visual score in the tumor region with pronounced signal enhancement and the susceptibility-artifact score. Linear regression and Pearson’s correlation were used to evaluate the association between quantitative results. The relationship between normalized perfusion values (i.e. ASL nTBF and DSC nTBF) was quantified by the coefficient of agreement (Bland and Altman, 1999), which is 1.96 × SD, where SD is the standard deviation of the difference between observed ASL nTBF and predicted ASL nTBF from DSC nTBF. These statistical analyses were performed using SigmaStat 3.5. (Systat Software, Inc., Point Richmond, CA, USA).

The absolute CBF value for all 28 tumors was 70.7 ± 46.3 ml/min/100 g (mean ± SD). Evaluating gliomas only, the same parameter was 61.5 ± 32.3 ml/min/100g, whereas the CBF value for nongliomas was 90.2 ± 65.5 ml/min/100g. Most (15/19) of the patients with gliomas had ongoing treatment at the time of the MR scan. Corresponding measurements in the cerebellum and the normal-appearing GM and WM in the hemisphere contralateral to the tumor were 30.5 ± 8.86 ml/min/100g, 44.4 ± 8.44 ml/min/100g, and 14.9 ± 6.21 ml/min/100g, respectively. Regarding normalized perfusion values, meningiomas showed the highest median ASL nTBF, DSC nTBF, and DSC nTBV values. The Pearson correlation coefficient for ASL nTBF and DSC nTBF was 0.82 (R2 = 0.67). Mean ASL nTBF (2.46 ± 1.67) in this study was approximately 67% (range: 58–75%) of the corresponding DSC nTBF (3.60 ± 2.24). Differences between ASL nTBF and adjusted (67% of) DSC nTBF values were on average 0.05 ± 0.97 (mean ± SD), resulting in a coefficient of repeatability of 1.91.

Results: pCASL Versus DSC-MRI Visual Scoring of Tumor with Pronounced Signal Enhancement The visual evaluation of signal enhancement in ASL CBF and DSC -MRI maps (both color-coded and gray-scale maps of rCBF and rCBV) showed that the total visual score in tumor regions with pronounced

Visual Assessment of Susceptibility Artifacts A lower total susceptibility-artifact score was seen for ASL CBF than for DSC-MRI. Susceptibility artifacts were less prominent in ASL CBF (Fig. 14.2). ASL CBF had a total susceptibility-artifact score of 15, whereas DSC rCBF yielded a score of 29. The difference in score for number of susceptibility artifacts was statistically significant (p < 0.01).

Discussion Three dimensional FSE pCASL can be used for the evaluation of perfusion in brain tumors (Jarnum et al.,

140

2010; Lehmann et al., 2010). A brain tumor with a rapidly arriving blood flow presents an ideal situation for ASL measurements. Several studies have demonstrated the strengths of ASL in measuring brain tumor perfusion (Chawla et al., 2007; Kim and Kim, 2007). Arterial spin labeling improves the diagnostic accuracy because of absolute quantification of perfusion and fewer susceptibility artifacts. Spiral acquisition, and FSE instead of single-shot GRE EPI, explains the low number of susceptibility artifacts seen in ASL compared with DSC GRE EPI maps (Wang et al., 2002). The ASL technique has, however, its drawbacks. Our simple quality scoring on signal intensity revealed a common ASL problem: low perfusion signal. Using pCASL, the measured CBF was 44.4 ± 8.44 ml/min/100g in GM. This GM CBF value is in accordance with other ASL studies (Petersen et al., 2010) but lower than GM CBF estimates obtained by other perfusion modalities. With PET, a mean GM CBF of 55 ml/min/100 g is reported (Leenders et al., 1990). A review article on absolute quantification of CBF with the use of DSC-MRI reported GM CBF values from 52 to 137 ml/min/100g (Knutsson et al., 2010). The large range is due to differences in experimental procedures and postprocessing methods. Arterial spin labeling and DSC-MRI acquired with spin-echo EPI are most sensitive to perfusion at the microvascular level (Boxerman et al., 1995), whereas GRE EPI DSC-MRI favors large blood vessels, leading to a larger perfusion signal. The control scan (null pulse) in the original method for pCASL perturbs the equilibrium magnetization diminishing the ASL signal (Nezamzadeh et al., 2010). However, Nezamzadeh et al. (2010) have recently developed a new, modified pCASL sequence, mpCASL, with a 25% larger perfusion signal in vivo due to an improved preservation of the equilibrium magnetization in the control experiment. Another challenge with ASL is perfusion measurements in WM. Perfusion estimates described in the literature are not consistent with each other or with WM CBF assessed with other modalities. White matter CBF is substantially affected by the limited image resolution and by signal losses caused by the long transit times in WM, which results in more label decay (van Gelderen et al., 2008). In our study, CBF measured by pCASL was 14.9 ± 6.21 ml/min/100g in WM, which is in good agreement with PET results in humans that indicate a

H. Järnum et al.

WM CBF of 15–18 ml/min/100g (Owler et al., 2004). This excellent result can be explained by improved pCASL labeling schemes and background suppression, which have led to an increase in signal-to-noise ratios (van Osch et al., 2009). Another possible solution to overcome CBF underestimation with ASL in areas with long transit times may be to combine ASL with DSC-MRI. In a previous study, a patient-specific correction factor (the ratio of ASL-and DSC-CBF) was calculated in short-arrivaltime regions (Zaharchuk et al., 2010). This combined CBF method was compared to gold standard xenon CT. Combined ASL and DSC CBF obtained a higher correlation with xenon CT in voxels with long transit times compare to ASL alone (Zaharchuk et al., 2010). Nevertheless, ASL is most often compared with DSC-MRI. Several studies evaluate ASL compared with DSC-MRI measurement of brain tumor perfusion (Lehmann et al., 2010; Warmuth et al., 2003). Our Pearson correlation coefficient of 0.82 for ASL nTBF and DSC nTBF values is in good agreement with results obtained with a Spearman’s nonparametric correlation test in a previous study that compared pCASL with DSC-MRI for measurement of brain tumor perfusion (Lehmann et al., 2010). A recent study states that ASL imaging more accurately distinguishes recurrent high-grade glioma from radiation necrosis, because DSC-MRI may underestimate true blood volume in “leaky” regions due to severe blood-brain barrier breakdown (Ozsunar et al., 2010). The consensus is that ASL provides a quantitative, non-invasive alternative for assessment of microvascular perfusion and a distinction between high- and low-grade gliomas. Although both ASL and DSC-MRI are MRI perfusion modalities, they are based on two totally different techniques, and their signal behavior/perfusion value is expected to differ because DSC-MRI uses a gadolinium-based contrast agent with a larger molecular weight and different diffusion behavior than those of water molecules (Ye et al., 2000). Indeed, researchers have tried to compare pCASL with PASL; the result is higher SNR with pCASL (Jung et al., 2010). However, pCASL cannot be widely applied in a clinical setting if the reproducibility is poor. The ASL technique is well suited for longitudinal studies since ASL enables repetitive follow-ups. At our department, we performed a pCASL reproducibility study in 29 healthy volunteers, 15 males and 14 females, mean age

14 Brain Tumors

43, without a history of neurological disease. The volunteers were scanned twice at an interval of 30 min. No caffeine or nicotine was permitted in the break between two measurements, because these substances are believed to influence CBF (Addicott et al., 2009; Zubieta et al., 2005). Reproducibility was analyzed according to Bland and Altman (1999) for mean CBF values for two measurements of whole brain, giving a repeatability coefficient of 15%. In 95% of the cases, the two perfusion measurements did not differ more than 5.7 ml/min/100g. This reproducibility result compares favorably with PET (Matthew et al., 1993). Despite promising results with pCASL, researchers continue to develop the ASL technique. One modification of the original scheme is the ability to obtain arterial blood volume (aBV). Brookes et al. (2007) have made it possible to assess regional aBV fractions by combining Look-Locker echo-planar imaging with ASL. The Look-Locker approach allows the measurement of signals across several TI values following each label because images are acquired by using low flip angle excitations (Brookes et al., 2007). Arterial blood volume could possibly be used as an alternative to CBV measurements in tumors. Another notion that may be useful in the evaluation of brain tumor perfusion is the capability of selective labeling of the external carotid artery to derive information about the vascular supply to hypervascular extra-axial brain tumors (Sasao et al., 2010). In conclusion, 3D FSE pCASL can assess brain tumor perfusion, facilitates the diagnosis of brain tumors, and distinguishes high-grade from low-grade tumors. The sequence may become a useful tool for the longitudinal assessment of treatment response.

References Addicott MA, Yang LL, Peiffer AM, Burnett LR, Burdette JH, Chen MY, Hayasaka S, Kraft RA, Maldjian JA, Laurienti PJ (2009) The effect of daily caffeine use on cerebral blood flow: How much caffeine can we tolerate? Hum Brain Mapp 30:3102–3114 Alsop DC, Detre JA (1996) Reduced transit-time sensitivity in noninvasive magnetic resonance imaging of human cerebral blood flow. J Cereb Blood Flow Metab 16:1236–1249 Bland JM, Altman DG (1999) Measuring agreement in method comparison studies. Stat Methods Med Res 8:135–160 Boxerman JL, Hamberg LM, Rosen BR, Weisskoff RM (1995) MR contrast due to intravascular magnetic susceptibility perturbations. Magn Reson Med 34:555–566

141 Brookes MJ, Morris PG, Gowland PA, Francis ST (2007) Noninvasive measurement of arterial cerebral blood volume using Look-Locker EPI and arterial spin labeling. Magn Reson Med 58:41–54 Chawla S, Wang S, Wolf RL, Woo JH, Wang J, O’Rourke DM, Judy KD, Grady MS, Melhem ER, Poptani H (2007) Arterial spin-labeling and MR spectroscopy in the differentiation of gliomas. AJNR Am J Neuroradiol 28:1683–1689 Dai W, Garcia D, de Bazelaire C, Alsop DC (2008) Continuous flow-driven inversion for arterial spin labeling using pulsed radio frequency and gradient fields. Magn Reson Med 60:1488–1497 Garcia DM, Duhamel G, Alsop DC (2005) Efficiency of inversion pulses for background suppressed arterial spin labeling. Magn Reson Med 54:366–372 Herscovitch P, Raichle ME (1985) What is the correct value for the brain–blood partition coefficient for water? J Cereb Blood Flow Metab 5:65–69 Jarnum H, Steffensen EG, Knutsson L, Frund ET, Simonsen CW, Lundbye-Christensen S, Shankaranarayanan A, Alsop DC, Jensen FT, Larsson EM (2010) Perfusion MRI of brain tumours: a comparative study of pseudo-continuous arterial spin labelling and dynamic susceptibility contrast imaging. Neuroradiology 52:307–317 Jung Y, Wong EC, Liu TT (2010) Multiphase pseudocontinuous arterial spin labeling (MP-PCASL) for robust quantification of cerebral blood flow. Magn Reson Med 64:799–810 Kim HS, Kim SY (2007) A prospective study on the added value of pulsed arterial spin-labeling and apparent diffusion coefficients in the grading of gliomas. AJNR Am J Neuroradiol 28:1693–1699 Kleihues P, Soylemezoglu F, Schauble B, Scheithauer BW, Burger PC (1995) Histopathology, classification, and grading of gliomas. Glia 15:211–221 Knutsson L, Stahlberg F, Wirestam R (2010) Absolute quantification of perfusion using dynamic susceptibility contrast MRI: pitfalls and possibilities. MAGMA 23:1–21 Larsson HB, Courivaud F, Rostrup E, Hansen AE (2009) Measurement of brain perfusion, blood volume, and bloodbrain barrier permeability, using dynamic contrast-enhanced T(1)-weighted MRI at 3 tesla. Magn Reson Med 62: 1270–1281 Leenders KL, Perani D, Lammertsma AA, Heather JD, Buckingham P, Healy MJ, Gibbs JM, Wise RJ, Hatazawa J, Herold S (1990) Cerebral blood flow, blood volume and oxygen utilization. Normal values and effect of age. Brain 113(Pt 1):27–47 Lehmann P, Monet P, de Marco G, Saliou G, Perrin M, StoquartElsankari S, Bruniau A, Vallee JN (2010) A Comparative Study of Perfusion Measurement in Brain Tumours at 3 Tesla MR: Arterial Spin Labeling versus Dynamic Susceptibility Contrast-Enhanced MRI. Eur Neurol 64:21–26 Lu H, Clingman C, Golay X, van Zijl PC (2004) Determining the longitudinal relaxation time (T1) of blood at 3.0 Tesla. Magn Reson Med 52:679–682 Matthew E, Andreason P, Carson RE, Herscovitch P, Pettigrew K, Cohen R, King C, Johanson CE, Paul SM (1993) Reproducibility of resting cerebral blood flow measurements with H2(15)O positron emission tomography in humans. J Cereb Blood Flow Metab 13:748–754

142 Nezamzadeh M, Matson GB, Young K, Weiner MW, Schuff N (2010) Improved pseudo-continuous arterial spin labeling for mapping brain perfusion. J Magn Reson Imaging 31: 1419–1427 Ohgaki H, Kleihues P (2005) Population-based studies on incidence, survival rates, and genetic alterations in astrocytic and oligodendroglial gliomas. J Neuropathol Exp Neurol 64:479–489 Owler BK, Momjian S, Czosnyka Z, Czosnyka M, Pena A, Harris NG, Smielewski P, Fryer T, Donovan T, Coles J, Carpenter A, Pickard JD (2004) Normal pressure hydrocephalus and cerebral blood flow: a PET study of baseline values. J Cereb Blood Flow Metab 24:17–23 Ozsunar Y, Mullins ME, Kwong K, Hochberg FH, Ament C, Schaefer PW, Gonzalez RG, Lev MH (2010) Glioma recurrence versus radiation necrosis? A pilot comparison of arterial spin-labeled, dynamic susceptibility contrast enhanced MRI, and FDG-PET imaging. Acad Radiol 17:282–290 Petersen ET, Mouridsen K, Golay X and all named co-authors of the QUASAR test-retest study (2010) The QUASAR reproducibility study, part II: results from a multi-center arterial spin labeling test-retest study. Neuroimage 49:104–113 Petersen ET, Zimine I, Ho YC, Golay X (2006) Non-invasive measurement of perfusion: a critical review of arterial spin labelling techniques. Br J Radiol 79:688–701 Rahman R, Smith S, Rahman C, Grundy R (2010) Antiangiogenic therapy and mechanisms of tumor resistance in malignant glioma. J Oncol 2010:251231 Sasao A, Hirai T, Nishimura S, Fukuoka H, Murakami R, Kitajima M, Okuda T, Akter M, Morioka M, Yano S, Nakamura H, Makino K, Kuratsu JI, Awai K, Yamashita Y (2010) Assessment of vascular supply of hypervascular extra-axial brain tumors with 3T MR regional perfusion imaging. AJNR Am J Neuroradiol 31:554–558

H. Järnum et al. van Gelderen P, de Zwart JA, Duyn JH (2008) Pittfalls of MRI measurement of white matter perfusion based on arterial spin labeling. Magn Reson Med 59:788–795 van Osch MJ, Teeuwisse WM, van Walderveen MA, Hendrikse J, Kies DA, van Buchem MA (2009) Can arterial spin labeling detect white matter perfusion signal? Magn Reson Med 62:165–173 Wang J, Alsop DC, Li L, Listerud J, Gonzalez-At JB, Schnall MD, Detre JA (2002) Comparison of quantitative perfusion imaging using arterial spin labeling at 1.5 and 4.0 Tesla. Magn Reson Med 48:242–254 Wang J, Zhang Y, Wolf RL, Roc AC, Alsop DC, Detre JA (2005) Amplitude-modulated continuous arterial spinlabeling 3.0-T perfusion MR imaging with a single coil: feasibility study. Radiology 235:218–228 Warmuth C, Gunther M, Zimmer C (2003) Quantification of blood flow in brain tumors: comparison of arterial spin labeling and dynamic susceptibility-weighted contrast-enhanced MR imaging. Radiology 228:523–532 Ye FQ, Berman KF, Ellmore T, Esposito G, van Horn JD, Yang Y, Duyn J, Smith AM, Frank JA, Weinberger DR, McLaughlin AC (2000) H(2)(15)O PET validation of steadystate arterial spin tagging cerebral blood flow measurements in humans. Magn Reson Med 44:450–456 Zaharchuk G, Straka M, Marks MP, Albers GW, Moseley ME, Bammer R (2010) Combined arterial spin label and dynamic susceptibility contrast measurement of cerebral blood flow. Magn Reson Med 63:1548–1556 Zubieta JK, Heitzeg MM, Xu Y, Koeppe RA, Ni L, Guthrie S, Domino EF (2005) Regional cerebral blood flow responses to smoking in tobacco smokers after overnight abstinence. Am J Psychiatry 162:567–577

Chapter 15

Cerebral Cavernous Malformations: Advanced Magnetic Resonance Imaging Robert Shenkar, Sameer A. Ansari, and Issam A. Awad

Abstract Cerebral vascular malformations (CCMs) are highly prevalent vascular lesions that can be detected by magnetic resonance imaging (MRI). MRI is the primary method for diagnosis of CCMs in patients. Conventional MRI sequences can detect features such as hemosiderin and ferritin deposit and can categorize CCMs into four types according to their respective T1 and T2 intensities. Contrast enhanced imaging, including use of gadolinium, is used for differential diagnosis. By increasing magnetic susceptibility, gradient recalled echo sequences can reveal otherwise undetectable lesions. Susceptibility weighted imaging (SWI) can depict CCMs with higher resolution. High field MRI can either shorten imaging time or reveal the microvascular anatomy of the CCM lesions. Experimental MRI of excised human CCMs show structural details such as “bland” and “honeycombed” regions, including the presence of smaller caverns along blood vessels or surrounding larger caverns. Experimental MRI of murine models for CCM may differentiate between simple single cavern containing pre-lesions and complex multi-cavernous mature CCM lesions that can lead to studies on the pathogenesis of these lesions. Future venues for advanced MRI include depicting additional feature of CCM lesions with SWI and higher field strength imaging and the use of contrast agents conjugated to antibodies against markers for endothelium and inflammatory cells in

I.A. Awad () Neurovascular Surgery Program, Section of Neurosurgery, University of Chicago Pritzker School of Medicine, 5841 S. Maryland Ave., Chicago, IL 60637, USA e-mail: [email protected]

order to reveal aggressive and quiescent states and other clinical behavior. Keywords CCMs · MRI · SWI · Proton · Murine model · Brain

Introduction Cerebral cavernous malformations (CCMs) are present in ∼0.5% of the population in the United States, as reviewed by Leblanc et al. (2009). The lesions consist of clusters of grossly dilated vascular spaces (“caverns”), lined by endothelium and filled with blood at various stages of thrombosis. The caverns lack mature vessel wall elements, and are often contiguous or separated by amorphous collagenous matrix or intervening brain parenchyma. As reviewed by Leblanc et al. (2009), the lesions and surrounding brain tissue include blood breakdown products and a robust inflammatory cell infiltration. Lesions may grow or bleed causing neurologic deficits or seizures. It has been reported by Zabramski et al. (1994) and reviewed by Maraire and Awad (1995) that the annual risk of hemorrhage ranges from 0.7 to 1.1% per lesion per year. Typically, patients with single lesions have a sporadic form of the disease, while those with multiple lesions (10–31% of all cases) often have an autosomal dominant form localizable to one of three known gene loci with respective encoded proteins (CCM1/KRIT1, CCM2/MGC4067, CCM3/PDCD10) (Leblanc et al., 2009). The hallmark of familial CCM, reported by Labauge et al. (1998) is the presence of multifocal lesions, in a volume distribution throughout the brain, with the appearance of new lesions over time.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_15, © Springer Science+Business Media B.V. 2011

143

144

The sporadic form of CCM is often characterized by a solitary lesion (or a cluster of lesions) in association with a developmental venous anomaly (DVA) (Abdulrauf et al., 1999; Gault et al., 2006). Prior to the widespread use of magnetic resonance imaging (MRI), CCMs were thought to be rare entities. Modern MRI sequences are highly sensitive for detecting CCMs as well as associated hemorrhage at various stages of thrombosis and reorganization. In this chapter, the authors describe the use of emerging imaging techniques for the diagnosis and treatment of this entity, and imaging applications in ongoing research on CCM.

R. Shenkar et al.

absence of calcification or hemorrhage, CCMs may be undetectable especially with respect to small lesions (<1 cm). High resolution MRI is the primary diagnostic modality in the evaluation of CCMs providing both high sensitivity and specificity that approaches 100% as described by Shenkar et al. (2008a). It was reported by Maraire and Awad (1995) that MRI allows the confident preoperative evaluation of symptomatic lesions as well as the identification, screening, and follow-up of incidental lesions to assess expansion, thrombosis, and acute/subacute hemorrhage.

Clinical Imaging of CCM

Imaging Features on Conventional MRI Sequences

Until the advent of MRI, evaluation of CCMs had been limited to diagnostic angiography and computerized tomography (CT). Described as being “angiographically occult”, CCMs inconsistently show a subtle capillary blush, venous pooling, or avascular mass effect on angiography. However, it was reported by Rigamonti et al. (1987) and Rivera et al. (2003), that due to the low sensitivity (30–60%) and invasiveness of the technique, conventional angiography has no current role in the diagnosis of CCMs, except for excluding arteriovenous shunting in atypical lesions. In contrast, according to Maraire and Awad (1995), CT scanning provides excellent diagnostic sensitivity ranging from 70 to 100%, but its low specificity (<50%) can be problematic for an accurate diagnosis. On noncontrast CT, CCMs are characterized as hyperattenuating nodular lesions secondary to calcification, intravascular slow flow, thrombosis and/or hemorrhage. Limited mass effect is seen unless associated with acute hemorrhage. It is important to note that differentiating thrombosis from hemorrhage in its various stages of evolution cannot be assessed by CT. Furthermore, it was reported by Rivera et al. (2003) the indistinct margins of CCMs with internal stippled or coarse calcifications are nonspecific CT findings requiring a differential diagnosis such as low grade calcified neoplasms, infectious or inflammatory granulomatous lesions, intracranial hemorrhage, or high flow arteriovenous malformations. Faint contrast enhancement has been described by Batra et al. (2009) on post-infusion CT scanning, but in the

The characteristic MRI appearance of CCMs is a well defined, lobulated lesion with a reticular core corresponding to the dilated, endothelial lined vascular channels. Utilizing standard MRI techniques, specifically T1 and T2 weighted imaging, CCMs can exhibit heterogeneous signal intensity due to variable hemorrhagic products in the acute/subacute (oxyhemoglobin, deoxyhemoglobin, intracellular/extracellular methemoglobin) or chronic (hemosiderin/ferritin) stages. Therefore, it is important to be familiar with the T1 - and T2 -weighted imaging characteristics of hemorrhage to classify and understand these lesions (Fig. 15.1). Conversely, more predictable MRI signal hypointensity is observed on both sequences corresponding to the chronic thrombosis, fibrosis, calcification, and adjacent gliosis associated with CCMs. A classic and typical feature of CCMs is the hemosiderin rim of T2 hypointensity or “hemosiderin ring”, secondary to repeated subclinical hemorrhages resulting in peripheral hemosiderin and ferritin deposition. Although this finding is nearly pathognomonic for CCMs, the hemosiderin rim may not be evident in intraventricular lesions that can expand rapidly and mimic a neoplasm rather than a vascular malformation according to Rivera et al. (2003). In the setting of acute/subacute hemorrhage or thrombosis, perilesional hemorrhage, edema, and mass effect may obscure the hemosiderin ring. It was reported by Maraire and Awad (1995) and Rivera et al. (2003) that an internal fluid/fluid level can be associated with an active CCM or acute/subacute hemorrhagic products, in contrast to

15 Cerebral Cavernous Malformations

145

Fig. 15.1 Subtle changes in appearance of solitary CCM with different MRI sequences, reflecting differential sensitivity of blood breakdown products at different ages, and low flow in

dilated cavernous channels. Reproduced with permission from Campbell et al. (2010)

chronic cysts that represent the sequelae of resolved hemorrhage and thrombus in previously expanded caverns. Zabramski et al. (1994) presented a classification of four types of CCM lesions based on MR signal characteristics correlating it with surgical pathology from a population of familial CCMs. Type 1 lesions are characterized by a T1 hyperintense and T2 hyperintense or a T2 hypointense core (depending on the state of methemoglobin) consistent with subacute hemorrhage. A T2 hypointense hemosiderin rim surrounds the margins of the lesion representing chronic hemorrhagic products, staining the peripheral macrophages and gliotic brain. Type 2 lesions exhibit mixed T1 and T2 signal intensity of a reticulated core correlating with loculated areas of hemorrhage and thrombosis of varying ages, the typical “popcorn” appearance. Internal calcifications in larger lesions may represent

components of the hypointense reticulated core, but the T2 hypointense hemosiderin rim is maintained in the peripheral gliotic tissue. Type 3 lesions demonstrate a T1 iso- or hypointense and T2 hypointense core as well as a T2 hypointense rim compatible with chronic resolved hemorrhage or hemosiderin within and surrounding the lesion. Type 4 lesions may be occult on T1 -weighted imaging, but are best depicted as small, punctuate hypointense foci on T2 or T2 ∗ gradient recalled echo (GRE)-weighted imaging highly sensitive for paramagnetic susceptibility effects from even minimal deposits of hemosiderin. Pathologically, type 4 lesions may represent capillary telangiectasias or early stage CCMs seen frequently in the familial form, as reported by Maraire and Awad (1995), Rivera et al. (2003), and Zabramski et al. (1994). Because type 1 and type 2 lesions are indicative of subacute hemorrhagic products, it was inferred by

146

Zabramski et al. (1994) that they would be more likely to bleed or be symptomatic as suggested in earlier reports. However, according to Rivera et al. (2003) there has been no significant correlation between the types of lesions or the natural history of the disease. Furthermore, there is no evidence of type 1 and 2 lesions evolving into type 3 lesions as would be expected with the gradual organization of hemorrhage within the CCM, suggesting recurrent and intermittent bleeding persists.

The Role of Contrast Enhanced MRI Variable or faint contrast enhancement of CCMs has been described on T1 post-gadolinium weighted sequences, but does not appear to correlate with size or histopathological findings as reported by Pinker et al. (2006). Gadolinium-enhanced studies may be more valuable in delineating associated developmental venous anomalies (∼30%) or capillary telangiectasias than the primary CCM lesions. To our knowledge, there are no studies that have investigated delayed phase contrast enhancement in these slow flow CCMs, analogous to extra-axial or peripheral cavernous hemangiomas. Contrast enhanced imaging is particularly useful in the diagnostic evaluation of CCM, and in clarifying

Fig. 15.2 Multiple MRI sequences in patient presenting with temporal lobe seizures. The T2 -weighted sequence (left) illustrates subtle abnormality in the left posterior mesiotemporal region, consistent with non-specific hemosiderin deposition. Gadolinium enhanced T1 -weighted image (center) delineates

R. Shenkar et al.

differential diagnosis. According to Abdulrauf et al. (1999) and Gault et al. (2006), the presence of an overt associated DVA is more likely to define the nongenetic non-familial form of the disease (Fig. 15.2). Also, the presence of an associated DVA may influence surgical decisions, especially with regard to surgical maneuvers aimed at avoiding injury to the DVA and consequences of venous ischemia. The concept of lesion cure with surgical resection must be tempered, when resecting a solitary CCM but leaving behind an overt DVA (which could later contribute to CCM recurrence). Contract enhancement my delineate patterns of overt enhancement consistent with other pathology than CCM, particularly tumors (homogeneous enhancement), or arteriovenous malformation (serpiginous enhancement). Finally, punctate enhancement in association with CCM’s on gadolinium enhanced T1 weighted images, without hemosiderin “blooming” on T2 ∗ -GRE-weighted sequences, has been suggested by Pozzati et al. (2007) and Abla et al. (2008) to represent capillary telangiectasia, most commonly reported in the pons, and in the bed of DVAs.

Gradient Echo Sequences Labauge et al. (1998) reported the higher sensitivity for detecting CCM lesions with GRE pulse sequences

a prominent venous structure with “caput medusae” pattern, associated with the T2 -weighted signal, likely suggesting an associated DVA. The T2 ∗ -GRE-weighted images reveal much better delineation of multiple foci of CCM. Reproduced with permission from Campbell et al. (2010)

15 Cerebral Cavernous Malformations

or T2 ∗ -weighted imaging in comparison to standard T1 - and T2 weighted sequences. GRE-weighted sequences are obtained by omitting the 180 degree refocusing pulse rendering it highly sensitive to magnetic susceptibility or the field inhomogeneities caused by the products of hemoglobin degradation (hemosiderin or ferritin) and calcification. In fact, it was reported by Brunereau et al. (2001) and Labauge et al. (1998) that multiple type 4 hypointense CCM lesions were documented on GRE-weighted sequences in familial CCM patients, which were undetectable or inconspicuous on conventional spin-echo sequences. Also, GRE-weighted sequences may better delineate multiple CCMs associated with DVA (Fig. 15.2).

Susceptibility Weighted Imaging High resolution susceptibility weighted imaging (SWI) is a three dimensional gradient echo technique based on blood oxygen level dependent (BOLD) induced phase effects. Initially described by Lee et al. (1999), it provides submillimeter resolution of venous anatomy capitalizing on the signal loss in venous (deoxygenated) blood due to phase and T2 -weighted changes. The technique is extremely powerful in detecting the venous vasculature, blood products, and vascular malformations analogous to GRE imaging. However, SWI

Fig. 15.3 Case with known familial CCM disease, presents for routine MRI screening. The T2 -weighted sequences (right) reveal two suspected CCM lesions, better delineated on T2 ∗ -GRE-weighted sequences (center), the latter also

147

provides increased resolution better depicting the margins of peripheral CCMs and with less “blooming” artifact. Recently, SWI has been shown by de Souza et al. (2008) to significantly enhance the sensitivity of detecting CCMs in patients with the familial form of the disease when compared to T2 -weighted fast spin echo or even GRE-weighted sequences, with approximately 40% lesions seen exclusively by SWI. The criteria for detecting type 4 CCM lesions should include SWI as part of the imaging protocol for the most accurate assessment, superseding the previous standard of GRE-weighted sequences. In the largest study to date, including 23 cases performed by our group and colleagues in Montpellier, France, confirmed nearly twice the number of lesions detected by SWI as compared to T2 ∗ -GRE sequences, but only in the 14 familial cases (Fig. 15.3). In none of the 9 cases with solitary CCM or clustered lesions in the bed of a DVA (Fig. 15.4) did the SWI contribute additional lesions than those noted on T2 ∗ /GRE (Menjot et al., manuscript in preparation). Hence, SWI seems to increase the sensitivity of lesion detection in familial multifocal CCM lesions, it does not per se appear to reveal lesion multiplicity that had not been already demonstrated by T2 ∗ -GRE (Figs. 15.3 and 15.4). The SWI images are highly sensitive to delineation of associated venous anomalies, and possibly telangiectasiae, without requiring gadolinium contrast enhancement, and this may be a significant advantage

suggesting perhaps one or two additional subtle lesions. The SWI images reveal many additional lesions throughout the brain. Reproduced with permission from Campbell et al. (2010)

148

R. Shenkar et al.

Fig. 15.4 Case with solitary sporadic CCM discovered incidentally in the workup of unrelated neoplasm. The T1 -weighted contrast enhanced images reveal a small suspected CCM in the right frontal cortex (left), and subtle abnormal venous prominence in superior and medial to the lesion (not shown).

The T2 ∗ -GRE weighting better delineates the same lesion (center). SWI sequences (right) reveal no additional lesions, although they also demonstrate the suspected venous anomaly. Reproduced with permission from Campbell et al. (2010)

in pregnant patients and in cases with nephrotoxicity or gadolinium allergy. While the SWI technique is not yet widely available it is possible that, given the early clinical data suggesting its enhanced sensitivity in familial cases, it should be included in the routine imaging assessment of vascular malformations. These sequences might provide endophenotypic markers of disease burden in familial CCM that should be correlated with disease penetrance and aggressiveness in different individuals and kindreds, and with the response to potential therapeutic interventions. The ultimate applicability of SWI is limited by several factors. Firstly, as with T2 ∗ -GRE, it is difficult to differentiate small venous structures from small hemorrhages and thrombosis. However, sequential SWI imaging before and after gadolinium administration, could ameliorate this deficiency. As pointed out above, the higher sensitivity of SWI sequences may not apply to sporadic or solitary CCMs, or CCM clusters associated with DVA. While SWI has shown greater ability to identify lesions in familial CCM, the need to apply this imaging modality to sporadic CCMs has yet to be demonstrated. The authors are not aware of a case whereby a solitary lesion was detected on T2 ∗ -GRE that was later found to be associated with occult lesion multiplicity on the more sensitive

SWI MRI (Fig. 15.4). As such, future studies should specifically address SWI sensitivity in cases of sporadic CCM, those associated with DVA and radiation induced CCMs. Finally, it is unclear what is the nature of the additional lesions delineated on SWI and occult on T2 ∗ -GRE (Fig. 15.3). Some may be better resolved in the 3D sequence acquisition of SWI, while they may have been attenuated by “volume averaging” in the typically two dimensional acquisition of T2 ∗ -GREweighted images. According to Awad et al. (1993), the occult punctate lesions may also represent nonhemorrhagic capillary telangiectasiae, often reported in the same case as CCM, and which could represent precursors of more mature CCMs.

Clinical Imaging with High Field MRI Although the diagnostic power of MRI with GREweighted and SWI sequences allows identification of small CCMs, the spatial resolution on lower field strengths (<3 T) limits evaluation of their angioarchitecture and early developing lesions. It is at this time when CCMs may be most susceptible to medical treatment and prevention of clinical sequelae such as

15 Cerebral Cavernous Malformations

a gliotic response or hemorrhage. High field MRI increases the signal to noise ratio and the phase difference between blood and the surrounding tissues for greater susceptibility effects, also providing the ability to shorten echo times (TE) and repetition times (TR). As a result, as reported by Pinker et al. (2007), optimal high resolution 3T SWI can be performed with shorter scan times to increase the sensitivity of the technique and reveal the microvascular anatomy of CCMs such as intralesional tubular structures. Ultra high-field MRI at 7 T has been shown by Schlamann et al. (2010) to detect new CCMs not previously visualized on 1.5 T GRE-weighted sequences. Ex vivo analysis of human CCM tissue specimens on 9.4 T and 14.1 T micro-scanners reported by Shenkar et al. (2008a) demonstrated the structural details of smaller caverns surrounding larger caverns that were organized on tubular structure resembling a mature blood vessel. The MR microscopy results correlated with the histopathological findings using confocal microscopy, confirming the angioarchitecture of CCMs.

Experimental MRI of CCM: Excised Human Lesions In order to show that high field imaging with greater spatial resolution can reveal more details of the structure of CCMs than those seen in clinical images at lower field strength, CCM specimen were surgically resected from four patients, diagnosed previously by clinical MRI at 1.5 Tesla (T). The subjects (one male and three females) were 23–59 years of age. Two subjects had recent hemorrhages from the CCM within 3 months before surgery; two subjects had clinically quiescent lesions. One subject harbored multiple lesions; three subjects had single lesions. Three subjects experienced seizure disorders. All CCM lesions, ranging in size from 10 to 50 mm, were located supratentorially. After surgical removal of the CCMs, lesions were rinsed in saline and immersed in 2% paraformaldehyde for 2–4 weeks and rinsed with phosphate buffered saline (PBS) immediately before MRI. High spatial resolution MRI was acquired on a 9.4 T Bruker vertical axis imager (400 MHz proton frequency) or a 14.1 T Bruker Avance imaging spectrometer (600 MHz proton frequency), using a

149

20-mm volume resonator tuned to the appropriate proton frequency. Two-dimensional multi-slice spin-echo proton density-weighted images were acquired using repetition time/echo time (TR/TE): 2,000 ms/10 ms. Three-dimensional gradient-recalled echo images were acquired using TR/TE 50–200 ms/4 ms and 60–90 μm isotropic pixel size. Proton density weighted images (Fig. 15.5, center) acquired at high field strength, with high spatial resolution reveal the caverns comprising the lesion, reflecting histopathologic CCM appearance, as well as a markedly heterogeneous microstructure within the specimen. Large caverns, greater than 500 μm in diameter and present at various locations within the lesion, represent the most striking feature of the CCM image. Besides these caverns, numerous smaller hypointense spots of varying sizes less than 100 μm in diameter are seen throughout the landscape (arrows). High field three dimensional gradient-recalled echo T2 ∗ -weighted MRI obtained at 14.1 T of CCM lesions revealed both “bland” and “honeycombed” regions illustrated in Fig. 15.5 (left). We define “bland” regions as hyperintense areas with larger caverns (0.5 – 1 mm) devoid of other details, and “honeycombed” regions as hypointense areas with smaller caverns or capillaries [<100 μm in diameter (arrows)] surrounding larger caverns (Fig. 15.5, left) and along apparent blood vessels (Fig. 15.5, center). After MRI, the tissue was prepared for confocal microscopy by rinsing in PBS, immersing into 2% paraformaldehyde, with crosslinks were reversed in 0.01 mol/L citrate buffer, pH 6.0, 95◦ C for 25 min before it was cut into 200-μm sections with a Vibratome Model G sectioning system (Oxford Laboratories, Foster, CA). Sections were placed on microscope slides, blocked with goat immunoglobulin G antibody (Santa Cruz Biotechnologies, Santa Cruz, CA) at 1:20 dilution for 1 h, treated with Image-iT FX signal enhancer (Molecular Probes, Eugene, OR). To stain endothelium, the specimen was incubated overnight with streptavidin-alexa-568 (Molecular Probes) at 1 μg/μl or with rabbit antihuman von Willebrand factor (Sigma-Aldrich, St. Louis, MO) at 1:300 dilution followed by incubation for 1 h with goat anti-rabbit-Alexa-568 (Molecular Probes) at a 1:300 dilution. It was then incubated with 70% ethanol for 2 min, treated with 2% Sudan Black-B in 70% ethanol for 3 min, incubated with

150

R. Shenkar et al.

Reconstituted 3D T2*-weighted MRI scan from a stack of 2D images acquired at 14.1 T (field of view: 15 x 15 x 13 mm, matrix size: 256 x 256 x 200, scale bars: 0.5 mm).

Caverns < 100 µm (arrows) along apparent blood vessels Protein density MRI scan acquired at 9.4 T (field of view:12 mm, matrix size: 256 x 256)

Nine regions of interest (green “roi”) for T1 and T2 relaxation time measurements on a proton density-weighted spin echo image acquired at 14.1 T (Slice thickness: 500 µm, field of view: 18 x 18 mm, matrix size: 256 x 256, in-plane pixel size: 705 µm)

Confocal microscopy images stained for von Willebrand factor (red) show caverns (arrows). Scale bars: 100 µm.

Confocal microscopy image

Diagrams indicating two region types

BLAND

1 mm

HONEYCOMBED

Fig. 15.5 Experimental imaging of CCM lesions resected from human patients. MRI and confocal microscopy images show “bland” and “honeycombed” regions (left) and small caverns (<100 μm) along apparent blood vessels (center) in human

CCMs. T1 and T2 relaxation times were measured for nine regions of interest in a human CCM (right). Reproduced with permission from Shenkar et al. (2008a)

70% ethanol for 1 min and rinsed twice in deionized water for 3 min. Coverslips were mounted with Prolong Gold antifade reagent (Molecular Probes). A D-eclipse C1 confocal microscope (Nikon Instruments Inc., Melville, NY) was used to view the sections using a 543 nm helium neon laser. Confocal microscopy confirmed the presence of similar features as identified by reconstructed high resolution MR imaging at 9.4 T or 14.1 T (Fig. 15.5, center and left). T1 and T2 relaxation times were measured simultaneously in each pixel of an image from one slice of a human CCM using progressive saturation reported by Freeman and Hill (1971) and Carr-Purcell-MeiboomGill (CPMG) described by Carr and Purcell (1954) and

Meiboom and Gill (1958), respectively. T2 ∗ was measured in a human CCM specimen acquired at 14.1 T by varying echo delay time at constant recycle time in a gradient-recalled echo imaging sequence. We measured relaxation times were measured in a CCM specimen from one case acquired at 14.1 T demonstrating regional differences in relaxation times within the specimen. Regional relaxation times were calculated by averaging T1 or T2 for all the pixels in the regions of interest (roi) shown in Fig. 15.5 (right). Two distinct regions are characterized in this specimen, one by T1 relaxation time >1 s and the other <1 s. Notably, all four “honeycombed” areas (roi 1–4) have larger T1 and T2 relaxation times than the five “bland” regions (roi

15 Cerebral Cavernous Malformations

5–9). The average T1 of the roi with the larger T1 is 1.55 s. The average T1 of the roi with the shorter T1 is 0.40 s. The larger T1 roi have an average T2 of 29 msec, while the shorter T1 roi have a shorter T2 average of 14.7 ms. T2 ∗ relaxation times varied between 8.0 and 19.1 ms in different regions of the sample. Large standard deviations of T2 ∗ relaxation times were measured in regions with “honeycombed” appearance, limiting the accuracy of those measurements in those regions. The observed regional differences in T1 , T2 and T2 ∗ relaxation times suggest the presence of iron in different oxidation states, in different breakdown products as noted by Bradley (1993), and perhaps in different amounts in the angioarchitectural features seen in human CCM.

Experimental MRI of CCM: Murine Models The detection and MR features of CCM lesions in the murine models, and present techniques of lesion localization and pathobiologic sampling are described. Double mutant mice and littermate controls were bred and genotyped at Duke University according to procedures previously published by Plummer et al. (2004, 2006). In this study we imaged eighteen animals, including Ccm1+/– Trp53–/– (six males, four females) and Ccm2+/– Trp53–/– mice (six males, two females) and 28 controls, including Ccm2+/– Trp53+/+ (four males, two females), Ccm1+/– Trp53+/+ , (nine males, five females), Ccm1+/+ Trp53–/– (four males, one female) and three male C57BL6 mice. Mice, shipped to Evanston Hospital at 1–3 months of age, were quarantined for 6 weeks before MRI. The Ccm+/– Trp53+/+ controls were imaged in vivo at age 10–19 months. Previous work by Plummer et al. (2004) revealed that these mice do not develop CCM lesions at an earlier age. The advanced age was chosen for imaging the brains of these mice to maximize the potential detection of occult CCM lesions by imaging, since more lesions and larger lesions are expected with time. The brains of the other control and all double mutant mice were subjected to in vivo MRI at age 3–7 months, because of the tendency of these mice to die from systemic tumors by 6–7 months of age.

151

Animals were anesthetized for imaging by intraperitoneal injection of Nembutal Sodium solution (Abbott Laboratories Abbott Park, IL) at 70 mg/kg, followed by inhalation of isoflurane (2% in oxygen/air) via nose cone. Eyes were treated with erythromycin ointment (Fougera, Melville, NY) to prevent dryness. Mice were allowed to recover spontaneously from anesthesia upon completion of imaging. Brains were removed from mice that died naturally or upon planned euthanasia by an overdose of Nembutal after in vivo imaging. Brains were rinsed in phosphate buffered saline (PBS) to remove excess blood, and were placed in 2% paraformaldehyde for 2–4 weeks before ex vivo MRI. Immediately prior to MRI, brains were rinsed in PBS and placed securely into 12 mm vials. Fomblin Y LVAC 06/6 perfluoropolyether (Sigma-Aldrich, St. Louis, MO) surrounded each brain sample to prevent dehydration and reduce magnetic susceptibility differences. MRI was conducted in vivo or ex vivo by using a 4.7 T Bruker Biospec (200 MHz) and/or a 14.1 T Bruker Avance (600 MHz) imaging spectrometer, with 25 mm birdcage coils. Three dimensional-GRE images were acquired in vivo with the 4.7 T spectrometer using TR/TE 100 ms/10 ms, with slice thickness 250 μm and in-plane pixel size 113 μm. Three dimensional-GRE images of brains were acquired ex vivo with the 14.1 T spectrometer using TR/TE 25 ms/7.8 ms, with slice thickness 117 μm and in-plane pixel size 27 μm. Lesions were identified in three planes on in vivo and ex vivo images, and excluded if they appeared to communicate with a brain slit (open trauma) in ex vivo images. MRI was conducted in both control and mutant animals, and lesion detection was adjudicated (in control and mutant animals) by review of all serial images of the whole respective brains by an MRI investigator not familiar with CCM biology, and blinded to animal genotype. Suspected CCM lesions were present in 83% (5 out of 6) of the Ccm1+/– Trp53–/– and 62% (5 out of 8) of the Ccm2+/– Trp53–/– mutant mice when subjected to in vivo MRI. These lesions appear as dark spots in GRE images (Fig. 15.6, right). H&E staining confirmed the histopathology of CCM lesions (N = 3 lesions, one lesion in each of three brains) as shown in Fig. 15.6 (right). In contrast, no lesions were found in the brains of 23 control mice similarly imaged in vivo. All lesions found in the in vivo images (N = 3 lesions,

152

R. Shenkar et al.

Staging of lesions Hematoxylin and eosin (H&E)-stained brain sections

Artist Description

Normal Capilaries

Dilated Capilaries (Stage 1 CCM)

Multiple Caverns (Stage 2 CCM)

MRI of CCMs (arrows) within mouse brains. In vivo T2*-weighted at 4.7 T (Bar = 1 mm)

3D reconstruction from ex vivo below (Bar: 0.5 mm)

T2*-weighted acquired ex vivo at 14.1 T (Bar: 1 mm)

Enlarged (Bar: 0.5 mm)

H&E –stained (Bar: 1 mm)

Enlarged (Bar: 0.5 mm)

Fig. 15.6 Experimental imaging of CCM lesions within brains of murine models. MRI and hematoxylin and eosin images are shown for a CCM in a murine model (right), with staging

of murine lesions indicated (left). Reproduced with permission from Shenkar et al. (2008b)

one lesion in each of three brains) were also seen in the ex vivo images. Lesions were present in the brains of all eight Ccm1+/– Trp53–/– mice imaged ex vivo (N = 1, 1, 3, 3, 5, 12, 14, 36 lesions/brain). In comparison, no lesions were found in any of the eight control brains that were removed from three Ccm1+/– Trp53+/+ , two Ccm1+/+ Trp53–/– and three C57BL6 mice, and imaged ex vivo using same MRI technique. In the core of one CCM lesion (Fig. 15.6, right), there appeared several dark foci interspersed in heterogeneous gray intensity. The core is circumscribed by a thin dark rim which is surrounded by a diffuse, bright outer rim. Smaller dark lesions appear to be budding from the thin dark rim. The dark spots and dark rim are most likely from iron in hemosiderin and other blood breakdown products as described by Bradley (1993). According to this author, the enhanced signal intensity of the bright outer rim indicates the presence of iron as methemoglobin, a T1 -shortening

compound and regions of intermediate signal intensity may represent varying levels of iron in different oxidation states. After the suspected CCM lesions were identified and located by MRI, the location and maximal diameter of the lesion were plotted on stereotactic grids as described by Shenkar et al. (2008b). Lesions were located within 1-mm coronal sections starting with the most caudal section at the cerebellar hindbrain and ending with the olfactory bulbs at the frontal rostrum, and within one of four quadrants in the coronal slice of interest. Brains were removed from the vials, rinsed in PBS, immersed in neutral buffered formalin, cut into 1-mm thick coronal slices and embedded in paraffin. The 1-mm thick slices suspected of containing lesions by MRI were sectioned at 5 μm with a microtome, and the thin coronal sections were whole mounted either onto charged microscope slides for staining with hematoxylin and eosin (H&E). Figure 15.6 (right)

15 Cerebral Cavernous Malformations

illustrates case where a stained coronal section from a Ccm1+/– Trp53–/– mouse brain reveals the presence of CCM lesions (consisting of blood filled dilated caverns, surrounded by a single layer of endothelial cells). There are smaller lesions near the larger lesion, and apparent budding of smaller cavern form a larger one. No true CCM lesions were detected on histology that was missed on MRI. Elsewhere in the brains of these same animals, there were areas of dilated capillaries indicating potential pre-lesional sites, many unsuspected on MRI (not shown).

Future Directions: Opportunities and Challenges Prior to the advent of MRI, evaluation of CCMs was limited to diagnostic angiography and computerized tomography. Presently, MRI is the best imaging method to evaluate CCMs, with GRE sequences being described as the “gold standard”. As the use of more advanced imaging techniques continues to achieve widespread distribution, high-field MRI and SWI are likely to become commonplace for the diagnosis and follow-up of these lesions. The nature of additional lesions revealed by SWI and other advanced sequences is not known. They may represent dilated capillaries (capillary telangiectasia), long recognized to be associated with CCM disease, or could indeed be revealing primordial lesions that could develop into full-blown clinically relevant CCM lesions. The SWI and other sequences may represent a more sensitive marker of “disease burden”, when characterizing penetrance and aggressiveness of familial CCM disease, defining natural history, or assessing response to novel therapeutic interventions. Murine models of heterozygous CCM disease are likely to reveal lesions at various stages of development, including the earliest putative CCM lesions that are rarely available for study in human pathologic material (Fig. 15.6, left). This will open new venues for studying CCM pathogenesis by characterizing various facets of lesion histology, immunohistochemical and other biomarkers, and MRI correlates, at various stages of lesion genesis and development, and in response to therapeutic manipulations. Clinical imaging may become possible in vivo at higher field strengths. This may reveal other features

153

of lesion phenotype in sporadic versus familial lesions, and in lesions associated with previous brain irradiation, venous developmental anomaly, or different familial genetic substrates. These same features of lesion behavior will need to be sought and followed over time, in correlation with clinical lesion activity. Most CCM lesions remain clinically quiescent in the host’s lifetime, except for brief periods of apparent proliferation and hemorrhage. It will be useful to define biologic states underlying clinical activity in these lesions. T1 contrast agents conjugated to specific antibodies against markers for endothelium and inflammatory cells could provide a platform for the development and validation of molecular imaging, as was done for intracellular adhesion molecule-1 by Lee et al. (2009) and for inducible nitric oxide synthase by Towner et al. (2010), giving another dimension for validating molecular profiles in different regions of these lesions, hopefully reflecting aggressive and quiescent states and clinical behavior of CCMs. The advent of MRI has in many ways helped define CCM disease. Advanced MRI techniques still hold many opportunities for improving our understanding of CCM biology and clinical behavior. Acknowledgments We thank Palamadai M. Venkatasubramanian and Alice M. Wyrwicz for assistance in MR imaging and Douglas A. Marchuk for the murine models for CCMs.

References Abdulrauf SI, Kaynar MY, Awad IA (1999) A comparison of the clinical profile of cavernous malformations with and without associated venous malformations. Neurosurgery 44:41–46. discussion 46–47 Abla A, Wait SD, Uschold T, Lekovic GP, Spetzler RF (2008) Developmental venous anomaly, cavernous malformation, and capillary telangiectasia: spectrum of a single disease. Acta Neurochir (Wien) 150:487–489 Awad IA, Robinson JR Jr, Mohanty S, Estes ML (1993) Mixed vascular malformations of the brain: clinical and pathogenetic considerations. Neurosurgery 33:179–188 Batra S, Lin D, Recinos PF, Zhang J, Rigamonti D (2009) Cavernous malformations: natural history, diagnosis and treatment. Nat Rev Neurol 5:659–670 Bradley WG Jr (1993) MR appearance of hemorrhage in the brain. Radiology 189:15–26 Brunereau L, Leveque C, Bertrand P, Tranquart E, Cordoliani Y, Rouleau P, Labauge P (2001) Familial form of cerebral cavernous malformations: evaluation of gradient-spin-echo

154 (GRASE) imaging in lesion detection and characterization at 1.5 T. Neuroradiology 43:973–979 Campbell PG, Jabbour P, Yadla S, Awad IA (2010) Emerging clinical imaging techniques for cerebral cavernous malformations: a systematic review. Neurosurg Focus 29:E6, 1–8 Carr HY, Purcell EM (1954) Effects of diffusion on free precession in nuclear magnetic resonance experiments. Phys Rev 94:630–638 de Souza JM, Domingues RC, Cruz LC Jr., Domingues FS, Iasbeck T, Gasparetto EL (2008) Susceptibility-weighted imaging for the evaluation of patients with familial cerebral cavernous malformations: a comparison with t2-weighted fast spin-echo and gradient-echo sequences. AJNR Am J Neuroradiol 29:154–158 Freeman R, Hill HDW (1971) Fourier transform study of NMR spin-lattice relaxation by “progressive saturation”. J Chem Phys 54:3367–3377 Gault J, Sain S, Hu LJ, Awad IA (2006) Spectrum of genotype and clinical manifestations in cerebral cavernous malformations. Neurosurgery 59:1278–1285 Labauge P, Laberge S, Brunereau L, Levy C, Tournier-Lasserve E (1998) Hereditary cerebral cavernous angiomas: clinical and genetic features in 57 French families. Societe Francaise de Neurochirurgie. Lancet 352:1892–1897 Leblanc GG, Golanov E, Awad IA, Young WL (2009) Biology of vascular malformations of the brain. Stroke 40: e694–702 Lee SI, Lee SY, Yoon KH, Choi KS, Jang KY, Yoo WH, Kim SH, Choi TH, Park JG (2009) Molecular MR imaging for visualizing ICAM-1 expression in the inflamed synovium of collagen-induced arthritic mice. Korean J Radiol 10: 472–480 Lee BC, Vo KD, Kido DK, Mukherjee P, Reichenbach J, Lin W, Yoon MS, Haacke M (1999) MR high-resolution blood oxygenation level-dependent venography of occult (low-flow) vascular lesions. AJNR Am J Neuroradiol 20:1239–1242 Maraire JN, Awad IA (1995) Intracranial cavernous malformations: lesion behavior and management strategies. Neurosurgery 37:591–605 Meiboom S, Gill D (1958) Modified spin-echo method for measuring nuclear relaxation times. Rev Sci Instrum 29: 688–691 Pinker K, Stavrou I, Knosp E, Trattnig S (2006) Are cerebral cavernomas truly nonenhancing lesions and thereby distinguishable from arteriovenous malformations? MRI findings and histopathological correlation. Magn Reson Imag 24:631–637

R. Shenkar et al. Pinker K, Stavrou I, Szomolanyi P, Hoeftberger R, Weber M, Stadlbauer A, Noebauer-Huhmann IM, Knosp E, Trattnig S (2007) Improved preoperative evaluation of cerebral cavernomas by high-field, high-resolution susceptibility-weighted magnetic resonance imaging at 3 Tesla: comparison with standard (1.5 T) magnetic resonance imaging and correlation with histopathological findings–preliminary results. Invest Radiol 42:346–351 Plummer NW, Gallione CJ, Srinivasan S, Zawistowski JS, Louis DN, Marchuk DA (2004) Loss of p53 Sensitizes Mice with a Mutation in Ccm1 (KRIT1) to Development of Cerebral Vascular Malformations. Am J Pathol 165:1509–1518 Plummer NW, Squire TL, Srinivasan S, Huang E, Zawistowski JS, Matsunami H, Hale LP, Marchuk DA (2006) Neuronal expression of the Ccm2 gene in a new mouse model of cerebral cavernous malformations. Mamm Genome 17:119–128 Pozzati E, Marliani AF, Zucchelli M, Foschini MP, Dall’Olio M, Lanzino G (2007) The neurovascular triad: mixed cavernous, capillary, and venous malformations of the brainstem. J Neurosurg 107:1113–1119 Rigamonti D, Drayer BP, Johnson PC, Hadley MN, Zabramski J, Spetzler RF (1987) The MRI appearance of cavernous malformations (angiomas). J Neurosurg 67:518–524 Rivera PP, Willinsky RA, Porter PJ (2003) Intracranial cavernous malformations. Neuroimaging Clin N Am 13:27–40 Schlamann M, Maderwald S, Becker W, Kraff O, Theysohn JM, Mueller O, Sure U, Wanke I, Ladd ME, Forsting M, Schaefer L, Gizewski ER (2010) Cerebral cavernous hemangiomas at 7 Tesla: initial experience. Acad Radiol 17:3–6 Shenkar R, Venkatasubramanian PN, Wyrwicz AM, Zhao JC, Shi C, Akers A, Marchuk DA, Awad IA (2008b) Advanced magnetic resonance imaging of cerebral cavernous malformations: part II. Imaging of lesions in murine models. Neurosurgery 63:790–797. discussion 797–798 Shenkar R, Venkatasubramanian PN, Zhao JC, Batjer HH, Wyrwicz AM, Awad IA (2008a) Advanced magnetic resonance imaging of cerebral cavernous malformations: part I. High-field imaging of excised human lesions. Neurosurgery 63:782–789. discussion 789 Towner RA, Smith N, Doblas S, Garteiser P, Watanabe Y, He T, Saunders D, Herlea O, Silasi-Mansat R, Lupu F (2010) In vivo detection of inducible nitric oxide synthase in rodent gliomas. Free Radic Biol Med 48:691–703 Zabramski JM, Wascher TM, Spetzler RF, Johnson B, Golfinos J, Drayer BP, Brown B, Rigamonti D, Brown G (1994) The natural history of familial cavernous malformations: results of an ongoing study. J Neurosurg 80:422–432

Chapter 16

Nosologic Imaging of Brain Tumors Using MRI and MRSI Jan Luts, Teresa Laudadio, Albert J. Idema, Arjan W. Simonetti, Arend Heerschap, Dirk Vandermeulen, Johan A.K. Suykens, and Sabine Van Huffel

Abstract A new technique is presented to create nosologic images of the brain based on magnetic resonance imaging and magnetic resonance spectroscopic imaging (MRSI). A nosologic image summarizes the presence of different tissues and lesions in a single image by color coding each voxel or pixel according to the histopathological class it is assigned to. The proposed technique applies advanced methods from image processing as well as pattern recognition to segment and classify brain tumors. First, a registered brain atlas and a subject-specific abnormal tissue prior, obtained from MRSI data, are used for the segmentation. Next, the detected abnormal tissue is classified based on supervised pattern recognition methods. Class probabilities are also calculated for the segmented abnormal region. Compared to previous approaches, the new framework is more flexible and able to better exploit spatial information leading to improved nosologic images. The combined scheme offers a new way to produce high-resolution nosologic images, representing tumor heterogeneity and class probabilities, which may help clinicians in decision making. Keywords Nosologic imaging · MRI · Brain · MRSI · Tumor · Pattern recognition

Introduction Contrast-enhanced magnetic resonance imaging (MRI) is an important tool for the anatomical assessment J. Luts () Department of Electrical Engineering (ESAT/SISTA), K.U. Leuven, Leuven, Belgium e-mail: [email protected]

of brain tumors. As a consequence, automatic brain tumor segmentation based on MRI is a widely studied topic (Kaus et al., 2001; Warfield et al., 2000). Nowadays, techniques based on thresholds, region growing, clustering, Markov random fields, classification algorithms, and artificial neural networks have been used and accurate segmentation results have been reported after comparison with manual tumor segmentation by an expert. However, several diagnostic questions, such as the type and grade of the tumor, are difficult to address using conventional MRI. The histopathological characterization of a tissue specimen remains the gold standard, despite the associated risks of surgery to obtain a biopsy. In recent years, the use of magnetic resonance spectroscopy (MRS), which provides metabolic information, is increasingly being used for more detailed and specific noninvasive evaluation of brain tumors. In particular magnetic resonance spectroscopic imaging (MRSI), which provides quantitative metabolite maps of the brain, is attractive as this may also enable to visualize the heterogeneous spatial extent of tumors, both inside and outside the MRI detectable lesion. However, the individual inspection and analysis of the many spectral patterns, obtained by MRSI, remains extremely time-consuming and requires specific spectroscopic expertise. Therefore, it is not practical in a clinical setting, where automated processing and evaluation of the MRSI data as well as easy and rapid display of the results as images or maps are needed for routine clinical interpretation of an exam. As such, the term nosologic image was introduced for an image which indicates a specific tissue type in a certain color (Szabo de Edelenyi et al., 2000). These authors used pattern recognition techniques and combined one specific contrast from

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_16, © Springer Science+Business Media B.V. 2011

155

156

MRI with spectroscopic information. Simonetti et al. (2003) combined additional image variables with the metabolic information to construct nosologic images. Both studies used the combination of MRI and MRSI for new, more relevant brain tumor classifiers, but treated each pixel or voxel of the nosologic image independently, only exploiting the spectral information and image intensities. Therefore, Laudadio et al. (2008) proposed the application of canonical correlation analysis (CCA) to MRSI data as a way to include spatial information when producing nosologic images. Afterwards, Kelm (2007) suggested to apply conditional random fields to MRSI data since the technique by Laudadio et al. (2008) uses a sliding window which can cancel out effects from isolated voxels and some long-range interactions are lacking because of the fixed window size. Kelm (2007) claimed that his approach, based on a global method, is able to increase the performance of single voxel classifiers by at least 15%. In this study, we extend previous work to construct improved, high-resolution nosologic images. The new scheme benefits from both advanced image processing methods, producing an improved segmented image, and sophisticated classification methods from pattern recognition. The framework comprises a segmentation step and a classification step where MRSI and MRI data are simultaneously exploited, thereby yielding improved segmentation and classification results. Furthermore, spatial information is included in both steps and class probabilities are also generated.

Experimental Data Data coming from 24 patients with a brain tumor and 4 healthy volunteers were selected from the INTERPRET project database. The MRI and MRSI data were acquired in the University Medical Center Nijmegen (UMCN). The study was approved by the ethical committee of the UMCN and followed the rules of the World Health Organization (WHO). All data passed a strict quality control and the tumor type was determined by histopathological consensus. To create a sufficiently large training data set for the pattern recognition methods, several voxels, situated in the tumor area, were selected from each patient

J. Luts et al.

based on histological, radiological, and spectroscopic information (Devos et al., 2005). Voxels for a specific tissue class were only selected from patients for which this tissue type was confirmed by a histopathological consensus, performed by three independent pathologists. Because of the heterogeneity of brain tumors, the MRI data and the spectral information were taken into account during the selection of voxels. For this purpose, for each patient all the high-resolution MR images were plotted together with the MRSI spectra and a segmented image in which voxels from MRSI were clustered by a model-based algorithm using a graphical user interface. Because the clustering provided an objective segmentation, this was considered to be helpful for voxel selection. Based on all these data sources, an expert in MR spectroscopy selected voxels if they were clearly within the affected brain region, avoiding contamination from normal cerebral tissue, such that spectra under consideration were found to be typical for that pathology. Normal brain MR spectra typically have a large NAA peak around 2 ppm and almost equal peaks around 3.2 and 3.03 ppm, respectively, choline and creatine. Cerebrospinal fluid (CSF) would show no signal in the spectra, but due to partial volume effects, profiles, similar to normal MR spectra, are often visible but with lower intensities. On the other hand, tumor MR spectra have elevated choline compared with creatine. Grade II gliomas have a large myo-inositol peak at 3.57 ppm, and elevated choline compared with creatine. Grade III glioma MR spectra have an intermediary pattern between grade II glioma and glioblastoma MR spectra. Glioblastoma MR spectra reveal dominant lipid/lactate peaks around 0.9 and 1.3 ppm. Some high-grade tumors have higher choline, and in general there is a large range of lipid and metabolite concentrations. Meningiomas are characterized by a number of almost equally sized peaks, broad peaks centered around 2.3 and 3.78 ppm, which most likely comprise glutamate, glutamine and macromolecules, choline (3.2 ppm), and various amounts of signal at 0.9 and 1.3 ppm, macromolecules/lipid/lactate. Following standard radiological MRI features, tumor is normally hypointense on a T1-weighted image and hyperintense on a T2- and proton density-weighted image, while contrast enhancement usually indicates higher grade tissue for gliomas. CSF is hypointense on a T1-weighted image, hyperintense on a T2-weighted image, and isointense

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

on a proton density-weighted image. Necrosis is usually hypointense and shows no contrast enhancement. Although this method was subjective, it was chosen because tumors are known to be heterogeneous and because there is no “ground truth” in the diagnosis of brain tumors at the voxel level. Also, spectra from CSF and normal tissue measured from volunteers and patients were selected. Six classes of tissue were considered: normal brain tissue from volunteers and apparently normal tissue from the contralateral half of the brain of patients (218 voxels from 8 persons), CSF (100 voxels from 8 patients), grade II glioma (176 voxels from 10 patients), grade III glioma (69 voxels from 5 patients), grade IV glioma (70 voxels from 7 patients), and meningioma (48 voxels from 3 patients). The data were acquired on a 1.5 T Siemens Vision Scanner with CP-head coil. Four MR images were acquired, namely T1-weighted image (TE/TR = 15/644 ms), T2-weighted image (TE/TR = 98/3,100 ms), proton density-weighted image (TE/TR = 16/3,100 ms) and gadolinium-enhanced T1-weighted image (15 ml 0.5 M Gd-DTPA), with matrix size equal to 256 × 256. Both water-suppressed and unsuppressed proton MR spectroscopic images were measured using a 2D STEAM sequence with the STEAM box positioned totally in the brain (16 × 16 × 1,024 samples, TR/TE/TM = 2,000 or 2,500/20/30 ms, slice thickness = 12.5 or 15 mm, FOV = 200 mm, spectral width = 1,000 Hz and NS = 2). Disturbing signals arising from the fat tissue surrounding the skull were avoided. The location of the STEAM box was determined using the gadolinium-enhanced T1-weighted image showing the largest tumor area. The MRSI slice was centered around an MRI slice of 5 mm. MR image registration was performed based on the algorithm described in Maes et al. (1997). It was assumed that the MRSI data were registered with the proton density-weighted image since they were acquired in a consecutive manner. Preprocessing of MRSI included filtering of k-space data by a Hanning filter of 50% using the LUISE software package (Siemens, Erlangen, Germany), zero filling to 32 × 32, spatial 2D Fourier transformation to obtain time domain signals for each voxel, correction for eddy current effects by a technique which prevents occasional occurrence of eddy current correction induced artifacts (Simonetti et al., 2002), frequency alignment, water removal using HLSVD from 4.3 to 5.5 ppm

157

(Pijnappel et al., 1992), and a simple baseline correction using an exponential filter with a width of 5 ms followed by subtraction of the residual of the original signal. All first order phases were corrected by first manually optimizing the mean spectrum which was calculated from all spectra in the STEAM box of each patient’s MRSI data. Next, this correction was applied to each separate signal of the patient’s MRSI data. Finally, the spectra were normalized using the water signal. For each metabolite signal, a corresponding water-unsuppressed signal was available, acquired with the same acquisition parameters as the metabolite signal and originating from the same voxel. The amplitude of the water-unsuppressed signal was estimated by a nonlinear least squares algorithm (Devos et al., 2004). The water-unsuppressed signal was modeled in the time domain by a Voigt model, with an additional first order term that corrected for eddy currents. HSVD was used to obtain initial parameter estimates, which were subsequently used as starting values in a nonlinear least squares algorithm to obtain the final parameter estimates. Each spectral value in the preprocessed water-suppressed spectrum was divided by the obtained estimate of the intensity of the water peak. In this study, an input pattern for pattern recognition was a set of features extracted from the spectrum in combination with MRI pixel intensities. First, the magnitude spectrum was computed and the spectral points within the frequency range (0.5–4 ppm) were considered. Figure 16.1 depicts the water-normalized mean magnitude spectra, along with the standard deviation, for each of the tissue types in the data set. Next, ten different features were extracted, using peak integration within a spectral range, from each spectrum (Govindaraju et al., 2000): lipids (0.835–0.965 ppm), lipids (1.2 ppm) and lactate + alanine (1.265– 1.395 ppm), N-acetylaspartate (1.955–2.085 ppm), glutamate + glutamine (2.135–2.265 ppm), creatine (2.955–3.095 ppm), choline (3.135–3.265 ppm), taurine + glucose (3.375–3.505 ppm), myo-inositol + glycine (3.495–3.625 ppm), glutamate + glutamine + alanine (3.685–3.815 ppm), and creatine (3.885– 4.015 ppm). These ten values, referred to as the peak-integrated values in the rest of this chapter, were combined with four MRI variables obtained by lowering the resolution of the four MR images to the one of the MRSI grid by averaging pixel intensities within each voxel as in Simonetti et al. (2003).

158

J. Luts et al.

Fig. 16.1 Mean water-normalized magnitude MR spectra for normal tissue, CSF, grade II gliomas, grade III gliomas, meningiomas, and glioblastomas. The solid lines are the means, while the dotted lines are the means plus the standard deviations

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

A Two-Step Segmentation-Classification Framework In this chapter, we propose a two-step segmentationclassification framework able to simultaneously exploit MRI and MRSI data. In particular, the first step was performed based on the segmentation method presented by Prastawa et al. (2004). The required input for that method was a T2-weighted image. No contrast-enhanced images were needed, but additional types of MR images could be used to improve the segmentation. More precisely, Prastawa et al. (2004) proposed a method consisting of three stages: abnormality detection using a digital brain atlas and outlier detection; edema detection based on the T2-weighted image intensities; application of geometric and spatial constraints to the detected edema and tumor regions. In our new framework, we included additional information from MRSI to improve the abnormal tissue detection for segmentation. Specifically, as described in the next section, in addition to the standard brain atlas, a subject-specific abnormal tissue prior was generated by applying a supervised pattern recognition technique to the available MRSI (and MRI) data set. In the second step of the segmentation-classification framework all pixels from the detected abnormal region were classified based on supervised pattern recognition techniques and nosologic images were provided as final result, where each tissue type was denoted by a specific color. The two-step framework is explained in the following paragraphs.

MRI Segmentation Generating a Subject-Specific Abnormal Tissue Prior from MRSI and MRI In a digital brain atlas, only probabilities are provided for the healthy tissue classes: white matter, gray matter, and CSF. The study in Prastawa et al. (2003) proposed the inclusion of a subject-specific abnormal tissue prior by means of the contrast-enhanced image. In this study, we generated a subject-specific prior for the abnormal tissue as well, but based on a linear discriminant analysis (LDA) classifier. Indeed, in Luts et al. (2007) it was shown that a simple linear classifier was sufficient to separate abnormal tissue from healthy

159

tissue. Therefore, we applied LDA (trained on all available training data except data coming from the case under study) to the grid of feature vectors containing 14 entries, that is ten bicubically interpolated peakintegrated values (such that it matched with the size of the MRI grid) and the four types of MRI pixel intensities. For each pixel within the area of the STEAM box, LDA predicted whether there was abnormal or healthy tissue based on all peak-integrated values and the MRI pixel intensities. The lower left panel of Fig. 16.2 depicts the resulting grid that was generated by the LDA classifier for case 4. For illustration purposes, the choline over creatine ratio, obtained by the interpolated peak-integrated values, is depicted in the right panel of the middle row. Next, the output from LDA was integrated into the original digital brain atlas by setting the abnormal prior, if abnormal tissue was predicted by the LDA method, to a probability derived from that classifier. If no abnormal tissue was predicted by LDA, the abnormal prior was set to a fraction of the white matter and gray matter atlas probabilities as in Prastawa et al. (2004). Since the MRSI information was restricted to the STEAM box, the abnormal tissue prior was only available within that area. Every prior probability outside that box was obtained as in Prastawa et al. (2004). In the results section, it is demonstrated that the addition of the abnormal tissue prior improves the segmentation procedure.

Segmentation Based on the Modified Atlas After the prior was generated, part of the segmentation approach from Prastawa et al. (2004) was followed. Abnormal tissue was detected using an iterative procedure. First a threshold was used on the posterior probabilities, or atlas probabilities in the first iteration, and pixels in high confidence regions were sampled. Next, the samples from normal tissue that exceeded a distance threshold, based on the minimum covariance determinant (MCD) estimator, were removed. A nonparametric model was used to estimate the probability density functions from the samples. The posterior probabilities were calculated based on the subject-specific atlas priors and the estimated densities. To conclude an iteration of abnormality detection, a bias field correction was performed because of the presence of the image inhomogeneity.

160

J. Luts et al.

Fig. 16.2 Case 4, grade II/III glioma. Top row: T1weighted MR image, T2-weighted MR image, proton densityweighted MR image, and gadolinium-enhanced T1-weighted MR image. The arrow identifies the enhanced area. Middle row: T2-weighted MR image and MR spectra originating from the identified area, choline over creatine ratio map (high ratios are white). Bottom row: probabilities for abnormal tissue generated by the LDA classifier (high probabilities are white) to

create subject-specific abnormal tissue prior, selected samples for healthy tissue classes, nosologic image (light blue = white matter, dark blue = gray matter, green = CSF, yellow = grade II glioma, orange = grade III glioma), contour plot for grade III glioma (blue = high probability, red = low probability). Grade III glioma probabilities are higher at the frontal part of the abnormal tissue region with the enhancement after gadolinium and smaller at the occipital part of the tumor

The bias field was estimated from the white and gray matter probabilities and the full procedure was repeated. After the separation of abnormality, an edema detection stage was proposed in the framework in Prastawa et al. (2004). The detection was done using the T2-weighted image, assuming that edema has high fluid content and appears brighter than the tumor. We decided to remove this stage from our segmentation step since, after validation of the results, it seemed

that cysts, which were also brighter on the T2 image, were often segmented as edema. Also, it is known that edema in glioma tumors is often infiltrated and therefore this tissue should be detected as abnormal. The last stage of the segmentation step was reclassification with spatial and geometric constraints. For this purpose region competition snakes were used (Ho et al., 2002). First, a threshold was used on the posterior probabilities and a sampling was performed. The

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

tumor samples were pruned based on the level-set evolution after which the densities were estimated. New posterior class probabilities were calculated; the bias field was corrected and the procedure was repeated.

Classification of the Abnormal Region Determination of the Type of Abnormal Tissue After segmentation, white matter, gray matter, CSF, and abnormal tissue were separated. To provide additional information about the tumor type that was present within the abnormal tissue region, a supervised pattern recognition step was introduced. As for the subject-specific abnormal tissue prior, the MRSI data (e.g. spectra, peak-integrated values) were interpolated to match the size of the MRI data. Naturally, only pixels within the STEAM box could be classified. More precisely, all pixels, found to be abnormal by the segmentation step, were classified into one of the tumor categories that were represented in the training set for pattern recognition. This means that for this specific study, we were able to detect white matter, gray matter, CSF, and, based on spectroscopic profiles, grade II glioma, grade III glioma, grade IV glioma, and meningioma.

Pattern Recognition Methods Concerning classification, basically every (semi-) supervised pattern recognition method can be incorporated, which makes the framework very flexible. Here, we decided to focus on kernel logistic regression (KLR) (Karsmakers et al., 2007), a multiclass classifier using Bayesian least squares support vector machines (LS-SVMs) (Luts et al., 2007) and CCA. Bayesian LS-SVMs (Van Gestel et al., 2002) and KLR were used since these kernel-based methods were able to provide class probabilities instead of “black or white” decisions. CCA was included because it was able to exploit spatial information (Laudadio et al., 2008). The methods are briefly discussed in the following paragraph. Standard LS-SVMs (Suykens and Vandewalle, 1999) can only handle binary classification problems and produce binary output scores. In Luts et al. (2007) it was proposed to combine binary Bayesian LSSVMs with radial basis function (RBF) kernel into a

161

multiclass classifier system. Based on binary classifiers and pairwise combination algorithms multiclass probabilities were retrieved. Following the findings in Luts et al. (2007), the combination method of Wu et al. (2004) was used to combine pairwise class probabilities. Direct multiclass probabilities were generated using KLR. In Karsmakers et al. (2007) it was shown that the optimization problem for KLR could be solved by a weighted version of LS-SVMs. From the practical point of view this means that KLR does not need to construct binary classifiers; one direct multiclass solution can be obtained. Finally, CCA was included since the method exploits spatial information, that is based on the information in the neighboring pixels, a decision is made for the center pixel. Since CCA was applied to the interpolated grid (within the abnormal tissue region), we reduced the risk to average out the effect of isolated tumor voxels by the information that was present in the neighborhood as compared to (Laudadio et al., 2008). The symmetric 3 × 3 model without corner voxels was used as spatial model and principal component analysis (PCA) was applied as subspace model.

Results A leave-one-person-out validation analysis was performed to evaluate the framework. For the creation of the abnormal tissue prior and for the classification step, all data from the test subject were removed from the training set. As such, model selection was performed on all data except for the voxels coming from the subject under study. In the next paragraph, nosologic images are reported and the use of class probabilities is illustrated. Finally, the use of the abnormal tissue prior is explained. Table 16.1 provides an overview of the patients’ data. Table 16.1 Pathology, tumor grade, age, and sex for 9 subjects Case

Pathology

Grade

Age

Sex

1 2 3 4 5 6 7 8 9

Diffuse astrocytoma Diffuse astrocytoma Oligoastrocytoma Oligodendroglioma Oligodendroglioma Glioblastoma Glioblastoma Glioblastoma Meningioma

II II II II/III III IV IV IV I

58 39 42 41 30 53 54 50 45

Female Male Male Female Female Male Male Male Female

162

Analysis of the Patient Data The framework was applied to generate nosologic images for nine patients (Figs. 16.2 and 16.3). Light blue reflects white matter, dark blue gray matter, green CSF, yellow grade II glioma, orange grade III glioma, dark red glioblastoma, and purple meningioma. All

Fig. 16.3 MR image and nosologic image for case 1, 2, 3, 5, 6, 7, 8, and 9. Light blue reflects white matter, dark blue gray matter, green CSF, yellow grade II glioma, orange grade III glioma,

J. Luts et al.

classification results in this section were generated using feature vectors where peak-integrated values were combined with image intensities. To demonstrate the flexibility of the framework, and more specifically the flexibility of the classification step, the use of various pattern recognition methods is illustrated. Nevertheless, it is worth stressing that the goal of the

dark red glioblastoma, and purple meningioma. The hematoma for case 8 is identified by an arrow

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

present study is not to compare the different pattern recognition methods, but to show that any type of classifier can be integrated in the classification step, or can be used for generating the additional subject-specific abnormal tissue prior in the segmentation step. The flexibility property shown here is very important as, based on the specific problem under investigation, ad hoc classification methods could be applied. Cases 6–8 represent patients with a glioblastoma multiforma tumor as assigned by the pathologists. The nosologic images, obtained by applying a Bayesian LS-SVM classifier, show that the method was able to detect the tumor lesions and glioblastoma tissue (Fig. 16.3). These examples also show that edema and/or tumor infiltration is difficult to assess. In diffuse astrocytoma, the peritumoral edema is presumed infiltrated by glioma tumor cells. In these cases, the differentiation between pure infiltration and pure edema cannot be made. As discussed before the segmentation of edema was not consistent and was left out of the procedure and so no edema is shown in the nosologic images. In heterogeneous tumors, like gliomas, this distinction cannot be made with conventional MRI techniques. Therefore, the most likely assignment for this abnormal tissue is low grade glioma. Low grade tumor next to glioblastoma is pathologically possible. In case 8, there are two tumor areas. The posterior part could include a hematoma, identified by an arrow, within the tumor. For the surrounding tissue it is difficult to distinguish edema from infiltrating tumor. Cases 3, 4, and 5 are good examples of the difficult differentiation between WHO grade II and grade III glioma brain tumor. The histopathological diagnosis for case 3 was grade II glioma. The framework, applying CCA in the classification step, found the correct tissue type, despite the enhancement after contrast for small parts of the tumor at the occipital parts of the insula (Fig. 16.3). For case 4, the pathologists disagreed between a grade II and a grade III glial tumor. The nosologic image, created using a Bayesian LSSVM classifier, shows a heterogeneous tumor with mainly grade III glioma, according to the spectroscopic classifier, especially in the part with enhancement after gadolinium (Fig. 16.2). The differentiation between grade II glioma and edema at the frontal and medial parts cannot be made. The use of class probabilities is illustrated by means of a contour plot. Blue contour lines show higher grade III glioma probabilities as opposed to red contour lines. The grade III

163

glioma probabilities are higher at the frontal part of the abnormal tissue region with the enhancement after gadolinium and smaller at the occipital part of the tumor in case 4. For case 5 with an anaplastic oligodendroglioma, CCA was used (Fig. 16.3). Almost only low grade glioma, WHO grade II was found with a small region of anaplastic glioma, WHO grade III. This would be an underestimation of the real situation. At surgery, a debulking of the tumor was performed. Therefore, the pathological diagnosis was not from only the small part with a nosological anaplastic tumor. Case 2 is a diffuse low grade astrocytoma, grade II, with no enhancement after contrast and good segmentation of the tumor/tumor infiltration/edema based on KLR (Fig. 16.3). For case 1 the histopathological agreement was grade II glioma, despite the enhancement after gadolinium. The Bayesian LS-SVM method was used and it detected the correct tissue type (Fig. 16.3). However, in this case no edema and/or tumor infiltration was detected which was very reliable around the enhancing spot. The nosologic image for case 9 with a meningioma tumor was obtained by the CCA classifier (Fig. 16.3). In this case, the small edema part at the occipital part of the tumor was not classified as abnormal tissue. In the case of a low grade meningioma there is no infiltration.

Subject-Specific Abnormal Tissue Prior During the detection of abnormality step a sampling for gray matter, white matter, and CSF was done based on the atlas priors. The bottom row (second panel from the left) of Fig. 16.2 depicts the selected samples for the healthy tissue classes (i.e., gray matter, white matter, and CSF) following the modified digital brain atlas for case 4. It can be observed that, in contrast to the approach in Prastawa et al. (2004), almost no samples for healthy tissue were picked from the tumor area after the subject-specific abnormal tissue prior was added to the standard atlas. The extra prior clearly improves the sampling and, as a consequence, reduces the contamination reported in Prastawa et al. (2004). As a result, improved sampling leads to higher quality prototypes for all the tissue classes and better density estimates in the detection of the abnormality step. Finally, an example (case 3) of the difference between the two approaches with respect to the final segmentation is

164

J. Luts et al.

Fig. 16.4 The effect of using a subject-specific abnormal tissue prior based on MRSI for case 3. The results on the left are generated without an abnormal tissue prior in contrast to the results on the right. The top row shows the prior probabilities from the atlas, the bottom row the final segmentation result. The abnormal tissue is only correctly recognized by the segmentation algorithm that includes the extra abnormal tissue prior from MRSI

provided (Fig. 16.4). In contrast to the panels on the right, no abnormal tissue prior was added to the digital brain atlas to produce the left images. The top row summarizes the information that was available in the atlas before starting segmentation. This result was generated by choosing the tissue type with the highest prior probability based on the (modified) atlas. The lower row shows the final result after segmentation. CSF is colored green, white matter light blue, gray matter dark blue, and abnormal tissue is red. Before starting segmentation, one can observe that there was only a region for which the probabilities of abnormal tissue were higher than the ones of the normal classes in the digital brain atlas that was modified with the subject-specific abnormal tissue prior. After segmentation, and when comparing with the MRI of case 3 (Fig. 16.3) only a small part of the abnormal tissue region was detected by the original segmentation algorithm that just included healthy tissue priors. By introducing the extra prior from MRSI, the abnormal

tissue region was better recognized. The MR spectra confirmed the existence of abnormal tissue in that area.

Discussion In MRI-based segmentation of tissue in the brain, a digital brain atlas is an important, additional, tool to resolve the problem of overlap between intensity distributions of different tissue types. Because a digital brain atlas only contains priors for the healthy tissue classes, researchers attempted to include a prior for the abnormal tissue as well. Typically, the contrastenhanced T1-weighted image is used for this purpose (Prastawa et al., 2003). However, it is known that blood vessels also generally appear enhanced. In addition, tumors may show only partially signal enhancement and some tumors do not enhance at all. For this reason, we decided to use the MRSI data to generate the prior

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

for tissue abnormality. Although the spatial resolution of MRSI is low compared to MRI, it appeared that MRSI was very useful for defining the prior and that it could improve the original segmentation algorithm described in Prastawa et al. (2004). Indeed, in Prastawa et al. (2004) it was explained that the segmentation algorithm could only deal with a limited amount of brain deformation. During the detection of abnormality step sampling is initially based on the atlas. The outliers are detected based on the MCD method and the density functions are estimated. However, when there is a serious brain deformation, the brain atlas leads to incorrect sampling since pixels for the healthy tissue classes are sampled from the abnormal tissue regions. In this case, the samples are severely contaminated and the MCD algorithm may not yield correct results, leading to incorrect estimated densities. Figure 16.2 and 4 illustrate this and show that the use of the abnormal tissue prior, generated from MRSI, is advantageous. In general, the proposed framework highlights the advantages of a method able to exploit and combine different types of information. In particular, our study shows the value of adding MRSI over MRI data, of adding MRI over MRSI information, and the improvements achieved in the segmentation and classification approaches with respect to their previously proposed versions. For instance, several diagnostic questions, such as the type and grade of the tumor, are difficult to address using MRI, but MRSI data permit taking the corresponding metabolic differences into account. Therefore, using MRSI information over just MRI has the additional advantage that nosologic images (specifying the type of the tumor) can be obtained, instead of the results (which only localize the tumor) obtained with classical MRI segmentation approaches. The added value of MRI over just MRSI has already been evaluated in a number of studies, which concluded that combining MRI with MRSI features improved the performance of the classifiers (Devos et al., 2005; Galanaud et al., 2006; Simonetti et al., 2005). For this reason, in the present study, the four MR images were also used in the classification step. Furthermore, adding MRI over MRSI data, allowed to perform the segmentation step based on highresolution MRI data. We found that this improved the detection of CSF compared to pure segmentation of MRSI data and also that gray matter and white matter were detected based on MRI. In contrast, the CCA approach by Laudadio et al. (2008) only performed a

165

segmentation on the MRSI data. CCA is an interesting method since it is able to simultaneously exploit the spectral-spatial information provided by the MRSI data and is computationally efficient. On the other hand, the disadvantage of this approach is that the same spatial constraints (i.e., a fixed spatial model) are used for every voxel of the MRSI grid, without taking into account whether the voxel is on the border or the middle of the tumor. In simulation studies, we found that this can lead to wrong classifications for rather isolated tumor voxels. In the proposed framework, CCA was applied after the segmentation step, that is after the tumor was detected, reducing the risk of misclassification for rather isolated tumor voxels. Therefore, the above issue is improved by making use of image processing methods that incorporate more advanced spatial constraints, based on the high-resolution MRI data. The use of these spatial constraints in the current method is also an improvement compared to the approaches where every pixel or voxel was classified separately, ignoring this prior knowledge (Luts et al., 2007; Simonetti et al., 2003; Szabo de Edelenyi et al., 2000). On the other hand, as proven by biopsies, outside the abnormal tissue on the conventional MRI image, infiltrating tumor cells can be found with and without abnormal spectroscopic findings (Croteau et al., 2001; Ganslandt et al., 2005; Li et al., 2002; Price et al., 2006). Therefore, the whole infiltrating tumor is not always taken into account. Price et al. (2006) showed a biopsy-guided improved delineation of glioma tumors with diffusion tensor imaging. Also, the segmentation of edema in glioma tumors is not straightforward. To distinguish edema from tumor the intensity of a T2-weighted image is not sufficient. This does not always differentiate tumor cysts, resection cysts, or cerebral fluid from pure edema. Next to this, in glioma tumors the peritumoral edema is mostly infiltrated by tumor cells and the intensity differences between edema and tumor are on a sliding scale. With the available information in this study no specific segmentation of peritumoral edema and tumor could be made for glioma tumors. Therefore, this segmentation step was not used. Ricci et al. (2007) showed metabolic changes in peritumoral edema with MR spectroscopy. The combination of diffusion tensor imaging and MR spectroscopy might be helpful to improve the delineation of glioma tumors and differentiation of edema and infiltrating tumor.

166

The differentiation between WHO grade II and grade III is radiologically very difficult. According to the WHO-2007 classification (Louis et al., 2007), the main histopathological difference between WHO grade II and III diffuse astrocytic neoplasms is increased mitotic activity in the latter. Therefore, it is not surprising that one third of the nonenhancing diffuse gliomas are histopathologically diagnosed as high-grade lesions at the time of biopsy (Barker et al., 2000). The value of the addition of MR spectroscopy for the differentiation of WHO grade II and grade III gliomas is currently not clear (McKnight et al., 2007; Stadlbauer et al., 2006). However, in part this could be due to the underestimation of the brain tumor malignancy in tumors diagnosed by biopsies (Jackson et al., 2001). By including perfusion and diffusion MR data the differentiation may be improved (Di Costanzo et al., 2006; Law et al., 2003). Although precise estimation of peak integrals is difficult due to peak overlap and noise, features were extracted from the spectra based on simple peak integration. Therefore, spectral fitting methods might seem more appropriate. However, prior to the application of the two-step framework, the classification accuracies of different spectral fitting methods and peak integration were compared. Since the quantification approaches did not improve classifier performance and since these methods were more time-consuming, we decided to use peak integration. Some previous studies also reported no increase in classifier performance for quantification methods compared to automated approaches as for instance PCA (Kelm, 2007; Simonetti et al., 2005). A related issue is determining the number of features to extract from the spectra. In this study, we extracted a common set of features which are in general known to be relevant for the classification problems under study. Although the focus of the present work was the development of the methodology of the framework, this is an issue that can be improved and it is still open for discussion. As such, future work could focus on the optimization of these data preprocessing steps as this can potentially further increase the performance of the total framework that is presented here. In this study, only one of the MRI slices was totally included within the MRSI slice. For this reason, the framework only combined one MRI slice with the MRSI data. However, in practice, depending on its size, multiple MRI slices may be needed to cover the whole tumor. As such, a potential improvement of the

J. Luts et al.

existing framework could be the extension to threedimensional MRSI data with the inclusion of multiple MRI slices. Also, in the present study the standard image orientation protocol in the radiology service was used in order to limit the acquisition time. If a different orientation is used, a better enclosure of the brain tumor mass may be obtained, such that the framework that is presented here provides further improved nosologic images. In conclusion, this study proposes a new method to generate nosologic images of the brain by exploiting and combining MRI and MRSI data. Advanced methods from image processing and pattern recognition are combined in a two-step approach. In the first step, segmentation is performed by means of a modified version of the method described in Prastawa et al. (2004), by including a subject-specific abnormal tissue prior, derived from MRI and MRSI, in the original brain atlas. The subject-specific abnormal tissue prior is found to improve the segmentation procedure. The segmentation step provides an image of the slice of interest where white matter, gray matter, CSF, and the abnormal region are clearly visualized. Next, the abnormal tissue is classified using pattern recognition and class probabilities are generated for the diverse tissue types. The method is applied to various patients with different types of brain tumors. Compared to previous approaches that generate nosologic images, the use of spatial information and prior knowledge is further improved in the proposed method. The framework offers a new way to visualize tumor heterogeneity and class probabilities in a single nosologic image, assisting clinicians in decision making.

Acknowledgements NMR in Biomedicine and John Wiley and Sons are gratefully acknowledged for the license agreement. Jan Luts is a Postdoctoral Fellow of the Research Foundation-Flanders (FWO-Vlaanderen); Research Council KUL: GOA-MaNet, Centers-of-excellence optimisation, GOA/2004/05 (Mixing and Analyzing Real and Virtual Environments and Lighting); Flemish Government: FWO: PhD/postdoc grants, projects, G.0302.07 (Support vector machines and kernel methods), G.0566.06 (Computational strategies for shape modeling and matching and their application in medical image analysis); Belgian Federal Government: DWTC (IUAP IV-02 (1996–2001), IUAP V-22 (2002–2006): Dynamical Systems and Control: Computation, Identification & Modelling) and Belgian Federal Science Policy Office IUAP P6/04 (Dynamical systems, control and optimization, 2007–2011); EU: eTUMOUR (contract no. FP6-2002-LIFESCIHEALTH 503094), FAST (contract no. FP6-019279-2).

16 Nosologic Imaging of Brain Tumors Using MRI and MRSI

References Barker FG II, Chang SM, Huhn SL, Davis RL, Gutin PH, McDermott MW, Wilson CB, Prados MD (2000) Age and the risk of anaplasia in magnetic resonance-nonenhancing supratentorial cerebral tumors. Cancer 80:936–941 Croteau D, Scarpace L, Hearshen D, Gutierrez J, Fisher JL, Rock JP, Mikkelsen T (2001) Correlation between magnetic resonance spectroscopy imaging and image-guided biopsies: Semiquantitative and qualitative histopathological analyses of patients with untreated glioma. Neurosurgery 49: 823–829 Devos A, Lukas L, Suykens JAK, Vanhamme L, Tate AR, Howe FA, Majos C, Moreno-Torres A, van der Graaf M, Arus C, Van Huffel S (2004) Classification of brain tumours using short echo time 1H MR spectra. J Magn Reson 170: 164–175 Devos A, Simonetti AW, van der Graaf M, Lukas L, Suykens JAK, Vanhamme L, Buydens LMC, Heerschap A, Van Huffel S (2005) The use of multivariate MR imaging intensities versus metabolic data from MR spectroscopic imaging for brain tumour classification. J Magn Reson 173:218–228 Di Costanzo A, Scarabino T, Trojsi F, Giannatempo GM, Popolizio T, Catapano D, Bonavita S, Maggialetti N, Tosetti M, Salvolini U, d’angelo VA„ Tedeschi G (2006) Multiparametric 3T MR approach to the assessment of cerebral gliomas: tumor extent and malignancy. Neuroradiology 48:622–631 Galanaud D, Nicoli F, Chinot O, Confort-Gouny S, FigarellaBranger D, Roche P, Fuentes S, Le Fur Y, Ranjeva J-P, Cozzone PJ (2006) Noninvasive diagnostic assessment of brain tumors using combined in vivo MR imaging and spectroscopy. Magn Reson Med 55:1236–1245 Ganslandt O, Stadlbauer A, Fahlbusch R, Kamada K, Buslei R, Blumcke I, Moser E, Nimsky C (2005) Proton magnetic resonance spectroscopic imaging integrated into image-guided surgery: correlation to standard magnetic resonance imaging and tumor cell density. Neurosurgery 56:291–298 Govindaraju V, Young K, Maudsley AA (2000) Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed 13:129–153 Ho S, Bullitt E, Gerig G 2002. Level-set evolution with region competition: Automatic 3-D segmentation of brain tumors. In Proceedings of the 16th International Conference on Pattern Recognition. Quebec, 532–535 Jackson RJ, Fuller GN, Abi-Said D, Lang FF, Gokaslan ZL, Shi WM, Wildrick DM, Sawaya R (2001) Limitations of stereotactic biopsy in the initial management of gliomas. Neuro Oncol 3:193–200 Karsmakers P, Pelckmans K, Suykens JAK 2007. Multi-class kernel logistic regression: a fixed-size implementation. In Proceedings of the 20th International Joint Conference on Neural Networks. Orlando, 1756–1761 Kaus M, Warfield SK, Nabavi A, Black PM, Jolesz FA, Kikinis R (2001) Automated segmentation of MRI of brain tumors. Radiology 218:586–591 Kelm BM 2007. Evaluation of vector-valued clinical image data using probabilistic graphical models: quantification and pattern recognition. Ph.D. Thesis, University of Heidelberg

167 Laudadio T, Martinez-Bisbal MC, Celda B, Van Huffel S (2008) Fast nosological imaging using canonical correlation analysis of brain data obtained by two-dimensional turbo spectroscopic imaging. NMR Biomed 21:311–321 Law M, Yang S, Wang H, Babb JS, Johnson G, Cha S, Knopp EA, Zagzag D (2003) Glioma grading: Sensitivity, specificity, and predictive values of perfusion MR imaging and proton MR spectroscopic imaging compared with conventional MR imaging. Am J Neuroradiol 24:1989–1998 Li X, Lu Y, Pirzkall A, McKnight TR, Nelson SJ (2002) Analysis of the spatial characteristics of metabolic abnormalities in newly diagnosed glioma patients. J Magn Reson Imaging 16:229–237 Louis DN, Ohgaki H, Wiestler OD, Cavenee WK (2007) World Health Organization classification of tumours of the central nervous system. Lyon, IARC Luts J, Heerschap A, Suykens JAK, Van Huffel S (2007) A combined MRI and MRSI based multiclass system for brain tumour recognition using LS-SVMs with class probabilities and feature selection. Artif Intell Med 40:87–102 Maes F, Collignon A, Vandermeulen D, Marchal G, Suetens P (1997) Multimodality image registration by maximization of mutual information. IEEE Trans Med Imaging 16: 187–198 McKnight TR, Lamborn KR, Love TD, Berger MS, Chang S, Dillon WP, Bollen A, Nelson SJ (2007) Correlation of magnetic resonance spectroscopic and growth characteristics within grades II and III gliomas. J Neurosurg 106: 660–666 Pijnappel WWF, van den Boogaart A, de Beer R, van Ormondt D (1992) SVD-based quantification of magnetic resonance signals. J Magn Reson 97:122–134 Prastawa M, Bullitt E, Ho S, Gerig G (2004) A brain tumor segmentation framework based on outlier detection. Med Image Anal 8:275–283 Prastawa M, Bullitt E, Moon N, Van Leemput K, Gerig G (2003) Automatic brain tumor segmentation by subject specific modification of atlas priors. Acad Radiol 10: 1341–1348 Price SJ, Jena R, Burnet NG, Hutchinson PJ, Dean AF, Pena A, Pickard JD, Carpenter TA, Gillard JH (2006) Improved delineation of glioma margins and regions of infiltration with the use of diffusion tensor imaging: an image-guided biopsy study. Am J Neuroradiol 27:1969–1974 Ricci R, Bacci A, Tugnoli V, Battaglia S, Maffei M, Agati R, Leonardi M (2007) Metabolic findings on 3T 1H-MR spectroscopy in peritumoral brain edema. Am J Neuroradiol 28:1287–1291 Simonetti AW, Melssen WJ, Szabo de Edelenyi F, van Asten JJA, Heerschap A, Buydens LMC (2005) Combination of feature-reduced MR spectroscopic and MR imaging data for improved brain tumor classification. NMR Biomed 18:34–43 Simonetti AW, Melssen WJ, van der Graaf M, Heerschap A, Buydens LMC (2002) Automated correction of unwanted phase jumps in reference signals which corrupt MRSI spectra after eddy current correction. J Magn Reson 159:151–157 Simonetti AW, Melssen WJ, van der Graaf M, Heerschap A, Buydens LMC (2003) A new chemometric approach for brain tumor classification using magnetic resonance imaging and spectroscopy. Anal Chem 75:5352–5361

168 Stadlbauer A, Gruber S, Nimsky C, Fahlbusch R, Hammen T, Buslei R, Tomandl B, Moser E, Ganslandt O (2006) Preoperative grading of gliomas by using metabolite quantification with high-spatial-resolution proton MR spectroscopic imaging. Radiology 238:958–969 Suykens JAK, Vandewalle J (1999) Least squares support vector machine classifiers. Neural Process Lett 9(3): 293–300 Szabo de Edelenyi F, Rubin C, Esteve F, Grand S, Decorps M, Lefournier V, Le Bas JF, Remy C (2000) A new approach for analyzing proton magnetic resonance spectroscopic images of brain tumors: nosologic images. Nat Med 6: 1287–1289

J. Luts et al. Van Gestel T, Suykens JAK, Lanckriet G, Lambrechts A, De Moor B, Vandewalle J (2002) Bayesian framework for least squares support vector machine classifiers, Gaussian processes and kernel Fisher discriminant analysis. Neural Comput 14:1115–1147 Warfield SK, Kaus M, Jolesz FA, Kikinis R (2000) Adaptive template moderated spatially varying statistical classification. Med Image Anal 4:43–55 Wu T, Lin C, Weng R (2004) Probability estimates for multiclass classification by pairwise coupling. J Mach Learn Res 5:975–1005

Chapter 17

Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes Kosuke Shimizu and Naoto Oku

Abstract Because brain cancer has a poor prognosis and difficulty of complete removal, development of diagnosis technique to detect the brain cancer in early stage is expected for decreasing the human death from brain cancer. Positron emission tomography (PET) is one of the innovative devices for cancer diagnosis and typical PET probe, [18 F]fluorodeoxyglucose ([18 F]FDG) is applicable to detect the internal organ tumors noninvasively. However, application of the FDG-PET diagnosis for detection of brain cancer is controversial, since brain tissue consequently requires the higher amount of glucose than the other organs. On the other hand, we previously advocated a novel concept for cancer therapy named antineovascular therapy (ANET), in which angiogenic vessels in tumor tissue were targeted and disrupted by utilizing liposomal DDS technology, and demonstrated the therapeutic benefit in the treatment of not only common but also difficult-to-treat cancers. In this chapter, we first introduce the concept of ANET and the application in the treatment of many tumors. Then, we prove the usefulness of a novel positron-labeling method, “solidphase transition” (SophT) method, for PET imaging with liposomes. Finally, we demonstrate the applicability of PET imaging with angiogenic vessel-targeting liposomes for detection of brain cancer. Keywords PET · Liposomes · Angiogenesis · Antineovascular therapy · Peptide · Tumor

N. Oku () Department of Medical Biochemistry, School of Pharmaceutical Sciences, University of Shizuoka, Shizuoka, 422-8526 Japan e-mail: [email protected]

Introduction Tumor Angiogenesis and Targeting Therapy Using Liposomes It is known that the microenvironment surrounding tumor cells is continuously changed for the convenient growth of them. Angiogenesis, an event of construction of new blood vessels, is critical for tumor growth, since the supply of oxygen and nutrients through the blood vessels is essential for tumor cell growth. Angiogenesis is also involved in blood-borne metastasis, since intravasation of tumor cells through these angiogenic vessels is an initial step of tumor metastasis to distal organs (Furuya et al., 2009). Thus, it is known that malignancy of tumor correlates with the development of tumor angiogenesis. A number of previous studies on tumor angiogenesis have been elucidated some pro-angiogenic factors such as vascular endothelial growth factor (VEGF) and the biological process of angiogenesis (Ferrara, 2002). As the results, the specific characteristics of angiogenic vessels have been identified: The structure of angiogenic blood vessels is different from that of the pre-existing blood vessels. Since some cytokines secreted by tumor cells change the morphology of the endothelial cells, the blood vessel permeability is enhanced: Endothelial cells making up angiogenic vessels express some specific molecules on the surface of the cells. These specificities of angiogenic vessels enable the drug nanocarriers such as liposomes to be efficiently delivered to a tumor tissue. Long-circulating liposomes such as polyethyleneglycol (PEG)-modified liposomes passively accumulate in a tumor tissue following the extravasation of liposomes through the permeable

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_17, © Springer Science+Business Media B.V. 2011

169

170

blood vessels and staying there for long period of time. This phenomenon is called “enhanced permeability and retention (EPR) effect” and is a fundamental mechanism of passive targeting of nanocarrier to tumor tissue. On the other hand, liposomes modified with affinity probes against the molecules specifically expressed on angiogenic vessels also enable the effective delivery to tumor tissue, since these activetargeting liposomes specifically target the angiogenic vessels in the tumor. Thus, targeting strategies focusing on the specific structure of tumor angiogenic vessels promise specific accumulation of the liposomes in a tumor.

Strategy of Antiangiogenic Targeting Concept of Antineovascular Therapy Permeability of angiogenic vessels has been enhanced in compared with that of pre-existing vessels, since structure of angiogenic vessels is immature. Thus, macromolecules such as liposomes passively accumulate in tumor tissues as a reflection of the feature. Especially, liposomes possessing long-circulating characteristics are known to accumulate passively in tumor tissue by EPR effect. Therefore, liposomalization of anti-cancer cytotoxic agents could enhance their R cytotoxic activity with decreased side effect. Doxil R and DaunoXome are typical liposomal anti-cancer agents that have been used in clinical chemotherapy. However, cancer therapeutic effect of these liposomal agents often depends on the density of angiogenic vessels in tumors. In the case of low-vascularized tumors such as pancreas tumor, macromolecules could not accumulate effectively because of little EPR effect. In fact, the commercialized anti-cancer liposomes are applied to the treatment of high-vascularized tumors such as breast and ovarian cancer. As described above, phenotype of angiogenic endothelial cells is different from that of pre-existing vessels due to the activation by many pro-angiogenic cytokines. For example, angiogenic endothelial cells stimulated with pro-angiogenic cytokines grow quite rapidly in comparison with endothelial cells of preexisting vessels. Therefore, traditional anti-cancer agents could be potent against angiogenic endothelial cells, since these anti-cancer agents effectively

K. Shimizu and N. Oku

eradicate growing cells. On the other hand, tumor angiogenic vessels are thought to be an ideal target site for a drug delivery system, since drugs or drug nanocarriers injected into bloodstream firstly meet angiogenic vessels prior to extravasate into the tumor tissue. Therefore, targeting to angiogenic endothelial cells may be easier than targeting to specific cells that exist outside of bloodstream. In addition, angiogenic endothelial cells also express specific molecules such as VEGF receptor (VEGFR) on their surface of plasma membrane. Thus, these molecules may provide an active targeting guide for angiogenic vessel targeting of liposomes. Based on these backgrounds, we established novel cancer chemotherapy that directly eradicates angiogenic vessels by allowing cytotoxic agents to deliver to angiogenic endothelial cells with liposomal DDS technology, named antineovascular therapy (ANET). ANET is one of cancer therapies targeting tumor angiogenesis, although the concept is different from conventional antiangiogenic cancer therapy in which small molecules inhibit neovessel formation through inhibiting certain steps of angiogenesis: ANET targets newly formed vessels in tumor and causes lethal damage of the vessels, leading long-term starvation for tumor cells and indirect lethal damages for them. ANET has some advantages in the treatment of cancer; ANET may be useful for not only hypervascular tumors but also hypovascular ones, since they could meet objective cells in bloodstream; ANET could apply to the treatment of drug-resistant tumor, which is severe problem in cancer chemotherapy, since angiogenic endothelial cells would not be expected to acquire drug-resistance. Thus, ANET is likely to be an ideal cancer chemotherapy in which the effect is not dependent on types of tumor.

Isolation of Angiogenic Vessel-Targeting Peptide To deliver nanocarriers to specific organ, some molecules for probing the targeting molecules or tissues are developed, and antibody is one of them. Since antibodies have their quite high specificity and affinity for antigens and/or targeting molecules, antibodies often used as not only therapeutic agents for some diseases but also targeting probes of drug nanocarriers

17 Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes

(Sapra et al., 2005). Peptide also has enough specificity and affinity for the targeting molecules. There are some methods to determine an active targeting probe against the specific molecule and the method of biopanning with phage displaying peptide library is one of them. Among biopanning method, in vivo biopanning is useful to obtain peptides specific for various organs and disease area including tumor tissues (Asai and Oku, 2010). To obtain the peptides specifically homing to angiogenic vessels, we performed in vivo biopanning by using dorsal air sac (DAS) mice, which were one of the model animals to induce in vivo angiogenesis. Therefore, the obtained phage clones express the peptides specific for angiogenic endothelial cells but not for tumor cells. Moreover, in vivo biopanning on DAS model mice eliminates the phage clones having affinity peptides for pre-existing vessels, since the phages passed through the pre-existing vessels during the blood circulation and arrived at angiogenic site. When the selected phage clones were intravenously injected into B16BL6 melanoma-bearing mice, the accumulation of PRPGAPLAGSWPGTS, DRWRPALPVVLFPLH, and ASSSYPLIHWRPWAR peptide-presented phage clones in the tumor tissue was over 20-fold higher than that of the original phage library (Oku et al., 2002). These clones also showed the higher accumulation in solid tumor of Meth A sarcoma-bearing mice. Furthermore, the accumulation of the phage clones was competitively suppressed by pre-injection of the synthetic peptides having the corresponding sequences. As a result of competitive inhibition assay by using fragment peptides of those three kinds of pentadecapeptides with flanking amino acid moiety, we determined epitope sequence as AlaPro-Arg-Pro-Gly (APRPG), which showed the highest affinity for angiogenic vessels among the fragment peptides tested. On the other hand, further experiment of histochemical analysis using biotinylated peptide demonstrated that PRPGAPLAGSWPGTS specifically bound to angiogenic endothelial cells in human islet of the pancreas tumor and glioblastoma. In this case also, pre-treatment with an excess of synthetic PRPGAPLAGSWPGTS inhibited the binding of biotinylated PRPGAPLAGSWPGTS on the glioblastoma specimens (Oku et al., 2002). These data indicated that APRPG-containing peptides had affinity for any molecule specifically expressed in human angiogenic endothelial cells and enabled to apply angiogenic targeting in human tumors.

171

Antineovascular Therapy for Various Tumors To deliver cytotoxic agent to angiogenic endothelial cells, we selected liposome as a drug nanocarrier, and angiogenic vessel-targeting liposomes were prepared by modifying angiogenic vessel-homing peptide, APRPG, on the surface of liposomes following synthesizing phospholipid derivative conjugated with APRPG peptide. To examine the specific delivery of APRPG-Lip to angiogenic vessels, the binding of APRPG-Lip to human umbilical vein endothelial cells (HUVECs) was observed under a confocal laser scanning microscope. Fluorescently labeled APRPGLip significantly bound to VEGF-stimulated HUVECs, whereas little fluorescence was observed after incubation of unmodified liposomes with HUVECs (Oku et al., 2002). Moreover, binding of APRPG-Lip to HUVECs was inhibited in the presence of excess APRPG peptide. Interestingly, the binding of APRPGLip was not observed in the absence of VEGF stimulation to HUVECs. Furthermore, when the biodistribution of radiolabeled APRPG-Lip in tumor-bearing mice was examined, APRPG-Lip significantly accumulated in both Meth A and Colon26 solid tumors in comparison with unmodified liposomes. Moreover, intratumoral distribution of APRPG-Lip in Colon26 solid tumor after intravenous injection to the tumorbearing mice was dominantly observed at vessel like structure in the tumor, and most of them were co-localized with CD31-positive cells (Maeda et al., 2006). In contrast, unmodified liposomes were observed in entire tumor tissue. These results suggested that APRPG-Lip possesses the ability of drug delivery to angiogenic vessels in tumor. Organ-selective targeting of nanocarrier promises the enhancement of encapsulated-drug activity and reduction of the side effects. Since APRPG-Lip showed the significant accumulation in angiogenic vessels, APRPG-Lip is suitable for the application of ANET. We encapsulated traditional anticancer drug, adriamycin (ADM) into the APRPG-Lip (APRPG-LipADM) and tried to apply ANET for many tumors. In therapeutic experiments using Meth A sarcomaand Colon26 carcinoma-bearing mice, injection of APRPG-LipADM resulted in the remarkable suppression of tumor growth with reduced side effect of ADM,

172

and the prolongation of survival time of the mice. Since APRPG-Lip delivers ADM to angiogenic endothelial cells but not to tumor cells and causes lethal damage on tumor cells, ANET would be effective for tumor cells acquired drug resistance. Therefore, we prepared both ADM-sensitive and ADM-resistant P388 leukemia cells and compared the therapeutic effects in each tumor-bearing mice group after injection of free ADM, Cont-LipADM or APRPG-LipADM. In the case of ADM-sensitive P388-bearing mice, the suppression curve of tumor growth in each treatment group was similar to that in Meth A- or Colon26-bearing mice: Treatment of free ADM or Cont-LipADM suppressed tumor growth, although the suppression of tumor growth by APRPG-LipADM was the most effective. In contrast, only APRPG-LipADM treatment was effective to ADM-resistant P388-bearing mice, whereas the treatment of other agents failed to suppress the tumor growth (Shimizu et al., 2005). Actually, the treatment of APRPG-LipADM induced apoptosis of ADMresistant P388 tumor cells, although the apoptosis was rarely observed in the solid tumor treated with ContLipADM. Thus, it is considered that APRPG-LipADM damaged ADM-sensitive angiogenic endothelial cells in ADM-resistant P388 solid tumor and showed indirect damage to the tumor cells, and that the effective delivery of ADM to angiogenic endothelial cells is beneficial in cancer therapy. For the enhancement of circulation time of the liposomes, we endowed angiogenic vessel-targeting liposomes with long circulating characteristics (APRPGPEG-Lip) by conjugating APRPG peptide to the edge of polyethylene glycol (PEG) of PEG-coated liposomes (Maeda et al., 2004). APRPG-PEG-LipADM accumulated more in tumor than PEG-uncoated liposomes after intravenous injection and showed enhanced therapeutic effect by effective delivery of ADM to angiogenic endothelial cells and lethal damage of angiogenic vessels. Then, the application of ANET to the treatment of pancreas cancer was performed. Since pancreatic cancer is known as a hypovascular cancer, delivery of anti-cancer drug to tumor area is not easy and the most of chemotherapy fails to cure pancreatic cancers. To this problem, we hypothesized that dependence of supplies of oxygen and nutrients on blood vessels in pancreas cancer was the same as that in the case of hypervascular cancers, and the lethal damage of small number of blood vessels may affect on large extent of tumor cells. Firstly,

K. Shimizu and N. Oku

orthotropic tumor model with human pancreatic tumor cell line SUIT-2 was prepared, which was confirmed to be similar histology of human pancreatic cancer. Biodistribution study after intravenous injection of APRPG-PEG-Lip or PEG-Lip into SUIT-2-bearing mice indicated similar accumulation of PEG-Lip and APRPG-PEG-Lip in the tumor, although intratumoral distribution was much different: PEG-Lip accumulated around angiogenic vessels, and APRPG-PEG-Lip associated with angiogenic vessels in pancreatic tumor tissue. Moreover, significant suppression of tumor growth by the treatment of APRPG-PEG-LipADM was observed, whereas PEG-LipADM showed little regression of the tumor (Yonezawa et al., 2007). Immunohistochemical analysis also supported the result that vessel-like structures were disappeared in the tumor after treatment with APRPG-PEG-LipADM, suggesting that APRPG-PEG-LipADM degenerated angiogenic vessels in the tumor and induced apoptosis of tumor cells. Similar approaches to ANET using liposomes were investigated by other research groups. Pastorino et al. (2003) reported anti-angiogenic chemotherapy by use of NGR peptide-modified long-circulating liposomes encapsulating doxorubicin (DOX). NGR peptide targets aminopeptidase N on angiogenic endothelial cells. The idea is quite similar to ours. They showed not only drastic therapeutic efficacy against tumor-bearing mice but also pronounced destruction of the tumor vasculature by use of NGR peptide-modified liposomal anti-cancer drug. Schiffelers et al. (2003) used RGD peptide as a targeting tool for integrin αv β3 that expressed on angiogenic endothelial cells. They developed angiogenic vasculature-targeting liposomes by conjugating the RGD peptide with the edge of PEG-coated long circulating liposomes. They demonstrated that RGD-long circulating liposomes specifically bound to endothelial cells in vitro and RGD-long circulating liposomes encapsulating doxorubicin effectively suppressed tumor growth in Colon26 carcinomabearing model mice. We recently proposed novel strategy to enhance the delivery of anti-cancer agent to angiogenic vessels, named “dual-targeting” strategy in which two different kinds of targeting probes are modified on liposomes. In general, a probe molecule for active targeting of drug nanocarrier was individually used. We hypothesized that two different probes on the surface of liposomes enable to enhance the potential of active

17 Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes

targeting, cooperatively. We prepared the targeting liposomes modified with both APRPG and GNGRG peptide, which selectively bind to angiogenic vessels, since targeting molecules of both peptides are predominantly expressed on angiogenic endothelial cells, and examined efficiency of dual-targeting liposomes on ANET. When binding of dual-targeting liposomes to HUVECs was examined, dual-targeting liposomes significantly bound to them in comparison with each single peptide-modified liposome, even if total amount of peptide modification was the same between dual targeting liposomes and single peptide-modified liposomes, suggesting that two different ligands synergistically contributed to association with angiogenic vessels (Murase et al., 2010). This synergistic binding reflected the therapeutic effect of tumor-bearing mice: Dual-targeting liposomes encapsulating ADM remarkably suppressed tumor growth in Colon26-bearing mice in comparison with each single peptide-modified liposomes encapsulating ADM. Thus, dual-targeting strategy could be a potential to enhance the efficiency of ANET. These accumulated data suggest that ANET can apply to the treatment with various kinds of cancers including intractable cancer and be a novel approach for complete cure of them.

Brain Tumor Imaging by Positron Emission Tomography Development of Positron PET Probe for Liposome Labeling Since active targeting of liposomes to angiogenic vessels of a tumor tissue is achieved by using APRPGPEG-modified ones, these liposomes are also useful for tumor imaging when the liposomes are radiolabeled with appropriate radionuclides. For this purpose, we firstly developed a method to radiolabel liposomes. Positron emission tomography (PET) is a noninvasive technology, which has been used for clinical application for tumor diagnosis, as well as for functional evaluation. This technology can also be used to pharmacokinetic studies to monitor the biodistribution and quantification of drugs more quickly and easily than conventional techniques that need to take time and effort with a large number of animals.

173

Application of PET to pharmacokinetic study is also expected in the field of development of new drugs, since elucidation of pharmacokinetics in human body before phase I study can assume the side effects or any problems in following clinical trials, which is often called as “ phase 0 study”. In the development of liposomal DDS drug, noninvasive PET technology is much useful to monitor the biodistribution (Urakami et al., 2007). We previously demonstrated the methodology for detecting noninvasive liposomal trafficking by PET. In those studies, we used a watersoluble [2-18 F]2-deoxy-2-fluoro-glucose ([18 F]FDG) as a positron tracer, which was encapsulated into aqueous phase of liposomes and prepared [18 F]-labeled liposomes. Then, the real-time trafficking of [18 F]labeled liposomes after intravenous injection in tumorbearing mice was monitored by PET. As a result the accumulation of targeted liposomes in the tumor was significantly higher than that untargeted liposomes (Kondo et al., 2004; Maeda et al., 2006). [18 F]FDGlabeling of liposomes, however, has several practical problems: The encapsulation efficiency of [18 F]FDG into liposomes is quite low (<1%), and the preparation method for liposomes encapsulating [18 F]FDG is complicated including reconstruction of the lipid bilayer by repeating freeze-thaw cycles, adjustment of the liposomal size after the free-thawing, and removal of untrapped [18 F]FDG. Thus, this method was not sophisticated in terms of efficiency of radiolabeling and prevention of occupational irradiation. Marik et al. (2007) developed a [18 F]-labeled amphiphilic probe ([18 F]fluorodipalmitin) for determining liposomal distribution by PET, although their method was not applicable for preformed liposomes. To resolve these problems of positron-emitter labeling for liposomes, we synthesized a novel 18 F-labeled amphiphilic compound known as 1-[18 F]fluoro-3,6-dioxatetracosane ([18 F]SteP2), and developed a method to label liposomes with [18 F]SteP2, which is named as “solidphase transition” (SophT) method (Fig. 17.1, Urakami et al., 2007). The advantage of SophT method is that it labels liposomes rapidly with a simple step and is applicable to preformed liposomes with quite high labeling efficiency (>60%). Moreover, [18 F]SteP2 in DSPC-based liposomes is stable in the presence of serum. This universal method for radiolabeling of lipidic nanocarriers can be used for not only liposomes but also polymer micelles (Koide et al., 2008). We applied this positron-emitter-labeling method and

174

K. Shimizu and N. Oku

Fig. 17.1 Solid-phase transition (SophT) method for positron labeling of liposomes. Solvent-free positron emitter probe ([18 F]SteP2) was prepared by semiautomated synthesizing system. Preformed liposomes were added to the tube and agitated gently. The solution was transferred and centrifuged to removed unlabeled probe. During the agitation, the [18 F]SteP2 is transited to lipid phase of liposomes directly

following PET analysis to determine the in vivo behavior of liposome encapsulated hemoglobin (LEH) in brain ischemia model rat (Urakami et al., 2009a). As the result, labeling LEH with [18 F]Step2 provided a high labeling efficiency in a short period of time, and allowed the real-time visualization of LEH in the rat brain by PET. Therefore, this positron-emitter labeling probe and PET scanning may be useful for the imaging of specific accumulation of liposomes in inflammatory or tumor sites.

Brain Tumor Diagnosis by PET with the Targeted Liposomes Since brain cancers such as glioblastoma are known to be poor prognosis and surgical removal of the tumor tissue requires the advanced operation, diagnosis of the brain cancer at the early stages is quite important. PET is one of innovative tools for diagnosis, therapeutic evaluation, and prognostic evaluation of cancer and [18 F]FDG is the widely used as positron emitter for cancer diagnosis by PET. However, the application of FDG-PET diagnosis for detection of brain tumor is controversial, since brain tissue consequently requires the higher amount of glucose than the other organs (Takeda et al., 2003; Chen, 2007). To reduce this bothersome background, various positron-emitter-labeled compounds such as [11 C]choline, [11 C]acetate, [11 C]methionine, L-[methyl-11 C]methionine, O-[18 F]fluoromethylL-tyrosine, O-[18 F]fluoroethyl-L-tyrosine, and

O-[18 F]fluoropropyl-L-tyrosine have been synthesized and evaluated as PET imaging agents for the diagnosis and detection of recurrence of brain cancers. Among them, the amino acid analogues have showed low accumulation in normal peripheral tissue and rapid blood clearance, resulting to use for detecting brain tumors and other tumors as well. We also demonstrated that the D-amino-acid analogue, O-[18 F]fluoromethyl-D-tyrosine was useful for tumor imaging by PET (Urakami et al., 2009b). To detect the brain tumor more specifically and sensitively, we applied positron-emitter-labeled, angiogenic vessel-targeting liposomes for PET scanning of brain tumor detection, since we demonstrated both the universal methodology of positron-emitter labeling for preformed liposomes and the availability of angiogenic vessel-targeting liposomes for the treatment of various tumors. We firstly examined the whole brain distribution of fluorescently labeled APRPG-PEG-Lip after intravenous injection to orthotropic brain tumor model rats by using the in vivo imaging system (IVIS). The result of ex vivo imaging indicated the accumulation of APRPG-PEG-Lip in region of brain tumor was about fourfold higher than that of normal brain tissue (Oku et al., 2011). To determine the accumulation of APRPG-PEG-Lip in brain tumor region, we examined the intratumoral distribution of fluorescence-labeled APRPG-PEG-Lip after intravenous injection by using the brain slice. As shown in Fig. 17.2, the liposomes were accumulated in tumor region and not so much fluorescence was observed in normal brain tissues neighboring to the tumor region.

17 Brain Tumor Diagnosis Using PET with Angiogenic Vessel-Targeting Liposomes

175

Fig. 17.2 Intratumoral distribution of fluorescently labeled liposomes in brain tumor after intravenous injection. Fluorescently labeled APRPG-PEG-Lip was intravenously injected to the rats implanted with C6 glioma cells into the left midbrain. At 1 h after the injection, liposomal distribution

in the brain was observed under a fluorescence microscope. Left picture shows the image stained with hematoxylin and eosin, corresponding to the fluorescent image (right). The dotted lines show the borders between normal and tumor tissue. Deep purple region in H&E staining indicates the tumor tissue

Then, glioma-bearing model rats were injected with angiogenic vessel-targeting liposomes labeled with positron-emitter and PET scanning was performed. Our PET study indicated the specific accumulation of [18 F]SteP2-labeled APRPG-PEG-Lip in the brain tumor region. As shown in Fig. 17.3, PET images obtained by the liposomes were clearer than those obtained with FDG-PET, since the background of

FDG-PET, [18 F]FDG distribution in the normal brain region, was rather high due to the glucose demand in brain. Similar results were observed when we used [18 F]SteP2-labeled non-targeting PEG-Lip for the detection of brain tumor, although the intratumoral distribution was different from that by APRPG-PEGLip. It should be also noted that a tumor with a diameter of only 1 mm could be successfully imaged

Fig. 17.3 PET imaging of brain tumor with [18 F]SteP2APRPG-PEG-Lip and [18 F]FDG. [18 F]SteP2-labeled APRPGPEG-Lip (upper panel) or [18 F]FDG (lower panel) were intravenously injected with brain tumor model rat of C6 glioma. The biodistribution pattern of each sample was determined for 1 h with the Clairvivo PET, and the horizontal (left panel)

and vertical (center panel) PET images averaged from 40 to 60 min are shown. After PET imaging, the brains were sliced into 2-mm sections, and the autoradiograms (right lower panels) were obtained, and the pictures were taken (right upper panels). The ovals formed by the white dotted line show the estimated brain position

176

by using [18 F]-labeled APRPG-PEG-Lip (Oku et al., 2011). In conclusion, we could get specific and distinct PET image by using angiogenic vessel-targeting liposomes labeled with novel [18 F]-labeling method. Advantages of this strategy for detection of brain cancer are as follow: (1) Since the structure of brain blood vessels are tightly closed by BBB function, the extravasation of circulating liposomes to normal brain tissues merely occurs, and this causes extremely low background. (2) The angiogenic vessels in brain tumor are quite permeable similar to other kinds of tumor, resulting long-circulating PEG-liposomes enable to extravasate into tumor tissues. (3) Angiogenic vesseltargeting APRPG peptide can specifically access the liposomes modified with this peptide to the angiogenic vessels. (4) This kind of liposomes is useful for not only imaging of brain tumor but also treatment of the tumor when the liposomes carrying anticancer drugs. These advantages may bring our idea that application of PET imaging with angiogenic vessel-targeting liposomes for detection of brain cancer is promising for clinical use.

References Asai T, Oku N (2010) Angiogenic vessel-targeting DDS by liposomalized oligopeptides. Methods Mol Biol 605:335–347 Chen W (2007) Clinical applications of PET in brain tumors. J Nucl Med 48:1468–1481 Ferrara N (2002) Role of vascular endothelial growth factor in physiologic and pathologic angiogenesis: therapeutic implications. Semin Oncol 29:10–14 Furuya M, Yonemitsu Y, Aoki I (2009) III. Angiogenesis: complexity of tumor vasculature and microenvironment. Curr Pharm Des 15:1854–1867 Koide H, Asai T, Hatanaka K, Urakami T, Ishii T, Kenjo E, Nishihara M, Yokoyama M, Ishida T, Kiwada H, Oku N (2008) Particle size-dependent triggering of accelerated blood clearance phenomenon. Int J Pharm 362:197–200 Kondo M, Asai T, Katanasaka Y, Sadzuka Y, Tsukada H, Ogino K, Taki T, Baba K, Oku N (2004) Anti-neovasucular therapy by liposomal drug targeted to membrane tyre-1 matrix metalloproteinase. Int J Cancer 108:301–306 Maeda N, Miyazawa S, Shimizu K, Asai T, Yonezawa S, Kitazawa S, Namba Y, Tsukada H, Oku N (2006) Enhancement of anticancer activity in antineovasculasr therapy is based on the intratumoral distribution of the active targeting carrier for anticancer drugs. Biol Pharm Bull 29:1936–1940

K. Shimizu and N. Oku Maeda N, Takeuchi Y, Takada M, Namba Y, Oku N (2004) Synthesis of angiogenesis-targeted peptides and hydrophobized polyethylene glycol conjugate. Bioorg Med Chem Lett 14:1015–1017 Marik J, Tartis MS, Zhang H, Fung JY, Kheirolomoom A, Sutcliffe JL, Ferrara KW (2007) Long-circulating liposomes radiolabeled with [18 F]fluorodipalmitin ([18 F]FDP). Nucl Med Biol 34:165–171 Murase Y, Asai T, Katanasaka Y, Sugiyama T, Shimizu K, Maeda N, Oku N (2010) A novel DDS strategy, “dualtargeting”, and it’s application for antineovscular therapy. Cancer Lett 287:165–171 Oku N, Asai T, Watanabe K, Kuromi K, Nagatsuka M, Kurohane K, Kikkawa H, Ogino K, Tanaka M, Ishikawa D, Tsukada H, Momose M, Nakayama J, Taki T (2002) Anti- neovasular therapy using novel peptides homing to angiogenic vessels. Oncogene 21:2662–2669 Oku N, Yamashita M, Katayama Y, Urakami T, Hatanaka K, Shimizu K, Asai T, Tsukada H, Akai S, Kanazawa H (2011) PET imaging of brain cancer with positron emitter-labeled liposomes. Int J Pharm 403:170–177 Pastorino F, Brignole C, Marimpietri D, Cilli M, Gambini C, Ribatti D, Longhi R, Allen TM, Corti A, Ponzoni M (2003) Vascular damage and anti-angiogenic effects of tumor vessel-targeted liposomal chemotherapy. Cancer Res 63: 7400–7409 Sapra P, Tyagi P, Allen TM (2005) Ligand-targeted liposomes for cancer treatment. Curr Drug Deliv 2:369–381 Schiffelers RM, Koning GA, ten Hagen TL, Fens MH, Schraa AJ, Janssen AP, Kok RJ, Molema G, Storm G (2003) Anti-tumor efficacy of tumor vasculature-targeted liposomal doxorubicin. J Control Release 91:115–122 Shimizu K, Asai T, Fuse C, Sadzuka Y, Sonobe T, Ogino K, Taki T, Oku N (2005) Applicability of anti-neovascular therapy to drug-resistant tumor: Suppression of drug-resistant P388 tumor growth with neovessel-targeted liposomal adriamycin. Int J Pharm 296:133–141 Takeda A, Tamano H, Oku N (2003) Alteration of zinc concentrations in the brain implanted with C6 glioma. Brain Res 965:170–173 Urakami T, Akai S, Katayama Y, Harada N, Tsukada H, Oku N (2007) Novel amphiphilic probes for [18 F]-radiolabeling preformed liposomes and determination of liposomal trafficking by positron emission tomography. J Med Chem 50:6454–6457 Urakami T, Kawaguchi AT, Akai S, Hatanaka K, Koide H, Shimizu K, Asai T, Fukumoto D, Harada N, Tsukada H, Oku N (2009a) In vivo distribution of liposome-encapsulated hemoglobin determined by positron emission tomography. Artif Organs 33:164–168 Urakami T, Sakai K, Asai T, Fukumoto D, Tsukada H, Oku N (2009b) Evaluation of O-[18 F]fluoromethyl-D-tyrosine as a radiotracer for tumor imaging with positron emission tomography. Nucl Med Biol 36:295–303 Yonezawa S, Asai T, Oku N (2007) Effective tumor regression by antineovascular therapy in hypovascular orthotopic pancreatic tumor model. J Control Release 118: 303–309

Chapter 18

Frozen Section Evaluation of Central Nervous System Lesions Richard Prayson

Abstract Intraoperative consultation is an important part of the management of patients with both neoplastic and nonneoplastic central nervous system (CNS) lesions which are targeted for surgical biopsy or intervention. The primary goal of intraoperative consultation is to ensure that adequate tissue is available for an accurate final diagnosis. This may, at times, necessitate the need for multiple biopsies, because many lesions which are the target of biopsy are heterogeneous in nature. Effective communication between the pathologist and the neurosurgeon at the time of surgery is an important component of successfully employing this process. Although the role of frozen section is the particular focus of this review, cytologic preparations in some cases have much to offer and can complement frozens sections. A systematic approach to evaluating a frozen section slide is useful. Basic clinical information is helpful in the beginning to frame a differential diagnosis. An assessment for location, if possible, by review of the microscopic section can be useful. Because most of the lesions targeted by frozen sections are neoplastic, the pathologist needs to be aware of flags that might be indicative of a nonneoplastic diagnosis (such things as increased number of macrophages, increased neutrophils, granulomatous inflammation, chronic inflammation, and abnormal blood vessels). Pathologists are cautioned not to overinterpret or overgrade lesions on frozen sections. Neurosurgeons are reminded that frozen section diagnoses represent a preliminary result, and are subject to discrepancy with the final diagnosis for a R. Prayson () Department of Anatomic Pathology, Cleveland Clinic Foundation, CCLCM, Cleveland, OH 44195, USA e-mail: [email protected]

variety of reasons including heterogeneity of lesions being sampled, experience of pathologists, and technical issues impacting on the quality of the frozen section slide. Keywords CNS · Lesions · Intraoperative consultation · Frozen section · WHO · Cytologic preparation

Introduction An intraoperative pathologic consultation is almost always obtained in the setting of a patient who is having a brain biopsy performed. The primary goal is to ensure that the tissue biopsied is adequate to make an accurate diagnosis. Successful intraoperative consultation of nervous system lesions requires good communication between the neurosurgeon and the pathologist at the time of the surgery. This chapter will focus on the use of frozen sections in the intraoperative consultation arena for the evaluation of lesions in the CNS. The role of intraoperative consultation from the neurosurgeon’s as well as the pathologist’s perspective will be reviewed. The relative merits and drawbacks to the utilization of smear/cytologic preparations versus frozen section preparations will be discussed. An approach to the evaluation of a specimen submitted for frozen section, from a pathologist perspective, will be described.

Role of Intraoperative Consultation The primary goal of the intraoperative consultation is to obtain tissue representative of the targeted lesion that will ultimately yield an accurate diagnosis

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_18, © Springer Science+Business Media B.V. 2011

177

178

on permanent sections. To this end, communication between the neurosurgeon and the pathologist is important. Many lesions in the CNS, particularly tumors, may be heterogeneous in nature. For example, a biopsy at the edge of a glioblastoma (WHO grade IV) may sometimes look more like a low grade astrocytoma (WHO grade II). Without communication of the imaging findings to the pathologist and an awareness of the sampling issues that may be associated with evaluation of a frozen section biopsy specimen, a lower grade lesion may be diagnosed and the patient undertreated. In such a scenario, an additional biopsy may be taken from a different area of the tumor in order to establish a diagnosis that is more in keeping with the imaging impression. Again, the goal is to obtain tissue that is a representative of the lesion; the need to take multiple biopsies to reach this goal is not necessarily unusual. Brainard et al. (1997) evaluated 188 stereotactic brain biopsy procedures from 185 patients to assess the diagnostic yield of a biopsy taken at the sterotactic target. In 67% of cases, an accurate diagnosis was obtained at the stereotactic target in the first frozen section submitted for evaluation. In 16% of cases, an accurate diagnosis was made on a subsequent frozen section; and in some cases, up to 4 biopsies were required before an accurate diagnosis could be rendered. Of the 131 neoplastic cases evaluated in this series, 11% of the cases yielded an inaccurate diagnosis related to sampling error; these most commonly included a diagnosis of necrosis in a setting of a glioblastoma or metastatic carcinoma or undergrading a higher grade glioma. Given the sophistication of the current intraoperative navigating systems, pathologists sometimes find it hard to understand why a biopsy taken at the target or center of the lesion may not be diagnostic. It is important to remember that the location of lesions can shift during the procedure due to swelling induced by the surgical procedure itself. The accuracy of the procedure is also somewhat operator-dependent. As mentioned before, the heterogeneity of many lesions including gliomas, lymphomas, and demyelinating lesions may be quite striking, even within a very short distance. Despite this, the diagnostic yield and accuracy of these stereotactic biopsy procedures are quite high (Alesch et al., 1991; Colbassani et al., 1988; diDivitiis et al., 1983; Feiden et al., 1991; Grunert et al., 1994; Kleihues et al., 1984; Taratuto et al., 1991).

R. Prayson

Another important role of the frozen section consultation is guidance of surgical approach and management in a given case. There are situations where a diagnosis made at frozen section may mitigate against a gross total resection of a lesion. For example, diagnosis of primary CNS lymphoma or a demyelinating lesion at frozen section should not result in an attempt at gross total resection of the lesion. Another example when a frozen section diagnosis may dictate surgical approach arises with small spinal cord biopsies of intramedullary lesions, in which the differential diagnosis of an astrocytoma versus ependymoma comes into play. A frozen section diagnosis of ependymoma may result in an attempt to excise the entire lesion, because these tumors are more likely to have a circumscribed border. In contrast, a diagnosis of low grade fibrillary or diffuse astrocytoma, which is usually not curable by excision because it is widely infiltrative, may generate no further tissue. Occasionally, documentation of recurrent or residual, viable tumor at frozen section may precipitate utilization of treatment protocols at the time of surgery (e.g., the use of Gliadel wafers to treat high grade gliomas). Neurosurgeons will also be speaking to families immediately following the surgery and may want to be able to communicate some information to the patient’s family and to the patient regarding a potential diagnosis. In this setting, it is important to remember that the frozen section diagnosis should be considered preliminary or tentative. It is clearly not always accurate, and it may not always coincide with the final diagnosis due to the heterogeneity of many of the lesions that are potential biopsy targets. Frozen sections also lacks the benefit of ancillary studies (such as immunostains or election microscopy), which sometimes are necessary to render certain diagnoses. And although not intended to satisfy a surgeon’s or pathologist’s curiosity, sometimes frozen sections have been employed to do just that. From a pathologist perspective, frozen sections inform the decisions that are made with regard to triaging of the remaining tissue that is not submitted for intraoperative consultation evaluation, as well as deciding what test may need to be employed to further evaluate the lesion. In cases of suspected infection, a recommendation may be made to the surgeon to obtain additional tissue, in a sterile fashion, for microbiologic evaluation such as culture. In a lesion with unusual or uncertain histological features at the time frozen section diagnosis or a lesion which may

18 Frozen Section Evaluation of Central Nervous System Lesions

be potentially problematic, a small amount of tissue may be triaged into glutaraldehyde for possible electron microscopic evaluation. Although the amount of tissue often obtained in small biopsies to evaluate for primary CNS lymphoma is usually not enough for an extensive lymphoma workup by flow cytometry, there may be instances when enough tissue is present and triage of tissue into a culture media (RPMI) may be useful. Many of the molecular techniques that are currently employed in the evaluation of brain tumors utilize FISH (florescence in situ hybridization) or PCR (polymerase chain reaction) methodologies that can utilize formalin-fixed, paraffin-embedded tissues. In the case of suspected prion disease, tissue should be snap frozen for Western blot analysis, which remains the gold standard for diagnosis. There is no indication for frozen section evaluation for suspected prion disease. One cannot make a definitive diagnosis of prion disease on a frozen section and the equipment in the process of obtaining a frozen section can be contaminated; decontamination of a cryostat is quite difficult, if not impossible.

Frozen Section Versus Cytologic Preparations The methodology that a pathologist decides to employ in the intraoperative consultation arena is dependent on a variety of factors. Although this chapter focuses on the frozen section approach, many pathologists use a cytologic preparation either exclusively or in conjunction with the frozen section (Burger and Nelson, 1997; DiStefano et al., 1998; Yachnis, 2002). To generate a squash preparation, a small fragment of tissue is placed between two slides and then with pressure the two slides are drawn apart and immediately placed into 95% alcohol for ∼45–60 sec (Burger and Nelson, 1997). Touch preparations can be useful in the evaluation of certain lesions such as metastatic carcinoma and lymphoma, but are generally not as useful in the evaluation of gliomas. There are a number of factors which come into play in determining which methodology one decides to employ. One advantage of the cytologic preparation is that it is a tissue sparing procedure; in cases where the amount of tissue is very small, this may be a preferred method. It also provides better cytologic

179

detail (Fig. 18.1a). The quality of frozen sections at one’s institution is also an important factor. If experience with performing frozen sections on brain tissue is lacking, the quality of a frozen section slide may be poor, presenting additional challenges for interpretation (Fig. 18.1b, c). Often, what dictates the decision is how the pathologist has been trained to evaluate neuropathologic cases. If one is more comfortable with frozen section, then one is more likely to employ that methodology; if one is a “cytologist” by training and in practice, the cytologic preparation approach is going to be preferred. Sometimes, the decision is based on what one is looking for. Clearly certain pathologies are easier to appreciate on a cytologic preparation. For example, a lesion in which one is anticipating a lot of macrophages (such as a demyelinating lesion), it may be easier to appreciate on a cytologic preparation. In this interpretation of a frozen section, misinterpreting macrophages as glial cells is not an uncommon error. It is certainly fine to employ both methodologies, if there is sufficient tissue. One disadvantage of a smear preparation is that it does not preserve tissue architecture. There are also certain tumors, such as the schwannomas and meningiomas, which may be difficult to smear and may be much easier to handle them on a frozen section. Again, remember the potential heterogeneity of lesions; tissue chosen for cytologic preparation or frozen section may end up looking different than the tissue which was not processed (Fig. 18.1d).

An Approach to Frozen Section A recent review presents an algorithmic approach to handling a frozen section case (KleinschmidtDeMasters et al., 2006). Before one even touches the tissue or looks at the microscopic slide, the specimen usually arrives with some history. The history is important for evaluation in that it helps the pathologist to begin to formulate a differential diagnosis. Critical pieces of information include the age of the patient, location of the lesion, imaging findings, clinical course, and prior history. There are a number of lesions that can look quite similar to one another, and based upon location may result in a different diagnosis or pathologic impression. With receipt of the tissue, an assessment needs to be made with regard to how much tissue is received and how best

180

R. Prayson

Fig. 18.1 (a) Crush preparation of an anaplastic astrocytoma showing atypical appearing astrocytic cells arranged against a fibrillary background. Cytologic preparations often yield a better look at individual cell morphology at the expense of tissue architectural changes which may not be as evident (original 400X). (b) Severe frozen section artifact resulting in a pseudomicrocystic change and a distortion of nuclear morphology is a problem with poorly preformed frozen. The degree of cellularity here might suggest a low grade astrocytoma (original 200X). (c) This is the permanent section from the same tumor the frozen section of which section is in part B. This underscores the extent to which the frozen section artifacts can distort the histomorphology. This tumor was diagnosed as an anaplastic astrocytoma with increased gemistocytes (WHO grade III) (original magnification 200X). (d) The specimen in this case consisted of multiple biopsies from a ring-enhancing lesion. Two side by side tissue

fragments show very different degrees of hypercellularity. The fragment at the bottom has a cellularity more akin to an anaplastic astrocytoma, whereas the mild increased cellularity in the top fragment is more consistent with a low grade astrocytoma. This underscores the degree of heterogeneity that is not uncommonly encountered in many high grade glial neoplasms (original 100X). (e) An infiltrating low grade astrocytoma marked by atypical tumor cells arranged around a neuron (secondary structure of Sherer). Reactive astrocytes typically do not satellite around preexisting structures (original 400X). (f) A section from the cerebellum where the orientation yields an image the shows a molecular layer around the perimeter and a focus of granular cells in the middle. If one is not aware of the location of the biopsy, one may misconstrue this as an inflammatory process or encephalitis; whereas, this simply represents an unusual section of normal tissue (original 100X)

to process it. An approach to generating a cytologic preparation has been previously discussed. Regardless of what methodology is employed (cytologic or frozen section), an attempt should be made to reserve some

tissue for permanent sections i.e., all the tissue should not be submitted for frozen sectioning unless the biopsy is too small to adequately bisect. Tissue should be quickly frozen in order to minimize ice crystal

18 Frozen Section Evaluation of Central Nervous System Lesions

formation, which may make interpreting the biopsy quite difficult. Once a slide is generated, a quick gross inspection of the slide to make sure that the tissue seen on the slide roughly corresponds to the amount of tissue that was submitted for frozen section evaluation is in order. An assessment for freeze artifacts can be easily made. Again, this can severely limit the pathologists ability to render an accurate assessment of cellularity and can result in severe nuclear distortion. When one begins examining the slide, trying to ascertain the location, if possible, should be made. In small biopsies where a high grade tumor has completely obliterated any normal architectural clues or landmarks, this may not be possible. In lower grade processes, one can frequently discern areas of gray matter (normal neurons present) versus white matter (Fig. 18.1e). This may be important in making sure that the tissue in the biopsy is from where the surgeon thinks it is from. If one is trying to make a diagnosis of demyelinating disease, a biopsy from the cortex will not suffice. One also needs to be wary of sites that can mimic tumors. A biopsy from the periventricular zone, for example, may be mildly hypercellular due to an accumulation of subependymal glial cells, and can mimic a low grade glioma. A normal pineal gland can very much look like a low or intermediate grade glioma. A poorly oriented section of cerebellum may show a collection of small granular cells in the middle of the parenchyma, resembling an inflammatory process (Fig. 18.1f). Occasionally, a biopsy may be taken off a target and may represent a normal tissue (although this is a relatively rare occurrence). Sometimes when multiple tissue fragments are present, comparing the fragments to each other can help direct one’s attention to the pathology (for example, one tissue fragment is more cellular than the others) as well as underscore the potential heterogeneity of a lesion. Because the majority of frozen sections involve targeting a lesion which usually represents a neoplasm, the majority of diagnoses will end up being in the neoplastic arena. One needs to be aware of potential clues as to the possibility of a nonneoplastic process. These clues include such findings as increased numbers of macrophages, increased neutrophils, granulomatous inflammation, numerous lymphocytes, and abnormal blood vessels (Figs. 18.2a–f). A review by Prayson and Kleinschmidt-DeMasters (2006) discusses these

181

various findings and what the consideration should be in evaluating such findings. The presence of reactive astrocytes usually indicates that the biopsy is near a lesion. At times, distinction of reactive astrocytosis or gliosis from a low grade glioma can be quite difficult. In gliosis, the reactive astrocytes are evenly distributed and are marked by an increased amount of eosinophilic cytoplasm and generally round to oval nucleus (Fig. 18.3a). Sometimes, areas of gliosis may be marked by increased Rosenthal fibers or eosinophilic granular bodies (Fig. 18.3b). The typical glioma is marked by an unevenly distributed, increased cellularity and cells with high nuclear to cytoplasmic ratio, nuclear hyperchromasia, nuclear irregularity, and nuclear pleomorphism (Fig. 18.3c). Microcalcifications, true microcystic degenerative changes (not frozen section artifact), and satellitosis of cells around preexisting structures, such as neurons or blood vessels, are all features more typically observed in a glioma rather than gliosis (Fig. 18.3d). When faced with a tumor, certain tumor types are more commonly encountered than others. Certain neoplasms, such as metastases and lymphoma, are common to multiple organ systems and are generally more easily recognizable by pathologists who may not be as familiar with neuropathology (Fig. 18.3e, f). Other common tumors, such as glioblastoma and meningioma, generally have fairly recognizable features and are straightforward. The challenge often lies in the limited biopsy, a biopsy distorted by artifact, or a biopsy taken from a less commonly encountered tumor or a variant of tumor. In some cases, it is alright to indicate that a neoplasm is present, to present a differential diagnosis and to defer definitive classification to permanent sections. Although traditionally, a light microscope has been employed for the evaluation of intraoperative consultation cases, recent advancements in telepathology have resulted in the potential use of such systems for the remote transmission of images or scanned slides to another location for interpretation (Oberholzer et al., 1995; Kaplan et al., 2002; Lagwinski and Prayson, 2008). There are a variety of issues related to use of telepathology in the evaluation of neuropathologic frozen sections. In technology which is dependent on someone at the transmitting site to move the microscopic slide or to take images or pictures for transmission, the ability of that person to identify the

182

R. Prayson

Fig. 18.2 The images in this panel represent examples of nonneoplastic conditions that may often be confused with tumor. (a) A patient with multifocal white matter lesions has a biopsy marked by increased number of white matter macrophages and occasional reactive astrocytes. The presence of numerous macrophages should raise the possibility of a nonneoplastic diagnosis (original 200X). (b) Higher magnification appearance of the same lesion biopsied in part A showing an intranuclear Cowdry type A viral inclusion consistent with a diagnosis of progressive multifocal leukoencephalopathy (original 400X). (c) This patient had a previous diagnosis of anaplastic astrocytoma and underwent a course of radiotherapy. This frozen section shows an increased number of macrophages corresponding to necrosis that was likely radiation induced. A focus of

vascular sclerosis and rare atypical appearing tumor cells which show some radiation associated atypia are also present (original 200X). (d) This patient had a ring-enhancing lesion on imaging studies and showed evidence of an organizing abscess, seen here marked by an increased number of predominately chronic inflammatory cells and a spindled fibroblastic cells (original 200X). (e) Focus of necrosis as seen on the left aspect of this field accompanied by a granulomatous inflammation to the right, suggestive of a necrotizing granulomatous process. The presence of granulomatous inflammation is more commonly associated with a nonneoplastic process (original 200X). (f) A proliferation of arterial and venous vessels with intervening brain tissue, consistent with a diagnosis of an arteriovenous malformation (original 100X)

lesional tissue can be an issue. Regardless of which methodology is employed, there often is a substantial increase in the amount of time that is required to systematically evaluate a frozen section slide via telepathology, which makes utilization in real time problematic unless this can be addressed.

Frozen Section Discrepancies Frozen section evaluation is not perfect. There is a well known error and discrepancy rate that is due to a variety of factors. One fairly recent study evaluated

18 Frozen Section Evaluation of Central Nervous System Lesions

183

Fig. 18.3 (a) Reactive astrocytosis with an increased number reactive astrocytes marked by an abundant eosinophilic cytoplasm and eccentrically placed, round to slightly oval nuclei (original 200X). (b) An area of gliosis showing increased numbers of brightly eosinophilic Rosenthal fibers. Rosenthal fibers may be seen in areas of long-standing gliosis but have been associated with a certain low grade neoplasms such as pilocytic astrocytoma (original 200X). (c) Low grade astrocytoma marked by atypical appearing cells with high nuclear to cytoplasmic ratio, nuclear enlargement and nuclear hyperchromasia (original 200X). (d) Microcystic changes in the

background of a mild hypercellular tissue on permanent section are consistent with a low grade glioma. True microcystic changes are usually not a feature of reactive astrocytosis or gliosis (original 100X). (e) Metastatic adenocarcinoma showing a sharp interphase between the tumor at the bottom of the field and the nontumorous tissue with reactive astrocytosis at the top of field. Metastases represent the most common tumor encountered in the central nervous system (original 200X). (f) Frozen section slide showing an increased angiocentric cellularity (on the left) in a primary central nervous system lymphoma (original 200X)

2,156 cases of brain tumor looking at frozen section diagnoses for evidence of discrepancies when compared with final diagnosis for the case (Plesec and Prayson, 2007). A discrepancy rate of 2.7% was found. The most common categories of discrepancy involved errors in the classification of spindle cell lesions (usually confusing fibrous meningiomas with schwannomas), errors in differentiating oligodendrogliomas

from astrocytomas (probably not a critical error at frozen section), discrepancies involving errors in the diagnosis of central nervous system lymphoma, errors in differentiating reactive gliosis from neoplastic processes (gliosis versus glioma), and errors in overgrading tumors. Overgrading tumors may be particularly problematic in that the purpose of the procedure is to obtain representative lesional tissue; overgrading a

184

tumor may result in no further biopsies and ultimately a lower grade than what was reported at the time of frozen section consultation. Another study by Plesec and Prayson (2009) examined the issue of frozen section discrepancy in the evaluation of nonneoplastic CNS samples. Of 303 cases identified in a 10 year period of time at one institution, discrepancies were noted in 39 (12.9%) cases, a rate much higher than that seen with tumors. This underscores the increased difficulty and unfamiliarity of most pathologists with nonneoplastic lesions. The most common types of errors noted were cases in which a nonneoplastic lesion was misdiagnosed as a neoplasm, one nonneoplastic process was misdiagnosed as another nonneoplastic process, and normal tissue was misdiagnosed as abnormal. Most common final diagnoses in the discrepant group included inflammatory lesions, malformations of cortical development or cortical dysplasia, gliosis, and vascular malformations. In conclusion, frozen section intraoperative consultation remains an important part of the evaluation of biopsied CNS lesions. It is important to remember that the primary goal of a frozen section is to ensure that lesional tissue has been obtained, but not necessarily make a definitive diagnosis. It is important that the lines of communication remain open between the neurosurgeon and the pathologist at the time of intraoperative consultation. From a pathologist’s perspective, it is better to undercall or underdiagnose lesions than to overcall or overdiagnose. The pathologist needs to be careful not to succumb to persuasion in diagnosing something that is not clearly there. Admittedly, neuropathology of frozen sections is difficult, requiring knowledge of both normal histology as well as neuropathology.

References Alesch F, Kitz K, Koos WT, Ostertag CB (1991) Diagnostic potential of sterotactic biopsy of midline lesions. Acta Neurochir Suppl 53:33–36 Brainard JA, Prayson RA, Barnett GH (1997) Frozen section evaluation of sterotactic brain biopsies. Diagnostic yield at

R. Prayson the sterotactic target position in 188 cases. Arch Pathol Lab Med 121:481–484 Burger PC, Nelson JS (1997) Sterotactic brain biopsies. Specimen preparation and evaluation. Arch Pathol Lab Med 121:477–480 Colbassani HJ, Nishio S, Sweeney KM, Bakay RAE, Takei Y (1988) CT-assisted sterotactic brain biopsy: value of intraoperative frozen section diagnosis. J Neurol Neurosurg Pschiat 51:332–341 diDivitiis E, Spaziante R, Cappabianca P, Caputi F, Pettinato G, delBassode Caro M (1983) Reliability of sterotactic biopsy: a model to test the value of diagnosis obtained from small fragments of nervous system tumors. Appl Neurophysiol 46:295–303 DiStefano D, Scucchi LF, Cosentino L, Bosman C, Vecchione A (1998) Intraoperative diagnosis of nervous system lesions. Acta Cytologica 42:346–356 Feiden W, Steude U, Bise K, Gundisch O (1991) Accuracy of sterotactic brain tumor biopsy: comparison of the histologic findings in biopsy cylinders and resected tumor tissue. Neurosurg Rev 14:51–56 Grunert P, Ungersback K, Bohl J, Kitz K, Hopf N (1994) Results of 200 intracranial sterotactic biopsies. Neurosurg Rev 17:59–66 Kaplan KJ, Burgess JR, Sandberg GD, Myers CP, Bigott TR, Greenspan RB (2002) Use of robotic telepathology for frozen-section diagnosis: A retrospective trial of a telepathology system for intraoperative consultation. Mod Pathol 15:1197–1204 Kleihues P, Volk B, Anagnostopoulos J, Kiessling M (1984) Morphologic evaluation of sterotactic brain tumour biopsies. Acta Neurochir 33:171–181 Kleinschmidt – DeMasters BK, Prayson RA (2006) An algorithmic approach to the brain biopsy – part I. Arch Pathol Lab Med 130:1630–1638 Lagwinski N, Prayson R (2008) Real – time telepathology for intraoperative neuropathology consultations. Arch Pathol Lab Med 132:1519–1520 Oberholzer M, Fischer HR, Christen H, Gerber S, Brhlmann M, Mihatsch MJ, Gahm T, Famos M, Winkler C, Fehr P, Hosch HJ, Bächtold L (1995) Telepathology: frozen section diagnosis at a distance. Virch Arch 426:3–9 Plesec TP, Prayson RA (2007) Frozen section discrepancy in the evaluation of central nervous system tumors. Arch Pathol Lab Med 131:1532–1540 Plesec TP, Prayson RA (2009) Frozen section discrepancy in the evaluation of nonneoplastic central nervous system samples. Ann Diagn Pathol 13:359–366 Prayson RA, Kleinschmidt-DeMasters BK (2006) An algorithmic approach to the brain biopsy – part II. Arch Pathol Lab Med 130:1639–1648 Taratuto AL, Sevlever G, Piccardo P (1991) Clues and pitfalls in sterotactic biopsy of the nervous system. Arch Pathol Lab Med 115:596–602 Yachnis AT (2002) Intraoperative consultation for nervous system lesions. Semin Diagn Pathol 19:192–206

Chapter 19

Clinical Role of MicroRNAs in Different Brain Tumors Richard Hummel, Jessica Maurer, and Joerg Haier

Abstract MicroRNAs (miRNAs, miRs) are small, non-coding RNA-molecules which regulate gene expression at a posttranscriptional level. Since their first discovery in the early 1990s it has become clear, that they play a crucial role in various physiological and pathological processes. Especially in the context of cancer, miRNAs have been identified as powerful modulators of disease development and progression. In brain tumors, however, this group of molecules has not yet been studied thoroughly, but the first results are very promising and implicate that miRNAs might open a new dimension for the control and treatment of this disease. The present article provides an overview about the current knowledge about miRNAs in the two most common brain tumors in adults (gliomas/glioblastomas) and children (medulloblastomas) with a focus on clinical relevant questions highlighting the importance of these molecules for a potential diagnostic and therapeutic use. Keywords miRNA · Brain tumor · microRNA · Glioblastoma · Medulloblastoma · Drug resistance

Introduction The term “brain tumor” includes a variety of malignant and benign tumors of the brain and the spinal cord, which originate either primarily in the central

J. Haier () Comprehensive Cancer Centre Muenster, International Patient Management, University Hospital Muenster, 48149 Muenster, Germany e-mail: [email protected]

nervous system or represent metastases from other tumors. Unlike most malignancies, brain tumors occur also at a very young age, in fact, they represent one of the most common tumors in children. The worldwide overall incidence of brain tumors is relatively low (3.5 per 100 000) with slightly higher rates in developed countries as for example the USA (6.5 per 100 000). While the overall 5-year survival rate of brain tumors in the United States lies at 35.1%, the mortality of the different subtypes varies widely. The most frequent primary malignant tumor in adults, the glioblastoma (which belongs to the family of the gliomas), presents an overall 5-year survival of only < 5%, whereas the medulloblastoma, which is besides not otherwise specified malignant gliomas the most frequent primary malignant tumor in children, demonstrates an overall 5-year survival of about 60% (for detailed information about the epidemiology of brain tumors please see the homepages of the Central Brain Tumor Registry of the United States (CBTRUS) and the International Agency for Research on Cancer (IARC)). In general, the standard therapies include initial surgery, often followed by radiotherapy and chemotherapy with variable outcome in the different tumor subtypes. On the search for a better understanding of the molecular basis of these tumors and for better diagnostic and therapeutic options, a new class of molecules, the microRNAs, gained increasing attention within the last years. MicroRNAs (miRNAs, miRs) are naturally occurring, small (19 to 25-nucleotides) RNA molecules that regulate gene expression at a post transcriptional level via negative regulation of messenger RNA by an antisense complementarity to an mRNA molecule. After their initial discovery in 1993, this new class of molecules did not attract too much attention in the scientific community, but a few years later miRNAs

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_19, © Springer Science+Business Media B.V. 2011

185

186

became increasingly interesting as epigenetic modulators of gene expression. Nowadays, these molecules are known to be involved in many physiological and pathological processes and the development and progression of various benign and malignant diseases including cancer. The latest release of miRBase (Release 15; April 2010), a database with information about sequence and putative targets of miRNAs, amounted the current number of known miRNAs to a total of 14,197 in different species such as animals, plants and viruses. Alone in Homo sapiens there are 940 miRNAs detected (http://microrna.sanger.ac.uk/) (Griffiths-Jones et al., 2008). In brain tumors, the investigations of the role of miRNAs in the different tumor subtypes are still in its infancy, and most publications refer to the most common tumor type, the glioma (mostly glioblastoma). However, the data available so far indicate that these molecules play a crucial role in tumor pathogenesis in brain tumors as already shown for other tumor types. The authors of this article aim to summarize the clinical relevant data about a potential diagnostic and therapeutic use of miRNAs in brains tumors available so far. For this purpose, we focus on the two most frequent malignant tumor types in adults (glioma/glioblastoma) and in children (medulloblastoma).

Clinical Importance of MicroRNAs in Brain Cancer The involvement of miRNAs in the pathogenesis, the development and the progression of malignant tumors is an exciting and promising new approach in the battle against cancer. In order to develop better therapeutic strategies against malignacies, we need to understand the molecular basis of the disease, and the pathways involved in this process. The discovery of the epigenetic modulation of gene expression via miRNAs within the last years presents an important step in this direction and opens the opportunity to potentially target tumors on a more individual basis. Unfortunately, the knowledge about molecular pathways and functions affected by miRNAs does often not help directly in the daily clinical practice. Clinicians are interested in a practical use of miRNAs and ask therefore different questions as for example: Can miRNAs improve diagnostic or staging investigations of

R. Hummel et al.

malignant tumors? Do miRNAs provide information about prognosis and outcome? Are miRNAs useful tools for the response control of established anticancer therapies? And of course even more important: Can miRNAs help improving the treatment of cancer? Do these molecule affect the outcome of conventional therapies such as irradiation or chemotherapy? The following sections focus on these questions and provide an overview about the “hard clinical facts” available so far for glioma/glioblastoma and medulloblastoma.

Glioblastoma Diagnostic Use of miRNAs The first crucial step for a clinical use of miRNAs as a diagnostic tool in cancer is the evaluation of the miRNA expression profile in tumorous tissue. The identification of a typical miRNA expression signature in the malignant tissue could allow distinguishing between tumorous and normal tissue or between different tumor types or tumor grades and might therefore contribute to the improvement of the diagnostic workup. The next step includes the assessment of correlations between miRNA expression and clinical relevant prognostic factors and outcome such as for example metastasis (which is extremely rare in glioblastomas), disease recurrence or survival. A very popular approach in this context is the miRNAmicroarray analysis, a high-throughput analysis of miRNA expression, usually followed by validation via RT-PCR or Northern Blotting. Several scientific groups used this technique in the last years and reported an altered miRNA expression in glioblastoma (either human tumor samples or cell lines) when compared to non-tumorous controls. Ciafrè et al. (2005) found 9 (miR-10b, miR-130a, miR-221, miR-125b-1, miR-125b-2, miR-9-2, miR-21, miR-25, miR-123) respectively 4 out of 245 miRNAs (miR-128a, miR181c, miR-181a, miR-181b) to be up-/downregulated in human glioblastoma samples, and an analysis of human glioblastoma cell lines showed an upregulation of 9 miRNAs (miR-221, miR-23a, miR-242, miR-24-1, miR-23b, miR-21, miR-222-prec, miR191, miR-220) and a downregulation of 7 miRNAs (miR-181a, miR-181b, miR-128b, miR-197, miR181c, miR-125b-2, miR-125b-1). Chan et al. (2005)

19 Clinical Role of MicroRNAs in Different Brain Tumors

demonstrated 5 (miR-21, miR-138, miR-347, miR291-5 , miR-135) respectively 3 out of 180 miRNAs (miR-198, miR-188, miR-202) to be up- and downregulated in glioblastoma samples. Sasayama et al. (2009) found miR-10b, miR-21, miR-183, miR-92b, miR-106b to be upregulated and miR- 302c∗ , miR-379, miR-329, miR-134, miR-369-3p to be downregulated in human glioblastoma samples. Other studies reported several miRNAs to be significantly deregulated in glioma samples of Chinese patients (including miR34a, -15b, -200a, and -146b) (Xia et al., 2009), or miR29b, miR-125a, and miR-149 to be downregulated in glioblastomas (Cortez et al., 2010). Furthermore, Silber et al. (2008) found in an array study with 192 miRNAs 13 miRNAs (miR-101, miR-128a, miR-132, miR-133a, miR-133b, miR-149, miR-153, miR-154∗ , miR-185, miR-29b, miR-323, miR-328, miR-330) to be downregulated in glioblastoma multiforme, and 3 miRNAs to be upreguated (miR-21, miR-155, miR210). Another only recently published work using an array with 756 miRNAs to profil glioblastoma, anaplastic astrocytoma and normal brain samples identified several differentially regulated miRNAs between these groups (Rao et al., 2010). Regarding the different glioma grades, several authors described a correlation between the expression of various miRNAs and the tumor grade. For example, miR-10b was demonstrated to be highly overexpressed in glioblastomas, and miR10b expression was correlated with the tumor grade and malignancy in astrocytic tumors. Most interestingly, this study provided first evidence that miR-10b expression is correlated with multifocal lesions of malignant glioma (Sasayama et al., 2009). Another study found miR-21 and miR-221 to be overexpressed in different gliomas. But while miR-21 was homogeneously overexpressed in low and high-grade tumors, miRNA-221 overexpression was more evident in highgrade tumors when compared to low-grade gliomas or anaplastic astrocytoma (Conti et al., 2009). Ernst et al. (2010) demonstrated that the expression of several members of miR-17-92 cluster was significantly higher in primary astrocytic gliomas compared to the normal brain and in addition significantly increased with tumor grade progression. miR-128 was demonstrated to be downregulated in low and high grade gliomas when compared to controls, but the expression level of this miRNA was lower in high grade versus low grade tumors (no statistical analysis provided) (Zhang et al., 2009). Furthermore, miR-181a

187

was as well shown to correlate with tumor grade in human glioma samples (Shi et al., 2008), and another study identified several differentially regulated miRNAs between glioblastoma, anaplastic astrocytoma and normal brain samples, which could differentiate glioma grades and normal brain (Rao et al., 2010). Malzkorn et al. (2010) further reported that a set of 14 miRNAs was associated with malignant progression of astrocytic gliomas. A number of additional studies demonstrated that even more miRNAs are deregulated in glioblastomas, and there is now striking evidence that a high number of these miRNAs play a key role in glioblastoma pathogenesis, cell proliferation, apoptosis, cell cycle regulation, invasion or other essential tumor characteristics. Regarding a possible correlation between miRNA expression and clinico-pathologic features of glioblastomas like recurrence of the disease after treatment or survival, there are so far only very few data available. We found only three studies describing an impact of miRNAs on outcome in these patients. On the one hand, Dou et al. (2010) demonstrated that a certain polymorphism in the miR-196a region (miR-196a rs11614913 (CC) genotype) was associated with a decreased risk of glioma in the Chinese population. Kim et al. (2010) performed a genome analysis and found that miR-26a was located within an amplicon at 12q13.3–14.1 which often contained as well CDK4 and CENTG1, two oncogenes. The median survival for glioblastoma patients with tumors harboring the miR-26a amplicon was significantly lower than that of patients lacking this amplicon. The authors concluded that hsa-miR-26a, CDK4, and CENTG1 comprise a functionally integrated oncomir/oncogene DNA cluster which impacts on survival of glioblastoma patients. And finally, Guan et al. (2010) demonstrated a significant difference between the expression levels of 16 miRNAs (miR-196a, miR-15b, miR-105, miR-367, miR-184, miR-196b, miR-363, miR-504, miR-302b, miR-128b, miR-601, miR-21, miR-517c, miR-302d, miR383, miR-135b) in glioblastoma versus anaplastic astrocytoma. Furthermore, classification of malignant gliomas via these 16 miRNAs showed significant correlations with the WHO grade of the tumor. Most importantly, two of these miRNAs, miR-196a and miR-196b, showed a highly statistically significant correlations between the overall survival of 46 malignant glioma patients (anaplastic astrocytomas and glioblastomas) and the expression levels of both

188

miRNAs. A sub-analysis revealed that glioblastoma patients with a lower expression of both, miR-196a and miR-196b (lower than average), showed a better overall survival than patients with high (higher than average) expression of either miRNA. Furthermore, a multivariate analysis demonstrated that a high level of miR-196 was an independent and significant predictor of short overall survival in glioblastoma patients.

Therapeutic Use of miRNAs One of the most promising approaches of a use of miRNAs in therapeutic settings is the modification of well-established and approved anticancer therapies such as chemotherapy or irradiation. This is especially interesting in a tumor entity like the glioblastoma that is highly resistant to various chemotherapeutic agents. And in this context, first encouraging results have been reported about the impact of miRNA modification on different anti-neoplastic drugs. Feng et al. (2010) for example used small interfering RNAs (siRNA), a synthetic form of microRNAs made of short double stranded RNA, against bFGF (Basic fibroblast growth factor) and treated glioblastoma cells with both, siRNA and chemotherapeutics. The found a significant improved effect of the combined treatment compared to a treatment with either the siRNA or chemotherapeutics. Other groups investigated a number of miRNAs which were previously mentioned to be differentially expressed in glioblastomas. MiR-21 seems to play an extraordinary role in this context: Li et al. (2009a) transfected glioblastoma cell lines with miR-21 inhibitors and treated them subsequently with VM-26 (Teniposide, a topoisomerase II inhibitor). They could show that the knock-down of miR-21 led to a dose-dependant reduction in cell survival after VM-26 therapy and found LRRFIP1 (which can inhibit NF-κB activation) to be a direct target of miR21. Another group demonstrated miR-21 inhibitors to cause a decrease of proliferation in PTEN-mutant and PTEN-wild-type human glioblastoma cell lines. More importantly, inhibition of miR-21 increased the cells’ sensitivity to taxol (paclitaxel, an antimicrotubule agent) and improved the effect of this treatment. The authors concluded that this effect might be independent from the PTEN pathway (PTEN is a target of miR-21) and hypothized that the observed effect depends on the inhibition of STAT3 (Ren et al., 2010a).

R. Hummel et al.

In addition, the co-delivery of miR-21 inhibitors and 5-fluorouracil improved significantly the cytotoxicity of 5-FU and led to a dramatic increase of apoptosis and a diminished migration in glioblastoma cells (Ren et al., 2010b). Another study showed that a pretherapeutic miR-21 overexpression in glioblastoma cells leads to a decrease of apoptosis and to an increase of the cell survival after treatment with TMZ (temozolomide, a DNA-methylating agent). As a mechanism, the authors found miR-21 overexpression to be associated with a significant decrease in the Bax/Bcl-2 ratio and caspase 3 activity. (Shi et al., 2010). But not only miR-21 seems to have an effect on sensitivity to anticancer treatment. Mir-451, which itself showed a considerable neurosphere inhibition and a negative effect on cell viability and cell growth, was reported to generate a synergistic effect in inhibiting neurosphere formation in glioblastoma cell lines when applied in combination with Imatinib mesylate (Gal et al., 2008). So far, we found only one article presenting data about miRNA expression analyses in resistant variants of glioblastoma multiforme cell lines. Ujifuku et al. (2010) found 12 (hsa-miR-455-3p, hsa-miR195, hsa-miR-10a∗ , hsa-miR-502-3p, hsa-miR-193b∗ , hsa-miR-584, hcmv-miR-US25-2-5p, hsa-miR-500∗ , hsa-miR-193a-5p, hsa-miR-452, hsa-miR-132, hsamiR-503) respectively 2 miRNAs (hsa-miR-106b∗ , hsa-miR-210) to be up-/downregulated in TMZ resistant cells. While the inhibition of miR-455-3p and miR-10a∗ in these cells did not affect the cell survival, miR-195 knockdown led to a moderate growth inhibition. More importantly, a combined treatment with TMZ and miR-455-3p or miR-10a∗ inhibitors showed a modest negative effect on cell survival, the combination of miR-195 inhibition with TMZ on the other hand enhanced cell death strongly. The authors provided a number of promising targets for the described miRNAs by in silico identification and cDNA microarray analysis.

Medulloblastoma Compared to glioma/glioblastoma, only few studies investigated so far the role of miRNAs in this tumor type. But these first studies present already very encouraging results regarding an enormous impact of miRNAs in this tumor type.

19 Clinical Role of MicroRNAs in Different Brain Tumors

Diagnostic Use of miRNAs Again, first microarray studies revealed that a large number of miRNAs is differentially expressed in medulloblastoma cell lines or human tumor samples when compared to controls. Ferretti et al. (2008) demonstrated that 30 (Let-7a, let-7e, let-7f, miR-7, miR-9, miR-25, miR-30b, miR-100, miR-103, miR124a, miR-125b, miR-132, miR-135a, miR-135b, miR-142-5p, miR-143, miR-150, miR-153, miR-181c, miR-190, miR-191, miR-203, miR-324-3p, miR-3245p, miR-326, miR-331, miR-338, miR-425) out of 250 miRNAs were downregulated in a subset of human primary medulloblastomas with high Hedgehog (Hh) signaling strength. The Hedgehog signaling pathway is believed to play an important role in the malignant transformation into medulloblastoma. A subsequent publication of the same group only 1 year later described the results of a high throughput profiling of miRNA expression in a large sample set of human medulloblastomas. 78 miRNAs out of selected 86 miRNAs (which had been reported to be expressed in neuronal tissues or to be associated with different tumors (onco-miRs)) were demonstrated to be differentially expressed between tumors and either adult or fetal controls. Interestingly, most of these miRNAs were downregulated in medulloblastomas, what leads to the suggestion that miRNAs might have a predominantly tumor-suppressive function in this tumor entity. In addition, the authors could show that a subset of four miRNAs (let7g, miR-19a, miR-106b and miR-191) allowed a classification of medulloblastoma samples into three different histotypes (anaplastic, classic and desmoplastic), and that groups of miRNAs showed different expression between tumors overexpressiong for example ErbB2 (miR-10b, miR-135a, miR-135b, miR-125b, miR-153, miR-199b) or c-Myc (miR-181b, miR-128a, miR128b) and corresponding “not-overexpressing” controls (Ferretti et al., 2009). Uziel et al. (2009) reported that 26 miRNAs (including 9 miRNAs that were encoded by the miR-17/92 cluster and its paralogs, and the miR-106b/25 paralog clusters) showed an increased expression in mouse medulloblastoma, and another 24 miRNAs (including miR-124a, miR-128, miR-138, miR-300, miR-381, miR-487b, miR-382, miR-433, miR-127, miR-434 and miR-136) showed decreased expression. Interestingly, investigations on

189

human medulloblastoma samples demonstrated that 3 miR-17/92-cluster-miRNAs (namely miR-92, miR19a, and miR-20) were also overexpressed in human medulloblastomas with a constitutively activated Sonic Hedgehog (SHH) signaling pathway. These results were widely confirmed by another microarray study with 427 miRNAs that showed that components of miR-17/92 polycistron, including miR-18a, miR19b, and miR-20a, as well as the paralogous miR106a (miR-106a/363 cluster) were overexpressed in human and murine medulloblastomas. Interestingly, Expression of miR-17/92 was highest in the subgroup of medulloblastomas associated with activation of the sonic hedgehog (Shh) signaling pathway when compared with other subgroups of medulloblastoma (Northcott et al., 2009). Liu et al. (2009) reported that a microRNA expression signature allows a discrimination between medulloblastoma and controls as well, they found 4 miRNAs (miR-17, miR-99a, miR100, miR-106b) to be overexpressed and 5 miRNAs (miR-218, miR-29a, miR-29c, miR-128a, miR-1273p) to be underexpressed in tumors. They validated their microarray data for miR-17, miR-100, miR-106b, and miR-218. Notably, most of the predicted target genes of these validated miRNAs had been shown to be involved in medulloblastoma carcinogenesis. Another just recently published study found 30 miRNA to be consistently and statistically significant downregulated in all examinated medulloblastoma samples. Some of these 30 miRNAs were already mentioned in previous studies (miR-124, miR-129, miR-138, miR-150 and miR-323 were mentioned by Ferretti et al. (2009) and Northcott et al. (2009), another 12 respectively 3 were only mentioned in one of those studies), but 10 miRNAs (let-7 g, miR-9, miR-124, miR-125, miR-128, miR-139, miR-181, miR-194, miR-324, miR-32) had not been reported so far. The authors validated some of these new miRNAs and proofed a downregulation for miR-125, miR-128a, miR -139 and let-7 g as well as an upregulation for miR-17-5p (Venkataraman et al., 2010). Several additional articles focused on specific miRNAs and investigated their expression pattern in medulloblastoma. MiR-124 or miR-129 for example were shown to be consistently downregulated in various medulloblsatoma cell lines and/or patient samples when compared to normal controls (Pierson et al., 2008; Li et al., 2009b; Wu et al., 2010). Another

190

very interesting work described the search for novel recurrent regions of genomic amplification in medulloblastoma cell lines, and the authors identified one of these regions at 8q24.22–q24.23. When looking for critical disease genes in that region they found the encoding genes for two miRNAs, miR-30b and miR-30d, harbored in this area, And indeed, it could be demonstrated that miR-30b respectively miR-30d were overexpressed in 54% respectively 12% of human tumor samples (Lu et al., 2009). Only two groups took the investigation on a possible diagnostic use of miRNAs in medulloblastomas further and looked at clinical relevant outcomes. Ferretti et al. (2009) identified the expression of miR31 and miR-153 to be inversely correlated with the disease risk, and Garzia et al. (2009) found that patients with non-metastatic medulloblastoma presented a significant higher expression of miR-199b-5p than patients with metastatic disease. Furthermore, there was a (not significant) trend towards a better overall survival in the group of the high-expressing patients.

Therapeutic Use of miRNAs So far, there are only very limited data regarding a possible therapeutical use of miRNAs in this tumor type available. Several groups investigated the effects of an ectopic modification of selected miRNAs in vitro. In general, a number of miRNAs had been proven to impact on tumor cell proliferation and growth. For example, miR-9 and miR-125a were shown to promote growth arrest and apoptosis in medulloblastoma by targeting the truncated isoform of the neurotropin receptor TrkC (t-TrkC) (Ferretti et al., 2009). Or an overexpression of miR-125b, miR-324-5p or miR-326 was reported to inhibit meduloblastoma cell growth in vitro by targeting Hedgehog signalling pathway (Ferretti et al., 2008). Unfortunately, none of the mentioned studies assessed so far a possible impact of miRNA modification on other well-established anticancer strategies such as chemotherapies or irradiation. But as there is a convincing body of evidence that miRNAs can affect these therapies in a variety of other cancer types and especially in other brain cancers (see above), further investigations have to show if these molecules can affect outcome of these treatments in medulloblastoma patients as well.

R. Hummel et al.

Summary MicroRNAs are an exciting class of small, non-coding RNA-molecules which regulate gene expression at a posttranscriptional level. As it has been shown in a variety of other tumor entities, these molecules have an enormous impact on cancer pathogenesis. In the context of brain tumors, we presented with this article an overview about the meaning of miRNAs in this tumor type with a clear focus on clinical questions. And despite the fact that there is so far only very limited data available, the already existing knowledge about miRNAs in gliomas/glioblastomas and medulloblastomas, the two most common primary brain tumors in adults and children, is highly encouraging. We demonstrated that miRNA expression profiles can help distinguishing tumors from normal brain tissue in both histological entities and even separate between different tumor grades. Furthermore, there is first evidence that the expression of different miRNAs correlates with clinical outcome in glioblastoma and medulloblastoma. And finally and most importantly, there is a hope that miRNAs might help to overcome resistance to conventional anticancer strategies in glioblastoma. Taken together, these results highlight the enormous potential of this class of molecules in brain tumors, and support the need for further intensive investigations in the field.

References Chan JA, Krichevsky AM, Kosik KS (2005) MicroRNA-21 is an antiapoptotic factor in human glioblastoma cells. Cancer Res 65:6029–6033 Ciafrè SA, Galardi S, Mangiola A, Ferracin M, Liu CG, Sabatino G, Negrini M, Maira G, Croce CM, Farace MG (2005) Extensive modulation of a set of microRNAs in primary glioblastoma. Biochem Biophys Res Commun 334: 1351–1358 Conti A, Aguennouz M, La Torre D, Tomasello C, Cardali S, Angileri FF, Maio F, Cama A, Germanò A, Vita G, Tomasello F (2009) miR-21 and 221 upregulation and miR-181b downregulation in human grade II-IV astrocytic tumors. J Neurooncol 93:325–332 Cortez MA, Nicoloso MS, Shimizu M, Rossi S, Gopisetty G, Molina JR, Carlotti C Jr., Tirapelli D, Neder L, Brassesco MS, Scrideli CA, Tone LG, Georgescu MM, Zhang W, Puduvalli V, Calin GA (2010) miR-29b and miR-125a regulate podoplanin and suppress invasion in glioblastoma. Genes Chromosomes Cancer 49:981–990

19 Clinical Role of MicroRNAs in Different Brain Tumors Dou T, Wu Q, Chen X, Ribas J, Ni X, Tang C, Huang F, Zhou L, Lu D (2010) A polymorphism of microRNA196a genome region was associated with decreased risk of glioma in Chinese population. J Cancer Res Clin Oncol 136:1853–1859 Ernst A, Campos B, Meier J, Devens F, Liesenberg F, Wolter M, Reifenberger G, Herold-Mende C, Lichter P, Radlwimmer B (2010) De-repression of CTGF via the miR-17-92 cluster upon differentiation of human glioblastoma spheroid cultures. Oncogene 29:3411–3422 Feng X, Zhang B, Wang J, Xu X, Lin N, Liu H (2010) Adenovirus-mediated transfer of siRNA against basic fibroblast growth factor mRNA enhances the sensitivity of glioblastoma cells to chemotherapy. Med Oncol 28:24–30 Ferretti E, De Smaele E, Miele E, Laneve P, Po A, Pelloni M, Paganelli A, Di Marcotullio L, Caffarelli E, Screpanti I, Bozzoni I, Gulino A (2008) Concerted microRNA control of Hedgehog signalling in cerebellar neuronal progenitor and tumour cells. EMBO J 27:2616–2627 Ferretti E, De Smaele E, Po A, Di Marcotullio L, Tosi E, Espinola MS, Di Rocco C, Riccardi R, Giangaspero F, Farcomeni A, Nofroni I, Laneve P, Gioia U, Caffarelli E, Bozzoni I, Screpanti I, Gulino A (2009) MicroRNA profiling in human medulloblastoma. Int J Cancer 124:568–577 Gal H, Pandi G, Kanner AA, Ram Z, Lithwick-Yanai G, Amariglio N, Rechavi G, Givol D (2008) MIR-451 and Imatinib mesylate inhibit tumor growth of Glioblastoma stem cells. Biochem Biophys Res Commun 376:86–90 Garzia L, Andolfo I, Cusanelli E, Marino N, Petrosino G, De Martino D, Esposito V, Galeone A, Navas L, Esposito S, Gargiulo S, Fattet S, Donofrio V, Cinalli G, Brunetti A, Vecchio LD, Northcott PA, Delattre O, Taylor MD, Iolascon A, Zollo M (2009) MicroRNA-199b-5p impairs cancer stem cells through negative regulation of HES1 in medulloblastoma. PLoS One 4:e4998 Griffiths-Jones S, Saini HK, van Dongen S, Enright AJ (2008) MiRBase: tools for microRNA genomics. Nucleic Acids Res 36:D154–158 Guan Y, Mizoguchi M, Yoshimoto K, Hata N, Shono T, Suzuki SO, Araki Y, Kuga D, Nakamizo A, Amano T, Ma X, Hayashi K, Sasaki T (2010) MiRNA-196 is upregulated in glioblastoma but not in anaplastic astrocytoma and has prognostic significance. Clin Cancer Res 16:4289–4297 Kim H, Huang W, Jiang X, Pennicooke B, Park PJ, Johnson MD (2010) Integrative genome analysis reveals an oncomir/oncogene cluster regulating glioblastoma survivorship. Proc Natl Acad Sci USA 107:2183–2188 Li Y, Li W, Yang Y, Lu Y, He C, H,u G, Liu H, Chen J, He J, Yu H (2009a) MicroRNA-21 targets LRRFIP1 and contributes to VM-26 resistance in glioblastoma multiforme. Brain Res 1286:13–18 Li KK, Pang JC, Ching AK, Wong CK, Kong X, Wang Y, Zhou L, Chen Z, Ng HK (2009b) miR-124 is frequently downregulated in medulloblastoma and is a negative regulator of SLC16A1. Hum Pathol 40:1234–1243 Liu W, Gong YH, Chao TF, Peng XZ, Yuan JG, Ma ZY, Jia G, Zhao JZ (2009) Identification of differentially expressed microRNAs by microarray: a possible role for microRNAs gene in medulloblastomas. Chin Med J 122:2405–2411 Lu Y, Ryan SL, Elliott DJ, Bignell GR, Futreal PA, Ellison DW, Bailey S, Clifford SC (2009) Amplification

191 and overexpression of Hsa-miR-30b, Hsa-miR-30d and KHDRBS3 at 8q24.22-q24.23 in medulloblastoma. PLoS One 4:e6159 Malzkorn B, Wolter M, Liesenberg F, Grzendowski M, Stühler K, Meyer HE, Reifenberger G (2010) Identification and functional characterization of microRNAs involved in the malignant progression of gliomas. Brain Pathol 20: 539–550 Northcott PA, Fernandez-L A, Hagan JP, Ellison DW, Grajkowska W, Gillespie Y, Grundy R, Van Meter T, Rutka JT, Croce CM, Kenney AM, Taylor MD (2009) The miR17/92 polycistron is up-regulated in sonic hedgehog-driven medulloblastomas and induced by N-myc in sonic hedgehogtreated cerebellar neural precursors. Cancer Res 69: 3249–3255 Pierson J, Hostager B, Fan R, Vibhakar R (2008) Regulation of cyclin dependent kinase 6 by microRNA 124 in medulloblastoma. J Neurooncol 90:1–7 Rao SA, Santosh V, Somasundaram K (2010) Genome-wide expression profiling identifies deregulated miRNAs in malignant astrocytoma. Mod Pathol 23:1404–1417 Ren Y, Kang CS, Yuan XB, Zhou X, Xu P, Han L, Wang GX, Jia Z, Zhong Y, Yu S, Sheng J, Pu PY (2010b) Co-delivery of as-miR-21 and 5-FU by poly(amidoamine) dendrimer attenuates human glioma cell growth in vitro. J Biomater Sci Polym Ed 21:303–314 Ren Y, Zhou X, Mei M, Yuan XB, Han L, Wang GX, Jia ZF, Xu P, Pu PY, Kang CS (2010a) MicroRNA-21 inhibitor sensitizes human glioblastoma cells U251 (PTEN-mutant) and LN229 (PTEN-wild type) to taxol. BMC Cancer 10:27 Sasayama T, Nishihara M, Kondoh T, Hosoda K, Kohmura E (2009) MicroRNA-10b is overexpressed in malignant glioma and associated with tumor invasive factors, uPAR and RhoC. Int J Cancer 125:1407–1413 Shi L, Chen J, Yang J, Pan T, Zhang S, Wang Z (2010) MiR-21 protected human glioblastoma U87MG cells from chemotherapeutic drug temozolomide induced apoptosis by decreasing Bax/Bcl-2 ratio and caspase-3 activity. Brain Res 1352:255–264 Shi L, Cheng Z, Zhang J, Li R, Zhao P, Fu Z, You Y (2008) hsamir-181a and hsa-mir-181b function as tumor suppressors in human glioma cells. Brain Res 1236:185–193 Silber J, Lim DA, Petritsch C, Persson AI, Maunakea AK, Yu M, Vandenberg SR, Ginzinger DG, James CD, Costello JF, Bergers G, Weiss WA, Alvarez-Buylla A, Hodgson JG (2008) miR-124 and miR-137 inhibit proliferation of glioblastoma multiforme cells and induce differentiation of brain tumor stem cells. BMC Med 6:14 Ujifuku K, Mitsutake N, Takakura S, Matsuse M, Saenko V, Suzuki K, Hayashi K, Matsuo T, Kamada K, Nagata I, Yamashita S (2010) miR-195, miR-455-3p and miR10a(∗ ) are implicated in acquired temozolomide resistance in glioblastoma multiforme cells. Cancer Lett 296: 241–248 Uziel T, Karginov FV, Xie S, Parker JS, Wang YD, Gajjar A, He L, Ellison D, Gilbertson RJ, Hannon G, Roussel MF (2009) The miR-17∼92 cluster collaborates with the Sonic Hedgehog pathway in medulloblastoma. Proc Natl Acad Sci USA 106:2812–2817 Venkataraman S, Alimova I, Fan R, Harris P, Foreman N, Vibhakar R (2010) MicroRNA 128a increases intracellular

192 ROS level by targeting Bmi-1 and inhibits medulloblastoma cancer cell growth by promoting senescence. PLoS One 5:e10748 Wu J, Qian J, Li C, Kwok L, Cheng F, Liu P, Perdomo C, Kotton D, Vaziri C, Anderlind C, Spira A, Cardoso WV, Lü J (2010) miR-129 regulates cell proliferation by downregulating Cdk6 expression. Cell Cycle 9: 1809–1818

R. Hummel et al. Xia H, Qi Y, Ng SS, Chen X, Li D, Chen S, Ge R, Jiang S, Li G, Chen Y, He ML, Kung HF, Lai L, Lin MC (2009) microRNA146b inhibits glioma cell migration and invasion by targeting MMPs. Brain Res 1269:158–165 Zhang Y, Chao T, Li R, Liu W, Chen Y, Yan X, Gong Y, Yin B, Liu W, Qiang B, Zhao J, Yuan J, Peng X (2009) MicroRNA-128 inhibits glioma cells proliferation by targeting transcription factor E2F3a. J Mol Med 87:43–51

Part III

Therapy

Chapter 20

Electrochemotherapy for Primary and Secondary Brain Tumors Mette Linnert, Birgit Agerholm-Larsen, Faisal Mahmood, Helle K. Iversen, and Julie Gehl

Abstract There is an increasing clinical challenge in the treatment of primary and secondary brain tumors, mostly because of a rise in the number of patients diagnosed with brain metastases and the limited treatment results with the standard treatments available today. A novel and promising treatment modality, electrochemotherapy, which is a combination of the technique of electroporation and the chemotherapeutic drug bleomycin, could be a new treatment option for these patients. This chapter elucidates the background and experience with electrochemotherapy and looks into the possibilities for use in the treatment of primary and secondary brain tumors. Keywords Electrochemotherapy · Metastases · Electroporation · Chemotherapy · Bleomycin · Brain

Introduction In this chapter we will outline the clinical challenges regarding brain metastases, and introduce a novel and upcoming treatment in this field: electrochemotherapy.

Brain Metastases Metastasis to the brain is an increasing problem for cancer patients today, and Smedby et al. (2009) found a doubling of hospital admissions for brain metastases

J. Gehl () Department of Oncology, Copenhagen University Hospital Herlev, 2730 Herlev, Denmark e-mail: [email protected]

from 1987 to 2006. One obvious reason is the generally growing proportion of elderly people in the population leading to an increase in the cancer incidence overall, as the risk of cancer increases with age. Another reason is the overall improvement in cancer treatments today, resulting in longer survival. This leads to an increased risk of patients developing brain metastases during the course of the disease. Also, there are more sensitive diagnostic methods available today such as MRI that detect multiple brain metastases in 75% versus 50% of the cases diagnosed with a CT scan (Kuhn et al., 1994), leading to increased detection of the presence of brain metastases. Finally, there is the blood brain barrier, which may limit penetration of anti-cancer drugs, leading to a limited effect in the treatment of brain metastases in both the adjuvant and the palliative setting.

Electrochemotherapy Patients with brain metastases have an unfavourable prognosis, and there is need for more efficient treatment options. Electrochemotherapy could be a new option in this setting. Electrochemotherapy is a treatment based on the method of electroporation. Electroporation is a technique that permeabilizes the cell membrane using electric pulses (see section “Electroporation”). The permeabilized state of the membrane can be exploited for various purposes, for example to gain access to the cytosol for a chemotherapeutic drug or even DNA (Fig. 20.1). Electrochemotherapy is a treatment that until now primarily has been used in the treatment of cutaneous metastases in the palliative treatment setting.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_20, © Springer Science+Business Media B.V. 2011

195

196 Fig. 20.1 Electroporation. a By application of brief electric pulses it is possible to transiently permeabilise the cell membrane (electroporation). This allows free passage of molecules to the cell cytosol over a matter of minutes. Bleomycin cytotoxicity is augmented over 300 fold when administration is combined with delivery of electric pulses. b As the electric pulses are subsequently applied, cells are permeabilised and the drug enters. After a few minutes, cells reseal and the extracellular drug is washed out while the internalised molecules remain trapped intracellularly. c Bleomycin is cytotoxic once in the cytosol and it works by cleaving DNA strands, quantified as 10–15 DNA strand breaks per DNA molecule. Bleomycin induces single-strand and double-strand DNA breaks with a ratio of 10:1

M. Linnert et al.

a

b

c

20 Electrochemotherapy for Primary and Secondary Brain Tumors

Treatment response for smaller cutaneous metastases is reported to be consistently high, with complete response rates (CR) after only one treatment of for example 73% (Marty et al., 2006) and 91% of the cases (Heller et al., 1998). So even if electrochemotherapy is only used in the clinic as a palliative treatment, it is in fact eliminating most of the smaller tumors locally. Experience with larger tumors is underway, and an ongoing clinical study of electrochemotherapy as a palliative treatment of large chest wall recurrences from breast cancer is also showing promising results (Matthiessen et al., 2009). To perform electrochemotherapy, it is necessary to apply electric pulses to the tissue via electrodes. The electrodes available up until now have primarily been designed to be used on skin (Fig. 20.2). Different

a

197

groups are now working at developing electrodes for use in internal organs, and we have in corporation with a medical device company developed an electrode suitable for brain tissue (Gehl and Videbæk, 2009; Mahmood and Gehl, 2011). Indeed, for the initial definition of treatment parameters, important data has been obtained by treating cutaneous metastases (Fig. 20.3). This chapter reviews the basis for electrochemotherapy as well as preliminary results on its use in the brain and lists perspectives for the technology.

Clinical Challenges of Brain Metastases It is a paradox that general improvement of cancer treatment today is both leading to prolonged survival of

b

c

Fig. 20.2 The electrochemotherapy procedure. a The cliniporator equipment allows monitoring of voltage and current during the pulse. b The application of pulses to skin tumors must be preceeded by local or general anestesia. c A treatment situation is

shown where a patient is receiving local injection of bleomycin followed by application of pulses under local anesthesia. The application of pulses lasts only a few minutes in total

198

M. Linnert et al.

b

a

c

Fig. 20.3 Treatment result of electrochemotherapy in the skin. Course of treatment for patient with malignant melanoma metastases, one of eight treated is shown. The patient was treated with electrochemotherapy in general anaesthesia and intravenous injection of bleomycin. Pictures shows (a) Before treatment the metastases was ulcerated and caused haemorrhage, pain and discomfort, (b) 1 month after treatment the

lesion is covered with a crust, needle marks in normal tissue are visible due to treatment of the tumor margin as well. Note that there is no necrosis of normal skin, and (c) 6 months after treatment the treated metastases is in CR (complete response) showing normal skin that had healed underneath the nodule. From Gehl, Ugeskrift for laeger, 2005, with permission

cancer patients and to an increased number of patients who live long enough to develop brain metastases. The risk of developing brain metastases increases in breast cancer patients, when there are metastases to the liver, lungs and lymph nodes (Ryberg et al., 2005), i.e., with the advancing stage of the cancer. The treatment of brain metastases are, for single or a limited number of metastases, surgery or stereotactic radiosurgery. For multiple brain metastases the evidence based treatment is whole brain radiation therapy (WBRT) and in some cases systemic chemotherapy or other anti-cancer drugs. Still, the overall prognosis is poor with a reported median survival of 10–12 months after treatment with surgery or stereotactic radiosurgery and a reported median survival of 4–6 months after WBRT (Brennum et al., 2002; Videtic et al., 2009). The patient’s quality of life is related to their neurocognitive function, which is the ability to do daily

activities (ADL) such as recognizing safe and unsafe behaviour, memory and compliance with medical treatments. It has been established, that tumor regression of brain metastases after whole brain radiation therapy (WBRT) is correlated with an increased survival and preservation of neurocognitive function (Li et al., 2007). Another study concluded that the tumor volume before treatment was the only predictor of decreased neurocognitive function (Meyers et al., 2004). Some patients have a better prognosis than others, and several positive prognostic factors have been found, including good performance status, age > 60 years; less than three brain metastases and no extracranial metastases (Sperduto et al., 2008). Some cancer histologies also have a better prognosis, for example, in a study of 43 patients with metastatic breast cancer the median survival was 23 months after the diagnosis of brain metastases (Gori et al., 2007).

20 Electrochemotherapy for Primary and Secondary Brain Tumors

Still, the effect of the available treatments is limited, and when the brain metastases progresses after whole brain radiation therapy, there are no additional standard treatments left to offer for a large number of the patients. WBRT is not advisedly repeated because only a few patients show neurological improvement after retreatment and, actually, most patients deteriorate during or soon following retreatment, resulting in a reported median survival of only 8 weeks (Hazuka and Kinzie, 1988). Additionally, 2 out of 42 patients in this trial most likely died as a direct consequence of the re-irradiation treatment (Hazuka and Kinzie, 1988). Previously, cancer patients died of general disease progression, but today it is not uncommon that the only site of progression is the brain, while the cancer elsewhere in the body is controlled. For example, in the previously mentioned study of 43 breast cancer patients, 60% of the patients only had disease progression in the brain (Gori et al., 2007). It is therefore increasingly important to prevent brain metastases and explore better treatment options for these patients.

Electroporation Electroporation, sometimes referred to as electropermeabilization, is basically a technique used to transfer exogenous molecules into cells, by applying a voltage difference across the target tissue. The technique stems from observations made in the 1960s and 1970s showing that cell membranes are made transiently more permeable by the action of voltage pulses thus allowing migration of hydrophilic molecules and ions from the extracellular space to the cell cytosol and eventually to the cell nucleus (Gehl, 2003). With electroporation a substantial enhancement of transport across cellular membranes can be achieved for certain molecules.

Electroporation and the Transmembrane Potential The biophysical basis of electroporation is often introduced in terms of the evoked change in the electrical potential across the membrane the transmembrane potential. The resting transmembrane potential of the cell is mostly between –90 and –70 mV. However, if the cell is exposed to an external electric field, the

199

transmembrane potential is altered accordingly. The exact relation between the external electric field and the change in the transmembrane potential is quite complicated, and for simple practical calculations an approximation is used. Two important implications need to be emphasized: (1) the transmembrane potential of the cell responds linearly to the electric field, e.g., a doubling of the electric field strength doubles the induced change of the transmembrane potential. (2) the change in the transmembrane potential for a given electric field strength, increases with the diameter of the cell. For example, twice as large cells experience twice as large transmembrane potential changes (spherical cells), meaning that it is generally easier to electroporate large cells compared to small cells. Electroporation occurs when the transmembrane potential of the cells in the treated tissue exceeds about 200 mV, depending on the cell type (Teissie and Rols, 1993). If the applied electric field is very strong, producing a transmembrane potential above approximately 1,000 mV, the changes in the phospho-lipid bilayer of the membrane will be very pronounced, causing what is known as irreversible electroporation, which results in cell death due to prolonged adverse ion concentrations (Hojman et al., 2008). Transient or reversible EP is found to last for minutes at physiological temperature, depending on the voltage pulse parameters and tissue type, after which the cells regain their molecular homeostasis (Saulis et al., 1991).

Electroporation Affecting Parameters The Electrode Device To deliver the electric field to the target tissue a mechanical electrode arrangement (electrode device) is applied. The size and shape of the electrodes have a large impact on the distribution of the electric field. It is not possible to state a generalized formula that relates the electric field to the geometry of the electrodes. Two types of electrode devices, the parallel plate electrode device and the needle electrode device, should be mentioned since they are very common in clinical applications, and commercially available. The parallel plate electrode device consists of two plates electrodes, typically 1–2 cm2 each, and delivers the most uniform electric field, however, is applicable only

200

in non-invasive procedures, for example treatment of small prominating tumors. The needle electrode device is available with different numbers of needles electrodes arranged in different formations. The needle electrode device is typically used to treat subcutaneous tumors because of its ability to penetrate up till about 3 cm tissue at a time. The electric field generated by the needle electrode is, however inhomogeneous.

The Pulse Parameters The voltage generator charges the electrodes according to pulse parameters suitable for the particular clinical situation (Fig. 20.2a). The pulses are specified by (1) the number of pulses in a single treatment, (2) the amplitude of the pulse voltage, (3) the duration of the pulse and (4) the frequency of the pulses (number of pulses per second). In clinical application only rectangular pulses are considered. In electrochemotherapy a complete treatment session typically consists of 8 pulses delivered as 1 pulse per second (1 Hertz), and with pulse duration of 100 ms. The pulse amplitude voltage setting is chosen such that the strength of the electric field surpasses the electroporation threshold level, which is empirically based and normally standardized unless specially designed electrodes are employed.

Tissue Parameters Besides the cell diameter, other tissue related parameters may influence electroporation of the target tissue, for example the cellular density of the tissue, the shape of the cells and the general condition of the cells. For example necrotic cells within the treated tissue may distort the electric field distribution, because of local differences in the electric properties e.g., conductivity of the tissue. Tissue anisotropy is another important parameter to consider, since anisotropic tissue is highly sensitive to the orientation of the electrodes, with regard to the distributed electric field. The advantage of electroporation based drug and gene delivery is that by applying the optimal parameters for a given tissue, it is possible to customize delivery of a given drug or gene to a particular region encompassed by the electrodes. In neurological disease, the obvious targets are cancer

M. Linnert et al.

in the brain and delivery of chemotherapy or non-viral gene delivery for e.g., Parkinsons disease. Electrodes for use in the brain are described in a subsequent section.

Choice of Chemotherapy The chemotherapeutic drug of choice in the performance of electrochemotherapy is in our opinion bleomycin. Other chemotherapeutic drugs have been tested regarding the magnitude of enhancement of the cytotoxicity when using electroporation. When testing drugs commonly used in the clinical setting, results show an increase of cytotoxicity by a factor 300–700 for bleomycin (Gehl et al., 1998; Orlowski et al., 1988), a factor 3 and 2.3 for carboplatin and cisplatin, and for the drugs daunorubicin, doxorubicin, etoposide and paclitaxel no effect of electroporation was found (Gehl et al., 1998). In concordance with this, other results show an enhancement of the cytotoxicity of bleomycin by a factor 5000, and confirmed a much more limited effect of electroporation when using the drugs carboplatin and cisplatin (Jaroszeski et al., 2000). The next paragraph will clarify the nature of bleomycin, and explain additional reasons why we prefer this drug to others for electrochemotherapy.

Bleomycin Bleomycin is an antibiotic produced from the fungus Streptomyces verticillus and was discovered by Umezawa et al. (1966). It is formed by a mixture of peptides and contains a unique structural component, the bleomycinic acid, and a terminal alkylamine group (Lazo and Chabner, 1996). Bleomycin is a hydrophilic and charged molecule with a molecular weight of 1,500 Da, and it passes the intact plasma membrane poorly (Poddevin et al., 1991). Bleomycin is a good chelator of metals and in the presence of oxygen it can bind to ions of iron, cobalt, zinc, and copper. The bleomycin-Fe2+ complex is the most active complex (Gothelf et al., 2003). When bleomycin chelates with iron in the presence of oxygen, a production of free radicals induce DNA

20 Electrochemotherapy for Primary and Secondary Brain Tumors

breaks and mediate lipid peroxidation (Bokemeyer, 2008; Lazo and Chabner, 1996). Bleomycin induces single-strand and double-strand DNA breaks with a ratio of 10:1 (Lazo and Chabner, 1996). In particular, the double strand breaks and resulting loss of chromosome fragments have a cytotoxic effect. Bleomycin has to be internalized in the cell cytosol to have an effect and the outer cell membrane is the limiting factor (Gothelf et al., 2003). Once in the cell cytosol it will effectuate its toxicity although its way to the cell nucleus is relatively unknown. There are two scenarios depending on how many molecules of bleomycin enter the cell: (i) the cells are arrested in the G2–M phase of the cell cycle and die in approximately three doubling times with low concentration of bleomycin, and (ii) pseudoapoptosis is induced and kills the cell within minutes with high concentrations of bleomycin (Tounekti et al., 1993). Bleomycin’s mechanism of action is effective in all steps of the cell cycle, though cells in G2/M phase are considered more sensitive because the DNA is more accessible in this phase (Mir et al., 1996). Bleomycin is mostly used in the treatment of lymphoma (Hodgkin and non-Hodgkin) and testicular cancer in combination with other antineoplastic drugs (Bokemeyer, 2008; Lazo and Chabner, 1996; Morantz et al., 1983). It is also the preferred drug in treating tumors with electrochemotherapy (Gothelf et al., 2003; Heller et al., 1998; Marty et al., 2006; Mir et al., 2006). Bleomycin is eliminated from the blood by renal excretion (Mir et al., 1996). The most important toxic reactions affect the lungs and skin, causing pulmonary fibrosis in about 10%, with a mortality rate of 1%, and erythema, induration, hyperkeratosis, and peeling of the skin (Lazo and Chabner, 1996; Morantz et al., 1983). The toxicity of bleomycin increases with and is directly related to the cumulative dose received. It should be noted that normally fever occurs 48 h after intravenous drug administration in 25% of the patients (Lazo and Chabner, 1996). Because electrochemotherapy is a once-only treatment, the doses necessary are not even near the cumulative doses that are reported to cause serious adverse effects. Therefore, the anticipated adverse effects related to bleomycin should be of a mild nature, such as flu-like symptoms and a slight fever. In the treatment of electrochemotherapy bleomycin can be administered both intravenously and intratumorally,

201

and the adverse effects of course depends on the route of administration. The mentioned adverse effects are seen when bleomycin is administered intravenously. In the next paragraph we will get into the adverse effects seen with intratumoral administration of bleomycin in the brain. To summarize, bleomycin is the ideal drug for electrochemotherapy as it is very effective once inside the cell, probably because of the ability to induce DNA strand breaks. At the same time bleomycin is a large molecule with less effect on the healthy, non- electroporated cells, where the cell membrane is intact. Bleomycin usually has only few and mild adverse effects in the doses that we use for electrochemotherapy. These facts make bleomycin the drug of choice, when performing a specific and localised cancer treatment with electrochemotherapy.

Clinical Experience with Bleomycin in the Brain In the preparation for introducing electrochemotherapy to the human brain we have looked in the literature for reported adverse effects of treatment with bleomycin in the brain. Bleomycin has been used in the treatment of brain tumors for more than 30 years. A review of adverse effects of bleomycin as a direct injection into a solid tumor or cyst showed, that only 5 out of 189 patients (3%) had serious adverse effects and 6 patients (3%) had moderate adverse effects (Linnert and Gehl, 2009). In general, treatment with bleomycin for brain tumors was well tolerated. The most common adverse effects was transient fever in 19%, headache in 16%, nausea and vomiting in 10% and peritumoral edema swelling and fluid around the tumor in 4% of the patients. Fatigue was reported in 5% of the patients and 3% (5 patients) had epileptic seizures, out of which two patients already had epilepsy. Five patients had serious adverse effects: 2 patients developed vision loss on one and both eyes respectively; 2 patients had hearing loss and one patient developed generalized brain edema and died. The patient who developed brain edema received an unusually large dose of bleomycin 56 mg in daily doses over only 8 days. All cases with severe and moderate adverse effects except one were patients with

202

craniopharyngiomas and the adverse effects were probably caused by the tumor localization in the deep brain. In conclusion, bleomycin injection into the brain has been fairly well tolerated at doses much higher than those used for electrochemotherapy.

Concept of Electrochemotherapy Electrochemotherapy is the electroporation-mediated transfer of antineoplastic drugs to tumors. Electrochemotherapy was invented in the early 1990s proving the potentiation of anti-tumor effect of chemotherapeutics by applying local electrical pulses. This is a very useful method for delivery of molecules to the cytosol such as bleomycin that would otherwise not be able to pass the outer cell membrane. Bleomycin is as mentioned a large, hydrophilic molecule, that passes the cell membrane poorly and there are no cellular uptake mechanisms for bleomycin. Normally the uptake of bleomycin has to depend on diffusion, which limits the efficacy of the drug. However, once inside the cell, bleomycin is a highly effective chemotherapeutic drug that works by inducing DNA strand breakage, quantified as 10–15 DNA strand breaks per DNA molecule (Tounekti et al., 2001). So when bleomycin and electroporation are combined (electrochemotherapy), bleomycin can enter the cell easily, and the cytotoxicity increases over 300 fold (Gehl et al., 1998; Jaroszeski et al., 2000; Orlowski et al., 1988). Because the increase of cytotoxicity is so pronounced, the treatment is successful even in cancer diseases, such as malignant melanoma, where chemotherapy has been abandoned because of a poor response. Electrochemotherapy can also be applied to areas that are previously irradiated (Marty et al., 2006; Matthiessen et al., 2009). Bleomycin can be administered either intravenously or directly into tumor as an injection. The drug enters the cells in sufficient quantities only where the electric pulses are applied and cells are electroporated, hence treatment with electrochemotherapy is always a local treatment regardless of the method of administration. It has been shown, that electrochemotherapy is very effective in the treatment of cutaneous metastases, where 73–91% of the tumors show a complete response (CR) after only one treatment (Heller et al., 1998; Marty et al., 2006).

M. Linnert et al.

Preclinical Experience with Electrochemotherapy in the Brain In the early 1990s the first reported initiative was made to explore electrochemotherapy as a treatment modality for brain tumors in a rat model (Salford et al., 1993). Acupuncture needles were used as electrodes to treat rats inoculated with tumor cells with electrochemotherapy, intravenously injected bleomycin being the chemotherapeutic drug. Salford et al. reported an almost double survival time for the electrochemotherapy treated rats than for the bleomycin only treated rats. However, most of the rats given electrochemotherapy (n = 17) had only on average 3–4 days delayed sign of symptoms for late stage tumor growth, and were terminated (n = 15), whereas only a few (n = 2) were without symptoms when terminated, eventually. Recently electrochemotherapy of primary brain tumors have successfully been performed in an animal model (Agerholm-Larsen et al., 2010). The experimental setup was basically to inoculate rat brains with glial derived tumor cells to obtain primary tumor growth for later once-only electrochemotherapy. The tumor cell line inoculated (Siesjo et al., 1993) was rather potent, and once the tumor appeared at MRI, it would, if untreated, progressively grow for only 2–3 weeks before termination had to be initiated due to tumor size. The tumors, however, in spite of their relative short life time before termination, did show several pathological features similar to human glioblastomas such as palisades and pseudo-palisades. From the time of initial growth of the primary tumor to the termination of the animal, the tumor size was estimated and/or treatment effects evaluated from in vivo magnetic resonance imaging MRI. Tissue reactions due to the electric field have also been investigated using diffusion weighted MRI (Mahmood et al. 2011). This way, the regression of the tumor could be followed in vivo once the treatment effects oedema and diffuse contrast upload in the target area had declined within the first week after treatment. The electrochemotherapy was performed with bleomycin, intracranially injected targeting the tumor, and with a newly developed electrode device for electrochemotherapy in soft tissue like the brain, the electrode device being a minor scale of a newly developed electrode device for electrochemotherapy

20 Electrochemotherapy for Primary and Secondary Brain Tumors

in human brain tissue (Agerholm-Larsen et al., 2010). The results so far have been promising in terms of treatment effect, survival, tolerability of electrochemotherapy and safety issues. For the treatment effect to be 100% successful the targeted area with the tumor must be permanently eliminated. The preclinical results, however, showed a wide range of steps on the ladder of success, reaching all the way from 50 to 100% success of covering the whole tumor leading to either partial or total elimination of the tumor. The data are based on MRI that was used to follow the stages of deterioration of the brain tissue in the targeted area in vivo, and immuno- and histochemical stainings performed in vitro to provide information of necrotic tissue, remains of tumor cells, and effect on neurons and glial cells in the targeted area. The data in summary were, that necrotic tissue was obtained in brain areas targeted with once-only electrochemotherapy and clearly distinguishable from the controls, where almost no effects were observed in brain areas targeted with bleomycin or electroporation only as shown in Fig. 20.5a, b, c. These data suggest that bleomycin may be taken up by the tumor cells due to the transient permeabilization of the cell membrane caused by electroporation, and that neither intracranially injected bleomycin nor electroporation alone can cause elimination of a tumor in this animal model. These preclinical data reflect promising preliminary results with the potentials of improvements of success. That is, in the animal model the tumor is accessed by electrodes deployed through the burr hole already made for inoculation of the tumor cells. In the clinic brain scans exposing the tumor position will be matched to stereotaxic coordinates to ensure that a burr hole or craniotomy is positioned optimally for treatment. Rats undergoing electrochemotherapy have successfully been followed up until 8 weeks after treatment before termination, showing no obvious sign of basic malfunction either physically or mentally.

Electrochemotherapy in the Human Brain Electrochemotherapy has been performed in a variety of different tissues, but mostly skin and other easily

203

Fig. 20.4 Electrochemotherapy in the human brain. Schematic drawing of the proposed electroporation procedure in the human brain

accessible tissue. Because of advances in electrode technology, it is now possible to pursue this treatment modality in other types of tissue. We have developed an electrode especially suited for use in the brain, which we find to be a suitable target organ for several reasons. Electrochemotherapy is known to be a treatment that is quick, effective, localized and relatively lenient to adjacent healthy tissue. These qualities make electrochemotherapy a new and interesting treatment modality in the brain, where it is very important, that the malignant cells are killed effectively at the same time as the healthy neurons are spared to the greatest extent possible. In our research group we are planning a clinical trial, where brain metastases will be treated as a palliative treatment with electrochemotherapy using bleomycin and the novel brain electrode (Fig. 20.4). The treatment will be performed with the patient in general anaesthesia as a once-only procedure. A neurosurgeon will plan the treatment from an MRI of the brain and use stereotactic equipment to guide the brain electrode into the tumor through a burr hole.

The Blood-Brain Barrier The blood brain barrier (BBB) is made of nonfenestrated endothelial cells, which makes it highly impermeable, only allowing passage of small, hydrophobic and uncharged molecules such as water. The endothelial cells are held together by tight

204

M. Linnert et al.

b

a

c

Fig. 20.5 Preclinical treatment results. Rat brain tissue (coronal slices) stained with H&E. All treatment modalities took place 2 weeks after inoculation of tumor cells, and all rats were

terminated 1 week after treatment. a Rat brain treated with electroporation, b Rat brain treated with bleomycin, c Rat brain treated with electrochemotherapy

junctions in the blood vasculature throughout the brain. Therefore, the BBB often keep anti-cancer drugs from penetrating into the brain tissue. The BBB is disrupted in different degrees by pathological processes, for instance after a stroke, a malignant brain tumor, infection or trauma. After whole brain radiation therapy (WRBT) the BBB is also under continuous break down for weeks to months (De Vita et al., 1993; Jiang et al., 2010). The patients in the upcoming trial will have both a pathological process in the form of brain metastases and will have received WBRT as first line treatment, so both factors will influence the BBB integrity. Additionally, deployment of electrodes and electroporation of the brain tissue will lead to increased permeability. Another perspective of using electrochemotherapy in the brain could

be to treat the margins of a primary brain tumor after surgical removal of the bulk of the tumor. Previous studies have pointed out, that it is very important to treat the well-vascularized, actively proliferating, infiltrating edge of the tumor with anti-cancer drugs, but problematic due to the intact BBB. In contrast, it is easier to treat the central, leaky and hypoxic part with anticancer drugs. This problem is termed the sink effect and can be the reason for chemotherapy failure in the brain (Neuwelt, 2004). Thus, in electrochemotherapy, drug delivery and precision of the expansion of the electric field are key elements. Using electrochemotherapy for treatment of the margins could be suitable for infiltrating tumors such as glioblastoma multiforme and is an important priority in our future research.

20 Electrochemotherapy for Primary and Secondary Brain Tumors

Conclusion Because patients with brain cancer generally have an unfavourable prognosis, there is a need for better treatment options. Electrochemotherapy is a good candidate because it is a quick, once-only and effective treatment that may also be applied to previously irradiated tissue. Preclinical studies show, that electrochemotherapy can be effective and tolerable in rat brain. Additionally, this treatment modality is relatively lenient to adjacent healthy tissue and should be able to overcome the problems of the blood brain barrier. Results from coming clinical trials will reveal the future implications of electrochemotherapy for the treatment of brain cancer.

References Agerholm-Larsen B, Iversen HK, Jensen KS, Møller J, Ibsen P, Mahmood F, Gehl J (2010) Electrochemotherapy for brain tumors: preclinical studies with bleomycin. Society of Neuro-Oncology, Neuro Oncol 12(suppl 4):88 Bokemeyer C (2008) Bleomycin in testicular cancer: will pharmaco genomics improve treatment regimens? J Clin Oncol 26:1783–1785 Brennum J, Kosteljanetz M, Roed H (2002) Hjernemetastaser. Ugeskr Læger 164:3522 De Vita V, Hellmann S, Rosenberg S (1993) Cancer, principles and practice of oncology, 5th edn. Lippincott-Raven Publishers, Philadelphia Gehl J (2003) Electroporation: theory and methods, perspectives for drug delivery, gene therapy and research. Acta Physiol Scand 177:437–447 Gehl J, Skovsgaard T, Mir LM (1998) Enhancement of cytotoxicity by electropermeabilization: an improved method for screening drugs. Anti-Cancer Drugs 9:319–325 Gehl Videbæk K (2009). (WO/2007/144004) Electrode Introducer Device. Gori S, Rimondini S, De Angelis V, Colozza M, Bisagni G, Moretti G, Sidoni A, Basurto C, Aristei C, Anastasi P, Crino L (2007) Central nervous system metastases in HER-2-positive metastatic breast cancer patients treated with trastuzumab: Incidence, survival, and risk factors. Oncologist 12:766–773 Gothelf A, Mir LM, Gehl J (2003) Electrochemotherapy: results of cancer treatment using enhanced delivery of bleomycin by electroporation. Cancer Treatment Rev 29:371–387 Hazuka MB, Kinzie JJ (1988) Brain Metastases – Results and Effects of Reirradiation. Int J Rad Oncol Bio Phys 15:433– 437 Heller R, Jaroszeski MJ, Reintgen DS, Puleo CA, DeConti RC, Gilbert RA, Glass LF (1998) Treatment of cutaneous and subcutaneous tumors with electrochemotherapy using intralesional bleomycin. Cancer 83:148–157

205

Hojman P, Gissel H, Andre FM et al (2008) Physiological effects of high- and low-voltage pulse combinations for gene electrotransfer in muscle. Hum Gene Ther 19:1249–1260 Jaroszeski MJ, Dang V, Pottinger C, Hickey J, Gilbert R, Heller R (2000) Toxicity of anticancer agents mediated by electroporation in vitro. Anticancer Drugs 11:201–208 Jiang J, Wei WH, Feng YL, Zhou YC, Luo WJ, Yuan JW, Zhang GY, Lu ZQ (2010) Application of 99mTc-DTPA in evaluation of blood-brain barrier permeability in patients receiving whole brain irradiation. Nan Fang Yi Ke Da Xue Xue Bao 30:329–330 Kuhn MJ, Hammer GM, Swenson LC, Youssef HT, Gleason TJ (1994) MRI Evaluation of solitary brain metastases with triple-dose gadoteridol – Comparison with contrastenhanced CT and conventional-dose gadopentetate dimeglumine MRI studies in the same patients. Comput Med Imaging Graph 18:391–399 Lazo Chabner B (1996) Bleomycin. In: Chabner B, Longo D (eds) Cancer chemotherapy and biotherapy, 2nd edn. Lippincott-Raven Publishers, Philadelphia, pp 379–393 Li J, Bentzen SM, Renschler M, Mehta MP (2007) Regression after whole-brain radiation therapy for brain metastases correlates with survival and improved neurocognitive function. J Clin Oncol 25:1260–1266 Linnert M, Gehl J (2009) Bleomycin treatment of brain tumors: an evaluation. Anti-Cancer Drugs 20:157–164 Mahmood F, Gehl J (2011) New clinical electrode device for electroporation of intracranial tumors– emiempirical designoptimization and geometrical tolerance assessment. Bioelectrochemistry 81:10–16. doi:10.1016/j.bioelechem.2010.12.002 Mahmood F, Hansen RH, Agerholm-Larsen B, Jensen KS, Iversen HK, Gehl J (2011) Diffusion-weighted MRI for verification of electroporation-based treatments. J Membr Biol. doi:10.1007/s00232-011-9351-0 Marty M, Sersa G, Garbay JR, Gehl J, Collins CG, Snoj M, Billard V, Geertsen PF, Larkin JO, Miklavcic D, Pavlovic I, Paulin-Kosir SM, Cemazar M, Morsli N, Rudolf Z, Robert C, O’Sullivan GC, Mir LM (2006) Electrochemotherapy – An easy, highly effective and safe treatment of cutaneous and subcutaneous metastases: results of ESOPE (European Standard Operating Procedures of Electrochemotherapy) study. EJC Suppl 4:3–13 Matthiessen L, Kamby C, Hendel H, Johannesen H, Gehl J (2009) Electrochemotherapy as palliative treatment for chest wall recurrence of breast cancer - initial results. EJC 7(Suppl 2):273 Meyers CA, Smith JA, Bezjak A, Mehta MP, Liebmann J, Illidge T, Kunkler I, Caudrelier JM, Eisenberg PD, Meerwaldt J, Siemers R, Carrie C, Gaspar LE, Curran W, Phan SC, Miller RA, Renschler MF (2004) Neurocognitive function and progression in patients with brain metastases treated with whole-brain radiation and motexafin gadolinium: results of a randomized phase III trial. J Clin Oncol 22: 157–165 Mir LM, Tounekti O, Orlowski S (1996) Bleomycin: revival of an old drug. Gen Pharmacol 27:745–748 Mir LM, Gehl J, Sersa G et al (2006) Standard operating procedures of the electrochemotherapy: instructions for the use of bleomycin or cisplatin administered either systemically or locally and electric pulses delivered by the Cliniporator (TM)

206 by means of invasive or non-invasive electrodes. EJC Suppl 4:14–25 Morantz RA, Kimler BF, Vats TS, Henderson SD (1983) Bleomycin and brain tumors. A review. J Neurooncol 1:249– 255 Neuwelt EA (2004) Mechanisms of disease: the blood-brain barrier. Neurosurgery 54:131–140 Orlowski S, Belehradek J Jr., Paoletti C, Mir LM (1988) Transient electropermeabilization of cells in culture. Increase of the cytotoxicity of anticancer drugs. Biochem Pharmacol 37:4727–4733 Poddevin B, Orlowski S, Belehradek J, Mir LM (1991) Very high cytotoxicity of bleomycin introduced into the cytosol of cells in culture. Biochem Pharmacol 42:S67–S75 Ryberg M, Nielsen D, Osterlind K, Andersen PK, Skovsgaard T, Dombernowsky P (2005) Predictors of central nervous system metastasis in patients with metastatic breast cancer. A competing risk analysis of 579 patients treated with epirubicin-based chemotherapy. Breast Cancer Res Treat 91:217–225 Salford LG, Persson BR, Brun A, Ceberg CP, Kongstad PC, Mir LM (1993) A new brain tumour therapy combining bleomycin with in vivo electropermeabilization. Biochem Biophys Res Commun 194:938–943 Saulis G, Venslauskas MS, Naktinis J (1991) Kinetics of Pore Resealing in Cell-Membranes After Electroporation. Bioelectrochem Bioenerg 26:1–13 Siesjo P, Visse E, Lindvall M, Salford L, Sjogren HO (1993) Immunization with (mutagen-treatedtum-) cells causes

M. Linnert et al. rejection of nonimmunogenic rat glioma isografts. Cancer Immunol Immun 37:67–74 Smedby KE, Brandt L, Backlund ML, Blomqvist P (2009) Brain metastases admissions in Sweden between 1987 and 2006. Br J Cancer 101:1919–1924 Sperduto PW, Berkey B, Gaspar LE, Mehta M, Curran W (2008) A new prognostic index and comparison to three other indices for patients with brain metastases: an analysis of 1,960 patients in the RTOG database. Int J Rad Oncol Bio Phys 70:510–514 Teissie J, Rols MP (1993) An experimental evaluation of the critical potential difference inducing cell-membrane electropermeabilization. Bio J 65(l):409–413 Tounekti O, Kenani A, Foray N, Orlowski S, Mir LM (2001) The ratio of single- to double-strand DNA breaks and their absolute values determine cell death pathway. Br J Cancer 84:1272–1279 Tounekti O, Pron G, Belehradek J Jr., Mir LM (1993) Bleomycin, an apoptosis-mimetic drug that induces two types of cell death depending on the number of molecules internalized. Cancer Res 53:5462–5469 Umezawa H, Maeda K, Takeuchi T, Okami Y (1966) New antibiotics, bleomycin A and B. J Antibiot (Tokyo) 19:200–209 Videtic GM, Gaspar LE, Aref AM, Germano IM, Goldsmith BJ, Imperato JP, Marcus KJ, McDermott MW, McDonald MW, Patchell RA, Robins HI, Rogers CL, Suh JH, Wolfson AH, Wippold FJ (2009) American College of Radiology appropriateness criteria on multiple brain metastases. Int J Radiat Oncol Biol Phys 75:961–965

Chapter 21

Brain Tumors: Convection-Enhanced Delivery of Drugs (Method) Anne-Laure Laine, Emilie Allard, Philippe Menei, and Catherine Passirani

Abstract Delivery of therapeutic agents into the brain has been an ongoing challenge for many years. The poor prognosis for patient with primary malignant brain tumors treated with conventional techniques (surgery, radiotherapy and chemotherapy) has motivated the development of new strategies to deliver drugs into the brain. Local intracranial delivery of antineoplastic agents has appeared to be the most effective drug delivery technique into the central nervous system by circumventing the limitations imposed by the blood brain barrier (BBB). Convection-enhanced delivery (CED) is an alternative strategy to directly infuse drugs into brain tissue. This approach is based on continuous injection of the therapeutic agent under positive pressure via a catheter implanted into the brain. Convective transport driven by pressure gradient allows a widespread distribution of small and large drugs within the brain. In vivo experiments in rodents, cats and primates proved the efficacy of CED to deliver drugs into a targeted zone. However, clinical trials have reported frequent leakage phenomenon leading to mixed results for this delivery technique. A better optimization of operational parameters including infusion rate, catheter design, catheter placement and drug pharmacological formulation should allow achieving accurate and efficient delivery. In conjunction with CED, the use of nanocarriers to enhance drug pharmacokinetic behavior may help to achieve higher therapeutic index against tumor cells

C. Passirani () INSERM, U646, Universite d’Angers, Angers F-491000, France e-mail: [email protected]

over healthy tissues. Additionally, the development of computer simulation to predict drug distribution and the real-time imaging for immediate assessment of convection efficiency may contribute to the CED improvement. Keywords Therapeutic agents · Convection-enhanced delivery · Catheter · Infusion rate · Infusate · CNS

Introduction Despite considerable advances of research in the area of Central Nervous System (CNS) disorders, the prognosis for patients with brain tumors remains poor. Median survival for patients with the most severe form treated with surgical resection, radiation and the addition of systemic chemotherapy is 12–15 months (Stupp et al., 2005). Finding new therapeutic strategies that will provide efficient treatment against CNS tumors remains a major challenge. Conventional methods for the treatment of brain tumors usually involve delivery of drugs via the systemic circulation. Failure of those therapies is attributed to the presence of the blood brain barrier (Pardridge, 2007). The BBB is a physical barrier that separates circulating blood from cerebrospinal fluid in the CNS. This separation protects the brain from the penetration of toxic substances throughout the CNS. The restricted entry into the brain is imposed by a layer of endothelial cells joined by tight junctions. This structure confers a limited diffusion through the BBB that prevents the entry of most pharmaceuticals into the CNS.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_21, © Springer Science+Business Media B.V. 2011

207

208

Despite a variety of approaches tested for enhancing BBB permeability of drugs, no significant improvement was showed for the treatment of CNS tumors (Misra et al., 2003). The poor results obtained by all conventional therapies and the unchanged prognoses for patient with GBM have motivated more direct approaches of drug delivery. Intracranial drug delivery has appeared to be the most adequate method of overcoming the BBB (Sawyer et al., 2006). In theory, regional drug approach constitutes a way to deliver high concentration of drug directly into the tumor. Such delivery can also reduce systemic exposure to drug-induced toxicity. A local delivery technique was achieved implanting a degradable or non-degradable polymer delivery system into the brain (Raza et al., 2005). This approach presents some advantages such as sustained controlled drug release. However, polymer-controlled release as intracerebral bolus injection suffers from restricted drug distribution due to limited drug diffusion within the brain (Fung et al., 1998). An alternative local delivery based on convective transport has been proposed to overcome the limitation imposed by the diffusion observed with polymer-delivery system. This technique called Convection-Enhanced Delivery (CED) was introduced by researchers from the US National Institutes of Health (NIH) by the early 1990s (Bobo et al., 1994). In CED method, the infusate is administered via a small catheter connected to a pump. The continuous injection of a convective flow directly in the brain results in a widespread distribution of small as well as large molecules into it (Debinski and Tatter, 2009). By bypassing the BBB, this technique allows the local delivery of a variety of agents, ranging from conventional therapeutics agents, to monoclonal antibodies, targeted toxins, nanocarriers (Allard et al., 2009) and viruses (Bidros and Vogelbaum, 2009). However, several parameters can affect therapeutic efficacy of CED, including infusion rate, cannula shape, infused tissue structure and physico-chemical properties of the infusate. Those parameters have to be carefully investigated and set up for an optimal delivery. This chapter aims at giving an overview of the CED approach. The following paragraphs will describe the technical aspects and critical parameters of this promising local delivery technique. Optimal setting for efficient delivery will be proposed according to the literature.

A.-L. Laine et al.

Convection Enhanced Delivery Convection-enhanced delivery is a local delivery method in which drugs are infused directly into brain tissue. The therapeutic agents are administered via one to several catheters implanted into the brain (Fig. 21.1a). An external pump produces a positive pressure which pushes the infusate through the catheter. With this delivery system, the drug is directly infused into the target tissue at a predetermined concentration, rate and duration. The use of a catheter allows a stereotaxic placement into a CNS targeted zone. Stereotaxy uses a threedimensional coordinates system to localize structures and targets within the brain. Thanks to this technique, the catheter can be accurately implanted within the tumor mass or around the tumor or the resection cavity. The ability to target the delivery placement limits structural damage to the infused tissue, except at the site of the catheter, and allows avoiding systemic toxicity. Most of local delivery techniques including polymer systems and bolus injection rely on diffusive transport to distribute a therapeutic agent within the CNS. Diffusion corresponds to the spread of molecules from regions of higher concentration to regions of lower concentration. The diffusion of a compound depends mainly on the concentration gradient and its diffusivity properties. Many therapeutic agents such as neurotrophic factors, antibodies, growth factors and enzymes are not able to diffuse over large distances into the brain due to high molecular weight. This poor diffusion severely limits drug distribution and penetration into the brain tissue. For example, an antibody IgG requires 3 days to diffuse 1 mm from its delivery site in tumor (Jain, 1989). That is why most antineoplastic substances show limited effects after direct administration into brain via diffused-based techniques. In contrast to those delivery techniques, CED relies on convection, a molecular weight independent transport, to overcome the limitation imposed by diffusion. The term convection refers to the mass transfer by bulk motion of the fluid. In CED, the convective flow, also called bulk flow, is driven by pressure gradient induced by the difference between the interstitial pressure and the pressure at the tip of the catheter. Drugs are carried along with the fluid flow and distributed through the brain tissue by convective and diffusive transport.

21 Convection-Enhanced Delivery of Drugs (Method)

209

A

B

Fig. 21.1 a Convection-enhanced delivery in a human brain. The drug is infused via a catheter implanted into the brain. An external pump connected to the catheter provides positive pressure pushing the drug through the interstitial space of the

central nervous system. b Brain section with identification of the gray matter (GM) and white matter (WM) location in humans and rats

The superposition of transports by random motion of the molecules (diffusion) and by the bulk motion of the fluid (convection) offers a greater volume of distribution than simple diffusion. Morrison et al. (1994) showed that CED can provide distribution volumes in brain that are 2- to 10-fold larger than those obtained after direct interstitial injection (Morrison et al., 1994). Additionally, data from CED studies showed enhanced distribution of various therapeutic agents in rat, cat and non human primate brains with an order of magnitude greater than simple diffusion-based delivery (Bobo et al., 1994; Yokosawa et al., 2010). However, despite proven efficacy, CED shows some limitations. Most common issues with infusions of

drug into the brain are reflux along the catheter track and leakage outside the targeted area. It has been reported that leakage phenomenon affects 20% of CED experiments (Varenika et al., 2008). While preventing continued pressure in the interstitial space, reflux and leakage can also increase the risk of neurotoxicity to adjacent tissue and induce adverse effects. The occurrence of leakage and reflux may explain the mixed results obtained in clinical trials (Kunwar, 2003, 2007). Indeed, for optimal CED efficiency, it is essential to cover the entire target zone while avoiding exposure of healthy tissue. The extent of drug distribution achieved via CED depends on the type of tissue infused (that is, tumor, gray matter or white matter), infusion volume

210

and rate, catheter design but also on drug elimination mechanism including metabolism, binding, efflux and leakage. These different parameters are discussed in the following paragraphs.

Key Parameters Catheter Placement During CED, the distribution volume for a given agent is highly dependent on the type of tissue being infused and its structural properties, such as hydraulic conductivity, vascular volume fraction, and extracellular fluid fraction. Clinical trials outcomes have highlighted the importance of accurate catheter placement for optimal drug distribution (Kunwar et al., 2007). Two main areas can be distinguished in the CNS; the gray matter and the white matter (Fig. 21.1b). The gray matter is mainly composed of cell bodies and glial cells and the transport within this structure is isotropic. In contrast, the white matter consists mostly of axons leading to the peripheral nervous system. This structure composed of fiber pathways makes the transport anisotropic. It is worth noting that gliomas invasion preferentially occurs along these white bundles. It has been reported that a significant natural bulk flow with an interstitial velocity of 10 μm/mL may exist in porous white matter near the ventricular surface (Rosenberg et al., 1980). Bobo et al. (1994) demonstrated that an infusion via CED into the corona radiate (white matter) of cats allowed an effective and large distribution of the drug within the brain tissue. Similar observations were reported by Lieberman et al. (1995) regarding a drug infusion via CED into the gray matter. Both gray and white matters have the capability to distribute the infused drug within the CNS. However, the distribution profile differs between both structures and the difference is even more significant compared to brain tumor. When the infusion takes place into a homogeneous brain structure (e.g., gray matter) the infusion is initially driven by pressure gradient from the source of injection toward surrounding tissue producing spherical volume (Fig. 21.2a). The concentration profile in this volume is relatively uniform and proportional to the delivered concentration. At the flow front, the pressure gradient reaches zero and the distribution

A.-L. Laine et al.

is no longer driven by convection but by diffusion. The concentration declines precipitately and exponentially within adjacent tissues in accordance with the mathematics of diffusion (Morrison et al., 1994). If the infused drug encounters a structure containing fiber pathways (eg. white matter), the convective flow can be modified by the natural bulk flow within that pathway. It has been reported that interstitial velocities into the white matter may be used to increase the rate of the infusate (Rosenberg et al., 1980). However, this natural movement in white matter can also lead to unpredictable flow and leakage outside the target zone if the direction of the natural bulk flow is opposed to that of the infusate flow (Krauze et al., 2005). It is apparent that inhomogeneities in the tissue may cause unpredictable distribution. The presence of oedemas, often observed in brain cancer and also induced by infusion, reduces volume of treated tissue and can disturb the flow of the infused drug. Moreover, the distribution profile into brain tumors differs markedly from the normal brain. This variation may be due to the difference of interstitial fluid pressure. In normal brain tissue the interstitial pressure is relatively low, 1–2 mmHg, whereas in brain tumor tissue it can be more than 25 times greater. The increased interstitial pressure in peritumoral tissue and within the tumor produces a counterproductive pressure gradient that can drive infusate to follow the path of lowest interstitial pressure, out of high pressure areas such as tumor tissue. The complex anatomy of brain tumor and its high intracranial pressure can lead to drug distribution in undesired area and leakage into the subarachnoid space.

Infusion Volume and Rate The flow of injection is a critical parameter for forming an efficient convection and is related to the resistance of the infused tissue (gray and white matter). Experimental results have demonstrated that the higher the infusion rate, the greater the reflux induced (Morrison et al., 1999). It has been shown that infusion rate below 0.5 μL/min in normal brain limits the spread of the drug but results in homogeneous distribution whereas high flow (above 5 μL/min) facilitates backflow (Bobo et al., 1994; Morrison et al., 1999). In most cases, the optimal infusion rate is that which allows the delivery of the therapeutic volume over the

21 Convection-Enhanced Delivery of Drugs (Method) Fig. 21.2 a Schematic representation of drug distribution in brain tissues via CED. The drug distribution is initially driven by convective flow producing a spherical volume around the infusion point. At the flow front, the infusion is governed by diffusion. b Relationship between the volume of infusion (Vi ) and the volume of distribution (Vd ). In a successful delivery (a), Vd linearly increases in proportion to Vi . Once undesirable phenomenon begins such as leakage or reflux (b), there is a plateau to Vd despite increases in Vi

211

A

B

least amount of time without associated reflux. The optimal infusion rate is also dependent on the cannula size. Krauze et al. (2005) reported that CED infusion flow rates less than 0.5 μL/min can prevent reflux in catheters up to 0.6 mm (outer diameter). However, higher infusion rates are preferred in order to shorten the CED treatment. A mean to achieve high rate is to use modified catheter such as stepped catheter or hollow fiber catheter. Those different types of catheters and their delivery characteristics are described in the following chapter. A correlation between the infusion volume (Vi ) and the distribution volume (Vd ) has been largely

proved from several CED outcomes (Bobo et al., 1994; Varenika et al., 2008). In an efficient CED without reflux or leakage, the relationship between Vi and Vd is linear which means that Vd increases linearly in proportion to Vi (Fig. 21.2b). The slope of the curve Vd /Vi is dependent on the structural properties of the tissue, the characteristics of the infusate and the rate of infusion (Bobo et al., 1994; Varenika et al., 2008). Once reflux or leakage begins, there is a plateau to Vd . Despite increases in Vi , Vd does not increase. The extra infused volume is lost in leak or reflux along the catheter track, away from the targeted zone.

212

Catheter Design The first infusion tool for CED was a needle implanted into the brain of experimental animals (Bobo et al., 1994). The technique has been improved replacing the needle by a catheter which can be stereotactically placed in the CNS. The material used for the catheter fabrication has to be biocompatible to allow a long implantation time. Such material can be plastic, teflon or fused silica. Standard catheters have only one port through that the drug is released. It has been recognized that the larger the catheter diameter, the more easily reflux occurs (Morrison et al., 1999). Additionally, data from experiments with agarose gelatin have shown that small diameter catheters can produce repeatable and uniform spherical volume of distribution (Bauman et al., 2004). New types of CED catheters (Fig. 21.3) are currently under development to increase the efficiency and the reliability of drug delivery. Fiandaca et al. (2008) developed a stepped cannula that allows increasing the infusion rate up to 5 μL/min without any reflux. This reflux-preventing catheter consists of a needle extended with a narrower tip at its end, the differing diameter making the step. This design prevents the reflux along the catheter track by restricting initial backflow of fluid flow beyond the step. The stepped cannula was recently compared with a stepless cannula by Yin et al. (2010). They showed that reflux

Fig. 21.3 Novel catheter designs for CED: stepped cannula, multiple-hole catheter, balloon-tipped catheter, hollow fiber catheter

A.-L. Laine et al.

occurred at only 0.5 μL/min in agarose gel for the stepless cannula leading to a small infusion volume of 4 μL. In contrast, with the 1-mm step cannula no reflux occurred even at high infusion rate of 3.0 μL/min allowing a wide distribution volume of 79 μL. Moreover, the infusion of a recombinant growth factor, GDNF via the step design cannula allowed achieving a great distribution volume of 97.2% in rat striatum with little evidence of reflux. The distribution volume percentage reached via the cannula without step was only 70.7% with significant reflux and leakage in the corpus callosum and the white matter (Yin et al., 2010). Catheters with a single opening for the drug release have shown their limitation for the delivery into brain tumor. The pressure gradient resulting from high interstitial pressure in brain tumor drives the drug flow into low-pressure area. With one port catheter, the infusate flow will follow the path of lowest interstitial pressure which will restrict the distribution and may lead to leakage. Thus, multiple-hole catheters have been designed to provide infusate flow in all the directions around the catheter overcoming the limitation of the one port catheter. The design consists of multiple openings of a few millimeters along the distal end of the catheter. However, a previous study performed in agarose gel showed that the drug release is effective only via the proximal port, rending the other ports useless (Raghavan et al., 2006). This issue can be overcome by reducing the hole size. A porous hollow fiber

21 Convection-Enhanced Delivery of Drugs (Method)

catheter composed of millions of nano-openings (0.45 μm) along its surface has been developed to provide multiple pathways for the drug to infuse within the tissue. Seunguk et al. (2009) compared the distribution of Evans blue dye into the mouse brain through CED via a hollow fiber catheter versus a conventional 28-gauge needle. Experimental results showed a threefold increase in the dye distribution volume with the hollow fiber catheter relative to the single-lumen needle. Since recently, a new type of catheter named balloon-tipped catheter has been proposed. This catheter is a one-port-opening with a balloon proximal to its tip which can be inflated. The inflated form has been designed to fill a resection cavity and prevent reflux phenomenon. A CED experiment using balloon-tipped catheter into canine model showed an efficient drug penetration within the brain parenchyma all around the balloon (Olson et al., 2008).

Infusate Formulation In a typical CED application, the therapeutic agent is dissolved in saline solution and delivered by infusion under positive pressure (Yang et al., 2011). It has been observed that infusion of some drugs achieves poorer efficiency than others due to their low viscosity. Mardor et al. (2005) have shown that low-viscosity infusates tend to backflow along the catheter and into the ventricles whereas high-viscosity infusates limit backflow event and tend to form efficient convection. The infusate viscosity can be increased by adding sucrose, human serum albumin or polyethylene glycol in the aqueous phase of the infusate. A linear correlation between infusate viscosity and distribution volume has been established showing the impact of the viscosity on the CED efficacy. Additionally, infusion of high-viscosity infusate via CED allows covering larger volumes of distribution in less time, thus leading to shorter treatment. Highly viscous formulation can be applied to a wide range of infusate including gene therapy, proteins, and nanoparticles (Perlstein et al., 2008). It has also been shown that the use of co-infusate (e.g., heparin, basic fibroblast growth factor or mannitol) expedites the spread of the drug within the CNS by reducing the affinity of therapeutic agent to the

213

brain environment. For example, the tissue penetration of GDNF and GDNF-homologous trophic factors was significantly enhanced with heparin co-infusion, probably by binding and blocking heparin-binding sites in the extracellular matrix (Hamilton et al., 2001). Despite these progresses, CED of free drug shows severe limitations. First of all, the drug concentration that can be delivered is limited by the solubility of the drug. Additionally, efficiency of free drug CED is restricted by rapid clearance of the therapeutic agent from the tumor interstitium (Kunwar et al., 2007) and no selective accumulation in targeted tissue. Other issues inherent to the physico-chemical properties of the drug can undermine CED efficiency. Although being a promising therapeutic agent, CED of free synthetic retinoid Am80, did not show any improvement on the mean survival days of rats bearing intracranial glioblastoma xenografts (Yokosawa et al., 2010). The failure was attributed to the Am80 hydrophobicity which increases the retention within the brain tissue resulting in a poor distribution volume. A solution to overcome solubility issue and to protect therapeutic agents against elimination and in vivo degradation is to formulate them into nanocarriers. Different drug carrier systems have recently emerged and been tested in CED technique including liposomes (Saito et al., 2006), polymer micelles (Inoue et al., 2009), lipid nanocapsules (Allard et al., 2009). . . Inoue et al. (2009) compared the therapeutic efficacy of free doxorubicin (DOX) and polymer micellar DOX infused by CED against 9L brain tumors in rat model. They showed that polymer micellar doxorubicin (DOX) achieved much wider distribution in brain tissue than free DOX. Accordingly, the median survival obtained with micellar DOX was 36 days compared to 19.6 days obtained from the free DOX treatment. Nanocarriers are constructed systems in nanometer size that can carry drugs (Yokosawa et al., 2010) and imaging agents (Fiandaca et al., 2008). Carrier systems include mainly polymer or lipid-based carriers such as liposomes, nanoparticles, micelles and dendrimers. These differ in structure composition, drugloading capacity, carrier stability, targeting possibility and ability to encapsulate hydrophobic or hydrophilic molecule. Nanocarrier systems are of great interest as they are inert until the drug is released from the confine of the carriers, thus minimizing side effects and toxicity to normal brain tissue. Additionally, they

214

offer several advantages, including enhancement of drug pharmacokinetic behavior and convection properties. Once a drug is encapsulated, the pharmacokinetic properties are no longer related to the drug but are determined by the nanocarrier. Furthermore, the nanocarrier architecture offers the possibility to adjust drug release rate. Improving sustained drug release allows the prolongation of drug’s half-life leading to increased exposure of brain tumor tissue to the therapeutic agent. It also limits the high peak concentration of free drug potentially associated with toxicity and provides an effective inhibition of drug absorption by cells. Surface properties have a considerable impact on the spread of the infusate distribution. The poorer brain distribution observed with hydrophobic or cationic infusate can be circumvented by adding polyethylene glycol or dextran onto the nanocarrier. Indeed, PEG or dextran coating make the infusate more hydrophilic leading to less tissue affinity and increased distribution. Saito et al. (2006) have reported that athymic nude rats bearing U87MG human glioma cells survived significantly longer after treatment with topotecan encapsulated into PEGylated liposome given by CED compared with free topotecan. PEG encapsulation also provides steric stabilization and reduces surface charge (Yokosawa et al., 2010). As the pore size of the extracellular matrix (ECM) is estimated as about 50 nm, nanocarrier size should be in the range of 20–50 nm for an effective transit through the interstitial space (MacKay et al., 2005). It has been showed that the infusion of a hyperosmolar solution of mannitol, either before the CED treatment or in co-infusion with the drug, may increase the size of channels of the interstial space enhancing the nanoparticle transport trough the ECM (Neeves et al., 2007). For instance, CED of 40 nm-liposomes co-infused with 25% mannitol showed an improvement of the Vd from 52.5 ± 2.1 to 78.5 ± 5.5 units immediately after CED (Mamot et al., 2004).

Ced Patient-Specific Simulation and Real-Time Imaging Imaging of brain tumor can be realized via CED. The co-infusion of tracers with the drug allows monitoring its distribution as it is being delivered (Fiandaca et al., 2008). The most common technique currently used to track real-time distribution of therapeutics is the

A.-L. Laine et al.

Magnetic Resonance Imaging (MRI). Varenika et al. (2008) used gadoteridol-loaded liposomes infused by CED to visualize the distribution in various anatomical structures of the CNS of nonhuman primates and canines with MR imaging. This technique evidenced the unexpected high frequency of leakage phenomenon (20% of infusions) into cerebral ventricles. Monitoring of CED is of crucial importance to ensure optimal delivery of therapeutic agents into target sites while minimizing exposure of healthy tissue. Adjustment and alteration to the infusion plan can be carried out in case of undesirable events during the process. In the event of reflux, reducing the infusion rate can allow CED process to continue while backflow is limited. More drastic changes would include altering the cannula placement or inserting a second cannula at the onset of leakage. Those adjustments could allow overcoming leakage issues and covering the targeted area without stopping the process. Additionally, direct visualization can help to determine whether a treatment failed because of the drug’s lack of therapeutic efficacy or because of inadequate delivery. Monitoring information allows optimizing the process including catheter placement in order to achieve robust and reliable delivery of therapeutics. As for any invasive procedure, it would be of great interest to base decision on how to plan CED with patient-specific simulation. The entry of CED in clinical trial has motivated the development of computational methods to aid in planning CED treatments. Such software developed by BrainLAB AG and approved by FDA is now commercially available for neuro-oncology. This program simulates drug distribution into the human brain based on the individual patient’s tractography MRI data and brain tissue anatomy. It is worth noting that tractography allows visualizing the white matter bundles which is of great interest for the simulation of drug distribution. Treatment of specific tumor can be visualized in 3D including the number and position of catheters. The distribution concordance between the actual distribution of ILI3-PE38QQR co–infused with 123 I-albumin by CED and the distribution prediction has proved the software specificity and reliability in treatment planning (Sampson et al., 2007). Additionally, this simulation allows observing infusate leakage into the subarachnoid cerebrospinal fluid space when catheters cross deep sulci (Sampson et al., 2007). Computational simulation provides therefore a useful tool to optimize CED catheter placement.

21 Convection-Enhanced Delivery of Drugs (Method)

Conclusion Convection enhanced-delivery has been developed as a local drug delivery strategy. Using convective transport, it offers the advantage of better drug distribution than the regional delivery technology based on diffusion. CED represents a potential powerful methodology for targeted therapy in the field of neuro-oncology. Bypassing the blood-brain barrier, this approach allows tumor and other brain tissue to be exposed to high drug concentration that could not be achieved via systemic application. The potential of this technique has been clearly demonstrated in preclinical studies (Yang et al., 2011; Yokosawa et al., 2010). However, adverse events including drug backflow along the catheter track and leakage in undesired zone can occur during infusion. Such events undermine CED efficacy and can imply side effects including neurotoxicity and adjacent structure damages. Some improvements are under development for making CED an effective way of delivering antineoplastic drugs in human brain. The use of nanocarrier technology to deliver drugs via CED seems to be a promising combination to improve the efficacy of the treatment against brain tumor. Further development of targeting nanocarriers would allow the accumulation in the desired area and limit extensive distribution into the CNS. It mays also help to penetrate the tumor which shows relative inaccessibility. However, there is a need to better understand the drug release mechanism out of the carrier system. It would be also of great interest to pay special attention to others parameters which may impact the distribution such as the manipulation of physiological parameters including blood pressure and heart rate. Indeed, it was shown that larger Vd could be obtained by increasing blood pressure/heart rate induced by epinephrine (Hadaczek et al., 2006). There is also a need for a better understanding of the drug delivery into tumor brain, markedly different from normal brain. Clinical trials have also reported that target tissue anatomy and patient-specific anatomy play a major role in drug distribution using CED. Gaining more information about the brain’s anatomy and fluid dynamics that govern delivery would allow predicting the time, direction and eventually, the efficiency of therapeutic molecule distribution. Finally, the transposition from rodent or non human primate brain to human brain remains an ongoing challenge

215

due to significant differences between brain structures (Fig. 21.1b) and CED intra-operational parameters have to be adjusted by taking precisely anatomical criteria into account.

References Allard E, Huynh NT, Vessières A, Pigeon P, Jaouen G, Benoit JP, Passirani C (2009) Dose effect activity of ferrocifenloaded lipid nanocapsules on a 9L-glioma model. Int J Pharm 379(2):317–323 Allard E, Passirani C, Benoit JP (2009) Convection-enhanced delivery of nanocarriers for the treatment of brain tumors. Biomaterials 30:2302–2318 Bauman MA, Gillies GT, Raghavan R, Brady ML, Pedain C (2004) Physical characterization of neurocatheter performance in a brain phantom gelatin with nanoscale porosity: steady-state and oscillatory flows. Nanotechnology 15:92–97 Bidros DS, Vogelbaum MA (2009) Novel drug delivery strategies in neuro-oncology. J Am Soc Exp Neuroth 6:539–546 Bobo RH, Laske DW, Akbasak A, Morrison PF, Dedrick RL, Oldfield EH (1994) Convection-enhanced delivery of macromolecules in the brain. Proc Natl Acad Sci USA 91: 2076–2080 Debinski W, Tatter ST (2009) Convection-enhanced delivery for the treatment of brain tumors. Expert Rev Neurother 9:15719–1527 Fiandaca MS, Forsayeth JR, Dickinson PJ, Bankiewicz KS (2008) Image-guided convection-enhanced delivery platform on the treatment of neurological diseases. Neurotherapeutics 5:123–127 Fung LK, Ewend MG, Sills A, Sipos EP, Thompson R, Watts M, Colvin OM, Brem H, Saltzman WM (1998) Pharmacokinetics of interstitial delivery of carmustine, 4-hydroperoxycyclophosphamide, and paclitaxel from a biodegradable polymer implant in the monkey brain. Cancer Res 58:672–684 Hadaczek P, Mirek H, Tamas L, Bohn MC, Noble C, Park JW, Bankiewicz K (2006) The “perivascular pump” driven by arterial pulsation is a powerful mechanism for the distribution of therapeutic molecules within the brain. Mol Therapy 14:69–71 Hamilton JF, Morrision PF, Chen MY, Harvey-White J, Pernaute RS, Phillips H, Oldfield E, Bankiewicz KS (2001) Heparin coinfusion during convection-enhanced delivery (CED) increases the distribution of the glial-derived neurotrophic factor (GDNF) ligand family in rat striatum and enhances the pharmacological activity of neurturin. Exp Neurol 168: 155–161 Inoue T, Yamashita Y, Nishihara M, Sugiyama S, Sonoda Y, Kumabe T, Yokoyama M, Tominaga T (2009) Therapeutic efficacy of a polymeric micellar doxorubicin infused by convection-enhanced delivery against intracranial 9L brain tumor models. Neuro-Oncology 11:151–157 Jain RK (1989) Delivery of novel therapeutic agent in tumors: physiological barriers and strategies. J Natl Cancer Inst 81:570–576

216 Krauze MT, Saito R, Noble C, Bringas J, Forsayeth J, Mcknight TR, Park J, Bankiewicz KS (2005) Effects of the perivascular space on convection-enhanced delivery of liposomes in primate putamen. Exp Neurol 196:104–111 Kunwar S (2003) Convection enhanced delivery of IL13PE38QQR for treatment of recurrent malignant glioma: presentation of interim findings from ongoing phase 1 studies. Acta Neurochir Suppl 88:105–111. Kunwar S, Prados MD, Chang SM, Berger MS, Lang FF, Piepmeier JM, Sampson J, Ram Z, Gutin PH, Gibbons RD, Aldape KD, Croteau DJ, Sherman JW, Puri RK (2007) Direct intracerebral delivery of cintredekin besudotox (IL13PE38QQR) in recurrent malignant glioma: a report by the Cintredekin Besutodox Intraparenchymal Study Group. J Clin Oncol 25:837–844 Lieberman DM, Laske DW, Morrison PF, Bankiewicz KS, Oldfield EH (1995) Convection-enhanced distribution of large molecules in gray matter during interstitial drug infusion. J Neurosurg 82:1021–1029 MacKay JA, Deen DF, Szoka C Jr (2005) Distribution in brain of liposomes after convection enhanced delivery; modulation by particle charge, particle diameter, and presence of steric coating. Brain Res 1035:139–153 Mamot C, Nguyen JB, Pourdehnad M, Hadaczek P, Saito R, Bringas JR, Drummond DC, Hong K, Kirpotin DB, McKnight T, Berger MS, Park JW, Bankiewicz KS (2004) Extensive distribution of liposomes in rodent brains and brain tumors following convection-enhanced delivery. J Neurooncol 68:1–9 Mardor Y, Rahav O, Zauberman Y, Lidar Z, Ocherashvilli A, Daniels D, Roth Y, Maier SR, Orenstein A, Ram Z (2005) Convection-enhanced drug delivery: increased efficacy and Magnetic Resonance Image Monitoring. Cancer Res 65:6858–6863 Misra A, Ganesh s, Shahiwala A, Shah SP (2003) Drug Delivery to the central nervous system: a review. J Pharm Pharm Sci 6:252–273 Morrison PF, Chen MY, Chadwick RS, Lonser RR, Oldfield EH (1999) Focal delivery during direct infusion to brain: role of flow rate, catheter diameter, and tissue mechanics. Am J Physiol 277:1218–1229 Morrison PF, Laske DW, Bobo H, Oldfield EH, Dedrick RL (1994) High-flow microinfusion: tissue penetration and pharmacodynamics. Am J Physiol 266:292–305 Neeves KB, Sawyer AJ, Foley CP, Saltzman WM, Olbricht WL (2007) Dilation and degradation of the brain extracellular matrix enhances penetration of infused polymer nanoparticles. Bain Res 1180:121–132 Olson JJ, Zhang Z, Dillehay D, Stubbs J (2008) Assessment of a balloon-tipped catheter modified for intracerebral convection-enhanced delivery. J Neurooncol 89:159–168 Pardridge WM (2007) Blood-brain barrier delivery. Drug Discov Today 12:54–61 Perlstein B, Ram Z, Daniels D, Ocherashvilli A, Roth Y, Margel S, Mardor Y (2008) Convection-enhanced delivery of maghemite nanoparticles: increased efficacy and MRI monitoring. Neuro-Oncology 10(2):153–161

A.-L. Laine et al. Raghavan R, Brady ML, Rodríguez-Ponce MI, Hartlep A, Pedain C, Sampson JH (2006) Convection-enhanced delivery of therapeutics for brain disease, and its optimization. Neurosurg Focus 20:E12 Raza SM, Pradilla G, Legnani FG, Thai QA, Olivi A, Weingart JD, Brem H (2005) Local delivery of antineoplastic agents by controlled-release polymers for the treatment of malignant brain tumours. Expert Opin Biol Ther 5:477–494 Rosenberg GA, Kyner WT, Estrada E (1980) Bulk flow of brain interstitial fluid under normal und hyperosmolar conditions. Am J Physiol 238:42–49 Saito R, Krauze MT, Noble CO, Drummond DC, Kirpotin DB, Berger MS, Park JW, Bankiewicz KS (2006) Convectionenhanced delivery of Ls-TPT enables an effective, continuous, low-dose chemotherapy against malignant glioma xenograft model. Neuro-Oncology 8:205–214 Sampson JH, Raghavan R, Brady ML, Provenzale JM, Herndon JE II, Croteau D, Friedman AH, Reardon DA, Coleman RE, Wong T, Bigner DD, Pastan I, Rodríguez-Ponce MI, Tanner P, Puri R, Pedain C (2007) Clinical utility of a patient-specific algorithm for simulating intracerebral drug infusions. Neuro-Oncology 9:343–353 Sawyer AJ, Piepmeier JM, Saltzman WM (2006) New methods for direct delivery of chemotherapy for treating brain tumors. Yale J Biol Med 79:141–152 Seunguk O, Odland R, Wilson SR, Kroeger KM, Liu C, Lowenstein PR, Castro MG, Hall WA, Ohlfest JR (2009) Improved distribution of small molecules and viral vectors in the murine brain using a hollow fiber catheter. NeuroOncology 9:343–353 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphroorn MJB, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 Varenika V, Dickinson P, Bringas J, LeCouteur R, Higgins R, Park J, Fiandaca M, Berger M, Sampson J, Bankiewicz K (2008) Detection of infusate leakage in the brain using realtime imaging of convection-enhanced delivery. J Neurosurg 109:874–880 Yang W, Huo T, Barth RF, Gupta N, Weldon M, Grecula JC, Ross BD, Hoff BA, Chou TC, Rousseau J, Elleaume H (2011) Convection enhanced delivery of carboplatin in combination with radiotherapy for the treatment of brain tumors. J Neurooncol 101:379–390 Yin D, Forsayeth J, Bankiewicz KS (2010) Optimized cannula design and placement for convection-enhanced delivery in rat striatum. J Neurosci Meth 187:46–51 Yokosawa M, Sonoda Y, Sugiyama S, Saito R, Yamashita Y, Nishihara M, Satoh T, Kumabe T, Yokoyama M, Tominaga T (2010) Convection-enhanced delivery of a synthetic retinoid Am80, loaded into polymeric micelles, prolongs the survival of rats bearing intracranial Glioblastoma xenografts. Tohoku J Exp Med 221:257–264

Chapter 22

Brain Metastases: Clinical Outcomes for Stereotactic Radiosurgery (Method) Ameer L. Elaimy, Alexander R. MacKay, Wayne T. Lamoreaux, Robert K. Fairbanks, John J. Demakas, Barton S. Cooke, Benjamin J. Arthurs, and Christopher M. Lee

Abstract Stereotactic radiosurgery (SRS) is a form of radiation therapy that delivers a focused, highly conformal dose of radiation to a single volume, while minimizing damage to the adjacent nervous tissue. Historically, surgical resection followed by whole brain radiation therapy (WBRT) has offered patients with a single metastatic brain tumor an improved quality of life, as well as an improved longevity, when compared to patients treated with WBRT alone or surgical resection alone. However, tumor resection with WBRT is not always the optimal treatment for all patients. Other clinical factors that impact treatment decisions are tumor location, patients who have multiple brain metastases, and patients who are in a poor medical state and are unable to undergo surgery. Due to the limitations of surgical resection alone and sideeffects of WBRT, the efficacy of SRS has been examined in the treatment of brain metastases for multiple clinical scenarios. Stereotactic radiosurgery is capable of targeting any region in the brain and can be utilized to irradiate multiple tumors in the same treatment setting in a non-invasive fashion. For many clinical situations, radiosurgery alone or radiosurgery in combination with WBRT or surgery can be an optimal treatment approach. Although many questions remain unanswered, this chapter reviews current management options for the treatment of brain metastases, along with describing the patient selection criteria, treatment planning, methods, and outcomes associated with SRS.

C.M. Lee () Department of Oncology, Cancer Care Northwest and Gamma Knife of Spokane, Spokane, WA 99204, USA e-mail: [email protected]

Keywords Metastasis · Stereotactic radiosurgery · Dose · Toxicity · Survival · Local control rate

Introduction Brain metastases are the most frequently observed cancerous lesions in the brain. Metastasis to the brain will occur in 20–40% of patients with systemic cancer and in each incidence the brain metastases have the potential to pose a serious threat to the patient’s quality of life and longevity (Smedby et al., 2009). Lung cancer is the most common form of primary cancer that has the ability to metastasize to the brain. However, melanoma, breast cancer, colorectal cancer, renal-cell carcinoma, and carcinoma of multiple other origins may also lead to brain metastases (Suh, 2010). Areas of the brain that receive a larger blood supply are more likely to develop one or more metastatic brain tumors; thus, ∼80% of brain metastases are hemispheric, while only 15% arise within the cerebellum and 5% arise within the brain stem (Delattre et al., 1988; Suh, 2010). Significant improvements in both imaging techniques and the treatment of extracranial cancer have led to early detection and an increase in the overall life expectancy of cancer patients, leaving them more susceptible to the development of brain metastases (Hazard et al., 2005). Due to the high rate of morbidity and mortality that results from metastatic brain tumors, the efficacy of multiple treatment regimens have been analyzed and compared with one another to determine the most advantageous course of treatment in selected patient groups. In general, patients who suffer from brain metastases have a poor outlook, and curative treatment

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_22, © Springer Science+Business Media B.V. 2011

217

218

is not achievable in most situations. Patients have an estimated survival time of 1–2 months when treated with corticosteroids alone and no radiation therapy or surgery (Andrews et al., 2004). Whole brain radiation therapy (WBRT) is a time-honored treatment protocol that targets rapidly dividing tumor cells in all areas of the brain with photon beams, while eliciting minimal harmful effects on the healthy surrounding brain tissue. As a result, patients treated with WBRT alone have an average survival time of 4–7 months (Hazard et al., 2005). Despite the advantages of WBRT, 50% of these patients will still die from their neurological disease (rather than their extracranial disease) (Delattre et al., 1988). Surgical resection followed by WBRT for appropriate patients with single, surgically accessible brain metastases has proven to be a superior treatment modality in studies comparing WBRT alone and surgical resection alone (Patchell et al., 1990, 1998; Vecht et al., 1993). However, for patients with multiple brain metastases, the relative efficacy and safety of tumor resection followed by WBRT has not been evaluated in a randomized trial; thus, remains a questionable and most experts would argue, an excessive course of treatment (Schackert, 2002). Stereotactic radiosurgery (SRS) is a highly technical image-guided form of radiation therapy that delivers a focused, highly conformal dose of radiation to a single intracranial volume, thereby minimizing damage to the adjacent nervous tissue. Stereotactic radiosurgery ensures precise tumor localization by immobilizing the patient’s skull in a specified fixed position and in turn precisely aiming a high dose of radiation at the tumor mass. Stereotactic radiosurgery can be delivered to the tumor volume via 3 therapeutic devices: gamma knife (GK) radiosurgery, linear accelerator (LINAC) based treatment, and a cyclotron-based proton beam. Of the 3 modalities, GK radiosurgery and LINAC-based treatment are the most frequent methods utilized (Hazard et al., 2005). Very few institutions use a cyclotronbased proton beam due to the high price of the machine and space requirements (Suh, 2010). As SRS has proven to be a viable course of treatment, numerous studies have been conducted on patient groups with single and multiple brain metastases to evaluate their treatment-specific outcomes compared with groups treated with the traditional modalities. As the evidence examining the role of SRS in the treatment of brain metastases accumulates, it is of utmost importance for physicians to understand the criteria associated with

A.L. Elaimy et al.

SRS, so that the optimal course of treatment for their patients can be prescribed. This chapter reviews the current therapies for the treatment of brain metastases, along with describing the patient selection criteria, treatment planning, methods, and outcomes associated with SRS. This chapter will not address specific treatment recommendations for primary brain tumors.

Indications and Patient Counseling Patient Selection Patients with brain metastases are frequently assessed using the Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis (RPA) prognostic system (Gaspar et al., 1997). This system categorizes patients into 3 different prognostic groups (classes 1, 2, or 3) based on the age of the patient, presence of extracranial metastases, Karnofsky Performance Score (KPS), and control of primary cancer. Statistically, a higher class (class 2 and 3) represents a worse prognosis for the patient. Phase III evidence suggests that surgical resection followed by WBRT for patients with a single, surgically accessible brain metastasis who have a good performance status (functionally independent and spend no more than 50% of their day in bed) is a more effective treatment modality when compared to WBRT alone or surgical resection alone (Patchell et al., 1990, 1998; Vecht et al., 1993). However, for patients with multiple brain metastases, surgical resection of all lesions followed by WBRT is thought to be an extreme course of treatment, usually with unacceptable outcomes, and has not been analyzed in a randomized study. For patients with multiple metastases, a combination of WBRT alone, SRS alone, or a combination of SRS with WBRT can be the optimal approach for select patient subsets. In addition to the 4 criteria described by the RTOG partitioning analysis, the outcomes of SRS are also influenced by the number, location, and size of metastatic brain tumors the patient has at the time of diagnosis. Literature reviews have demonstrated that patients with a KPS ≥ 70 and a single brain metastasis, who receive SRS treatment combined with WBRT survive a longer period of time than patients treated with WBRT alone (Linskey et al., 2010). Surgical resection followed by WBRT for a single brain metastasis is

22 Clinical Outcomes for Stereotactic Radiosurgery (Method)

219

recommended for patients who present with rapid neurologic deterioration due to a dominant brain tumor, a ventricular obstruction, or a tumor of a large diameter of ≥ 40 mm (which many times leads to a mass effect) (Hazard et al., 2005). Stereotactic radiosurgery is warranted when the patient has controlled neurological symptoms (with or without steroids), a tumor/s ≤ 30 mm in diameter, a single brain metastasis, a surgically inaccessible brain metastasis, or multiple brain metastases in specific clinical situations (Hazard et al., 2005). In a review analyzing the role of SRS in the management of patients with newly diagnosed brain tumors, Linskey et al. (2010) found evidence that SRS with WBRT leads to an increased level of local tumor control and functional independence for patents with 1–4 brain metastases who have a KPS ≥ 70 and an increased survival time for patients with 2–3 brain metastases, when compared to patients treated with WBRT alone.

hair loss, and hyperpigmentation (Kaal et al., 2005; Peacock and Lesser, 2006). The long-term side effects include hearing loss, decrease in neurocognitive function, alopecia, changes in behavior, ataxia, urinary incontinence, necrosis from radiation, and potential somnolence syndrome (Kaal et al., 2005; Peacock and Lesser, 2006). Gaspar et al. (2010) conducted a review evaluating WBRT in the management of patients with brain metastases and concluded that altered dose/fractionation schedules do not exhibit a compelling difference in survival, local control, or neurocognitive outcomes when compared with “customary” dosing (i.e., biologically effective dose of 30 Gy in 10 fractions).

Whole Brain Radiation Therapy for Single and Multiple Brain Metastases

After the diagnosis of a single brain metastasis with magnetic resonance (MR) imaging, neurosurgical intervention in appropriate patients, with open-skull craniotomy or tumor resection followed by WBRT has proven to be a more effective treatment method when compared with WBRT alone. In 1990, Patchell et al. published a study investigating the role of surgical resection followed by WBRT in the treatment of patients with single brain metastases. The authors randomized a total of 48 patients, where 25 were treated with neurosurgery followed by WBRT, while the other 23 were treated with WBRT alone. The prescribed radiation dose for both groups was uniform, with a total of 36 Gy being delivered in 12 daily fractions of 3 Gy each. Patients eligible for the study were able to care for themselves independently, with a KPS ≥ 70. Patients treated with surgery and WBRT experienced a statistically significant increase in median survival time (P < 0.01), living 40 weeks, whereas patients in the WBRT alone group lived a median of 15 weeks following treatment. Also, the addition of surgery substantially increased the time for the brain metastases to reoccur (P < 0.02) and functional independence (P < 0.005). In 1993, another study was performed by Vecht et al. which randomized 63 patients with single brain metastases to a surgical resection followed by WBRT group and a WBRT alone group. The authors delivered

Whole brain radiation therapy is historically recognized to be the most common treatment utilized for approximately 70–80% of patients diagnosed with brain metastases (Kirsch and Loeffler, 2004; Peacock and Lesser, 2006). The aim of WBRT is to destroy metastatic tumor cells, while sparing as much as possible of the neighboring, functional brain tissue from acute and late side-effects from radiation therapy (Peacock and Lesser, 2006). Patients who possess either single or multiple inoperable brain tumors, or brain tumors that are thought to be too large for SRS are often treated with WBRT. Whole brain radiation therapy is an effective treatment modality due to the fact that tumor cells undergo mitosis much more rapidly than nervous tissue, making them more susceptible to radiation cell-kill (Goodhead, 1994). This difference in radiobiology provides an advantageous therapeutic ratio. However, even though radiation therapy has the capability to eradicate tumor cells more frequently than healthy cells, it is still possible for the patient to experience a variety of acute, subacute, or late side-effects from treatment. The acute side-effects from WBRT include fatigue, headache, nausea, impaired sense of taste, erythema, alopecia,

Surgical Resection with Whole Brain Radiation Therapy for a Single Brain Metastasis

220

a novel, but constant, radiation schedule between the 2 groups. Specifically, a total of 40 Gy were prescribed to the patients, with 2 Gy being delivered twice per day over 10 treatment days. The patients who participated in the study had a moderate quality of life and did not spend more than 50% of their day in bed. It was reported that the surgery and WBRT group survived a median of 10 months, which was a statistically significant (P = 0.04) improvement over the WBRT alone group, which survived a median of 6 months. Although the tendency did not reach full statistical significance (P = 0.06), it was observed that the surgery and WBRT group experienced an increased period of functional independence when compared with the WBRT alone group. Due to these data, surgical resection followed by WBRT was considered to be the benchmark treatment modality for patients with a single brain metastasis.

Radiosurgery with Whole Brain Radiation Therapy for a Single Brain Metastasis After it was demonstrated that surgery combined with WBRT improved outcomes for patients with a single brain metastasis, the role of SRS with WBRT was soon investigated. As discussed earlier, neurosurgical resection followed by WBRT is the suggested course of treatment when the patient experiences a severe neurologic deficit, ventricular obstruction, or a substantial mass effect (Hazard et al., 2005). Because large tumors (≥ 40 mm) tend to cause mass effect, surgery is also the desired modality for patients with metastatic brain tumors of a large volume (Hazard et al., 2005; Suh, 2010). In appropriate patients that exhibit manageable neurocognitive symptoms, have a tumor diameter ≤ 30 mm, or possess an unressectable brain tumor, SRS has proven to be a respected treatment modality. Andrews et al. (2004) evaluated 333 patients with 1–3 brain metastases that were treated with either SRS with WBRT or WBRT alone. Precisely, 167 patients were randomized into the SRS with WBRT group, while 164 patients were randomized into the WBRT alone group. Stereotactic radiosurgery was performed with either GK radiosurgery or LINAC-based treatment. It was found that there was no statistical significance in survival between the 2 treatment groups (P = 0.1356). Even though the 2 groups, when compared

A.L. Elaimy et al.

as a whole, did not differ in terms of median survival in subset analysis, those patients with single metastatic brain tumors in the SRS with WBRT group experienced a statistically significant (P = 0.0393) increase in survival time (6.5 vs. 4.9 months). Other sub-groups within the SRS with WBRT group that had an increased survival time were patients categorized in the RPA class 1, patients with nonsmall cell lung cancer, or patients with squamous cell lung cancer. Also, the authors observed that patients in the SRS with WBRT group were more inclined to have a stable or improved KPS score at 6 months of follow-up.

Radiosurgery with Whole Brain Radiation Therapy for Multiple Brain Metastases Other clinical research has reported positive results when comparing SRS with WBRT and WBRT alone in patients who suffer from multiple brain metastases. A study published by Kondziolka et al. in 1999 investigated the role of SRS with WBRT versus WBRT alone in patients with 2–4 brain metastases. A total of 27 patients with a KPS ≥ 70, 2–4 metastatic brain tumors, and tumor diameters ≤ 25 mm were randomized into the SRS with WBRT and WBRT alone groups. Precisely, 13 patients were placed in the radiosurgery group, while 14 patients were placed in the WBRT alone group. All 27 patients received a total radiation dose of 30 Gy delivered in 12 daily fractions of 2.5 Gy each. Stereotactic radiosurgery was performed using GK radiosurgery. The 13 patients that underwent GK radiosurgery received a tumor marginal dose of 16 Gy. Because the authors witnessed a drastic difference in tumor control between the 2 groups, the study was stopped at the 60% accrual point. The authors observed that the SRS with WBRT group lived a median of 11 months, whereas the WBRT alone group lived a median of 7.5 months. Differences in survival times did not reach statistical significance (P = 0.22) because there were a limited number of patients in the study. The radiosurgery group exhibited a substantially better local failure rate at 1 year (8%) and median time of recurrence (36 months) compared with the WBRT alone group, which had a 100% local failure rate at 1 year and a median time for recurrence of 6 months at the original site.

22 Clinical Outcomes for Stereotactic Radiosurgery (Method)

221

Sanghavi et al. (2001) evaluated 502 patients treated with SRS with WBRT from databases from 10 institutions. The authors excluded patients who had previously underwent WBRT or surgical resection, those with incomplete follow-up data, those not being treated for all brain metastases with radiosurgery, and those who had treatment after the specified cut-off date. Of the 10 institutions, GK radiosurgery was used in 3, LINAC-based treatment was used in 6, and 1 institution utilized both radiosurgical techniques. Each of the institutions prescribed radiation doses based on preference. The authors concluded that the overall median survival time was 10.7 months. Specifically, patients with a higher KPS (P = 0.0001), a controlled primary cancer (P = 0.0023), the absence of extracranial cancer (P = 0.0001), and a lower RPA class (P = 0.000007) were likely to survive an increased period of time when compared with those who did not possess those characteristics. Although this trial was not randomized, the results indicate that radiosurgery improves patient outcomes in RPA classes 1 through 3. Also, the authors found that the addition of radiosurgery improved survival times when compared with patients treated with solely WBRT in another study previously performed by the RTOG (Borgelt et al., 1980).

resection with WBRT group and the surgical resection alone group. Cumulative evidence suggests that SRS alone may provide equivalent levels of survival when compared with patients treated with both SRS and WBRT (Linskey et al., 2010). However, there is conflicting data regarding the 2 treatment modalities when analyzing the risk of both local and distant tumor control. Aoyama et al. (2006) randomized 132 patients into either a SRS with WBRT (65 patients) group or a SRS alone group (67 patients). There were no statistically significant differences in the 2 groups, with respect to survival, death due to neurological causes, and toxicity. It was observed that the SRS and WBRT group experienced a statistically significant (P < 0.001) lower local and distant 12-month actuarial brain tumor recurrence rate when compared with the SRS alone group. Sneed et al. (2002) performed a retrospective study evaluating 569 patients from 10 institutions treated with either SRS with WBRT (301 patients) or SRS alone (268 patients). After adjusting median survival time for each RPA class, it was found that there were no statistically significant differences in overall survival between the 2 groups, with the SRS and WBRT group surviving a median of 8.6 months and the SRS alone group surviving a median of 8.2 months. It was documented that only 7% of patients in the SRS with WBRT group required salvage treatment, whereas 37% of patients in the SRS alone group required salvage treatment. This indicates an increase in the local control rate in the SRS with WBRT group. Li et al. (2000) analyzed 29 patients treated with WBRT alone, 23 patients treated with SRS alone, and 18 patients treated with SRS with WBRT. Only patients with small-cell lung cancer (SCLC) and nonsmall-cell lung cancer (NSCLC), a single metastatic brain tumor ≤ 45 mm in diameter, and a KPS ≥ 60 were eligible for the study. Of the SRS with WBRT group and SRS alone group, the authors reported that there were no statistically significant differences in median survival. In contrast to the previous studies mentioned, adding WBRT to SRS did not improve local control levels when compared to the SRS alone group. Although this combination of data indicates patients who undergo SRS alone require salvage therapy more often than patients treated with SRS with WBRT, patients treated with both modalities should be monitored closely following treatment

Surgical Resection or Radiosurgery Alone After the efficacy of surgical resection followed by WBRT was demonstrated in patients with single brain metastases, questions arose pertaining to the clinical outcomes of patients treated with solely surgical resection. In 1998, Patchell et al. performed a study where 95 patients with single brain metastases were randomized to either surgical resection followed by WBRT (49 patients) or surgical resection alone (46 patients). The authors reported that patients in the postoperative WBRT group experienced less frequent tumor recurrence at the site of the original metastasis (P < 0.001) and less frequent tumor recurrence anywhere in the brain (P < 0.001). In addition, patients who received WBRT were less likely to die from neurological causes when compared with the surgical resection alone group (P = 0.003). Interestingly, there was not a statistically significant difference in median survival time and functional independence between the surgical

222

so appropriate patients can begin salvage procedures at the earliest possible time.

Impact of Various Histologic Subtypes The influence of primary tumor histology on the effectiveness of SRS is an area of controversy in the treatment of patients with brain metastases. Historically, melanoma and renal cell carcinoma have been identified as “radioresistant” histological types because they have displayed a poor response to standard radiation therapy when compared with cancers of other origins. Nieder et al. (1997) analyzed 108 patients (336 brain metastases) before and after WBRT using contrast-enhanced CT scans. The authors observed that complete response after WBRT occurred in 37% of metastases from small-cell carcinoma, 35% of metastases from breast cancer, 25% of metastases from squamous-cell carcinoma, and 14% of metastases from nonbreast adenocarcinoma. The complete response for melanoma and renal cell carcinoma was found to be 0% for each. On the contrary, Flickinger et al. (1994) conducted a study where 116 patients from 5 institutions underwent GK radiosurgery for the treatment of solitary brain metastases. It was reported that tumor histology influenced survival time, with breast cancer patients experiencing a statistically significant increase in median survival (P < 0.004). Interestingly, it was also observed that patients with melanoma and renal cell carcinoma primaries exhibited a statistically significant higher local tumor control rate (P < 0.0003) than primary cancers of other histologies. Continued research on the influence of tumor histology on SRS efficacy and the specific radiobiology of tumors will lead to a greater understanding of the outcomes following SRS treatment.

Treatment Planning and Methods Types of Radiosurgery As previously discussed, there are 3 main modalities by which SRS can be delivered (GK radiosurgery, LINAC-based treatment, and a cyclotron-based proton beam). The cyclotron-based proton beam is rarely

A.L. Elaimy et al.

used because the machine is extremely expensive and requires a great deal of space (Hazard et al., 2005). The GK device is a cobalt-60-based machine, with 201 separate 4–18 mm collimator openings that emits multiple gamma rays that converge on a target specified by computer planning (Suh, 2010). The LINAC device functions by accelerating an electron, which generates high-energy x-ray beams that are focused on a specific target at different angles by micro-multileaf collimators (Suh, 2010). Studies by Sneed et al. (2002) and Andrews et al. (2004) concluded that patient outcome is independent of mode of SRS delivery.

Dose Selection A study from the RTOG evaluating the maximum tolerated dose of single fraction radiosurgery in patients with metastatic brain tumors (64% of patients) or primary brain tumors (36% of patients) which were previously irradiated, was conducted by Shaw et al. (2000). 156 patients with cerebral or cerebellar solitary non-brainstem tumors which were ≤ 40 mm in diameter participated in the study. The maximum tolerated doses for tumors ≤ 20 mm was 24 Gy, 18 Gy for tumors 21–30 mm in diameter, and 15 Gy for tumors 31–40 mm in diameter. The authors reported 3 variables associated with grade 3–5 toxicity: maximum tumor diameter, KPS, and tumor dose. It was found that compared to tumors < 20 mm, tumors 21–40 mm in diameter were 7.3–16 times more likely to encounter grade 3–5 toxicity. Grade 3 toxicity was edema, which occurred a median of 4.5 months following treatment. Grade 4 toxicity was radionecrosis, which occurred at an incidence of 5% at 6 months, 8% at 12 months, 9% at 18 months, and 11% at 24 months after SRS was administered. Additionally, risk of local progression was higher in patients with primary brain tumors than in patients with metastatic brain tumors. It was concluded that large tumors (> 20 mm) were more likely to encounter toxicities, while local control was associated with the histologic type of brain tumor. In 2004, Shehata et al. published a study assessing 160 patients (468 brain metastases ≤ 20 mm in diameter). The SRS dose ranged from 7 to 30 Gy (median of 20 Gy). 240 patients received WBRT (49%), in which the dose ranged from 6.75 to 50.4 Gy (median of 40.5 Gy). Patients who receieved WBRT in addition to

22 Clinical Outcomes for Stereotactic Radiosurgery (Method)

223

SRS, with a SRS dose of ≥ 20 Gy, had a local control rate of 99%, whereas patients who received a SRS dose < 20 Gy had a local control rate of 91%. Interestingly, SRS doses > 20 Gy did not increase the patient’s local control rates, and increasing doses resulted in a near statistically significant increased level of grade 3 or 4 toxicities (P = 0.078).

(2004) evaluating the efficacy of WBRT with or without SRS found that the group treated with SRS with WBRT had a statistically significant improvement in local control rate (P = 0.01) when compared to the WBRT alone arm, along with patients having an improved survival time with a single brain metastasis, RPA class 1, NSCLC, or squamous-cell lung cancer. Evidence from the studies by Li et al. (2000) and Andrews et al. (2004), along with the findings already discussed by Kondziolka et al. (1999), who witnessed a substantially improved local control rate (P = 0.0016) in the SRS with WBRT arm, indicates that SRS combined with WBRT produces significantly better local tumor control when compared with patients treated with WBRT alone. This improvement in local tumor control is more pronounced in patients with a KPS ≥ 70 who have 1–4 brain metastases (Linskey et al., 2010). A subject matter that requires further experience and research is the effect of SRS with or without WBRT on local and distant tumor control. Aoyama et al. (2006) found that patients treated with SRS alone were more likely to have an increased 12-month brain tumor recurrence rate than patients treated with both SRS and WBRT (P < 0.001). However, several retrospective studies have reported that the SRS alone and SRS with WBRT treatment groups do not exhibit compelling statistical differences, with respect to local tumor control (Jawahar et al., 2002; Sneed et al., 1999; Wang et al., 2002). Muacevic et al. (2008) found that patients treated with GK radiosurgery do not differ in terms of survival when compared to patients treated with surgical resection followed by WBRT. The authors did conclude that postoperative WBRT led to less frequent distant tumor recurrences than within the GK radiosurgery alone group (P = 0.04). This finding that WBRT reduces the risk of distant tumor recurrences correlates with those found by Patchell et al. (1990). This data permits the close monitoring of patients treated with either modality following treatment, so that salvage therapy may proceed at the earliest possible time.

Dose Limitations of Adjacent Structures When Performing Treatment Planning At our institution (the Gamma Knife of Spokane), we follow the RTOG guidelines and adjust the dose range for tumors ≤ 20 mm in size to 20–24 Gy, depending on previous SRS and WBRT the patient has undergone in the past. Marginal doses may also be adjusted based on the vicinity of adjacent, healthy brain tissue. The RTOG guidelines (Shaw et al., 2000) for tumors 21–30 mm in diameter and tumors 31–40 mm are followed, with patients being prescribed a marginal dose of 18 and 15 Gy, respectively. In general, the normal tissue maximum tolerance doses for the brainstem ranges from 12 to 13 Gy, while the optic chiasm can receive a dose of 7–9 Gy. A dose of 8–10 Gy to the optic chiasm is possible if the patient has not received previous surgical or radiation therapy intervention. After prior external beam radiation therapy (EBRT), the normal tissue tolerance dose for the chiasm is 7 Gy. Alterations may be taken based on the patient history and tumor location. Of course, these general guidelines need to be tailored to each patient’s specific circumstances and cannot be applied safely to every patient or in every clinical situation.

Treatment Outcomes Local and Regional Tumor Control In the three-arm study already discussed by Li et al. (2000), the SRS with WBRT group did not show an increased local control rate when compared to the SRS alone group, but the SRS with WBRT group did show a statistically significant improvement in local control (P < 0.0001) when compared to the WBRT alone group. The RTOG study performed by Andrews et al.

Impact of Treatment on Survival The evidence previously examined pertaining to the addition of WBRT to SRS demonstrated an increased

224

median survival for patients with a single metastatic brain tumor (P = 0.0393) in the RTOG randomized trial by Andrews et al. (2004). For patients with a single brain metastasis who are treated with SRS with WBRT, survival is more pronounced if the patient has a KPS ≥ 70 (Linskey et al., 2010). The study by Kondziolka et al. (1999) evaluating patients with 2–4 brain metastases showed a trend in survival that favored the SRS with WBRT group (P = 0.22), but did not reach statistical significance due to the low number of patients. This data still supports the use of SRS with WBRT in patients with 2–4 brain metastases, but this evidence is still in evolution when compared to the evidence in treating a single brain metastasis with SRS with WBRT. SRS with WBRT and surgical resection with WBRT are known to produce equivalent survival rates. Schoggl et al. (2000) compared neurosurgery and SRS in the treatment of single brain metastases. A total of 133 patients were treated, where 67 were treated with SRS and 66 were treated with neurosurgery. The authors included all patients treated with WBRT. There was not a statistically significant difference in survival (P = 0.19) between the treatment arms. Likewise, SRS with WBRT and SRS alone both represent effective treatments, with respect to patient survival. The studies previously discussed by Aoyama et al. (2006), Sneed et al. (2002), and Li et al. (2000) all reported no statistically significant differences in median survival between the 2 treatment arms. Other studies have demonstrated that SRS alone is a superior treatment modality when compared to WBRT alone. Rades et al. (2007) investigated the outcomes of 186 patients treated with either SRS alone (95 patients) or WBRT alone (91 patients). It was found that the SRS treatment arm exhibited a longer median survival (13 months) than the WBRT alone arm (7 months), with a P value of 0.045. Even though the groups exhibited similar outcomes in distant brain control and toxicity, it was observed that the SRS alone group had substantially better 1-year local and overall brain control levels.

Neurological Acute and Late Toxicity The most frequent acute side-effects following SRS include headaches after the stereotactic head-frame is removed and screw-site soreness at the areas where the

A.L. Elaimy et al.

head-frame was attached to the patient’s skull (Suh, 2010). Other acute side-effects that are not as common range from seizures, infection at the screw-site, and the worsening of neurological symptoms for a relatively short period of time (Suh, 2010). Sub-acute and late reactions to radiosurgery are not as commonly observed. Lutterbach et al. (2003) analyzed radiosurgery alone in patients with 1–3 brain metastases and found that acute side-effects, including seizures and increased severity of pre-existing neurocognitive symptoms, were present in 9% of patients, while late side-effects, paresis and decreased visual awareness, were present in only 4% of patients. In general, late side-effects following radiosurgery include radiation necrosis, edema, the development of new neurological deficits, and the exacerbation of neurological deficits the patient has previously suffered from (Suh, 2010). Patients who are treated with steroids must be closely monitored because steroid therapy puts patients at risk for weight gain, insomnia, diabetes, psychosis, and suppression of the patient’s immune system (Suh, 2010).

Discussion In suitable patients, SRS is a beneficial modality in the treatment of brain metastases. For patients who suffer from a single brain metastasis, the RTOG randomized trial by Andrews et al. (2004) determined that those treated with SRS with WBRT experience a significantly longer period of survival than those treated with WBRT alone. In addition to improving survival time for patients with a single brain metastasis, the RTOG 95–08 trial also demonstrated an enhancement in KPS and local control in patients treated with SRS with WBRT who have 1–3 metastatic brain tumors. Although it has shown relatively successful outcomes, the survival rate for patients with ≥ 2 brain metastases ≤ 30 mm remains a controversial treatment modality and needs to be evaluated further in a Phase III study with more compelling statistical evidence than reported by Kondziolka et al. (1999). Recent publications have shown that specific patients with > 4 brain metastases have also benefited from SRS. At our institution, we do advocate treatment of selected patients with > 4 brain metastases if their performance status is high and they are undergoing active treatment

22 Clinical Outcomes for Stereotactic Radiosurgery (Method)

225

for systemic disease, or are free of known extracranial disease. This decision is based on the lack of other treatment options and because of the encouraging, evolving literature, which shows benefits for select patients. Despite the fact that the study by Schoggl et al. (2000) found equivalent survival rates for patients treated with surgical resection following WBRT versus SRS with WBRT, it is still unclear which is a superior treatment modality for specific patients with a single brain metastasis who are qualified candidates for both procedures. Investigation into this matter in the form of a randomized trial would provide the best statistical evidence in terms of survival rates to help answer this question. The most controversial subject matter in the treatment of patients with brain metastases is whether or not the addition of WBRT will provide a favorable patient prognosis when compared to those who are treated with SRS alone. It is now known that the combination of SRS with WBRT does not produce a survival advantage for all patients, however, physicians who support the use of SRS alone emphasize the fact that if tumor recurrence occurs, then salvage treatment, with either SRS or WBRT is warranted and can provide the patient with further tumor control and does impact survival outcomes for specific patients (Suh, 2010). Those who advocate the addition of WBRT to SRS note that patients who undergo SRS alone exhibit local and distant tumor recurrence within the brain more often and that disease progression can produce further neurological deficits (Suh, 2010). These can be more damaging to the patient than the potential side-effects that can occur following WBRT. In conclusion, SRS can be of clinical benefit to specific patient cohorts when utilized after surgery, with WBRT, or in combination with either or both of the treatment modalities. We agree that many questions remain unanswered, and look forward to additional clinical outcome studies, which will further guide us with future clinical treatment decisions.

radiosurgery boost for patients with one to three brain metastases: phase III results of the RTOG 9508 randomised trial. Lancet 363:1665–1672 Aoyama H, Shirato H, Tago M, Nakagawa K, Toyoda T, Hatano K, Kenjyo M, Oya N, Hirota S, Shioura H, Kunieda E, Inomata T, Hayakawa K, Katoh N, Kobashi G (2006) Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial. JAMA 295: 2483–2491 Borgelt B, Gelber R, Kramer S, Brady LW, Chang CH, Davis LW, Perez CA, Hendrickson FR (1980) The palliation of brain metastases: final results of the first two studies by the radiation therapy oncology group. Int J Radiat Oncol Biol Phys 6:1–9 Delattre JY, Krol G, Thaler HT, Posner JB (1988) Distribution of brain metastases. Arch Neurol 45:741–744 Flickinger JC, Kondziolka D, Lunsford LD, Coffey RJ, Goodman ML, Shaw EG, Hudgins WR, Weiner R, Harsh GR 4th, Sneed PK, larson DA (1994) A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys 28:797–802 Gaspar LE, Mehta MP, Patchell RA, Burri SH, Robinson PD, Morris RE, Ammirati M, Andrews DW, Asher AL, Cobbs CS, Kondziolka D, Linskey ME, Loeffler JS, McDermott M, Mikkelsen T, Olson JJ, Paleologos NA, Ryken TC, Kalkanis SN (2010) The role of whole brain radiation therapy in the management of newly diagnosed brain metastases: a systematic review and evidence-based clinical practice guideline. J Neurooncol 96:17–32 Gaspar L, Scott C, Rotman M, Asbell S, Phillips T, Wasserman T, McKenna WG, Byhardt R (1997) Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys 37:745–751 Goodhead DT (1994) Initial events in the cellular effects of ionizing radiations: clustered damage in DNA. Int J Radiat Biol 65:7–17 Hazard LJ, Jensen RL, Shrieve DC (2005) Role of stereotactic radiosurgery in the treatment of brain metastases. Am J Clin Oncol 28:403–410 Jawahar A, Willis BK, Smith DR, Ampil F, Datta R, Nanda A (2002) Gamma knife radiosurgery for brain metastases: do patients benefit from adjuvant external-beam radiotherapy? An 18-month comparative analysis. Stereotact Funct Neurosurg 79:262–271 Kaal EC, Niel CG, Vecht CJ (2005) Therapeutic management of brain metastasis. Lancet Neurol 4:289–298 Kirsch DG, Loeffler JS (2004) Treating brain metastases: current approaches and future directions. Expert Rev Neurother 4:1015–1022 Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC (1999) Stereotactic radiosurgery plus whole brain radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat Oncol Biol Phys 45:427–434 Li B, Yu J, Suntharalingam M, Kennedy AS, Amin PP, Chen Z, Yin R, Guo S, Han T, Wang Y, Yu N, Song G, Wang L (2000) Comparison of three treatment options for single brain metastasis from lung cancer. Int J Cancer 90:37–45 Linskey ME, Andrews DW, Asher AL, Burri SH, Kondziolka D, Robinson PD, Ammirati M, Cobbs CS, Gaspar LE, Loeffler

References Andrews DW, Scott CB, Sperduto PW, Flanders AE, Gaspar LE, Schell MC, Werner-Wasik M, Demas W, Ryu J, Bahary JP, Souhami L, Rotman M, Mehta MP, Curran WJ Jr (2004) Whole brain radiation therapy with or without stereotactic

226 JS, McDermott M, Mehta MP, Mikkelsen T, Olson JJ, Paleologos NA, Patchell RA, Ryken TC, Kalkanis SN (2010) The role of stereotactic radiosurgery in the management of patients with newly diagnosed brain metastases: a systematic review and evidence-based clinical practice guideline. J Neurooncol 96:45–68 Lutterbach J, Cyron D, Henne K, Ostertag CB (2003) Radiosurgery followed by planned observation in patients with one to three brain metastases. Neurosurgery 52:1066– 1073. discussion 1073–1064 Muacevic A, Wowra B, Siefert A, Tonn JC, Steiger HJ, Kreth FW (2008) Microsurgery plus whole brain irradiation versus Gamma Knife surgery alone for treatment of single metastases to the brain: a randomized controlled multicentre phase III trial. J Neurooncol 87:299–307 Nieder C, Berberich W, Schnabel K (1997) Tumor-related prognostic factors for remission of brain metastases after radiotherapy. Int J Radiat Oncol Biol Phys 39:25–30 Patchell RA, Tibbs PA, Regine WF, Dempsey RJ, Mohiuddin M, Kryscio RJ, Markesbery WR, Foon KA, Young B (1998) Postoperative radiotherapy in the treatment of single metastases to the brain: a randomized trial. JAMA 280:1485–1489 Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ, Maruyama Y, Kryscio RJ, Markesbery WR, Macdonald JS, Young B (1990) A randomized trial of surgery in the treatment of single metastases to the brain. N Engl J Med 322:494–500 Peacock KH, Lesser GJ (2006) Current therapeutic approaches in patients with brain metastases. Curr Treat Options Oncol 7:479–489 Rades D, Pluemer A, Veninga T, Hanssens P, Dunst J, Schild SE (2007) Whole-brain radiotherapy versus stereotactic radiosurgery for patients in recursive partitioning analysis classes 1 and 2 with 1 to 3 brain metastases. Cancer 110: 2285–2292 Sanghavi SN, Miranpuri SS, Chappell R, Buatti JM, Sneed PK, Suh JH, Regine WF, Weltman E, King VJ, Goetsch SJ, Breneman JC, Sperduto PW, Scott C, Mabanta S, Mehta MP (2001) Radiosurgery for patients with brain metastases: a multi-institutional analysis, stratified by the RTOG recursive partitioning analysis method. Int J Radiat Oncol Biol Phys 51:426–434

A.L. Elaimy et al. Schackert G (2002) Surgery of brain metastases – pro and contra. Onkologie 25:480–481 Schoggl A, Kitz K, Reddy M, Wolfsberger S, Schneider B, Dieckmann K, Ungersbock K (2000) Defining the role of stereotactic radiosurgery versus microsurgery in the treatment of single brain metastases. Acta Neurochir (Wien) 142:621–626 Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, Farnan N (2000) Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90–05. Int J Radiat Oncol Biol Phys 47:291–298 Shehata MK, Young B, Reid B, Patchell RA, St Clair W, Sims J, Sanders M, Meigooni A, Mohiuddin M, Regine WF (2004) Stereotatic radiosurgery of 468 brain metastases < or =2 cm: implications for SRS dose and whole brain radiation therapy. Int J Radiat Oncol Biol Phys 59:87–93 Smedby KE, Brandt L, Backlund ML, Blomqvist P (2009) Brain metastases admissions in Sweden between 1987 and 2006. Br J Cancer 101:1919–1924 Sneed PK, Lamborn KR, Forstner JM, McDermott MW, Chang S, Park E, Gutin PH, Phillips TL, Wara WM, Larson DA (1999) Radiosurgery for brain metastases: is whole brain radiotherapy necessary? Int J Radiat Oncol Biol Phys 43:549–558 Sneed PK, Suh JH, Goetsch SJ, Sanghavi SN, Chappell R, Buatti JM, Regine WF, Weltman E, King VJ, Breneman JC, Sperduto PW, Mehta MP (2002) A multi-institutional review of radiosurgery alone vs. radiosurgery with whole brain radiotherapy as the initial management of brain metastases. Int J Radiat Oncol Biol Phys 53:519–526 Suh JH (2010) Stereotactic radiosurgery for the management of brain metastases. N Engl J Med 362:1119–1127 Vecht CJ, Haaxma-Reiche H, Noordijk EM, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR et al (1993) Treatment of single brain metastasis: radiotherapy alone or combined with neurosurgery? Ann Neurol 33:583–590 Wang LG, Guo Y, Zhang X, Song SJ, Xia JL, Fan FY, Shi M, Wei LC (2002) Brain metastasis: experience of the Xi-Jing hospital. Stereotact Funct Neurosurg 78:70–83

Chapter 23

Noninvasive Treatment for Brain Tumors: Magnetic Resonance-Guided Focused Ultrasound Surgery Ernst Martin and Ferenc A. Jolesz

Abstract Over the last decade, the development of transcranial MR-image guided focused ultrasound surgery (tcMRgFUS) has, as a non-invasive methodology, initiated interesting alternatives to traditional invasive tumor surgery. The tcMRgFUS therapy delivery system integrates advanced MR image acquisition modes, e.g., precise anatomical information, local temperature maps and tissue stiffness, with complex, computer controlled large phased array ultrasound transducers to provide closed-loop therapy planning, soft tissue thermal ablation, without damaging surrounding normal tissue, treatment monitoring and assessment of the tissue coagulation. These procedures can be accomplished fully automated and first clinical trials are in progress at specialized centers. Moreover, besides treating tumors by high energy MRgFUS thermal ablation, low energy tcMRgFUS combined with microbubbles (e.g., ultrasound contrast agents) injected into the systemic circulation offers new therapeutic solutions for adjuvant tumor pharmacotherapy such as blood-brain-barrier opening and targeted drug delivery or even gene therapy. The combination of these technical advancements may generate a paradigm shift in the field of neuro-oncology by opening a novel, comprehensive and non-invasive strategy for brain tumor treatment with locally enhanced chemotherapy. Keywords Focused Ultrasound · MRI · tcMRgFUS · HIFU Ablation · BBB · Tumor

E. Martin () University Children’s Hospital, CH-8032 Zurich, Switzerland e-mail: [email protected]

Introduction In 1942 Lynn and colleagues reported for the first time on the ability of High Intensity Focused Ultrasound (HIFU) energy to produce focal lesions deep inside the brain and spinal cord without damaging surrounding non-targeted tissue. While sonicating various cortical, subcortical and cerebellar structures through the intact skull in experimental animals they produced well defined tissue changes which resulted in reversible and irreversible specific behavioural and neurological dysfunctions. However, the energy delivery invariably damaged the skin and underlying tissue, including scalp, muscles and even meninges in the area where they placed the apparatus (Lynn et al., 1942). As a consequence, during the following years HIFU was applied to the brain mostly through craniectomies to avoid skull heating, beam distortion and reflection, and extremely strong attenuation when ultrasound is passing through the skull bone. With this technique it became possible to successfully coagulate well defined tissue volumes in the brain by heat induced protein denaturation and coagulation, selectively and precisely, and without damaging the surrounding structures, a procedure used to be called “trackless surgery” (Lele, 1962) and now called as Focused Ultrasound Surgery (FUS). The establishment of computer control and image guidance with b-mode ultrasound led to the development of a new generation of FUS devices that has been used in treating various organs, yet with two major limitations: (1) low image resolution and image quality, which made accurate lesion localization and targeting difficult and (2) inability to monitor and/or measure temperature induced tissue changes at the focal spot (Fig. 23.1).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_23, © Springer Science+Business Media B.V. 2011

227

228

E. Martin and F.A. Jolesz

a

skull

tumor skin skull window ultrasound transducer

water

W.J. Fry in the 1970s Early b-mode US-image guided HIFU system developed by the Fry team for brain interventions in animals and humans.

b

c

B-mode real time imaging during HIFU intervention on human brain tumor by F.J. Fry and R. Heimburger 1974

Fig. 23.1 (a) Schematic diagram of a HIFU system for open skull human brain surgery used in the early 1970s and 1980s. (b) Early b-mode guided HIFU system for open skull brain tumor

surgery constructed and used by the Fry team in the 1970s. (c) Ultrasound images of a brain tumor during FUS treatment by Heimburger and Fry 1974 (from Heimburger, 2005)

Without any available monitoring and control method extensive experimental studies were necessary to elucidate the correct HIFU parameters, such as the optimal ultrasound frequency, power and duration of

sonication. In their pioneering research William and Francis Fry conveyed experimental and clinical studies in collaboration with colleagues and neurosurgeons for many decades (Fry et al., 1955, 1958; Fry and

23 Magnetic Resonance-Guided Focused Ultrasound Surgery

Fry 1960; Fry 1958). While continuously improving this promising technology interesting and potentially groundbreaking results were reported for both brain tumor surgery and for stereotactic functional neurosurgery in animals and humans. Rigid head fixation and the use of skull x-ray with direct ventriculography was used to obtain landmarks allowing stereotactic functional neurosurgery with astonishingly high precision. Small targets in the basal ganglia, thalamus and subthalamic area (ansa lenticularis, medial globus pallidus and substantia nigra) were targeted predominantly for hyperkinetic and hypertonic movement disorders. They also perfomed amygdalotomy for violent behaviour and medial thalamotomy against intractable pain and paresthesia with good clinical results (Meyers et al., 1960; Fry et al., 1962; Heimburger, 2005). These results represented major progress in treating a variety of neurological diseases and brain tumors with FUS and demontrated the great potential of brain FUS as an alternative to more invasive brain surgery. Despite this success, it took decades and significant changes in the technology before a broader community of clinicians became interested in this attractive procedure. Possible reasons for the reluctance to accept this novel technology for interventional neurology and non-invasive brain tumor surgery were: (a) the fact that ultrasound is strongly attenuated by the skull bone, which made a craniectomy necessary leaving FUS mediated interventions still invasive procedures. (b) the lack of an imaging modality that can localize the tumors or other anatomical targets, (c) the lack of temperature measurements and continuous temperature control at the focal spot and in the immediate adjacent tissue, and (d) the difficulty of steering and controlling the ultrasound beam to obtain a precise focus. In an attempt to make the FUS interventions fully non-invasive, Fry et al. (1981) and colleagues tried to sonicate through the intact skull without making an acoustic window by a craniectomy but they were unable to obtain sharp focus. It became clear that the phase aberration induced by the human skull resulted in distorted and shifted foci. A solution to this problem should only become realistic in the 1990s with the development of phased array transducers. This new transducer technology together with the integration of the FUS device with magnetic resonance imaging (MRI) over the last decade resolved most of the limitations of FUS and represented a major

229

step towards fully non-invasive, image guided and temperature controlled brain interventions (Cline et al., 1992; Hynynen et al., 1998; Tanter et al., 1998; Jolesz and Hynynen, 2002; Aubry et al., 2003; Pernot et al., 2003).

Focused Ultrasound Technology and Clinical Applications Principles of Focused Ultrasound Magnetic Resonance guided focused ultrasound surgery (MRgFUS) is a quickly developing technology that uses heat and mechanical vibration from therapeutic, high-intensity focused ultrasound (HIFU) along with MR-based image guidance to cause desirable biological effects. It is a highly promising and newly emerging method for non-invasive interventions in the brain with potential applications for the treatment of benign and malignant tumors, vascular diseases, e.g., thrombolysis in case of thrombosis, or arterial occlusion in arteriovenous malformations, and for targeted drug delivery. After more than half a century of technical development, numerous animal experiments and pioneering clinical studies transcranial MRgFUS stays at the threshold for multiple clinical applications. Ultrasound produces mechanical and elastic vibrations in tissues resulting in energy absorption. This energy deposition may lead to heat generation, mechanical force induced rupture of tissue structures, or even cavitation with sudden explosive release of dissolved gases. Whereas the thermal effects have been well described, the nonthermal processes especially the acoustic streaming and the formation of gaseous micro-bubbles as a result of negative pressure waves are less well understood. Thermal heating of tissue is dependent on the intensity of the acoustic energy applied and the absorption coefficient of the tissue involved. In addition, ultrasound beam focusing results in much higher energy density at the focus than outside the focus, which permits temperature elevations of 20–30◦ C above body temperature at the focal target. Therefore, in combination with beam focusing and optimum parameter selection of (a) ultrasound frequency or wave length, respectively, (b) radiation intensity and (c) exposure time, called sonication duration, focused high intensity ultrasound produces precise focal heat induced

230

lesions, i.e. tissue denaturation and coagulation, in the depth of the body within seconds and without injuring intervening non-targeted normal tissue. This entire procedure is called Focused Ultrasound Surgery (FUS).

Magnetic Resonance Guided Brain-Interventions with HIFU In addition to focusing based thermal ablation, modern FUS device for non-invasive, transcranial interventions, such as the Exablate 4000 (InSightec TiratCarmel, Israel) make use of two important technical developments: First, multielement (up to 1000) ultrasound phased array technology that distributes the necessary energy along the larger skull surface area. This phased array arrangement (Fig. 23.2) not only helps to avoid unnecessary skull heating but also can correct for beam aberration by the cranial bone with uneven thickness. Second, intraoperative real time magnetic resonance image guidance (MRI) and magnetic resonance thermometry (MRT) that assures localization, targeting, monitroring of the procedure and closed-loop control of energy deposition. Earlier trials of transcranial FUS mostly resulted in unpredictable ultrasound energy absorption by the skull with focal scalp and bone heating, while the phase aberration of the ultrasound beam prevented perfect focusing. The development of multichannel phased array transducer technology in combination with ultrasound field modelling based on data obtained from computerized tomography allows to correct distorted ultrasound beams and reduces the need of mechanical beam steering (Thomas and Fink, 1996; Hynynen and Jolesz, 1998; Clement and Hynynen, 2002). In addition, the specific energy per transducer element could be brought down to just a few watts per element compared to hundreds of watts in a single focused applicator. The large region of the transducer filed makes the phased array transducer an ideal applicator, since the totally applied energy per unit time can be distributed over a much larger surface area of the skull. The Exablate 4000 consists of a 30 cm diameter hemispherical phased array transducer with 1024 elements operating at 650 kHz. This device is coupled with a 1024-channel rf-amplifier, which allows phase

E. Martin and F.A. Jolesz

and amplitude control of each singe transducer element in the phased array (Fig. 23.2). To compensate for beam aberration induced by the skull bone which has uneven thickness and bone density, the planning software calculates phase offsets according to the acoustic field modeling with CT scan data showing geometry and density of the cranium. The integration of the high energy focused ultrasound system into a magnetic resonance scanner not only provided improved imaging capabilities with excellent soft tissue contrast of MRI, but it was a real break through and an important step towards the concept of “ideal surgery” (Jolesz and Hynynen, 2002). This concept involves the complete removal of the tumor, without damaging the surrounding tissue. MRI provides the full visual control of the entire procedure from correctly localizing the tumor, exactly evaluating the tumor margins, calculating the tumor volume, monitoring the temperature and energy deposition at the target, and controlling the thermal dose within the entire tumor volume, all in real time. The noninvasive treatment of brain tumors without disturbing adjacent, functionally intact structures can be achieved using software, which integrates the MR scanner with the FUS system in a unified console allowing the physician to generate the entire treatment plan and to monitor the whole procedure online. In addition, MRT allows accurate measurement of the temperature at the focal spot and to create a thermal map for estimating the thermal gradient in the immediate vicinity with accuracy in the range of ± 1◦ C at 3T field strength. This is accomplished by rapidly measuring changes in water proton resonance frequency shifts, estimated from phase difference images at intervals of 3–5 s (Ishihara et al., 1995; Hynynen et al., 1997; Kuroda et al., 1998). Calculating temperature maps during the sonication provides information about the regions of the tumor, which have received enough energy that the tumor cells are beyond doubt coagulated, and thus, allows the operator to guide the next step of the procedure. Image guidance with MRI and quantifying temperature changes with MRT permits visualization of the whole non-invasive brain intervention and makes closed loop feedback control over the entire ablation procedure with HIFU possible. This significantly enhances the safety and efficacy of the FUS procedure for brain tumor surgery and other interventions enormously (Jolesz et al., 2005).

23 Magnetic Resonance-Guided Focused Ultrasound Surgery

231

a phased array ultrasound transducer

water circulation, cooling & degassing system

MRI room

MRI bore

rubber membrane

stereotactic frame

intact skull

FUS-workstation for treatment planning, thermometry & dosimetry

1024 channel driving system mechanical positioning system

degassed water

MRI-Table

MRI-workstation for imaging and targeting

b

Fig. 23.2 (a) Schematic diagram of the transcranial MRgFUS device. (b) Patient positioned on MR-table in US transducer with stereotactic frame

Brain Tumor Ablation with Transcranial Magnetic Resonance Guided Focused Ultrasound Brain tumor ablation with little or minimally invasive stereotactic procedures, such as laser, gammaknife, or radiofrequency have been performed for many decades (Arbour, 1994; Anzai et al., 1995). The use of focused ultrasound, however, was hampered by the fact that a craniectomy was necessary to

avoid beam scattering and skull bone heating. Using this approach, Ram et al. (2006) successfully treated three patients with histologically confirmed recurrent glioblastoma multiforme. However, the feasibility and safety of this “old technology” although established over 20 years ago had to be newly proven with extracranial applications first. Then, after several years of compilation of preclinical experience of MRguided focused ultrasound brain surgery (Hynynen et al., 1997; Hynynen and Jolesz, 1998; Clement et al., 2000) that included technical ex-vivo and in-vivo

232

experiments using various prototype FUS brain systems, the FDA approved the first feasibility study worldwide for transcranial application (tcMRgFUS) using the clinical prototype ExAblate 3000, Insightec Carmel-Tirat. This study was subsequently initiated at Brigham and Women’s Hospital in Boston, in February 2005.

Patient Preparation After a thorough clinical examination and all necessary laboratory and radiological work-up, patients selected by the tumor board receive a high resolution CT, which is going to be co-registered with a volumetric 3D-T1 weighted MR imaging scan. This is needed for estimating the skull thickness at the location of each of the 1024 transducer elements to be used for phase correction of the ultrasound beam. The patient’s head has to be fully shaved in order to avoid gas bubbles in the degassed water that circulates in the space between the transducer and the head. This water interface, which is held back by a flexible membrane seal that is fitted over the patient’s head, is used as transport medium of the US beams and for cooling (17◦ C) the head surface. Before positioning the patient in the helmet-like cavity of the ultrasound transducer on the FUS treatment device, which is integrated into a standard MR table, a MR-compatible stereotactic frame (Radionics, Burlington MA) is mounted to the patient’s head for rigid fixation to the table (Fig. 23.2b).

Intervention Procedure Intervention starts with a series of contrast enhanced MR-images for coregistration and volumetric tumor segmentation. After exact tumor targeting, the first sonication is placed at a central spot in the tumor with a low dose, sub-lethal energy level sonication to confirm the targeting accuracy on the MR images. Focal point position and/or transducer location is adjusted as necessary to move the focal point to within ± 2 mm of the desired target. One sonication at therapeutic power level is delivered to confirm thermal dosimetry. The energy level and/or sonication time can be adjusted as necessary to achieve a sufficient level of energy at the focal point to coagulate the targeted volume of tissue as determined

E. Martin and F.A. Jolesz

by temperature sensitive MRI scans obtained during the sonication. Following confirmation of the proper location and dose level for treatment, the tumor is successively sonicated point by point. Typically the whole tumor volume is treated by multiple sonications of 15–25 s each until the entire volume is coagulated. The acoustic power is adjusted based on the MRI thermometry, to result in a maximum temperature between 58 and 62◦ C at the target. The subsequent cooling time between two successive sonications varies according to the results of temperature monitoring at the focus area. MR-images are collected during the entire sonication procedure for visualization and real time temperature measurements at the focal spot and the surrounding tissue (Fig. 23.3a).

Initial Results Initial findings of this worldwide first, non-invasive, transcranial brain tumor intervention in humans have been reported by the Boston group (McDannold et al., 2010). Three patients with glioblastoma multiforme have been treated with multiple ultrasound exposures. Although the intervention was limited by the maximum acoustic power provided by the prototype ultrasound device working at the frequency of 650 kHz, it marks a major step towards a completely non-invasive alternative to surgical resection of brain tumors. The fourth patient was treated with a lower frequency system (220 kHz) and with higher energy levels. In that case the tumor (thalamus glioma) was successfully coagulated but the patients expired 5 days after the treatment from delayed intraventricular hemorrhage. Based on theses experiences, further clinical studies on non-invasive brain tumor surgery using a new generation tcMRgFUS device with 1024 phased array transducer elements are presently under way in Boston, USA and Zurich, Switzerland.

Blood-Brain-Barrier Opening with Focused Ultrasound for Targeted Drug Delivery State of the art therapy of most malignant brain tumors comprises (a) gross tumor resection and (b) adjuvant chemo- and/or radiotherapy. The latter often fail to

23 Magnetic Resonance-Guided Focused Ultrasound Surgery

233

a HIFU-Ablation

tumor targeting

b

tumor targeting

HIFU-Ablation

LIFU focal opening of BBB for targeted drug delivery

Fig. 23.3 (a) tcMRgFUS treatment of brain tumors. (b) Future combined approach of tcMRgFUS brain tumor ablation followed by focal BBB opening for targeted adjuvant chemotherapy. Adapted from McDannold et al. (2010)

eradicate infiltrating glioma cells, because the effectiveness of chemotherapy is hampered by the hindrance of the therapeutic agent to enter the brain tissue through the blood-brain barrier (BBB), which results in relatively low drug concentrations at the target. This is particularly problematic in the treatment of malignant gliomas which are characterized by diffuse infiltration of brain tissue well beyond the radiologically and intraoperatively visible tumor margins. Widely infiltrating tumor cells in the brain parenchyma are supplied with oxygen and nutrients by the normal brain vasculature, while they are well protected by a BBB. Site specific and temporal opening of the BBB for targeted drug delivery represents an alternative

approach to widespread BBB disruption for delivery of chemotherapeutic and other neuroeffective agents into the brain for a number of reasons: (a) Non-selective opening of the BBB exposes the brain to undesirable substances, (b) in most situations, the BBB remains open for only very short times (minutes), so that the procedure has to be repeated. A number of laser-based techniques, such as photodynamic therapy, or focused ultrasound have accomplished the need for selective, highly localized and temporal disruption of the BBB. MR-guided, low intensity focused ultrasound (MRgLIFU) in combination with microbubbles, which are injected as ultrasound contrast agents into the bloodstream, have been proven precise site-specific

234

opening of the BBB for several hours to allow multi-fractionated drug delivery (Hynynen et al., 2005; Hynynen, 2008). This technique is particularly promising in neuro-oncology with regards to its ability to facilitate delivery of chemotherapeutic agents to infiltrating glioma cells (Kinoshita et al., 2006a; Treat et al., 2007; Mei et al., 2009).

Future Approach for Brain Tumor Therapy: Combination of Tumor Ablation and BBB-Opening Using tcMRgFUS TcMRgFUS not only has the potential of ablating tumors, but specific acoustic energy deposition can also be used to create changes in tissue permeability, e.g. opening the BBB. The future therapeutic approach for non-invasive, radical brain tumor treatment, such as malignant gliomas, will be a combination of: First, thermal ablation of core tumor tissue with tcMRgHIFU, and subsequently MRgLIFU together with ultrasound contrast agents (microbubbles) to open the BBB in order to increase the local concentration of chemotherapeutic agents (Liu et al., 2010) (Fig. 23.3a, b).

Discussion Transcranial thermal ablation of brain tumors with MRgFUS certainly marks a major milestone towards image guided, fully non-invasive neurosurgical procedures. As such MRgFUS has the potential to be a disruptive technology in the context of traditional radiation therapy like stereotactic radiosurgery, with the intent to locally destroy targets using heat, while largely sparing normal tissue. In the past, success in focused ultrasound brain interventions has been hindered by its invasive nature due to the necessity of performing a large craniectomy to obtain sufficient acoustic window. The lack of imaging prevented correct localization of tumor margins and insufficient precision in positioning the focal volume within the tumor. Without visual control of energy deposition and without any reliable temperature measurements the method was not reaching its full potential. Integrating the focused ultrasound method

E. Martin and F.A. Jolesz

with advanced image guidance and real time temperature monitoring provided by MRI resulted in a unique non-invasive complex technology that enables precise, controlled surgical interventions which can be performed even in various organs. However, treating the brain presented another seemingly irresolvable problem: the inability to focus the US beams through the intact skull. The real potential of MRgFUS has been realized since this problem was resolved by the correction of phase aberration caused by the irregular bone. Transcranial MRgFUS is a complex procedure with a rather narrow safety window. The Exablate 4000 brain system (InSightec Carmel-Tirat, Israel), however, has been brought to a very high standard device over the years allowing safe interventions without collateral damage. The procedure itself is carried out on awake, fully responsive patients, without anesthesia, so that neurological symptoms and pain from undesirable skull heating can be monitored during the whole treatment session. All brain interventions in the small number of patients treated so far have been well tolerated. However, for the time being, the degree of freedom of the electronic steering is such that centrally located tumors are best reached and peripheral tumors close to the skull bone or large vessels are not amendable with the present device configuration. To overcome these physical limits, new phased array transducers with lower ultrasound frequencies and reduced power outputs in combination with cavitation (with and without the use of microbubbles ultrasound contrast agents) will decrease the peak sound pressure amplitudes, and thus, the amount of heat deposition. There are several potential advantages of focused ultrasound surgery, when compared to traditional neurosurgical interventions: • Transcranial MRgFUS is a non-invasive procedure that uses non-ionizing radiation, and thus is a gentle and safe procedure to the patients. • MRgFUS allows volumetric targeting of the tumor under constant visual control. • MRgFUS permits real time temperature monitoring of the whole procedure for readjusting sonication parameters, i.e. closed loop feedback control. • MRgFUS minimizes damage to surrounding tissue, since ultrasound passing through tissue has not been found to have any cumulative effect on that tissue. Thus, tissue that has been heated to sub-ablative

23 Magnetic Resonance-Guided Focused Ultrasound Surgery

temperatures shows no ill-effects, and treatments may be repeated any time if necessary to complete treatments. • FUS-based heat-induced ablation of brain tumors results in the immediate, irreversible destruction of the targeted tissue. This is in contrast to radiotherapy, where the desired effect can take weeks to months to develop and also detrimental adverse effects may ensue with long delays. • The procedure has the potential to be performed on an out-patient basis with corresponding improvement of QoL for patients. However, there are also limitations and potential complications of MRgFUS: • A limited lesion size per sonication makes multiple sonications per session necessary to treat the whole tumor volume. This may prolong the overall treatment time. • Long cooling times of several minutes between successive sonications are necessary to avoid skull heating. This may also add to long lasting treatment interventions, and solutions to this problem are presently tested. • The tumor vascularity might affect the lesion size and the shape of the sonication lesion. • Immediate tissue swelling may limit the treatment volume. In view of the numerous animal experiments and the first experience of the authors with human interventions, this doesn’t seem to be a major problem. The first clinical results of non-invasive brain tumor coagulation with transcranial MRgFUS are promising (McDannold et al., 2010). As this novel technology matures, it may become especially interesting for the treatment of “benign” brain tumors, such as meningiomas, DENET, pituitary adenomas and even craniopharyngiomas. Moreover, research is also focusing on use of MRgFUS as an adjuvant therapy to traditional radiation or chemotherapy (Guthkelch et al., 1991) and for the targeted delivery of chemotherapeutic agents (Kinoshita et al., 2006b) that can be used by themselves or in tandem with radiation. To conclude: Transcranial MR-guided FUS will “revolutionize” current brain tumor surgery, targeted chemotherapy and radiation oncology.

235

Glossary US HIFU LIFU FUS tcMRgFUS

BBB

Ultrasound High Intensity Focused Ultrasound Low Intensity Focused Ultrasound Focused Ultrasound Surgery transcranial Magnetic Resonance-guided Focused Ultrasound Surgery Blood-Brain-Barrier

References Anzai Y, Lufkin R, DeSalles A, Hamilton DR, Farahani K, Black KL (1995) Preliminary experience with MR-guided thermal ablation of brain tumors. AJNR Am J Neuroradiol 16:39–48. discussion 49–52 Arbour R (1994) Laser and ultrasound technology in aggressive management of central nervous system tumors. J Neurosci Nurs 26:30–35 Aubry JF, Tanter M, Pernot M, Thomas JL, Fink M (2003) Experimental demonstration of noninvasive transskull adaptive focusing based on prior computed tomography scans. J Acoust Soc Am 113:84–93 Clement GT, Hynynen K (2002) A non-invasive method for focusing ultrasound through the human skull. Phys Med Biol 47:1219–1236 Clement GT, White J, Hynynen K (2000) Investigation of a large-area phased array for focused ultrasound surgery through the skull. Phys Med Biol 45:1071–1083 Cline HE, Schenck JF, Hynynen K, Watkins RD, Souza SP, Jolesz FA (1992) MR-guided focused ultrasound surgery. J Comput Assist Tomogr 16:956–965 Fry WJ (1958) Intense ultrasound in investigations of the central nervous system. Adv Biol Med Phys 6:281–348 Fry FJ, Ades HW, Fry WJ (1958) Production of reversible changes in the central nervous system by ultrasound. Science 127:83–84 Fry WJ, Barnard JW, Fry EJ, Krumins RF, Brennan JF (1955) Ultrasonic lesions in the mammalian central nervous system. Science 122:517–518 Fry WJ, Fry FJ (1960) Fundamental neurological research and human neurosurgery using intense ultrasound. IRE Trans Med Electron ME-7:166–181 Fry FJ, Goss SA, Patrick JT (1981) Transkull focal lesions in cat brain produced by ultrasound. J Neurosurg 54:659–663 Fry WJ, Meyers R (1962) Ultrasonic method of modifying brain structures. Confin Neurol 22:315–327 Guthkelch AN, Carter LP, Cassady JR, Hynynen KH, Iacono RP, Johnson PC, Obbens EA, Roemer RB, Seeger JF, Shimm DS et al (1991) Treatment of malignant brain tumors with focused ultrasound hyperthermia and radiation: results of a phase I trial. J Neurooncol 10:271–284 Heimburger RF (2005) An encounter with stereotactic brain surgery. Neurosurgery 56:1367–1373. discussion 1373–1364

236 Hynynen K (2008) Ultrasound for drug and gene delivery to the brain. Adv Drug Deliv Rev 60:1209–1217 Hynynen K, Jolesz FA (1998) Demonstration of potential noninvasive ultrasound brain therapy through an intact skull. Ultrasound Med Biol 24:275–283 Hynynen K, McDannold N, Sheikov NA, Jolesz FA, Vykhodtseva N (2005) Local and reversible blood-brain barrier disruption by noninvasive focused ultrasound at frequencies suitable for trans-skull sonications. Neuroimage 24:12–20 Hynynen K, Vykhodtseva NI, Chung AH, Sorrentino V, Colucci V, Jolesz FA (1997) Thermal effects of focused ultrasound on the brain: determination with MR imaging. Radiology 204:247–253 Ishihara Y, Calderon A, Watanabe H, Okamoto K, Suzuki Y, Kuroda K (1995) A precise and fast temperature mapping using water proton chemical shift. Magn Reson Med 34:814–823 Jolesz FA, Hynynen K (2002) Magnetic resonance image-guided focused ultrasound surgery. Cancer J 8(Suppl 1):S100–S112 Jolesz FA, Hynynen K, McDannold N, Tempany C (2005) MR imaging-controlled focused ultrasound ablation: a noninvasive image-guided surgery. Magn Reson Imaging Clin N Am 13:545–560 Kinoshita M, McDannold N, Jolesz FA, Hynynen K (2006a) Noninvasive localized delivery of Herceptin to the mouse brain by MRI-guided focused ultrasound-induced bloodbrain barrier disruption. Proc Natl Acad Sci USA 103:11719–11723 Kinoshita M, McDannold N, Jolesz FA, Hynynen K (2006b) Targeted delivery of antibodies through the blood-brain barrier by MRI-guided focused ultrasound. Biochem Biophys Res Commun 340:1085–1090 Kuroda K, Chung AH, Hynynen K, Jolesz FA (1998) Calibration of water proton chemical shift with temperature for noninvasive temperature imaging during focused ultrasound surgery. J Magn Reson Imaging 8:175–181 Lele P (1962) A simple method for production of trackless focal lesions with focused ultrasound: physical factors. J Physiol 160:494–512

E. Martin and F.A. Jolesz Liu HL, Hua MY, Chen PY, Chu PC, Pan CH, Yang HW, Huang CY, Wang JJ, Yen TC, Wei KC (2010) Blood-brain barrier disruption with focused ultrasound enhances delivery of chemotherapeutic drugs for glioblastoma treatment. Radiology 255:415–425 Lynn JG, Zwemer RL, Chick AJ (1942) The biological application of focused ultrasonic waves. Science 96: 119–120 McDannold N, Clement GT, Black P, Jolesz F, Hynynen K (2010) Transcranial magnetic resonance imaging- guided focused ultrasound surgery of brain tumors: initial findings in 3 patients. Neurosurgery 66:323–332. discussion 332 Mei J, Cheng Y, Song Y, Yang Y, Wang F, Liu Y, Wang Z (2009) Experimental study on targeted methotrexate delivery to the rabbit brain via magnetic resonance imaging-guided focused ultrasound. J Ultrasound Med 28:871–880 Meyers R, Fry FJ, Fry WJ, Eggleton RC, Schultz DF (1960) Determination of topologica human brain representations and modifications of signs and symptoms if some neurologic disorders by the use of hight level ultrasound. Neurology 10:271–277 Pernot M, Aubry JF, Tanter M, Thomas JL, Fink M (2003) High power transcranial beam steering for ultrasonic brain therapy. Phys Med Biol 48:2577–2589 Ram Z, Cohen ZR, Harnof S, Tal S, Faibel M, Nass D, Maier SE, Hadani M, Mardor Y (2006) Magnetic resonance imaging-guided, high-intensity focused ultrasound for brain tumor therapy. Neurosurgery 59:949–955. discussion 955–946 Tanter M, Thomas JL, Fink M (1998) Focusing and steering through absorbing and aberrating layers: application to ultrasonic propagation through the skull. J Acoust Soc Am 103:2403–2410 Thomas JL, Fink MA (1996) Ultrasonic beam focusing through tissue inhomogeneities with a reversal mirror: application to transskull therapy. IEEE Trans Ultrason Ferroelectr Freq Control 43:1122–1129 Treat LH, McDannold N, Vykhodtseva N, Zhang Y, Tam K, Hynynen K (2007) Targeted delivery of doxorubicin to the rat brain at therapeutic levels using MRI-guided focused ultrasound. Int J Cancer 121:901–907

Chapter 24

Radioguided Surgery of Brain Tumors Laurent Menard

Abstract Surgery is still considered as the primary therapeutic procedure for brain tumors. The precise delineation and excision of brain tumor extent allows one to improve survival outcome and quality of life of surgically treated patients. In that context, many technical adjuncts to surgery, such as neuronavigation, ultrasound or intraoperative MRI have been explored to achieve the most complete removal of tumors. To date, none of these methods has evolved into a standard within the practice of surgery on brain tumors and there is room for new intraoperative tools to refine their resection. The radioguided surgery of brain tumors using gamma or beta-sensitive probes has the potential to fulfill this need by providing real time information to the surgeon regarding the brain tumor extend and the assessment of surgical resection margins. This paper presents a current review on radioguided surgery of brain tumors including technical aspects, preliminary clinical experiences and future perspectives to increase its impact on the surgical management of patients. Keywords Tumorectomy · Radiotracers · Gamma probes · Beta probes · Surgery · Meningiomas

Introduction Surgery plays a preponderant role in the management of patients suffering from solid cancers. The impact of surgery is still more important for brain tumors, L. Menard () Laboratoire Imagerie et Modelisation en Neurobiologie et Cancerologie, IMNC- UMR 8165, Universita Paris Diderot 7, Paris, France e-mail: [email protected]

such as gliomas, that represent 40% of all primary brain tumors, or meningiomas. The aim of surgery in such cases is to totally resect the tumor mass so as to relieve the adjacent neurological structures, to reduce the potential functional lesions and to prevent recurrence (Behin et al., 2003). Removing as much tumor tissue as possible also improves the efficiency of associated treatment, such as radiotherapy or chemotherapy. Several recent clinical studies have shown that the extent of surgical resection is a determinant prognostic factor of outcome and that total brain tumor resection provides patients with a longer and better quality of life when compared to incomplete resection (Carpentier, 2008; Sanai and Berger, 2008). Although complete resection is generally possible in cases of low-grade gliomas and in numerous cases of meningiomas, it is more challenging in cases of invasive tumors, such as malignant gliomas. The infiltration of normal brain tissue by malignant tissues makes it almost very difficult to delineate intraoperatively the tumor extent beyond the visible boundaries of the nidus. Thus, the prognosis of patients suffering of high-grade gliomas is extremely poor, with expected survival being less than 3 years after diagnosis (Behin et al., 2003). The surgical resection of brain tumors is also complicated by the potential neurological deficits that may be incurred with wide resection when the tumor is located in or near eloquent areas of the brain. Many technical advances have been introduced into the operating room to help neurosurgeons to achieve the most complete removal of the tumor by providing information during surgery on the position of the tumor relative to the functional areas. The most commonly used monitoring tool is the stereotactic navigation station that provides the location of surgical excision in real time. This information is registered

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_24, © Springer Science+Business Media B.V. 2011

237

238

with preoperative anatomical or functional MRI image sequences to plan and control the tumor resection relative to eloquent brain structures (Willems et al., 2006). When combining to other surgical tools like robotized microscopes or cortical mapping with electrodes to determine the functional areas, stereotactic image-guided surgery reduces the risk of morbidity and improves removal even for deep-seated tumors. However, this technique is intrinsically limited by the brain shift occurring during the removal of the gross tumor mass that is due to the loss of cerebral fluid or the release of pressure. This phenomenon makes anatomical data acquired prior to surgery useless and significantly degrades the accuracy of the excision. The problem can be overcome by using intraoperative MRI or ultrasound to update preoperative brain images (Schneider et al., 2005; Unsgaard et al., 2006). Nevertheless, setting-up intraoperative MRI systems involves expensive structural modifications to the operating room and requires the use of MRI-compatible surgical instruments, and ultrasound is prone to artifacts and difficulties of image interpretation. Furthermore, the lack of sensitivity of anatomical imaging techniques makes them improper to confidently detect small tumor remnants after the bulk has been excised. In fact, the pathological diagnosis of suspicious tissue can only be obtained from histological analysis of biopsy samples. These analyses are costeffective and also significantly extend the duration of the surgery. Thus, despite all the available monitoring techniques, the precise delineation of tumor boundaries remains very difficult and mainly depends on the tactile and visual feedback of the surgeons. In that context, intraoperative image-guided surgery based on the detection of radioactive or fluorescent tumor-seeking tracers opens up new prospective to help surgeons to discriminate with higher sensitivity and specificity the malignant brain tissues from surrounding normal tissues. These new intraoperative technologies should be able to guide with more accuracy the tumor resection by analyzing in real time the tumor boundaries so as to localize visually indistinct residual lesions. In addition, they should allow one to minimize the surgical invasiveness. Main new diagnostic tools for intraoperative delineation of brain tumors are based on fluorescent tracers. These techniques involve the injection of exogenous fluorescent contrast agents like fluorescein, indocianine green or 5-aminolevulinic acid-induced

L. Menard

porphyrins, or the use of the intrinsic autofluorescence properties of biological tissues in terms of spectral shape, signal intensity and lifetime that can be affected by pathological processes (Pogue et al., 2010). The clinical impact of such techniques is currently under investigation for several types of brain tumors, including glioblastoma, astrocytoma and meningioma (Pogue et al., 2010). Alternatively, the intraoperative detection of radiolabeled tissues using hand-held probes represents a promising adjunctive technique to refine the resection of brain tumor. The concept of radioguided cancer surgery emerged in the early 1950s and is now considered an established surgical procedure for many clinical applications, including the sentinel lymphatic node mapping of melanoma and breast cancer, colorectal cancer or parathyroid disease (Povoski et al., 2009). Surprisingly, if the first description of radioguided surgery involved the detection of brain tumors using 32 P-phosphate (Selverstone et al., 1949), this application has since been investigated in a very limited fashion. The major limitation to its development compared to other intraoperative detection techniques, like MRI, ultrasound or laserinduced fluorescence, is partly due to the lack of specific radiotracers for brain tumors but also to the need of dedicated detection probes. However, the recent emergence of new radiopharmaceuticals such as those labeled with positron emitters, and the development of novel radiation detection techniques are giving rise to a renewed interest for radioguided brain tumor surgery.

Radiopharmaceutical Agents for Radioguided Brain Tumorectomy Radioguided surgery uses specific tumor properties to label tissues with a radiopharmaceutical agent before surgery. A radiopharmaceutical is a biochemical compound that involves a radionuclide coupled to a molecule, which can be able to target a large spectrum of biological functions, including increased metabolism receptor expression or protein synthesis. These agents are labeled with gamma-emitting or beta-emitting radionuclides (including either electron or positron emitters). Most gamma-emitting radionuclides used for radioguided surgery, like 99m Tc, 123 I or 111 In, have predominant gamma photons emission

24 Radioguided Surgery of Brain Tumors

energies in the 140–250 keV range, which represents an ideal compromise between tissue penetration, minimal background counts secondary to scatter and optimal stopping power among the currently available detection technologies. These radionuclides also exhibit relative absence of beta particle emissions, thereby minimizing the absorbed dose of radiation by the patient (Povoski et al., 2009). More recently, radiotracers labeled with positron radionuclides, like 18 F, have been evaluated for radioguided surgery (Gulec et al., 2006; Piert et al., 2007). These elements produce a positron emission from the nucleus decay and two secondary high-energy 511 keV gamma photons arising from the positron-electron annihilation in tissue. All tumor-seeking agents used in nuclear medicine imaging of brain tumors can potentially be applied to radioguided brain tumorectomy. The most important feature for radioguided surgery is the biological uptake ratio of the radiotracer between tumor and surrounding normal tissue, which mainly defines the lesion detectability. The nature (beta or gamma) and the energy of the predominant emission of the radionuclide also deal with a trade-off between optimization of detection parameters such as sensitivity, spatial resolution, background contamination and patient radiation absorbed dose. For example, the short range of charged particles, such as positron, is potentially well suited to reduce the background noise and ensure an accurate localization of residual tumor sites in the surgical resection margins of brain tumors while minimizing the loss of brain material. Moreover, the physical halflife of the radionuclide should be both suited to the clearance kinetics of the radiopharmaceutical to reach optimal tumor-to-normal tissue uptake ratio, and to the duration of surgery. For example, the short physical half-life of 11 C (20 min) compared to the typical duration of a brain surgery is a strong limiting factor to the use of radiopharmaceuticals, such as 11 C-Methionine, for radioguided surgery of brain tumor.

The Gamma-Emitting Radiotracers The most widely used gamma-emitting radiopharmaceutical for radioguided surgery of brain tumors is the myocardial perfusion imaging agent 99m Tchexakis-2-methoxy-isobutyl-isonitrile (99m Tc-MIBI).

239 99m Tc-MIBI

is a lipophilic cationic substance, which passes through the cells membrane through electric diffusion mechanism and is retained by binding to mitochondrias, whose density is higher in tumors due to their greater metabolic activity. The MIBI uptake in brain tumors is poorly understood but it was shown to correlate with leakage from blood vessels and thus, is mainly dependant on the blood-brain barrier damage (Bénard et al., 2003). 201 Tl retention in tumors is also due to the increase of perfusion and this radiopharmaceutical, injected as thallium chloride, has similar applications to that of 99m Tc-MIBI. When applied to the imaging of brain tumors, 201 Tl shows lower detectability than 99m Tc-MIBI and worse discrimination between tumor and normal tissue, mainly due to the lower energy of its predominant gamma emission (Bénard et al., 2003). 111 Indium-labeled somatostatin analogues, like octreotide or pentetreotide, bind specifically to somatostatin receptors that are expressed at a high level in nearly all meningiomas. In that context, 111 In-(DTPA) octreotide and 111 In-(DTPA)D-Phe1 pentetreotide have been recently evaluated as markers in the surgical management of cranial base meningiomas (Gay et al., 2005; Dammers et al., 2009). Other promising agents labeled with gamma photon emitters are currently under validation for brain tumor detection. This includes 123 I-iodo-Lphenylalanin or 123 I-2-iodo-tyrosin, which target the increased tumor metabolism.

The Beta-Emitting Radiotracers The accumulation of 32 P-labeled sodium phosphate in tumor tissue has been known for more than 70 years. Radiophosphorus is a pure beta-emitter and a metabolic non-specific substrate. It is incorporated to a high amount into DNA, RNA, phospholipids and some enzymes of dividing and growing cells (Reinhardt, 1989). As a consequence, tumors with high mitotic activity show a larger uptake than normal tissue, like normal gray and white matters. 32 P-labeled sodium phosphate was used at a variety of anatomic sites, such as brain, to detect superficial tumors. However, the radiation dose of 32 P is not trivial due to its high accumulation in some normal organs, such as liver, muscle and viscera and to the short range of beta particles in tissue. For this reason, 32 P is almost no more

240

used today and has been replaced by more specific radiopharmaceuticals. Most of the available radiopharmaceuticals used for brain tumor detection in nuclear medicine department or currently under development are positron emitters preliminary dedicated to positron emission tomography (PET) (Basu and Alavi, 2009; Chen and Silverman, 2008). 18 F-Fluorodeoxyglucose (FDG) is a positron-emitting nonphysiological analog of glucose. 18 F-FDG uptake in the cell correlates with glycolitic activity, which is increased in tumor compared to normal tissue with both an over-expression of the glucose transporters and a higher activity of the enzyme hexokinase. This accumulation mechanism forms the basis of the wide clinical application of FDG-PET for the detection of tumors and more especially for staging and evaluation of their response to therapies. The potential interest of 18 F-FDG was also successfully recently evaluated for the intraoperative localization of tumor tissue in various contexts, including breast lesions, gastric tumors, melanomas and thyroid and colorectal cancers (Povoski et al., 2009). However, the high background glucose metabolism of normal gray matter strongly degrades the in situ brain tumor-tonormal tissue uptake ratio of 18 F-FDG (between 1.2:1 and 2:1 for low and high-grade gliomas) and is a major limitation for its use during radioguided brain tumors surgery. Other positron radiopharmaceuticals are more specific of brain tumors (Basu and Alavi, 2009; Chen and Silverman, 2008). 18 F-FluoroethylL-Tyrosine (18 F-FET) and 18 F-fluoro-L-phenylalanine (18 F-FDOPA) are amino acids and amino acid analogues tracers, whose uptake in gliomas correlates to increased amino acid transport and protein synthesis and is related to cellular proliferation activity. 18 F-Fluorothymidine avidity reflects the tumor cell proliferation through the activity of the enzyme thymidine kinase-1. The accumulation of 18 F-Fluorocholine in tumors is related to the increased synthesis of phosphatidylcholine, a major phospholipid component of the cell membrane. Compared to 18 F-FDG, these radiotracers exhibit markedly lower uptake in normal brain (tumor-to-normal tissue uptake ratios for high-grade gliomas ranging from 3:1 to 9:1 for the most specific radiotracers, 18 F-Fluorothymidine and 18 F-Fluorocholine), which allows detection of smaller lesions and low-grade tumors (Basu and Alavi, 2009). Among all these radiotracers, only 18 F-FDG, 18 Ffluoro-L-phenylalanine and 18 F-Fluorocholine have

L. Menard

marketing authorizations. The others are currently used during clinical trials, but hold exciting promises for radioguided brain tumor delineation.

Radiation-Sensitive Detection Probes for Intraoperative Tumor Localization After injection of the radiopharmaceuticals, the surgeon uses a radiation-sensitive probe to localize the radiolabeled tissue. From a practical point of view, tumor-labeling process involved in cancer surgery usually requires administration of radiopharmaceuticals several hours prior to excision in order to minimize non-specific activity. The radioguided detection is usually performed following three steps. First, the surgeon carefully surveys the surgical cavity with the tip of the detection probe to localize area of high radiotracer uptake. The tissue is then excised and its tumoral involvement is controlled with an ex vivo measurement of the radiotracer concentration. The accuracy of the resection can also be verified by controlling the residual radioactivity in the operative wound. Two categories of detection probes are currently available depending of the type of radiation detected (Mariani et al., 2008). Gamma probes are dedicated to radiotracers labeled with gamma-emitting radionuclides or to the indirect detection of positron-emitting radionuclides through secondary 511 keV gamma photons resulting from the annihilation of positron in tissue. Alternatively, beta probes detect beta radiations, i.e. either electrons or positrons. Optimal lesion detectability (contrast) during radioguided surgical resection is mainly related to the sensitivity, i.e. the ratio between detected count rate and activity of a source at a fixed distance, and the capability of the probe to reject background radiations. These features set the ability of the probe to localize small amount of radiolabeled tumor tissue against the background activity from non-specific radiotracer accumulation sources. A high level of background contamination is generally encountered when the radiotracer uptake ratio between tumor and normal tissue is low, when the target lesion is placed in close proximity with a physiological accumulation area of the radiotracer or when using high-energy gamma photon producing radionuclides, such as positron emitters. In those situations, a poor suppression of background

24 Radioguided Surgery of Brain Tumors

noise can cause small lesions to be missed, such as residual tumor sites in the surgical resection margins of brain tumors. Thus, the rejection methods of background noise play an extremely important role in radioguided cancer surgery. The weight, size (from small to laparoscopic) and shape (straight or angled) of the probe are also critical features to facilitate the surgical exploration and to shorten the distance between the radiation detector and the surveyed tissue in order to improve the sensitivity and spatial resolution of the detection procedure. The small size of the craniotomy performed during brain tumorectomy tightens the need of small and ergonomic devices to optimize probe manipulation into the surgical cavity.

The Gamma-Sensitive Intraoperative Probes To date, most detection probe systems that have been used for radioguided resection of brain tumors are gamma-counting probes based on similar nuclear detection techniques. The counting system consists of a detection head, built around scintillation crystal/photosensors assemblies or room temperature semiconductor materials, which collect gamma rays (Zanzonico and Heller, 2000). The acceptance cone in which radiation is being detected is defined by the geometry (aperture and length) of the collimator around the gamma sensitive detector. The output electrical pulse produced at the end of the probe for each interaction event is sent to an external acquisition module for pulse height analysis. The control unit processes and displays in real time the integral counting rate in the region pointed to by the tip of the probe. The output information is provided to the surgeon as a visual digital display and/or an auditory signal, which enables him to evaluate the radiotracer concentration in tissue without taking his eyes off the surgical cavity. The rejection methods of gamma-counting probes are usually based on both spatial selection of detected events and electronic suppression of scattered radiations that enter the detector after single or multiple Compton events in the surrounding tissue. First, collimation and lateral shielding of the detection head improves the radial spatial selectivity of the probe (acceptance cone) and allow one to reduce the

241

contribution of background radiations coming from the sides of the detector. A longer or narrower collimator improves the spatial selectivity and thus, reduces the background noise signal from surrounding nontarget tissue compared to the target site in front of the detector. However, such collimation also decreases the sensitivity of the probe and a trade-off has to be found. The second rejection method is based on the pulse height electronic discrimination of primary photons from scattered photons with an energy acceptance window, which is set according to the particular gamma photon energy of the radionuclide. The efficiency of the spectral discrimination is limited by the energy resolution of the detection system, i.e. the ability of the detector to distinguish two events with close energy levels. Detector material is one of the key elements to optimize the sensitivity and energy resolution of the counting probes. Scintillations crystals (NaI(Tl) or CsI(Tl)) offer high efficiency for medium and highenergy gamma emissions (greater than 150 keV), but their corresponding energy resolution (between 20 and 40% at 140 keV) and scatter rejection are rather poor. On the other hand, semi-conductor probes are compact and offer a good energy resolution (between 5 and 15% at 140 keV), which allows very efficient rejection of scattered photons but are primarily dedicated to the detection of low-energy gamma rays (below 150 keV). Most commercially available devices are pencilthin, light-weight probes, whose field of view diameter ranges between 10 and 15 mm. More recently, the growing interest for the radioguided detection of tumors labeled with 18 F-FDG has led to the development of gamma counting probes specifically dedicated to the detection of high-energy 511 keV gamma photons (Mariani et al., 2008; Piert et al., 2007; Gulec et al., 2006). These probes are designed for optimal count rate and rejection of scattered 511 keV photons, including scintillating crystals like GSO with high stopping power and thicker shielding, which increases their overall dimensions and weight. The major limitation of gamma-counting probes is their low spatial selectivity due to the long range of gamma rays in tissue (4 and 24 cm at 140 and 511 keV, respectively). If the radial spatial selectivity can be optimized by using collimation, gamma counting probes remain very sensitive to deep non-specific area of radiotracers, which may interfere with the detection of small tumor lesions. Because the radiolabeled tumors are usually embedded in a non-uniform background activity distribution

242

due to variations in the amount taken up by normal tissues, localization of small lesions is strongly influenced by the position (inclination, distance) of the probe with respect to the positions of the tumor and the non-specific accumulation areas. By mapping the local distribution of background activity, recently developed intraoperative imaging probes are able to spatially resolve target signal from background noise and thus, are very attractive to enhance the signal on noise ratio that is the critical parameter for radioguided surgery. It is expected that imaging probe may help the surgeon more effectively in the task of tumor localization, with respect to non-imaging probes, especially when the radiotracer specificity is poor or for small, deep-seated tumors. Many miniaturized high resolution gamma cameras are currently under evaluation, especially for sentinel lymph node protocols (Mariani et al., 2008; Pitre et al., 2003; Scopinaro et al., 2008; Vermeeren et al., 2009). However, these devices are usually built around a basic detection design, namely a parallel hole collimator and a scintillation crystal (array or continuous plate) coupled to a position-sensitive photomultiplier tube, which are too bulky to be introduced inside the small operative wound encountered during brain tumor surgery.

The Beta-Sensitive Intraoperative Probes A beta-sensitive probe has many advantages over the traditional low and high-energy gamma probes for localizing small amount of radiolabeled tissues. First of all, the low range of electrons and positrons in soft tissue (∼1 mm in water for positron from 18 F and ∼7 mm for electrons from 32 P) allows one to increase the tumor signal-to-noise ratio by avoiding count rate contamination from distal non-specific accumulation of radiotracers. The short depth of penetration of beta particles in matter can also be used to develop imaging or non-imaging detection systems with overall small dimensions and high sensitivity, because no collimation is needed. For example, a 1 mm thick plastic scintillator is adequate to absorb all positrons emitted by a 18 F source (Emax =650 keV, <E>=240 keV) placed in contact. As a consequence, the overall sensitivity of beta detection probes is roughly one order of magnitude higher than that of gamma detection probes. The sensitivity of commercially available gamma counting

L. Menard

probes ranges from 1 to 40 cps/MBq in contact of a point source (122 keV) depending on the geometry of the collimation (Mariani et al., 2008). For a beta probe, the sensitivity is usually higher than 200 cps/MBq for a contact 18 F source. These probes can therefore be used to detect few milligram quantities of tumor when they are placed on the lesion. However, the short range of beta particles in soft tissue also constrains the detector to operate within less than 1 mm of the tumor. This shallow detection makes a beta probe of no value for the localization of deep-seated tumors. Otherwise, this method is particularly suited to radioguided protocols, such as brain tumor surgery, where the main issue is to detect the presence of residual lesion in the resection bed of the gross tumor mass. The first intraoperative beta probe was used in 1949 by Selverstone et al. to localize brain tumors labeled with 32 P. The detector was a miniature Geiger Müller tube. This kind of gas-filled detector is no more used today and the beta probes currently available or under development are based on, schematically, two detection designs. The first one consists of plastic scintillators directly coupled or by means of a fiber light guide to a photosensor, such as a photomultiplier tube, an avalanche photodiode or a CCD, in order to be detected and counted (Daghighian et al., 1994; Tipnis et al., 2004; Yamamoto et al., 2005). The second design is based on direct interaction of charged particles within semiconductor material, such as silicon photodiodes (Raylman, 2000; Lauria et al., 2007). To ensure the detection of positrons in the operative wound, most of imaging or non-imaging beta probes utilizing scintillator detectors or solid-state devices are able to reject the background 511 keV gamma rays. In fact, despite the low intrinsic sensitivity of beta probes to gamma photons, the high flux of 511 keV photons coming from physiological radiotracer accumulation areas can still produce a significant noise. This noise may reduce the in vivo tumor–to-normal tissue ratio relative to the true radiotracer uptake and strongly hampers the intraoperative detection of small tumor lesions (Piert et al., 2007). Various background rejection set-ups can be implemented to eliminate the gamma photon contamination. Some groups proposed phoswich detector configurations for coincidence detection of positron and one of the two associated annihilation gamma rays (Yamamoto et al., 2005; Tornai et al., 1998). To minimize the reduction in sensitivity due to the coincidence rejection scheme, these positron probes 57 Co

24 Radioguided Surgery of Brain Tumors

use bulky high-Z scintillators such as BGO or GSO to detect the 511 keV gamma photons. Other beta probes use real-time subtraction methods between two detectors (Raylman, 2000; Daghighian et al., 1994; Bogalhas et al., 2008). The first one predominantly detects positrons and the second one is made sensitive only to gamma rays by using a beta shielding. By subtracting count rates measured by the second detector from that counted with the first detector, one obtains an estimate of the pure beta signal. When coupled to a background rejection system, positrons probes enable the detection of small lesions even in a noisy environment (Raylman, 2000; Bogalhas et al., 2008). Background compensated probes can also be used as dual probes detecting both beta-emitting and gammaemitting radiotracers simultaneously (e.g. 18 F-FDG and 99m T-Tc) (Raylman, 2001; Tipnis et al., 2004). In that context, Bogalhas et al. (2008, 2009) have developed an intraoperative positron imaging probe specifically dedicated to the real-time localization of residual brain tumors after the bulk has been excised. The probe was designed to be directly coupled to the excision tool leading to simultaneous detection and removal of radiolabeled tumors. This association should help to overcome localization errors due to brain shift or to bad correlations between the intraoperative mapping of radiotracer distribution and the true and precise position of the tumor in the operative wound, especially when no anatomical reference marker is available. The probe, built around clear and plastic scintillating fibers, was also designed to detect positrons emitted from radiolabeled brain tissue and laser induced fluorescence simultaneously. This association should allow one to discriminate more specifically neoplastic from normal tissues due to the complementarity of the biological information coming from radiolabeled radioactive tracers and endogenous fluorophores. A first prototype of the positron probe has been built and evaluated (Bogalhas et al., 2009) (Fig. 24.1). It consists of a detection head composed of detection elements held around the excision tool in a closed packed annular arrangement. Each detection element was composed of a thin piece of scintillating fiber of 1.5 mm diameter thermally fused to a 10 cm long clear fiber. The detection head is coupled to a fiber light guide that exports the scintillating light to an external detection and processing module based on a multi-channel photomultiplier tube. The data are visualized as a two-dimensional image showing the count

243

Fig. 24.1 An intraoperative positron imaging probe specifically dedicated to radioguided resection of brain tumors (Bogalhas et al., 2009)

rate on each detection fiber placed around the surgical tool. The tip of the probe has outer diameters ranging from 8.6 to 12.7 mm, depending on the geometry of the exchangeable detection heads. The gamma ray background is eliminated by a real-time subtraction method. More precisely, the amount of gamma contamination is measured by several detection elements shielded to beta particles by a 400-μm thick sheet of stainless steel. The detector exhibits a beta sensitivity of 217 cps/kBq (3.4 cps/kBq/ml) and 70 cps/Mbq (1.1 cps/kBq/ml) for the large and small head configurations, respectively. The gamma ray rejection efficiency measured by realistic brain phantom modeling of the surgical cavity and the boundaries of the tumor was 99.4%. This good rejection ability allows the true tumor-to-normal tissue uptake ratio to be recovered during phantom measurements (Bogalhas et al., 2009). Phantom studies also demonstrated that tumor tissue blocks as small as 5 mm in diameter and 1 mm thick (20 mg) can confidently be detected, with spatial accuracy better than 2 mm, for tumor-to-normal tissue uptake ratios of fluorinated tracers greater than 3:1. This ratio is achieved with radiopharmaceuticals like 18 F-Fluoroethyl-L-Tyrosine or 18 F-choline. The minimal amount of detectable tumor tissue measured with this positron probe can favorably be compared to the detection threshold of PET systems. For example, the minimal diameter of tumor tissue that can be detected during whole-body 18 F-FDG scans of torso lesions is about 8 mm (equivalent to 300 mg) for radiotracer uptake ratio greater than 10:1 (Piert et al., 2007; Raylman et al., 1999).

244

Radioguided Surgery of Brain Tumors Clinical Experiences with Gamma Probes The clinical concept of intraoperative detection of brain tumor with a gamma counting probe was first introduced by the neurosurgeon William Sweet in 1951 (Sweet, 1951). 50 years later, in 2002, Vileha Filho and Carneiro Filho described the first radio-guided microsurgical resection of a metastatic renal cell carcinoma to the right parietal lobe with a gamma probe. Five hours before the operation, a dose of 30 mCi of 99m Tc-MIBI was intravenously injected to the patient. Accumulation of the radiotracer in the brain tumor was controlled by a preoperative SPECT. At surgery, the gamma probe based on a CdTe crystal (Europrobe, Eurorad) was used to measure radioactive counts from the expected tumor site, as identified on the preoperative MIBI SPECT, and from the adjacent normal tissue. A signal-to-background activity ratio equal or greater than 2:1 was set as a reliable indicator of the pathological nature of the target tissue, as commonly used in radioguided protocols (Gulec et al., 2006). The subcortical area with the highest counting rate corresponding to the highest MIBI uptake was chosen as the best place to perform corticotomy, respecting eloquent areas. Once gross total tumor resection was done, the completeness of removal was controlled by using again the gamma probe to scan the bed of resection in order to ascertain that no radiolabeled residual tissue was left behind. At that time, a 3 mm diameter piece of tumor with tumor-to-normal tissue ratio of 5:1 was detected and removed. This residual tumor was not visually identified during the first survey. A post-operative CT confirmed complete tumor resection and a follow-up of 3 months showed no evidence of tumor recurrence. This single-case study indicates the feasibility of radio-guided brain tumorectomy that may facilitate localization of superficial tumors before the excision to restrict the size of the craniotomy, but above all, to assure completeness of its removal in real time. Following a similar protocol, Bhanot et al. (2007) reported on the use of 99m Tc-MIBI in a dose of 10 mCi, for gamma-probe assisted resection in 13 patients with high-grade supratentorial gliomas. The radiotracer was administered 2 h before surgery. In accordance to Vileha Filho and Carneiro Filho (2002),

L. Menard

the intraoperative probe (Euro 4 probe, Euro Medical Instruments) was used to guide the location of craniotomy by picking up radioactivity through the scalp, to visually identify indistinct tumor from normal tissue near or in eloquent areas during the excision and to provide assessment of completeness of tumor resection. In 9 patients, the post-operative CT/MRI showed no enhancing area and in 4 patients, only small residual enhancing areas. These cases were associated to doubtful positive readings from the probe inside the cavity (signal-to-background activity ratio less than 2:1). The authors concluded that the radio-guided resection of brain tumor with a gamma counting probe is an inexpensive and easy to use technique that provides realtime intraoperative information about the tissue that is being excised and thus, a reliable proof to confirm the presence or absence of residual tumors. They also emphasized several limits of the method. First of all, the retention of MIBI in gliomas is dependant on the damage of the blood-brain barrier and thus, has no value for the localization of low-grade lesions where new methodologies to monitor the resection margins might be particularly helpful. The intraoperative detection of radiolabed tissue was also hampered by the radioactivity in the blood that contaminates the probe readings in the vicinity of large veins, venous sinuses or pools of blood in the surgical cavity. Finally, the large size of the gamma counting probe restricted the probe manipulation in the resection bed and prevented to perform a detailed search for tumor remnants (oversampled exploration). In 2008, Serrano et al. reported on the use of a different radiotracer, 201 Tl, for radioguided resection of high-grade astrocytomas in 6 patients (Serrano et al. 2008). In all cases, residual activity uptake was found with a scintillating gamma probe (Scintiprobe MR100, Pol.Hi.tech) in the surgical bed after visual resection. These areas were confirmed as pathological by examination of biopsy samples. In 3 cases, the suspicious tissue was removed and the final activity was similar to the activity found in the normal brain. In the other 3 patients, the removal of all the residual activity was impossible due to eloquent structures involved. Despite these encouraging results that showed the feasibility of radioguided surgery of highgrade astrocytomas with 201 Tl, authors also underline the limitation of this radiotracer compared to 99m TcMIBI, in terms of detectability, signal-to-background ratio and availability. Because the high background activity in the normal brain can make difficult to

24 Radioguided Surgery of Brain Tumors

discriminate residual tumors from surrounding normal tissue with the gamma probe, they concluded that this technique requires a good expertise of the medical staff in both radioguided surgery and neuroimaging. To improve the accuracy of the intraoperative detection of radiolabeled brain tissue, Kojima et al. (2004) reported on the use of a mobile gamma camera during radioguided surgery of brain tumors, as a complementary tool to a conventional gamma counting probe. Thirteen patients with primary or recurrent high-grade astrocytomas or metastatic brain lesions received an intravenous dose of 99m Tc-MIBI just before surgery. During the operation, the tumor was localized with a semiconductor counting probe (Navigator, United State Surgical Corporation) and the mobile gamma camera (2020tc imager, Digirad). The camera was placed in close proximity with the brain surface to acquire scintigraphic images for 5 min. The probe was also coupled to an optical stereotactic navigation system in order to provide real-time indication on its location relative to the target tumor and other eloquent brain structures based on preoperative MRI scan. The probe and the gamma camera correctly identified tumors in all patients, according both to the preoperative MRI and 99m Tc-MIBI SPECT. The tumors ranged in size from 12 to 70 mm. The tumor-to-background activity ratio measured with the mobile gamma camera showed a significant correlation with values found in the pre-operative 99m Tc-MIBI-SPECT images (mean value of 6.9:1). After tumor resection was considered to be finished, residual tumors were seeked with the mobile gamma camera. In 9 out of 13 patients, scintigraphy images showed no accumulation of radioactivity in the tumor sites. The absence of residual tumors was confirmed by histologic results of tissue samples taken around the resection bed. Residual tumor tissue was found in 4 patients with high-grade primary gliomas (astrocytoma and glioblastoma). This study indicates the potential impact of intraoperative scintigraphy imaging for monitoring in real time the extent of brain tumor resection. Compared to open intraoperative MRI or ultrasound, scintigraphy images are not affected by artifacts due to postoperative bleeding that can hamper the detection of residual tumor during surgery. The technique can also be easily implemented in any operating room and only slightly prolongs the operation duration (less than 10 min for the whole imaging procedure). Authors also underline the difficulty to discriminate target signal from background

245

activity with the counting probe when the tumor uptake is low and located close to physiological area of MIBI accumulation in the scalp, choroid plexus or base of the skull. As expected, the better spatial selectivity of the mobile gamma camera allows one to overcome this limitation. If most clinical experiences with gamma-probe assisted resection was focused on high grade gliomas, Gay et al. (2005) also investigated the use of a handheld gamma probe to guide the resection of bone invasive or “en plaque” meningiomas labeled with 111 In(DTPA)-octreotide. Ten patients with sphenoid wing meningiomas without any cavernous sinus involvement and meningiomas with invasion of the skull convexity or into the mastoid process were enrolled. A dose of 3 mCi of 111 In-(DTPA)-octreotide was injected the day before surgery and its good accumulation in the tumor was controlled with a pre-operative scintigraphy. During the surgery, the scintillating-based gamma probe (Tec probe 2000, Stratec electronic) was used to compare the counting rates of the invaded bone and of adjacent normal skull in order to help define the tumor margins. In all patients, intraoperative detection was able to identify radiolabeled tumor invasion on bone, dura matter and periorbital involvement of sphenoid wing meningiomas. The high affinity of octreotide for somatostatin receptors of meningiomas provided high in vivo tumor-to-background ratio ranging from 2:1 to 12:1 with a mean value of 4.4:1. In 6 patients, the high level of radiotracer uptake measured with the probe demonstrated good correlation with the margins of the invaded bone defined with a computer-aided navigation system using CT images. This led to assist the delineation of bone and dural resection. The post-operative control of the counting rates of the bony margins with the gamma probe was used again to assess the accuracy of the resection. The absence of residual radioactivity confirmed the complete removal of meningiomas of the skull convexity, but was more difficult to implement for sphenoid wing tumors due to the size of the probe and the background contamination from the pituitary gland that also binds the 111 In-(DTPA)-octreotide. Post-operative somatostatin receptor scintigraphies performed during the follow-up period showed no sign of recurrence in all patients. This study demonstrates the feasibility of intraoperative detection of somatostatin receptors in meningiomas. The high specificity of radiolabeled somatostatin analogues offers the benefit of high

246

tumor-to-background ratio, which is a critical issue for radioguided resection of tumors, and low radiation exposure to patient. The method was more efficient for bone invasive meningiomas of the skull convexity and may increase the probability of complete resection of these tumors that are difficult to control surgically. In 2009, Dammers et al. (2009) conducted successfully a similar study on a patient with a left sphenoid wing meningioma labeled with 111 In-(DTPA)-D-Phe1 pentetreotide.

Clinical Experiences with Beta Probes Although Vilela Filho and Carneiro Filho (2002) reported the first radioguided surgery of brain tumor using a gamma probe in 2002, many attempts have been performed in the 1950s with 32 P-labeled sodium phosphate. Selverstone et al. (1949) first used this radioactive phosphorus to define the localization and demarcation of brain tumors during surgery. The radiotracer was intravenously injected 1 day before the surgery, as a buffered phosphate solution of 0.5–2 mCi, to 33 patients with various types of cerebral tumors (meningiomas, astrocytomas, glioblastomas. . .). During the operation, a miniature Geiger-Müller counter in the end of a needle probe with external diameter of 2 mm was sunk into the brain and counting rates were obtained at several areas and depths beneath the cerebral cortex. The position and extend of tumors were estimated from the regions showing an increased counting rate when the probe is inserted into or very close to the lesion. Tumors were localized in 29 of 33 cases. In 23 patients, the tumor was subcortical. In 12 patients, the operative mapping with the probing counter was used to estimate the gross delineation of the tumor boundary margins in order to facilitate its complete resection. Three falsenegative detections were reported. Two of them were reported obtained when the Geiger-Müller counter could not be placed close enough to the tumor to detect it. Finally, one additional case gave a false-positive response, which was attributed to a diffuse gliomatosis. Morley and Jefferson (1952) demonstrated similar results in 32 cases of brain tumor following the protocol laid down by Selverstone et al. (1949). Both studies underline the potential value of the intraoperative detection of brain tumors labeled with 32 P for two

L. Menard

main purposes: to map the gross extend of tumors in order to plan the excision into the brain (accessibility of the tumor and potential extensions in eloquent areas) and to pin-point the tumor location in order to obtain a valid biopsy specimen from the active portion of a tumor along a cannula introduced through a solitary burr hole. However, these studies also highlight the intrinsic limitations of 32 P as a tumor-seeking radiotracer. First, the unspecific accumulation of 32 P in tissue with increased metabolic activity, such as cerebral inflammatory lesion, can cause false-positive detections. Second, its short beta range in white or grey matters can also strongly prevent the detection of small deeply situated tumors. This requires that their locations are known by other tools before the beta probe is used. In the 1950s, ventriculography and angiography were the main techniques for imaging the central nervous system. Nowadays anatomical or functional MRI and stereotactic navigation systems enable the accurate localization of tumors relative to eloquent brain structures. Therefore, the main current application of radioguided brain tumorectomy is no longer to locate the lesion prior to the excision, but to detect the presence of residual tumor tissue in the resection bed of the gross tumor mass identified on preoperative scans. In that context, the shallow detection achieved with betaemitting radionuclides is not a limitation, but rather an advantage. However, the assessment of surgical resection margins requires a degree of sensitivity far beyond that of the Geiger-Müller counter employed in these studies could provide. More recently, Reinhardt (1989) reported the use of a more suitable beta probe based on a solid-state technology and demonstrated its benefit for intraoperative detection of radiolabeled brain tumors after 32 P infusion (1.5 mCi). A reliable discrimination between tumor and normal tissue was achieved, especially in the border area of meningiomas, where high accumulation of 32 P was found within the matrix zone. The intraoperative detection of tumor remnants was also possible in brain metastases and high-grade gliomas. In accordance to Morley and Jefferson (1952), the contrast between tumors and normal brain tissue was only insufficient for low-grade gliomas. The recent emergence of promising specific tumorseeking agents labeled with positron emitters for glioma delineation is giving rise to a renewed interest for radioguided brain surgery using beta probes. In that context, Leston et al. (2009) investigated the feasibility

24 Radioguided Surgery of Brain Tumors

of using their positron-sensitive intraoperative imaging probe to guide detection and excision of brain tumors in a primate model (Bogalhas et al., 2009). The main purposes were to test the ability of the probe to localize radiolabeled tissue in vivo and to evaluate the direct coupling between the probe and the excision tool to detect and remove the target tissue at once. Because no model of gliomas is currently available in primate, they chose to mimic a brain tumor surgery by using the accumulation of 18 F-FDOPA in the striatum and the globus pallidus of a rhesus monkey’s brain and by implementing the beta probe to accurately remove the left part of this cerebral structure. A dose of 3 mCi of 18 F-FDOPA was injected after the craniotomy and 40 min before the excision. The beta probe was coupled to an ultrasonic aspirator (Dissectron, Integra) and used to map the radioactivity distribution in the brain (Fig. 24.2). A first counting rate was obtained from the cerebellum to determine the threshold signal. Once the left lobe of the striatum has been exposed, all cerebral tissue showing a signal-to-background activity ratio equal or greater than 1.5:1 was removed. After gross total removal was thought to be complete, the bed of resection was probed again in order to seek any residual tissue. The accuracy of the resection was controlled by post-mortem histological examination of the brain. All targeted cerebral structures labeled with 18 F-FDOPA were correctly removed, including putamen and globus pallidus, without significant damage to the surrounding tissues. The caudate nucleus was left behind due to its bad accessibility. Small residual

Fig. 24.2 Evaluation on a primate model of the association between the intraoperative positron probe developed by Bogalhas et al. (2009) and an ultrasonic aspirator (Leston et al., 2009)

247

regions in the putamen were attributed to the difficulty to move the probe inside the small surgical cavity. This preliminary study indicates that the ability to map the distribution of residual radiolabeled lesions in the close vicinity of the excision tool should facilitate and improve brain tumor removal when associated with a specific tumor-seeking agent labeled with positron emitters.

Radiation Exposure During Radioguided Brain Surgery An important issue when studying the clinical potential of radioguided surgery for brain tumor relative to nonradioisotopic monitoring techniques is the additional radiation exposure to the surgical team. The annual occupational exposure limits for radiation workers set by the International Commission on Radiological Protection (ICRP) is a total effective dose equivalent to 20 mSv per year for the whole body (100 mSv averaged over a 5-year period) and 500 mSv per year for the hands. Due to the small number of clinical experiences, data on radiation exposure from radiotracers used during radioguided brain surgery surgical is currently very limited. Kojima et al. (2004) evaluated the whole body dose to surgical personnel during radioguided brain tumorectomy using 99m TcMIBI. The radiotracer was administered on the day of surgery. The radiation dose to the hands was not reported. The mean exposure time was 6.1 h. The mean whole body dose per operation was 27.9 μSv for the surgeon, 25.8 μSv for the nurse and 14.9 μSv for the anesthetist. Bhanot et al. (2007) reported a similar value (22.9 μSv), 2 h after injection of 99m Tc-MIBI, in a dose of 10 mCi. These values are quite below the allowable limit. However, the radiation exposition to the operating team is strongly dependant on the features of the radiopharmaceutical agent (gamma or beta emitters, biological distribution in the human body, physical half-life of the radionuclide) and higher absorbed doses of radiation are expected with beta-emitting radiopharmaceuticals. Bogalhas et al. (2009) investigated the radiation exposure to the surgeon’s hand during brain phantom experiments with their positron-sensitive intraoperative probe. Lithium fluoride thermoluminescent dosimeters were placed around the probe. The radiation dose to the surgeon’s

248

hand was assumed to be mainly due to the radiotracer accumulation in the brain. The measurements were scaled to extrapolate the radiation dose received by the hands of the surgeon due to the accumulation of 18 F-FDG, 18 F-FET or 18 F-Choline in tumor and normal brain tissue, 1 h after injection of a typical injected dose of 300 MBq. The dose rate to the surgeon’s hand was significant (300, 60 and 12 μSv per hour for 18 F-FDG, 18 F-FET and 18 F-Choline, respectively), but well within the exposure limits. Moreover, several recent studies have also demonstrated that the mean body dose received by the surgical staff involved in radio-guided cancer surgery using 18 F-FDG and high-energy gamma probes was small compared to the ICRP recommendations (range 20–80 μSv per hour) (Povoski et al., 2009). As a conclusion, although additional comprehensive evaluations of occupational radiation exposure to intraoperative and perioperative personnel are required, the absorbed radiation dose appears to present no serious limitation to the development and the repetition of radioguided brain tumor surgical procedures.

Discussion Radical removal of brain tumors, such as gliomas and meningiomas, remains the best mean of providing increased survival, lesser morbidity and better quality of life. Intraoperative surgical techniques have been developed to improve the completeness of local tumor removal while sparing normal tissue. At this time, neurosurgical navigation systems and intraoperative MRI or ultrasound are the standard of care to guide the neurosurgeon in the operative room. However, there is a need for novel intraoperative tools that could provide a more specific discrimination between tumor and normal brain tissue during operations in order to facilitate tumor detection and its complete removal, especially for invasive lesions. Some feasibility studies demonstrated that radioguided resection using gamma or beta-sensitive probes is a safe, easy to use and reliable technique to enhance the ability of the surgeon to delineate more accurately the brain tumor extend and to confirm in real time the presence or absence of residual tumor in the surgical resection margins. Despite these preliminary encouraging results that must be completed on large cohorts of patients, several

L. Menard

issues have to be addressed in order to increase the impact of radioguided surgery on the surgical management of brain tumors. First of all, future efforts should aim to increase the specificity of brain tumor markers. New positron-emitting radiopharmaceuticals, preliminary developed for PET, such as 18 F-Choline or 18 FFluorothymidine, or radiolabeled analogues of tumor specific receptors have the potential to fulfill this goal and to extend the application field of radioguided brain cancer surgery beyond meningiomas and high-grade gliomas. From a technological point of view, the detection of brain tumor remnants in the resection margin of the gross tumor mass requires dedicated miniaturized probes with high sensitivity and background noise rejection capabilities. In that context, miniaturized beta imaging probes are very attractive to map the distribution of small and weak radioactive sources and to spatially resolve target signal from background noise (Bogalhas et al., 2009; Lauria et al., 2007; Tipnis et al., 2004; Tornai et al., 1998). Background compensated imaging probes used as dual probes detecting both positrons arising from the decay of positron-emitting radiotracers and the associated 511 keV annihilation γ-rays simultaneously seem also very promising to further improve the detection accuracy (Raylman, 2001; Tipnis et al., 2004; Tornai et al., 1998). The gamma signal could be used to rapidly identify residual uptake within the boundaries of the tumor, even for deepseated lesions, and the beta signal to evaluate surgical margins of these tumor remnants. Another important issue to improve the accuracy of brain tissue removal is to combine the various surgical and monitoring devices used during brain tumorectomy. This may include the direct coupling between the intraoperative probe and the excision tool (Bogalhas et al., 2009) and the convergence of the information provided by radiation-sensitive probes, neuronavigation systems and intraoperative MRI, when available (Yamamoto et al., 2004). For example, the complete exploration of the tumor boundaries could be facilitated by coupling the probe to a frameless stereotaxic neuronavigation system to track its intraoperative location and thereby reconstruct the three-dimensional activity distribution in the cavity (Wendler et al., 2006). This could provide a reliable verification of the completeness of resection in real time. Bogalhas et al. (2008) also proposed an intraoperative bimodal probe, which will combine simultaneous localization of radiolabeled tissues with measurement of laser-induced fluorescence

24 Radioguided Surgery of Brain Tumors

so as to discriminate neoplastic from normal tissues more specifically (Siebert et al., 2008).

References Basu S, Alavi A (2009) Molecular imaging (PET) of brain tumors. Neuroimaging Clin N Am 19(4):625–646 Behin A, Hoang-Xuan K, Carpentier AF, Delattre JY (2003) Primary brain tumours in adults. Lancet 361(9354):323–331 Bénard F, Romsa J, Hustinx R (2003) Imaging gliomas with positron emission tomography and single-photon emission computed tomography. Semin Nucl Med 33(2):148–162 Bhanot Y, Rao S, Parmeshwaran RV (2007) Radio-guided neurosurgery (RGNS): early experience with its use in brain tumour surgery. Br J Neurosurg 21(4):382–388 Bogalhas F, Charon Y, Duval MA, Lefebvre F, Palfi S, Pinot L, Siebert R, Menard L (2009) Development of a positron probe for localization and excision of brain tumours during surgery. Phys Med Biol 54(14):4439–4453 Bogalhas F, Ménard L, Bonzom S, Palfi S, Siebert R, Duval MA, Lefebvre F, Pinot L, Pitre S, Charon Y (2008) Physical performance of an intraoperative beta probe dedicated to glioma radioguided surgery. IEEE Trans Nucl Sci 55(3):833– 841 Carpentier AC (2008) Surgical resection of gliomas in 2008. Cancer Radiother 12(6–7):676–686 Chen W, Silverman DH (2008) Advances in evaluation of primary brain tumors. Semin Nucl Med 38(4):240–250 Daghighian F, Mazziotta JC, Hoffman EJ, Shenderov P, Eshaghian B, Siegel S, Phelps ME (1994) Intraoperative beta probe: a device for detecting tissue labeled with positron or electron emitting isotopes during surgery. Med Phys 21(1):153–157 Dammers R, Hsu SP, Krisht AF (2009) Radioguided improved resection of a cranial base meningioma. Neurosurgery 64 (3 Suppl):84–85 Gay E, Vuillez JP, Palombi O, Brard PY, Bessou P, Passagia JG (2005) Intraoperative and postoperative gamma detection of somatostatin receptors in bone-invasive en plaque meningiomas. Neurosurgery 57:107–113 Gulec SA, Daghighian F, Essner R (2006) PET-Probe: evaluation of technical performance and clinical utility of a hand-held high energy gamma probe in oncologic surgery. Ann Surg Oncol 1–7 Kojima T, Kumita S, Yamaguchi F, Mizumura S, Kitamura T, Kumazaki T, Teramoto A (2004) Radio-guided brain tumorectomy using a gamma detecting probe and a mobile solid-state gamma camera. Surg Neurol 61(3):229–238 Lauria A, Mettivier G, Montesi MC, Aloj L, Lastoria S, Aurilio M, Russo P (2007) Experimental study for an intraoperative probe for 18F imaging with a silicon pixel detector. Nucl Instr Meth 576:198–203 Leston JM, Bogalhas F, Plafi S, Siebert R, Duval MA, Lefebvre F, Pinot L, Charon Y, Ménard L 2009. Intra-operative positron probe to address surgical challenges in brain surgery. International symposium on advanced intraoperative imaging of radioisotopes and presymposium workshop TOF PET, Baia delle Zagare, Italy, September 2009

249 Mariani G, Giuliano AE, Strauss HW (2008) Radioguided surgery: a comprehensive team approach. Springer Science, New York, NY. ISBN: 978–0–387–33684–8 Morley TP, Jefferson G (1952) Use of radioactive phosphorus in mapping brain tumours at operation. Br Med J 2:575–578 Piert M, Burian M, Meisetschlager G, Stein HJ, Ziegler S, Nahrig J, Picchio M, Buck A, Siewert JR, Schwaiger M (2007) Positron detection for the intraoperative localisation of cancer deposits. Eur J Nucl Med Mol Imaging 34(10):1534–1544 Pitre S, Menard L, Ricard M, Solal M, Garbay JR, Charon Y (2003) A hand-held imaging probe for radio-guided surgery: physical performance and preliminary clinical experience. Eur J Nucl Med Mol Imaging 30(3):339–343 Pogue BW, Gibbs-Strauss S, Valdés PA, Samkoe K, Roberts DW, Paulsen KD (2010) Review of Neurosurgical Fluorescence Imaging Methodologies. IEEE J Sel Top Quantum Electron 16(3):493–505 Povoski SP, Neff RL, Mojzisik CM, O’Malley DM, Hinkle GH, Hall NC, Murrey DA Jr, Knopp MV, Martin EW Jr (2009) A comprehensive overview of radioguided surgery using gamma detection probe technology. World J Surg Oncol 7:11 Raylman RR (2000) A solid-state intraoperative beta probe system. IEEE Trans Nucl Sci 47(4):1696–1703 Raylman RR (2001) Performance of a dual solid state intraoperative probe system with 18 F, 99m Tc, and 111 In. J Nucl Med 42:352–360 Raylman RR, Kison PV, Wahl RL (1999) Capabilities of twoand three-dimensional FDG-PET for detecting small lesions and lymph nodes in the upper torso: a dynamic phantom study. Eur J Nucl Med 26(1):39–45 Reinhardt H (1989) Surgery of brain neoplasm using 32-P tumour marker. Acta Neurochir 97:89–94 Sanai N, Berger MS (2008) Glioma extent of resection and its impact on patient outcome. Neurosurgery 62(4):753–764. discussion 264–266 Schneider JP, Trantakis C, Rubach M, Schulz T, Dietrich J, Winkler D, Renner C, Schober R, Geiger K, Brosteanu O, Zimmer C, Kahn T (2005) Intraoperative MRI to guide the resection of primary supratentorial glioblastoma multiforme – a quantitative radiological analysis. Neuroradiology 47(7):489–500 Scopinaro F, Tofani A, di Santo G, Di Pietro B, Lombardi A, Lo Russo M, Soluri A, Massari R, Trotta C, Amanti C (2008) High-resolution, hand-held camera for sentinel-node detection. Cancer Biother Radiopharm 23(1):43–52 Selverstone B, Sweet WH, Robinson CV (1949) The clinical use of radioactive phosphorus in the surgery of brain tumors. Ann Surg 130(4):643–651 Serrano J, Rayo JI, Infante JR, Dominguez L, Garcia-Bernardo L, Duran C, Fernandez Portales I, Cabezudo JM (2008) Radioguided surgery in brain tumors with thallium-201. Clin Nucl Med 33(12):838–840 Siebert R, Vu Thi MH, Jean F, Charon Y, Collado-Hilly M, Duval MA, Mandat T, Menard L, Palfi S, Tordjmann T (2008) Development of a new autofluorescence probe for the analysis of normal and tumour brain tissues. Proc SPIE 2008 6991:699122 Sweet WH (1951) The uses of nuclear disintegration in the diagnosis and treatment of brain tumors. N Engl J Med 245(875):887

250 Tipnis SV, Nagarkar VV, Shestakova I, Gaysinskiy V, Entine G, Tornai MP, Stack BC (2004) Feasibility of a beta-gamma digital imaging probe for radioguided surgery. IEEE Trans Nucl Sci 51(1):110–116 Tornai MP, Levin CS, MacDonald LR, Holdsworth CH, Hoffman EJ (1998) A miniature phoswich detector for gamma ray localization and beta imaging. IEEE Trans Nucl Sci 45(3):1166–1173 Unsgaard G, Rygh OM, Selbekk T, Muller TB, Kolstad F, Lindseth F, Hernes TA (2006) Intra-operative 3D ultrasound in neurosurgery. Acta Neurochir (Wien) 148(3): 235–53 Vermeeren L, Valdes Olmos RA, Meinhardt W, Bex A, van der Poel HG, Vogel WV, Sivro F, Hoefnagel CA, Horenblas S (2009) Intraoperative radioguidance with a portable gamma camera: a novel technique for laparoscopic sentinel node localisation in urological malignancies. Eur J Nucl Med Mol Imaging 36(7): 1029–1036

L. Menard Vilela Filho O, Carneiro Filho O (2002) Gamma probe-assisted brain tumor microsurgical resection: a new technique. Arq Neuropsiquiatr 60(4):1042–1047 Wendler T, Traub J, Ziegler SI, Navab N (2006) Navigated three dimensional beta probe for optimal cancer resection. Med Image Comput Comput Assist Interv 9(Pt 1):561–569 Willems PW, van der Sprenkel JW, Tulleken CA, Viergever MA, Taphoorn MJ (2006) Neuronavigation and surgery of intracerebral tumours. J Neurol 253(9):1123–1136 Yamamoto S, Kuroda K, Senda M (2004) Development of an MR-compatible gamma probe for combined MR/RI guided surgery. Phys Med Biol 49(15):3379–3388 Yamamoto S, Matsumoto K, Sakamoto S, Tarutani K, Minato K, Senda M (2005) An intra-operative positron probe with background rejection capability for FDG-guided surgery. Ann Nucl Med 19(1):23–28 Zanzonico P, Heller S (2000) The intraoperative gamma probe: basic principles and choices available. Semin Nucl Med 30:30–48

Chapter 25

Implications of Mutant Epidermal Growth Factor Variant III in Brain Tumor Development and Novel Targeted Therapies Murielle Mimeault and Surinder K. Batra

Abstract The mutant epidermal growth factor receptor variant III (EGFRvIII) oncogene has attracted much attention due to its critical functions in primary brain cancer development, treatment resistance and disease relapse. The EGFRvIII mutant is frequently overexpressed during brain cancer initiation and progression in conjunction with the wild-type EGFR amplification while no expression level of its truncated receptor is detected in normal brain tissue specimens and any distant tissues. The constitutively activated EGFRvIII mutant can substantially contribute in cooperation with other receptor tyrosine kinases, including wild-type EGFR, to the sustained growth, migration and local metastases of brain cancer cells and resistance to current therapies. Therefore, the truncated EGFRvIII mutant represents a therapeutic target of great clinical interest for developing new effective treatments against primary brain tumors. In this chapter, we summarize the recent advancements on the characterization of key functions supplied by the EGFRvIII mutant and wild-type EGFR in brain tumor development and the most important findings about the novel therapeutic strategies developed for their effective molecular targeting. The emphasis is on the specific roles played by

M. Mimeault () Department of Biochemistry and Molecular Biology, College of Medicine, Eppley Cancer Institute, 7052 DRC, University of Nebraska Medical Center, 985870 Nebraska Medical Center, Omaha, NE 68198-5870, USA e-mail: [email protected] S.K. Batra () Department of Biochemistry and Molecular Biology, College of Medicine, Eppley Cancer Institute, 7052 DRC, University of Nebraska Medical Center, 985870 Nebraska Medical Center, Omaha, NE 68198-5870, USA e-mail: [email protected]

the EGFRvIII mutant and its cooperative interactions with wild-type EGFR in brain tumor cells, treatment resistance and disease relapse. The implications of targeting the EGFRvIII mutant and/or wild-type EGFR to develop novel combination therapies for improving the current treatments against aggressive and recurrent medulloblastomas and glioblastoma multiforme are also discussed. Keywords EGFR · Mutant · Tyrosine kinase · Inhibitor · Antibody · Vaccination

Introduction Malignant primary brain tumors, including pediatric medulloblastomas and adult glioblastoma multiforme (GBMs), are among the most frequent, rapidly growing and lethal tumors of the central nervous system (CNS) (Furnari et al., 2007; Sampson et al., 2008; Stupp et al., 2005). The highly aggressive and locally invasive medulloblastomas and GBMs are generally refractory to current clinical therapies by tumor resection, radiotherapy and/or adjuvant chemotherapeutic treatments (Furnari et al., 2007; Halatsch et al., 2006; Stupp et al., 2005). Especially, the GBM patients treated with radiotherapy plus adjuvant temozolomide have a poor median survival time of about 14.6 months after diagnosis (Stupp et al., 2005). The inefficacy of current therapies for treating patients with medulloblastomas and gliomas has been associated with the accumulation of genetic and epigenetic alterations that can contribute to the acquisition of more malignant phenotypes and survival advantages by brain cancer cells.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_25, © Springer Science+Business Media B.V. 2011

251

252

M. Mimeault and S.K. Batra

Fig. 25.1 Schematic structure of human wild-type EGFR protein and truncated EGFRvIII mutant. The scheme shows the extracellular domains divided in four subdomains I–IV, transmembrane region (TM), juxtamembrane segment (JM), tyrosine kinase domain and C-terminal tail. Moreover, the in-frame

deletion of the 6–273 amino acid segment of wild-type EGFR in the extracellular domain of truncated EGFRvIII mutant, which results in the generation of a novel glycine residue at the fusion point that is specific to the mutant receptor, is also indicated

Among the frequent genetic alterations that can contribute to primary brain tumor development, the wild-type EGFR (erbB1) transmembrane receptor tyrosine kinase (RTK) is amplified, overexpressed and/or mutated in approximately 40–60% of GBM patients (Fig. 25.1) (Aldape et al., 2004; Furnari et al., 2007; Halatsch et al., 2006). Moreover, the immunohistochemical and real-time PCR analyses have revealed that an enhanced expression of constitutively actived EGFRvIII mutant frequently occurs in approximately 30–60% of GBM patients while no expression of this mutant is observed in the normal adult brain and any other tissues (Fig. 25.1) (Aldape et al., 2004; Pelloski et al., 2007; Yoshimoto et al., 2008). Of particular interest, it has also been shown that the wild-type EGFR and/or EGFRvIII mutant can cooperate with other genetic alterations to the malignant transformation of neural stem cells (NSCs) or their early progenies during medulloblastomas and GBM development, treatment resistance and disease relapse (Batra et al., 1995; Lammering et al., 2004). Consequently, it appears that the development of multitargeted strategies directed against wild-type EGFR and/or EGFRvIII mutant with the current clinical treatments by radiation and chemotherapies, might represent promising therapies for treating the patients diagnosed with aggressive and recurrent primary brain tumors (Fig. 25.2). In this regard, we reviewed the specific functions of the wild-type EGFR and EGFRvIII mutant in cancer cells during the development of medulloblastomas and GBMs. Of clinical interest, recent studies supporting the therapeutic benefit of targeting the wild-type EGFR and EGFRvIII signaling pathways, alone or in combination, for improving the current treatments against highly aggressive and recurrent medulloblastomas and GBMs are also discussed.

Functions of the Wild-Type EGFR and EGFRvIII Mutant in Primary Brain Tumor Development Several investigations have revealed that the sustained activation of the wild-type EGFR and EGFRvIII mutant can cooperate for the malignant transformation of GBM cells, angiogenic process and tumor development. More specifically, the enhanced expression of wild-type EGFR may lead to its activation by its secreted ligands, including EGF and transforming growth factor-α (TGF-α), through an autocrine loop or in a juxtacrine or a paracrine manner and the stimulation of diverse intracellular signaling elements (Halatsch et al., 2006; Lo et al., 2008). These intracellular signaling components include phosphatidylinositol 3 kinase (PI3K)/Akt, Ras/mitogen-activated protein kinases (MAPKs), Janus-activated kinase 2 (JAK2)/signal transducers and activators of transcription 3 (STAT3) and phospholipase-Cγ (PLC-γ) (Fig. 25.2) (Halatsch et al., 2006; Lo et al., 2008). The activation of these intracellular pathways might result in the induction of mitotic effects and promote the growth, migration and local invasion of GBM cells. In the case of the EGFRvIII mutant, which lacks a N-terminal 267 amino acid-portion of the full-length EGFR’s extracellular ligand-binding domain, this truncated receptor acts as a constitutively autophosphorylated and activated receptor that mediates its oncogenic effects in the absence of a ligand (Fig. 25.1) (Batra et al., 1995). The constitutively activated EGFRvIII mutant can stimulate in a persistent manner diverse intracellular cascades and up-regulate the expression of numerous target genes such as metalloproteinases (MMPs), tissue factor (TF) and TATA-binding protein

25 Implications of Mutant Epidermal Growth Factor Variant III

253

Fig. 25.2 Schematic representation showing the signaling elements induced through the wild-type EGFR and EGFRvIII mutant involved in the malignant behavior of brain cancer cells and novel targeted strategies. This scheme shows the stimulatory effect induced through the activation of EGF/EGFR and constitutively activated EGFRvIII mutant on the Ras/mitogen activated protein kinases (MAPKs) and phosphatidyl 3 inositol (PI3K)/Akt which might lead to an mTOR, inhibition of cyclin-dependent kinase inhibitors p27kip1 and p21cip1 and enhanced expression of target gene products. In addition, the potential inhibitory effect induced by diverse pharmacological agents, such as a monoclonal antibody (mAb) or immunotoxins directed against EGF or TGF-α ligand, wild-type EGFR or EGFRvIII mutant, selective inhibitors of EGFR/EGFRvIII tyrosine kinase activity such as gefitinib and erlotinib and

vaccination based on the EGFRvIII-specific peptide are also indicated. Moreover, the therapeutic strategy consisting of using a mAb directed against HGF, a specific PI3K (LY294002), mTOR (rapamycin), dual inhibitor of PI3K/mTOR (PI-103), and SREBP-1 (25-HC) in combination therapies is also indicated. c-MET, hepatocyte growth factor receptor; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; EGFRvIII, mutant epidermal growth factor receptor variant III; 25-HC, 25-hydroxycholesterol; HGF, hepatocyte growth factor; mAb, monoclonal antibody; MMPs, matrix metalloproteinases; mTOR, mammalian target of rapamycin; SREBP-1, serol regulatory element-binding protein 1; TBP, TATA-binding protein; TF, tissue factor; TGF-α, transforming growth factor-α and VEGF, vascular endothelial growth factor

(TBP), that can contribute to malignant behavior of GBM cells. The downstream signaling elements that can be induced through the EGFRvIII mutant in a cell type-dependent manner include PI3K/Akt, Ras/MAPK, JAK2/STAT3, protein kinase C-α (PKCα)/myristoylated alanine-rich protein kinase C substrate (MARCKS) and abnormal spindle-like microcephaly associated (ASPM) protein (Fig. 25.2) (Lal et al., 2002; Fan et al., 2006; Wang et al., 2006; Fromm et al., 2008; Micallef et al., 2009; Mukherjee et al., 2009a; Bikeye et al., 2010). Hence, the EGFRvIII mutant can play critical roles in the proliferation and migration of GBM cells, and enhanced tumorigenecity and angiogenic process, and cooperate with the wild-type EGFR and other oncogenic products for GBM tumor development. Moreover, the expression of the EGFRvIII mutant in the GBM patients has also been associated with a decreased expression level

of collapsing response mediator protein 1 (CRMP1), which in turn can promote the invasion of EGFRvIIIexpressing GBM cells (Mukherjee et al., 2009a). In addition, the co-expression and heterodimerization of the wild-type EGFR and EGFRvIII mutant in GBM cells might also contribute to their acquisition of a more malignant behavior (Johns et al., 2007; Martens et al., 2008; Patel et al., 2007).

Implications of Wild-Type EGFR and EGFRvIII Mutant in Treatment Resistane of Brain Tumor Cells The enhanced expression of the wild-type EGFR and EGFRvIII mutant has also been shown to contribute to radiation and chemotherapy resistance of GBM cells

254

and has been associated with a poor survival of GBM patients (Golding et al., 2009; Halatsch et al., 2006; Kim et al., 2008; Lammering et al., 2004). The cytoprotective effects induced through the activation of these RTKs might be mediated, at least in part, by an up-regulation of the Ras/MAPK and PI3K/Akt survival pathways and anti-apoptotic factor expression such as Bcl-2 and Bcl-xL that leads to an inhibition of the apoptosis of GBM cells (Fig. 25.2). Moreover, it has also been reported that the expression of the wild-type EGFR and/or EGFRvIII mutant might promote the DNA double-strand break repair, and thereby contribute to the radioresistance of malignant gliomas (Golding et al., 2009). In addition, a growing body of experimental evidence has indicated that the wild-type EGFR and EGFRvIII mutant are expressed in brain tumor-initiating cells which can provide critical roles in tumor development, treatment resistance and disease relapse (Griffero et al., 2009; Mimeault and Batra, 2010). Hence, on the basis of these observations, it appears that the molecular targeting of the wild-type EGFR and EGFRvIII mutant, and other oncogenic products that are frequently deregulated during human primary brain cancer development might constitute more promising therapeutic strategies as monotherapies to eradicate total brain tumor cell mass and prevent disease relapse. In this matter, we are reporting the recent investigations supporting the therapeutic interest of targeting the wild-type EGFR and EGFRvIII mutant, and other oncogenic signaling elements, alone or in combination with current therapies, for overcoming the treatment resistance of GBM cells and preventing disease relapse.

Novel Therapeutic Strategies Against Primary Brain Cancers by Molecular Targeting of the Wild-Type EGFR and EGFRvIII Mutant Signaling Pathways Different therapeutic strategies have been designed to block the wild-type EGFR and/or EGFRvIII mutant cascades. Among them, there are the use of a specific tyrosine kinase activity inhibitor (TKI) such as gefitinib and erlotinib, monoclonal antibody (mAb) directed against these receptors or EGF and TGF-α ligands and their silencing by small interference (siRNA)

M. Mimeault and S.K. Batra

or short hairpin RNA (shRNA) (Fig. 25.2) (Halatsch et al., 2006; Mimeault and Batra, 2010). Moreover, the inhibition of the downstream signaling elements, such as PI3K/Akt/mTOR, JAK2/STAT3, MARCKS and ASPM, also might constitute an alternative therapeutic approach for improving the antitumoral efficacy induced through the wild-type EGFR and EGFRvIII blockade (Lo et al., 2008; Mukherjee et al., 2009b; Bikeye et al., 2010). The data from numerous studies have revealed that the blockade of the wild-type EGFR and/or EGFRvIII pathways with these agent types, alone or in synergy with radiation and chemotherapeutic drugs, results in an inhibition of growth and invasion and/or the apoptotic death of the primary brain cancer cells in vitro and in vivo (Fig. 25.2) (Cemeus et al., 2008; Halatsch et al., 2006; Lal et al., 2002; Lo et al., 2008; Loew et al., 2009; Mukherjee et al., 2009b; Pillay et al., 2009).

Targeting of EGFR and/or EGFRvIII Using Specific Tyrosine Kinase Activity Inhibitors Numerous studies have been made to improve the current treatments of brain tumors by using specific inhibitors of EGFR and/or EGFRvIII tyrosine kinase activity, alone or in combination with other cytotoxic agents. In particular, the results from a recent investigation have indicated that the treatment of EGFRvIIIexpressing astrocytes with gefitinib or PI3K inhibitor, LY294002 improved the cytotoxic effects induced by radiation through the inhibition of DNA doublestrand break repair enzyme, DNA-dependent protein kinase catalytic subunit (DNA-PKsc) (Mukherjee et al., 2009b). Moreover, gefitinib plus a specific inhibitory agent of JAK2/STAT3 cascade, JSI-124 also inhibited the growth and induced the apoptosis in the GBM and medulloblastoma cell lines expressing the wild-type EGFR or EGFRvIII mutant in vivo (Lo et al., 2008). Importantly, it has also been reported that a combination of gefitinib plus a 3-hydroxy-3methyl-glutaryl-coenzyme A (HMG-CoA) reductase inhibitor, lovastatin synergistically induced the cytotoxic effects on U87MG human malignant glioma cells engineered to overexpress EGFRvIII (U87MGEGFRvIII) and PTEN irrespective of EGFRvIII and PTEN status (Cemeus et al., 2008). In the same way, a decreased expression of the cell-surface EGFRvIII mutant and certain target gene products involved in the

25 Implications of Mutant Epidermal Growth Factor Variant III

255

invasion of GBM cells has also been observed after a long-term exposure of these cancer cells to erlotinib in vitro (Lal et al., 2002). It has however been observed that erlotinib as single agent did not significantly inhibit the tumor growth of U87MG overexpressing EGFRvIII mutant in vivo (Lal et al., 2002). In contrast, a combination of erlotinib with an anti-hepatocyte growth factor (HGF) antibody, L2G7 synergistically induced a marked reduction of the tumor growth and an increase of the survival of mice (Fig. 25.2) (Lal et al., 2002).

after short-term culture into the brains of nude mice (Martens et al., 2008). In contrast, all the unresponsive GBM tumors lacked amplified and/or mutated EGFR expression (Martens et al., 2008). Based on these observations, it has been proposed that the mAb528 or cetuximab treatment of GBM cells might prevent the heterodimer formation between the wild-type EGFR and EGFRvIII molecules by sterically impairing the adoption of an extended conformation by these receptors which is necessary for their dimerization (Fig. 25.2) (Martens et al., 2008; Patel et al., 2007).

Targeting of EGFR and/or EGFRvIII Using Specific Antibodies

Clinical Trials Involving the Molecular Targeting of Wild-Type EGFR and EGFRvIII Mutant

It has been reported that he treatment of U87MGEGFRvIII cell-derived xenografts with an anti-EGFR antibody, panitumumab partially inhibited the tumor growth whereas a combination of panitumumab and an anti-HGF antibody, AMG 102 substantially suppressed the tumor growth through the induction of the apoptotic death of glioma cells (Pillay et al., 2009). In addition, other mAbs specifically recognizing the wild-type EGFR and/or EGFRvIII mutant, including mAb528, chimeric IMC-C225 (cetuximab or Erbitux), mAb 806 and 3C10 have also been shown to induce antitumoral effects on GBM cells in vitro and/or in vivo (Halatsch et al., 2006; Martens et al., 2008; Patel et al., 2007). More specifically, it has been reported that cetuximab can induce its cytotoxic effects on GBM cells expressing the wild-type EGFR and/or EGFRvIII mutant by inhibiting the ligand binding to the wildtype EGFR and homodimerization as well as by causing the internalization of the cetuximab-EGFRvIII complexes (Patel et al., 2007). Moreover, it has also been observed that mAb528 inhibited the in vivo tumor growth of U87MG cells expressing endogenous EGFR, which have been engineered for also overexpressing EGFRvIII mutant while be this mAb does not suppress the tumor growth of xenografted fibroblasts overexpressing EGFRvIII alone (Johns et al., 2007). In the same way, the intracranial delivery of cetuximab also induced the tumor growth inhibitory, anti-invasive and apoptotic effects on three of seven cases of diffusely invasive xenografts established from different human GBM spheroids exhibiting EGFR amplification and EGFRvIII expression, which have been implanted

Several clinical trials undertaken to investigate the therapeutic interest of using the RTKs such as gefitinib and erlotinib or mAbs directed against wild-type EGFR and EGFRvIII murant for the treatment of GBM patients have also given promising results, and more particularly in combination with radiation or chemotherapies (Rich et al., 2004; Halatsch et al., 2006; Prados et al., 2009). In general, it has been observed that these agents used as monotherapy may be effective to control the disease, while their combination with the current treatments by radiation and/or chemotherapies, might lead to an improvement of the survival in certain GBM patients (Rich et al., 2004; Halatsch et al., 2006; Prados et al., 2009). It has also been noticed that the response to the RTK inhibitors was positively associated with the expression of EGFRvIII and PTEN in certain cases of GBM patients (Halatsch et al., 2006; Mellinghoff et al., 2005). For instance, the data from a recent phase II study have revealed that the inclusion of erlotinib plus temolomide during and after radiation therapy of newly diagnosed GBM patients improved median survival to 19.3 months as compared to 14.1 months observed for a treatment consisting of radiation plus temolomide without erlotinib (Prados et al., 2009). In this matter, it has also been observed that the gefitinib or erlotinib induced the anti-proliferative and cytotoxic effects on EGFR+ /CD133+ tumor-initiating cells from five patients with glioblastomas (GBM TICs) while two cases of GBM TICs with high Akt activation were insensitive to both drugs or only sensitive

256

to high concentrations of erlotinib (Griffero et al., 2009). Altogether, these observations suggest then that the combined use of specific inhibitors of the wild-type EGFR, EGFRvIII mutant and other tumorigenic cascades, including PI3K/Akt, could be more effective in certain GBM patients than the single agents. Consistently, the data from several studies have revealed that the targeting of PI3K/Akt/mammalian target of rapamycin (mTOR) or SREBP-1 signaling components improved the cytotoxic effects induced by different antitumoral agents including the wild-type EGFR and mutant EGFRvIII inhibitors on GBM cells in vitro and in vivo (Fig. 25.2) (Fan et al., 2006; Guo et al., 2009; Wang et al., 2006). In addition to targeting PI3K/Akt pathway, diverse immunotherapies have also been designed to specifically target the EGFRvIII mutant.

EGFRvIII Mutant-Based Immunotherapies and Vaccination Distinct anti-EGFRvIII antibodies and EGFRvIIIspecific peptides have been developed, which may serve as potential carriers for radioconjugate- and immunotoxin-based therapies and therapeutic tools for vaccination of patients diagnosed with GBM tumors overexpressing the EGFRvIII mutant (Choi et al., 2009; Sampson et al., 2008). In fact, the EGFRvIII mutant, which is not expressed in normal tissues, constitutes an ideal and attractive tumor-specific target for vaccination due to its unique extracellular epitope that is formed subsequent to an in-frame genomic deletion creating a unique antigenic site that might be targeted using antibody-based antitumor vaccines (Fig. 25.2) (Beers et al., 2000; Choi et al., 2009; Sampson et al., 2008). Numerous preclinical and clinical studies aimed at the immunologic targeting of the tumor cell-expressing EGFRvIII mutant have been carried out with cultured cells engineered to overexpress EGFRvIII mutant, in animal models or selected cancer patients. The investigations consisting of the vaccination with dendritic cells and EGFRvIII-specific peptide or immunotoxins, have indicated that these therapeutic strategies might be effective to eradicate the brain cancer cells expressing EGFRvIII mutant and enhance the median survival time of GBM patients (Beers et al., 2000; Choi et al.,

M. Mimeault and S.K. Batra

2009; Sampson et al., 2008). Different antibodies specifically reacting at the fusion junction of deletionmutant EGFRvIII have also been developed and in certain cases reached the clinical trials (Fig. 25.1) (Beers et al., 2000; Choi et al., 2009; Sampson et al., 2008). For instance, it has been observed that an immunotoxin MR1(Fv)-PE38 obtained from an antibody phage display library consisting of a single-chain antibody variable domain (scFv2 ), which specifically binds the EGFRvIII mutant, fusioned with a truncated form of Pseudomonas exotoxin induced the cytotoxic effects on GBM cells (Beers et al., 2000). More specifically, the treatment of U87MG cells and tumor spheres expressing EGFRvIII mutant and CD133 with BsAb and human CD16-expressing natural killer (NK) cells used as the effectors resulted in a greater lysis of the U87MG cells and tumor spheres expressing both antigens as compared to normal neurospheres expressing only CD133 (Wong et al., 2009). Hence, this strategy is of great therapeutic interest to specifically target CD133+ /EGFRvIII+ brain tumorinitiating cells, and thereby prevent the secondary effects on the normal CD133+ neural stem/progenitor cells. In addition, several EGFRvIII mutant-based immunotherapies and vaccine therapies have also given promising results in preclinical and clinical trials. For instance, several studies have been performed using an EGFRvIII mutant-based cancer vaccine designated as PEPvIII-KLH. The PEPvIII-KLH construct consists of an EGFRvIII-specific peptide PEPvIII (LEEKKGNYVVTDHC) which corresponds to a 13 amino acid sequence that spans the EGFRvIII fusion junction combined with an additional cystein residue to facilitate the chemical conjugation to a keyhole limpet hemocyanin (KLH). It has been reported that PEPvIII-KLH can generate EGFRvIII-specific antibodies in the patients with high grade gliomas (Choi et al., 2009; Sampson et al., 2008). It has also been observed that the brain tumor resection and sequential treatment of GBM patients with radiation plus TMZ followed by intradermal injections of the cancer vaccine PEPvIII-KLH was accompanied by a specific T- and B- cell-induced immune response and elimination of tumor cells expressing the EGFRvIII mutant (Choi et al., 2009). The overall survival of newly-diagnosed GBM patients overexpresing the EGFRvIII mutant after this treatment type was significantly enhanced to about 26.0 months as compared to 15 months observed for the patients treated with

25 Implications of Mutant Epidermal Growth Factor Variant III

257

radiation plus temozolomide alone (Sampson et al., 2008). Importantly, it has also been noticed that the temozolomide-induced lymphopenia associated with this treatment improved the efficacy of the peptide vaccination by inhibiting regulatory T cells or their delayed recovery (Choi et al., 2009). In the same way, the results from a phase II multi-center trial with 22 cases of newly-diagnosed GBM patients overexpresing the EGFRvIII consisting of external beam radiation therapy followed by vaccinations with PEPvIII-KLH and granulocyte macrophage-colony stimulating factor (GM-CSF) have also indicated that the humoral and immune responses were manifested in patients. The median time-toprogression of these patients was of 14.2 months which is superior to only 7.1 months observed for the patients who have not been treated with temozollonide (Sampson et al., 2008). Although these are interesting results, it has been noticed that a down-regulation of the EGFRvIII mutant and stimulation of diverse tumorigenic cascades might occur in certain GBM patients during disease progression, and contribute to the disease recurrence (Sampson et al., 2008). These observations suggest that it is necessity to also include other potential tumor-specific antigens to immunotherapy such as interleukin-13 receptor α2 in addition to the EGFRvIII mutant-specific peptide in the future investigations in order to prevent treatment resistance and disease relapse.

brain-initiating cells versus their differentiated progenies during primary brain cancer development. Particularly, it will be important to further determine the specific functions supplied by homodimers and heterodimers formed by the wild-type EGFR and EGFRvIII mutant and interactive cross-talks between these RTKs and other growth factor pathways during primary brain cancer progression to locally advanced and invasive stages and their implications in the resistance to current therapies. Hence, these additional studies should lead to the development of more effective multi-targeted approaches to inhibit the wild-type EGFR and mutant EGFRvIII tumorigenic pathways and other key signaling elements that cooperate for the acquisition of a more malignant behavior and survival advantages by brain tumor-initiating cells and their progenies during disease progression. These multi-targeted strategies could be used to improve the current radiation and chemotherapeutic treatments against highly aggressive and lethal primary brain cancers, including pediatric medulloblastomas and adult GBMs, and thereby prevent disease relapse and the death of cancer patients.

Conclusions and Perspectives

References

Together these recent investigations suggest that the enhanced expression of the wild-type EGFR and truncated EGFRvIII mutant and up-regulation of PI3K/Akt represent frequent transforming events occurring during primary brain cancer development, and more particularly during GBM progression to aggressive and invasive disease stage. Especially, oncogenic pathways induced through EGFRvIII mutant can provide critical roles for the cell proliferation, survival and migration of cancer cells, and thereby cooperate with the wildtype EGFR for tumor formation and local invasion, treatment resistance and disease relapse. In spite of these important advancements, future studies are required to more precisely establish the specific intracellular pathways induced through the wild-type EGFR and EGFRvIII mutant in the

Aldape KD, Ballman K, Furth A, Buckner JC, Giannini C, Burger PC, Scheithauer BW, Jenkins RB, James CD (2004) Immunohistochemical detection of EGFRvIII in high malignancy grade astrocytomas and evaluation of prognostic significance. J Neuropathol Exp Neurol 63:700–707 Batra SK, Castelino-Prabhu S, Wikstrand CJ, Zhu X, Humphrey PA, Friedman HS, Bigner DD (1995) Epidermal growth factor ligand-independent, unregulated, cell-transforming potential of a naturally occurring human mutant EGFRvIII gene. Cell Growth Differ 6:1251–1259 Beers R, Chowdhury P, Bigner D, Pastan I (2000) Immunotoxins with increased activity against epidermal growth factor receptor VIII-expressing cells produced by antibody phage display. Clin Cancer Res 6:2835–2843 Bikeye SN, Colin C, Marie Y, Vampouille R, Ravassard P, Rousseau A, Boisselier B, Idbaih A, Calvo CF, Leuraud P, Lassalle M, El Hallani S, Delattre JY, Sanson M (2010) ASPM-associated stem cell proliferation is involved in malignant progression of gliomas and constitutes an attractive therapeutic target. Cancer Cell Int 10:1

Acknowledgements The authors on this work are supported in part by U.S. Department of Defense grants PC04502 and PC074289 and the National Institutes of Health [Grants CA78590, CA111294, CA133774, CA131944 and CA138791].

258 Cemeus C, Zhao TT, Barrett GM, Lorimer IA, Dimitroulakos J (2008) Lovastatin enhances gefitinib activity in glioblastoma cells irrespective of EGFRvIII and PTEN status. J Neurooncol 90:9–17 Choi BD, Archer GE, Mitchell DA, Heimberger AB, McLendon RE, Bigner DD, Sampson JH (2009) EGFRvIII-targeted vaccination therapy of malignant glioma. Brain Pathol 19:713– 723 Fan QW, Knight ZA, Goldenberg DD, Yu W, Mostov KE, Stokoe D, Shokat KM, Weiss WA (2006) A dual PI3 kinase/mTOR inhibitor reveals emergent efficacy in glioma. Cancer Cell 9:341–349 Fromm JA, Johnson SA, Johnson DL (2008) Epidermal growth factor receptor 1 (EGFR1) and its variant EGFRvIII regulate TATA-binding protein expression through distinct pathways. Mol Cell Biol 28:6483–6495 Furnari FB, Fenton T, Bachoo RM, Mukasa A, Stommel JM, Stegh A, Hahn WC, Ligon KL, Louis DN, Brennan C, Chin L, DePinho RA, Cavenee WK (2007) Malignant astrocytic glioma: genetics, biology, and paths to treatment. Genes Dev 21:2683–2710 Golding SE, Morgan RN, Adams BR, Hawkins AJ, Povirk LF, Valerie K (2009) Pro-survival AKT and ERK signaling from EGFR and mutant EGFRvIII enhances DNA double-strand break repair in human glioma cells. Cancer Biol Ther 8: 730–738 Griffero F, Daga A, Marubbi D, Capra MC, Melotti A, Pattarozzi A, Gatti M, Bajetto A, Porcile C, Barbieri F, Favoni RE, Lo CM, Zona G, Spaziante R, Florio T, Corte G (2009) Different response of human glioma tumor-initiating cells to EGFR kinase inhibitors. J Biol Chem 284:7138–7148 Guo D, Prins RM, Dang J, Kuga D, Iwanami A, Soto H, Lin KY, Huang TT, Akhavan D, Hock MB, Zhu S, Kofman AA, Bensinger SJ, Yong WH, Vinters HV, Horvath S, Watson AD, Kuhn JG, Robins HI, Mehta MP, Wen PY, DeAngelis LM, Prados MD, Mellinghoff IK, Cloughesy TF, Mischel PS (2009) EGFR signaling through an Akt-SREBP-1dependent, rapamycin-resistant pathway sensitizes glioblastomas to antilipogenic therapy. Sci Signal 2:ra82 Halatsch ME, Schmidt U, Behnke-Mursch J, Unterberg A, Wirtz CR (2006) Epidermal growth factor receptor inhibition for the treatment of glioblastoma multiforme and other malignant brain tumours. Cancer Treat Rev 32:74–89 Johns TG, Perera RM, Vernes SC, Vitali AA, Cao DX, Cavenee WK, Scott AM, Furnari FB (2007) The efficacy of epidermal growth factor receptor-specific antibodies against glioma xenografts is influenced by receptor levels, activation status, and heterodimerization. Clin Cancer Res 13:1911–1925 Kim K, Brush JM, Watson PA, Cacalano NA, Iwamoto KS, McBride WH (2008) Epidermal growth factor receptor VIII expression in U87 glioblastoma cells alters their proteasome composition, function, and response to irradiation. Mol Cancer Res 6:426–434 Lal A, Glazer CA, Martinson HM, Friedman HS, Archer GE, Sampson JH, Riggins GJ (2002) Mutant epidermal growth factor receptor up-regulates molecular effectors of tumor invasion. Cancer Res 62:3335–3339 Lammering G, Valerie K, Lin PS, Hewit TH, Schmidt-Ullrich RK (2004) Radiation-induced activation of a common variant of EGFR confers enhanced radioresistance. Radiother Oncol 72:267–273

M. Mimeault and S.K. Batra Lo HW, Cao X, Zhu H, li-Osman F (2008) Constitutively activated STAT3 frequently coexpresses with epidermal growth factor receptor in high-grade gliomas and targeting STAT3 sensitizes them to Iressa and alkylators. Clin Cancer Res 14:6042–6054 Loew S, Schmidt U, Unterberg A, Halatsch ME (2009) The epidermal growth factor receptor as a therapeutic target in glioblastoma multiforme and other malignant neoplasms. Anticancer Agents Med Chem 9:703–715 Martens T, Laabs Y, Gunther HS, Kemming D, Zhu Z, Witte L, Hagel C, Westphal M, Lamszus K (2008) Inhibition of glioblastoma growth in a highly invasive nude mouse model can be achieved by targeting epidermal growth factor receptor but not vascular endothelial growth factor receptor-2. Clin Cancer Res 14:5447–5458 Mellinghoff IK, Wang MY, Vivanco I, Haas-Kogan DA, Zhu S, Dia EQ, Lu KV, Yoshimoto K, Huang JH, Chute DJ, Riggs BL, Horvath S, Liau LM, Cavenee WK, Rao PN, Beroukhim R, Peck TC, Lee JC, Sellers WR, Stokoe D, Prados M, Cloughesy TF, Sawyers CL, Mischel PS (2005) Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors. N Engl J Med 353:2012–2024 Micallef J, Taccone M, Mukherjee J, Croul S, Busby J, Moran MF, Guha A (2009) Epidermal growth factor receptor variant III-induced glioma invasion is mediated through myristoylated alanine-rich protein kinase C substrate overexpression. Cancer Res 69:7548–7556 Mimeault M, Batra SK (2010) New promising drug targets in cancer- and metastasis-initiating cells. Drug Discov Today 15:354–364 Mukherjee J, DeSouza LV, Micallef J, Karim Z, Croul S, Siu KW, Guha A (2009a) Loss of collapsin response mediator protein1, as detected by ITRAQ analysis, promotes invasion of human gliomas expressing mutant EGFRvIII. Cancer Res 69:8545–8554 Mukherjee B, McEllin B, Camacho CV, Tomimatsu N, Sirasanagandala S, Nannepaga S, Hatanpaa KJ, Mickey B, Madden C, Maher E, Boothman DA, Furnari F, Cavenee WK, Bachoo RM, Burma S (2009b) EGFRvIII and DNA doublestrand break repair: a molecular mechanism for radioresistance in glioblastoma. Cancer Res 69:4252–4259 Patel D, Lahiji A, Patel S, Franklin M, Jimenez X, Hicklin DJ, Kang X (2007) Monoclonal antibody cetuximab binds to and down-regulates constitutively activated epidermal growth factor receptor vIII on the cell surface. Anticancer Res 27:3355–3366 Pelloski CE, Ballman KV, Furthm AF, Zhang L, Lin E, Sulman EP, Bhat K, McDonald JM, Yung WK, Colman H, Woo SY, Heimberger AB, Suki D, Prados MD, Chang SM, Barker FG, Buckner JC, James CD, Aldape K (2007) Epidermal growth factor receptor variant III status defines clinically distinct subtypes of glioblastoma. J Clin Oncol 25:2288–2294 Pillay V, Allaf L, Wilding AL, Donoghue JF, Court NW, Greenall SA, Scott AM, Johns TG (2009) The plasticity of oncogene addiction: implications for targeted therapies directed to receptor tyrosine kinases. Neoplasia 11: 448–458 Prados MD, Chang SM, Butowski N, DeBoer R, Parvataneni R, Carliner H, Kabuubi P, yers-Ringler J, Rabbitt J, Page M, Fedoroff A, Sneed PK, Berger MS, McDermott MW, Parsa AT, Vandenberg S, James CD, Lamborn KR, Stokoe D,

25 Implications of Mutant Epidermal Growth Factor Variant III Haas-Kogan DA (2009) Phase II study of erlotinib plus temozolomide during and after radiation therapy in patients with newly diagnosed glioblastoma multiforme or gliosarcoma. J Clin Oncol 27:579–584 Rich JN, Reardon DA, Peery T, Dowell JM, Quinn JA, Penne KL, Wikstrand CJ, Van Duyn LB, Dancey JE, McLendon RE, Kao JC, Stenzel TT, Ahmed Rasheed BK, Tourt-Uhlig SE, Herndon JE, Vredenburgh JJ, Sampson JH, Friedman AH, Bigner DD, Friedman HS (2004) Phase II trial of gefitinib in recurrent glioblastoma. J Clin Oncol 22:133–142 Sampson JH, Archer GE, Mitchell DA, Heimberger AB, Bigner DD (2008) Tumor-specific immunotherapy targeting the EGFRvIII mutation in patients with malignant glioma. Semin Immunol 20:267–275 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff

259 RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996 Wang MY, Lu KV, Zhu S, Dia EQ, Vivanco I, Shackleford GM, Cavenee WK, Mellinghoff IK, Cloughesy TF, Sawyers CL, Mischel PS (2006) Mammalian target of rapamycin inhibition promotes response to epidermal growth factor receptor kinase inhibitors in PTEN-deficient and PTEN-intact glioblastoma cells. Cancer Res 66:7864–7869 Wong A, Mitra S, Gupta P (2009) Targeting brain tumor stem cells using a bispecific antibody directed against CD133+ and EGFRvIII+. J Clin Oncol 27:15s Yoshimoto K, Dang J, Zhu S, Nathanson D, Huang T, Dumont R, Seligson DB, Yong WH, Xiong Z, Rao N, Winther H, Chakravarti A, Bigner DD, Mellinghoff IK, Horvath S, Cavenee WK, Cloughesy TF, Mischel PS (2008) Development of a real-time RT-PCR assay for detecting EGFRvIII in glioblastoma samples. Clin Cancer Res 14: 488–493

Chapter 26

Endoscopic Port Surgery for Intraparenchymal Brain Tumors Pawel G. Ochalski and Johnathan A. Engh

Abstract Intraparenchymal endoscopic port surgery (EPS) is a minimally-invasive technique for the removal of brain tumors and other mass lesions. The operation is made feasible through the combination of multiple technologies including imageguidance, parallel endoscopy, and cylindrical brain retraction. Tumors with an overlying cuff of normal brain parenchyma, soft consistency, and low or moderate vascularity are the ideal candidates. Techniques for tumor removal using the endoscopic port and prevention of peri-operative morbidity are discussed. In selected cases, EPS can provide superior visualization and decreased white matter manipulation during intraparenchymal tumor resection when compared with conventional approaches using the operating microscope. Keywords EPS · Brain tumor · Lesions · CNS · Neuroendoscopy · Hemostasis

Introduction Brain Tumor Demographics and Challenges Although brain tumors are not particularly common adult tumors, accounting for 2% of all cancer deaths

J.A. Engh () Department of Neurological Surgery, University of Pittsburgh Medical Center, UPMC Presbyterian, Pittsburg, PA 15213, USA e-mail: [email protected]

in the United States, they are among the most deadly. Most primary central nervous system (CNS) tumors in adult patients are malignant. The incidence of such tumors is approximately 22,000 cases/year in the United States, with approximately 13,000 deaths/year attributable to brain tumors (American Cancer Society, 2009). Brain metastases, all of which are malignant, have an incidence greater than 170,000 cases/year (Suh, 2010). Therapy for CNS primary and metastatic tumors is multi-modal, often beginning with surgical resection in selected cases followed by adjuvant radiation and/or chemotherapy. Neurological morbidity from brain tumors is often attributed to regional pressure (i.e. mass effect) from the lesion disrupting the normal function of surrounding neurons. Surgical removal of the tumor can facilitate relaxation of the surrounding brain, facilitating the subsequent return of improved neurologic function. However, surgical manipulation also requires dissection through surrounding brain, which often includes eloquent neurological tissue. As a result of such manipulation, patients may develop hemiparesis, dysphasia, cognitive impairment, seizures, or stroke. In an effort to facilitate the resection of intraparenchymal brain tumors with minimal injury to the surrounding brain, EPS was developed. This technique combines a minimally invasive cylindrical retraction system, light and magnification as delivered by parallel endoscopy, and frameless image-guidance. In selected cases, this method of brain tumor resection facilitates neurological recovery by minimizing the amount of normal tissue disruption required for tumor removal (Fig. 26.1).

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_26, © Springer Science+Business Media B.V. 2011

261

262

P.G. Ochalski and J.A. Engh

Fig. 26.1 Comparison of the visualization corridor provided by the operating microscope (a) versus rod-lens endoscope (b). Post-contrast axial T1-weighted magnetic resonance imaging showing a large rim-enhancing lesion arising from the putamen

and extending into the deep cerebral white matter. Note the difference in the visualization corridor provided by the source light (white line) when using the operating microscope (a) versus the rod-lens endoscope (b)

History

for visualization of skull base pathology. Resection is performed in an air medium, and all scarring is invisible to the patient. More importantly, particularly for lesions of the sella, tuberculum, and clivus, this approach exploits a natural corridor which avoids critical blood vessels and cranial nerves. Starting with the demonstration of the feasibility of endoscopic pituitary surgery in 1996 (Carrau et al., 1996), the expanded endonasal approach has now become an accepted approach to a large variety of skull base lesions (Kassam et al., 2005). In addition, the endoscope has often been used as an adjunct to the operating microscope for seeing around corners and inspecting operative cavities following skull base tumor resection. In addition, the endoscope may be used as the sole source of light and magnification within the subarachnoid space, so-called “endoscope-assisted microsurgery” (Badie et al., 2004; Perneczky and Fries, 1998). Endoscopy has been much less widely applied to intra-axial brain tumors, mainly because the intraparenchymal space provides no natural medium for light dispersion, thus limiting visualization. However, there are scattered reports of pioneering attempts using an endoscope mounted on a stereotactic frame for biopsy and limited resection of brain tumors

Endoscopic Brain Surgery The endoscope was initially adapted for neurosurgical use by Lespinasse in 1910, who performed intraventricular choroid plexus fulguration in the treatment of two cases of infantile hydrocephalus (Grant, 1996; Prevedello et al., 2007). Subsequently, Dandy refined the technique of intraventricular endoscopy, mainly for choroid plexus extirpation in the treatment of hydrocephalus, and coined the term “neuroendoscopy” (Dandy, 1932). Subsequent modifications and technological improvements have facilitated the development of other common intraventricular endoscopic operations, such as endoscopic third ventriculostomy. The large fluid medium for visualization that is provided by the ventricular system makes neuroendoscopy an intuitive approach for the treatment of obstructive hydrocephalus and other ventricular lesions. More recently, the endoscope has become a critical component in the growth and development of endonasal surgery for lesions of the pituitary region and surrounding skull base. In the anterior and middle skull base, the sinuses provide a natural corridor

26 Endoscopic Port Surgery for Intraparenchymal Brain Tumors

through a tulip-shaped retractor, dating as far back as 1980 (Jacques et al., 1980; Shelden et al., 1980). A decade later, the use of a bullet-shaped dilator for stereotactic-guided endoscopic tumor resection was reported (Otsuki et al., 1990). However, these approaches remained limited in their ability to handle larger and more vascular tumors, as they lacked the versatility of the operating microscope as well as its ability to be dynamically manipulated to provide multiple angles of tumoral visualization.

Cylindrical Retractors for Brain Surgery The introduction of the operating microscope to the field of neurosurgery revolutionized the treatment of brain tumors as well as numerous other disorders. The illumination and magnification provided by the use of the operating microscope significantly improved the ability to visualize and dissect tissue planes. However, because the microscope uses a funneling cone of light for binocular visualization, removal of deep intraparenchymal lesions can require extensive dissection of the overlying brain. In order to minimize tissue trauma inherent to the dissection and resection of deep seated tumors, Dr. Patrick Kelly pioneered a 20-mm-diameter tubular retraction system for the stereotactically-guided microscopic resection of deep brain tumors (Kelly et al., 1986, 1988; Kelly et al., 1986). Dr. Kelly’s technique combined computer technology, frame-based stereotaxy, and microsurgical dissection to deploy a cylindrical retractor as a transcortical conduit into the tumor. Visualization was provided by the operating microscope, and retractor diameter was typically ~20 mm. His approach made the resection of deep subcortical brain tumors far more feasible and far less traumatic than it had been in the past.

263

the challenges inherent to the removal of deep-seated brain tumors, both intraparenchymal and intraventricular. The technique is a combination of elements of traditional endoscopy and traditional microsurgery. Goals include minimization of tissue trauma, maximal precision, and the versatility to address a high variety of lesions. In addition, dynamic manipulation of the port is intended to facilitate resection of lesions larger than the port itself. Initially reported in 2005 (Harris et al., 2005), stereotactic-guided EPS was originally applied to intraventricular lesions alone. This technique was a modification of a previous application which consisted of rolling up a 1 cm vinyl tube, stereotactically placing it into the ventricle, and then unrolling the tube, creating a conduit into the ventricle (Jho and Alfieri, 2002). The ventricles provide a natural conduit for light and magnification, and they can be converted into an air medium for bimanual microsurgery by aspirating cerebrospinal fluid from the ventricle. Visualization is provided by parallel endoscopy through a rod-lens endoscope. Colloid cysts and intraventricular tumors are both quite amenable to this technique of resection (Engh et al., 2010). Application of EPS to intraparenchymal lesions appeared shortly thereafter (Kassam et al., 2009). By converting the endoscopic port itself into a cylindrical retractor, EPS can be performed quite similarly to the manner of Dr. Kelly’s pioneering work. Critical differences include the smaller port size (11.5 mm), completely endoscopic visualization, and the use of frameless image-guidance. Dynamic mobilization of the endoscopic port facilitates visualization and removal of lesions much larger than the port itself. Most importantly, although the working space is small, bimanual microsurgery is performed within the port, which allows for complex tumor resections to be possible.

Technique Endoscopic Port Surgery

Patient Selection

Despite the success with the cylindrical retractor system demonstrated by Dr. Kelly’s group, widespread adoption of this technology was limited. However, the marriage of parallel endoscopy with cylindrical retraction techniques led to the adoption of EPS to address

Careful patient selection is essential to identify those who may be suitable for EPS, based on clinical, radiographic and anatomical considerations. From a clinical perspective, EPS should generally be considered in those patients who develop evidence of a significant

264

neurologic deficit combined with regional tumoral mass effect. In addition, one can consider EPS whenever tissue sampling is required for diagnosis, and when there is a need for neo-adjuvant surgical debulking for cytoreduction prior to the start of systemic therapy. The goals of EPS are similar to the goals of conventional microsurgical tumor resection: maximal tumor resection with functional neurologic preservation. Primary brain tumors, metastases, and cavernous malformations are often suitable candidates for this procedure. Radiographic features which favor EPS are related to the anatomic configuration of the tumoral mass relative to the skull and overlying cortex. A significant cortical cuff certainly favors using EPS for resection. In addition, if the long axis of the tumor is perpendicular to the skull, the port can be advantageous by limiting the amount of white matter dissection required for tumor removal. On the other hand, since cortical and subcortical tissue preservation is the paramount goal of EPS, tumors which feature a wide, shallow cortical component are poor candidates. In addition, tumors which cross a large pial surface, such as the Sylvian fissure are not favorable, and should be addressed with standard microsurgical techniques. Furthermore, tumors which can be removed piecemeal are far more favorable than tumors which require en bloc resection. Finally, for tumors which are amenable to a cisternal approach or a trans-sulcal dissection directly into tumor, EPS does not afford the same benefit of minimizing tissue trauma as for tumors which are completely surrounded by white matter.

Operative Planning and Surgical Technique Pre-operative evaluation and imaging consists of an image-guided fine-cut MRI or CT scan for the purposes of identifying the appropriate site of the craniotomy and a safe cortical entry. A CT angiogram can be particularly useful in delineating important cortical and subcortical vascular structures prior to surgical debulking. In selected cases, fiber tractography or functional MRI scans can further delineate the planned operative trajectory. All patients are placed under general anesthesia. The head is immobilized via three-point fixation using a standard rigid head holder. Surgical head positioning is such that the long

P.G. Ochalski and J.A. Engh

axis of the tumor is elevated within the field, in order to maximize the surgeon’s comfort. Using image guidance, the tumor’s location and the best trajectory into the tumor are marked out on the patient’s scalp. A linear incision is typically performed (except for lesions anterior to the hairline), large enough to allow for a craniotomy of approximately 2.5 cm in diameter. A cruciate dural opening is performed, and the trajectory into the tumor is planned using image-guidance. The process of cannulation and the typical intraoperative view are demonstrated in Fig. 26.2. The decision to enter via a transgyral or trans-sulcal entry point is based on regional vascular anatomy, the location of the tumor, and the eloquence of the surrounding brain. Once the entry point is defined, a small cortisectomy is performed, and a 2 mm brain needle (Elekta, Inc., Stockholm, Sweden) is advanced through the cortical-pial surface. Then, a bullet-shaped dilator (Omni Services, Inc., Wilmerding, PA) is passed over the needle, using gentle rotation to facilitate cannulation of the brain. Directly overlying the dilator is a clear plastic tube, 11.5 mm in outer diameter with lengths varying from 5.0 to 8.5 cm, depending on lesional depth. The port length used during an operation is based on measurements performed using the pre-operative MRI scan. Typically the goal of cannulation is to arrive at approximately 2/3 the depth of the tumor from the pial surface. The brain needle and dilator are then removed, and the port is secured to the scalp using sutures. A 4 mm zero degree rodlens endoscope (Karl Storz, Inc., Tuttlingen, Germany) is brought into the port for visualization. Resection is performed using bimanual microsurgical technique facilitated by tear-drop suction, pituitary forceps, scissors, and bipolar electrocautery. In many cases, two suction devices are used at the same time to facilitate gentle tumoral aspiration. The scope and port angles are adjusted multiple times throughout the resection in order to facilitate dynamic lesion visualization. This technique facilitates removal of lesions much larger than the 11.5 mm port itself. Following resection, meticulous hemostasis is achieved, then the cavity is irrigated with warm saline and lined with Surgicel (Ethicon, Inc., Somerville, NJ). The port is removed, the cortical entry site is also lined with Surgicel, and the dura is then closed in watertight fashion, with or without a dural graft. The bone flap is plated and replaced, and the wound is closed in standard fashion, following copious antibiotic irrigation. Post-operative

26 Endoscopic Port Surgery for Intraparenchymal Brain Tumors

265

Fig. 26.2 (a–c) Technique of cannulation during endoscopic port surgery via a small cortisectiomy. (d) Endoscopic view during resection of a high grade glioma demonstrating bimanual

microsurgical technique within the port. Note the transparent tube which provides an excellent view of the surrounding white matter

care is essentially identical to that for a standard craniotomy. Most patients stay in the intensive care unit overnight, then are discharged from the hospital approximately on post-operative day 3. Sutures are removed at approximately post-operative day 10.

delineate tissue planes and promote hemostasis. In selected cases, especially for larger tumors, intratumoral decompression is performed prior to extracapsular dissection. However, the philosophy is generally to work around the outside of the tumor whenever possible. In contrast, EPS represents an “inside-out” approach to tumor surgery. The regional mass effect of the tumor is exploited in order to facilitate delivery of the tumor into the port, and white matter planes are defined with minimal manipulation. Hemostasis is obtained in slightly more difficult fashion, but the dural opening and white matter manipulation are absolutely minimized. Except for the most vascular of brain tumors (e.g. renal cell carcinoma metastases, solid hemangioblastomas), blood loss does not tend to be excessive, and can be adequately managed through the port.

Discussion Philosophy of Endoscopic Port Surgery The most essential aspect of EPS which represents a departure from microscopic approaches to intraaxial tumors is intra-tumoral cannulation. Typical microsurgical approaches to brain tumors advocate extracapsular dissection whenever possible, both to

266

P.G. Ochalski and J.A. Engh

The application of the operating microscope to neurosurgery was a landmark achievement (Yasargil, 2010). Improvements in the ability to visualize and dissect around brain tumors and other pathological entities have resulted in significant improvements in patient outcomes. Despite the advantages of the operating microscope, in selected cases, even this device can be associated with a high degree of tissue trauma for the sake of tumoral visualization. In contrast, EPS creates a constant corridor width at increasing depths, minimizing tissue dissection both because of the shape of the port and the “flashlight” effect of the

endoscope. This distinction is particularly relevant in an era in which the biological cost of brain surgery is of paramount interest to patients. No longer is the success of tumor surgery defined merely by the degree of resection alone; preservation of neurologic function is equally, if not more critical. Figure 26.3 provides a case illustration of a peri-ventricular high grade glioma which was resected using the endoscopic port with minimal brain manipulation, as well as a schematic of intra-tumoral cannulation. Despite the recent success with EPS, further research and developments in the field are necessary

Fig. 26.3 Schematic and radiographic overview of endoscopic port surgery. (a, b) Example of pre- and post-operative axial contrast-enhanced images of a high-grade glioma bordering the occipital horn of the left lateral ventricle which was resected

using EPS. (c) Schematic of endoscopic port surgery for a deep-seated tumor in the coronal plane, demonstrating intratumoral cannulation and the two-suction technique of tumor removal

26 Endoscopic Port Surgery for Intraparenchymal Brain Tumors

in order to overcome certain limitations with the technique. Improvements in instrumentation and port design are both in the midst of development. For example, a dilatable endoscopic port may be able to limit white matter disruption during cannulation to a greater degree than the current dilatation technique. Furthermore, the ability to differentially dilate the distal end of the port once it has been docked within lesional tissue may improve visualization and resection of tumors. One of the biggest advantages of the port is that it allows bimanual dissection with instruments working in parallel to the endoscope. However, at times there can be a limited working corridor within the port. A design that incorporates a camera into the port could significantly increase working space, thereby facilitating instrument manipulation during tumor resections. Further research is underway to compare the effects and structural footprint of the endoscopic port on the surrounding white matter versus the white matter trauma inherent to standard transcortical approaches using the operating microscope and blade retraction systems.

Conclusion In appropriately selected patients, EPS offers a viable option to achieve the goals of tumor surgery – that is, cytoreduction or complete tumor removal with avoidance of morbidity. Nevertheless, proficiency in microsurgical techniques is an essential component to performing these procedures. Furthermore, a stepwise and methodical introduction of endoscopic techniques in conjunction with conventional approaches can help inexperienced surgeons to gain familiarity with the technology and better understand its limitations. Most importantly, however, correct patient selection is paramount in ensuring good outcomes with EPS. In the senior author’s current practice, approximately one-third of brain tumors that are candidates for resection are approached via EPS; the remainder are treated with conventional approaches. The most advantageous tumors for EPS have a significant cortical cuff (> 1 cm), with significant regional mass effect on the surrounding brain, loose consistency, and low to moderate vascularity. In many such cases, the EPS approach allows a significant reduction to the amount of brain trauma inherent to tumoral resection.

267

References American Cancer Society (2009) Cancer facts & figures. The Society, Atlanta, GA Badie B, Brooks N, Souweidane MM (2004) Endoscopic and minimally invasive microsurgical approaches for treating brain tumor patients. J Neurooncol 69:209–219 Carrau RL, Jho HD, Ko Y (1996) Transnasal-transsphenoidal endoscopic surgery of the pituitary gland. Laryngoscope 106: 914–918 Dandy W (1932) The brain. In: Lewis D (ed) Practive of surgery. WF Prior, Hagerstown, MD, pp 247–252 Engh JA, Lunsford LD, Amin DV, Ochalski PG, FernandezMiranda J, Prevedello DM, Kassam AB (2010) Stereotactically guided endoscopic port surgery for intraventricular tumor and colloid cyst resection. Neurosurgery 67:198–204. discussion 204–195 Grant JA (1996) Victor Darwin Lespinasse: a biographical sketch. Neurosurgery 39:1232–1233 Harris AE, Hadjipanayis CG, Lunsford LD, Lunsford AK, Kassam AB (2005) Microsurgical removal of intraventricular lesions using endoscopic visualization and stereotactic guidance. Neurosurgery 56:125–132. discussion 125–132 Jacques S, Shelden CH, McCann GD, Freshwater DB, Rand R (1980) Computerized three-dimensional stereotaxic removal of small central nervous system lesions in patients. J Neurosurg 53:816–820 Jho HD, Alfieri A (2002) Endoscopic removal of third ventricular tumors: a technical note. Minim Invasive Neurosurg 45:114–119 Kassam AB, Engh JA, Mintz AH, Prevedello DM (2009) Completely endoscopic resection of intraparenchymal brain tumors. J Neurosurg 110:116–123 Kassam AB, Snyderman CH, Mintz A, Gardner P, Carrau RL (2005) Expanded endonasal approach: the rostrocaudal axis. part I. crista galli to the sella turcica. Neurosurg Focus 19:E3 Kelly PJ, Goerss SJ, Kall BA (1988) The stereotaxic retractor in computer-assisted stereotaxic microsurgery. Technical note. J Neurosurg 69:301–306 Kelly PJ, Kall BA, Goerss S, Earnest F 4th (1986) Computerassisted stereotaxic laser resection of intra-axial brain neoplasms. J Neurosurg 64:427–439 Otsuki T, Jokura H, Yoshimoto T (1990) Stereotactic guiding tube for open-system endoscopy: a new approach for the stereotactic endoscopic resection of intra-axial brain tumors. Neurosurgery 27:326–330 Perneczky A, Fries G (1998) Endoscope-assisted brain surgery: part 1: evolution, basic concept, and current technique. Neurosurgery 42:219–224 Prevedello DM, Doglietto F, Jane JA Jr, Jagannathan J, Han J, Laws ER Jr (2007) History of endoscopic skull base surgery: its evolution and current reality. J Neurosurg 107:206–213 Shelden CH, McCann G, Jacques S, Lutes HR, Frazier RE, Katz R, Kuki R (1980) Development of a computerized microstereotaxic method for localization and removal of minute CNS lesions under direct 3-D vision. Technical report. J Neurosurg 52:21–27 Suh JH (2010) Stereotactic radiosurgery for the management of brain metastases. N Engl J Med 362:1119–1127 Yasargil MG (2010) Editorial. Personal considerations on the history of microneurosurgery. J Neurosurg 112:1347

Chapter 27

Intracranial Tumor Surgery in Elderly Patients Paul Ronning, Torstein Meling, Siril Rogne, and Eirik Helseth

Abstract The western world is facing an aging population and this will present challenges for the medical profession insofar as most studies regarding therapy has been undertaken in a younger population. In this chapter we present our experience with intracranial tumor surgery in patients aged more than 70 years old. We find that in our selected aging patients undergoing surgery the results are comparable to the results in younger patients given that the same adjuvant therapy is given. Hence, we believe that the indication for surgery should be based on the physiological age rather than the chronological age of the patient. This and the potential sources of surgical morbidity must always be weighted against the natural history of the disease. Keywords Meningiomas · Brain · Intracranial tumors · ECOG · Astrocytomas · Macroadenoma

Introduction “Primum non noncere”- All medical care is a matter of weighing positive vs. negative effects of treatment for the patient against the natural history of the disease. Applying this basic tenet to intracranial tumor surgery entails that the clinician must contrast the natural history of the tumor against whether the possible positive effects of symptom relief and cyto-reduction outweigh the risk of infection, hematomas, anesthesia and approach related morbidity.

There is an increasing incidence of intracranial tumors up until the age of 75, whereafter the incidence falls steeply (Johannesen et al., 2004). There is no biological rationale for this age-threshold and we suspect that it is due to a negative referral bias due to preconceptions about intracranial surgery in old patients, that old patients often suffer from multiple systemic diseases that increase the risk of surgery and due to reduced difference between the natural history of the disease and the remaining life years (the marginal effect of surgery will diminish as the patient grows older). In this chapter we present the results of surgery in patients aged above 70 years and discuss whether this practice is worthwhile, based on a recently published article by Rogne et al. (2009).

Methodology Data were retrieved from a prospectively collected tumor database containing information on all tumors operated in the department of neurosurgery, Oslo University Hospital, Norway, between 2003 and 2007. A subset of the database, comprising patients aged over 70 years, was further scrutinized by a retrospective systematic review of their medical records to supplement the information in the database. The data was analyzed using Kaplan-Meier curves and multivariate Cox regression models after verifying that the standard assumptions of the models were fulfilled.

Results P. Ronning () Department of Neurosurgery, OSLO University Hospital, Oslo, Norway e-mail: [email protected]

A total of 289 patients were included with a median age of 74.9 (range 70–89.1) years at the time of

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_27, © Springer Science+Business Media B.V. 2011

269

270

surgery. Ninety percent of the tumors could be categorized as meningioma (total n = 79, n = 70 WHO grade I, n = 8 WHO grade II, n = 1 WHO grade III), astrocytoma (total n = 87, n = 82 WHO grade IV, n = 4 WHO grade III, n = 1 WHO grade II), metastases (n = 20 lung cancer, n = 17 malignant melanoma, n = 9 GI cancer, n = 4 breast cancer, n = 12 others) or pituitary adenomas (n = 33). Eight patients died within 30 days of surgery (two patients with meningioma, four with astrocytoma and two with metastases) yielding a surgical mortality of 2.8% in this series. Overall survival rates at 6, 12, 24, and 60 months were 73, 57, 46 and 38% respectively. A multivariate Cox model was fitted with histology, preoperative ECOG score, age, sex, ASA score and resection as opposed to biopsy as independent variables. Increasing preoperative ECOG and biopsy, compared to resection, were significantly (p < 0.05) associated with increasing hazard ratio (1.33 and 1.94 respectively). Astrocytomas and brain metastases had a significantly (p < 0.01) worse prognosis than meningiomas (hazard ratio 17.7 and 12.7, respectively). Furthermore, with regard to functional status at 6 months, we found that 85% were still alive and that stable or improved ECOG scores were witnessed in all patients with pituitary adenomas, >90% of patients with meningiomas, >80% of patients with brain metastases and more than 70% of astrocytoma patients alive at 6 months. Both patients with astrocytomas and metastases showed a significantly decreased hazard ratio with adjuvant treatment (p < 0.01). The astrocytomas and metastases undergoing adjuvant treatment had a median survival of approx 14 and 10 months, respectively, compared to 4 and 5 months without adjuvant treatment.

Discussion In selected patients we find that surgery clearly is worthwhile, in that it offers prolonged survival compared to the natural history of the disease. Furthermore we find that patients respond to adjuvant therapy to a similar extent as younger patients. The main limitation of our data, that must be kept in mind, is the nonrandomized nature of this series, i.e. there is a strong selection bias towards older patients that are good candidates for surgery: patients with good neurological

P. Ronning et al.

function, non-eloquent location and low co-morbidity. Hence, our results are not generalizable to the entire geriatric population, but only to a restricted subset of patients demonstrating these favorable preoperative conditions. Age is a well-known risk factor for poor outcome in glioblastoma patients as detailed by multiple authors (Antonio, 2008; Barnholtz-Sloan et al., 2008; Carson et al., 2007; de Robles and Cairncross, 2008; Iwamoto et al., 2009, 2008; Kurimoto et al., 2007; Lutterbach et al., 2005; Mukerji et al., 2008; Piccirilli et al., 2006). At the same time geriatric glioblastoma patients are known to receive limited treatment due to the preconception that the marginal effect of treatment is low compared to the effect in younger patients. In our series only 18% of our elderly glioblastoma patient underwent both radiation and chemotherapy. However, in our glioblastoma patients undergoing both temodal and radiation therapy we find survival close to what was reported by Stupp et al. (2005), despite the patients in the Stupp trial being limited to an age interval between 18 and 70 years old. Furthermore, in another analysis not yet published, we do not find evidence of a significant interaction between age and treatment. This further supports the notion that elderly patients can benefit from radiation and chemotherapy to a similar extent as younger patients. Similar findings have been published by other authors (Mukerji et al., 2008; Piccirilli et al., 2006). Toxicity can be reduced by offering 40 Gy radiotherapy concurrent with temodal instead of 60 Gy with a marginal effect on survival and with 2 weeks less radiotherapy (Minniti et al., 2008, 2009; Roa et al., 2004). EORTC trial 26062 (http://www.eortc. be/protoc/details.asp?protocol=26062) started accruing patients in 2009 specifically investigating the effect of temodal in patients aged above 70 receiving short course radiation EORTC (2009). Pending the publication of this trial we advocate that biological, instead of chronological age, should be the criterion for offering adjuvant therapy. Due to improved treatment for many common cancers, increased availability of CT and MRI scanners and an aging population there is an increasing trend in overall incidence of brain metastases. The treatment modalities available for brain metastases are surgery, whole brain radiation, chemotherapy (in certain cancers) and radio-surgery. The general consensus is that single, large lesions not amenable to radiosurgery (>3 cm), and lesions of unknown origin with

27 Intracranial Tumor Surgery in the Elderly Patients

negative work-up (approx 10%) should be considered for surgery if the patient has a favorable physiological status and limited systemic disease. Since metastatic disease often confers limited survival time (median survival 8–16 months) patients undergoing surgery should have a short postoperative course and incur limited need for rehabilitation. Hence, deep seated lesions in the basal ganglia, thalamus and brain stem are less than ideal candidates for surgery. Our results indicate again that biological age should not in itself be a limiting factor in offering treatment for metastatic brain disease, but that the general consensus should be followed irrespective of age. Both meningiomas and pituitary adenomas are usually benign slow growing tumors that can create neurological symptoms due to compressive effects and edema. Whether surgery is indicated is a balance between the tumors’ potential for growth, nature of the symptoms and potential morbidity of surgical approach against expected survival. Surgical outcome in the geriatric population has also been reported by other groups claiming good outcomes in the majority of elderly meningioma patients (Black et al., 1998; Cohen-Inbar et al., 2010; D’Andrea et al., 2005; Nakamura et al., 2005; Proust et al., 1997; Riffaud et al., 2007; Roser et al., 2007). We also find that surgery is well tolerated both with regard to survival and functional outcome. However, we believe that these tumors should be symptomatic before surgery is contemplated in this age group. Old asymptomatic patients with radiological growth probably should be referred for radio-surgery (Sonoda et al., 2005). Pituitary adenomas in elderly patients are most often non-secreting macroadenomas (80%), followed by GH- and prolactin secreting tumors (Minniti et al., 2005). A decision to offer surgery should be based on weighing local compressive effects and endocrine disturbances against approach related morbidity. Transsphenoidal surgery has proved to be safe and effective in pituitary surgery (Ferrante et al., 2002; Hong et al., 2008; Nakamura et al., 2007). Macroadenomas and intrasellar GH secreting adenomas are usually well controlled by transsphenoidal surgery (Minniti et al., 2005), whilst the rare prolactin secreting macroadenoma in this population can be controlled by dopamine agonists (Minniti et al., 2005). In our material, we found the life expectancy of elderly patients with pituitary adenomas to be almost identical with the age adjusted survival curves of the

271

general population. Hence, we advocate that patients with symptoms that can be attributed to the pituitary adenoma should undergo transsphenoidal surgery irrespective of their age as long as their physiological reserve is adequate. In conclusion we believe that the indication for surgery should be based on the physiological age rather than the chronological age of the patient. This and the potential sources of surgical morbidity must always be weighted against the natural history of the disease.

References Antonio E (2008) Being old is no fun: treatment of glioblastoma multiforme in the elderly. J Neurosurg 108:639–640 Barnholtz-Sloan JS, Williams VL, Maldonado JL, Shahani D, Stockwell HG, Chamberlain M, Sloan AE (2008) Patterns of care and outcomes among elderly individuals with primary malignant astrocytoma. J Neurosurg 108:642–648 Black P, Kathiresan S, Chung W (1998) Meningioma surgery in the elderly: a case-control study assessing morbidity and mortality. Acta Neurochir 140:1013–1020 Carson KA, Grossman SA, Fisher JD, Shaw EG (2007) Prognostic factors for survival in adult patients with recurrent glioma enrolled onto the new approaches to brain tumor therapy CNS consortium phase I and II clinical trials. J Clin Oncol 25:2601–2606 Cohen-Inbar O, Soustiel JF, Zaaroor M (2010) Meningiomas in the elderly, the surgical benefit and a new scoring system. Acta Neurochir 152:87–97 D’Andrea G, Roperto R, Caroli E, Crispo F, Ferrante L (2005) Thirty-seven cases of intracranial meningiomas in the ninth decade of life: our experience and review of the literature. Neurosurgery 56:956–960 de Robles P, Cairncross G (2008) Glioblastoma in the elderly: an age-old problem. Ann Neurol 64:597–599 Ferrante L, Trillo G, Ramundo E, Celli P, Jaffrain-Rea ML, Salvati M, Esposito V, Roperto R, Osti MF, Minniti G (2002) Surgical treatment of pituitary tumors in the elderly: clinical outcome and long-term follow-up. J Neurooncol 60: 185–191 Hong JF, Ding XH, Lu YC (2008) Clinical analysis of 103 elderly patients with pituitary adenomas: transsphenoidal surgery and follow-up. J Clin Neurosci 15:1091–1095 Iwamoto FM, Cooper AR, Reiner AS, Nayak L, Abrey LE (2009) Glioblastoma in the elderly the memorial sloankettering cancer center experience (1997–2007). Cancer 115:3758–3766 Iwamoto FM, Reiner AS, Panageas KS, Elkin EB, Abrey LE (2008) Patterns of care in elderly glioblastoma patients. Ann Neurol 64:628–634 Johannesen TB, Angell-Andersen E, Tretli S, Langmark F, Lote K (2004) Trends in incidence of brain and central nervous system tumors in Norway, 1970–1999. Neuroepidemiology 23:101–109 Kurimoto M, Nagai S, Kamiyama H, Tsuboi Y, Kurosaki K, Hayashi N, Origasa H, Endo S (2007) Prognostic factors

272 in elderly patients with supratentorial malignant gliomas. Neurol Med-Chir 47:543–549 Lutterbach J, Bartelt S, Momm F, Becker G, Frommhold H, Ostertag C (2005) Is older age associated with a worse prognosis due to different patterns of care? A long-term study of 1346 patients with glioblastomas or brain metastases. Cancer 103:1234–1244 Minniti G, De Sanctis V, Muni R, Filippone F, Bozzao A, Valeriani M, Osti MF, De Paula U, Lanzetta G, Tombolini V, Enrici RM (2008) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma in elderly patients. J Neurooncol 88:97–103 Minniti G, De Sanctis V, Muni R, Rasio D, Lanzetta G, Bozzao A, Osti MF, Salvati M, Valeriani M, Cantore GP, Enrici RM (2009) Hypofractionated radiotherapy followed by adjuvant chemotherapy with temozolomide in elderly patients with glioblastoma. J Neurooncol 91:95–100 Minniti G, Esposito V, Piccirilli M, Fratticci A, Santoro A, Jaffrain-Rea ML (2005) Diagnosis and management of pituitary tumours in the elderly: a review based on personal experience and evidence of literature. Eur J Endocrinol 153:723–735 Mukerji N, Rodrigues D, Hendry G, Dunlop PRC, Warburton F, Kane PJ (2008) Treating high grade gliomas in the elderly: the end of ageism? J Neurooncol 86:329–336 Nakamura K, Iwai Y, Yamanaka K, Kawahara S, Ikeda H, Nagata R, Uda T, Ichinose T, Murata K, Sakaguchi M, Yasui T (2007) The surgical treatment of non-functioning pituitary adenomas in the ninth decade. Neurol Surg 35:371–375 Nakamura M, Roser F, Dormiani M, Vorkapic P, Samii M (2005) Surgical treatment of cerebellopontine angle meningiomas in elderly patients. Acta Neurochir 147:603–610 Piccirilli M, Bistazzoni S, Gagliardi FM, Landi A, Santoro A, Giangaspero F, Salvati M (2006) Treatment of glioblastoma

P. Ronning et al. multiforme in elderly patients. Clinico-therapeutic remarks in 22 patients older than 80 years. Tumori 92:98–103 Proust F, Verdure L, Toussaint P, Bellow F, Callonec F, Menard JF, Freger P (1997) Surgery of intracranial meningiomas in elderly patients. Prognosis factors: 39 cases. Neurochirurgie 43:15–20 Riffaud L, Mazzon A, Haegelen C, Hamlat A, Morandi X (2007) Surgery for intracranial meningiomas in patients older than 80 years. Pres Med 36:197–202 Roa W, Brasher PM, Bauman G, Anthes M, Bruera E, Chan A, Fisher B, Fulton D, Gulavita S, Hao C, Husain S, Murtha A, Petruk K, Stewart D, Tai P, Urtasun R, Cairncross JG, Forsyth P (2004) Abbreviated course of radiation therapy in older patients with glioblastoma multiforme: a prospective randomized clinical trial. J Clin Oncol 22:1583–1588 Rogne SG, Konglund A, Meling TR, Scheie D, Johannesen TB, Ronning P, Helseth E (2009) Intracranial tumor surgery in patients >70 years of age: is clinical practice worthwhile or futile? Acta Neurol Scand 120:288–294 Roser F, Ebner FH, Ritz R, Samii M, Tatagiba MS, Nakamura M (2007) Management of skull based meningiomas in the elderly patient. J Clin Neurosci 14:224–228 Sonoda Y, Sakurada K, Saino M, Kondo R, Sato S, Kayama T (2005) Multimodal strategy for managing meningiomas in the elderly. Acta Neurochir 147:131–136 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352: 987–996

Chapter 28

Intracranial Hemangiopericytoma: Gamma Knife Surgery Jason P. Sheehan and Edward M. Marchan

Abstract This review provides an overview of the biology of hemangiopericytomas as well as an overview of currently available treatment regimens for this highly vascular lesion. While extirpation is the gold standard for diagnosis, tumor control, and relief of mass effect, the high recurrence rate of this tumor type in addition to its proximity to venous channels or skull base locations, can makes resection much less re-resection less attractive. Therefore, stereotactic radiosurgery principally with the Gamma Knife has been used to deliver a steep dose gradient and minimizes the radiation delivered to the surrounding areas. Hence, it becomes possible to deliver a significantly larger and presumably more biologically effective dose to the tumor while limiting the side effects of radiation to normal brain tissue. These characteristics make Gamma Knife radiosurgery (GKS) a very useful tool in treating patients with recurrent hemangiopericytoma or tumors in surgically inaccessible locations. Keywords Hemangiopericytoma · GKS · Meningiomas · Embolization · Radiosurgery · Dose

Introduction Hemangiopericytomas (HPC) are highly vascular and rapidly growing lesions of the central nervous system (Bastin and Mehta, 1992). Embryologically they

J.P. Sheehan () Department of Neurological Surgery, Health Sciences Center, Charlottesville, VA 22908, USA e-mail: [email protected]

belong to the mesenchymal type of tumor class harboring pericytic differentiation (Stout and Murray, 1942). They represent a rare type of brain tumor, and they tend to be misdiagnosed as meningiomas (often mislabeled as angioblastic in the meningioma category) because they can share similar clinical and radiographic findings. It was Begg and Garret who recognized the similarity of the angioblastic meningioma described in 1938 to this soft tissue sarcoma in 1954 (Begg and Garret, 1954). However, more recent assessment of this issue has placed them in a separate category from meningiomas. They are recognized for their aggressive clinical behavior with high recurrence rates and distant metastases even after gross total resection (Goellner et al., 1978; Guthrie et al., 1989; Mena et al., 1991; Pitkethly et al., 1970). HPC and malignant meningioma patients share the potential to develop CNS or systemic metastases (Goellner et al., 1978; Guthrie et al., 1989; Mena et al., 1991). For instance, in Galanis et al.’s series (1998), 50% of patients (17/34) had extraneural recurrence: 14 in bones (82%), 7 in liver parenchymal areas (41%), and 5 had lung metastases (29%). Surgical resection is still considered to be the gold standard for accurate tissue diagnosis and control of mass effect (Fountas et al., 2006; Olson et al., 2010). Most HPCs can be totally removed in toto; however, local recurrence occurs often, with some series quoting as many as 91% of patients (Vuorinen et al., 1996). It is well established that lower cell proliferation indexes (e.g., a MIB-1 index) are associated with a better prognosis, longer intervals to recurrence, a lower rate of metastasis, and extended long term survival (Vuorinen et al., 1996). Those hemangiopericytomas which possess low-grade histology (i.e., Grade 2) usually have a lower MIB-1 index, while high proliferation indexes

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_28, © Springer Science+Business Media B.V. 2011

273

274

favor anaplastic WHO Grade 3 tumors (Vuorinen et al., 1996). It is well established that their proximity to venous sinuses or their highly aggressive, recurrent nature makes an initial gross total resection much less a reoperation difficult (Goellner et al., 1978; Guthrie et al., 1989; Sheehan et al., 2002) Because these tumor recurrences are often circumscribed and focal, stereotactic radiosurgery with the Gamma Knife has become a very attractive technique to achieve tumor control. This chapter focuses upon the current role of stereotactic radiosurgery for hemangiopericytomas.

Surgical Resection of Hemangiopericytomas As stated above, resection is usually the initial treatment for hemangiopericytomas (Olson et al., 2010). In surgery, grossly meningeal hemangiopericytomas are lobulated and have a firm consistency with a pinkgray to red color. They usually have a broad meningeal base but do not tend to spread en plaque or invade brain (Guthrie et al., 1989; Jaaskelainen et al., 1985). They are highly vascular but at surgery are usually separated from surrounding brain without difficulty. Nonetheless, it is this highly vascular nature of this lesion in addition to its origin from deep skull base regions that can make surgery a highly morbid undertaking. Thus, mortality rates can range from 9 to 24% (Guthrie et al., 1989; Jaaskelainen et al., 1985; Olson et al., 2010; Sheehan et al., 2002) Moreover, recurrence is the rule rather than the exception even after gross total resection. Furthermore, distant metastases have been noted to appear between a mean of 63 and 99 months after the first diagnosis (Dufour et al., 2001; Guthrie et al., 1989; Jaaskelainen et al., 1985; Pitkethly et al., 1970; Sheehan et al., 2002). The incidence of local recurrence has varied from 26 to 80%, depending on the quality of resection, the length of follow-up, and the delivery of postoperative RT (Mena et al., 1991; Guthrie et al., 1989). While multiple resections can be feasible, they are nevertheless not usually performed because of the known morbidity of re-operation. For instance, Guthrie et al. (1989) found that if microsurgery was the sole treatment for hemangiopericytoma, the tumor recurred after an average of 29 months. Moreover,

J.P. Sheehan and E.M. Marchan

subsequent recurrences following additional microsurgery occurred at progressively shorter intervals (Guthrie et al., 1989). Therefore, the added clinical setbacks caused by the emergence of iatrogenic neurological deficits in the pursuit of improved resection should therefore be discouraged (Payne et al., 2000). In many cases, preoperative embolization of the feeding vessels can be helpful in controlling bleeding during surgery. Nonetheless, its usefulness is not as evident as it is for meningiomas because of the propensity of hemangiopericytomas to invade cortical vessels. Payne et al. (2000) described how, in his series, one patient was embolized three times over the course of 3 years and the therapy was successful in postponing an additional craniotomy for 3 years after the last embolization. However, a Horner’s syndrome ensued from that procedure. Another patient had two simultaneous recurrences, both of which were embolized, but only one was surgically extirpated. Neither tumor was apparent at 1 year on surveillance MRIs. A third patient was embolized but the intervention was unable to prevent a recurrence and the lesion was treated with Gamma surgery 9 months following embolization. 2 years after the radiosurgical intervention, the tumor had decreased in volume by 35%. The incidence of metastasis increases with time and has been reported as 13%, 33 and 64% at 5, 10 and 15 years respectively (Guthrie et al., 1989; Sheehan et al., 2002) Thus, it is vital that long term follow-up be obtained in order to maximize available adjuvant therapies and to prevent tumor recurrence.

Conventional Radiotherapy in the Management of Hemangiopericytomas Conventional radiotherapy has been proposed for post-surgical treatment of hemangiopericytomas, even when a gross total resection is performed, due to the tumor’s propensity to recur. The purpose of radiation therapy is to delay recurrence of tumors. Dube and Paulson (1974) published the first account of postoperative control of a hemangiopericytoma, and they witnessed a complete radiologic response. Meanwhile, Dufour et al. (2001) showed that post-operative radiotherapy decreased the recurrence rate from 88% after surgical removal alone to 12.5% with adjuvant fractionated radiotherapy. The authors delivered a dose of 50–64 Gy to patients in this series. It has been

28 Intracranial Hemangiopericytoma: Gamma Knife Surgery

275

established that a regional minimum of 50 Gy should be used to prevent early recurrence (Guthrie et al., 1989). Moreover, Mira et al. (1977) reported on a series of 11 patients where hemangiopericytomas were treated with around 29 courses of radiation, and they noticed a strong clinical response in 26 of the 29 courses with complete regression following 14 courses (Mira et al., 1977). Comparatively, a series from our institution published by Payne et al. (2000) described how three patients had received 54 Gy regional fractionated radiation therapy following the first hemangiopericytoma surgical procedure and five had not received any radiation therapy. Those who were not irradiated averaged 2.4 years between their 26 procedures while those that were irradiated post operatively averaged 7.5 years between their 9 procedures. Furthermore, no patient had died or developed systemic metastases in the latter group. This corroborates Guthrie et al.’s (1989) finding that radiotherapy after the first operation extends the mean time before first recurrence from 34 to 75 months while extending survival from 62 to 92 months. In their series, Jaaskelainen et al. (1985) concluded that radiation therapy should be only used after the initial resection. For instance, two patients irradiated after an initial gross total resection were disease free at 167 and 263 months respectively. In contrast, three patients who received conventional radiotherapy for recurrent non-resectable tumors had progression of disease. Interestingly, there have been studies such as from Uemura et al. (1992) who clearly have shown that

radiotherapy can provide a benefit if delivered prior to resection. He hypothesized that radiation can make this type of tumor less vascular while allowing the resection to proceed with less blood loss.

Role of Stereotactic Radiosurgery in the Management of Hemangiopericytomas With radiosurgery, one is able to maximize the efficacy that resection can achieve while using radiation to minimize the potential morbidity associated with re-operation. Radiosurgery can thus achieve a steep dose gradient that minimizes the radiation delivered to the surrounding areas. Consequently, it is possible to deliver a significantly larger and presumably more biologically effective dose to the tumor and minimize other undesired side effects of radiation to normal tissue (Olson et al., 2010). By performing surveillance MRIs, early detection of tumor recurrence or distant tumor formation is achieved. With early detection comes the potential to target smaller tumor volumes and make adjuvant treatment with radiosurgery a more efficient option. Coffey et al. (1993) has postulated that the highly vascular and focal component of the hemangiopericytoma makes it an excellent radiosurgical target Table 28.1. In his series he reported a small subset of five patients with eleven tumors treated with Gamma Knife radiosurgery. A marginal dose of 12–18 Gy

Table 28.1 Stereotactic radiosurgery (LINAC or gamma knife) for hemangiopericytoma

Author

Dosage (Gy) (range)

# of patients

# of lesions

Median F/U (mo)

Median survival after GKS (mo, %)

Coffey et al. (1993) Galanis et al. (1998)

12–18 12–18

5 10

11 20

14.8 36

n/a, 80 n/a, n/a

Payne et al. (2000)

14–37

10

12

25

28, n/a

Sheehan et al. (2002) Olson et al. (2010) Kano et al. (2008)

11–20 2.8–22 11–20

14 21 20

15 28 29

21 68 46

21, 93 68, 33 46, 60

Chang and Sakamoto (2003) n/a: not available

16–24

8

8

44

44, 88

Local control (complete regression or partial regression) 82% (9/11 lesions) 3 patients, no prior XRT: 100% at 36 months 7 patients, + prior XRT: 100% at 12 months 67% (8/12 lesions) at 22 months 80% (12/15 lesions) 46% (13/28 lesions) 72% (21/29 lesions) at 23.3 months 75% (6/8 lesions) at 44 months

276

was used, and 3 patients had undergone prior radiotherapy with doses ranging from 50 to 53 Gy. Nine tumors had post-radiosurgery imaging for comparison. This cohort showed an impressive reduction of size after a mean follow-up of 14.8 months. It is interesting to note that in one of these patients, no tumor was visualized in the post radiosurgery neuro-imaging study. In another report, Galanis et al. (1998) described how in 10 patients with 20 hemangiopericytomas treated with doses ranging from 12 to 18 Gy, there was a dramatic reduction of 14 of them, and a complete eradication of the lesion in three of them. The effect however lasted less than a year. Most of the tumors had also received fractionated radiation therapy and averaged 32 mm in greatest dimension. A subgroup of three patients with solitary tumors less than 25 mm in greatest diameter that had not been treated with radiation therapy all showed complete imaging response that had persisted for a median of 36 months. Finally, he reported a 5-year HPC distant metastasis rate of 33%. Moreover, Payne et al. (2000) followed 10 hemangiopericytoma patients who had 12 stereotactic radiosurgery treatments for these lesions and monitored them for 24.8 months. A dose of 14–37 Gy was given to these tumors. Nine of these hemangiopericytomas initially decreased in size while three remained stable. Local control at a span of 22 months was 67% as four of the nine tumors that shrank later progressed. Two new tumors occurred in patients previously treated. Of the tumors that decreased in volume and did not progressed, the response had totaled 11 months at the time of publication. The follow-up for two tumors that remained unchanged was 10 and 34 months (average 22 months). There were no complications and the quality of life following the procedure was maintained or improved in every case. The senior author (Sheehan et al., 2002) has also obtained positive results in reduction of tumor size with the use of Gamma Knife Radiosurgery. In one study, a series of 14 patients with 15 hemangiopericytomas was analyzed. Seven patients had undergone prior radiotherapy (dose 30–61 Gy), while 27 had prior craniotomies conducted. Doses of 11–20 Gy were implemented with a mean follow-up of 31.3 months. On follow up MR data, 12 of 15 lesions showed reduction in size. Nonetheless, there was a 29% distal failure rate emphasizing the poor ability of radiosurgery to prevent distant metastases.

J.P. Sheehan and E.M. Marchan

In another study of a cohort of patients from the University of Virginia and published by the senior author (Olson et al., 2010), the tumor control rate was found to be lower compared to those of earlier published series, i.e., a 46.4% tumor control rate (13 out of 28 tumors). The mean prescription dose to the tumor was 16.8 Gy. A mean clinical and imaging followup period of 69 months was established (range 2–138 months). Of note, this mean follow-up period was significantly greater than previous series assessing the role of radiosurgery in the treatment of hemangiopericytomas (Chang and Sakamoto, 2003; Coffey et al., 1993; Galanis et al., 1998; Kano et al., 2008; Payne et al., 2000; Sheehan et al., 2002). Hence, it is conceivable that the longer follow-up period provided a greater chance for hemangiopericytoma recurrence or metastasis. Seventeen of the 28 tumors decreased after the first GKS. Six tumors later demonstrated re-growth. Thirteen cases were treated two times or more with GKS. Two of these tumors that underwent multiple GKS demonstrated regression below the initial tumor volume at the last imaging follow-up. The mean time to progression of the tumor after the first GKS was 48 months. The occurrence of new intracranial tumors was 19%; with only one of these tumors demonstrating tumor control after subsequent additional GKS. Nineteen percent of the patients (4 of 21 patients) developed extracranial metastases. These findings of local recurrence rate and extracranial metastases are comparable to previous studies (Chang and Sakamoto, 2003; Kano et al., 2008; Payne et al., 2000; Sheehan et al., 2002). Recently, Kano et al. (2008) demonstrated that a greater marginal dose (≥14 Gy) was significantly associated with better progression-free survival. In their series, 20 patients were reviewed who had undergone GKS for 29 tumors. They reported a median time to the development of intracranial or systemic metastasis at 79.2 months (range, 12.2–158.3 months) after the initial diagnosis. At last assessment, twelve patients (60%) were alive and eight (40%) had died at an average of 62.6 months (range, 13.8–99.4) after GKS and an average of 135.5 months (range, 40.6–255.7) after the initial diagnosis. Four patients (20%) died secondary to dissemination throughout the neuraxis and one (5%) died of liver and lung metastasis. Three patients (15%) developed local tumor progression and died. The follow-up imaging studies demonstrated tumor control in 21 (72.4%) of 29 tumors. However,

28 Intracranial Hemangiopericytoma: Gamma Knife Surgery

277

in this report, the length of imaging follow-up was relative short with a median and mean of 23.3 and 37.9 months after GKS, respectively. Complete eradication was observed in five hemangiopericytomas (Kano et al., 2008). Of note, of the 29 tumors, 21 (72.4%) had received a marginal dose of ≥14 Gy and 8 (27.6%) had received <14 Gy. Consequently, one limitation of this study which the authors do allude to is the existence of a bias between the lower dose group and the higher dose group, especially in relationship to the tumor location, because the prescribed dose was reduced, in most cases, to avoid location-specific toxicity. Hence, of the 21 lesions treated with a marginal dose of ≥14 Gy, 4 (19.0%) exhibited in-field progression. In contrast, four (50.0%) of eight lesions treated to <14 Gy exhibited lack of local control. The marginal dose of ≥14 Gy resulted in a 1- and 5-year progression-free survival rate of 93.3 and 75.4%, respectively. On the other hand, the marginal dose of <14 Gy resulted in a 1- and 5-year progressionfree survival rate of 75.0 and 56.3%, respectively. The mean progression-free survival time after a marginal dose of ≥14 Gy was 79.4 months while it was 45.2 months after a marginal dose of <14 Gy. This difference in progressional free-survival was statistically significant between the two doses (univariate, p = 0.0023; multivariate, p = 0.0185) (Kano et al., 2008). Of note, the mean target volume of lesions treated with ≥14 Gy was 16.5 cm3 and for those treated with <14 Gy was 5.6 cm3 . However, the target volume was not significantly associated with better progressionfree survival. Hence, since the difference in tumor volume was not significant between the two groups, the extension of progression-free survival was most likely related to the greater marginal dose. The dichotomy described above between doses highlights a persistent issue in the field of radiosurgery, i.e., the greater SRS dose is correlated with enhanced tumor control (De Salles et al., 2008). Thus, it is clear from Kano’s et al. (2008) series that a marginal dose of <14 Gy correlated with poor outcome, while the greater dose enhanced tumor control. Chang’s series from Stanford (Chang and Sakamoto, 2003) shows a 75% tumor control rate during a mean follow-up period of 44 months in eight patients with nine hemangiopericytomas. This is comparable to the control rates noticed in other series. A mean dose range of 20.8 Gy was used. The

3-year overall survival rate was 88%. They noted no SRS-related complications. It is interesting to note that this higher tumor control rate was not indicative of a greater rate of lesion eradication or control. Two patients reported clinical improvement, while six described no change in symptoms. Of note, all patients had undergone one or multiple resections prior to radiosurgery. The disadvantage of Gamma surgery, as Payne et al. (2000) suggest, is the delay that can occur waiting for treatment results to be seen as well as the theoretical limitation of repeated GKS treatments because of the increase integral dose that needs to be applied. Thus, in order to minimize the chance of a radiation induced complication, it follows that the tumors should be treated when they are small. Consequently, this should reduce the risk of radiation associated complications and allow for higher marginal doses to be used. If in fact a recurrent tumor grows to a large size, a subtotal debulking with preservation of neurological function, should then be considered prior to Gamma surgery (Payne et al., 2000). The recurrence of these tumors, both at the treated site and distant to it, following Gamma surgery is not uncommon which highlights the importance for the need to perform multiple treatments (Payne et al., 2000). Nonetheless, it is critical to point out that if recurrence occurs and the tumor can no longer be safely resected, that Gamma surgery should be done or repeated. Multiple series from this institution has illustrated the safety of performing repeated Gamma surgery without complications (Payne et al., 2000; Olson et al., 2010). Despite this, there is very little reported in the literature regarding repeated treatment of CNS neoplastic disease with GKS. In our experience when retreating metastases, pituitary tumors and meningiomas, the incidence of radiation associated complication is no higher than usual. Nonetheless, the importance of total integral dose received by the brain when multiple treatments are performed with the Gamma Knife still is unclear (Payne et al., 2000). This issue becomes less critical in cases with relatively limited life expectancy. In conclusion, while resection remains the initial treatment for most patients thereby allowing for diagnosis and control of mass effect as well as first attempt at eradication of the hemangiopericytoma, it is important to stress that recurrent or surgically inaccessible hemangiopericytomas can be treated with Gamma

278

Knife radiosurgery. The aggressive nature of these tumors makes for an unusually high rate of local recurrence and distant intracranial and extracranial metastasis. Long-term tumor control may necessitate repeat radiosurgery or further treatments.

References Bastin KT, Mehta MP (1992) Meningeal hemangiopericytoma: defining the role for radiation therapy. J Neurooncol 14(3):277–287 Begg CF, Garret R (1954) Hemangiopericytoma occurring in the meninges. Cancer 7:602–606 Chang SD, Sakamoto GT (2003) The role of radiosurgery for hemangiopericytomas. Neurosurg Focus 14(5):e14 Coffey RJ, Cascino TL, Shaw EG (1993) Radiosurgical treatment of recurrent hemangiopericytomas of the meninges: preliminary results. J Neurosurg 78:903–908 De Salles AA, Gorgulho AA, Selch M, De Marco J, Agazaryan N (2008) Radiosurgery from the brain to the spine: 20 years experience. Acta Neurochir Suppl 101:163–168 Dube VE, Paulson JF (1974) Metastatic hemangiopericytoma cured by radiotherapy. A case report. J Bone Joint Surg Am 56(4):833–835 Dufour H, Metellus P, Fuentes S, Murracciole X, Regis J, Figarella-Branger D, Grisoli F (2001) Meningeal hemangiopericytoma: a retrospective study of 21 patients with special review of postoperative external radiotherapy. Neurosurgery 48:756–762 Fountas KN, Kapsalaki E, Kassam M, Feltes CH, Dimopoulos VG, Robinson JS, Smith JR (2006) Management of intracranial meningeal hemangiopericytomas: outcome and experience. Neurosurg Rev 29:145–153 Galanis E, Buckner JC, Scheithauer BW, Kimmel DW, Schomberg PJ, Piepgras DG (1998) Management of recurrent meningeal hemangiopericytoma. Cancer 82: 1915–1920 Goellner JR, Laws ERJ, Soule EH, Okazaki H (1978) Hemangiopericytoma of the meninges: Mayo clinic experience. Am J Clin Pathol 70:375–380

J.P. Sheehan and E.M. Marchan Guthrie BL, Ebersold MJ, Scheithauer BW, Shaw EG (1989) Meningeal hemangiopericytoma: histopathological features, treatment, and long-term follow- up of 44 cases. Neurosurgery 25:514–522 Jaaskelainen J, Servo A, Haltia M, Wahlstrom T, Valtonen S (1985) Intracranial hemangiopericytoma: radiology, surgery, radiotherapy, and outcome in 21 patients. Surg Neurol 23:227–236 Kano H, Niranjan A, Kondziolka D, Flickinger JC, Lunsford LD (2008) Adjuvant stereotactic radiosurgery after resection of intracranial hemangiopericytomas. Int J Rad Oncol Biol Phys 72:1333–1339 Mena H, Ribas JL, Pezeshkpour GH, Cowan DN, Parisi JE (1991) Hemangiopericytoma of the central nervous system: a review of 94 cases. Hum Pathol 22:84–91 Mira JG, Chu FC, Fortner JG (1977) The role of radiotherapy in the management of malignant hemangiopericytoma: report of eleven new cases and review of the literature. Cancer 39(3):1254–1259 Olson C, Yen CP, Schlesinger D, Sheehan JP (2010) Radiosurgery for intracranial hemangiopericytomas: outcomes after initial and repeat Gamma Knife surgery. J Neurosurg 112(1):133–139 Payne BR, Prasad D, Steiner M, Steiner L (2000) Gamma surgery for hemangiopericytomas. Acta Neurochir 142: 527–537 Pitkethly DT, Hardman JM, Kempe LG (1970) Angioblastic meningiomas: clinicopathologic study of 81 cases. J Neurosurg 32:539–544 Sheehan JP, Kondziolka D, Flickinger J, Lunsford LD (2002) Radiosurgery for treatment of recurrent intracranial hemangiopericytomas. Neurosurgery 51(4):905–910 Stout AP, Murray MR (1942) Hemangiopericytoma: a vascular tumor featuring Zimmermann’s pericytes. Ann Surg 116: 26–33 Uemura S, Kuratsu J, Hamada J, Yoshioka S, Kochi M, Ushio Y (1992) Effect of radiation therapy against intracranial hemangiopericytoma. Neurol Med Chir (Tokyo) 32: 328–332 Vuorinen V, Sallinen P, Haapasalo H (1996) Outcome of 31 intracranial hemangiopericytomas: poor predictive value of cell proliferation indices. Acta Neurochir (Wien) 138: 1399–1408

Chapter 29

Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors Jesse J. Savage and Jason P. Sheehan

Abstract Cerebral metastases account for the majority of intracranial tumors. Digestive system malignancies contribute 3–8% of cerebral metastases and represent a late complication of systemic disease. Stereotactic radiosurgery employs linear accelerator technology (Linac) or multi-cobalt sources (Gamma Knife) to deliver a high dose of focal radiation to a precise target. The efficacy of stereotactic radiosurgery in treating brain metastases has been established in numerous clinical trials with reports of durable local control and median survival times of 7–12 months. To date, 3 significant retrospective studies have described the clinical outcomes, prognostic factors, tumor control rates and complications associated with stereotactic radiosurgical treatment of brain metastases from gastrointestinal cancer. Herein the authors explicate the natural history of digestive system cerebral metastases and describe the retrospective analyses that demonstrate utility of stereotactic radiosurgery in tumor control. Keywords SRS · Cerebral metastases · Gamma Knife · Cancer · GI · Dose

Introduction The technical limitations and morbidity associated with open neurosurgical approaches to intracranial pathology prompted the development of stereotactic J.P. Sheehan () Department of Neurological Surgery, Health Sciences Center, Charlottesville, VA 22908, USA e-mail: [email protected]

radiosurgery (SRS) pioneered at the Karolinska Institute. Although SRS was conceived by Dr. Lars Leksell in 1949 with the intention of treating relatively benign neurological syndromes, today this therapeutic modality is utilized in a considerably broader context. Stereotactic radiosurgical approaches to primary and metastatic cerebral tumors have been increasingly championed over the last decade secondary to the established rates of effective control of local tumor progression and low associated morbidity. Cerebral metastases represent the most common intracranial malignancies and are estimated to outnumber primary brain tumors 4:1. Digestive system malignancies account for 3–8% of cerebral metastases and represent a late complication of systemic disease (Flickinger et al., 1994; Salvati et al., 1995; Hammoud et al., 1996). The need for an effective, low morbidity therapy to treat brain metastases of digestive tract origin has prompted recent analyses of the utility of SRS in this regard.

Cerebral Metastases The frequency of cerebral metastatic disease has increased over the last 25 years. This trend reflects advancement in the fields of medical and surgical oncology as well as neuroradiology. Patients develop brain metastases in 10–40% of cancer cases and 60–75% of these lesions are symptomatic (Arnold and Patchell, 2001). Clinical findings attributable to cerebral metastases are the presenting symptom in 15% of undiagnosed cancer patients and result in significant morbidity and mortality (Voorhies et al., 1980). Symptoms at presentation include headache (40–50%),

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_29, © Springer Science+Business Media B.V. 2011

279

280

focal neurological deficit (40–50%), seizure (15–20%), and hemorrhagic stroke (5–10%) (Soffietti et al., 2002). Variables including Karnofsky performance status (KPS), tumor type, age, metastases number, and time from diagnosis have prognostic significance in patients with cerebral metastases (Gaspar et al., 1997; Soffietti et al., 2008). Hematogenous dissemination is considered to be the most common route of metastatic spread, though local extension and cerebrospinal fluid (CSF) dissemination are possible. At the time of radiologic diagnoses, 30–50% of cerebral metastases are solitary depending on the resolution of the imaging modality utilized (i.e. computed tomography [CT] vs. magnetic resonance imaging [MRI]) (Greenberg and Arredondo, 2006). Tissue sources of cerebral metastases generally correlate with the distribution of neoplasms in the population. In accordance with this principle, lung and breast cancer account for nearly 60% of brain metastases (Nguyen and Deangelis, 2004). An exception to this concept is observed when considering tumors of the digestive system.

Gastrointestinal Cancer The global incidence of gastrointestinal (G.I.) cancer is approximately 4.2 million cases each year. The American Cancer Society estimates 275,720 new digestive system cancers occurred in 2009. Cancers of the G.I. tract include, in order of incidence, colon, pancreas, rectum, liver & intrahepatic bile duct, stomach, esophagus, gallbladder, small intestine, anus, anal canal and anorectum. Although the G.I. tract is the second most common site for cancer incidence, after the respiratory system, digestive system malignancies account for only 3–8% of cerebral metastases (Flickinger et al., 1994; Salvati et al., 1995; Hammoud et al., 1996). The cerebral metastatic properties of the digestive cancers are varied; however the poor prognosis of individuals diagnosed with gastrointestinal metastases is quite uniform. Colorectal cancer maintains a cerebral metastasis rate of 0.4–1.8% and represents 3–5% of brain metastases (Hammoud et al., 1996). They analyzed 100 patients diagnosed with metastatic brain tumors secondary to colorectal carcinoma and found that primary tumors located in the sigmoid colon and rectum

J.J. Savage and J.P. Sheehan

accounted for 65% of identified lesions (Hammoud et al., 1996). The mean interval between diagnoses of the primary lesions and identification of cerebral metastases for this cohort was determined to be 26 months. The investigators found that the median survival time following diagnoses and treatment with either steroids only, radiotherapy, or surgery to be 1, 3, and 9 months, respectively (Hammoud et al., 1996). A similar study by Patanaphan and Salazar (1993) reported a median interval between diagnoses of colorectal cancer and brain metastases of 33 months with an average survival time of 5.5 months. Pancreatic cancer is the 4th leading cause of cancer deaths in the United States. However, the incidence of cerebral metastases from pancreatic cancer is rare. A report by Park et al. (2003) of 1,229 patients diagnosed with pancreatic cancer determined the rate of metastases to the brain to be 0.3%. The median survival of patients in this study was 2.9 months. Interesting, no patients/cerebral metastatic lesions responded to whole brain radiation therapy (Park et al., 2003). Hepatocellular carcinoma is the 22nd most frequent cancer in the United States; however it is considerably more common in Sub-Sahara Africa and Asia. As with other cancers of the digestive system, cerebral metastases of hepatocellular carcinoma are a rare occurrence. Reports have shown hepatocellular carcinoma cerebral metastases rates to range from 0.3 to 1.3% (Friedman, 1991). Of note, hepatocellular metastases have a high rate of presentation with intracranial hemorrhage (Tanabe et al., 1994). The prominent vascularity and coagulopathies associated with hepatocellular carcinoma provide a theoretical explanation for this clinical association (Friedman, 1991). Cerebral metastasis from gastric carcinoma is extremely rare. A study of 2,700 individuals with a known primary cancer undergoing post-mortem examination at Sloan-Kettering failed to demonstrate a single case of cerebral metastasis from gastric carcinoma (Greenberg and Arredondo, 2006). A report of 3,320 gastric cancer patients seen at the University of Texas, M.D. Anderson Cancer Center documented brain metastases in 0.7% of cases (York et al., 1999). Notably, the majority of metastases reported in this study originated in the proximal stomach as opposed to the more common distal location. York et al. (1999) reported a median survival of 2.4 months following diagnoses of cerebral metastases from gastric cancer.

29 Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors

In 2009, more than 16,000 Americans were diagnosed with esophageal cancer. A single institution series by Weinberg et al. (2003) of 1,588 patients with known esophageal carcinoma demonstrated a cerebral metastases rate of 1.7%. The median survival of the entire cohort was found to be 12.6 months, in contrast to the 3.8 months documented in patients with known metastases to the brain. Risk factors for development of cerebral metastases from esophageal cancer include stage and size of the primary tumor.

Principles of Stereotactic Radiosurgery Stereotactic radiosurgery utilizes linear accelerator technology (Linac) or multi-cobalt sources (Gamma Knife) to deliver a high dose of focal radiation to a precise target volume (typically <3 cm in diameter). The efficacy of SRS in treating brain metastases has been defined in numerous clinical trials with reports of local tumor control ranging from 82 to 90% and median survival times of 7–12 months (Adler et al., 1992; Mehta et al., 1992; Flickinger et al., 1994; Alexander et al., 1995; Auchter et al., 1996; Shiau et al., 1997; Mori et al., 1998). Brain metastases are an ideal application of SRS as these lesions are typically small in volume, spherical in shape, and classically exhibit discrete margins on MRI. Stereotactic radiosurgery allows for radiation dosing in excess of the established 30 Gy in 10 fractions that whole brain radiation therapy (WBRT) employees to achieve control of metastases with minimal toxicity to normal cerebral parenchyma. The principal objectives for SRS of cerebral metastases are to prevent tumor growth and to stabilize or improve the patient’s neurological function. In contrast to traditional fractionated radiation, SRS precisely and accurately delivers a high dose of focal radiation to a defined target volume. The steep dose gradient associated with SRS minimizes damage to adjacent normal brain parenchyma that is documented with the protracted “fall off ” of conventional radiotherapy. The intrinsic properties of SRS allow for control of metastatic lesions with reduced concern for the repair, recovery, redistribution and reoxygenation associated with fractionated radiation. Radiosurgical dose is defined by the maximum and marginal doses delivered to the tumor. Dose constraints for normal brain parenchyma at the tumor periphery limit the effective maximum dose that can be delivered.

281

Stereotactic radiosurgery dose planning for the treatment of cerebral metastases are predicated on a number of clinical studies including the pivotal Radiation Therapy Oncology Group (RTOG) study 90–05 (Shaw et al., 1996, 2000). In this report peripheral radiation doses starting at 18, 15 and 12 Gy for metastases with diameters of 2, 2–3, and 3–4 cm, respectively, were escalated in 3 Gy increments in consecutive cohorts until a 30% complication/grade 3 toxicity rate was observed. In metastases less than or equal to 2 cm, peripheral doses of 24 Gy failed to exceeded 30% grade 3 toxicity. Interestingly, further escalation of radiation dose beyond 24 Gy was not permitted in the RTOG study despite the low rate of observed toxicity. Complications associated with SRS are infrequent. The majority of available series report complication rates ranging from 3 to 15% for single session SRS (Gelblum et al., 1998). Stereotactic radiosurgery related complications are classified as acute or chronic. Examples of acute complications include headache, nausea, vomiting, decline in pre-SRS neurologic status, intratumoral bleeding and seizure. Acute complications occur within 2 weeks from treatment and are most often associated with peritumoral edema (Soffietti et al., 2002). Therefore, acute complications are typically amenable to treatment with steroids. Focal cerebral radionecrosis is a significant chronic complication associated with SRS and occurs in 5–11% of cases (Soussain et al., 2009). The risk of radionecrosis peaks within the first 2 years following treatment, however risk for this complication can persist for decades. Unfortunately it is difficult to differentiate radionecrosis from edema or tumor progression with conventional MRI. Treatment of radionecrosis includes steroids and surgical resection, which is the only method of making a definitive diagnosis. Other less proven approaches to the treatment of radionecrosis include anticoagulation (warfarin or heparin), antivascular endothelial growth factor antibody, trental, vitamin E, bevacizumab and hyperbaric oxygenation (Soussain et al., 2009). The most common chronic complication of WBRT is cognitive dysfunction secondary to radiation induced leukoencepalopathy; however there is limited data addressing neurocognitive outcomes following SRS of brain metastases. Studies by Chang et al. (2007) evaluated 5 patients 200 days following SRS and found 4 to have stable or improved neurocognitive indices of learning and memory, with 3 patients also

282

demonstrating recovery of dexterity and executive functioning (Chang et al., 2007). Another report by Aoyama et al. (2006)documented patients receiving WBRT and SRS demonstrated stable Mini-Mental Status Examination (MMSE) scores at 2 years in contrast to a decrease in MMSE scores observed in a SRS-only cohort. Nevertheless, the decline in MMSE appreciated in the SRS-only cohort was believed to be more attributable to disease progression than radiation induced neurocognitive decline. In contrast to the findings of Aoyama et al. (2006) and Chang et al. (2009) recently published studies that demonstrated decreased neurocognitive dysfunction in patients with brain metastases treated with SRS alone when compared to SRS+WBRT. This report documents a randomized control study of 58 patients (SRS alone group n = 30, SRS+WBRT group n = 28) that was stopped early secondary to results which showed the SRS alone patients had declines of 24% in Hopkins Verbal Learning Test-Revised (HVLT-R) recall at 4 months compared to declines of 52% in the SRS+WBRT group.

Stereotactic Radiosurgery for Brain Metastases of G.I. Tract Origin Studies examining the utility of SRS for the treatment of cerebral metastases of the digestive tract are rare. To date, 3 significant retrospective studies have described the clinical outcomes, prognostic factors that affect survival, local control rates, and complications associated with SRS treatment of brain metastases from G.I. cancer (Schoeggl et al., 2002; Hasegawa et al., 2003; Da Silva et al., 2009). The first of these reports, published by Schoeggl et al. (2002) examined 35 patients with single and multiple cerebral metastases from colorectal carcinoma with a total of 60 lesions treated using the Leksell Gamma Knife. Prior to SRS, diagnosis of brain metastases from digestive cancer was confirmed in all patients by gadolinium-enhanced MRI. Characteristics of the patient population analyzed in this study included ∼2:1 male to female ratio, ∼2:1 single to multiple tumor ratio, average tumor volume of 3.9 cm3 , median age 66 years, ∼6:1 preoperative Karnofsky performance scale (KPS) score more than

J.J. Savage and J.P. Sheehan

70 to less than 70 ratio. Thirteen of the 35 enrolled patients had cerebral-limited disease and 22 had additional metastatic sites including lung (n = 20), liver (n = 16), adrenal (n = 1), abdominal wall (n = 3) and bone (n = 2). Twenty-two patients received SRS-only with a median margin dose of 22 Gy and 13 patients were treated with SRS plus WBRT including a median margin dose of 17 Gy with 30 Gy fractionated in 3 Gy per day over 10 days. The efficacy of treatment was assessed with regular follow-up including radiological evaluation, assignment of KPS score, calculation of glucocorticoid dosing and neurological examination. Hasegawa and colleagues were the next group to publish retrospective data on SRS for brain metastases of the digestive system (Hasegawa et al., 2003). This report describes 39 patients with cerebral metastases from the G.I. tract including lesions originating from colorectal (n = 25), esophageal (n = 11), cholangiocarcinoma (n = 1), duodenal (n = 1) and jejunal tissue (n = 1). In total, 72 metastatic brain tumors were treated. Characteristics of the patient population analyzed in this study included ∼2:1 male to female ratio, median age of 63 years, mean preoperative KPS of 91.5 and ∼2:1 single to multiple tumor ratio. Twenty-two patients had brain-limited metastatic involvement whereas 17 patients had other extracranial metastases including liver (n = 7), spine (n = 2), bladder (n = 1), ovary (n = 1) and adrenal gland (n = 1). Eighty-two percent of patients enrolled in the study received WBRT with a mean fractionated dose of 32 Gy. All patients included in the study underwent SRS using the Gamma Knife for metastatic lesions confirmed with prior contrast-enhanced MRI. Mean maximum and marginal doses of 30.9 and 16.6 Gy, respectively, were utilized in the SRS treatment of metastatic brain tumors with an average volume of 4.9 cm3 . The efficacy of SRS was assessed with regular follow-up including radiological evaluation, assignment of KPS score and neurological examination. The most recent analyses regarding the utility of SRS for the treatment of cerebral metastases of digestive origin is a report published by Da Silva et al. (2009). This retrospective review details the treatment of 40 patients who had undergone SRS to treat a total of 118 metastatic cerebral lesions from G.I. cancers including colonic (n = 25), esophageal (n = 7), rectal (n = 5), pancreas (n = 2) and gastric

29 Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors

(n = 1). Characteristics of the patient population analyzed in this study included ∼1:1 male to female ratio, mean age of 58.7 years, ∼1:1 single to multiple cerebral metastases ratio and an average KPS score of 70. Thirteen patients analyzed in the study had brainlimited metastatic involvement, while 24 patients had extracranial metastases that included lung (n = 14), liver (n = 10), ovary (n = 1), omentum (n = 1) and lymph nodes (n = 6). Three individuals were determined to have a second primary tumor including breast (n = 1), ovary (n = 1), or lung (n = 1). Fifteen patients had SRS+WBRT with a mean fractionated dose of 33.7 Gy. All patients evaluated underwent SRS using the Leksell Gamma Knife for metastatic brain tumors confirmed by prior gadolinium-enhanced MRI. Mean maximum and marginal doses of 41.8 and 18.4 Gy, respectively, were utilized in the SRS treatment of metastatic cerebral lesions with an average volume of 4.3 cm3 . The efficacy of SRS was assessed with regular follow-up including radiological evaluation, assignment of KPS score and neurological examination. The median survival times following SRS in patients with brain metastases from digestive system cancers were determined to be 6, 5, and 6.7 months in the reports by Schoeggl et al. (2002), Hasegawa et al. (2003), and Da Silva et al. (2009) reports, respectively. These data, when compared to that of other reported survival times associated with cerebral metastases including lung (7–15 months), breast (7.8–13 months) and renal cell carcinoma (12.5–15 months), demonstrate that patients with G.I. cancer metastases to the brain have slightly poorer outcomes (Da Silva et al., 2009). The majority of patients with metastatic cerebral lesions have additional metastases, particularly in the lung and liver, and survival is likely limited by the extent of this extracranial disease. Statistical analyses of the potential prognostic factors affecting post-SRS survival were extensively reviewed in the aforementioned studies. A multitude of variables including sex, age, number of brain metastases, pre-SRS KPS, SRS-only, SRS+WBRT, extracranial disease, initial tumor volume and digestive cancer type have been evaluated utilizing the Kaplan-Meier estimator (Schoeggl et al., 2002; Hasegawa et al., 2003; Da Silva et al., 2009). The only variable that has demonstrated statistical significance in either univariate or multivariate analyses of patient survival is KPS score. Schoeggl et al. (2002) reported that

283

a KPS score greater than 70 was a positive prognostic indicator in both univariate and multivariate analyses (p < 0.005). Similarly, a KPS score above 80 demonstrated statistical significance in patient survival from both univariate and multivariate analyses in the report by Da Silva et al. (2009) (p = 0.018 and 0.002, respectively). Additional statistical powering of future studies will assist in elucidating additional significant prognostic factors affecting post-SRS survival for patients with brain metastases from digestive carcinomas. In each of the studies described above local tumor control following SRS was assessed with contrastedMRI studies at the time of patient follow-up (Schoeggl et al., 2002; Hasegawa et al., 2003; Da Silva et al., 2009). Schoeggl et al. (2002) reported local tumor control in 94% of patients at 12 weeks post-SRS. Hasegawa et al. (2003) obtained an 84% rate of local tumor control in cerebral metastases from G.I. tract cancers at the time of final MRI analyses. The data from the Da Silva et al. (2009) series demonstrated absence of local tumor growth following SRS in 91.1% of cases (Fig. 29.1). These results approximate local control rates of other metastatic brain lesions treated with SRS including breast (90–94%), lung (73–94%), and renal cell tumors (83–96%). Interestingly, the addition of WBRT failed to prolong survival time or improve the rate of local tumor control in any of the aforementioned retrospective studies thereby indicating that SRS alone may be effective in preventing tumor growth in brain metastases from digestive cancers. Preoperative symptoms in patients included in the studies above included headaches, motor weakness/hemiparesis, gait instability, speech difficulties, and seizure. Post-stereotactic radiosurgery complications included transient peritumoral edema in 7.2, 8.2 and 17.5% of patients in the Schoeggl et al. (2002), Hasegawa et al. (2003) and Da Silva et al. (2009) report, respectively. Steroid therapy was effective in treating all cases of peritumoral edema associated with SRS in these studies. To date, no cases of intratumoral hemorrhage or radionecrosis have been described in cerebral metastases of digestive cancers following SRS. In conclusion, brain metastases are a relatively common development in patients with cancer arising from the digestive tract. The incidence of brain metastasis is likely to increase as more effective systemic treatments

284

J.J. Savage and J.P. Sheehan

Fig. 29.1 Axial (a) and coronal (b) post-contrast T1 MR imaging obtained from a 75 year old patient with known metastastic G.I. cancer. The documented 4th ventricular tumor had a

pre-SRS volume of 1.8 cm3 and was treated with an 18 Gy margin dose. Six month post-Gamma Knife axial (c) and coronal (d) MR imaging show considerable reduction of tumor size

of the patient’s primary site of cancer and extracranial sites of disease extend survival. Stereotactic radiosurgery is a generally effective means of treating patients with brain metastasis arising from cancers of the digestive tract. Stereotactic radiosurgery, when appropriately indicated, affords effective tumor control and stabilization or improvement in the patient’s neurological examination.

References Adler JR, Cox RS, Kaplan I, Martin DP (1992) Stereotactic radiosurgical treatment of brain metastases. J Neurosurg 76(3):444–449 Alexander E 3rd, Moriarty TM, Davis RB, Wen PY, Fine HA, Black PM, Kooy HM, Loeffler JS (1995) Stereotactic radiosurgery for the definitive, noninvasive treatment of brain metastases. J Natl Cancer Inst 87(1):34–40

29 Stereotactic Radiosurgery for Cerebral Metastases of Digestive Tract Tumors Aoyama H, Shirato H, Tago M, Nakagawa K, Toyoda T, Hatano K, Kenjyo M, Oya N, Hirota S, Shioura H, Kunieda E, Inomata T, Hayakawa K, Katoh N, Kobashi G (2006) Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: a randomized controlled trial. JAMA 295(21):2483–2491 Arnold SM, Patchell RA (2001) Diagnosis and management of brain metastases. Hematol Oncol Clin North Am 15(6):1085–1107, vii Auchter RM, Lamond JP, Alexander E, Buatti JM, Chappell R, Friedman WA, Kinsella TJ, Levin AB, Noyes WR, Schultz CJ, Loeffler JS, Mehta MP (1996) A multiinstitutional outcome and prognostic factor analysis of radiosurgery for resectable single brain metastasis. Int J Radiat Oncol Bio Phys 35(1):27–35 Chang EL, Wefel JS, Hess KR, Allen PK, Lang FF, Kornguth DG, Arbuckle RB, Swint JM, Shiu AS, Maor MH, Meyers CA (2009) Neurocognition in patients with brain metastases treated with radiosurgery or radiosurgery plus wholebrain irradiation: a randomised controlled trial. Lancet Oncol 10(11):1037–1044 Chang EL, Wefel JS, Maor MH, Hassenbusch SJ 3rd, Mahajan A, Lang FF, Woo SY, Mathews LA, Allen PK, Shiu AS, Meyers CA (2007) A pilot study of 16 neurocognitive function in patients with one to three new brain metastases initially treated with stereotactic radiosurgery alone. Neurosurgery 60(2):277–283. discussion 283–284 Da Silva AN, Nagayama K, Schlesinger DJ, Sheehan JP (2009) Gamma Knife surgery for brain metastases from gastrointestinal cancer. J Neurosurg 111(3):423–430 Flickinger JC, Kondziolka D, Lunsford LD, Coffey RJ, Goodman ML, Shaw EG, Hudgins WR, Weiner R, Harsh GR 4th, Sneed PK (1994) A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys 28(4):797–802 Friedman HD (1991) Hepatocellular carcinoma with central nervous system metastasis: a case report and literature review. Med Pediatr Oncol 19(2):139–144 Gaspar L, Scott C, Rotman M, Asbell S, Phillips T, Wasserman T, McKenna WG, Byhardt R (1997) Recursive partitioning analysis (RPA) of prognostic factors in three Radiation Therapy Oncology Group (RTOG) brain metastases trials. Int J Radiat Oncol Biol Phys 37(4):745–751 Gelblum DY, Lee H, Bilsky M, Pinola C, Longford S, Wallner K (1998) Radiographic findings and morbidity in patients treated with stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 42(2):391–395 Greenberg MS, Arredondo N (2006) Cerebral metastases. In: Handbook of neurosurgery. Thieme Medical Publishers, New York, NY, pp 484–491 Hammoud MA, McCutcheon IE, Elsouki R, Schoppa D, Patt YZ (1996) Colorectal carcinoma and brain metastasis: distribution, treatment, and survival. Ann Surg Oncol 3(5): 453–463. 17 Hasegawa T, Kondziolka D, Flickinger JC, Lunsford LD (2003) Stereotactic radiosurgery for brain metastases from gastrointestinal tract cancer. Surg Neurol 60(6):506–514. discussion 514–515

285

Mehta MP, Rozental JM, Levin AB, Mackie TR, Kubsad SS, Gehring MA, Kinsella TJ (1992) Defining the role of radiosurgery in the management of brain metastases. Int J Radiat Oncol Biol Phys 24(4):619–625 Mori Y, Kondziolka D, Flickinger JC, Logan T, Lunsford LD (1998) Stereotactic radiosurgery for brain metastasis from renal cell carcinoma. Cancer 83(2):344–353 Nguyen T, Deangelis LM (2004) Treatment of brain metastases. J Support Oncol 2(5):405–410. discussion 411–416 Park KS, Kim M, Park SH, Lee KW (2003) Nervous system involvement by pancreatic cancer. J Neurooncol 63(3): 313–316 Patanaphan V, Salazar OM (1993) Colorectal cancer: metastatic patterns and prognosis. South Med J 86(1):38–41 Salvati M, Cervoni L, Paolini S, Delfini R (1995) Solitary cerebral metastases from intestinal carcinoma. Acta Neurochir (Wien) 133(3–4):181–183 Schoeggl A, Kitz K, Reddy M, Zauner C (2002) Stereotactic radiosurgery for brain metastases from colorectal cancer. Int J Colorectal Dis 17(3):150–155 Shaw E, Scott C, Souhami L, Dinapoli R, Bahary JP, Kline R, Wharam M, Schultz C, Davey P, Loeffler J, Del Rowe J, Marks L, Fisher B, Shin K (1996) Radiosurgery for the treatment of previously irradiated recurrent primary brain tumors and brain metastases: initial report of radiation therapy oncology group protocol (90–05). Int J Radiat Oncol Biol Phys 34(3):647–654. 18 Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, Farnan N (2000) Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90–05. Int J Radiat Oncol Biol Phys 47(2):291–298 Shiau CY, Sneed PK, Shu HK, Lamborn KR, McDermott MW, Chang S, Nowak P, Petti PL, Smith V, Verhey LJ, Ho M, Park E, Wara WM, Gutin PH, Larson DA (1997) Radiosurgery for brain metastases: relationship of dose and pattern of enhancement to local control. Int J Radiat Oncol Biol Phys 37(2):375–383 Soffietti R, Ruda R, Mutani R (2002) Management of brain metastases. J Neurol 249(10):1357–1369 Soffietti R, Ruda R, Trevisan E (2008) Brain metastases: current management and new developments. Curr Opin Oncol 20(6):676–684 Soussain C, Ricard D, Fike JR, Mazeron JJ, Psimaras D, Delattre JY (2009) CNS complications of radiotherapy and chemotherapy. Lancet 374(9701):1639–1651 Tanabe H, Kondo A, Kinuta Y, Matsuura N, Hasegawa K, Chin M, Saiki M (1994) Unusual presentation of brain metastasis from hepatocellular carcinoma–two case reports. Neurol Med Chir (Tokyo) 34(11):748–753 Voorhies RM, Sundaresan N, Thaler HT (1980) The single supratentorial lesion. An evaluation of preoperative diagnostic tests. J Neurosurg 53(3):364–368 Weinberg JS, Suki D, Hanbali F, Cohen ZR, Lenzi R, Sawaya R (2003) Metastasis of esophageal carcinoma to the brain. Cancer 98(9):1925–1933 York JE, Stringer J, Ajani JA, Wildrick DM, Gokaslan ZL (1999) Gastric cancer and metastasis to the brain. Ann Surg Oncol 6(8):771–776

Chapter 30

Malignant Brain Tumors: Role of Radioresponsive Gene Therapy Hideo Tsurushima and Akira Matsumura

Abstract Patients with malignant gliomas have a very poor prognosis. To explore a novel and more effective approach for the treatment of malignant gliomas, a strategy combining suicide gene therapy and radiation treatment (RT) was designed with insertion of a radio-inducible promoter. The plasmids used in this study included one containing the radio-inducible early growth response gene 1 (Egr-1) promoter, which yielded the best response with fractionated RT. Radio-inducible plasmids that included apoptosis-inducible therapeutic genes caused the apoptosis of glioma cells in response to RT in our in vitro experiments. The radio-inducible plasmids were transfected into subcutaneous glioma cells in nude mice by electroporation in vivo. The plasmids could function in vivo in response to RT and caused a reduction of the tumor volume only in mice that were treated with gene therapies and RT. Our research suggested that apoptosis-inducible gene therapies functioned well together with radio-inducible gene therapies based on the synergistic effects that were observed in combination with RT. Thus, radio-inducible gene therapy may have great potential as a novel treatment because this therapeutic system can be spatially or temporally controlled by exogenous RT and provides specificity and safety due to the limitation of area treated by gene therapies, resulting in decreasing the unpleasant effects to the surrounding normal tissues.

Keywords Genes · Apoptosis · Gene therapy · Malignant gliomas · Radiation treatment · Transgene

Introduction The ability to increase transfection rates in vivo remains one of the greatest challenges in gene therapy. Most researchers have focused on the development of novel gene transfer systems that provide higher transfection rates. In many studies, viral vectors have been used to increase the transfection rate. These conventional vectors transduce the surrounding normal cells in a non-specific manner. It has been reported that suicide gene therapies, including apoptosis-inducible genes under the control of strong viral promoters (CMV, cytomegalovirus immediate early promoter; RSV, Rous sarcoma virus promoter), are an effective strategy for the treatment of cancer. These strong viral promoters are used to compensate for low gene transfection rates. Therefore, the use of the combination with strong viral promoters and suicide gene therapies is limited by organ toxicity due to nonspecific gene transfer (Chen et al., 1994; O’Malley et al., 1995). In the treatment of tumors, via suicide gene transfer, it is essential to restrict the expression of suicide genes to tumor cells or to the treated areas.

Targeted Gene Therapies H. Tsurushima () Department of Neurosurgery, Clinical Medicine, Graduate School of Comprehensive Human Science, University of Tsukuba, Tsukuba, Ibaraki 305-8565, Japan e-mail: [email protected]

To overcome the above-described challenge, several strategies have been developed to localize suicide gene expression to tumor cells through the regulation of

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_30, © Springer Science+Business Media B.V. 2011

287

288

gene expression, which is known as targeted gene therapy. The approach used to target tumors is to express a suicide gene under the control of a cell-specific promoter. For example, the α-fetoprotein promoter has been used to target hepatocarcinoma cells (Huber et al., 1991), the tyrosinase promoter to target melanocytes (Vile and Hart, 1993), the ERBB2 and muc1 promoters to target breast cancer cells (Ring et al., 1996), the carcinoembryonic antigen gene promoter to target carcinoembryonic antigen-producing adenocarcinoma of the lung (Osaki et al., 1994), and the human telomerase reverse transcriptase (hTERT) promoter to target cancer cells; hTERT has been shown to be highly active in 85–90% of human cancers, whereas in most normal somatic cells, it demonstrates a lower or undetectable level of activity (Hodes, 2001). The viral vectors that are employed for gene therapy might engage in lateral gene transfer. For example, retroviral vectors stably transduce only actively dividing cells, and adenoviruses can infect both dividing and non-dividing cells. However, only specific promoters are used in targeted gene therapies, because it would be difficult to develop novel viral vectors that can be transfected into specific cells. In some reports, the activities of several cell-specific promoters were shown to be high in many kinds of stem cells and in cirrhotic liver tissue (Lee et al., 1998; Thomson et al., 1998; Youssef et al., 2001). Therefore, the use of cell-specific promotermediated gene therapies might cause unknown adverse effects in these cells and tissues. In some experiments, the expression of genes under the control of cell-specific promoters was found to be highly dependent on the cell type (Komata et al., 2002; Fukazawa et al., 2007). Despite these findings, the safety of targeted gene therapies with cell-specific promoters is not guaranteed because only some cell types have been examined, and most of the tested cell-specific promoters function in normal cells.

Radio-Inducible Gene Therapies Concerning the safety of gene therapy, we suspect that the use of promoters that are regulated by external factors might provide certain advantages with respect to avoidance of the suicide gene expression in the unexpected cells, such as normal cells. The use of external factor-mediated gene therapies would

H. Tsurushima and A. Matsumura

allow control of the spatial and temporal expression of genes. These external factors might include heat shock, hypoxia, chemotherapy, and radiation, among others. For example, hypoxic conditions activate a signaling pathway that is mediated by various transcriptional factors such as hypoxia inducible factor-1, which binds to a specific promoter known as the hypoxiaresponse element (HRE) (Rocha, 2007). Heat shock protein (hsp) promoters and human multidrug resistance gene promoters are also known as heat-inducible promoters (Plathow et al., 2005). Gene therapy that utilizes heat-inducible promoters has been combined with hyperthermia for the treatment of malignant tumors. The localized expression of therapeutic proteins limits systemic toxicity via hyperthermia induction, and the effect of the treatment is enhanced by combining it with gene therapy. Radiation therapy and chemotherapy are frequently performed in patients with malignant tumors, expecting a synergistic function of multiple therapies. Compared to other factors that are mediated externally, radiation would be particularly convenient because of its ease of use. In radiation treatment, the area and duration of irradiation can be controlled. To activate transgene expression in the radiation field, the use of a radiation-inducible promoter would permit the confined expression of a therapeutic gene within a specific irradiated volume for a selected period of time (Hallahan et al., 1995; Weichselbaum et al., 1994). This stringent spatial and temporal control of therapeutic gene expression potentially increases the efficiency and safety of the gene therapy. The early growth response-1 (Egr-1) gene was initially characterized in mouse fibroblasts as an early growth response gene that functions as a transcription factor, similarly to the c-fos and c-jun proto-oncogenes (Sukhatme et al., 1988; Sherman et al., 1990). Egr-1 expression is induced in many cell types in response to various mitogenic stimuli. The Egr-1 promoter possesses six CArG boxes that correspond to the radiationinducible element, which is responsive to ionizing radiation (Fig. 30.1a, b) (Datta et al., 1993) and DNAdamaging chemotherapeutic agents (Park et al., 2002). Therefore, the Egr-1 promoter can be used effectively in cytotoxic gene therapy in combination with RT to regulate the expression of tumor-sensitive genes within the targeted region (Staba et al., 1998). The Egr-1 promoter was inserted to an adenovirus vector, Ad.EGR-TNF, in which the pro-apoptotic cytokine

30 Role of Radioresponsive Gene Therapy

a

289

b

Fig. 30.1 Green fluorescence protein (GFP) cDNA was cloned into pCIneo (GIBCO-BRL, Grand Island, NY) to create pCIGFP. The CMV promoter in pCIneo and pCI-GFP was replaced with the Egr-1 promoter to produce pEGR and pEGR-GFP. U251 human glioma cells (1 × 105 cells) were seeded on day 0 in 24-well culture plates and transfected on day 1 with 1 μg of the plasmids using Lipofectamine 2000. The number of GFPpositive cells was measured on day 2 using a FACScan flow cytometer. Radiation treatment (RT) was performed on day 2 using a 60 Co γ-ray source, or fractionated administrations were conducted (days 2–4). On the day after the last RT, the number of GFP-positive cells was measured. Gene expression is presented as the fold increase in the percentage of GFP-positive cells compared with the results obtained for corresponding non-irradiated sample transfected with the plasmid. The cells were transfected

with pCI-GFP or pEGR-GFP on day 1 and subjected to 5 Gy of radiation on day 2. The data are expressed as the fold increase. Peak responses of pEGR-GFP were observed from 16 to 24 h after RT (Fig. 30.1a). The cells were transfected with pEGRGFP and pCI-GFP on day 1 and exposed to a single RT dose of 5, 10, or 15 Gy on day 2 or a fractionated dose of 5 Gy/day on days 2–3 and 5 Gy/day on days 2–4. The GFP-positive cells were counted 24 h after the last RT. The response to RT was dosedependent and a maximal response was observed using a single dose of 15 Gy (Fig. 30.1b). The fractionated RT induced a higher pEGR-GFP response compared to a single identical RT. A peak response was observed with 3 doses of 5 Gy/day in U251 cells. The results shown represent the average and standard deviation of three independent experiments. NP; not performed

tumor necrosis factor-α (TNF-α) was under the control of CArG elements, thus resulting in the induction of TNF-α in response to radiation or chemotherapy (Sukhatme et al., 1988). Ad.EGR-TNF has been tested in phase I and phase II clinical trials and demonstrated antitumor effects in response to ionizing radiation or anticancer agents, including resveratrol (Bickenbach et al., 2008). Recently, phase II and phase III clinical trials have been initiated to investigate Ad.EGR-TNF for local treatment of advanced pancreatic cancer. pE9, a synthetic promoter that contains CArG elements, was generated to drive the production of inducible nitric oxide synthase (iNOS), which resulted in the radiosensitization of hypoxic tumor cells (Coulter et al., 2008). Similarly, a radioinducible iNOS gene therapy approach utilizing the p21(WAF1)-promoter resulted in the sensitization of both p53 wild-type RIF-1 tumors and p53 mutant HT29 tumors to fractionated radiotherapy (McCarthy et al., 2007).

Application of Apoptosis-Inducible Genes to Radio-Inducible Gene Therapies The ability of external-mediated factors to control transgene expression and to localize the treatment area of therapeutic gene expression to avoid adverse treatment effects on the normal tissue that surrounds the cancer cells would be ideal. However, this type of localization does not produce consistent results due to the use of weak specific promoters (Hallahan et al., 1995; Kasahara et al., 1994). Most of the specific promoters used above function in normal cells and are not as strong as those derived from viruses, such as CMV or RSV. Gene therapies using specific promoters might fail to induce gene expression levels that generate a therapeutic effect. The achievement of high expression levels in response to a specific stimulation is desirable. For example, a plasmid including the

290

H. Tsurushima and A. Matsumura

Egr-1 promoter and green fluorescence protein (GFP) was transfected into U251 or U87 glioma cell lines, and the frequencies of GFP-positive cells were evaluated by flow cytometry. The expression levels achieved with 15 Gy of radiation treatment were just 1.5–1.6 times the maximum levels, which were insufficient for radiation-inducible gene therapies (Fig. 30.1a, b) (Tsurushima et al., 2007a, b, 2008). In another report describing a plasmid containing the Egr-1 promoter and the luciferase gene, the fold increase induced by radiation was approximately 5 to 20 times the optimal values (Takahashi et al., 1997). An approximately 20-fold increase was obtained with the addition of 3 MBq/mL of I-131 to human AsPc-1 pancreatic adenocarcinoma cells in 96-well culture plates. However, local radiation doses might not be feasible under these conditions. The Egr-1 promoter may be upregulated

modestly, which may be insufficient for gene therapy. Because most endogenous cell promoters function under physiological conditions, it is reasonable to conclude that these promoters are naturally weak. To achieve large increases in expression with specific stimuli, the entire treatment system must be considered. We focused on the caspase cascade to induce apoptosis. Sukhatme et al. (1988) combined the EGR-1 promoter with the TNF-α gene as a therapeutic gene by generating the Ad.EGR-TNF vector. We selected TRAIL and caspase-8 as therapeutic genes in our radio-inducible gene therapies (Figs. 30.2 and 30.3) (Tsurushima et al., 2007b, 2008), which resulted in activation of the caspase cascade. These gene therapies have been shown to have synergistic effects with respect to the induction of apoptosis in response to RT (Kawashita et al., 1999). These synergistic effects

Fig. 30.2 The CMV promoter in pCIneo was replaced with the Egr-1 promoter to produce pEGR. TRAIL and CSP8 were cloned and inserted into pCI-neo and pEGR to produce pCITRAIL, pCI-CSP8, pCI-TRAIL and pEGR-CSP8. On day 0, 1 × 105 of U251 cells were seeded in 24-well culture plates, and 1 μg of the plasmids were transfected into the cells on day 1. RT was performed on day 2, or fractionated administrations were conducted (days 2–4). The cells were assayed 24 h after the last RT. Annexin-V-fluorescein isothiocyanate and propidium iodide were used. All analyses were performed using a FACScan flow cytometer. pCI-TRAIL, pCI-CSP8, pEGR-TRAL, and pEGRCSP8 were transfected into cells on day 1. RT was performed using one dose of 15 Gy on day 2. RT increased the sensitivity to apoptosis in the TRAIL and CSP8 gene therapies. RT was

performed by administering a fractionated dose of 5 Gy/day on days 2, 3, and 4. The assays were performed on day 5. Apoptosis induction in the pEGR-TRAIL and pEGR-CSP8 transfection alone was similar to that observed for pEGR transfection alone. Furthermore, the ability to induce apoptosis with a combination of pEGR-TRAIL or pEGR-CSP8 and RT was similar to that observed with a combination of pCI-TRAL or pCI-CSP8 and RT. A significant difference in apoptosis induction was observed in comparisons between the treatment with pEGR-TRAIL or pEGR-CSP8 alone and its combination with RT. The results shown represent the average and standard deviation of three independent experiments. ∗ : p < 0.05, ∗∗ : p < 0.01, Student’s t test

30 Role of Radioresponsive Gene Therapy

Fig. 30.3 Xenograft tumor models were generated by subcutaneously injecting U251 cells (2 × 106 cells) into the hind limb of 8- to 9-week-old male BALB/c nude mice. Treatment was initiated when the tumor volume had reached 50–100 mm3 . Electroporation was performed for in vivo plasmid transfection, following conditions: 150 V/cm, eight 50-ms pulses at a frequency of 1 pulse/s of the electrical stimulus. Ten μg of the plasmids were injected into the subcutaneous tumors. Subsequently, electroporation was performed. Plasmid transfections were performed once a day for 3 days. The lower parts of their mouse bodies carrying the tumors were irradiated. Transfection was

might compensate for the weakness of the EGR-1 promoter. In conclusion, our results indicate that radioinducible gene therapy may have great potential as a novel therapy because this therapeutic system can be spatially or temporally controlled by exogenous RT and provides specificity and safety for use in humans.

References Bickenbach KA, Veerapong J, Shao MY, Mauceri HJ, Posner MC, Kron SJ, Weichselbaum RR (2008) Resveratrol is an effective inducer of CArG-driven TNF-a gene therapy. Cancer Gene Ther 15:133–139 Chen SH, Shine HD, Goodman JC, Grossman RG, Woo SL (1994) Gene therapy for brain tumors: regression of experimental gliomas by adenovirus-mediated gene transfer in vivo. Proc Natl Acad Sci USA 91:3054–3057 Coulter JA, McCarthy HO, Xiang J, Roedl W, Wagner E, Robson T, Hirst DG (2008) The radiation-inducible pE9 promoter driving inducible nitric oxide synthase radiosensitizes hypoxic tumor cells to radiation. Gene Ther 15:495–503 Datta R, Taneja N, Sukhatme VP, Qureshi SA, Weichselbaum R, Kufe DW (1993) Reactive oxygen intermediates target CC(A/T)6GG sequences to mediate activation of the early

291

performed by in vivo electroporation on days 1, 2, and 3, and RT was performed on days 2, 3, and 4. The animal models in which tumor growth was evaluated comprised the pEGR, pEGR & RT, pEGR-TRAIL, pEGR-TRAIL & RT, pEGR-CSP8, and pEGRCSP8 & RT groups. Tumor growth was observed. In the pEGR & RT group, transient regressions in tumor size were observed. In the pEGR-TRAIL & RT and pEGR-CSP8 & RT groups, a consistent decrease in tumor size was observed after day 7, and tumor regrowth was not observed on day 39. The results shown represent the average and standard deviation of eight mice per group. ∗ : p < 0.01, Student’s t test

growth response 1 transcription factor gene by ionizing radiation. Proc Natl Acad Sci USA 90:2419–2422 Fukazawa T, Maeda Y, Durbin ML, Nakai T, Matsuoka J, Tanaka H, Naomoto Y, Tanaka N (2007) Pulmonary adenocarcinoma-targeted gene therapy by a cancer- and tissue-specific promoter system. Mol Cancer Ther 6:244–252 Hallahan DE, Mauceri HJ, Seung LP, Dunphy EJ, Wayne JD, Hanna NN, Toledano A, Hellman S, Kufe DW, Weichselbaum RR (1995) Spatial and temporal control of gene therapy using ionizing radiation. Nat Med 1:786–791 Hodes R (2001) Molecular targeting of cancer: telomeres as targets. Proc Natl Acad Sci USA 98:7649–51 Huber BE, Richards CA, Krenitsky TA (1991) Retroviralmediated gene therapy for the treatment of hepatocellular carcinoma: an innovative approach for cancer therapy. Proc Natl Acad Sci USA 88:1550–1552 Kasahara N, Dozy AM, Kan YW (1994) Tissue-specific targeting of retroviral vectors through ligand-receptor interactions. Science 266:1373–1376 Kawashita Y, Ohtsuru A, Kaneda Y, Nagayama Y, Kawazoe Y, Eguchi S, Kuroda H, Fujioka H, Ito M, Kanematsu T, Yamashita S (1999) Regression of hepatocellular carcinoma in vitro and in vivo by radiosensitizing suicide gene therapy under the inducible and spatial control of radiation. Hum Gene Ther 10:1509–1519 Komata T, Kondo Y, Kanzawa T, Ito H, Hirohata S, Koga S, Sumiyoshi H, Takakura M, Inoue M, Barna BP, Germano IM, Kyo S, Kondo S (2002) Caspase-8 gene therapy using the human telomerase reverse transcriptase promoter for malignant glioma cells. Hum Gene Ther 13:1015–1025

292 Lee HW, Blasco MA, Gottlieb GJ, Horner JW 2nd, Greider CW, DePinho RA (1998) Essential role of mouse telomerase in highly proliferative organs. Nature 392:569–574 McCarthy HO, Worthington J, Barrett E, Cosimo E, Boyd M, Mairs RJ, Ward C, McKeown SR, Hirst DG, Robson T (2007) p21((WAF1))-mediated transcriptional targeting of inducible nitric oxide synthase gene therapy sensitizes tumours to fractionated radiotherapy. Gene Ther 14:246–255 O’Malley BW, Chen SH, Schwartz MR, Woo SL (1995) Adenovirus-mediated gene therapy for human head and neck squamous cell cancer in a nude mouse model. Cancer Res 55:1080–1085 Osaki T, Tanio Y, Tachibana I, Hosoe S, Kumagai T, Kawase I, Oikawa S, Kishimoto T (1994) Gene therapy for carcinoembryonic antigen-producing human lunf cancer cells by cell type-specific expression of herpes simplex virus thymidine kinase gene. Cancer Res 15:5258–5261 Park JO, Lopez CA, Gupta VK, Brown CK, Mauceri HJ, Darga TE, Manan A, Hellman S, Posner MC, Kufe DW, Weichselbaum RR (2002) Transcriptional control of viral gene therapy by cisplatin. J Clin Invest 110:403–410 Plathow C, Lohr F, Divkovic G, Rademaker G, Farhan N, Peschke P, Zuna I, Debus J, Claussen CD, Kauczor HU, Li CY, Jenne J, Huber P (2005) Focal gene induction in the liver of rats by a heat-inducible promoter using focused ultrasound hyperthermia: preliminary results. Invest Radiol 40:729–735 Ring CJA, Harris JD, Hurst HC, Lemoine NR (1996) Suicide gene expression induced in tumor cells transduced with recombinant adenoviral, retroviral, and plasmid vectors containing the ERBB2 promoter. Gene Ther 3:1094–1103 Rocha S (2007) Gene regulation under low oxygen: holding your breath for transcription. Trends Biochem Sci 32:389–397 Sherman ML, Datta R, Hallahan DE, Weichselbaum RR, Kufe DW (1990) Ionizing radiation regulates expression of the c-jun protooncogene. Proc Natl Acad Sci USA 87: 5663–5666 Staba MJ, Mauceri HJ, Kufe DW, Hallahan DE, Weichselbaum RR (1998) Adenoviral TNF-alpha gene therapy and radia-

H. Tsurushima and A. Matsumura tion damage tumor vasculature in a human malignant glioma xenograft. Gene Ther 5:293–300 Sukhatme VP, Cao XM, Chang LC, Tsai-Morris CH, Stamenkovich D, Ferrira PC, Cohen DR, Edwards SA, Shows TB, Curran T, Le Beau MM, Adamson ED (1988) A zinc finger-encoding gene coregulated with c-fos during growth and differentiation, and after cellular depolarization. Cell 53:37–43 Takahashi T, Namiki Y, Ohno T (1997) Induction of the sucide HSV-TK gene by activation of the Egr-1 promoter with radioisotope. Human Gene Ther 8:827–833 Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, Jones JM (1998) Embryonic stem cell lines derived from human blastocysts. Science 282:1145–1147 Tsurushima H, Yuan X, Dillehay LE, Leong KW (2007a) Radioresponsive gene therapy for malignant glioma cells without the radio-sensitive promoter. Caspase-3 gene therapy combined with radiation. Cancer Lett 246:318–323 Tsurushima H, Yuan X, Dillehay LE, Leong KW (2007b) Radioresponsive tumor necrosis factor-related apoptosisinducing ligand (TRAIL) gene therapy for malignant brain tumors. Cancer Gene Ther 14:706–716 Tsurushima H, Yuan X, Dillehay LE, Leong KW (2008) Radiation-inducible caspase-8 gene therapy for malignant brain tumors. Int J Radiat Oncol Biol Phys 71:517–525 Vile RG, Hart IR (1993) Use of tissue-specific expression of the herpes simplex virus thymidine kinase gene to inhibit growth of established murine melanomas following direct intratumoral injection of DNA. Cancer Res 53: 3860–3864 Weichselbaum RR, Hallahan DE, Beckett MA, Mauceri HJ, Lee H, Sukhatme VP, Kufe DW (1994) Gene therapy targeted by radiation preferentially radiosensitizes tumor cells. Cancer Res 54:4266–4269 Youssef N, Paradis V, Ferlicot S, Bedossa P (2001) In situ detection of telomerase enzymatic activity in human hepatocellular carcinogenesis. J Pathol 194:459–465

Part IV

Prognosis

Chapter 31

Brain Tumors: Quality of Life Cristina D’Angelo, Antonio Mirijello, Giovanni Addolorato, and Vincenzo Antonio D’Angelo

Abstract Brain tumors and their treatment may have significant physical, cognitive, emotional and social effects on patients. Patients affected by brain tumors frequently exhibit poor quality of life (QoL) because of tumor symptoms, diagnostic examinations, way of communication of the diagnosis and treatments. In brain tumor patients, palliating symptoms and maintaining or improving Health-Related quality of life is particularly relevant. This chapter will analyze several aspects related to QoL in patients affected by brain tumors. Keywords Brain · Tumor · QoL · Radiotherapy · Gliomas · Astrocytoma

Introduction Health is defined by the WHO as “a state of complete physical, social and mental well-being, and not merely the absence of disease or infirmity.” (Ottawa Charter for Health Promotion, 1986). Quality of life (QoL) is defined by the WHO as “individual’s perceptions of their position in life in the context of the culture and value system where they live, and in relation to their goals, expectations, standards and concerns. It is a broad ranging concept, incorporating in a complex way a person’s physical health, psychological

C. D’Angelo () Department of Internal Medicine, Catholic University of Rome, Gemelli Hospital, l.go Gemelli, 8, 00168 Rome, Italy e-mail: [email protected]

state, level of independence, social relationships, personal beliefs and relationship to salient features of the environment.” (WHO, 1996). Six broad domains were identified to describe core aspects of QoL cross-culturally: physical domain (e.g., energy, fatigue), psychological domain (e.g., positive feelings), level of independence (e.g., mobility), social relationships (e.g., practical social support), environment (e.g., accessibility of health care) and personal beliefs/spirituality (e.g., meaning in life). Basing on these concepts, the domains of health and QoL are complementary and overlapping. QoL reflects the perception of individuals that their primary needs are being satisfied. The goal of improving the QoL has become of increasing importance in health promotion, alongside preventing avoidable ill-health. This is particularly important in relation to meet the needs of older people, the chronically sick, terminally ill, and disabled populations (Orley and Kuyken, 1994). Brain tumors significantly impact QoL outcomes. This chapter will analyze several aspects related to QoL in patients affected by brain tumors.

Quality of Life QoL is a complex entity that involves several different aspects of life, including emotional functioning, physical functioning, cognitive functioning, social functioning and spiritual well-being (Locke et al., 2007). The term “health related” QoL refers to aspects of the disease manifestation such as pain, apprehension, depressed mood and functional impairment (Muldoon et al., 1998), which are able to modify the pre-existing set of satisfying or unsatisfying aspects of the QoL

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_31, © Springer Science+Business Media B.V. 2011

295

296

(Giovagnoli et al., 2005). At present the assessment of the patient’s experience of disease and treatment is a central component of health care and healthcare research (Muldoon et al., 1998).

Quality of Life in Oncologic Patients The outcome in oncologic patients is no longer restricted to mortality and/or morbidity. The objective of any anticancer therapy extends well beyond prolongation of survival. Palliation of symptoms and improvement or maintenance of QoL are important therapeutic goals. It is important to consider not only physical conditions, but also all the aspects related to the individual. Ignoring psychological, social and familial aspect, such as working and copying disability, reduces the therapeutic approach and possibility. Recent evidences indicate that psychosocial factors influence immune function; suppressed immune function has been considered as the biological link between psychosocial factors and cancer (Kaplan and Miner, 2000). A diagnosis of cancer represents a negative experience and could be associated with down regulation of the immune system (Kaplan and Miner, 2000). Moreover, interpersonal stressors have been related to greater immune alterations than non-social stressors (Kaplan and Miner, 2000). Consequently, patient’s coping strategy needs to be considered in treatment decisions, in particular to limit the global negative experience related the disease and to support the personal and familial relationships. The assessment of QoL has become common in oncological research and in daily clinical practice and has proved to be useful in symptom management and evaluation of oncological treatment (Giesinger et al., 2009). The benefits of new or existing cancer treatments that maintain or extend survival time need to be weighed against side effects and decrease in the patient’s Health-Related QoL (HRQoL). Thus, HRQoL has become an increasingly important endpoint in cancer studies, next to outcome measures such as overall and disease-free survival, and is most relevant in patients who cannot be radically cured of their disease. In brain tumor patients, palliating symptoms and maintaining or improving HRQoL is particularly relevant. Patients affected by primary brain tumors, mainly

C. D’Angelo et al.

gliomas, or by metastatic tumors in the brain have a poor prognosis and cannot be cured of their disease: have to cope with clinical symptoms, such as motor deficit and epilepsy, and to the decline of cognitive and emotional functioning as a result of cerebral disease. Moreover, side effects of treatments may have an even further negative impact on cerebral functioning. In the case of recurrence of a brain tumor, evaluation and support of QoL are of primary importance because treatments are not always expected to lengthen survival. In this regard, previous studies have examined the association of HRQoL with the burden of neurological symptoms (Osoba et al., 1997; Giovagnoli et al., 2005), showing that the emotional components of the QoL significantly worsen after brain tumor recurrence. To date few studies assessing QoL in patients affected by brain tumors are available. However the limited curative options for these diseases underline the importance of considering QoL.

Quality of Life in Brain Tumors Usually, the primary aim of measuring HRQoL in oncological patients is to better understand the impact of a specific tumor or a specific treatment on the functional, psychological and social health of the individual. With respect to other cancer diseases (i.e. lung or breast cancer), relatively little attention has been paid to the impact of primary brain tumors on HRQoL. The long-term outcome in these patients is largely dependent on tumor typology (e.g., low-grade or high grade tumor). In particular malignant brain tumors usually result in devastating morbidity for those affected and for their families and typically result in death of the patient within a few years or less (Pelletier et al., 2002). For example, even with aggressive radiation and chemotherapy regimes, the median survival for patients with grade 4 astrocytoma (glioblastoma multiforme or GBM) is about 12 months, and for grade 3 gliomas it is about 36 months (Pelletier et al., 2002). Malignant gliomas are considered incurable; less than 2% of GBM patients survive 3 years or more (Pelletier et al., 2002). Further, the side effects of the treatment itself can result in considerable morbidity so the QoL of brain cancer. A better understanding of the emotional distress experienced by brain tumor patients could lead to more

31 Brain Tumors: Quality of Life

effective interventions, thereby enhancing the ability of patients and families to cope with the illness and enjoy a better QoL (Pelletier et al., 2002). However published reports have demonstrated that a significant percentage of general oncology patients report anxiety and depressive symptoms (Arnold et al., 2008; D’Angelo et al., 2008). Failing to adequately treat depression and other psychological symptoms in these patient populations can compromise overall health and QoL. Tumors of the CNSrepresent a risk factor for the development of affective disorders (D’angelo et al., 2008). In particular, depression is common in patients with CNS tumors (D’angelo et al., 2008), and the prevalence of depressive symptoms in these patients can vary from 15 to 38% (Pelletier et al., 2002). The functional impact of a brain tumor diagnosis may result in more specific neuropsychiatric, cognitive and behavioral manifestations (e.g., depressive symptoms, irritability, anxiety). Anxiety may result from situational fear related to diagnosis and prognosis or may be directly related to the effects of the tumor (Arnold et al., 2008).

Quality of Life in Low-Grade Glioma Taphoorn et al. (1994) evaluated the impact of radiotherapy on QoL in long-term survivors of biopsyproved low-grade gliomas without signs of tumor recurrence. In this study, 20 patients had been treated with early radiotherapy, while the other 21 patients had undergone surgery or biopsy only. Nineteen patients with low-grade hematological malignancies, surviving 1–15 years without CNSinvolvement, were considered as controls. Authors concluded that radiotherapy had no negative impact on QoL in these patients. A comparison between high-dose (59.4 Gy in 6.5 weeks) and low-dose (45 Gy in 5 weeks) radiotherapy with conventional techniques was made in a randomized phase III trial in low-grade glioma patients (Heimans and Taphoorn, 2002). The primary endpoint was survival. QoL was evaluated by means of a questionnaire, constructed for this study, consisting of 47 items assessing a range of physical, psychological and social symptoms domains. Patients who received highdose radiotherapy reported more fatigue and insomnia

297

immediately after radiotherapy and poorer emotional functioning at 7–15 months post-randomisation. A study by Reijneveld et al. (2001) compared functional status, QoL and cognitive status of 24 patients suspected of having a low-grade glioma, in whom treatment was deferred and 24 patients with proven low-grade glioma, who underwent early surgery, matched with healthy control subjects for educational level, handedness, age and gender. Both patient groups scored worse on QoL scales than healthy control subjects. Unoperated patients with suspected low-grade glioma scored better on most items than patients with proven low-grade glioma. Conclusion of this study suggest that a wait and-see policy in patients with suspected low-grade has no negative effect on cognitive performance and QoL (Reijneveld et al., 2001).

Quality of Life in High-Grade Glioma A study by Klein et al. (2001) compared 68 newly diagnosed high-grade glioma patients with 50 lung cancer patients and matched healthy controls. In both glioma patients and lung cancer patients HRQoL was significantly lower than that of healthy controls. Glioma patients showed lower neurologic, objective and subjective neuropsychologic functioning than lung cancer patients. Cognitive impairment was observed in all glioma patients. Extent of surgical resection was not related to neuropsychological functioning and antiepileptic drug use was correlated negatively with working memory capacity. Bampoe et al. (2000) evaluated QoL in patients with GBM who participated in a randomised study of brachytherapy as boost treatment. No statistical difference between the two treatment arms (conventional radiation plus a brachytherapy boost versus conventional radiotherapy alone) was found regarding QoL parameters. However, a significant deterioration in Karnofsky Performance Scale (KPS) and in some HRQoL items was found during the first year of follow-up: self care, speech, concentration, cognitive functioning and physical experience deteriorated. In a recent Italian study (Giovagnoli, 1999), HRQoL was evaluated in 57 patients with high-grade malignant gliomas, who had stable disease after multimodalitytreatment (surgery, radiotherapy and chemotherapy). The Functional Living Index – Cancer (FLIC) was

298

employed. This is a self administered visual analogue scale exploring different dimensions of QoL (physical, emotional, social and occupational aspects as well as drug side-effects). It showed that QoL in this selected group of glioma patients was satisfactory and did not differ from the QoL in patients with chronic neurological illnesses. The study suggests that aggressive combination treatment (including adjuvant chemotherapy) does not necessarily affect QoL more than cancer therapies for other tumors. QoL in brain tumor patients was significantly associated with depression and state anxiety. Weitzner et al. (1996) evaluated QoL in a group of 50 brain tumor patients (mainly grade III anaplastic astrocytomas). They found that QoL in glioma patients is most affected by (1) the extent of tumor involvement, bilateral being worse than unilateral; (2) poor performance status; (3) being a woman; (4) having been divorced; (5) undergoing aggressive treatment, including chemotherapy; and (6) being not able to work. Tumor grade and age were not related to QoL in this study. The study shows how both tumor-related and non-tumor-related factors may influence QoL in individual patients. It also indicates that adjuvant chemotherapy may negatively affect HRQoL, although the effects of several treatment modalities on HRQoL need to be appreciated in prospective randomized trials. In this respect, it is important to realize that therapeutic nihilism, resulting in refraining from therapy, may have a negative impact on HRQoL as well and therefore should be rejected. The effect of treatment in recurrent glioma may be measured by tumor response but also by a change in HRQoL. HRQoL was assessed using the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ-C30) and the Brain Cancer Module-20 in two clinical trials enrolling a total of 366 patients with recurrent GBM; 288 patients provided HRQoL data that could be analysed. One hundred and nine patients received temozolomide in a phase II study; 89 patients received temozolomide and 90 patients received procarbazine in a randomized phase III trial. Before disease progression, patients who were treated with temozolomide were found to have an improvement in a number of HRQoL domain scores compared with their pretreatment scores. Patients who were treated with procarbazine reported deterioration in HRQoL (Osoba et al., 2000a).

C. D’Angelo et al.

The same instruments were used in a phase II study in patients with recurrent anaplastic astrocytoma who were treated with temozolomide. After 6 months of treatment, those patients who were free of disease progression reported either an improvement or maintenance of all the preselected HRQoL domains scores (Osoba et al., 2000b). The prognosis of patients with glioma greatly depends on age, performance status, cognitive status and tumor grade. Generally, gliomas cannot be cured by standard treatment. Therefore, experimental and complementary approaches are an important consideration. Moots states that for these reasons individualization of treatment planning for malignant glioma is important (Moots, 1998). Moreover, if the patient should have a clear concept of the nature of the disease and of the possibilities and limitations of various treatment modalities, is still now uncertain. It should be realized that cognitive impairment as well as emotional distress often make it difficult to convey a clear overview of the exact nature of the illness and the details of the treatment options. Generally, several specialists (neurologist, neurosurgeon, radiation oncologist, medical oncologist) are involved in the treatment of glioma patients. This places a great premium on the development of a team that can deal with all the (oncological and neurological) issues. The already mentioned therapeutic nihilism, which is sometimes associated with the treatment of primary brain tumors, may limit enrolment in clinical trials. Nevertheless, treatment efforts may have a positive (albeit temporary) impact on HRQoL.

Quality of Life Measurement in Brain Tumors Standardized HRQoL measurements are used in clinical practice to monitor changes in the symptom experience and in self-reported functioning. QoL measurement as a routine part of daily practice implies that both patients and physicians are aware of the value of discussing a wide range of issues which are not directly related to the illness, but also to emotional and social functioning. A variety of cancer-specific and non-specific tests are available for the measurement of QoL. Despite the specific relevance for measuring HRQoL in brain

31 Brain Tumors: Quality of Life

tumor patients, the interest in HRQoL emerged relatively late in this patient group compared with more common cancers. This could have been related to the low incidence of primary malignant brain tumors and their bed prognosis, and to the fact that the subjective nature of HRQoL assessment may be problematic in brain tumor patients with mental impairments. Measuring outcome in terms of tumor size, time to tumor progression, and overall survival is relatively simpler compared with outcome measures such as impairment, disability, or handicap, which require symptoms scales (Heimans and Taphoorn, 2002). These scales, however, do not adequately measure the patient’s HRQoL. HRQoL is an even more complex outcome measure, demanding a multidimensional instrument that should be filled out by the patient (self-report questionnaire). Both generic and disease-specific questionnaires have been developed and validated to assess HRQoL in cancer patients. In patients affected by brain tumor, the generic and wide used the EORTC QLQ-C30 is usually associated to the Brain Cancer Module-20 or the Functional Assessment of Cancer Therapy (FACT) generic questionnaire together with the FACT brain module (Osoba et al., 1996; Weitzner et al., 1995). The EORTC QLQ-BN-20 includes 20 items assessing visual disorders, motor dysfunction, communication deficit, various disease symptoms (e.g., headaches and seizures), treatment toxicities (e.g., hair loss) and future uncertainty. Filling out these questionnaires takes about 10–20 min (Osoba et al., 1996). The FACT addresses physical, family, social, emotional, and functional well-being. The FACT brain module is a subjective instrument which measures substantially different QoL items than the core instrument (Weitzner et al., 1995). Another widely used brain tumor-specific HRQoL questionnaire is the Functional Assessment of Cancer Therapy-Brain (FACT-Br), to be combined with the generic FACT module (FACT-G) (Cella et al., 1993; Weitzner et al., 1995). Compared to EORTC questionnaires, the focus of FACT is more on psychosocial aspects and less on symptoms. Additionally, a 15-item symptom index was constructed for brain tumor patients (FACT-BrSI). Alternatively, an independent living score (ILS) for use in glioblastomas patients has also been developed (Recht et al., 2003).

299

The Short-Form Health Survey (SF – 36) is a selfreport non cancer-specific questionnaire (Ware and Sherbourne, 1992). It is composed of 36 items. This test can be used as a routine part of daily practice. Other HRQol instruments are the Spitzer Quality of Life Index (SQLI), the Quality Adjusted Live Years (QALY) (Ware and Sherbourne, 1992). Recently, the psychometric properties of singleitem linear analogue scale assessments (LASAs) for measuring overall QoL, physical, emotional, spiritual and intellectual functioning in patients with newly diagnosed high-grade gliomas were investigated. Data suggest that the single-item LASA scales would be an appropriate alternative when a shorter tool is warranted (Locke et al., 2007). These tools are very simple: patients are asked to directly score their QoL or functioning on a 100 mm horizontal line, and the rating is measured from the left edge. Obviously, they do not capture the detailed information obtained from HRQoL questionnaires. The Brain Cancer Module, which is designed to be used with the QLQ–30 or other general questionnaires (de Haes et al., 2000; Osoba et al., 1996), assesses problems specific to brain tumor patients. The module has been modified into the Brain Cancer Module–20 in order to eliminate the overlap with emotional distress items on the QLQ – 30. The module contains 4 multi-item scales: future uncertainty, visual disorder, motor dysfunction, and communication deficit. Furthermore it contains seven single items: headache, seizures, drowsiness, hair loss, itching, weakness of both legs, and difficulties with bladder control. Mackworth et al. (1992) and Osoba et al. (1997) found a relationship between QoL and Karnofsky Performance scores. In the majority of patients, QoL was strongly related to the absence of physical limitations, cognitive function, sex-life and mood disorders. This represent an important feature, since mood disorders and psychological distress in patients with intracranial tumors may be sufficient to warrant psychological and/or pharmacological intervention. Prevalence of mood disorders in brain tumor patients varies widely between studies. Anderson et al. (1999) found that only two of 40 patients tested had clinically significant levels of anxiety as assessed by the Clinical Anxiety Scale. Six of these 40 patients had clinically significant levels of depression as assessed by the Hamilton Rating Scale for Depression. Psychological

300

morbidity was associated with high levels of physical disability and also with cognitive dysfunction, but not with the grade of the tumor or with the extent to which the patient was aware of the prognosis of his or her disease. In a recent longitudinal study D’Angelo et al. (2008) and co-workers assessed state and trait anxiety and depression before and after surgical treatment in consecutive patients with intracranial tumors; before brain surgery patients showed high levels of anxiety and low levels of current depression; during the follow-up period there was no significant variation in the percentage of patients with anxiety whereas a significant increase in the percentage of those with current depression was found.

C. D’Angelo et al.

Long term negative effects of the tumor and its treatment demand an increasing effort of doctors, nurses, and psychologists for supportive care. Psychological support for patients and families (such as counselling, family groups, self support, etc.) could be tailored according to the personal appraisal of QoL, reactions, and strategies of coping. Although these treatments do not necessarily lengthen patients’ survival, they could satisfy the psychological needs of the patients and their families, contributing to improve QoL (Giovagnoli et al., 2005). Nowadays in most of oncology centres there is a psycho/oncology unit, to grant psychological support to patients and to families.

References Conclusions Brain tumors and their treatment may have significant physical, cognitive, emotional, and social effects on the patient. The patient’s partner and family may also experience a negative emotional and social impact. The psychological experience of being affected by a brain tumor is not adequately taken into account by physicians (D’Angelo et al., 2008). Patients with brain tumors frequently exhibit poor QoL related to tumor symptoms, diagnostic examinations, way of communication of the diagnosis, treatment (surgery, radio, chemo-therapy) and finally to consequences of treatment. During the diagnostic period, the physician–patient interaction is generally focused on issues such as diagnosis of the tumor, treatment options, long-term prognosis, and neurological assessment, but evaluation of psychological assessment seems not to have a high priority (D’Angelo et al., 2008). Usually the neurosurgeon focuses attention only on the organic features. For these reasons the psychometric evaluation gains great importance in the global evaluation of patients affected by brain tumors, to recognize and treat psychological disorders quickly, thus improving the HRQoL of these patients (D’Angelo et al., 2008). Apart from treatment of the tumor itself, supportive treatment of patient and patient’s family may include medication (antiepileptic drugs, steroids, antidepressants) and/or psychological/cognitive support. As there is an increase in effective treatments for brain tumor patients, the number of long-term survivors will grow.

Anderson SI, Taylor R, Whittle IR (1999) Mood disorders in patients after treatment for primary intracranial tumours. Br J Neurosurg 13:480–485 Arnold SD, Forman LM, Brigidi BD, Carter KE, Schweitzer HA, Quinn HE, Guill AB, Herndon JE 2nd, Raynor RH (2008) Evaluation and characterization of generalized anxiety and depression in patients with primary brain tumors. Neuro-oncology 10:171–181 Bampoe J, Laperriere N, Pintilie M, Glen J, Micallef J, Bernstein M (2000) Quality of life in patients with glioblastoma multiforme participating in a randomised study of brachytherapy as a boost treatment. J Neurosurg 93:917–926 Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, Silberman M, Yellen SB, Winicour P, Brannon J (1993) The functional assessment of cancer therapy (FACT) scale: development and validation of the general measure. J Clin Oncol 11:570–579 D’Angelo C, Mirijello A, Leggio L, Ferrulli A, Carotenuto V, Scolaro N, Miceli A, D’Angelo V, Gasbarrini G, Addolorato G (2008) State and trait anxiety and depression in patients with primary brain tumors before and after surgery: 1-year longitudinal study. J Neurosurg 108:281–286 de Haes J, Curran D, Young T, Bottomley A, Flechtner H, Aaronson N, Blazeby J, Bjordal K, Brandberg Y, Greimel E, Maher J, Sprangers M, Cull A (2000) Quality of life evaluation in oncological clinical trials-the EORTC model. The EORTC Quality of Life Study Group. Eur J Cancer 36:821–825 Giesinger JM, Golser M, Erharter A, Kemmler G, SchauerMaurer G, Stockhammer G, Muigg A, Hutterer M, Rumpold G, Holzner B (2009) Do neurooncological patients and their significant others agree on quality of life ratings?. Health Qual Life Outcomes 7:87 Giovagnoli AR (1999) Quality of life in patients with stable disease after surgery, radiotherapy, and chemotherapy for malignant brain tumour. J Neurol Neurosurg Psychiatry 67:358–363 Giovagnoli AR, Silvani A, Colombo E, Boiardi A (2005) Facets and determinants of quality of life in patients with recurrent

31 Brain Tumors: Quality of Life high grade glioma. J Neurol Neurosurg Psychiatry 76: 562–568 Heimans JJ, Taphoorn MJB (2002) Impact of brain tumour treatment on quality of life. J Neurol 249:955–960 Kaplan CP, Miner ME (2000) Relationships: importance for patients with cerebral tumours. Brain Inj 14:251–259 Klein M, Taphoorn MJB, Heimans JJ, Van der Ploeg HM, Vandertop WP, Smit EF, Leenstra S, Tulleken CAF, Boogerd W, Belderbos JSA, Cleijne W, Aaronson NK (2001) Neurobehavioral status and health-related quality of life in newly diagnosed high-grade glioma patients. J Clin Oncol 19:4037–4047 Locke DE, Decker PA, Sloan JA, Brown PD, Malec JF, Clark MM, Rummans TA, Ballman KV, Schaefer PL, Buckner JC (2007) Validation of single-item linear analog scale assessment of quality of life in neuro-oncology patients. J Pain Symptom Manage 34:628–638 Mackworth N, Fobair P, Prados MD (1992) Quality of life selfreports from 200 brain tumor patients: comparisons with Karnofsky performance scores. J Neurooncol 14:243–253 Moots PL (1998) Pitfalls in the management of patients with malignant gliomas. Semin Neurol 18:257–265 Muldoon MF, Barger SD, Flory JD, Manuck SB (1998) What are quality of life measurements measuring? Br Med J 316: 542–545 Osoba D, Aaronson NK, Muller M, Sneeuw K, Hsu MA, Yung WKA, Brada M, Newlands E (1996) The development and psychometric validation of a brain cancer quality-of-life questionnaire for use in combination with general cancerspecific questionnaires. Qual Life Res 5:139–150 Osoba D, Aaronson NK, Muller M, Sneeuw K, Hsu MA, Yung WKA, Brada M, Newlands E (1997) Effect of neurological dysfunction on health-related quality of life in patients with highgrade malignant glioma. J Neurooncol 34:263–278 Osoba D, Brada M, Yung WKA, Prados M (2000a) Healthrelated quality of life in patients treated with temozolamide versus procarbazine for recurrent glioblastoma multiforme. J Clin Oncol 18:1481–1491 Osoba D, Brada M, Yung WKA, Prados MD (2000b) Healthrelated quality of life in patients with anaplastic astrocy-

301 tomas during treatment with temozolomide. Eur J Cancer 36: 1788–1795 Ottawa Charter for Health Promotion (1986) WHO/HPR/ HEP/95.1. WHO, Geneva Pelletier G, Verhoef MJ, Khatri N, Hagen N (2002) Quality of life in brain tumor patients: the relative contributions of depression, fatigue, emotional distress, and existential issues. J Neurooncol 57:41–49 Orley J, Kuyken W (eds) (1994) Quality of life assessment: international perspectives. Proceedings of the joint meeting organized by the world health organization and the foundation IPSEN in Paris, 2–3 July 1993. Springer, Berlin Recht L, Glantz M, Chamberlain M, Hsieh CC (2003) Quantitative measurement of quality outcome in malignant glioma patients using an independent living score (ILS). Assessment of a retrospective cohort. J Neurooncol 61: 127–136 Reijneveld JC, Sitskoorn MM, Klein M, Nuyen J, Taphoorn MJB (2001) Cognitive status and quality of life in patients with suspected versus proven low-grade gliomas. Neurology 56:618–623 Taphoorn MJB, Schiphorst AK, Snoek FJ, Lindeboom J, Wolbers JG, Karim ABMF, Huijgens PC, Heimans JJ (1994) Cognitive functions and quality of life in patients with low-grade gliomas: the impact of radiotherapy. Ann Neurol 36:48–54 Ware JE, Sherbourne CD (1992) The MOS 36-item short-form health survey (SF-36): I conceptual framework and item selection. Med Care 30:473–483 Weitzner MA, Meyers CA, Byrne K (1996) Psychosocial functioning and quality of life in patients with primary brain tumors. J Neurosurg 84:29–34 Weitzner MA, Meyers CA, Gelke CK, Byrne KS, Cella DF, Levin VA (1995) The functional assessment of cancer therapy (FACT) scale. development of a brain subscale and revalidation of the general version (FACT-G) in patients with primary brain tumors. Cancer 75:1151–1161 WHO (1996) What quality of life? The WHOQOL Group. World health organization quality of life assessment. World Health Forum 17:354–356 (Geneva, World Health Organization)

Chapter 32

Health-Related Quality of Life in Patients with High Grade Gliomas Eefje M. Sizoo and Martin J.B. Taphoorn

Abstract The concept of health-related quality of life (HRQOL) was founded to evaluate the effect of a disease and its treatment on the patients’ subjective functioning and well-being. Because patients with HGG cannot be cured, HRQOL as an outcome measure in clinical trials evaluating new treatment modalities has become increasingly important. Furthermore, assessment of HRQOL in daily clinical practice improves physician-patient communication and could thereby in turn improve the patient’s quality of life. More focussed HRQOL questionnaires are needed for common use in daily practice. Keywords QoL · HRQOL · Cognition · Tumour · High-grade gliomas · FACT

Introduction Patients diagnosed with a high grade glioma (HGG) have a poor prognosis and cannot be cured from their disease. Despite intensive treatment with surgery, chemotherapy, radiotherapy, tumour recurrence inevitably occurs and patients will decease from tumour progression. In this patient category, the aim of treatment is not only to prolong life, but also to maintain quality of life as long as possible (Efficace and Bottomley, 2002). Combined radiochemotherapy and other new treatment strategies may not only

E.M. Sizoo () Department of Neurology, VU University Medical Center, Amsterdam, The Netherlands e-mail: [email protected]

increase the duration of survival, but may also have severe side-effects including a risk of toxicity (Stupp et al., 2005; Hart et al., 2008). Therefore, the benefits of extended survival and/or progression delay have to be carefully balanced against side-effects of treatments and their potential negative impact on functioning and quality of life. Hence, the concept of health-related quality of life (HRQOL) should be included as an outcome measure supplementing traditional end points such as (progression-free) survival time in clinical trials evaluating the effect of treatment. Measuring HRQOL emerged in the early nineties in the medical oncology literature. In brain tumour patients, however, it has long been a neglected issue (Efficace and Bottomley, 2002). Since the beginning of this century, HRQOL has become a secondary outcome measure in a growing number of clinical trials evaluating glioma treatment (Bottomley et al., 2005).

Outcome Measures in Glioma Research Next to the classic outcome measures such as progression free survival and overall survival, the effect of a brain tumour and its treatment on the patient’s functioning and well-being should be assessed. It is important to make a distinction between impairment, disability, and handicap (Grant et al., 1994). Impairments are the direct consequences of disease demonstrated by physical examination. Disability is the impact of the impairment on a patients’ ability to carry out activities. Finally, handicap is the consequence of disability on the patients’ well-being. Impairment is considered to be a “hard” measure compared to disability and handicap, which are more

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_32, © Springer Science+Business Media B.V. 2011

303

304

relevant for the patient’s functioning. Impairment in a brain tumour patient can be evaluated using neurological and neuropsychological examination. Disability can be determined by using scales, such as the Barthel index (BI), an instrument on a persons’ ability for selfcare, or the Karnofsky Performance Scale (KPS), an assessment tool to measure a patient’s ability to carry out activities of daily living. The Modified Ranking Handicap Scale (MRHS) is frequently used to measure handicap. This is a six-point scale ranging from 0 (no symptoms) to 5 (severe handicap/ totally dependent; requiring attention day and night). It should be noted that there are no specific disability or handicap scales for brain tumour patients, besides the Spitzer scale (Spitzer et al., 1981) which is hardly used. Although these outcome measures provide information on the influence of the tumour on a patient’s functioning in daily life, they do not fully reflect the effect of the tumour on the patient’s quality of life.

Assessing Health-Related Quality of Life: A Patient-Reported Outcome Measure To measure quality of life, the concept of HRQOL was developed. HRQOL is defined as a person’s selfassessed ability to function in the physical, psychological, emotional, and social domains of day-to-day life (Aaronson, 1988). This complex patient-reported outcome (PRO) measure demands a multidimensional instrument, and preferably should be assessed using a self-reported questionnaire. As an alternative, a (semi)structured interview could be undertaken with the patient. At present, no single gold standard tool exists to measure HRQOL. Both generic and disease specific tools have been developed and validated to assess HRQOL, both for cancer patients and in the non-cancer population. For cancer patients, the most common tool in use was developed by the European Organisation for Treatment and Research of Cancer (EORTC) quality of life group: the EORTC QLQ-C30 (Aaronson et al., 1993). This is a 30-item measure designed to assess HRQOL of cancer patients. Table 32.1 shows the construction of this measure. The EORTC BN20 was specifically developed and validated for patients with brain cancer (Osoba et al., 1996). It includes

E.M. Sizoo and M.J.B. Taphoorn

20 items and assesses visual disorder, motor dysfunction, various disease symptoms, treatment toxicity, and future uncertainty (Table 32.2). This tool should be used in combination with the EORTC QLQ C30 and is often used in clinical trials in glioma patients undergoing chemotherapy and radiation therapy. The items on both the EORTC QLQC30 as well as the EORTC BN20 measures are scaled, scored, and transformed to a linear scale (0–100). Differences of at least 10 points are classified as a clinically meaningful changes in a HRQOL parameter. Changes over 20 points are classed as large effects. Another widely used (brain) cancer specific HRQOL tool is the Functional Assessment of Cancer Therapy (FACT). Next to a general FACT module (FACT-G), a brain cancer specific module was developed (FACT-Br) by Weitzner et al. (1995), combining the FACT-G with a brain subscale. Table 32.3 shows the construction of this measure. Compared to the EORTC questionnaires, the FACT modules are more focussed on psychosocial aspects and less on symptoms. When patients are unable to self-report, for example due to cognitive disturbances, one might consider using proxies or health care professionals to rate the patient’s quality of life. In the past, this method was regarded far from optimal. However, a recent review found moderate to good agreement in various studies evaluating the concordance between patient and proxy measures (Sneeuw et al., 2002). Mixed results have been reported for patients and health care providers. Proxies and health care providers tend to report more HRQOL problems than do patients themselves, and proxy ratings tend to be more in agreement with the patients’ physical HRQOL domains compared to the psychological domains. Also, the agreement between patients, and proxy HRQOL reports was evaluated specifically in brain tumour patients. The EORTC QLQ-C30, EORTC-BN20, and the FACT-Br showed moderate agreement between the patients’ and proxies assessment of HRQOL, provided cognitive functioning was not severely affected (Brown et al., 2008; Sneeuw et al., 2002). The use of a nonpatient-based report should, therefore, only be used when patients are incapable of self-report. One may anticipate that patients with more severe clinical symptomatology and quality of life difficulties are less likely to complete questionnaires because it is too burdensome. Because these patients (noncompliers) will be excluded from the analysis, this may

32 Health-Related Quality of Life in Patients with High Grade Gliomas

305

Table 32.1 Content of the EORTC QLQ c30 version 3.0 Number of items

Range item scores

Item numbers

Scale scores

Global health status/QOL Global health status/QOL

2

1–7

29, 30

0–100

Functioning scales Physical Role Emotional Cognitive Social

5 2 4 2 2

1–4 1–4 1–4 1–4 1–4

1–5 6, 7 21–24 20, 25 26, 27

0–100 0–100 0–100 0–100 0–100

Symptom scales Fatigue Nausea/vomiting Pain

3 2 2

1–4 1–4 1–4

10, 12, 18 14, 15 9, 19

0–100 0–100 0–100

Single-item scales Dyspnoea Sleep disturbance Appetite loss Constipation Diarrhoea Financial impact

1 1 1 1 1 1

1–4 1–4 1–4 1–4 1–4 1–4

8 11 13 16 17 28

0–100 0–100 0–100 0–100 0–100 0–100

Table 32.2 Content of the EORTC BN20 Number of items

Range item scores

Item numbers

Scale scores

Subscales Future uncertainty Motor dysfunction Communication deficits Visual disorder

4 3 3 3

1–4 1–4 1–4 1–4

1–3, 5 10, 15, 19 11–13 6–8

0–100 0–100 0–100 0–100

Single-item scales Headaches Seizures Drowsiness Bothered by hair loss Bothered by itching skin Weakness of legs Difficulty controlling bladder

1 1 1 1 1 1 1

1–4 1–4 1–4 1–4 1–4 1–4 1–4

4 9 14 16 17 18 20

0–100 0–100 0–100 0–100 0–100 0–100 0–100

Table 32.3 Content of the FACT-Br version 4 Number of items

Range item scores

Item numbers

Range scale scores

FACT-G subscales: Physical well-being Social well-being Emotional well-being Functional well-being

7 7 6 7

0–4 0–4 0–4 0–4

GP1–GP7 GS1–GS7 GE1–GE6 GF1–GF7

0–28 0–28 0–24 0–28

19

0–4

Br1–Br18, NTX6

0–76

Brain subscale

306

lead to an overestimation of the actual quality of life (Brown et al., 2008). Indeed, the interpretation of serial measurements of HRQOL is affected by missing data (Walker et al., 2003). Apart from the selection bias due to the clinical condition, in both patients and observers compliance with filling out questionnaires decreases over time. The main cause of missing data, however, is administrative failure. Administrative failure arises, for example, when questionnaires are not distributed by the doctor or nurse, distributed at the wrong moment or handed out without instructions. Methodological and patient-related factors can also lead to missing data. Methodological problems may arise due to the study design, for example, using HRQOL instruments unknown to the clinicians who are supposed to hand these out. Other patient-related factors than the clinical situation encompass lack of motivation on the part of the patient, misunderstanding instructions, and/or filling out questionnaires incorrectly. Several approaches can be undertaken to minimize avoidable loss of data on the quality of life (Walker et al., 2003). Of the utmost importance is that research staff and patients understand the relevance of these data to be collected. While writing a research protocol, HRQOL assessment should be explicitly defined as a trial endpoint, the way of data collection should be specified, and the analysis of HRQOL parameters should be described in order to prevent methodological problems. Administrative problems can be challenged by the training staff in charge of data collection to check for completeness of assessments at submission, document reasons for missing data, and structurally contact patients who miss appointments. To reduce patient-related missing data, it is important to motivate patients. At trial entry, patients should be fully informed regarding the importance of HRQOL assessments, how they will be done, and when they will be done. Multiple questionnaires addressing similar issues in a different format and/or a high frequency of assessments will result in a low overall compliance.

Health-Related Quality of Life in Patients with High Grade Glioma As one would expect, the majority of newly diagnosed HGG patients have a significant impaired level of HRQOL compared to healthy controls (Brown et al.,

E.M. Sizoo and M.J.B. Taphoorn

2006; Klein et al., 2001). In patients with a reduced level of HRQOL at the time of diagnosis, the quality of life will further decrease over time, while in patients not significantly distressed, the HRQOL scores may improve (Brown et al., 2005). In comparison to other neurological diseases of the central and peripheral nervous system, patients with HGG experience the same level of HRQOL (Giovagnoli, 1999). When comparing HGG patients to other cancer patients, such as lung cancer patients, again similar quality of life results were found in both patient groups (Klein et al., 2001). Several tumour-related factors in HGG patients can have an impact on perceived quality of life. Patients with HGGs experience worse quality of life than patients who have a low grade glioma (Salo et al., 2002). However, between patients diagnosed with glioblastoma multiforme (grade IV) and patients diagnosed with anaplastic astrocytoma (grade III), no differences in HRQOL scores exist at the time of diagnosis (Brown et al., 2006). Next to the grade, the size of the tumour and the location in the brain correlate with HRQOL. Large tumours, tumours in the nondominant hemisphere, and tumours located anteriorly in the brain are associated with poorer HRQOL scores (Salo et al., 2002). Disease-specific signs and symptoms have a major impact on quality of life. Neurological signs and symptoms as seizure frequency (Klein et al., 2003), motor deficits (Osoba et al., 1997) and functional status (Giovagnoli et al., 2005) have proven to diminish HRQOL. Surprisingly, no deleterious effect of dysphasia on HRQOL has been established (Osoba et al., 1997). As to nonspecific signs and symptoms in patients with systemic cancers, fatigue and depression are identified as the leading factors diminishing HRQOL (Gupta et al., 2007). Also, in high grade glioma patients, fatigue is one of the most common symptoms and, therefore, one of the leading symptoms of decreasing quality of life (Pelletier et al., 2002). Clinically significant symptoms of depression have shown to be present in a significant portion of HGG patients, and are probably higher than the prevalence in the general cancer population (Pelletier et al., 2002). Thus, depressive symptoms are a serious clinical issue negatively affecting HRQOL in these patients (Pelletier et al., 2002). Disease recurrence has a significantly deleterious impact on a patients’ life. Patients carry a significant symptom burden and neurological

32 Health-Related Quality of Life in Patients with High Grade Gliomas

deficits are more severe at the time of recurrence compared to the initial presentation (Giovagnoli et al., 2005). Not surprisingly, HRQOL of patients with tumour recurrence is more comprised compared to patients without recurrence at the same time from diagnosis (Bosma et al., 2009; Osoba et al., 2000a).

Cognitive Functioning Versus Health-Related Quality of Life Cognition encompasses functions such as language, memory, attention, and executive functioning. Disturbances in cognition are common in patients with brain tumours. These can be caused by the tumour itself, by tumour-related epilepsy, but also by tumour treatment (surgery, radiotherapy, chemotherapy) as well as supportive treatment (antiepileptic drugs, corticosteroids) (Taphoorn and Klein, 2004). Cognitive disturbances can cause burdensome symptoms for patients, and it is assumed that impaired cognitive function reduces quality of life. For example, cognitive deficits are associated with increased fatigue in HGG patients. Fatigue, in turn, has proven to diminish HRQOL. The direct relation between cognitive functioning and HRQOL in HGG patients has only been demonstrated in one study so far. Patients with cognitive disturbances reported lower HRQOL than did patients without cognitive deficits (Giovagnoli et al., 2005).

Effect of (Tumour) Treatment on Health Related Quality of Life Effect of Surgery on Quality of Life Reduction of tumour mass may alleviate neurological symptoms and cognitive deficits; thereby, improving quality of life. On the other hand, surgery and perioperative injuries may cause neurological deficits and focal cognitive deficits as a result of damage to normal surrounding tissue (Taphoorn and Klein, 2004). Although these deficits are often transient, they may result in a temporarily lower perceived quality of life.

307

In a nonrandomized study, patients who had undergone a gross-total resection had both a longer survival and a better HRQOL than patients who only had a biopsy (Brown et al., 2005). Clearly, these results have been biased because the selection of patients for resection versus biopsy depends on factors as tumour size, tumour location, multi-focality, and performance status. Finally, the HRQOL in patients who had undergone a gross-total resection increased over time. Therefore, it appeared from this study that the benefit of resection in terms of quality of life outweigh the early side-effects of surgery.

Effect of Radiotherapy on Quality of Life The benefit of radiotherapy is well-established in the treatment of HGG patients, because tumour progression is postponed and overall survival extended. By stabilizing disease and delay progression, quality of life can be maintained for a longer period than without radiation. Side-effects of cranial radiotherapy, however, of which cognitive deterioration is most feared, may negatively affect HRQOL. Radiation sideeffects in the brain can be divided in acute radiation encephalopathy, early-delayed radiation encephalopathy and late-delayed encephalopathy. Acute and earlydelayed radiation encephalopathy, occurring during or shortly following radiotherapy, may result in drowsiness and fatigue. Because these side-effects are nearly always completely reversible, they may only temporarily affect HRQOL. By contrast, late-delayed radiation encephalopathy which occurs months to years after radiotherapy, may result in progressive cognitive decline (Taphoorn and Klein, 2004). Two randomized studies evaluating the combination of chemotherapy and radiation versus radiation therapy alone included HRQOL as an outcome measure (Taphoorn et al., 2005, 2007). No negative effects of radiotherapy on quality of life were observed in anaplastic oligodendroglioma patients and patients with glioblastoma multiforme with a good performance status. On longer follow-up, >1.5 year after completion of radiotherapy, HRQOL scores of HGG patients without progression even improved compared to scores at the start of the treatment. In long term (i.e., >2 years from initial treatment) HGG survivors without disease progression who all had initial radiotherapy, even HRQOL scores were

308

observed meeting the level of healthy controls, which may partly be explained by response shift, i.e., that patients over time more readily accept their situation (Bosma et al., 2009). Specifically in the elderly population (age >70 years), a moderate survival benefit from radiotherapy has been established for patients who had a good performance status at the start of the treatment. More importantly, HRQOL, performance status and cognitive functions did not further deteriorate compared to the observation arm of this study, in which patients only received supportive care (Keime-Guibert et al., 2007). Reirradiation in HGG patients is increasingly applied because patients live longer following their initial treatment. Reirradiation should be considered in patients with an adequate performance status (KPS ≥70) applying focal radiation treatment after an interval from initial treatment of at least 6 months (Nieder et al., 2006). The effect of reirradiation, specifically on HRQOL, was only evaluated in one small study (Ernst-Stecken et al., 2007) with a median follow-up of 9 months. The majority of patients (80%) judged their general health status after reirradiation to be stable or even improved compared to before treatment; in 20% of patients, their perceived general health status declined. Scores for physical functioning, cognitive functioning and fatigue remained stable in nearly all patients.

Effect of Chemotherapy on Quality of Life In 2005, a large randomized controlled EORTC trial showed that the combination of temozolomide chemotherapy and radiotherapy significantly prolonged survival in patients with newly diagnosed glioblastoma multiforme compared to patients treated with radiotherapy alone (Stupp et al., 2005). The effect of this new treatment modality on HRQOL was evaluated separately (Taphoorn et al., 2005). During treatment and follow-up, in both treatment groups changes over time in 7 preselected HRQOL domains were not substantial during the first year of follow-up, provided there was no progression of disease. For several scales, scores even improved over time. However, during treatment, the patients in the combination treatment group reported more side effects (nausea, vomiting, appetite loss and constipation) compared to the

E.M. Sizoo and M.J.B. Taphoorn

radiotherapy only group, which can be attributed to the use of temozolomide and antiemetics. Furthermore, during adjuvant temozolomide treatment, social functioning was worse in the intensive treatment group. Overall, it can be concluded that the addition of temozolomide during and after radiotherapy significantly improved survival without a long-lasting negative effect on quality of life. As for treatment of patients with anaplastic oligodendroglioma, adjuvant treatment with combined procarbazine, CCNU (lomustine), and vincristine (PCV) chemotherapy after radiotherapy significantly prolongs progression-free survival, but not overall survival (van den Bent et al., 2006). With respect to HRQOL, patients receiving PCV chemotherapy show a significant increase in nausea/vomiting and appetite loss during and shortly following treatment compared to patients only receiving radiotherapy. Furthermore, patients on PCV chemotherapy report more drowsiness. These differences, however, resolve over time: after 1 year follow up, no longer differences were observed in HRQOL between treatment groups (Taphoorn et al., 2007). Overall, there is a short-lasting negative impact of PCV chemotherapy on HRQOL during and shortly after treatment, but no long term effects or HRQOL have been established. More importantly, because PCV chemotherapy postpones tumour progression, the impact of progression on well-being and HRQOL should be evaluated in future studies. In recurrent glioma, the median survival is short and treatment so far is only modestly effective. Because HRQOL measurements encompass assessment of both functioning ability and toxicity from therapy, HRQOL outcomes are of equal importance as survival in this patient group (Osoba et al., 2000a, b). Patients with recurrent anaplastic astrocytoma or glioblastoma multiforme succesfully treated with temozolomide achieve a statistically significant improvement in a portion of the HRQOL domains while patients with disease progression reported statistically significant deterioration in most HRQOL domains (Osoba et al., 2000a, b). Thus, there is HRQOL benefit from temozolomide treatment for the period of stable disease due to treatment before disease progression occurs. The effect of temozolomide on HRQOL in recurrent glioblastoma has been compared with the effect of procarbazine in a randomized study. Patients receiving procarbazine showed deterioration in most HRQOL domains during

32 Health-Related Quality of Life in Patients with High Grade Gliomas

treatment, whereas patients treated with temozolomide improved while on treatment (Osoba et al., 2000a). Although temozolomide chemotherapy has largely replaced PCV chemotherapy in glioma patients due to fewer side effects and improved tolerability, HRQOL data on chemotherapy in elderly HGG patients with poor performance status, as well as in the recurrent setting are scarce (Hart et al., 2008).

Effect of Supportive Treatment on Health Related Quality of Life Symptomatic medications prescribed for glioma patients often include anti-epileptic drugs (AED) and steroids (dexamethasone). Because the occurrence of seizures can diminish HRQOL, it can be assumed that treatment with AEDs would improve quality of life. Conversely, an adverse effect of AED on cognition has been demonstrated (Klein et al., 2002). This, in turn, can have a negative effect on the quality of life. A study examining the impact of seizures and AED on cognition and quality of life showed both cognitive functions as well as HRQOL to deteriorate. The cognitive deficits could primarily be ascribed to the use of antiepileptic drugs, whereas the low HRQOL scores were mainly related to poor seizure control (Klein et al., 2003). Dexamethasone reduces peritumoral edema and is prescribed to alleviate neurological symptoms, thereby improving quality of life. On the other hand, common side-effects are myopathy, gastro-intestinal complications, hyperglykemia, and psychiatric complications (mainly agitation or depression). Because these sideeffects are related to the prescribed dosage, steroids should be tapered or maintained at the lowest effective dose (Kaal and Vecht, 2004).

Value of Health Related Quality of Life in Daily Practice Old age and low functional status (Karnofsky Performance Status <70) have proven to be poor prognostic factors for survival in patients with HGGs. In daily practice, these prognostic factors are used to select patients who will probably benefit from

309

aggressive treatment and patients who will probably not. HRQOL parameters have shown to be independent prognostic factors in various types of cancers (Mauer et al., 2007a). At present, the prognostic value of baseline HRQOL data in predicting survival of HGG patients is questionable. Hitherho, four relatively large studies have been published about this subject. Two analyses using FACT scores for prognosis were performed. The first analysis has demonstrated patients with high scores on the FACT-G to have an enhanced survival compared to patients with low scores (Sehlen et al., 2003). The second one, using the FACT-Br in combination with a five-item linear analogue scale assessment (LASA) also found a relation between high HRQOL scores and improved survival in univariate analysis. However, HRQOL was closely related to functional status and after correction for this in a multivariate analysis, no prognostic significance of HRQOL scores remained (Brown et al., 2006). Two EORTC brain tumour studies regarding this issue were analyzed by Mauer et al. (2007a). Classical analysis of EORTC-QLQ C30 subscores, controlled for major prognostic factors as age and performance status, identified cognitive functioning, global health status, and social functioning as statistically significant prognostic factors for survival in glioblastoma patients. In patients with anaplastic oligodendrogliomas, emotional functioning, communication deficit, future uncertainty, and weakness of legs were found to be significant prognostic factors (Mauer et al., 2007b). In a more sophisticated boot-strap analysis, HRQOL scales were added to other predictive factors in a prognostic model. It came out that the HRQOL scales did not improve the prognostic value of known clinical factors. More importantly, fewer parameters are required in the prognostic model using clinical factors compared to the model using HRQOL data. From these analyses it can be concluded that, although various HRQOL scales have prognostic value, they have no additional value over already known clinical factors. However, HRQOL data may have value in daily clinical practice. Routine HRQOL measurements in oncology patients visiting the outpatient department, with information provided to physicians, have shown to have a positive effect on physician-patient communication. In some patients, these measurements improved HRQOL and emotional functioning. However, measurement of HRQOL, symptoms, and

310

functioning are still far from being implemented in daily practice. In the future a core set of standard and disease specific questions repeated at key points of the disease trajectory (beginning of treatment, midtreatment, during follow up, at relapse) should be implemented to allow comparison over time. A small set of focussed HRQOL questions could be used at each visit (for example, during treatment the focus could be on side effects). Furthermore, clear interpretation of scores is important and decision guidelines should be provided to the clinicians (Velikova et al., 2008).

References Aaronson NK (1988) Quality of life: what is it? How should it be measured? Oncology 2:69–76 Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, Filiberti A, Flechtner H, Fleishman SB, de Haes JC, Kaasa S, Klee M, Osoba D, Razavi D, Rofe PB, Schraub S, Sneeuw K, Sullivan M, Takeda F (1993) For the European Organization for Research and Treatment of Cancer Study Group on Quality of Life. The European Organization for Research and Treatment of Cancer QLQ-C30: a qualityof-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 85:365–376 Bosma I, Reijneveld JC, Douw L, Vos MJ, Postma TJ, Aaronson NK, Muller M, Vandertop WP, Slotman BJ, Taphoorn MJ, Heimans JJ, Klein M (2009) Health-related quality of life of long-term high-grade glioma survivors. Neuro-oncology 11:51–58 Bottomley A, Flechtner H, Efficace F, Vanvoorden V, Coens C, Therasse P, Velikova G, Blazeby J, Greimel E (2005) Health related quality of life outcomes in cancer clinical trials. Eur J Cancer 41:1697–1709 Brown PD, Ballman KV, Rummans TA, Maurer MJ, Sloan JA, Boeve BF, Gupta L, Tang-Wai DF, Arusell RM, Clark MM, Buckner JC (2006) Prospective study of quality of life in adults with newly diagnosed high-grade gliomas. J Neurooncol 76:283–291 Brown PD, Decker PA, Rummans TA, Clark MM, Frost MH, Ballman KV, Arusell RM, Buckner JC (2008) A prospective study of quality of life in adults with newly diagnosed highgrade gliomas: comparison of patient and caregiver ratings of quality of life. Am J Clin Oncol 31:163–168 Brown PD, Maurer MJ, Rummans TA, Pollock BE, Ballman KV, Sloan JA, Boeve BF, Arusell RM, Clark MM, Buckner JC (2005) A prospective study of quality of life in adults with newly diagnosed high-grade gliomas: the impact of the extent of resection on quality of life and survival. Neurosurgery 57:495–504 Efficace F, Bottomley A (2002) Health related quality of life assessment methodology and reported outcomes in randomised controlled trials of primary brain cancer patients. Eur J Cancer 38:1824–1831

E.M. Sizoo and M.J.B. Taphoorn Ernst-Stecken A, Ganslandt O, Lambrecht U, Sauer R, Grabenbauer G (2007) Survival and quality of life after hypofractionated stereotactic radiotherapy for recurrent malignant glioma. J Neurooncol 81:287–294 Giovagnoli AR (1999) Quality of life in patients with stable disease after surgery, radiotherapy, and chemotherapy for malignant brain tumour. J Neurol Neurosurg Psychiatry 67:358–363 Giovagnoli AR, Silvani A, Colombo E, Boiardi A (2005) Facets and determinants of quality of life in patients with recurrent high grade glioma. J Neurol Neurosurg Psychiatry 76: 562–568 Grant R, Slattery J, Gregor A, Whittle IR (1994) Recording neurological impairment in clinical trials of glioma. J Neurooncol 19:37–49 Gupta D, Lis CG, Grutsch JF (2007) The relationship between cancer-related fatigue and patient satisfaction with quality of life in cancer. J Pain Symptom Manage 34:40–47 Hart MG, Grant R, Garside R, Rogers G, Somerville M, Stein K (2008) Temozolomide for high grade glioma. Cochrane Database Syst Rev 8(4):CD007415 Kaal EC, Vecht CJ (2004) The management of brain edema in brain tumors. Curr Opin Oncol 16:593–600 Keime-Guibert F, Chinot O, Taillandier L, Cartalat-Carel S, Frenay M, Kantor G, Guillamo JS, Jadaud E, Colin P, Bondiau PY, Menei P, Loiseau H, Bernier V, Honnorat J, Barrie M, Mokhtari K, Mazeron JJ, Bissery A, Delattre JY (2007) Radiotherapy for glioblastoma in the elderly. N Engl J Med 356:1527–1535 Klein M, Engelberts NH, van der Ploeg HM, KasteleijnNolst Trenite DG, Aaronson NK, Taphoorn MJ, Baaijen H, Vandertop WP, Muller M, Postma TJ, Heimans JJ (2003) Epilepsy in low-grade gliomas: the impact on cognitive function and quality of life. Ann Neurol 54: 514–520 Klein M, Heimans JJ, Aaronson NK, van der Ploeg HM, Grit J, Muller M, Postma TJ, Mooij JJ, Boerman RH, Beute GN, Ossenkoppele GJ, van Imhoff GW, Dekker AW, Jolles J, Slotman BJ, Struikmans H, Taphoorn MJ (2002) Effect of radiotherapy and other treatment-related factors on midterm to long-term cognitive sequelae in low-grade gliomas: a comparative study. Lancet 360:1361–1368 Klein M, Taphoorn MJ, Heimans JJ, van der Ploeg HM, Vandertop WP, Smit EF, Leenstra S, Tulleken CA, Boogerd W, Belderbos JS, Cleijne W, Aaronson NK (2001) Neurobehavioral status and health-related quality of life in newly diagnosed high-grade glioma patients. J Clin Oncol 19:4037–4047 Mauer M, Stupp R, Taphoorn MJ, Coens C, Osoba D, Marosi C, Wong R, de Witte O, Cairncross JG, Efficace F, Mirimanoff RO, Forsyth P, van den Bent MJ, Weller M, Bottomley A (2007a) The prognostic value of health-related quality-of-life data in predicting survival in glioblastoma cancer patients: results from an international randomised phase III EORTC brain tumour and radiation oncology groups, and NCIC clinical trials group study. Br J Cancer 97:302–307 Mauer ME, Taphoorn MJ, Bottomley A, Coens C, Efficace F, Sanson M, Brandes AA, van der Rijt CC, Bernsen HJ, Frenay M, Tijssen CC, Lacombe D, van den Bent MJ (2007b) Prognostic value of health-related quality-oflife data in predicting survival in patients with anaplastic

32 Health-Related Quality of Life in Patients with High Grade Gliomas oligodendrogliomas, from a phase III EORTC brain cancer group study. J Clin Oncol 25:5731–5737 Nieder C, Adam M, Molls M, Grosu AL (2006) Therapeutic options for recurrent high-grade glioma in adult patients: recent advances. Crit Rev Oncol Hematol 60:181–193 Osoba D, Aaronson NK, Muller M, Sneeuw K, Hsu MA, Yung WK, Brada M, Newlands E (1996) The development and psychometric validation of a brain cancer quality-of-life questionnaire for use in combination with general cancerspecific questionnaires. Qual Life Res 5:139–150 Osoba D, Aaronson NK, Muller M, Sneeuw K, Hsu MA, Yung WK, Brada M, Newlands E (1997) Effect of neurological dysfunction on health-related quality of life in patients with high-grade glioma. J Neurooncol 34:263–278 Osoba D, Brada M, Yung WK, Prados M (2000a) Health-related quality of life in patients treated with temozolomide versus procarbazine for recurrent glioblastoma multiforme. J Clin Oncol 18:1481–1491 Osoba D, Brada M, Yung WK, Prados MD (2000b) Healthrelated quality of life in patients with anaplastic astrocytoma during treatment with temozolomide. Eur J Cancer 36: 1788–1795 Pelletier G, Verhoef MJ, Khatri N, Hagen N (2002) Quality of life in brain tumor patients: the relative contributions of depression, fatigue, emotional distress, and existential issues. J Neurooncol 57:41–49 Salo J, Niemela A, Joukamaa M, Koivukangas J (2002) Effect of brain tumour laterality on patients’ perceived quality of life. J Neurol Neurosurg Psychiatry 72:373–377 Sehlen S, Lenk M, Hollenhorst H, Schymura B, Aydemir U, Herschbach P, Duhmke E (2003) Quality of life (QoL) as predictive mediator variable for survival in patients with intracerebral neoplasma during radiotherapy. Onkologie 26:38–43 Sneeuw KC, Sprangers MA, Aaronson NK (2002) The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease. J Clin Epidemiol 55:1130–1143 Spitzer WO, Dobson AJ, Hall J, Chesterman E, Levi J, Shepherd R, Battista RN, Catchlove BR (1981) Measuring the quality of life of cancer patients: a concise QL-index for use by physicians. J Chronic Dis 34:585–597 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C,

311

Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352: 987–996 Taphoorn MJ, Klein M (2004) Cognitive deficits in adult patients with brain tumours. Lancet Neurol 3:159–168 Taphoorn MJ, Stupp R, Coens C, Osoba D, Kortmann R, van den Bent MJ, Mason W, Mirimanoff RO, Baumert BG, Eisenhauer E, Forsyth P, Bottomley A (2005) Health-related quality of life in patients with glioblastoma: a randomised controlled trial. Lancet Oncol 6:937–944 Taphoorn MJ, van den Bent MJ, Mauer ME, Coens C, Delattre JY, Brandes AA, Sillevis Smitt PA, Bernsen HJ, Frenay M, Tijssen CC, Lacombe D, Allgeier A, Bottomley A (2007) Health-related quality of life in patients treated for anaplastic oligodendroglioma with adjuvant chemotherapy: results of a European Organisation for Research and Treatment of Cancer randomized clinical trial. J Clin Oncol 25: 5723–5730 van den Bent MJ, Carpentier AF, Brandes AA, Sanson M, Taphoorn MJ, Bernsen HJ, Frenay M, Tijssen CC, Grisold W, Sipos L, Haaxma-Reiche H, Kros JM, van Kouwenhoven MC, Vecht CJ, Allgeier A, Lacombe D, Gorlia T (2006) Adjuvant procarbazine, lomustine, and vincristine improves progression-free survival but not overall survival in newly diagnosed anaplastic oligodendrogliomas and oligoastrocytomas: a randomized uropean Organisation for Research and Treatment of Cancer phase III trial. J Clin Oncol 24: 2715–2722 Velikova G, Awad N, Coles-Gale R, Wright EP, Brown JM, Selby PJ (2008) The clinical value of quality of life assessment in oncology practice-a qualitative study of patient and physician views. Psychooncology 17:690–698 Walker M, Brown J, Brown K, Gregor A, Whittle IR, Grant R (2003) Practical problems with the collection and interpretation of serial quality of life assessments in patients with malignant glioma. J Neurooncol 63:179–186 Weitzner MA, Meyers CA, Gelke CK, Byrne KS, Cella DF, Levin VA (1995) The Functional Assessment of Cancer Therapy (FACT) scale. Development of a brain subscale and revalidation of the general version (FACT-G) in patients with primary brain tumors. Cancer 75:1151–1161

Chapter 33

Epilepsy and Brain Tumours and Antiepileptic Drugs Sophie Dupont

Abstract Seizures are a common occurrence in patients with brain tumours and can contribute to undesirable side effects that can greatly impact quality of life. In the vast majority of cases, seizures are the presenting clinical sign of the tumour, however, late seizures may occur in the evolution of the disease. The tumours type and their locations are determining factors that significantly influence seizure frequency: developmental tumours, low-grade gliomas and multiple metastases are at higher risk of seizures. According to guidelines, antiepileptic treatment should be started only in patients who have already experienced seizures. Major difficulties in patients with epilepsy and brain tumours include refractory seizures, potential interactions between anticonvulsants and chemotherapeutic agents and enhanced risks of toxicity, including cognitive deterioration. For seizure control, levetiracetam, valproic acid, and lamotrigine can each be considered as agents of first choice. First-line prescription of enzyme-inducing antiepileptic drugs, especially phenytoin, is not mandatory. Managing seizures in brain tumour patients may be challenging and requires comprehensive and multidisciplinary approach. Keywords Epilepsy · Antiepileptic drug · Seizures · Meta-analysis · Enzyme

Introduction Epilepsy is a frequent manifestation of brain tumours that has important clinical, social and therapeutic implications. Seizures can contribute to many undesirable life-altering consequences that can greatly impact quality of life. Unfortunately, although patients with brain tumours are at high risk to develop subsequent seizures, there is no available evidence suggesting that prophylactic administration of antiepileptic drugs (AEDs) provides substantial benefit to prevent further epilepsy. In patients who have proven epileptic seizures, the choice of AEDs should be based as in other symptomatic localisation related epilepsies on efficacy and tolerability criteria but also on pharmacokinetic criteria because of potential interactions between AEDs and dedicated brain tumour treatments (chemotherapy, dexamethasone, . . .). Unfortunately, seizures related to brain tumours are frequently refractory to drugs due to various explanations: intrinsic epileptogenesis of tumoral and peritumoral tisssues, multi-drug resistance proteins, potential genetic factors, interactions between chemotherapeutic agents and AEDs . . . Add-on antiepileptic therapy may be thus required but we must keep in mind that inappropriate escalation in dosages or number of AEDs could lead to harmful side effects. Management of epilepsy in patients with brain tumours is a major issue requiring a comprehensive and multidisciplinary approach.

S. Dupont () Epilepsy Unit, Clinique Neurologique Paul Castaigne, Hopital de la Salpetriere, 75651 Paris cedex 13, France e-mail: [email protected] M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_33, © Springer Science+Business Media B.V. 2011

313

314

S. Dupont

Epidemiology of Seizures in Brain Tumours The frequency of seizures in people with brain tumours is high, oscillating between 30 and 50%. Most of patients present with seizures in the early stage of the malignant disease (30–86%) (Hildebrand et al., 2005; Van Breemen et al., 2007) whereas lateonset seizures are more uncommon (∼15%). Globally, 6–45% of patients who do not initially present with seizures will eventually develop them in the course of the disease (Lee et al., 2010). Overall, occurrence of seizures mainly depends on the histological type of the brain tumour (see below). The occurrence of seizures at diagnosis has been proven to be a strong predictor of subsequent seizures and refractoriness to AEDs (Moots et al., 1995). Variable frequency of seizures (early or late-onset seizures) may be explained by several factors in patients with brain tumours: (1) Mainly the histological type of tumours (Hildebrand et al., 2005; Hwang et al., 2001): a higher incidence of seizures occurs in patients with low-grade astrocytomas (about 75%) than in those with high-grade gliomas (about 40%),

anaplastic astrocytomas (20–40%), meningiomas (40%) or primary CNS lymphoma (about 20%). For patients with brain metastasis, incidence of seizures ranges from 20 to 35%. Dysembryoplasic neuroepithelial tumours (DNET) that are rare and presumed benign developmental glioneuronal tumours are consistently associated to epilepsy (Fig. 33.1). (2) The location of the tumour: location in the temporal or frontal cortical regions (especially the rolandic region) are more frequently associated to seizures and logically, infratentorial and sellar tumours rarely cause seizures. A recent study (Lee et al., 2010) has shown that the propensity of tumours to cause seizures according to location varied with the grade of the tumour. In low-grade gliomas, tumours located in the temporal lobe or the insular region were more likely to present with seizures, in high-grade gliomas, deepseated pericallosal tumours (so-called butterfly gliomas) were more likely to present with nonseizure neurological symptoms. (3) The size of the tumour: the influence of the tumour size on the frequency of seizures has been poorly studied. A recent study (Lee et al., 2010) suggests that, as location, it is influenced by the

100% 90% 80% 70% 60% 50% 40% 30% 20% 10%

Fig. 33.1 Estimated frequency of seizures according to the tumour type

primary CNS lymphoma

brain metastasis

anaplastic astrocytoma

meningioma

high-grade glioma

low-grade glioma

DNET

0%

33 Epilepsy and Brain Tumours and Antiepileptic Drugs

Tumour grade

315

Seizure frequency

Tumour size

Tumour location

Fig. 33.2 Factors influencing the seizure frequency

grade of the tumour. In this study, patients with larger low-grade gliomas and smaller high-grade gliomas were more likely to present with seizures (Fig. 33.2).

Clinical Manifestations Epileptic patients with brain tumours mainly exhibit focal seizures (simple partial seizures, complex partial seizures). Secondarily generalization is common and may occur so quickly that, in certain patients, the focal phase passes unnoticed. A recent study (Hildebrand et al., 2005) has shown that the presentation of epilepsy could vary throughout the course of the malignant disease. At the onset of the disease, secondarily generalized tonic-clonic seizures were observed in one-half of the patients, the other patients exhibited simple or complex partial seizures (40%) or not determined seizures (10%). With the course of the disease, a significant reduction in seizure generalization (attributed to medication) was observed (19% versus 50% at the

beginning) and patients with persisting seizures had mostly focal manifestations. Specific studies dedicated to the correlation between the location of the tumour and the presentation of focal symptoms are lacking. Usually, a clear anatomico-clinical correlation is found between the nature of epileptic symptoms and the location of the brain tumour. But in some cases, the epileptogenic focus does not correspond to tumour location, especially in temporal lobe tumours. Secondary epileptogenesis causing seizure activity (and subsequent clinical manifestations) in regions that are distant to the site of the tumour could explain this lack of anatomico-clinical correlations.

Severity of Epilepsy and Epileptogenesis of Brain Tumours A substantial number of patients with brain tumours continue to experience epileptic seizures despite treatment and the administration of AEDs. The exact percentage of patients who continue to experience

316

S. Dupont

Fig. 33.3 Factors related to drug resistance Tumour grade Genetic factors

Tumour factors

Refractory Seizures

Host factors

Lower efficacy of AEDs Peri-tumoural tissue

seizures throughout the disease varies among studies from 32 (Moots et al., 1995) to 89% (Hildebrand et al., 2005). Several factors may explain the difficulty to achieve seizure freedom in epileptic patients, involving tumour and host factors.

Tumour and Peritumoural Factors The underlying pathophysiology of tumour-related seizures is poorly understood (Riva, 2005) but recent findings suggest that both the tumoral and peritumoral microenvironments could exhibit epileptogenic properties (Shamji et al., 2009). The mechanisms that promote tumour-related seizure activity are multifactorial depending on the specific tumour histology and the characteristics of the peritumoural environment (Fig. 33.3). Specific Tumour Histology Factors Developmental tumours such as DNET or gangliomas are frequently associated to refractory epilepsy. One hypothesis (Van Breemen et al., 2007) is that these tumours consist of well-differentiated cells that can release neurotransmitters that will promote seizures.

Low-grade gliomas that grow slowly and invade normal surrounding brain tissue have one of the highest frequency of epilepsy. Experimental studies suggest that the lesion itself could be the source of seizure activity due to abnormal functional connectivity and impaired mechanisms of activity modulation interspersing a normally functioning neuronal tissue with altered glial tissue (Shamji et al., 2009). High-grade lesions present much less frequently with seizures. It is actually unclear whether this lower observed incidence is a consequence of shorter overall survival or less epileptogenic pathology. In highgrade gliomas, seizure activity is attributed to abrupt tissue damage such as necrosis or hemosiderin deposition (Shamji et al., 2009). Characteristics of the Peritumoural Environment Epileptogenic properties linked to peritumour factors involve various mechanisms: edema, vascular insufficiency, intercellular communication, enzymatic changes, inflammation, and release of metabolically active molecules that may also promote seizure activity. Experimental data also suggest that pathological disruption of blood brain barrier among brain tumour patients could contribute to observed seizure activity (Shamji et al., 2009).

33 Epilepsy and Brain Tumours and Antiepileptic Drugs

Host Factors Genetic Factors Epileptic patients and patients with brain tumours could share common gene mutations. For example, LGI1 (leucine-rich, glioma inactivated 1) gene mutations were evidenced in both tumorigenesis (formation and malignant progression of glial tumors) and epileptogenesis. The LGI1 gene was first cloned from a glioblastoma cell line, where it was interrupted by a t(10;19)(q24;q13) translocation event and was suggested to function as tumour suppressor for its ability to reduce malignant features of glioblastoma cells (Gabellini and Masola, 2009). Additional evidence for a possible role of LGI1 in tumorigenesis came from the observation that its expression was reduced in low-grade brain tumors, and significantly reduced or absent in advanced gliomas. Ottman et al. (2004) demonstrated that mutations in LGI1 also caused autosomal dominant partial epilepsy with auditory features (ADPEAF), a form of familial temporal lobe epilepsy with auditory ictal manifestations. Nevertheless, no link has been evidenced between ADPEAF and subsequent risk for malignancies (Brodtkorb et al., 2003).

317

(3) Role of the multidrug resistance proteins: expression of multidrug resistance proteins in tumour cells of patients with glioma is high, suggesting that they can decrease drug transport, i.e. chemotherapy, into the brain parenchyma and explain the poor effectiveness of chemotherapy in gliomas (Calatozzolo et al., 2005). Similarly, a recent pharmacogenomic study identified a genetic factor that consisted in a polymorphism in the drug-transporter gene ABCB1 associated with resistance to antiepileptic drugs (Siddiqui et al., 2003). In the future, the possible association of inhibitors of drug efflux pumps with AEDs and/or chemotherapy could be investigated to improve drugs delivery into the tumour.

Therapeutic Issues Prophylactic Treatment Because the number of newly diagnosed brain tumour patients is increasing yearly, the question of whether to administer prophylactic anticonvulsants is being posed to neurologists with increasing frequency by their patients and by colleagues especially in neurosurgery, oncology, and radiation oncology.

Lower Efficacy of AEDs General Guidelines The relative refractoriness to AEDs could be related to several additional factors: (1) Inadequacy between classical AEDs mechanisms of action and the presumed pathophysiology of tumour-related seizures: AEDs mainly act either on excitatory mechanisms by blocking and inactivating Na+ and Ca+ channels and glutamate receptors or enhance inhibition through an increase of GABAergic activity. However, they don’t act on mechanisms such as morphologic changes, altered receptor and connexion patterns, changes of cytokine expression . . . that are involved in tumoural epilepsy (Riva, 2005). (2) Pharmacodynamic and pharmacokinetic interactions between AEDs and concomittent medications, especially chemotherapy, leading to non optimal plasma levels of AEDs (see chapter below on AEDs).

A recent meta-analysis study (Glantz et al., 2000) reported the results of twelve studies with adequate methodology providing data on the frequency of first seizures in patients with brain tumours with or without prophylactic anticonvulsant. Of these, only four were randomized clinical trials that provided level I evidence. All these studies were performed with conventional antiepileptic drugs (valproate acid, phenytoin, phenobarbital). None of the studies revealed a beneficial effect on the frequency of first seizures in patients receiving anticonvulsant prophylaxis. In the setting of this meta-analysis, the following guidelines were proposed by the Quality Standards Subcommittee of the American Academy of Neurology: 1. In patients with newly diagnosed brain tumours, anticonvulsant medications are not effective in preventing first seizures. Because of their lack of

318

efficacy and their potential side effects, prophylactic anticonvulsants should not be used routinely in patients with newly diagnosed brain tumours (standard). 2. In patients with brain tumours who have not had a seizure, tapering and discontinuing anticonvulsants after the first postoperative week is appropriate, particularly in those patients who are medically stable and who are experiencing anticonvulsant-related side effects (guideline). These results were confirmed by a more recent metaanalysis of literature that did not find any difference between the treatment interventions and the control groups in preventing a first seizure in participants with brain tumours with a higher risk of adverse events for those on antiepileptic drugs (Tremont-Lukats et al., 2008). Perioperative Period While it is now clear that long term prophylactic antiepileptic treatment is not beneficial, the role for peri-operative (up to 1 week post-operative) prophylactic anticonvulsant therapy is less well defined. It is possible that some patients undergoing surgery for malignant glioma may be at particularly high risk for peri-operative seizures. Furthermore, patients with large tumors and/or raised intracranial pressure could be at risk for significant morbidity if seizures occurred (Lwu et al., 2010). Recently, a retrospective chart review (Lwu et al., 2010) performed in adult patients with newly diagnosed malignant gliomas undergoing surgery showed that patients receiving peri-operative AED prophylaxis had a trend to reduced peri-operative seizures, and had few adverse effects. However, 68% of these patients were maintained on prophylactic AED continued beyond the first peri-operative week, contradicting published guidelines. Similarly, retrospective uncontrolled studies (Lee et al., 1989) suggested that AEDs could be effective in reducing the incidence of early postoperative seizures. A meta-analysis conducted by Kuijlen in 1996 (Kuijlen et al., 1996) on the effectiveness of prophylactic AEDs with supratentorial craniotomies suggested that prophylactically used AEDs showed a tendency to prevent postoperative convulsions. Nevertheless, prospective controlled studies demonstrated that antiepileptic prophylactic medication was not effective in reducing the incidence of late

S. Dupont

postoperative seizures and a prospective randomized study (Foy et al., 1992) that specifically addressed the incidence of early post-craniotomy seizures in patients receiving or not prophylactic antiepileptic treatment did not find significant differences between the regimes in respect of the incidence of seizures or death.

Reality Despite the AAN recommendations, recent data suggests that most newly diagnosed malignant glioma patients in North America (Lwu et al., 2010) or Europe (Hildebrand et al., 2005) continue to receive prophylactic AED. In the Glioma Outcomes Project, a prospective longitudinal database tracking clinical practice patterns and outcomes for malignant glioma patients seen at 52 North American centers between 1997 and 2000, 89% of 565 newly diagnosed glioma patients received AED while only 32% presented with seizures (Lwu et al., 2010). This discrepancy between the evidence-based medicine guidelines and the daily practice may be explained by three main factors: (i) the lack of information about the actual guidelines on prophylactic antiepileptic treatment in patients with brain tumours, (ii) the fact that these recommendations are based on studies performed with conventional antiepileptic drugs and (iii) the frequent initiation of anticonvulsant medication during the peri-operative period without further discontinuation. This discrepancy between guidelines and daily practice shows that additional corroborative studies assessing the effectiveness of new antiepileptic drugs to prevent epilepsy in patients with brain tumours are warranted.

Choice of the Optimal Antiepileptic Drugs In tumour brain patients who experienced seizures, the need for anticonvulsant medication is clear. Three factors may influence the choice of the optimal antiepileptic treatment: (i) efficacy (refractory seizures are commonly associated with brain tumours), (ii) tolerability: side-effects appear more frequent in patients with brain tumour compared with the overall population of epileptic patients who receive the same drugs, the commonest and more disabling side-effects of AEDs are: cognitive impairment, sedation, liver and

33 Epilepsy and Brain Tumours and Antiepileptic Drugs

haematological dysfunction, bone-marrow suppression and dermatological reactions, and (iii) pharmacokinetic interactions.

Which AEDs Should We Avoid? An emerging problem is the possible interactions between AEDs and chemotherapeutic drugs, because many of these drugs are metabolized by the cytochrome P450. Carbamazepine, phenytoin, phenobarbital, and primidone are known to have prominent cytochrome P450 (CYP) enzyme-induction effects and in a lesser manner oxcarbazepine and eslicarbazepine. Topiramate that has specific CYP2C19 enzyme-induction effects does not seem to interact with chemotherapeutic agents. Enzyme induction might lead to an increased metabolic rate of chemotherapeutic drugs that are metabolized by the cytochrome P450 and decreased serum drug concentrations that may result in a loss of efficacy. In a retrospective study, Oberndorfer et al. (2005) studied the effects of enzyme inducing and non- enzyme inducing AEDs in patients with glioblastoma multiforme treated with standard chemotherapeutic agents (mainly CCNU and temozolomide) on survival. They found that overall survival of patients who received an enzyme inducing antiepileptic drug (mostly carbamazepine) was significantly shorter than for those who received non-enzyme inducing antiepileptic drug (mostly valproic acid) (13.9 month versus 10.8 month). On the other hand, many chemotherapeutic agents that induce coenzymes of the cytochrome P450 may decrease the plasma concentration of concomitantly prescribed AEDs and reduce their efficacy. A special emphasis on phenytoin (PHT) is necessary. PHT has been frequently used for seizure control in brain tumour patients despite many undesirable pharmacological qualities. Phenytoin should be avoided in brain tumour patients for many reasons: (i) its interaction with steroid dexamethasone (PHT induces the metabolism of the steroid dexamethasone whereas dexamethasone decreases PHT serum levels (Chalk et al., 1984), (ii) its interaction with numerous chemotherapeutic agents that lower PHT levels whereas PHT accelerate the metabolism of a variety of chemotherapy drugs, (iii) its adverse effects, especially an increased risk of drug rashes, including Stevens-Johnson syndrome, an immune-mediated

319

rash associated with severe morbidity and death, particularly in those patients who received cranial irradiation (Borg et al., 1995). We can thus recommend to avoid the first-line prescription of enzyme-inducing antiepileptic drugs, especially phenytoin, in epileptic brain tumour patients.

First-Line Antiepileptic Medication Levetiracetam Leveticaetam is a new antiepileptic drug that is not metabolized through the P-450 hepatic cytochrome system. Its efficacy and tolerability have already been demonstrated in symptomatic partial epilepsies. Previous studies (Wagner et al., 2003) have demonstrated that add-on of levetiracetam led to some patients being seizure free or having seizure reduction without significant side effects. A recent prospective study (Rosati et al., 2010) assessed the efficacy and safety of levetiracetam monotherapy in the management of epilepsy in patients with glioma. On 176 consecutive patients with a first diagnosis of glioma who were enrolled in the study, eighty-two patients received levetiracetam because of a diagnosis of epilepsy. On those, 75 patients (91%) were seizure free; in 2 of these patients, levetiracetam was withdrawn because of intolerable adverse effects. Prompt and long-lasting control of seizures was obtained in 49 of 82 patients (60%) with a dose of levetiracetam that ranged from 1500 to 3000 mg/d, and 9 (11%) of the treated patients needed an increase of levetiracetam dosage to 4000 mg/d to become seizure free.

Valproic Acid Valproic acid is an established drug for the treatment of epileptic seizures, but it is also an enzymeinhibiting antiepileptic drug that may increase the activity and toxic effects of a concomitantly given drug. Nevertheless, recent data (Duenas-Gonzalez et al., 2008) suggest that valproic acid could have potent antitumour effects via a histone deacetylase inhibitor effect. Histone deacetylase inhibitors represent a novel class of therapeutic agents that play an important role in the regulation of gene transcription and oncogenesis through remodeling of chromatin

320

structure and dynamic changes in nucleosomal packaging of DNA (Li et al., 2005). Using valproic acid as a first-line antiepileptic drug could thus combine both antiepileptic and antitumoural effects. Nevertheless, prescription of valproic acid should not be considered in women if a pregnancy is planned. Lamotrigine Lamotrigine is a new large broad-spectrum AED. A study performed by Meyer et al. (1999), that examined lamotrigine content in the brain and the tumour tissue showed a good penetration of the molecule in tumoural tissues. Unfortunately, specific trials are lacking to estimate its specific efficacy in brain tumour patients. Nevertheless, its excellent tolerability and its lack of cognitive or sedative effects or pharmacokinetic interactions are good arguments to recommend it as a first choice treatment. Second-Line Antiepileptic Medication Alternative Monotherapy In case of failure of the first line medication, a secondline antiepileptic medication is necessary. Usually, alternative monotherapy is recommended. In this situation, second-line antiepileptic monotherapy should be chosen among the first line treatments that have not already been employed.

S. Dupont

Zonisamide Zonisamide is a second generation non-enzyme inducing AED with multiple mechanisms of action. Preliminary data on the use of zonisamide in add-on in patients with brain tumour-related epilepsy indicate that this drug may represent a valid alternative as addon in this particular patient population (Maschio et al., 2009). Topiramate The efficacy and tolerability of topiramate in patients with brain tumour and epilepsy has been recently addressed in an observational prospective study in 47 patients (Maschio et al., 2008). Of 45 patients followed for more than 3 months, 25 were seizure free (55.6%), 9 had a reduction of seizure frequency higher than 50% (20%) and 11 were stable (24.4%). Three patients (6.4%) discontinued topiramate for severe side effects and 4 (8.5%) had mild and reversible side effects. Nevertheless, potential cognitive side-effects of topiramate must be considered before its prescription in tumour brain patients.

Antitumour Treatments Antitumour therapy by neurosurgery, or chemotherapy contribute substantially to reducing seizure activity. At the opposite, cranial radiation could have negative effects on epilepsy.

Immediate or Differed Add-on Therapy Nevertheless, some authors have proposed adding immediately a second antiepileptic drug, rather than switching to monotherapy with another anticonvulsant (Vecht and Wilms, 2010). If a second line add-on antiepileptic medication is considered, valproic acid could be a good choice based on its antitumoural effects. A recent study (van Breemen et al., 2009) suggested that combination of levetiracetam and valproic acid was synergistic, and produced few or no cognitive side effects. In this study, of all 99 patients with glioma and seizures, those treated with a combination of valproic acid and levetiracetam showed the highest percentage of responders (81.5%). Other add-on AEDs may be considered in a second-line therapy:

Neursurgery In patients with malignant brain tumours, surgery resects part or all of the tumour in order to cure the tumour without specific attention to the epileptogenic zone. In refractory epileptic patients with developmental tumours such as DNET or gangliomas, surgery specifically resects the epileptogenic zone in order to cure epilepsy. Chemotherapy The alkylating chemotherapeutic agent temozolomide is widely used in patients with glioma with beneficial

33 Epilepsy and Brain Tumours and Antiepileptic Drugs

effects on seizure progression. It has also been proven that temozolomide (Pace et al., 2003) improved seizure control in patients with progressive low-grade glioma. Similarly, nitrosoureas, another alkylating agent, was tested as first line antitumoural therapy in patients with non-resectable fibrillary low-grade astrocytomas (Frenay et al., 2005). In this study, all patients with epilepsy had a clinical improvement with reduction in seizure frequency and 60% became seizure-free. Cranial Radiation In a recent study, Khan and Onar (2006) showed that whole-brain radiation treatment was associated with seizure recurrence. It has also been suggested that cranial irradiation could contribute to late intractable epilepsy in children with acute lymphocytic leukemia (Fasano et al., 2009). Furthermore, the use of phenytoin and, to a lesser extent, phenobarbital and carbamazepine during cranial irradiation is associated with an increased risk for severe, potentially fatal, mucocutaneous reactions. In conclusion, quality of life represents a key issue among individuals with brain tumours. Seizures and side effects of AEDs that are unfortunately common in patients with brain tumours can negatively influence their quality of life. The choice of AEDs with the best profile of efficacy and tolerability is thus essential. But treatment of epilepsy does not only involve the use of antiepileptic medication and may require a multidisciplinary approach.

References Borg M, Probert J, Zwi L (1995) Is phenytoin contraindicated in patients receiving cranial irradiation? Australas Radiol 39:42–46 Brodtkorb E, Nakken K, Steinlein O (2003) No evidence for a seriously increased malignancy risk in LGI1-caused epilepsy. Epilepsy Res 56:205–208 Calatozzolo C, Gelati M, Ciusani E, Sciacca F, Pollo B, Cajola L, Marras C, Silvani A, Vitellaro-Zuccarello L, Croci D, Boiardi A, Salmaggi A (2005) Expression of drug resistance proteins Pgp, MRP1, MRP3, MRP5 and GST-pi in human glioma. J Neurooncol 74:113–121 Chalk J, Ridgeway K, Brophy T, Yelland J, Eadie M (1984) Phenytoin impairs the bioavaibility of dexamethasone in neurological and neurosurgical patients. JNNP 47:1087–1090 Duenas-Gonzalez A, Candelaria M, Perez-Plascencia C, PerezCardenas E, de la Cruz-Hernandez E, Herrera L (2008)

321 Valproic acid as epigenetic cancer drug: preclinical, clinical and transcriptional effects on solid tumors. Canc Treat Rev 34:206–222 Fasano R, Bergen D (2009) Intractable epilepsy in patients treated for childhood acute lymphocytic leukemia. Seizure 18:298–302 Foy P, Chadwick D, Rajgopalan N, Johnson A, Shaw M (1992) Do prophylactic anticonvulsant drugs alter the pattern of seizures after craniotomy? JNNP 55:753–757 Frenay M, Fontaine D, Vandenbos F, Lebrun C (2005) Firstline nitrosourea-based chemotherapy in symptomatic nonresectable supratentorial pure low-grade astrocytomas. Eur J Neurol 12:685–690 Gabellini N, Masola V (2009) Expression of LGI1 impairs proliferation and survival of HeLa cells. Int J Cell Biol 2009:417197 Glantz M, Cole B, Forsyth P, Recht L, Wen P, Chamberlain M, Grossman S, Cairncross J (2000) Practice parameter: anticonvulsant prophylaxis in patients with newly diagnosed brain tumors. Report of the quality standards subcommittee of the American academy of neurology. Neurology 54: 1886–1893 Hildebrand J, Lecaille C, Perennes J, Delattre J (2005) Epileptic seizures during follow-up of patients treated for primary brain tumors. Neurology 65:212–215 Hwang S, Lieu A, Kuo T, Lin C, Chang C, Huang T, Howng S (2001) Preoperative and postoperative seizures in patients with astrocytic tumours: analysis of incidence and influencing factors. J Clin Neurosci 8:426–429 Khan R, Onar A (2006) Seizure recurrence and risk factors after antiepilepsy drug withdrawal in children with brain tumors. Epilepsia 47:375–379 Kuijlen J, Teernstra O, Kessels A, Herpers M, Beuls E (1996) Effectiveness of antiepileptic prophylaxis used with supratentorial craniotomies: a meta-analysis. Seizure 5: 291–298 Lee S, Lui T, Chang C, Cheng W, Wang D, Heimburger R, Lin C (1989) Prophylactic anticonvulsants for prevention of immediate and early postcraniotomy seizures. Surg Neurol 31:361–364 Lee J, Wen P, Hurwitz S, Black P, Kesari S, Drappatz J, Golby A, Wells W III, Warfield S, Kikinis R, Bromfield E (2010) Morphological characteristics of brain tumours causing seizures. Arch Neurol 67:336–342 Li X, Shu Q, Su J, Perlaky L, Blaney S, Lau C (2005) Valproic acid induces growth arrest, apoptosis, and senescence in medulloblastomas by increasing histone hyperacetylation and regulating expression of p21Cip1, CDK4, and CMYC. Mol Cancer Ther 4:1912–1922 Lwu S, Hamilton M, Forsyth P, Cairncross J, Parney I (2010) Use of peri-operative anti-epileptic drugs in patients with newly diagnosed high grade malignant glioma: a single center experience. J Neurooncol 96:403–408 Maschio M, Dinapoli L, Saveriano F, Pompili A, Carapella C, Vidiri A, Muti P, Jandolo B (2009) Efficacy and tolerability of zonisamide as add-on in brain tumor-related epilepsy: preliminary report. Acta Neurol Scand 120:210–212 Maschio M, Dinapoli L, Zarabla A, Pompili A, Carapella C, Pace A, Giannarelli D, Occhipinti E, Jandolo B (2008) Outcome and tolerability of topiramate in brain tumor associated epilepsy. J Neurooncol 86:61–70

322 Meyer F, Bandit P, Schubert A, Schöche J (1999) Lamotrigine concentrations in human serum, brain tissue, and tumor tissue. Epilepsia 40:68–73 Moots P, Maciunas R, Eisert D, Parker R, Laporte K, AbouKhalil B (1995) The course of seizure disorders in patients with malignant gliomas. Arch Neurol 52:717–724 Oberndorfer S, Piribauer M, Marosi C, Lahrmann H, Hitzenberger P, Grisold W (2005) P450 enzyme inducing and non-enzyme inducing antiepileptics in glioblastoma patients treated with standard chemotherapy. J Neurooncol 72:255–260 Ottman R, Winawer M, Kalachikov S, Barker-Cummings C, Gilliam T, Pedley T, Hauser WA (2004) LGI1 mutations in autosomal dominant partial epilepsy with auditory features. Neurology 62:1120–1126 Pace A, Vidiri A, Galiè E, Carosi M, Telera S, Cianciulli A, Canalini P, Giannarelli D, Jandolo B, Carapella C (2003) Temozolomide chemotherapy for progressive lowgrade glioma: clinical benefits and radiological response. Ann Oncol 14:1722–1726 Riva M (2005) Brain tumoral epilepsy: a review. Neurol Sci 26:S40–S42 Rosati A, Buttolo L, Stefini R, Todeschini A, Cenzato M, Padovani A (2010) Efficacy and safety of levetiracetam in patients with glioma: a clinical prospective study. Arch Neurol 67:343–346

S. Dupont Shamji M, Fric-Shamji E, Benoit B (2009) Brain tumors and epilepsy: pathophysiology of peritumoral changes. Neurosurg Rev 32:275–285 Siddiqui A, Kerb R, Weale M, Brinkmann U, Smith A, Goldstein D, Wood N, Sisodiya S (2003) Association of multidrug resistance in epilepsy with a polymorphism in the drugtransporter gene ABCB1. N E J M 348:1442–1448 Tremont-Lukats I, Ratilal B, Armstrong T, Gilbert M (2008) Antiepileptic drugs for preventing seizures in people with brain tumors. Cochrane Database Syst Rev 16:CD004424 van Breemen M, Rijsman R, Taphoorn M, Walchenbach R, Zwinkels H, Vecht C (2009) Efficacy of anti-epileptic drugs in patients with gliomas and seizures. J Neurol 256: 1519–1526 Van Breemen M, Wilms E, Vecht C (2007) Epilepsy in patients with brain tumours: epidemiology, mechanisms, and management. Lancet Neurol 6:421–430 Vecht C, Wilms E (2010) Seizures in low- and high-grade gliomas: current management and future outlook. Expert Rev Anticancer Ther 10:663–669 Wagner G, Wilms E, Van Donselaar C, Vecht C (2003) Levetiracetam: preliminary experience in patients with primary brain tumours. Seizure 12:585–586

Chapter 34

Familial Caregivers of Patients with Brain Cancer Youngmee Kim

Abstract Brain cancer affects not only the quality of life of individuals with the disease but also that of their family members and close friends. The impact on various aspects of the family caregivers’ quality of life (QOL) is significant throughout the trajectory of the illness, such as the acute and mid- to long-term survivorship phases a well as the bereavement phase. Theoretically and methodologically rigorous research on various aspects of the family’s quality of life, including physical, spiritual, and behavioral adjustment to brain cancer in the family, will help improving not only the quality of life of family caregivers of brain cancer patients but also the quality care the family members will provide to enhance the quality of life of the patients. Keywords QOL · Brain cancer · Caregiver · Self-efficacy · Caregiving · Tumor

Introduction Cancer affects the quality of life: not only of the individuals with the disease but also of their family members and close friends. Informal cancer care carried out by family caregivers requires meeting the survivor’s multiple needs. These include treatment monitoring; treatment-related symptom management; emotional, financial, and spiritual support; and assistance with

Y. Kim () Department of Psychology, University of Miami, Coral Gables, FL 33146, USA e-mail: [email protected]

personal and instrumental care (Given et al., 2001). Family caregivers report various problems from their caregiving experiences, including conflict among their social roles, restrictions of activities, strain in marital and family relationships, psychological distress, and diminished physical health (Kim and Given, 2008). Although most research in this area has focused on the negative experiences of providing care, a number of studies have also reported on benefits of taking care of family members who are ill. Family members providing care have reported benefit finding, post-traumatic growth, an improved sense of self-worth, and increased personal satisfaction (Folkman et al., 1994; Kim et al., 2007). The degree to which family caregivers have negative and positive experiences in caregiving affects their ability to care for the survivor, which also related to their own quality of life (QOL). A thorough review of the various aspects of cancer care and psychosocial characteristics of individuals involved in cancer care both during the time they are providing care and throughout the trajectory of the illness provides solid evidence about the psychological impact of cancer on family caregivers (Kim and Given, 2008). The review concluded that the impact on psychological, mental, social, physical, spiritual, and behavioral aspects of the family caregivers’ QOL is significant throughout the trajectory of the illness. Further, the QOL of family caregivers of persons with cancer varies along the illness trajectory. This review highlights the importance of assessing the adjustment of the caregivers over time and it also points out significant gaps in our understanding of the effects of family caregiving.

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_34, © Springer Science+Business Media B.V. 2011

323

324

Quality of Life of Family Caregivers of Patients with Cancer at the Acute Survivorship Phase Providing care to a family member with cancer has common sequelae across different types of cancer including brain cancer. The existing body of work on family caregivers of adult cancer survivors focuses primarily on the caregiver’s adjustment during the acute survivorship phase, that is, from the time of diagnosis to 1–2 years post-diagnosis. Furthermore, although QOL is a multidimensional construct (Ferrell et al., 1995; Weitzner et al., 1999), the most studied aspect of QOL is psychological distress. For example, one study found that 13% of caregivers of advanced cancer patients met the criteria for a psychiatric disorder. Still, most family caregivers do not experience clinically significant levels of depression when providing care. They have, however, reported levels of depressive symptoms similar to or even greater than those of the person with cancer. Compared with husbands of persons with no cancer or other chronic or acute illness, husbands of persons undergoing chemotherapy for breast cancer reported poorer mental health (Pitceahly and Maguire, 2003). Demographic and psychosocial characteristics associated with the caregiver’s distress (i.e., depression, general psychological distress, and cancer-specific distress) include female gender, younger age, spousal caregivers, lower socioeconomic status, employed, and lack of personal and social support (Kim and Given, 2008). Factors specific to the caregiving situation that are also related to caregiver’s distress include caregiving burden, self-efficacy for caregiving, types of care provided, and the survivor’s functional status (Nijboer et al., 1998). Studies examining other aspects of the QOL of caregivers are sparse. Limited evidence on the physical health of cancer caregivers suggests that although their level of physical functioning is comparable to the U.S. population norm, they have impaired physiological reactions in a daily setting and in a laboratory setting. Likewise, an adverse impact on the caregiver’s physical health has been reported among those providing burdensome caregiving (Kiecolt-Glaser et al., 2003; Vitaliano et al., 2003). The spiritual aspect of QOL among caregivers is also inadequately studied (Ferrell et al., 1995). A few

Y. Kim

studies have found that caregivers expressed similar spiritual needs, such as finding meaning, hope, and preparing for death with cancer survivors. Some work has shown that the caregiver’s spirituality buffered the adverse effect of caregiving stress on mental functioning, whereas it aggravated the effect on physical functioning. No studies, to date, has examined the behavioral adjustment of family caregivers, namely, engaging in or maintaining healthy lifestyle behaviors, such as healthy diet, exercise, or routine health screening, during this phase of survivorship (Kim and Given, 2008).

Family Caregivers of Patients with Brain Cancer at the Acute Survivorship Phase In addition to these common sequelae of caregiving for a family member with cancer, family caregivers of patients with brain cancer have burden specifically related to the disease (Sherwood et al., 2004). For example, the location of the tumor affects the caregiving burden. Patients with brain cancer in lefthemisphere lesions may need assistance based on verbal deficits, whereas those in right-hemisphere lesions may need assistance in facial recognition; and those in frontal lobe may need assistance for impaired judgment, decreased reasoning ability, and changes in emotional control. Deficits in neurological functioning are apparent in neuropsychiatric symptoms, such as anxiety, depression, irritability, anger, apathy, hallucinations, and mania, among others. The severity and number of neuropsychiatric symptoms can increase the caregiving burden. Managing neuropsychiatric symptoms can be challenging since it involves time demands (either arranging or providing 24-h supervision) for the patients’ problematic and unsafe behaviours. Decreases in functional status among the brain cancer population may also be linked to paralysis, paresis, gaze disorders, sensory loss, decreased levels of consciousness, ataxia, and difficulty swallowing. Deficits in cognitive functioning including memory deficits, confusion, shortened attention span to perform a task, and increased difficulty in coordinate motor movement often require family caregivers to give to the patients multiple directions and reminders to perform routine activities. These changes in neurologic

34 Familial Caregivers of Patients with Brain Cancer

and cognitive functioning not only place additional demands for providing daily and instrumental activities to the family members but also lead the family members to have grief experience before the death of the patient.

Quality of Life of Family Caregivers of Patients with Cancer at the Mid- to Long-Term Survivorship Phase Few studies to date have examined the QOL of caregivers after the initial turmoil around the time of diagnosis and treatment, namely, at mid- to long-term survivorship. Studies on caregivers of brain cancer patients during these phases are particularly lacking, due in part to relatively poor survival rate of brain cancer. With the advance in treatment and early detection, growing number of brain cancer survivors and their family members has now experienced these phases of survivorship.

Mid- to Long-Term Survivorship Phase with Remission Persistent psychological distress and role adjustment problems among spousal caregivers have been found about 1 year after the completion of cancer treatment, with levels greater than healthy controls (Hagedoorn et al., 2000; Northouse et al., 2000). Heightened uncertainty and fear of recurrence remained as a major concern among family members, although patients showed physical improvement and no cancerrelated symptoms. Sexual problems have also persisted in this phase, although studies suggested a gradual improvement over time to a level comparable to that of healthy age-matched controls (Northouse et al., 2000). Other studies, however, have found no lasting or long-term psychological distress, moderately high to normal range of marital satisfaction, and reduced disruptions in social relationships and activities and social role conflict that were imposed at the onset of caregiving around 2 years post-diagnosis (Kim and Given, 2008; Nijboer et al., 1998). Pre-existing

325

poor relationship quality and negative communication patterns between caregivers and the patient; caregiver’s sociodemographic characteristics, such as female gender; carrying out multiple social roles, such as employee or parent; greater levels of fear of recurrence and less social support; and patient’s medical characteristics, such as certain type of cancer appeared to predict poor post-treatment adjustment of both patients and their family caregivers at the mid- to long-term survivorship phase (Kim and Given, 2008). Few studies have documented physical aspects of the QOL of caregiver at this phase. Approximately half of the family members of long-term cancer survivors (diagnosed an average 3.5 years ago) reported that they had health problems, such as heart disease, hypertension, and arthritis (Mellon, 2002). The extent to which these morbid conditions are exacerbated by cancer caregiving stress remains unknown. Furthermore, few studies have examined changes in family member’s lifestyle behaviors, such as cancer screening, physical activity, and healthy diet, that would reflect the patient’s cancer serving as a “wake-up call” to the family member. One cross-sectional study reported that older and White families of long-term cancer survivors practiced a high number of health maintenance activities. Other studies, however, found that compliance rates with cancer screenings among first degree relatives of cancer survivors decreased as time elapsed after the relative’s diagnosis (Kim and Given, 2008). Although the 5-year survival rate for all cancers has improved in recent years, people still perceive cancer as a life-threatening disease. Consequently, the existential apprehensions and spiritual concerns evoked immediately after the diagnosis persist beyond the earlier phase of survivorship. A small number of studies have found that spousal caregivers reported similar levels of existential experience from their partner’s cancer as the patient did, and also had personal growth experiences years after their partner’s cancer diagnosis (Manne et al., 2004). Furthermore, various domains of the experience of benefit finding were uniquely associated with life satisfaction and depression. For example, coming to accept what happened and appreciating new relationships with others related to greater adaptation. Becoming more empathic toward others and reprioritizing values related to greater symptoms of depression (Kim et al., 2007).

326

Mid- to Long-Term Survivorship Phase with Recurrence and End-of-Life Care Although cancer recurrence has been one of the most stressful events in the course of illness for both patients and their family, few studies have examined the impact of cancer recurrence on the caregivers’ QOL (Northouse et al., 1995). Caregivers of persons with recurrent breast cancer reported as many adjustment problems as the survivors did or even greater levels of fear of recurrence, less social support, and poorer QOL than survivors (Northouse et al., 1995). These caregivers were troubled by depressive moods and stress in marital relationships. Self-efficacy, social support, and family hardiness related positively to the QOL of family caregivers, whereas symptom distress, fear of recurrence, hopelessness, family stressors, and negative appraisal of illness or caregiving related negatively to the QOL of family caregivers. Unlike the findings at the acute phase of survivorship, small correlations between recurrent survivors’ and their family caregivers’ QOL have been found. In an intervention group, family caregivers of women with recurrent breast cancer reported a less negative appraisal of their caregiving role at the end of the intervention sessions compared with the group that received usual care only, but the difference became non-significant in 3 months (Mellon et al., 2007). In addition to a cancer survivor’s cancer recurrence, family members are likely to be involved in active care to some survivors for end-of-life care. At the start of the end-of-life (palliative) care period, which begins after the poor prognosis is given, caregivers report heightened levels of caregiving burden, which continue during the entire palliative care period (Grunfeld et al., 2004). Overall, the caregiving burden is the strongest predictor of caregiver psychological distress, even more than the patient’s physical and emotional status (Grunfeld et al., 2004). However, one study found that the effectiveness of the use of certain coping strategies on caregivers’ QOL depended on the level of patient’s symptom distress: use of avoidant coping strategies related to poorer mental health of caregivers when the patient had low levels of symptoms distress (Kershaw et al., 2004). Studies examining aspects of the QOL of caregivers other than psychological distress at the midto long-term cancer survivorship phase are also

Y. Kim

sparse. Decreases in the caregiver’s physical health as the patient’s illness progresses have been observed. Caregivers of patients with advanced-staged cancer have reported severe fluctuations in sleep patterns since the cancer diagnosis, which was related to changes in depressive symptoms. The poor physical health of caregivers with chronic strain from providing care to a relative with dementia has been associated with increases in health risk behaviors, such as smoking, alcohol consumption, and the use of prescription drugs, and with getting inadequate rest, lack of exercise, and forgetting to take prescription drugs to manage their own health conditions (Beach et al., 2000). It remains unknown, however, whether cancer caregivers at this phase would also display increases in health risk behaviors. Evidence about spiritual concerns among dying patients, their family members, and health care providers is salient. A broad range of perspectives on illness, death, and dying drawn from diverse religious, philosophical, and cultural orientations has been associated with managing symptoms, coping with illness and loss, and enhancing quality of life (Chochinov and Cann, 2005). For example, a greater use of positive religious coping, such as benevolent religious appraisal, was related to better quality of life, whereas a greater use of negative religious coping, such as anger at God, was related to poorer overall well-being, particularly in psychological aspects of quality of life (Chochinov and Cann, 2005).

Quality of Life of Family Caregivers of Patients with Cancer at the Bereavement Phase Although survivorship ends at the death of the person with the disease, the caregiver’s life continues. Another group to be considered is bereaved caregivers. The death of a close family member is one of the most stressful of life events, so not surprisingly, bereavement in general has been widely studied for several decades. Caregiving stress has a negative impact on the psychological and physical health of caregivers, even increasing mortality (Schulz and Beach, 1999). Bereavement among family caregivers of cancer patients has also been studied relatively extensively. Existing findings, although

34 Familial Caregivers of Patients with Brain Cancer

inconsistent, suggest that poor psychological adjustment to bereavement (i.e., depression, anxiety, and complicated grief) relates to numerous demographic and psychosocial factors. These include older age, female gender, spouses, losing a younger family member, past grief experience, close bonds to the deceased, lack of bereavement coping self-efficacy, lower religiousness, lack of social support, greater number of adverse life events, lack of self-efficacy, shorter time between diagnosis and death, greater severity of the patient’s illness, perceived caregiving burden, and being unprepared for the relative’s death (Gilbar and Ben-Zur, 2002; Hebert et al., 2006). Again, other than studies of psychological distress at the bereavement phase, studies examining aspects of the QOL of cancer caregivers are sparse. One study with recently bereaved older persons showed that health behaviors, such as consistent exercise, monitoring caloric intake, and proper amount of sleep at 6and 11-months post loss, were related to better QOL at 19-months post loss (Chen et al., 2005). Among recently bereaved adults (an average, 6 months post loss), greater use of religious/spiritual coping was associated with more functional disabilities and fewer outpatient physical health care visits at baseline, which was not related to the health status at 4-month follow-up (Pearce et al., 2002). Identifying particularly vulnerable family caregivers before the relative’s death based on the presence of a dysfunctional family system (Chen et al., 2005) and the demographic characteristics previously mentioned has helped in interventions to prevent these caregivers from experiencing severe levels of grief and bereavement symptoms at 4 months, 6 months, and 12 months after the loss. In addition, an intervention designed to provide psychosocial support and information to assist in the bereavement process for family members and friends of recently deceased cancer patients has demonstrated its efficacy in improving their QOL at 3 months after completion of the eightsession psychoeducational group (Chen et al., 2005).

Future Directions One major gap identified is that we do not know whether such the impact of cancer is equally significant across different ethnic groups. According to the

327

United States Census data (2005), Hispanics/Latinos have become the largest U.S. minority group (14.4%), followed by African American (12.8%); Asians (4.3%) are the fastest growing ethnic minority. Although disparities in cancer incidence and mortality by ethnicity have been documented (American Cancer Society, 2010), the consequent disproportionate burden of cancer on families from different ethnic origins and differing rates in caregiving involvement remains, however, uninvestigated. In addition, it is not clear whether the impact of cancer on the family differs across different types of cancer or differs from caregiving for other types of chronic illnesses, such as dementia. The problems associated with having only limited knowledge about the QOL of family caregivers also apply to the need to assess the impact of caregiving on caregivers other than spouses of the care recipient. For example, offspring, parents, siblings, and friends are often primary or secondary caregivers among ethnic minorities. Most of the existing research investigates spousal caregivers. As the U.S. population continues to age and as ethnic minorities continues to grow, the proportion of spousal care of older individuals is expected to decrease as the proportion of adult children’s care increases. This trend applies particularly for adult daughters, because the caregiving role falls disproportionately on women (Kim and Given, 2008). The few existing studies suggest that adult offspring caregivers, particularly adult daughters, report higher levels of caregiving stress than spouses or other caregivers. This is due in part to the carrying of multiple social roles (they are the so-called sandwich generation). What remains largely unexplored is how the late effects and long-term needs of the person with cancer affect young caregivers, who have multiple and competing role demands at work and with their own family. Thus, timely attention to this subgroup of caregivers is needed in order to improve their psychological adjustment as they juggle other social roles while providing care to a family member with cancer. The aging of the U.S. population as the baby boom generation approaches retirement age also means that the number of older caregivers will grow. Although older individuals report better mental health or psychological adjustment in general and in the cancer caregiving context in particular, caregiving stress has a disproportionately burdensome impact on their physical health (Baltes and Carstensen, 2003; KiecoltGlaser et al., 2003). The extent to which the acute but

328

intensive nature of cancer caregiving or chronic psychological concerns about their relative’s cancer recurring could impair the caregiver’s immune function or cause premature aging remains unknown. In addition, literature on the financial and economic burden that families bear as caregivers is sparse (Yabroff and Kim, 2009). A few studies have documented family caregivers’ losses of employment benefits and health insurance due to their involvement with cancer care. This information is generally limited to the acute phase of the survivorship, however. The economic ramifications for the family extend beyond the time of diagnosis and treatment. Families often must assist in managing survivors’ serious late effects, with the related financial burden. This, often together with the disability of the survivor, particularly among those who are socioeconomically disadvantaged, is a problem that warrants future studies. In conclusion, accumulating evidence has supported the concept that cancer affects not only the patients/survivors but also their family members. The psychological impact of cancer on family caregivers across different phases of survivorship appears to be salient. The evidence about the quality of life of family caregivers of brain cancer patients, however, is lacking, particularly after the acute phase of survivorship. Theoretically and methodologically rigorous research on various aspects of the family’s quality of life, including physical, spiritual, and behavioral adjustment to brain cancer in the family, will help improving not only the quality of life of family caregivers of brain cancer patients but also the quality care the family members will provide to enhance the quality of life of the patients.

References American Cancer Society (2010) Cancer facts and figures. American Cancer Society, Atlanta, GA Baltes MM, Carstensen LL (2003) The process of successful aging: Selection, optimization and compensation. In: Staudinger UM, Lindenberger U (eds) Understanding human development: dialogues with lifespan psychology. Kluwer, Dordrecht, The Netherlands, pp. 81–104 Beach SR, Schulz R, Yee JL, Jackson S (2000) Negative and positive health effects of caring for a disabled spouse: longitudinal findings from the caregiver health effects study. Psychol Aging 15:259–271 Chen JH, Gill TM, Prigerson HG (2005) Health behaviors associated with better quality of life for older bereaved persons. J Palliat Med 8:96–106

Y. Kim Chochinov HM, Cann BJ (2005) Interventions to enhance the spiritual aspects of dying. J Palliat Med 8(Suppl 1): S103–S115 Ferrell BR, Dow KH, Grant M (1995) Measurement of the quality of life in cancer survivors. Qual Life Res 4: 523–531 Folkman S, Chesney MA, Christopher-Richards A (1994) Stress and coping in caregiving partners of men with AIDS. Psychiatr Clin North Am 17:35–53 Gilbar O, Ben-Zur H (2002) Bereavement of spouse caregivers of cancer patients. Am J Orthopsychiatry 72:422–432 Given BA, Given CW, Kozachik S (2001) Family support in advanced cancer. CA Cancer J Clin 51:213–231 Grunfeld E, Coyle D, Whelan T, Clinch J, Reyno L, Earle CC et al (2004) Family caregiver burden: results of a longitudinal study of breast cancer patients and their principal caregivers. Can Med Assoc J 170:1795–1801 Hagedoorn M, Buunk BP, Kuijer RG, Wobbes T, Sanderman R (2000) Couples dealing with cancer: role and gender differences regarding psychological distress and quality of life. Psycho-Oncology 9:232–242 Hebert RS, Prigerson HG, Schulz R, Arnold RM (2006) Preparing caregivers for the death of a loved one: a theoretical framework and suggestions for future research. J Palliat Med 9:1164–1171 Kershaw T, Northouse L, Kritpracha C, Schafenacker A, Mood D (2004) Coping strategies and quality of life in women with advanced breast cancer and their family caregivers. Psychol Health 19:139–155 Kiecolt-Glaser JK, Preacher KJ, MacCallum RC, Atkinson C, Malarkey WB, Glaser R (2003) Chronic stress and agerelated increases in the proinflammatory cytokine IL-6. Proc Natl Acad Sci USA 100:9090–9095 Kim Y, Given BA (2008) Quality of life of family caregivers of cancer survivors across the trajectory of the illness. Cancer 112(11 Suppl):2556–2568 Kim Y, Schulz R, Carver CS (2007) Benefit-finding in the cancer caregiving experience. Psychosom Med 69:283–291 Manne SL, Babb J, Pinover W, Horwitz E, Ebbert J (2004) Psychoeducational group intervention for wives of men with prostate cancer. Psycho-Oncology 13:37–46 Mellon S (2002) Comparisons between cancer survivors and family members on meaning of the illness and family quality of life. Oncol Nurs Forum 29:1117–1125 Mellon S, Kershaw TS, Northouse LL, Freeman-Gibb L (2007) A family-based model to predict fear of recurrence for cancer survivors and their caregivers. Psycho-Oncology 16: 214–223 Nijboer C, Tempelaar R, Sanderman R, Triemstra M, Spruijt RJ, van den Bos GA (1998) Cancer and caregiving: the impact on the caregiver’s health. Psycho-Oncology 7:3–13 Northouse LL, Laten D, Reddy S (1995) Adjustment of women and their husbands to recurrent breast cancer. Res Nurs Health 18:515–524 Northouse LL, Mood D, Templin T, Mellon S, George T (2000) Couples’ patterns of adjustment to colon cancer. Soc Sci Med 50:271–284 Pearce MJ, Chen J, Silverman GK, Kasl SV, Rosenheck R, Prigerson HG (2002) Religious coping, health, and health service use among bereaved adults. Int J Psychiatry Med 32:179–199

34 Familial Caregivers of Patients with Brain Cancer Pitceathly C, Maguire P (2003) The psychological impact of cancer on patients’ partners and other key relatives: a review. Eur J Cancer 39:1517–1524 Schulz R, Beach SR (1999) Caregiving as a risk factor for mortality: the Caregiver Health Effects Study. JAMA 282: 2215–2219 Sherwood P, Given B, Given C, Schiffman R, Murman D, Lovely M (2004) Caregivers of persons with brain tumor: A conceptual model. Nurs Inquiry 11:43–53 U.S. Bureau of the Census (2005) American Community Survey. U.S. Bureau of the Census [On-line]. Available: www. census.gov/acs/www/index.html

329 Vitaliano PP, Zhang J, Scanlan JM (2003) Is caregiving hazardous to one’s physical health? A meta-analysis. Psychol Bull 129:946–972 Weitzner MA, McMillan SC, Jacobsen PB (1999) Family caregiver quality of life: differences between curative and palliative cancer treatment settings. J Pain Symptom Manage 17:418–428 Yabroff KR, Kim Y (2009) Time costs associated with informal caregiving for cancer patients. Cancer 115(18 Suppl): 4362–4373

Chapter 35

Pain Management Following Craniotomy Doug Hughes and Scott Y. Rahimi

Abstract Postoperative pain management in neurosurgical patients is a complex issue. Patients undergoing craniotomies have traditionally received opiates for the management of their postoperative pain. This limited approach to craniotomy pain issues yields results that are limited by the specificity of the opiate class. Additionally, opiate receptors lack specificity within the central nervous system. This results in the undesirable side effect of CNS depression which limits the dosing a patient can receive. Our elaborate pain pathway is a system that has been thoroughly studied and described in the literature. The Gate Control Theory, introduced by Melzak and Wall, was the first widely accepted pain theory. Our current understanding of the pain pathway also acknowledges a central modulation of pain primarily from the thalamus and the periaquductal gray. As our understanding of the complexity of the pain pathway has evolved, new medications such as cyclooxygenase-2 inhibitors and Tramadol have been utilized to more specifically modulate the pain pathway. Keywords Pain · Craniotomy · Receptors · Pathways · Opiates · Fiber

D. Hughes () Department of Neurosurgery, Medical College of Georgia, Augusta, GA 30912, USA e-mail: [email protected]

Introduction Like many other surgical specialties, neurosurgeons continually confront postoperative pain management issues. While post-operative pain issues after spine surgery are common, cranial surgery offers analgesia challenges as well. Insults to dura, periosteum, muscle, and skin during the surgery are all potential sources of postoperative pain. Subtemporal and suboccipital approaches are especially prone to postoperative pain issues due to muscle dissection. Patient age has been shown to be inversely proportional to the extent of postoperative pain. Failure to properly treat pain leads to increased heart rate and blood pressure. This can increase risk for postoperative hemorrhage (Mordhorst et al., 2010). Traditionally four classes of analgesics have been used for the treatment of postoperative pain. These include opioids, nonsteroidal anti-inflammatory drugs (NSAIDs), cyclooxygenase-2 (cox-2) inhibitors, and a miscellaneous category including acetaminophen and tramadol. Neurosurgical patients requiring craniotomies have historically been treated exclusively with narcotic pain medications. Opiates are effective and proven for pain relief, especially in the postoperative period, but they have many side effects including nausea and vomiting, constipation, respiratory depression, and altered mental status that may interfere with the neurological examination of these patients. As new classes of analgesics have become available, some patients may benefit from alternative medication regimens that may avoid the side effects of narcotics. Newer analgesics can potentially be a means of reducing post-operative sedation, hospital length-of-stay,

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_35, © Springer Science+Business Media B.V. 2011

331

332

postoperative pain, and costs to hospitals and patients related to exclusive use of opiates.

Pain Theory Gate-Control Theory has been the most widely documented theory for pain. First proposed by Melzack and Wall (1965), the gate-control theory of pain came from feline nerve studies. These studies showed that when slow velocity type-C fibers were stimulated there was a positive dorsal root potential created. Conversely, when large type-A fibers were stimulated negative dorsal root potentials were created. The theory surmises that negative potential comes from inhibitory interneurons being activated by large myelinated fibers. Once activated, these gates presynaptically inhibit central transmission of peripheral pain signals. Similarly, smaller unmyelinated fibers can cause presynaptic excitation of upper-motor neurons. The dorsal root ganglia is also under central modulation from the thalamus, limbic system, and periaquaductal gray. This theory led to use of dorsal column stimulation for treatment of chronic peripheral pain syndromes. Today’s pain theory acknowledges a central modulation of pain primarily from the thalamus and the periaquductal gray. This theory now states that there are specialized peripheral nociceptors, specialized nerve endings for specific interpretation of pain, that are carried centrally by fast, A-δ fibers, and slow, C fibers.

Pain Receptors Pain sensations are transmitted centrally by A-δ fibers (Type III, thinly myelinated, average diameter 2–5 μm, average transmission velocity 12–30 m/s) and C (Type IV, no myelin, average diameter 0.1–0.3 μm, average transmission velocity 0.5–2 m/s) whose cell bodies both lie in the dorsal root ganglia. The A-δ fibers transmit fast pain and temperature that are described as sharp, intense, and localized. The C fibers transmit slow pain and temperature that are described as dull, burning, and diffuse. Both fiber types largely utilize free nerve endings which are branched and unmyelinated. These unmyelinated nerve endings are located in skin and other organs and are associated with Schwann cells. These endings do have specializations

D. Hughes and S.Y. Rahimi

including; mechanoreceptors, thermoreceptors, and polymodal nociceptors. Type C nerve endings respond to thermal and mechanical stimulation. They are also highly sensitive to chemical signs of adjacent tissue damage. Type A-δ fibers are responsive to light touch, pressure and temperature. They have an ability to respond incrementally to progressively intense stimuli. Pain fibers mainly occupy the lateral aspect of the doral root before entering the dorsal column. The type C fibers travel segmentally in the tract of Lissauer.

Pain Pathways Post operative pain after craniotomy involves sensory distributions of the trigeminal nerve (face, dura, anterior scalp) and the second cervical dermatome (posterior scalp). Peripheral pain fibers enter the spinal cord via the dorsal root ganglia. Pain fibers travel in Lissauer’s tract and terminate in the ipsilateral and contralateral marginal zone of the dorsal gray matter. Type A-δ first order fibers terminate primarily in rexed lamina I. Some of these fibers terminate in the outer layers of II and V. Type C first order fibers mainly terminate in the substania gelatinosa (lamina II). Pain fibers also terminate in the ventral horn gray matter, laminas VII and VIII, which contribute to the pain reflexes. The largest grouping of second-order pain fibers crosses the anterior commisure after ascending 2–3 segments to join the lateral spinothalamic tract. Lesions cause contralateral trunkal loss of pain and temperature. The deficit begins 2–3 dermatomal segments below lesion. The lateral spinothalamic tract ascends in the contralateral spinal cord to the ventral posterolateral nucleus. The trigeminal nerve provides sensation, including pain, to the anterior third of the scalp and the entire face (Fig. 35.1). Three branches, the ophthalmic (V1), the maxillary (V2), and the mandibular (V3) send pain fibers that converge on the gasserian or semilunar ganglion in Meckel’s cave. First-order nerve cell bodies are located here making this ganglion analogous to the peripheral dorsal root ganglion. These fibers continue in the spinal tract of V, which is analogous to Lissauer’s tract, and synapse in the spinal nucleus of V. The second order neurons decussate and run parallel to the spinothalamic tract as the trigeminothalamic tract.

35 Pain Management Following Craniotomy

333

Fig. 35.1 Trigeminal sensory and pain pathway

These fibers terminate in the ventral posteromedial nucleus. Pain fibers from the thalamus distribute to varied central locations. The primary sensory cortex (S1) is responsible for the conscious perception of pain. There are projections to the insular and cingulate cortex which tie the sensation of pain with emotional, sensory and memory functions of the limbic system. The dorsomedial and interlaminar thalamic nuclei also have diffuse and varied cortical and diencephalic projections. These projections are involved with overall arousal as part of the reticular activating system.

Drug Receptors Opioid Receptors Opiate receptors are an active area of research with new receptors and receptor subtypes identified

regularly. The three most studied receptors are mu (μ), delta (δ), and kappa (κ) receptors. Most opioids selectively target μ receptors at standard pharmacologic doses. Morphine is the classic μ specific agonist. Opiate medications have the desired effect of analgesia and provide euphoria in some patients. These medications also produce many neurological side effects that can be especially troubling in the post-operative craniotomy patient. Miosis, respiratory depression, sedation, lower seizure threshold, nausea and emesis are the most concerning side effects in a neurosurgical setting. μ receptors are clinically responsible for the majority of patient’s analgesia. They are also noted to stimulate patient’s central reward centers, causing euphoria, which can lead to opiate dependency. μ receptors depress respiratory function, alter mood and level of consciousness. δ receptors produce a mild level of analgesia in humans and are clinically the least noteworthy class. In animal models, κ receptors cause analgesia mainly in the spinal cord. These receptors

334

are responsible for some of the unpleasant side effects of opioids including dysphoria. These receptors function centrally as antagonists to the μ receptors. The κ class of receptors also demonstrates side effects of miosis, decreased GI motility, and respiratory depression. These receptors all demonstrate tolerance; a phenomenon where increasing levels of a drug is required to produce the same effect (Brunton, 2006).

Drug Classes Opioids Narcotic pain medications including codeine, oxycodone, hydrocodone, propoxyphene, and morphine have traditionally been used successfully for pain management in neurosurgery patients following craniotomies. (Quiney et al., 1996) Narcotic pain medications exert their effect by stimulating the μ (analgesia), δ, and κ subtype opioid receptors that are widespread in the central and peripheral nervous system. These locations include the brainstem which mediates respiratory drive, pupillary response, cough, nausea, blood pressure control, and arousal level; the hypothalamus which controls neuroendocrine function; the limbic system which affects emotional behavior; the substantia gelatinosa of the spinal cord which is involved with reception and integration of incoming sensory information; receptors in the gastrointestinal (GI) tract, stimulation of which leads to decreased GI motility; and peripheral histamine receptors involved with pruritis and vasodilation (Mycek et al., 2000). The use of narcotics alone can have several side effects associated with activation of these receptors which in turn may delay recovery and ambulation leading to extended hospital stays.

Nonsteroidal Anti-inflammatory Drugs NSAIDs provide analgesia by a different mechanism. These medications acetylate and therefore inhibit the cyclooxygenase enzyme which has two distinct isomers: cox-1 and cox-2 (Hawkey, 1999). The cyclooxygenase enzyme is responsible for the conversion of arachidonic acid to prostaglandins (Fig. 35.2). It is the inhibition of prostaglandins that reduces pain and

D. Hughes and S.Y. Rahimi

inflammation (Ferreira et al., 1973). Cox-1 is constitutively active throughout the body and provides routine physiologic functions such as protection of gastric mucosa and vascular hemostasis (Moncada et al., 1976). Cox-2 in contrast is mainly expressed by polymorphonuclear leukocytes and macrophages becoming active in response to inflammatory stimuli (Seibert and Masferrer, 1994). Although NSAIDs are effective at providing analgesia (cox-2 isomer), they can lead to platelet dysfunction and increased bleeding times (cox-1 isomer) which can be devastating in neurosurgery patients. The administration of NSAIDs following craniotomies has therefore been described to be a major risk factor leading to perioperative bleeding in neurosurgical patients (Palmer et al., 1994).

Anti-inflammatory Drugs: Cox-2 Inhibitors With the introduction of new analgesic medications such as cox-2 inhibitors, many specialties have moved away from exclusively using narcotic medications for pain management. The successful use of cox-2 inhibitors for management of postoperative pain has been well described in orthopedic surgery patients who have undergone hip and knee arthroplasty, lumbar discectomies, and spinal fusions (Bekker et al., 2002; Gimbel et al., 2001; Hubbard et al., 2003). Patients who have undergone gynecological and oral surgery have also benefited from decreased postoperative pain and narcotic use when treated with adjuvant cox-2 inhibitors (Tang et al., 2002). The cox-2 inhibitors provide the option of an anti-inflammatory agent without the increased risk of bleeding due to their selective inhibition of the cox-2 enzyme (Leese et al., 2000, 2003). In addition, these drugs do not affect the central nervous system and thus preventing many of the side effects associated with opioids as described earlier (Mycek et al., 2000). The use of cox-2 inhibitor has been shown to decrease postoperative pain in patients following craniotomies without an increased risk of postoperative hemorrhage (Rahimi et al., 2006). Recently concern has been raised regarding the increased risk of cardiovascular disease due to thrombotic events in arthritic patients taking cox-2 inhibitors. This has led to the withdrawal of medications including rofecoxib (Vioxx, Merck) and valdecoxib (Bextra, GD Searle). Due to these events,

35 Pain Management Following Craniotomy

335

Fig. 35.2 Cyclooxygenase metabolism of arachidonic acid

the use of cox-2 inhibitors for pain management has been limited by many institutions. Recently concern has been raised regarding the increased risk of cardiovascular disease due to thrombotic events in arthritic patients taking Vioxx and similar medications for prolonged periods. This is hypothesized to be due to inhibition of prostacyclins (inhibits platelet aggregation) in the vascular endothelium without the simultaneous inhibition of the platelet cox-1 (promotes platelet aggregation) (Mukherjee et al., 2001). The use of Vioxx has also been associated with other side effects including abdominal pain, fatigue, nausea, hypertension, diarrhea, and musculoskeletal pain. The sale of Vioxx was discontinued following release of data from the APPROVe study (Adenomatous Polyp Prevention on Vioxx). The APPROVe study was a clinical trial designed to assess whether or not Vioxx could prevent the recurrence of colorectal polyps in patients who had a history of

colorectal cancer. The study which was terminated prior to its completion found that patients receiving Vioxx daily for greater than 18 months had a two fold increase in their risk of developing cardiovascular disease when compared to a placebo group. The relevance of this data to short term use of these medications for post-operative pain is unclear. The use of cox-2 inhibitors has not been associated with thromboembolic events when used over a short period. However due to recent concerns as outlined above cox-2 inhibitors should be avoided in neurosurgery patients with significant cardiac risk. The chronic use of cox-2 inhibitors in an outpatient environment may be associated with increased cardiovascular risk. However proposing the short term use of these medications in a controlled hospital setting is very effective in reducing postoperative pain and opioid medication use in neurosurgical patients (Rahimi et al., 2006). Additionally other studies evaluating cox-2 inhibitors

336

have shown no increased risk of cardiovascular complications when compared to other NSAIDs or placebo (Seibert and Masferrer, 1994).

Other Analgesics Acetaminophen is an analgesic that is effective for mild to moderate pain though it has no anti-inflammatory actions (Mycek et al., 2000). Additionally acetaminophen is already present in many oral analgesics which are used for postoperative pain such as Percocet, Vicodin, and Darvocet. Tramadol is a relatively new analgesic which is under utilized for the management of postoperative pain in neurosurgical patients. Tramadol has been used very effectively for the management of postoperative pain in patients following orthopedic, cardiothoracic, and obstetric procedures for several years (Bourne et al., 2005). Tramadol exerts its analgesic effects by inhibiting the reuptake of serotonin and norepinephrine though the exact mechanism of its analgesic effect is not completely understood. Tramadol has no effect on platelet or coagulation function therefore making it a safe medication to use for neurosurgical patients following craniotomies. Tramadol also has a weak interaction with opioid receptors which can lead to some similar side effects as opioids including nausea, vomiting, dry mouth, and dizziness (Leppert and Luczak, 2005). Tramadol has been proven to be very effective in managing postoperative pain following craniotomies in neurosurgical patients. When used in conjuction with opioids, tramadol reduces the need for narcotic medications thereby reducing the associated side effects as well (Rahimi et al., 2010). Steroids Postoperative pain is reduced by intraoperative corticosteroid administration. This class of drugs inhibits prostaglandin synthesis and may increase endogenous endorphin production. This anti-inflammatory action previously justified steroid use for prevention of cerebral edema postoperatively (Burchiel and De Beneditties, 2002). Steroids are also beneficial because they decrease postoperative nausea. In one prospective anesthesia study, absence of corticosteroids caused patient’s a 119% increase in average pain score.

D. Hughes and S.Y. Rahimi

Corticosteroids adversely affect glucose control, which can lead to higher risk of infection. Strict glucose control is a goal that may require an insulin drip in the setting of corticosteroid therapy (Mordhorst et al., 2010). In conclusion, pain control following craniotomies is a major issue facing neurosurgeons, particularly in the first 24 h following surgery. Type of approach, patient age, and use of intraoperative steroids all effect pain scores postoperatively. Opiates remain a mainstay of postoperative pain treatment. However, new opportunities for pain control have been shown to be effective with cox-2 inhibitors and tramadol. Clinical assessment of postcraniotomy pain by nursing staff is also an important factor in the proper dosing of pain medicines. The use of non-narcotic medication classes in conjunction with aggressive pain management by nursing staff and physicians can help reduce postoperative pain, cost of hospitalization, and narcotics side effects in patients following neurosurgical procedures. Acknowledgements Special Thanks to Michael Jensen for Illustrations.

References Bekker A, Cooper P, Frempong-Boadu A, Babu R, Errico T, Lebovits A (2002) Evaluation of preoperative administration of the cyclooxygenase-2 inhibitor rofecoxib for the treatment of postoperative pain after lumbar disc surgery. Neurosurgery 50(5):1053–1058 Bourne HM, Rosenthal NR, Xiang J, Jordan D, Kamin M (2005) Tramadol/acetaminophen tablets in the treatment of postsurgical orthopedic pain. Am J Orthop 35:592–597 Brunton L (2006) Chapter 21: opioids. Goodman and Gilman’s the pharmacological basis of therapeutics. McGraw-Hill, New York, NY. Burchiel K, De Benedittis G (2002) Chapter 19: management of postoperative pain in neurosurgery. Surgical management of pain. Thieme, New York, NY. Ferreira SH, Moncada S, Vane JR (1973) Prostaglandins and the mechanism of analgesia produced by aspirin-like drugs. Br J Pharmacol 49:86–97 Gimbel JS, Brugger A, Zhao W, Verburg KM, Geis GS (2001) Efficacy and tolerability of celecoxib versus hydrocodone/acetaminophen in the treatment of pain after ambulatory orthopedic surgery in adults. Clin Ther 23(2):228–241 Hawkey CJ (1999) Cox-2 inhibitors. Lancet 353:307–314 Hubbard RC, Naumann TM, Traylor L, Dhadda S (2003) Parecoxib sodium has opioid-sparing effects in patients

35 Pain Management Following Craniotomy undergoing total knee arthroplasty under spinal anaesthesia. Br J Anaesth 90(2):166–172 Leese PT, Hubbard RC, Karim A, Isakson PC, Yu SS, Geis GS (2000) Effects of celecoxib, a novel cyclooxygenase-2 inhibitor, on platelet function in healthy adults: a randomized, controlled trial. J Clin Pharmacol 40:124–132 Leese PT, Recker DP, Kent JD (2003) The COX-2 selective inhibitor, valdecoxib, does not impair platelet function in the elderly: results of a randomized controlled trial. J Clin Pharmacol 43:504–513 Leppert W, Luczak J (2005) The role of tramadol in cancer pain treatment-a review. Support Care Cancer 13:5–17 Melzack R, Wall PD (1965) Pain mechanism: a new theory. Science 150:971 Moncada S, Gryglewski R, Bunting S, Vane JR (1976) An enzyme isolated from arteries transforms prostaglandin endoperoxides to an unstable substance that inhibits platelet aggregation. Nature 263:663–665 Mordhorst C, Latz B, Kerz T, Wisser G, Schmidt A, Schneider A, Jahn-Eimermacher A, Werner C, Engelhard K (2010) Prospective assessment of postoperative pain after craniotomy. J Neurosurg Anesthesiol 22:202–206 Mukherjee D, Nissen SE, Topol EJ (2001) Risk of cardiovascular events associated with selective cox-2 inhibitors. JAMA 286:954–959

337 Mycek MJ, Harvey RA, Champe PC (2000) Nonsteroidal anti-inflammatory drugs. Pharmacology, 2nd ed. Lippincott Williams & Wilkins, Philadelphia, PA. Palmer JD, Sparrow OC, Iannotti F (1994) Postoperative hematoma: a 5-year survey and identification of avoidable risk. Neurosurgery 35:1061–1065 Quiney N, Cooper R, Stoneham M, Walters F (1996) Pain after craniotomy. A time for reappraisal? Br J Neurosurg 10: 295–299 Rahimi SY, Alleyne CH, Vernier E, Witcher MR, Vender JR (2010) Postoperative pain management with tramadol after craniotomy: evaluation and cost analysis. J Neurosurg 112:268–272 Rahimi SY, Vender JR, Macomson SD, French A, Smith JR, Alleyne CH (2006) Post-operative pain management following craniotomy: evaluation and cost analysis. Neurosurgery 59:852–857 Seibert K, Masferrer JL (1994) Role of inducible cyclooxygenase (Cox-2) in inflammation. Receptor 4:17–23 Tang J, Li S, White PF, Wender RH, Quon R, Sloninski A (2002) Effect of parecoxib, a novel intravenous cyclooxygenase type-2 inhibitor on the postoperative opioid requirement and quality of pain control. Anesthesiology 96:1305–1309

Chapter 36

Air Transportation of Patients with Brain Tumours Peter Lindvall and Tommy Bergenheim

Abstract Air transportation of patients to specialised health care services has become ever more important in modern health care. Air transport has the advantage of a swift transport and the possibility to cover large geographical areas. Air transport may be used for the pre and postoperative transport of patients with brain tumours. Preoperative transport of patients harbouring brain tumours seem to be safe in most cases even if there are a few reports reporting clinical deterioration during and after air transport. Results from microdialysis of normal brain tissue and tumour tissue have shown only minor metabolic changes during and after air transport. In case of postoperative air transport of patients operated for brain tumours the presence of intracranial air be associated with an increased risk. Intracranial air can be treated as an ideal gas and will expand as the cabin pressure in the aeroplane decreases. Air trapped in the intracranial cavity cannot easily expand, however, and this may result in an increased intracranial pressure. Therefore it may be recommended that patients with a large amount of intracranial air should be transported with ground transportation or with air transportation where the cabin is pressurized to sea level. Keywords Air transportation · Health care · Microdialysis · Tumour tissue · Intracranial air · ICP

P. Lindvall () Umea University Hospital, Umea, Sweden e-mail: [email protected]

Introduction Air-medical transport has become ever more important in regionalized health care systems to ensure that patients can be speedily transferred to specialised health care services like Neurosurgery. The advantage of air transport may be short transfer times of patients which may be especially valuable in medical emergencies and in covering large geographical areas where ground transportation would be too time consuming. Air transport can be provided by fixed wing aircrafts (aeroplanes) or helicopter air ambulances. Aeroplanes generally have the advantage of a longer operating radius compared to an aeroplane. Air transport has proven especially valuable when transporting trauma patients to trauma centres where a neurosurgical service is available. It has been proven that early detection and surgical evacuation of intracranial haematomas reduces mortality and morbidity (Haselsberger et al., 1988). Air ambulances and even commercial airliners may also be used for the transportation of patients harbouring brain tumours both pre or postoperatively. There are, however, several medical issues that need to be addressed when evaluating if it is appropriate to send a patient with a brain tumour with an aviation transport. There are also some possible hazards, depending on the patient’s condition, associated with the transport itself. Whereas the benefit of a swift aviation transport is obvious, relatively little has been written concerning possible risks associated with this form of transport. With increasing availability and use of air transport these issues needs to be addressed. These issues are primarily related to critical events that may arise during

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4_36, © Springer Science+Business Media B.V. 2011

339

340

transport of patients in a poor condition or barometric variations associated with air transport. It has previously been shown that critical events were observed in 5.1% of all urgent air-medical transports. These events consisted primarily of hypotension or airway management procedures during transport. Some predictors of those critical events were assisted ventilation or hemodynamic instability before transport, reflecting the patients condition prior to transport (Singh et al., 2009). Barometric variations will also occur during air transport, especially during transport with an aeroplane. In commercial airliners the cabin pressure is routinely pressurized to a minimum of 75 kPa, which is equivalent to an altitude of 2440 m (Andersson et al., 2003; Zrinzo et al., 2006). The resultant decrease in the partial pressure of oxygen may result in lower blood oxygen saturation. At cruising altitude in a commercial airliner with a cabin pressure of 75 kPa, the oxygen partial pressure is approximatively 15.4 kPa (Lindgren, 2003). This may lead to hypoxemia in normal cerebral tissue as well as tumour tissue. This may be especially evident in patients with critical intracranial circulation and increased intracranial pressure due to tumour mass effect and brain oedema. Death attributable to high altitude cerebral oedema (HACE) experienced by mountaineers have been reported in altitudes as low as 2500 m (Houston, 1976). Using air ambulance services there is usually an option keep the cabin pressure at sea level during transport. However, limitations of the aircraft pressurization system will reduce the possible

Fig. 36.1 Images from a postoperative CT scan of a patient operated with removal of a frontal bilateral falx meningeoma. The CT scan was performed 24 h after surgery and shows a postoperative resection cavity and a frontal low-attenuating collection consistent with a moderate amount of intracranial air

P. Lindvall and T. Bergenheim

operational flying altitude significantly resulting in higher risk of turbulence and higher fuel consumption (Andersson et al., 2003).

Intracranial Air and Postoperative Air Transport Intracranial air (pneumocephalus) may imply a communication between the intracranial cavity and the atmosphere or a paranasal sinus. The most common cause of pneumocephalus is head trauma even if there are other possible causes such as meningitis from gas forming organisms and a previous intracranial surgery (Markham, 1967). Following intracranial surgery there are in most cases some intracranial air postoperatively (Fig. 36.1). Pneumocephalus has also been reported in patients treated with continuous positive airway pressure (CPAP) after surgery through a transsphenoidal route (Haran and Chandy, 1997). Pneumocephalus is usually a benign condition and there is a spontaneous resorption even if intracranial air can be present up to 3 weeks postoperatively (Reasoner et al., 1994). In rare cases intracranial air will produce a mass effect known as tension pneumocephalus. In these cases operation may be necessary to relieve the pressure from trapped intracranial air. Normally the intracranial pressure (ICP) has been shown to remain stable during variation in ambient pressure. In one previous study of a patient it was shown that the ICP was unchanged during a simulated ascent to altitudes of

36 Air Transportation of Patients with Brain Tumours

341

49.000–50.000 ft (14,935–15,240 m) (Peterson et al., 1944). The presence of intracranial may, however, change these conditions. During air transport gas contained within distensible body cavities will increase in volume as the cabin pressure decreases. At pressures below 200 kPa air can be treated as an ideal gas and therefore the volume expansion ca be calculated using the Boyle-Mariottes’s law (P1 V1 = P2 V2 ) (Zemansky and Dittman, 1987). Air trapped in the intracranial cavity cannot easily expand, however, and this may result in an increased intracranial pressure. The actual altitude may also be less important than rate of change of altitude. The time frame may be crucial as spatial compensation mechanisms such as resorption of CSF or a decreased intracranial venous volume will be dependant on time. The slower the process, the more time for compensating mechanisms to work. In one study a mathematical computer model was used to estimate resultant ICP depending on different intracranial air volumes, altitude and rate of altitude change (Andersson et al., 2003). It was hypothetically shown, by using this model, that ICP values above 30 mmHg could be reached depending on clinical relevant factors such as ICP before air transport, air volume, altitude and rate of change of altitude. Even if rarely reported, a clinical exacerbation with headache and confusion has previously been reported in patients during commercial flights. In these patients subsequent radiological investigations showed the presence of either pneumocephalus or an expansive intracerebral pneumatocele that needed surgical evacuation (Chan et al., 2000; Mahabir et al., 2004).

transported with a commercial aeroplane or air ambulance for further treatment. These patients received no supplemental oxygen during transport. Microdialysis catheters were placed in normal brain tissue, or in brain adjacent to tumor (BAT) for two patients who had underwent resection. For two other patients with a recent biopsy catheters were placed in normal brain and tumor tissue. The levels of glucose, glycerol, glutamate and the lactate/pyruvate ratio were analysed for each patient in tumor/BAT or normal brain tissue before, during and after air transport. Using microdialysis there were only small differences detected in the levels of cerebral metabolites after air transport compared to a previous fasting sample. Clinically the air transport was also uneventful without any signs of deterioration. Analysing mean values there was a small but significant increase in the lactate/ pyruvate ratio in normal cerebral tissue after air transport compared to a previous fasting sample and for tumour tissue there was a small decrease in glucose and an increase in glutamate. The lactate/pyruvate ratio is a marker of changes in the redox state of cells due to cerebral ischemia whereas glutamate is a non – essential amino acid and also a neurotransmitter. Glutamate has been suggested to be a marker of brain injury and metabolic disturbance, since increased levels of glutamate have been shown in patients with traumatic head injury and subarachnoid haemorrhage (Gopinath et al., 2000; Unterberg et al., 2001).

Cerebral Metabolism in Brain and Tumour Tissue During Air Transport

Air ambulance transport may enable a swift transport of patients in need of an urgent neurosurgical procedure, which seem to be most important in patients with head trauma (Haselsberger et al., 1988). Increased availability has also led to an increasing use of air ambulance for other neurosurgical patients such as pre or postoperative transport of patients with brain tumours. At the neurosurgical department at Umeå University Hospital air ambulance transport of patients with helicopter or aeroplane pre and postoperatively has been routinely used for more than a decade. Our department covers a geographically large and sparsely populated area and air ambulance transport has proven invaluable especially for neurosurgical emergencies. In our experience air ambulance

Little has been written concerning the metabolism in tumour or normal brain tissue during air transport. If the presence of intracranial air, or the relative hypoxemia associated with a lower partial pressure of oxygen, during air transport would be clinically significant one would expect this to be reflected in the cerebral metabolism. In one short report microdialysis was used to evaluate metabolism or possible cellular damage in four patients before, during and after air transport (Lindvall et al., 2008). These patients were newly operated due to glioblastoma multiforme and

Discussion

342

transport has been proven safe and to our recollection we have not encountered any significant clinical deterioration of patients during transport that could be attributed to the transport itself. Clinical deterioration of patients during transport is not uncommon but this is mainly due to progression of the intracranial pathology. Nevertheless, there are reports of patients with or without pre-existing intracranial pathology who have deteriorated clinically during air transport. As discussed above barometric variations with a decreased cabin pressure can result in expansion of intracranial air that has been trapped during a previous intracranial procedure. Symptomatic pneumocephalus or even an intracerebral pneumatocele can also occur prior to surgery in patients with a discontinuity in the scull base. Erosion of the paranasal sinuses resulting in pneumocephalus or a pneumatocele during air transport has been reported to occur due to chronic sinusitis or osteomas (Chan et al., 2000; Mahabir et al., 2004). There are principally two mechanisms explaining the development of pneumocephalus. The first would require a leak of cerebrospinal fluid (CSF) and subsequently development a relatively negative intracranial pressure sufficient enough to cause air to be introduced in the intracranial cavity. The second mechanism that has been proposed is the presence of a valve mechanism. In these cases the presence of a positive endotympanic pressure (ie a vasalva manoeuvre) would force air from pneumatic cavities to the intracranial cavity. When the intracranial pressure then exceeds the pressure in the pneumatic cavities the valve would close. Rare cases of haemorrhage in brain tumours have also been reported to occur after air transport. Two cases have been reported with clinical deterioration soon after air transport due to haemorrhage in a melanoma metastasis and meningeoma (Goldberg and Hirschfeld, 2002). Among other suggestions it was speculated that haemorrhage in these cases would be the result of hypoxemia and ischemia in tumour tissue and a subsequent haemorrhage into necrotic tumour tissue. According to results from microdialysis in patients with glioblastomas there were, however, only minor metabolic changes during air transport (Lindvall et al., 2008). New neurological deficits presenting during or shortly after air transport have also been reported in patients with brain tumours without signs of haemorrhage. In one of these patients a radiological investigation showed signs of chronically elevated

P. Lindvall and T. Bergenheim

intracranial pressure such as tonsillar herniation, and in one other patient deterioration could be attributed to a tumour in the posterior fossa causing obstruction of the aqueduct and hydrocephalus (Zrinzo et al., 2006). In one of the cases the cause of deterioration was unclear (Perin et al., 2009). Brain tumours have previously been reported to become symptomatic at high altitudes (Shlim et al., 1991). The proposed mechanism may be similar to the development of HACE. Experiments in rats have shown that hypoxia may trigger the upregulation of vascular endothelial growth factor (VEGF). The increased VEGF expression may cause an increased capillary permeability and cerebral oedema (Xu and Severinghaus, 1998). The occurrence of an epileptic seizure, which is a common symptom associated with brain tumours, may also constitute a risk during air transport. It has previously been reported that there may be an increased seizure frequency in patients with known epilepsy after air transport (Trevorrow, 2006). The literature on air transportation of patients with space occupying lesions such as brain tumours is sparse. Although there are a few reports on clinical deterioration possibly related to the nature of the transport it seems that air transportation of brain tumour patients preoperatively is safe in most cases. Postoperatively, the common situation with intracranial air may be associated with an increased risk using air transportation. In cases with a large volume of intracranial air we would, to be on the safe side, recommend ground transportation or air transportation using a sea level cabin pressure.

References Andersson N, Grip H, Lindvall P, Koskinen LO, Brandstrom H, Malm J, Eklund A (2003) Air transport of patients with intracranial air: computer model of pressure effects. Aviat Space Environ Med 74:138–144 Chan YP, Yau CY, Lewis RR, Kinirons MT (2000) Acute confusion secondary to pneumocephalus in an elderly patient. Age Ageing 29:365–367 Goldberg CR, Hirschfeld A (2002) Hemorrhage within brain tumors in association with long air travel. Acta Neurochir (Wien) 144:289–293 Gopinath SP, Valadka AB, Goodman JC, Robertson CS (2000) Extracellular glutamate and aspartate in head injured patients. Acta Neurochir Suppl 76:437–438 Haran RP, Chandy MJ (1997) Symptomatic pneumocephalus after transsphenoidal surgery. Surg Neurol 48:575–578

36 Air Transportation of Patients with Brain Tumours Haselsberger K, Pucher R, Auer LM (1988) Prognosis after acute subdural or epidural haemorrhage. Acta Neurochir (Wien) 90:111–116 Houston CS (1976) High altitude illness. Disease with protean manifestations. JAMA 236:2193–2195 Lindgren T (2003) Cabin air quality in commercial aircraft: exposure, symptoms and signs. Medical dissertation, Uppsala University, Uppsala Lindvall P, Roslin M, Bergenheim AT (2008) Cerebral metabolism during air transport of patients after surgery for malignant glioma. Aviat Space Environ Med 79:700–703 Mahabir RC, Szymczak A, Sutherland GR (2004) Intracerebral pneumatocele presenting after air travel. J Neurosurg 101:340–342 Markham JW (1967) The clinical features of pneumocephalus based upon a survey of 284 cases with report of 11 additional cases. Acta Neurochir (Wien) 16:1–78 Perin A, Larosa F, Longatti P (2009) Barometric changes in patients with intracranial lesions: can they dive and fly? Surg Neurol 71:368–371. discussion 371 Peterson EW, Kent BS, Cone WV (1944) Intracranial pressure in the human subject at altitude. Arch Neurol Psychiatry 52:520–525

343 Reasoner DK, Todd MM, Scamman FL, Warner DS (1994) The incidence of pneumocephalus after supratentorial craniotomy. Observations on the disappearance of intracranial air. Anesthesiology 80:1008–1012 Shlim DR, Nepal K, Meijer HJ (1991) Suddenly symptomatic brain tumors at altitude. Ann Emerg Med 20:315–316 Singh JM, MacDonald RD, Bronskill SE, Schull MJ (2009) Incidence and predictors of critical events during urgent air-medical transport. CMAJ 181:579–584 Trevorrow T (2006) Air travel and seizure frequency for individuals with epilepsy. Seizure 15:320–327 Unterberg AW, Sakowitz OW, Sarrafzadeh AS, Benndorf G, Lanksch WR (2001) Role of bedside microdialysis in the diagnosis of cerebral vasospasm following aneurysmal subarachnoid hemorrhage. J Neurosurg 94:740–749 Xu F, Severinghaus JW (1998) Rat brain VEGF expression in alveolar hypoxia: possible role in high-altitude cerebral edema. J Appl Physiol 85:53–57 Zemansky M, Dittman R (1987) Heat and thermodynamics, 6th ed. McGraw-Hill, Singapore, pp 108–109 Zrinzo LU, Crocker M, Zrinzo LV, Thomas DG, Watkins L (2006) Commercial flight and patients with intracranial mass lesions: a caveat. Report of two cases. J Neurosurg 105: 627–630

Index

A AAN, see American Academy of Neurology (AAN) Aaronson, N. K., 304 Abbott Laboratories, 151 Abdulrauf, S. I., 144, 146 Abla, A., 146 aBV, see Arterial blood volume (aBV) Acetaminophen, 331, 336 Add-on antiepileptic therapy, 313, 320 Addicott, M. A., 141 Addolorato, G., 295–310 Ad.EGR-TN, 288–290 Adenomatous Polyp Prevention on Vioxx (APPROVe), 335 Adler, J. R., 281 ADM, see Adriamycin (ADM) Adriamycin (ADM), 171–173 Adult offspring caregivers, 327 AEDs, see Antiepileptic drugs (AEDs) Ahmed, A. A., 78 Air ambulances, 339–341 Air transportation of patients with brain tumours advantages, 339 air ambulances, hazards/issues, 339–340 critical events during transport, 340 HACE in mountaineers, 340 high risk of turbulence and higher fuel consumption, 340 cerebral metabolism in brain and tumour tissue glutamate, marker of brain injury, 341 microdialysis, 341 development of pneumocephalus, mechanisms, 342 haemorrhage after air transport, cases, 342 intracranial and postoperative air transport, 340–341

clinical deterioration of patients, 342 epileptic seizure, symptom, 342 pneumocephalus, causes, 340 postoperative resection cavity, CT image, 340f tension pneumocephalus, 340 variations in ICP with altitude, 340–341 Ala-Pro-Arg-Pro-Gly (APRPG), 171–176 Alavi, A., 240 Albagli, O., 110 Aldape, K. D., 252 Alesch, F., 178 Alexander, E., 56 Alfieri, A., 263 ALK, see Anaplastic lymphoma kinase (ALK) Allard, E., 207–215 Almeida, O. F., 112 Alonso, M. M., 114 Alsop, D. C., 137 Alternative monotherapy, 320 Altman, D. G., 139, 141 Altundag, K., 79 American Academy of Neurology (AAN), 317–318 American Cancer Society, 261, 280, 327 Amino acid-based radiopharmaceuticals, 2 Amino acid PET in brain tumors, clincial applications diagnostic assessment of recurrent tumors, 123 extent of tumor for biopsy and treatment planning, 119–120 MET/FET-PET, benefits over MRI, 118f, 119–120, 120f glioma grading and prognosis, 121–122 amino acid imaging, prognostic predictor for LGG/HGG, 122 amino acid uptake, effects on HGG/LGG, 122 FET-PET combined with MRI, diagnostic assessment, 122

M.A. Hayat (ed.), Tumors of the Central Nervous System, Volume 3, DOI 10.1007/978-94-007-1399-4, © Springer Science+Business Media B.V. 2011

345

346

Amino acid PET in brain tumors (cont.) imaging brain tumors in children, 124 intracranial tumors, differential diagnosis of, 120–121 FET-PET combined with 1 H-MRS, 121 FET-PET combined with MRI, study, 121 proton magnetic resonance spectroscopy (1 H-MRS), 121, 124 treatment monitoring, 123–124 MET-PET, usefulness in therapy assessment, 123–124 pseudoprogression, 123 Amino acids, 2, 117–125, 240 See also Amino acid-based radiopharmaceuticals; Amino acid PET in brain tumors, clincial applications; Radiolabelled amino acids 5-Aminolevulinic acid photodynamic therapy, 2 Analgesics, treatment of postoperative pain cox-2 inhibitors, 331, 334–336 increased risk of cardiovascular disease, 335 use of Vioxx, APPROVe study, 335 newer analgesics, benefits, 331–332 acetaminophen, 336 tramadol, 336 NSAIDs, 331, 334 opioids, 331, 334 steroids corticosteroid therapy, 336 prostaglandin synthesis, inhibition of, 336 Anaplastic lymphoma kinase (ALK), 64–65, 71 Anderson Cancer Center, 280 Anderson, S. I., 299 Andersson, N., 340–341 Andrews, D. W., 57–58, 218, 220, 222–223 Anghileri, M., 128–129 Angiogenesis, 29, 39, 79, 89, 98, 169–171 Angiogenic vessel-targeting liposomes, 169–176 peptide development, biopanning method, 171 DAS model mice, study, 171 histochemical analysis using biotinylated peptide, 171 Angiographically occult, see Cerebral cavernous malformation (CCM) Ang, K. K., 48 Ansari, S. A, 143–153 Anti-angiogenic drugs, 54 Antiangiogenic targeting strategy antineovascular therapy for various tumors, 171–173

Index

APRPG-Lip, ability of drug delivery to angiogenic vessels, 171 APRPG-LipADM injected in Meth A sarcoma/Colon26 carcinoma-bearing mice, study, 171–172 dual-targeting strategy, 172–173 long-circulating liposomes, therapeutic efficacy, 172 treatment of pancreas cancer, 172 concept of ANET advantages in treatment of cancer, 170 liposomal anti-cancer agents, therapeutic effect, 170 VEGFR, 170 isolation of angiogenic vessel-targeting peptide, 170–171 peptide development, biopanning method, 171 Antibody, 255 Antiepileptic drugs (AEDs) choice of optimal AEDs, 318–320 medication, see Antiepileptic medication side effects of AEDs, 318–319 Antiepileptic medication first-line medication lamotrigine, 320 leveticaetam, 319 valproic acid, 319–320 second-line medication alternative monotherapy, 320 immediate or differed add-on therapy, 320 topiramate, 320 zonisamide, 320 Antineovascular therapy (ANET), 170–173 APRPG-Lip, ability of drug delivery to angiogenic vessels, 171 APRPG-LipADM injected in carcinoma-bearing mice, study, 171–172 dual-targeting strategy, 172–173 long-circulating liposomes, therapeutic efficacy, 172 treatment of pancreas cancer, 172 Antitumour treatments, epilepsy, 320–321 chemotherapy temozolomide/nitrosoureas, effects, 320–321 cranial radiation, 321 neurosurgery, 320 Antoch, G., 132 Antonio, E., 270

Index

Anzai, Y., 231 Aoyama, H., 57, 221, 223–224, 282 Apoptosis, 2, 25f, 26–27, 89, 99, 110–112, 172, 187–188, 190, 254, 287, 289–290, 290f Apoptosis-inducible genes, 287, 289–291 APRPG, see Ala-Pro-Arg-Pro-Gly (APRPG) Arber, D. A., 64 Arbour, R., 231 Armelin, H. A., 111, 114 Armelin, M. C., 111, 114 Arnold, S. D., 297 Arnold, S. M., 279 231 Array experiment examination of endoglin in MDA-MB-231 cancer cells, 90 examination of MMPs in MDA-MB-231 cancer cells, 89–90 examination of MSG in cancer cells, 87–89 gene profiling of MDA-MB-231 breast cancer cells, 86–87 differentially expressed genes in brain/bone-selective clones, 87, 87f Arredondo, N., 280 Arterial blood volume (aBV), 141 Arterial spin labeling (ASL), 135–141 Arthurs, B. J., 217–225 Asai, T., 171 Asbell, S., 225 ASL evaluation of perfusion in brain tumors 3D-FSE-pCASL, 136–138 3D FSE imaging sequence, 137f Dai’s method, 137 assessment of aBV, 141 drawbacks low perfusion signal, 140 perfusion measurements in WM, 140 PASL/CASL approach, 136 perfusion analysis, 138–139 DSC-MRI maps, 138f results, pCASL vs. DSC-MRI quantitative perfusion analysis, 139 visual assessment of susceptibility artifacts, 139 visual scoring of tumor, 139 statistical analysis, 139 vs. DSC-MRI measurement, 140 Assoian, R. K., 114 Astrocytic tumors, 2, 114–115, 122, 187

347

Astrocytomas, 2, 11, 13, 23, 27, 110–111, 114–115, 115f, 118f, 122, 124, 138, 163, 178, 180f, 182f, 183, 187, 238, 244–246, 270, 296, 298, 306, 308, 314, 321 Aubry, J. F., 229 Auchter, R. M., 281 Awad, I. A., 143–153 B Bacchetti, S., 27 Badie, B., 262 Bahadori, M., 63 Bailey, S., 81 Balkwill, F. R., 79 Balloon-tipped catheter, 212f, 213 Baltes, M. M., 327 Bampoe, J., 297 Bao, S., 29 Barker, F. G., 166 Barnholtz-Sloan, J. S., 53, 79, 97, 270 Barthel index (BI), 304 Barth, R. F., 111 Bartsch, R., 49–50 Bastin, K. T., 273 Basu, S., 240 Batra, S. K., 251–257 Bauman, M. A., 212 Bayer Pharmaceuticals Corporation, 79 BBB, see Blood-brain barrier (BBB) BCSFB, see Blood-CSF barrier (BCSFB) Beach, S. R., 326 Beato, M., 111 Beers, R., 256 Begg, C. F., 273 Behin, A., 22, 237 Beier, D., 27 Bekker, A., 334 Bénard, F., 239 Bendell, J. C., 98 Benedetto, N., 37 Benign tumors, 37, 63, 131, 185 Benz, M. R., 127, 132 Ben-Zur, H., 327 Bereaved caregivers, 326 Bergenheim, A. T., 339–342 Berger, M. S., 237 Besse, B., 54

348

Beta-emitting radiotracers, 239–240 32 P-labeled sodium phosphate, 239–240 positron emitters, 240 18 F-FDG, intraoperative localization of tumor tissue, 240 18 F-FDOPA, 240 18 F-FET, 240 radiophosphorus, 239 Beta probes, 242–243, 243f, 246–247 Beyer, T., 131 Bhanot, Y., 244, 247 Bhargava, P., 41 Bhatt, S., 63 Bickenbach, K. A., 289 Bidros, D. S., 208 Bieche, I., 89 Bikeye, S. N., 253–254 Billing, P. S., 44 Binello, E., 36 Biopanning method, 171 Birindelli, A., 128 BI, see Barthel index (BI) Blackburn, E. H., 23 Black, P., 271 Blagosklonny, M. V., 114 Bland, J. M., 139, 141 Bleomycin bleomycin-Fe2+ complex, 200 clinical experience in brain, 201–202 entry into cell, action mechanism, 201 produced from Streptomyces verticillus, 200 toxicity, 201 in treatment of electrochemotherapy, efficacy, 201 treatment of lymphoma/testicular cancer, 201 Blood-brain barrier (BBB), 48, 55, 76, 78–80, 86, 90, 98–99, 103, 110, 117–120, 123, 140, 176, 203–204, 207–208, 232–234, 239, 244 Blood-CSF barrier (BCSFB), 78 Bobo, R. H., 208–212 Boccardo, F., 100 Bogalhas, F., 243, 247–248 Bohmer, R. M., 114 Bokemeyer, C., 201 Bone metastases, 85–86 Bonfanti, L., 24 Bonnet, D., 23 Bonn, V. E., 23 Boogerd, W., 50 Borgelt, B., 47–48, 221, 225

Index

Borg, M., 319 Boron neutron capture, 2 Bosma, I., 307–308 Bottomley, A., 303 Bourne, H. M., 336 Boxerman, J. L., 140 Boyle-Mariottes’s law, 341 Bradley, W. G. Jr., 151–152 Brainard, J. A., 178 Brain Cancer Module, 298–299 Brain metastases from breast cancer neuroimaging techniques, 47 prognostic factors, 50–51 patient-related factors, 50 recursive partitioning analysis (survival rates), 51 treatment-related factors, 50–51 treatments chemotherapy, 48–50 corticosteroids, 47 radiosurgery, 48 surgical resection, 48 trastuzumab use, 50 WBRT, 47–48 Brain metastases in RCC patients clinical presentation brain scans, importance, 54 signs/symptoms, 54 comparison of SRS and surgery, 56–57 multimodality treatment of brain metastases, 58–59 palliative treatment approaches blood-brain barrier, 55 treatment with steroids, 55 VEGF pathway/mTOR inhibitors, 55 patients with CNS involvement, 56 radiosurgery GK/SRS, limitation, 56 Japanese study of treatment with GK/SRS, 56 use of heavy particles as radiation source, 56 recommendations, 59–60 CNS control, improvements in treatment approaches, 59–60 multidisciplinary management of lesions, 59 periodic screening, 59 therapeutic challenges, 59 screening, 54–55 immunotherapy/anti-angiogenic drug treatment, risk, 54

Index

immunotherapy and targeted therapy agents, 55 SRS plus immediate WBRT vs. deferred WBRT, 57–58 phase II study (ECOG 6397), 57–58 SRS alone/with WBRT, randomized trials, 57 WBRT followed by RTOG, randomized trials, 57 surgical treatment neurosurgery, 55 stereotactic surgical approaches, 55 survival rate, 56 WBRT, 55–56 complications, 56 RPA, survival rates, 55–56 and surgical resection, survival rates, 56 Brain metastases, treatment with electrochemotherapy, 195 clinical challenges, 197–199 patient’s quality of life and neurocognitive function, 198 single/multiple brain metastases, prognosis, 198 WBRT treatment, efficacy, 198 incidence/survival, 195 treatment results in skin, 198f Brainshift, 93 Brain tumor classification using MRS automatic classification challenges, 11 Machine Learning methods, 11 automatic classification based on MR spectra, 12–13 interpretations, 16–17 one-class classification strategy, 13 performance estimation (ROC-AUC) by diagnosis, 13t bi-directional Kohonen network, 12 classification in children/adults compatibility of 3 T MRS with 1.5 T-based classifiers, 14 incremental learning in brain tumor classification, 14 computerized decision support systems for diagnosis, 17 ensemble models, 12 evaluation strategies of BT automatic classification, 15–16 scatter plot of performance, eTUMOUR/INTERPRET cases, 15f

349

feature extraction, 8 for brain tumor classification based on MR spectra, 10–11 pattern recognition approach, 8 PCA, 9 peak integration (PI), 8–9 quantification approach, 8 Relieff feature selection, 9 stepwise algorithm for feature selection in classification, 9 functional data analysis, 9–10 Gaussian parametric model, 11–12 ICA, 10 KNN, 12 machine learning approach design of BT classification with MRS, 6f objectives, 5–6 multilayer perceptron, 12 preprocessing MRS, 7–8 multicenter INTERPRET and eTUMOUR projects, 7 training phase and recognition phase, 6 wavelet transform and multi-resolution analysis, 10 Brain tumor development, role of EGFRvIII functions of wild-type EGFR and EGFRvIII mutant, 252–253 activation of intracellular signaling pathways, 252 intracellular signaling elements, 252, 253f role in angiogenesis, 253 upregulation of gene expression, 252–253 implications in treatment resistance, 253–254 inefficacy of therapies for treating medulloblastomas/gliomas genetic and epigenetic alterations, cause, 251–252 novel therapeutic strategies EGFRvIII mutant-based immunotherapies and vaccination, 256–257 molecular targeting of EGFR, clinical trials, 255–256 targeting EGFR using specific antibodies, 255 targeting EGFR using tyrosine kinase activity inhibitors, 254–255 schematic structure of human wild-type EGFR protein and EGFRvIII mutant, 252f

350

Brain tumor diagnosis using PET PET probe for liposome labeling, development, 173–174 LEH labeling in brain ischemia model rat, 174 phase 0 study, 173 SophT method, 173, 174f strategy of antiangiogenic targeting antineovascular therapy for various tumors, 171–173 concept of antineovascular therapy, 170 isolation of angiogenic vessel-targeting peptide, 170–171 tumor angiogenesis and targeting using liposomes, 169–170 angiogenic vessels, characteristics, 169 EPR effect, 169–170 liposome activity in angiogenic vessels, 169–170 pro-angiogenic factors, 169 tumor diagnosis with the targeted liposomes, 174–176 APRPG-PEG-Lip intravenous injection in glioma rat models, 174–175, 175f novel [18 F]-labeling method, advantages, 176 PET imaging with [18 F]SteP2-APRPG-PEG-Lip and [18 F]FDG, 175–176, 175f tumor imaging by PET development of positron PET probe for liposome labeling, 173–174 tumor diagnosis by PET with the targeted liposomes, 174–176 Brain tumors causes, 2 diagnosis, 3 prognosis, 1, 3 radiotherapy/chemotherapy, post-effects, 1 functional deficits, 1 high dose, causes, 1 survival rates 5-year relative survival rates, 1 poor rates, reasons, 1 symptoms, 3 therapeutic approaches, see Therapies for CNS tumors tumor grading Grade I/II/III/IV, 2–3 types, 2

Index

Brain tumors, cellular immortality in brain tumourigenesis and tumour heterogeneity, 23 CD133, role in tumour initiation, 23 characteristics of intracranial neoplasms, 21–22 developmental neurobiology multipotent NSCs and neurogenesis in CNS, 24–26 neuroepithelial cells, radial glia and early embryogenesis, 24 role of telomeres/telomerase during neurogenesis, 26 stem cell theory in postnatal neurogenesis and cancer, 25f multidisciplinary approach to a new synthesis, 23–24 neuro-developmental biology germinative layers in adult brain, 22 NSCs, 22 radial glial cells, 22 new therapeutic outlook, 29–31 targeted therapy, 29–30, 30f tumour resistance mechanisms, 30–31, 30f telomere/telomerase status, normal/dysregulated neurogenesis, 22–23 and tumour initiation brain tumour stem cell paradigm, 27–29 telomeres and telomerase in brain tumors, 26–27 Brain tumors, diagnostic impact of PET using radiolabelled amino acids amino acid PET 18 F-labelled amino acids, 119 perspectives, 124–125 clincial applications of amino acid PET diagnostic assessment of recurrent tumors, 123 differential diagnosis of intracranial tumors, 120–121 extent of tumor for biopsy and treatment planning, 119–120 glioma grading and prognosis, 121–122 imaging brain tumors in children, 124 treatment monitoring, 123–124 FDG-PET, 118 MRI diagnosis imaging with standard T1-weighted sequences, 117–118 T1-weighted sequences with FLAIR, 118 radiolabelled amino acids, properties, 119 MET/FET uptake by glioma cells, effects, 119

Index

Brain tumors, drug delivery conventional methods, treatment of brain tumors, see Convection-enhanced delivery (CED) intracranial drug delivery, 208 local delivery approaches CED, 208 polymer implantation into brain, 208 regional drug approach, 208 Brain tumors, evaluation of perfusion ASL 3D-FSE-pCASL, 136–138 PASL/CASL approach, 136 perfusion analysis, 138–139 results, pCASL vs. DSC-MRI, 139 statistical analysis, 139 CBF and CBV, perfusion parameters, 136 cerebral perfusion, 135 neovascularization/hypervascularity in gliomas, assessment CT, 136 gadolinium-based contrast agents, risk of renal failure, 136 MRI perfusion techniques, 136 PET, 135–136 Brain tumors, MRgFUS (noninvasive) treatment BBB, disruption to targeted drug delivery, 232–234 laser-based techniques, 233 MRgLIFU technology, 233–234 site specific and temporal opening of BBB, impact, 233 brain tumor ablation with tcMRgFUS, 231–232 initial results, 232 intervention procedure, 232 patient preparation, 232 focused ultrasound technology and clinical applications drawbacks/solutions, 229 magnetic resonance guided brain-interventions with HIFU, 230 principles, 229–230 future approach for brain tumor therapy, 233f, 234 HIFU parameters, 228 system for open skull human brain surgery, 228f through craniectomy, trackless surgery, 227 new generation of FUS devices, limitations, 227 See also Focused ultrasound surgery (FUS) Brain tumors, nosologic imaging using MRI/MRSI analysis of the patient data, 162–163, 161t, 162f

351

classification of the abnormal region determination of the type of abnormal tissue, 161 LS-SVMs/KLR, pattern recognition methods, 161 experimental data, 156–158 classes of pathology, 157 MR image acquisition/registration, 157 normal/tumor MR spectra and NAA peaks, 156–157 preprocessing of MRSI, 157 quality control/tumor type determination, 156 voxel selection, considerations, 156 water-suppressed spectrum, pattern recognition, 157, 158f nosologic image, construction, 155–156 CCA analysis of MRSI data, 156 high-resolution nosologic images, scheme benefits, 156 subject-specific abnormal tissue prior, 163 two-step segmentation-classification framework Prastawa’s method, stages, 159 See also MRI segmentation Brain tumors (primary/secondary), electrochemotherapy for BBB barrier to penetration of anti-cancer drugs, 203–204 sink effect, 204 WBRT effects on, 204 brain metastases, 195 clinical challenges, 197–199 incidence/survival, 198 choice of chemotherapy, 200–201 bleomycin, 200–201 concept, 202 intravenous/direct administration of bleomycin, 202 electroporation, 196f, 199 affecting parameters, 199–200 and the transmembrane potential, 199–200 preclinical experience in the brain, 202–203, 204f treatment of small cutaneous metastases electrochemotherapy procedure, 197f response rates, 195–197 Brain tumors, role of miRNAs in glioblastoma diagnostic use of miRNAs, 186–188

352

Brain tumors, role of miRNAs in (cont.) incidence/survival, 185 therapeutic use of miRNAs, 188 medulloblastoma diagnostic use of miRNAs, 189–190 incidence/survival, 185 therapeutic use of miRNAs, 190 role in tumor pathogenesis, 186 Brain tumour stem cell, 24, 25f, 27–30, 30f Brandsma, D., 123 Breast cancer brain metastases 231 array experiment examination of endoglin in MDA-MB-231 cancer cells, 90 examination of MMPs in MDA-MB-231 cancer cells, 89–90 examination of MSG in cancer cells, 87–89 gene profiling of MDA-MB-231 breast cancer cells, 86–87 brain as a special microenvironment, 86 breast cancer metastases to neurocranium, 85 bone metastases, 85 dural metastases, morphologies, 85 intraparenchymal brain metastases, 85 epidemiology, 85–86 incidence/survival rates, 85 leptomeningeal dissemination, 86 neurosurgical therapy adjuvant treatment, 94 epidemiology, 90–91 imaging, 91–92 location of brain metastases, 91 planning surgical treatment: neuronavigation, 93 primary tumor diagnosis and detection of brain metastases, 90–91 prognostic factors and survival, 93 solitary/singular/multiple brain metastases, 91 surgical treatment, 93 surgical treatment, complications, 94 surgical treatment, optional tools, 93–94 symptoms and findings, 91 treatment decision/indication for surgery, 92 seed and soil hypothesis (Paget), 86 molecular mechanisms, 86 selective MBC cells, 86 Breast cancer and RCC metastases to the brain BBB physiology, 78 claudin expression, role in cancer, 78

Index

inter endothelial junctions, 78 “Rule of Five,” 78 sanctuary to treatment, 76 breast cancer Her-2 positive tumors, risk analysis, 80 highest affinity for leptomeninges, 79 myriad intracellular signals, 80 risk factors for development of brain metastases, 80 treatment with trastuzumab, 80 chemokines, 78–79 role in tumor proliferation, metastases and angiogenesis, 78–79 factors involved in breast cancer metastases, 76f factors involved in RCC metastases, 76f glial microenvironment, 78 imaging, 76–77 findings for cerebral metastases, 77 metastases vs. primary neoplasms, 77 MRI, 77 MR spectroscopy, 77 PET, 77 incidence rates, 75 metastatic breast/renal carcinoma, MRI images, 76f RCC, 80–81 IFN-α/IL2 therapies, 80–81 mTOR inhibitors, role, 81 schema of VHL pathogenesis, 81 signs and symptoms clinical presentation, determinants of, 76 headache and focal neurological deficit, 76 hemorrhage, 76 seed-soil paradigm of metastases (Paget), 76 systemic therapies/imaging techniques, effects, 75 tumor biology, 77–78 trans-endothelial migration of cancer cells, 78 VEGF, 79 EDGF inhibitors, 79 lapatinib, treatment of metastatic breast and RCC, 79 tyrosine kinase inhibitors, 79 VEGF inhibitors, 79 Breast Cancer Study Group, 80 Brennum, J., 198 Bret, P., 36–38, 40–42 Bria, E., 99 Brodtkorb, E., 317 Brookes, M. J., 141

Index

Brown, J. M., 304, 306 Brown, P. D., 306–307, 309 Brunereau, L., 147 Brunn, H., 63 Brunton, L., 334 Bukowski, R. M., 59 Burchenal, J. H., 93 Burger, P. C., 179 Burstein, H. J., 98 C Cairncross, G., 270 Calatozzolo, C., 317 Cameron, D., 100 Campbell, L. V. Jr., 35–36, 44 CAMs, see Cell adhesion molecules (CAMs) Cancer specific HRQOL tools EORTC BN20, 304 FACT, 304 Cancer stem cell, 23, 27–30, 30f Cann, B. J., 326 Cannulation technique, 264–266, 265f–266f Canonical correlation analysis (CCA), 156, 161, 163, 165 Capecitabine, 97–104 Capecitabine and lapatinib therapy, 97–104 Capillary telangiectasia, 145–146, 148, 153 Capri, G., 100, 104 Carcinoembryonic antigen gene promoter, 288 Caregiver, see Familial caregivers of patients with brain cancer Caregiving, 323 impact on QOL of caregivers, 323 psychological distress, 324 positive/negative effects on patients, 323 self-efficacy, 324 Carneiro Filho, O., 244, 246 Caroli, E., 36–38, 40–41, 44 Carpentier, A. C., 237 Carrau, R. L., 262 Carr, H. Y., 150 Carson, K. A., 270 Carstensen, L. L., 327 CASL, see Continuous ASL (CASL) Catheter, 208–215, 212f, 341 Caverns, 143 CBF, see Cerebral blood flow (CBF)

353

CBTRUS, see Central Brain Tumor Registry of the United States (CBTRUS) CBV, see Cerebral blood volume (CBV) CCA, see Canonical correlation analysis (CCA) CCM, clinical imaging of diagnostic methods angiography, 144 CT, 144 high resolution MRI, 144 gradient echo sequences, 146–147 multiple MRI sequences of patients with temporal lobe seizures, 146f with high field MRI, 148–149 imaging features on conventional MRI sequences, 144–146 CCM lesions, types based on MR signal, 145 changes in appearance of solitary CCM with different MRI sequences, 145f hemosiderin ring, 144 T1 and T2 weighted imaging, 144 role of contrast enhanced MRI, 146 gadolinium-enhanced studies, 146 SWI case with known familial CCM disease, 147, 147f case with solitary sporadic CCM, 147, 148f detection of type 4 CCM lesions, criteria, 147 CCM detection by MRI familial/sporadic CCM, characteristics, 143–144 future directions, 153 imaging, see Clinical imaging of CCMs; Experimental imaging of CCMs risk of hemorrhage, 143 CCM lesions, types, 145 Type 1 lesions, 145 Type 2 lesions, 145 Type 3 lesions, 145 Type 4 lesions, 145 CCM, see Cerebral cavernous malformation (CCM) CDNA, 87, 110–111, 113–115, 188, 289f CDSSs, see Clinical decision support systems (CDSSs) Cella, D. F., 299 Cell adhesion molecules (CAMs), 78 Cellular immortality in brain tumors brain tumourigenesis and tumour heterogeneity, 23 CD133, role in tumour initiation, 23 characteristics of intracranial neoplasms, 21–22 developmental neurobiology

354

Cellular immortality in brain tumors (cont.) multipotent NSCs and neurogenesis in CNS, 24–26 neuroepithelial cells, radial glia and early embryogenesis, 24 role of telomeres/telomerase during neurogenesis, 26 stem cell theory in postnatal neurogenesis and cancer, 25f multidisciplinary approach to a new synthesis, 23–24 neuro-developmental biology germinative layers in adult brain, 22 NSCs, 22 radial glial cells, 22 new therapeutic outlook, 29–31 targeted therapy, 29–30, 30f tumour resistance mechanisms, 30–31, 30f telomere/telomerase status, normal/dysregulated neurogenesis, 22–23 and tumour initiation brain tumour stem cell paradigm, 27–29 telomeres and telomerase in brain tumors, 26–27 Central Brain Tumor Registry of the United States (CBTRUS), 1, 185 Central nervous system (CNS), 21, 39, 48, 63–64, 68, 75–76, 97–104, 114, 135, 177–184, 207, 246, 251, 261, 273, 334 Cerebral blood flow (CBF), 136 Cerebral blood volume (CBV), 136 Cerebral cavernous malformation (CCM), 143–153 See also CCM, clinical imaging of; CCM detection by MRI; CCM lesions, types Cerebral metabolism, 341 Cerebral metastases, 75, 77, 80, 279–284 See also SRS for cerebral metastases of digestive tract tumors Cerebral perfusion, 135 Cerebrospinal fluid (CSF), 78, 80, 93, 156, 207, 214, 263, 280, 342 Chalk, J., 319 Chambers, P. W., 35, 40–41 Chang, E. L., 47, 58, 281–282 Chang, J., 80 Chang, S. D., 275–277 Chan, J. A., 186–187 Chan, Y. F., 65–66 Chan, Y. P., 341–342

Index

Chawla, S., 140 Chemokines, 39, 78–79 Chemotherapy, 3, 30, 44, 48–51, 49t, 55, 67t, 68–69, 79, 93–94, 103f, 117, 123, 129, 132–133, 135, 170, 172, 185–186, 188, 195–205, 207, 233, 233f, 235, 237, 253, 261, 270, 288–289, 296–298, 303–304, 307–309, 313, 317, 319–321, 324 See also Electrochemotherapy for primary/secondary brain tumors; Radiochemotherapy Cheng, A., 26 Chen, J. H., 327 Chen, S.H., 287 Chen, W., 118, 174, 240 Chetty, R., 63, 70 Child wavelet, 10 Chochinov, H. M., 326 Choi, B. D., 256–257 Ciafrè, S. A., 186 Cifone, M. A., 114 Clark, A. W., 35–45 Claudins, 78 Clayton, A. J., 98–99 Clement, G. T., 230–231 Clinical Anxiety Scale, 299 Clinical decision support systems (CDSSs), 6, 17–18 CADS, 17 CURIAM-BT, 17 HEALTHAGENTS distributed CDSS, 17 SV INTERPRET GUI, 17 Clinical imaging of CCMs diagnostic methods angiography, 144 CT, 144 high resolution MRI, 144 gradient echo sequences, 146–147 multiple MRI sequences of patients with temporal lobe seizures, 146f with high field MRI, 148–149 imaging features on conventional MRI sequences, 144–146 CCM lesions, types based on MR signal, 145 changes in appearance of solitary CCM with different MRI sequences, 145f hemosiderin ring, 144 T1 and T2 weighted imaging, 144

Index

role of contrast enhanced MRI, 146 gadolinium-enhanced studies, 146 SWI case with known familial CCM disease, 148, 147f case with solitary sporadic CCM, 148, 148f detection of type 4 CCM lesions, criteria, 147 CNS lesions, 57–59, 100, 184 See also CNS lesions, frozen section evaluation CNS lesions, frozen section evaluation approach to frozen section, 179–182 algorithmic approach, 179 identification of lesions as gliomas, 181–182, 183f nonneoplastic process, findings, 181, 182f patient history, importance, 179 telepathology, issues, 181–182 tissue assessment/reserve, 179–180, 180f frozen section discrepancies, 182–184 categories of discrepancy, 183 in nonneoplastic CNS samples, study, 184 overgrading tumors, problems, 183–184 frozen section vs. cytologic preparations, 179, 180f factors influencing the method of choice, 179 intraoperative consultation, 177–179 communication between neurosurgeon and pathologist, importance, 177–178 goal, 177–178 guidance of surgical approach and management, example, 178 stereotactic biopsy procedures, diagnostic yield evaluation, 178 triaging of tissue, FISH/PCR analysis, 178 CNS lymphomas, 2, 68, 178–179, 314, 314f CNS, see Central nervous system (CNS) CNS tumors in women with capecitabine/lapatinib therapy for breast cancer clinical data, 99–100 HER2 targeted strategies, treatment of CNS disease, 99 current trials, 100–102 ALTTO trials, 102 CEREBRAL trial, 101 COMPLETE trial, 102 LANTERN study, 100–101 the future HER2+ disease, preventive treatments, 104 pre-clinical data, 99

355

combination of lapatinib with trastuzumab, effects, 99 mice treated with lapatinib, study/results, 99 treatment of CNS metastases HER2+ patients treated with lapatinib, effects, 98 HER2+ patients treated with trastuzumab, effects, 98 HER2 overexpression, risk factors, 98 lantern trial flow diagram, 102f lapatinib plus capecitabine vs. trastuzumab plus capecitabine, CEREBRAL flow diagram, 103f response to treatment with lapatinib/ capecitabine in LEAP study, 101f, 98–99 SRS, 98 surgery, treatment of choice, 98 WBRT, 98 Coffey, R. J., 275–276 Coffin, C. M., 64–65 Cognition, 307, 309 Cognitive functioning, 295, 297, 304, 307–309, 324–325 Cohen-Inbar, O., 271 Cohen-Solal, C., 51 Colbassani, H. J., 178 Collision tumor, 36 Colorectal cancer, 57, 91–92, 110, 112, 217, 238, 240, 280, 335 Computerized tomography (CT), 3, 41, 44, 47, 49t, 54, 65, 68–69, 91–93, 119–120, 123, 127–133, 136, 140, 144, 153, 195, 222, 230, 232, 244–245, 264, 270, 280, 340f Constans, J. P., 85 Conti, A., 187 Continuous ASL (CASL), 136 See also Pseudo-continuous arterial spin labeling (pCASL) Continuous positive airway pressure (CPAP), 340 Convection, 208 Convection-enhanced delivery (CED), 207–215, 211f BBB, prevention of drug delivery into CNS, 207 catheter placement, 208 convective/diffusive transport, 208–209 diffused-based techniques, limitations, 208 in human brain, 209f infusate formulation, 213–214

356

Convection-enhanced delivery (CED) (cont.) limitations, 209–210 parameters affecting therapeutic efficacy, 208 catheter design, 212–213, 212f catheter placement, 210 infusion volume and rate, 210–211 patient-specific simulation and real-time imaging, 214 Cooke, B. S., 217–225 Cortez, M. A., 187 Corticosteroids, 47, 55–56, 59, 111, 218, 307, 336 Coulter, J. A., 289 Cox-2 inhibitors, 331, 334–336 Cox, J. D., 48 CPAP, see Continuous positive airway pressure (CPAP) Cranial radiation, 321 Croteau, D., 165 Cruz, L. C. Jr., 149 CSF, see Cerebrospinal fluid (CSF) CT-PET, 132–133, 133f CT, see Computerized tomography (CT) Culine, S., 54 Cuppone, F., 104 Cushing, H., 81 Customary dosing, 219 Cyclotron-based proton beam, 218, 222 Cylindrical retractors, 263 Cytologic preparation, 177, 179–180, 180f D Dagash, H., 63 Daghighian, F., 242–243 D’Agostino, A. N., 128 Dai, W., 137 Dammers, R., 239, 246 D’Andrea, G., 271 Dandy, W., 262 D’Angelo, C., 295–300 D’Angelo, V. A., 295–300 Da Silva, A. N., 282–283 Datta, R., 288 R DaunoXome , 170 Davie, J. R., 110 Davies, N., 13 Dawood, S., 99 Deangelis, L. M., 280 De Angelis, V., 98

Index

Debies, M. T., 88–89 Debinski, W., 208 Decision support systems, see Clinical decision support systems (CDSSs) Deen, H. G. Jr., 36 Degaki, T. L., 109–115 De Haes, J., 299 Delattre, J. Y., 217–218 Demakas, J. J., 217–225 Demasi, M. A. A., 109–115 De Raedt, T., 128 De Robles, P., 270 De Salles, A. A., 277 De Souza, J. M., 147 Detre, J. A., 137 Developmental venous anomaly (DVA), 144 De Vita, V., 204 Devos, A., 11–12, 156–157, 165 Dexamethasone, 55, 100, 112, 309, 313, 319 Dick, J. E., 23 Di Costanzo, A., 166 DiDivitiis, E., 178 Diffusion, 208 Digestive tract tumors, SRS for, 279–284 cerebral metastases, 279–280 clinical symptoms, 279–280 hematogenous dissemination, 280 prognostic factors, 280 gastrointestinal cancer, 280–281 cancer types in G.I. tract, 280 cerebral metastasis from gastric carcinoma, 280 colorectal cancer, 280 digestive system malignancies, 280 esophageal carcinoma, risk factors, 281 global incidence, 280 hepatocellular carcinoma, 280 median survival time, 280 pancreatic cancer, 280 G.I. tract origin, 282–284 efficacy of SRS, assessment, 282 local tumor control, MRI study, 283, 284f median survival times, 283 patient characteristics for study analysis, 282 post-SRS survival, prognostic factors, 283 preoperative symptoms/post-operative complications, 283 single/multiple cerebral metastases from colorectal carcinoma, study, 282

Index

principles of SRS, 281–282 acute/chronic complications, examples, 281–282 dose planning, RTOG study, 281 principal objectives, 281 steep radiation dosing, 281 treatment of radionecrosis, approaches, 281 Dirks, P. B., 27 Disability, 303 Disseminated tumor cells (DTCs), 39 DiStefano, A., 48, 50 DiStefano, D., 179 Dittman, R., 341 Dogra, V. S., 63 Donor tumors, 36, 37t, 40 Donzelli, R., 85 Doorn, P. F., 129 Dou, T., 187 R Doxil , 170 Doxorubicin (DOX), 67t, 129, 172, 200, 213 DSC-MRI, see Dynamic susceptibility contrast MRI (DSC-MRI) Dual-targeting liposomes, 173 Dual-targeting strategy, 172–173 See also Antiangiogenic targeting strategy Dube, V. E., 274 Ducatman, B. S., 130 Duenas-Gonzalez, A., 319 Dufour, H., 274 Dupont, S., 313–321 Dural metastases, morphologies, 85 DVA, see Developmental venous anomaly (DVA) Dynamic contrast enhanced (DCE) MRI, 136 Dynamic susceptibility contrast MRI (DSC-MRI), 136, 138–140 E Ebstein-Barr Virus (EBV), 2, 64 EBV, see Ebstein-Barr Virus (EBV) ECOG, 57, 270 EDGF, see Epidermal growth factor (EDGF) Efficace, F., 303 EGFR, see Epidermal growth factor receptor (EGFR) EGFRvIII in brain tumor development functions of wild-type EGFR and EGFRvIII mutant, 252–253 activation of intracellular signaling pathways, 252 intracellular signaling elements, 252, 253f

357

role in angiogenesis, 253 upregulation of gene expression, 252–253 implications in treatment resistance, 253–254 inefficacy of therapies for treating medulloblastomas/gliomas genetic and epigenetic alterations, cause, 251–252 novel therapeutic strategies EGFRvIII mutant-based immunotherapies and vaccination, 256–257 molecular targeting of EGFR, clinical trials, 255–256 targeting EGFR using specific antibodies, 255 targeting EGFR using tyrosine kinase activity inhibitors, 254–255 schematic structure of human wild-type EGFR protein and EGFRvIII mutant, 252f EGFRvIII mutant-based immunotherapies, 256–257 Eichler, A. F., 76, 80, 98 Eirik, H., 269–271 Ekenel, M., 99 Elaimy, A. L., 217–225 Electrochemotherapy for primary/secondary brain tumors BBB barrier to penetration of anti-cancer drugs, 203–204 sink effect, 204 WBRT effects on, 204 brain metastases, 195 clinical challenges, 197–199 incidence/survival, 195 choice of chemotherapy, 200–201 bleomycin, 200–201 concept, 202 intravenous/direct administration of bleomycin, 202 electroporation, 196f, 199 affecting parameters, 199–200 and the transmembrane potential, 199 preclinical experience in the brain, 202–203, 204f treatment of small cutaneous metastases electrochemotherapy procedure, 197f response rates, 197 Electropermeabilization, see Electroporation Electroporation, 195, 196f, 199–200 affecting parameters electrode device, 200

358

Electroporation (cont.) electroporation based drug and gene delivery, advantage, 200 pulse parameters, 200 tissue parameters, 200 and transmembrane potential, 199 and external electric field, relation/implications, 199 irreversible/reversible electroporation, 199 See also Electrochemotherapy for primary/ secondary brain tumors Ellison, D. W., 39–40, 44 Embolization, 35, 274 Embryonal tumors, 2, 21 ENC-1/PIG10, see Nuclear restrict protein in brain (NRP/B) Endoglin (CD105), 87, 90 Endoscope-assisted microsurgery, 262 Endoscopic port surgery (EPS), 261–267 See also EPS for intraparenchymal brain tumors Endoscopic ventriculostomy, 262 Engelhardt, B., 78 Engh, J. A., 261–267 Enhanced permeability and retention (EPR), 170 En plaque meningiomas, 245 Ensemble models, 12 Entschladen, F., 40 Enzyme-inducing/inhibiting antiepileptic drugs, 319 EORTC, see European Organization for Research and Treatment of Cancer (EORTC) Ependymal tumors, 2 Ependymoma, 13, 22–23, 27, 37t, 178 Epidermal growth factor (EDGF), 79–81, 251–257 Epidermal growth factor receptor (EGFR), 79, 98–99, 110, 170, 251–257 Epidermal growth factor receptor Erb-2 (or Her-2/neu), 80 Epilepsy, 201, 296, 307, 313–321, 342 Epilepsy and brain tumours and antiepileptic drugs clinical manifestations, 315 seizures in brain tumours, epidemiology, 314–315 frequency of, influencing factors, 314–315, 314f–315f severity of epilepsy/epileptogenesis of brain tumours host factors (genetic/lower efficacy of AEDs), 316f, 317 tumour and peritumoural factors, 316, 316f

Index

therapeutic issues, see Therapeutic issues in epileptic patients EPR, see Enhanced permeability and retention (EPR) EPS for intraparenchymal brain tumors brain tumor demographics and challenges, 261 tumor visualization by operating microscope vs. rod-lens endoscope, 262f cylindrical retractors for brain surgery, 263 endoscopic brain surgery, 262–263 endoscope-assisted microsurgery, 262 intraventricular endoscopic operations, 262 pituitary surgery, endonasal approach, 262 treatment of hydrocephalus (neuroendoscopy), 262 tumoral visualization, 262–263 EPS application to intraparenchymal lesions, 263 philosophy of EPS, 265–267 EPS, inside-out approach, 265 extracapsular dissection, 265 hemostasis, 265 intra-tumoral cannulation, 265 operating microscope to neurosurgery vs. EPS, 266 port design, considerations, 267 resection of peri-ventricular high grade glioma by EPS, case, 266, 266f technique operative planning and surgical technique, 264–265 patient selection, 263–264 EPS, see Endoscopic port surgery (EPS) EPS technique operative planning and surgical technique cannulation technique, 264, 265f cortisectomy following cannulation, 264 post-operative care, 264–265 pre-operative evaluation by MRI/CT, 264 suction devices, gentle tumoral aspiration, 264 surgical head positioning, 264 patient selection, 263–264 candidates/radiographic features favorable for EPS, 264 ERBB2 and muc1 promoters, 288 Ernst, A., 187 Ernst, S., 118 Ernst-Stecken, A., 308 Esophageal carcinoma, 41–42, 43f, 281

Index

Estrogen and progesterone receptor (ER/PR), 104 eTUMOUR projects, 7, 12–13, 15, 15f, 17 European Organization for Research and Treatment of Cancer (EORTC), 270, 298–299, 304, 305t, 308–309 Evaluation (frozen section) of CNS lesions approach to frozen section, 179–182 algorithmic approach, 179 identification of lesions as gliomas, 181–182, 183f nonneoplastic process, findings, 181, 182f patient history, importance, 179 telepathology, issues, 181–182 tissue assessment/reserve, 179–181, 180f frozen section discrepancies, 182–184 categories of discrepancy, 183 in nonneoplastic CNS samples, study, 184 overgrading tumors, problems, 183–184 frozen section vs. cytologic preparations, 179, 180f factors influencing the method of choice, 179 intraoperative consultation, 177–179 communication between neurosurgeon and pathologist, importance, 178 goal, 177–178 guidance of surgical approach and management, example, 178 stereotactic biopsy procedures, diagnostic yield evaluation, 178 triaging of tissue, FISH/PCR analysis, 179 Evans, A. J., 80 Everolimus, 81 Ewing, J., 38, 76 Ewing’s theory, 38, 76 Exablate 4000, 230, 231f, 234 Experimental imaging of CCMs excised human lesions, 149–151 measurement of T1 and T2 relaxation times, 150–151 MRI and confocal microscopy images, 150f murine models, 151–153, 152f Extracranial to intracranial metastasis, 35–45 metastasis, 55, 57–59, 91–94, 98, 198, 217–218, 221, 276, 278, 282–284 organs, 55 F FACT-Br, see Functional assessment of cancer therapy-brain (FACT-Br)

359

FACT-G, see General FACT module (FACT-G) FACT, see Functional assessment of cancer therapy (FACT) Fairbanks, R. K., 217–225 Familial caregivers of patients with brain cancer caregiving impact on QOL of caregivers, 323 positive/negative effects on patients, 323 future directions, 327–328 informal cancer care by family, factors, 323 QOL of, see QOL of familial caregivers for brain cancer patients Fan, Q. W., 253, 256 Fast fourier transform (FFT), 7 3D Fast spin echo (FSE), 135–141 FDG, see Fluorodeoxyglucose (FDG) 18-FDG PET, 127–133 Feiden, W., 178 Feng, X., 188 Ferner, R. E., 133 Ferrante, L., 271 Ferrara, N., 169 Ferreira, S. H., 334 Ferrell, B. R., 324 Ferretti, E., 189–190 α-Fetoprotein, 288 FFT, see Fast fourier transform (FFT) Fiandaca, M. S., 212–214 Fibromatosis, 63 See also Neurofibromatosis Fidler, I. J., 38–39 FID, see Free induction decay (FID) Fink, M., 230 FISH, see Florescence in situ hybridization (FISH) Fitzgerald, D. P., 78 Flash light effect, 266 Fletcher, C. D., 63 Flickinger, J. C., 222, 279–281 FLIC, see Functional living index – cancer (FLIC) Floeth, F., 117–125 Florescence in situ hybridization (FISH), 179 Fluid-attenuated inversion recovery (FLAIR), 118 Fluorescent tracers, 238 Fluorodeoxyglucose (FDG), 118–119, 124, 131–133, 173–175, 175f, 240–241, 243, 248 Focused ultrasound surgery (FUS), 227–235 advantages/limitations, 234–235 brain tumor ablation with tcMRgFUS, 231–232 initial results, 232

360

Focused ultrasound surgery (cont.) intervention procedure, 232 patient preparation, 232 magnetic resonance guided brain-interventions with HIFU, 230 concept of ideal surgery, 230 Exablate 4000 (transcranial MRgFUS device), 230, 231f, 234 principles, 229–230 beam focusing and optimum parameter selection, 229–230 MRgFUS, applications, 229 Fokas, E., 35, 38–40 Folkman, S., 79, 323 Fonsatti, E., 87, 90 Forsthoefel, K. F., 103 Fountas, K. N., 273 Foy, P., 318 Fraggetta, F., 37 Frank, D., 127–128, 130–131 Freedman, V. H., 114 Free induction decay (FID), 7 Freeman, R., 150 Frenay, M., 321 Friedman, H. D., 280 Fries, G., 262 Fromm, J. A., 253 Frozen section, see CNS lesions, frozen section evaluation Fry, F. J., 228–229 Fry, W. J., 228–229 Fuchs, I. B., 99 Fujita, M., 111 Fukazawa, T., 288 Functional assessment of cancer therapy-brain (FACT-Br), 299, 304, 305t, 309 Functional assessment of cancer therapy (FACT), 299, 304, 309 Functional living index – cancer (FLIC), 297 Fung, L. K., 208 Furnari, F. B., 251–252 Furuya, M., 169 FUS, see Focused ultrasound surgery (FUS) G Gabellini, N., 317 Gadolinium-based contrast agents, 136, 140

Index

Gage, F. H., 24 Galanaud, D., 165 Galanis, E., 273, 275–276 Gal, H., 188 Galli, R., 23 Gamma-emitting radiotracers 111 Indium-labeled somatostatin analogues, expression in meningiomas, 239 99m Tc-MIBI, application, 239 Gamma-knife (GK) system, 56 Gamma knife surgery (GKS), 56, 218, 223, 231, 273–278, 281, 283, 284f Gamma probes, 240, 242, 244–246, 248 Ganslandt, O., 165 Garcia, D. M., 138 García-Gómez, J. M., 5–18 Garczarek, U. M., 9 Garret, R., 273 Garzia, L., 190 Gaspar, L. E., 55, 218, 225, 280 Gastrointestinal cancer, 280–281 Gastro intestinal (GI), 270, 279–284, 334 See also Digestive tract tumors, SRS for Gate-control theory, see Pain theory Gault, J., 144, 146 Gaussian parametric model, 11 Gavrilovic, I. T., 75 Gay, E., 239, 245 GBM, see Glioblastoma multiforme (GBM) Gehl, J., 195–205 Gehl Videbæk, K., 197 Gelblum, D. Y., 281 General FACT module (FACT-G), 299, 304, 309 Gene therapy, 2, 213, 287–291 Germanidis, G., 68 Germ cell tumors, 2 Germinative layers in adult brain SGZ, 22 SVZ, 22 Geyer, C., 100 Giesinger, J. M., 296 Gilbar, O., 327 Gill, D., 150 Gimbel, J. S., 334 Giovagnoli, A. R., 296–297, 306–307 Given, B. A., 323–325, 327 GKS, see Gamma knife surgery (GKS) Glantz, M., 317 Glaser, R., 324, 327

Index

Glial cells (radial), 22, 24, 25f Glial fibrillary acidic protein (GFAP), 24, 25f, 26 Glioblastoma diagnostic use of miRNAs, 186–188 array study with 756 miRNAs, 187 miRNA microarray analysis, 189 patient outcomes/survival, study, 187–188 up-/downregulation of miRNAs, study, 186–187 therapeutic use of miRNAs chemotherapy/irradiation, 188 miR-21 overexpression in glioblastoma cells, study effects, 188 mir-451, role, 188 treatment with siRNA and chemotherapeutics, 188 treatment with TMZ, 188 Glioblastoma multiforme (GBM), 22–23, 27, 135, 187–188, 204, 231–232, 251, 296, 306–308, 319, 341 Gliomas, see Glioblastoma Glucocorticoid, 55, 59, 111–112, 282 Goellner, J. R., 273–274 Goldberg, C. R., 342 Golding, S. E., 254 Goldmann, S., 119 Goldspiel, B. R., 79 Goodhead, D. T., 219 Gopinath, S. P., 341 Gori, S., 98, 198–199 Gothelf, A., 200–201 Govender, D., 63, 70 Govindaraju, V., 157 Gradient recalled echo (GRE), 140, 145–151, 146f, 148f, 153 Graesslin, O., 104 Grandinetti, C. A., 79 Grant, J. A., 262 Grant, R., 303 Greenberg, M. S., 280 Greene, H. S. N., 38 Greider, C. W., 23 Greig, N. H., 48 Greiner, C., 65–66, 68 Griffero, F., 254, 256 Griffin, C. A., 64 Griffiths-Jones, S., 186 Gril, B., 99 Grobmyer, S. R., 127–131 Gross, C. G., 24

361

Grossi, P. M., 80, 103 Grosu, A. L., 120 Grundy, R. G., 21–31 Grunert, P., 178 Grunfeld, E., 326 Guan, Y., 187 Guha, A., 127–133 Gulati, S., 63–71 Gulec, S. A., 239, 241, 244 Guo, D., 256 Gupta, D., 306 Guthkelch, A. N., 235 Guthrie, B. L., 273–275 H HACE, see High altitude cerebral oedema (HACE) Hadaczek, P., 215 Hagedoorn, M., 325 Hagenstad, C. T., 64 Hakyemez, B., 77 Halatsch, M. E., 251–252, 254–255 Hallahan, D. E., 288–289 Hallahan, A. R., 28 Hamilton, J. F., 213 Hamilton rating scale for depression, 299 Hammarsund, M., 111 Hammoud, M. A., 279–280 Handicap, 303 Hankel-Lanczos singular value decomposition (HLSVD), 7, 157 Harada, Y., 53 Haran, R. P., 340 Harms, J. F., 87–88 Harris, A. E., 263 Hartfuss, E., 24 Hart, I. R., 288 Hart, M. G., 303, 309 Hartyánszky, I. L., 63 Hasegawa, T., 282–283 Haselsberger, K., 339, 341 Hawkey, C. J., 334 Hayat, M. A., 1–3 Hayflick, L., 23 Hazard, L. J., 217–220, 222 Hazuka, M. B., 199 Health care, 295–296, 304, 326–327, 339 Health, definition (WHO), 295 Health related QoL (HRQoL), 295–300, 303–310 See also HRQoL in patients with HGG

362

Heat-inducible promoters, 288 Heat shock protein (hsp), 288 Hebert, R. S., 327 Heerschap, A., 12, 155–166 Heimans, J. J., 297, 299 Heimburger, R. F., 228–229 Heinemann, V., 50, 80 Heldwein, 79, 81 Heller, R., 197, 201–202 Heller, S., 241 Hellwig, D., 2 Helseth, E., 269–271 Hemangiopericytomas (HPC), 273–278 See also Intracranial HPC, GKS for Hematogenous dissemination, 65, 280 Hemmati, H. D., 23 Hemosiderin ring, 144 Hemostasis, 264–265, 334 Hepatocellular carcinoma, 280 HER2 disease, preventive treatments, 103–104 capecitabine monotherapy, 99 EGF100151 study capecitabine monotherapy vs. combination of capecitabine and lapatinib, 100 lapatinib monotherapy, phase II study, 99–100 LEAP/EGF100151 study, clinical efficacy, 100 LEAP study, 100 PCI in SCLC, 104 PI3-kinase, AKT, mTOR and HSP90 inhibitors, 103–104 Herholz, K., 120, 124 Herman, T. E., 63 Hernandez, M. C., 110, 115 Herscovitch, P., 138 He, S., 110 HGG, see High-grade gliomas (HGG) HHV-8, see Human herpesvirus 8 (HHV-8) Hicks, D. G., 80 HIFU, see High intensity focused ultrasound (HIFU) High altitude cerebral oedema (HACE), 340, 342 High-grade gliomas (HGG), 77, 118, 122, 140, 237, 240, 246, 248, 266f, 297, 299, 303–310, 314–316 High intensity focused ultrasound (HIFU), 227, 229 Hildebrand, J., 314–316, 318 Hill, K. L. Jr., 79 Hill, H. D. W., 150 Hirschfeld, A., 342 Histiocytosis, 68, 70

Index

HLSVD, see Hankel-Lanczos singular value decomposition (HLSVD) Hochberg, F. H., 98 Hodes, R., 288 Hoffman-La Roche, 80 Hojman, P., 199 Holland, E. C., 110 Ho, M., 160 Hong, J. F., 271 Horak, C. E., 88 Horneff, G., 70 Horner’s syndrome, 274 Horska, A., 18 Ho, S., 160 Houston, C. S., 340 HPC, see Hemangiopericytomas (HPC) HRE, see Hypoxia response element (HRE) HRQoL in patients with HGG assessment, a patient-reported outcome measure, 304–306 administrative failure, cause for missing data, 306 approaches to minimize loss of data, 306 cancer specific HRQOL tools, 304 content of EORTC BN20, 304, 305t content of EORTC QLQ c30 version 3.0, 304, 305t content of FACT-Br version 4, 304, 305t HRQOL, definition, 304 patient and proxy measures (self-report), 304 effect of treatments on HRQoL effect of chemotherapy on quality of life, 308–309 effect of radiotherapy on quality of life, 307–308 effect of supportive treatment, 309 effect of surgery on quality of life, 307 outcome measures in glioma research, 303–304 impairment/disability/handicap, distinction and evaluation, 303–304 radiochemotherapy, risk of toxicity, 303 value of HRQoL in daily practice, 309–310 vs. cognitive functioning, 307 HRQoL, see Health related QoL (HRQoL) Hruban, R. H., 129–130 Hsp, see Heat shock protein (hsp) HTERT, see Human telomerase reverse transcriptase (hTERT) Hubbard, R. C., 334

Index

Huber, B. E., 288 Hudes, G., 81 Huffel, S. V., 155–166 Hughes, D., 331–336 Human cytomegalovirus, 2 Human epithelial growth factor receptor (HER), 40, 50, 99–104, 102f Human herpesvirus 8 (HHV-8), 64 Human telomerase reverse transcriptase (hTERT), 26–27, 28f, 288 Hung, M. C., 80 Huse, J. T., 110 Hwang, S., 314 Hwang, S. W., 57 Hynynen, K., 229–231, 234 Hypovascular cancer, see Pancreatic cancer Hypoxia response element (HRE), 288 I Iagaru, A., 131 IARC, see International Agency for Research on Cancer (IARC) ICA, see Independent component analysis (ICA) ICP, see Intracranial pressure (ICP) Ideal surgery, concept, 230 Idema, A. J., 155–166 ILS, see Independent living score (ILS) Imaging, see individual entries Immediate or differed add-on therapy, 320 Immunotherapy, 2, 54–55, 59–60, 257 See also Radioimmunotherapy Impairment, 303 IMT in lung and brain, coexsistence clinical features cranial imaging findings, 65 intracranial inflammatory myofibroblastic tumors, 65 pulmonary and brain inflammatory myofibroblastic tumor, cases, 65, 66t–67t systemic features, 64–65 differential diagnosis CNS tumors, 68 plasmacytoma, 68 tuberculomas, 68 etiopathogenesis HHV-8/EBV, role, 64 neoplastic nature of myofibroblastic tumors, 64

363

histopathology, 63–64 inflammatory pseudotumor, 64 inflammatory tumors, subgroups, 64 illustrative case, 68–71 CECT thorax showing lesions, 69f photomicrographs showing fascicles of spindle cells, 69f treatment, recommendations, 70–71 inflammatory cells/myofibroblastic spindle cells, 63 inflammatory myofibrohistiocytic proliferation, 63 myofibroblastic tumors, classification, 63 sites of occurrence, 63 treatment and prognosis combination of methotrexate and 6-mercaptopurine therapy, 68 complete surgical resection, 68 immunomodulation and combination chemotherapy, 68 Independent component analysis (ICA), 8, 10–11 Independent living score (ILS), 299 Indications/patient counseling for SRS impact of various histologic subtypes on melanoma/RCC (radioresitant), 222 on patients who underwent GK radiosurgery, 222 patient selection, 218–219 clinical outcomes of SRS, inluencing factors, 218 RTOG partitioning analysis, criteria, 218 SRS/WBRT/SRS and WBRT, recommendations, 218–219 radiosurgery with WBRT for multiple brain metastasis, 220–221 radiosurgery with WBRT for single brain metastasis, 220 surgical resection or radiosurgery alone, 221–222 retrospective study of patients with SRS and WBRT/SRS alone, 221 risk of local and distant tumor control, 221 study of patients with SRS and WBRT/SRS alone/WBRT alone, 221–222 surgical resection with WBRT for single brain metastasis, 219–220 randomized study of patients with WBRT alone/WBRT plus surgery, 219–220 WBRT for single/multiple brain metastases acute side-effects, 219

364

Indications/patient counseling for SRS (cont.) altered dose/customary dosing, outcomes, 219 long-term side effects, 219 Inducible nitric oxide synthase (iNOS), 153, 289 Infantile hydrocephalus, 262 Inflammation, 64–65, 177, 181, 182f, 316, 334 Inflammatory cells, 63–64, 68, 69f, 119, 121, 131, 143, 153, 182f Inflammatory myofibroblastic tumor (IMT), 63–71 Inflammatory myofibrohistiocytic proliferation, 63 Inflammatory pseudotumor, 63–64 Infusate, 208, 210–214 Infusion rate, 208, 210–212, 214 iNOS, see Inducible nitric oxide synthase (iNOS) Inoue, T., 213 International Agency for Research on Cancer (IARC), 185 International Breast Cancer Study Group (IBCSG), 80 International Commission on Radiological Protection (ICRP), 247–248 Intracranial air, 340–342, 340f Intracranial HPC, GKS for conventional radiotherapy, management of, 274–275 Mira, Payne and Guthrie’s study, findings, 275 post-surgical treatment of HPC, 274 purpose of radiation therapy, 274 radiation dose, prevention of recurrence, 274–275 radiation therapy before/after resection, study, 275 HPC and angioblastic meningioma, similarity, 273 HPC recurrence, 273 role of SRS in management of, 275–278, 275t Gamma surgery, disadvantage, 277 lower tumor control rate by GKS, study (Olson), 276 marginal dose (≥14 Gy) and survival rates, 276–277 radiation associated complications, 277 reduction of tumor size by GKS, study (Sheehan), 276 steep dose gradient, efficacy, 275 surgical resection of, 274 incidence of local recurrence, 274 incidence of metastasis, 274 preoperative embolization, control of bleeding, 274

Index

Intracranial inflammatory myofibroblastic tumors, 65 Intracranial neoplasms, 21–22 ependymoma, 22 GBM, 22 medulloblastoma, 21–22 sPNET, 21 Intracranial pressure (ICP), 40, 54, 65, 91, 210, 218, 340–342 Intracranial tumors, 35–45, 36, 47, 91, 120, 135, 269–271, 276, 299–300 See also Surgery for intracranial tumors in elderly patients Intraoperative consultation, 177–179 goal, 177–178 guidance of surgical approach and management, example, 178 stereotactic biopsy procedures, diagnostic yield evaluation, 178 triaging of tissue, FISH/PCR analysis, 179 Intraparenchymal brain tumors, EPS for brain tumor demographics and challenges, 261 tumor visualization by operating microscope vs. rod-lens endoscope, 262f cylindrical retractors for brain surgery, 263 endoscopic brain surgery, 262–263 endoscope-assisted microsurgery, 262 intraventricular endoscopic operations, 262 pituitary surgery, endonasal approach, 262 treatment of hydrocephalus (neuroendoscopy), 262 tumoral visualization, 262–263 EPS application to intraparenchymal lesions, 263 philosophy of EPS, 265–267 EPS, inside-out approach, 265 extracapsular dissection, 265 hemostasis, 265 intra-tumoral cannulation, 265 operating microscope to neurosurgery vs. EPS, 266 port design, considerations, 267 resection of peri-ventricular high grade glioma by EPS, case, 266, 266f technique operative planning and surgical technique, 264–265 patient selection, 263–264 Intraventricular endoscopic operations, 262 Irreversible/reversible electroporation, 199

Index

Ishihara, Y., 230 Ishiuchi, S., 47–51 Iversen, H. K., 195–205 Iwamoto, F. M., 270 J Jaaskelainen, J., 274–275 Jackson, R. J., 166 Jacques, S., 263 Jager, P. L., 118 Jain, R. K., 208 Jansen, E. P., 118 Jarnum, H., 135–140 Jaroszeski, M. J., 200, 202 Jarrell, S. T., 35–38, 40, 42, 44 Jawahar, A., 223 Jeba, J., 65, 67 Jefferson, G., 246 Jemal, A., 53, 80 Jho, H. D., 263 Jiang, J., 204 Johannesen, T. B., 269 Johns, T. G., 253, 255 Johnston, S., 97–104 Jolesz, F. A., 227–235 Jung, Y., 140 K Kaal, E. C., 219, 309 Kano, H., 275–277 Kaplan, C. P., 296 Kaplan, K. J., 181 Kapoor, A., 81 Karabatsou, K., 127, 132–133 Karnak, I., 64, 68 Karnofsky, D. A., 93 Karnofsky performance score (KPS), 50, 55, 93, 218–224, 280, 282–283, 297, 304, 308 Karsmakers, P., 161 Kasad, 48 Kasahara, N., 289 Kaschten, B., 122 Kassam, A. B., 262–263 Kato, T., 79 Kaur, B., 111 Kaus, M., 155 Kawashita, Y., 290 Keime-Guibert, F., 308

365

Kelly, Dr. P., 263 Kelly, P. J., 263 Kelm, B. M., 156, 166 Kernel logistic regression (KLR), 161 Kershaw, T., 326 Khan, J., 87 Khan, R., 321 Kiecolt-Glaser, J. K., 324, 327 Kim, H., 187 Kim, H. S., 140 Kim, K., 254 Kim, S. G., 111 Kim, S. Y., 140 Kim, T. A., 110–111, 114–115 Kim, W. Y., 39 Kim, Y., 323–328 King, A. A., 128 Kinoshita, M., 234–235 Kinoshita, Y., 18 Kinzie, J. J., 199 Kirsch, D. G., 98, 219 Kiss1, 87–88 Klapper, W., 26 Kleihues, P., 117, 135, 178 Klein, C. A., 39 Klein, M., 297, 306–307, 309 Kleinschmidt-DeMasters, B. K., 179, 181 Klose algorithm, 7 KLR, see Kernel logistic regression (KLR) K-nearest neighbors (KNN), 9, 11–12, 14 Knutsson, L., 135–141 Kohonen network, 11–12 Koide, H., 173 Koirala, R., 63 Kojima, M., 63 Kojima, T., 245, 247 Komata, T., 288 Kondo, M., 173 Kondziolka, D., 57, 220, 223–224 Konecny, G. E., 99 KPS, see Karnofsky performance score (KPS) Krauze, M. T., 210–211 Kripke, M. L., 38 Kuhn, M. J., 195 Kuijlen, J., 318 Kunwar, S., 209–210, 213 Kurimoto, M., 270 Kuroda, K., 230 Kurtz, J. M., 48

366

Kuyken, W., 295 Kvale, P. A., 44 L Labauge, P., 143, 146–147 Lagwinski, N., 181 Laine, A.-L., 207–215 Lai, P., 79–80 Lai, R., 79–80 Lal, A., 253–255 Lam, J. S., 54 Lammering, G., 252, 254 Lamoreaux, W. T., 217–225 Lamotrigine, 320 Langen, K.-J., 117–125 LANTERN flow diagram, 102f Lapatinib, 79, 97–104, 100 Lapatinib expanded access program (LEAP) study, 99–100, 101f Lapidot, T., 23 Larsen, B. A., 195–205 Larsson, E.-M., 135–141 Larsson, H. B., 136 LASA, see Linear analogue scale assessment (LASA) Laudadio, T., 155–166 Laufman, L. R., 103 Lauria, A., 242, 248 Law, M., 77 Laws, E. R. Jr., 36 Lawson, S. L., 63 Lazennec, 79 Lazo Chabner, B., 201 LDA, see Linear discriminant analysis (LDA) Least squares support vector machines (LS-SVMs), 11–12, 161, 163 Leblanc, G. G., 143 Lee, B. C., 147 Lee, C. M., 217–225 Lee, H. Y., 39 Lee, J., 314 Leenders, K. L., 140 Lee, S., 318 Leese, P. T., 334 Lee, S. I., 153 Lee, S. S., 50, 79 Lee, S. Y., 153 Lehmann, P., 140 LEH, see Liposome encapsulated hemoglobin (LEH) Leksell, D. G., 56

Index

Lele, P., 227 Le Marc’hadour, F., 65–66, 68 Leppert, W., 336 Le Roux, A. A., 127–133 Le Scodan, R., 51 Lesser, G. J., 219 Leston, J. M., 246–247 Leveticaetam, 319 Levy, D. A., 54 Leyland-Jones, B., 98 LGG, see Low-grade gliomas (LGG) Liang, X. Q., 110–112, 114–115 Li, B., 221, 223–224 Lieberman, 210 Liebow, A. A., 63 Ligand-toxin conjugate administration, 2 Li, J., 198 Li, K. K., 189 Lindau, A., 81 Lindgren, T., 340 Lindvall, P., 339–342 Linear accelerator (LINAC) based treatment, 56, 218, 220–222, 275t, 281 Linear analogue scale assessment (LASA), 299, 309 Linear discriminant analysis (LDA), 11–12, 14, 17, 159, 160f Linnert, M., 195–205 Lin, N. U., 79, 99 Linskey, M. E., 218–219, 221, 223–224 Liposomal anti-cancer agents, 170 Liposomal DDS technology, see Antineovascular therapy (ANET) Liposome encapsulated hemoglobin (LEH), 174 Liposome labeling, 173–174 Liposomes, see Angiogenic vessel-targeting liposomes; Liposomal anti-cancer agents; Liposome labeling Liu, H. L., 234 Liu, W., 189 Li, X., 165, 320 Li, Y., 188–189 Local control rate, 48, 56, 221, 223, 282–283 Locke, D. E., 295, 299 Loeffler, J. S., 47, 54, 219 Loew, S., 254 Lo, H. W., 252, 254 Long TE spectra, 7, 11, 16 Look-Locker echo-planar imaging, 141

Index

Lo, S., 47 Loudyi, A., 53–60 Louis, D. N., 114, 166 Low-grade gliomas (LGG), 122, 124, 140, 187, 237, 246, 297, 314–316 LS-SVMs, see Least squares support vector machines (LS-SVMs) Luciferase gene, 290 Luczak, J., 336 Lu, H., 137 LUISE software, 157 Lu, J.-Q., 35–45 Lukas, L., 11 Lustig, R. A., 118, 123 Luts, J., 7–8, 11, 13, 155–166 Lutterbach, J., 224, 270 Lu, Y., 190 Lwu, S., 318 Lyden, D., 38–39, 45 Lymphoma, 1–2, 13, 37t, 64, 68, 135, 138, 178–179, 181, 183, 201, 314 Lynn, J. G., 227 M Maas, N., 89 Macedo, A. F. A., 112 MacKay, J. A., 214 MacKay, A. R., 217–225 Mackworth, N., 299 Macroadenoma, 271 Maeda, N., 171–173 Maes, F., 157 Magnetic resonance guided focused ultrasound surgery (MRgFUS), 229, 231f, 234–235 See also Transcranial MR-image guided focused ultrasound surgery (tcMRgFUS) Magnetic resonance spectroscopic imaging (MRSI), 7–8, 11, 13, 155–166 Magnetic resonance spectroscopy (MRS), 5–18, 41, 77, 121, 155 Magnetic resonance thermometry (MRT), 230 Maguire, P., 324 Mahabir, R. C., 341–342 Mahmood, F., 195–205 Majos, C., 6, 11, 18 Makino, K., 65 Malhotra, V., 65–66 Malignant brain tumors, role of radioresponsive gene therapy

367

apoptosis-inducible genes application in radio-inducible gene therapies, 289–291 in treatment of cancer, viral promoters used, 287 radio-inducible gene therapies, 288–289 Ad.EGR-TNF, antitumor effects, 289 Egr-1 promoter, expression of tumor-sensitive genes, 288, 289f external factor-mediated gene therapies, 288 pE9, radiosensitization of hypoxic tumor cells, 289 radiation treatment, 288 radioinducible iNOS gene therapy approach, 289 targeted gene therapies, 287–288 ability to increase transfection rates, challenge, 287 cell-specific promoter to target tumors, examples, 288 Malignant gliomas, 2, 42, 110, 114, 122, 124, 135, 185, 187, 233–234, 237, 254, 296–297, 318 Malignant neurofibromas, 131 Malignant peripheral nerve sheath tumors (MPNST) 18-FDG PET, 131–132 controversies in PET imaging, 133 CT-PET in MPNST, usefulness, 132 intra-tumoral heterogeneity/MPNST, distinguishment, 131–132 management strategy based on CT-PET, 133f non-cancer specificity, limitation, 132 PET vs. normal imaging, 131 definition, 127 imaging modalities, 131 benign or malignant neurofibromas, features, 131 CT scans, 131 MRI for soft tissue tumors, 131 macroscopic, microscopic and cytogenetic features, 130–131 NF1 neurofibromin, TSG, 128 risk of malignant transformation, 128 specific mutations in, 128 presentation, treatment and prognosis, 128–130 chemotherapy, efficacy, 129 NF1 and non-NF1 MPNSt patients, outcomes, 129

368

Malignant peripheral nerve sheath tumors (cont.) prognostic factors, 129 radical surgical excision of the nerve, 128 radiotherapy regimes, 128 site of occurrence, 128 surgical margins, detrminant of survival, 129–130 symptoms, 128 risk factors, 127 Malzkorn, B., 187 Manne, S. L., 325 Manon, R., 57 Maor, M. H., 56 Maraire, J. N., 143–145 Marchan, E. M., 273–278 Mardor, Y., 213 Mariani, G., 240–242 Marik, J., 173 Markham, J. W., 340 Martens, T., 253, 255 Martin, E., 227–235 Martin, S. E., 37, 40, 42 Marty, M., 197, 201–202 Maschio, M., 320 Masferrer, J. L., 334, 336 Masola, V., 317 Mathieu, D., 57 Matrix metalloproteinases (MMPs), 87, 89, 253f Matsubara, O., 64 Matsuda, S., 61 Matsumura, A., 287–291 Matthew, E., 141 Matthiessen, L. W., 197, 202 Mauer, M., 309 Mauer, M. E., 309 Maurer, J., 185–190 MBC, see Metastatic breast cancer (MBC) McCarthy, H. O., 289 McClain, K. L., 68 McCormick, P. C., 45 McDannold, N., 232–233, 235 McKnight, T. R., 166 MDA-MB-231, 85–87, 87f, 89–90, 99, 114–115 Medulloblastoma, 21–22, 251 diagnostic use of miRNAs, 189–190 Hedgehog signaling pathway, role, 189 miRNA downregulation patterns, study, 189–190

Index

patient outcomes/survival, 190 therapeutic use of miRNAs, 190 Mehdorn, H. M., 85 Mehta, M., 56 Mehta, M. P., 273, 281 Meiboom, S., 150 Mei, J., 234 Meling, T., 269–271 Mellinghoff, I. K., 255 Mellon, S., 325–326 Melssen, W., 12 Melzack, R., 332 Melzak, 331 Memorial Sloan-Kettering Cancer Center, 75 Mena, H., 273–274 Menard, L., 237–249 Mendes, O., 90 Meningeal tumors, 2 Menze, B. H., 8, 11 Meric, F., 98 Meric-Bernstam, 80 Metastasis genes, classes, 40 Metastasis suppressor genes (MSG), 40, 87–89 Metastatic breast cancer (MBC), 50, 78, 85–90, 97–104, 198 Metastatic niche model, 39 Meyer, F., 320 Meyers, C. A., 198 Meyers, R., 229 Micallef, J., 253 Microdialysis, 341–342 MicroRNA, see MiRNAs in brain tumors, clinical role Miles, F. L., 78 Miller, K. D., 79 Miller, S. J., 128 Mimeault, M., 251–257 Miner, M. E., 296 Minniti, G., 270–271 Mira, J. G., 275 MiRBase, 186 Mirijello, A., 295–300 MiRNAs in brain tumors, clinical role glioblastoma diagnostic use of miRNAs, 186–188 incidence/survival, 185 therapeutic use of miRNAs, 188 medulloblastoma diagnostic use of miRNAs, 189–190 incidence/survival, 185

Index

therapeutic use of miRNAs, 190 role in tumor pathogenesis, 186 Misra, A., 208 Miura, F. K., 111 Mixed gliomas, 2, 122 MMPs, see Matrix Metalloproteinases (MMPs) Models of metastasis linear progression model, 39 parallel progression model, 39 Modified ranking handicap scale (MRHS), 304 Moncada, S., 334 Moorhead, P. S., 23 Moots, P., 314, 316 Moots, P. L., 298 Morantz, R. A., 201 Mordhorst, C., 331, 336 Mori, T., 22 Mori, Y., 56, 281 Morley, T. P., 246 Mother wavelet, 10 Motzer, R. J., 59, 81 MPNST, see Malignant peripheral nerve sheath tumors (MPNST) MRgFUS (noninvasive) treatment for brain tumors BBB, disruption to targeted drug delivery, 232–234 laser-based techniques, 233 MRgLIFU technology, 233–234 site specific and temporal opening of BBB, impact, 233 brain tumor ablation with tcMRgFUS, 231–232 initial results, 232 intervention procedure, 232 patient preparation, 232 focused ultrasound technology and clinical applications drawbacks/solutions, 229 magnetic resonance guided brain-interventions with HIFU, 230 principles, 229–230 future approach for brain tumor therapy, 233f, 234 HIFU parameters, 228 system for open skull human brain surgery, 228f through craniectomy, trackless surgery, 227 new generation of FUS devices, limitations, 227 See also Focused ultrasound surgery (FUS) MRgFUS, see Magnetic resonance guided focused ultrasound surgery (MRgFUS)

369

MRgLIFU, see MR-guided, low intensity focused ultrasound (MRgLIFU) MR-guided, low intensity focused ultrasound (MRgLIFU), 233 MRHS, see Modified ranking handicap scale (MRHS) MRI perfusion techniques, 136 MRI segmentation generating subject-specific abnormal tissue, 159, 160f results, 161 segmentation based on the modified atlas, 159–160 MRSI, see Magnetic resonance spectroscopic imaging (MRSI) MRS, see Magnetic resonance spectroscopy (MRS) MRT, see Magnetic resonance thermometry (MRT) MSG, see Metastasis suppressor genes (MSG) 99m Tc-MIBI, 239 MTOR inhibitors, 81 Muacevic, A., 223 Mukerji, N., 270 Mukherjee, B., 254 Mukherjee, D., 335 Mukherjee, J., 253 Muldoon, M. F., 295–296 Muller, A., 79 Multicenter evaluation study, 5–8, 13, 15–16, 18 Multilayer perceptron (MLP) model, 12 Multiple brain metastases, 48, 57, 91, 195, 198, 218–220 Murase, Y., 173 Murayama, S., 47–51 Murine models, imaging of CCMs, 151–153, 152f Murray, M. R., 273 Mycek, M. J., 334–336 Myofibroblastic spindle cells, 63 Myofibroblastic tumors, classification, 63 N Nagase, H., 87, 89 Nahaczewski, A. E., 111 Nakajima, Y., 63 Narla, L. D., 64–65, 68 Narla, S., 64–65, 68 National Cancer Institute, 86 National Human Genome Research Institute, 86 Neeves, K. B., 214 Nelson, J. S., 179 Neoplasms, 2, 21–23, 35–36, 64, 68, 77, 117, 144, 148f, 166, 180f, 181, 183f, 184, 280

370

Nephrogenic systemic fibrosis (NSF), 136 Neural stem cells (NSCs), 22–24, 25f, 27–30, 28f, 30f, 252 Neurocranium, 85 Neuroendoscopy, 262 Neuroepithelium, 24 Neurofibromatosis, 2, 127–128 Neurofibromin, 128 Neurogenesis, 22, 24, 25f, 26–29, 28f Neurological acute and late toxicity, 224 Neurologic functioning, 324 Neuronavigation, 93, 248 Neurosphere, 24 Neurosurgery, 55–56, 59, 219, 224, 229, 263, 266, 269, 317, 320, 334–335, 339 Neurosurgical therapy for breast cancer brain metastases adjuvant treatment, 94 epidemiology, 90 imaging MRI/CT, 91–92 location of brain metastases, 91 planning surgical treatment: neuronavigation, 93 primary tumor diagnosis and detection of brain metastases, 91 prognostic factors and survival, 93–94 KPS for evaluation of patient performance, 93 leptomeningeal carcinomatosis, poor prognosis, 93 solitary/singular/multiple brain metastases, 91 surgical treatment, 93 surgical treatment, complications, 94 surgical treatment, optional tools, 93–94 symptoms and findings, 91 treatment decision/indication for surgery, 92 Neuwelt, E. A., 204 Nezamzadeh, M., 140 NGR peptide, 172 Nguyen, D. X., 37–40 Nguyen, T., 280 Nieder, C., 222, 308 Nijboer, C., 324–325 Nitrosoureas, 321 Niwinska, A., 49–51 Noctor, S. C., 24 Nonsteroidal anti-inflammatory drugs (NSAIDs), 331, 334 Northcott, P. A., 189

Index

Northouse, L. L., 325–326 Nosologic image, 155 Nosologic imaging of brain tumors using MRI/MRSI analysis of the patient data, 162–163, 161t, 162f classification of the abnormal region determination of the type of abnormal tissue, 161 LS-SVMs/KLR, pattern recognition methods, 161 experimental data, 156–158 classes of pathology, 156 MR image acquisition/registration, 157 normal/tumor MR spectra and NAA peaks, 156 preprocessing of MRSI, 157 quality control/tumor type determination, 156 voxel selection, considerations, 156 water-suppressed spectrum, pattern recognition, 157, 158f nosologic image, construction, 155–156 CCA analysis of MRSI data, 156 high-resolution nosologic images, scheme benefits, 156 subject-specific abnormal tissue prior, 163–164 two-step segmentation-classification framework Prastawa’s method, stages, 159 See also MRI segmentation Notch and Sonic-hedgehog pathways, 28 Novartis Pharmaceuticals, 81 Novel [18 F]-labeling method, 176 NRP/B, see Nuclear restrict protein in brain (NRP/B) NSAIDs, see Nonsteroidal anti-inflammatory drugs (NSAIDs) NSF, see Nephrogenic systemic fibrosis (NSF) Nuclear restrict protein in brain (NRP/B), 109–115 gliomas, 109–110 genetic alterations during malignant progression, 109–110 grade I/II/III/IV tumors, 109 strategies to overcome molecular defects, 110 isolation by single pass sequencing technology, 110 member of BTB/Kelch repeat family, 110 nuclear matrix, role in tumor development, 110 rat brain tumor model, 111 cloning/functional characterization of the rat full length cDNA, 111–112 GC as anti-tumor agent, mechanisms, 111 rat C6 glioma cell, treatment with GC, 111

Index

retroviral-mediated transfer of NRP/B in C6/ST1 cells, effects, 112, 113f structural analysis, 110–111 BTB-domain, regulator of neuronal differentiation, 110 Nrf2-dependent NQO1 induction in cellular protection, mechanism, 111 NRP/B mutations in the kelch domain, 110–111 tumor suppressor expression in human cell lines and astrocytic tumors qRT-PCR analysis of NRP/B expression, 114–115, 115f use of cDNAs to identify NRP/B expression pattern, 114–115 as a tumor suppressor gene, role, 112–114 dexamethasone treatment of rat hippocampus, 112 p16/pRb/E2F pathway, 114 p53-induced genes, 112 tumor formation assays in nude mice, 113 Nussbaum, 76 O Oberholzer, M., 181 Oberndorfer, S., 319 Ochalski, P., 261–267 Ogawa, K., 47–51 Ogura, M., 50 Ohgaki, H., 117, 135 Oku, N., 169–176 Oligodendroglial tumors, 2 Olson, C., 273–277 Olson, J. J., 213 O’Malley, B. W., 287 Onar, A., 321 One-class classification strategy, 13 O’Neill, 56 Opiate receptors mu (μ), delta (δ), and kappa (κ) receptors, 333–334 role in pain management cyclooxygenase metabolism of arachidonic acid, 334f side effects, 333 Opstad, K. S., 8, 11 Orley, J., 295 Orlowski, S., 200, 202 Osaki, T., 288 Oslo University Hospital, 269 Osoba, D., 296, 298–299, 304, 306–309

371

Otsuki, T., 263 Ottawa Charter for Health Promotion, 295 Ottman, R., 317 Ouban, A., 78 Overstreet-Wadiche, L., 22 Owler, B. K., 140 Oxford Laboratories, 149 Ozsunar, Y., 140 Ozturk, E., 131 P P53-induced genes (PIG), 112 Pace, A., 321 Packer, R. J., 22 Padma, M. V., 118, 121 Paget, S., 35, 38–39, 45, 76, 86 Paik, S., 98 Pain fibers, 332–333 Pain management following craniotomy analgesics, treatment of postoperative pain cox-2 inhibitors, 331, 335 newer analgesics, benefits, 331–332 NSAIDs, 331, 334 opioids, 331, 334 steroids, 336 drug receptors opiate receptors, 333–334 failure to treat pain, effects, 331 pain pathways, 332–333 and trigeminal sensory, 333f pain receptors, 332 C fibers, 332 nerve endings, specializations, 332 A-δ fibers, 332 pain theory, 332 gate-control theory, feline nerve studies, 332 peripheral nociceptors, 332 treatment of chronic peripheral pain syndromes, 332 sources of postoperative pain, 331 subtemporal and suboccipital approaches, issues, 331 Pain receptors C fibers, 332 A-δ fibers, 332 Pain theory, 332 Palma, V., 29 Palmer, J. D., 334 Palmieri, D., 78–79

372

Pamphlett, R., 35–36, 41 Pancreatic cancer, 172, 280, 289 Pardee, A. B., 114 Pardridge, W. M., 207 Park, B. B., 49–50 Park, J. O., 288 Park, K. S., 280 Park, Y. H., 49–50, 98 PASL, see Pulsed ASL (PASL) Passirani, C., 207–215 Pastorino, F., 172 Patanaphan, V., 79, 280 Patchell, R. A., 48, 55–56, 218–219, 221, 223, 225, 279 Patel, D., 253, 255 Patel, A. S., 75–81 Pathogenesis, 22, 36–40, 45, 64, 80–81, 153, 186–187, 190 Pattern recognition (PR) methods, 6, 8–9, 11, 155, 157, 159, 161, 162, 166 Paulson, J. F., 274 Payne, B. R., 274–277 PCA, see Principal component analysis (PCA) pCASL, see Pseudo-continuous arterial spin labeling (pCASL) PCI, see Prophylactic cranial irradiation (PCI) PCR, see Polymerase chain reaction (PCR) Peacock, K. H., 219 Peak height of typical resonances (PPM), 8 Peak integration (PI), 8–9 Pearce, M. J., 327 PEG, see Polyethyleneglycol (PEG) Pellegrini, A. E., 63 Pelletier, G., 296–297, 306 Pelloski, C. E., 252 Peptide, 170–173, 176, 202, 253f, 256–257 Peretto, P., 24 Perfusion in brain tumors, ASL evaluation 3D-FSE-pCASL, 136–137 3D FSE imaging sequence, 137f Dai’s method, 137 assessment of aBV, 141 drawbacks low perfusion signal, 140 perfusion measurements in WM, 140 PASL/CASL approach, 136 perfusion analysis, 138–139 DSC-MRI maps, 138f results, pCASL vs. DSC-MRI

Index

quantitative perfusion analysis, 139 visual assessment of susceptibility artifacts, 139 visual scoring of tumor, 139 statistical analysis, 139 vs. DSC-MRI measurement, 140 Perifocal edema, 69, 70f Perin, A., 342 Perlstein, B., 213 Perneczky, A., 262 Pernot, M., 229 Perrone, F., 128 Pestalozzi, B. C., 80 Peter Dirks, 27 Petersen, E. T., 136, 139, 140 Peterson, E. W., 341 PET hot spots, 131–132 Petraki, C., 35–38 Petridis, A. K., 63, 65, 67 PET, see Positron emission tomography (PET) Piccirilli, M., 270 Pierson, J., 189 Piert, M., 239, 241–243 Piette, C., 112 PIG, see P53-induced genes (PIG) Pijnappel, W. W. F., 157 Pillay, V., 254–255 Pineal parenchymal tumors, 2 Pinker, K., 146, 149 Piroth, M. D., 123–124 Pirotte, B. J., 119, 122, 124 PI, see Peak integration (PI) Pitkethly, D. T., 273–274 Pitre, S., 242 Pituitary adenomas, 36, 37t, 38, 40–41, 44, 235, 270–271 Pituitary tumors, 2, 277 Plaat, B. E., 130 Plastic cells, 22 See also Neural stem cells (NSCs) Plathow, C., 288 Platini, C., 103 Plesec, T. P., 183–184 Plexiform neurofibromas, 127, 131 Plotkin, S. R., 76 Plummer, N. W., 151 PNET, see Primitive neuroectodermal tumor (PNET) Pneumocephalus, see Intracranial air Poddevin, B., 200 Pogue, B. W., 238

Index

Polli, J. W., 99 Polyak, K., 110, 112 Polyethyleneglycol (PEG), 169 Polymerase chain reaction (PCR), 87–89, 114–115, 179, 186, 252 Pöpperl, G., 122–123 Positron emission tomography (PET), 169–176 See also Amino acid PET in brain tumors, clincial applications Positron emitters, 173–175, 174f, 238, 240, 246–247 Posner, J. B., 75 Pouessel, D., 54 Povoski, S. P., 238–240, 248 Powell, S. Z., 36–38, 41, 44 Pozzati, E., 146 PPM, see Peak height of typical resonances (PPM) Prados, M. D., 255 Prastawa, M., 159–160, 163–164, 166 Prayson, R., 177–184 Preul, M. C., 5 Prevedello, D. M., 262 Price, S. J., 165 Primary brain tumors, 2 Primitive neuroectodermal tumor (PNET), 138 Principal component analysis (PCA), 8, 10, 166 Prophylactic cranial irradiation (PCI), 104 Prophylactic treatment, therapeutic issues in epilepsy general guidelines (AAN), 317–318 meta-analysis study, 317 perioperative period, 318 reality, 318 evidence-based guidelines and daily practice, discrepancies, 318 Proton magnetic resonance spectroscopy (1 H-MRS), 121, 124 Proust, F., 271 Przybylowska, K., 90 Psaila, B., 38–39, 45 Pseudo-continuous arterial spin labeling (pCASL), 135–141 Pseudoprogression, 123 Pseudotumor, 63–64 Puduvalli, V. K., 86 Pulsed ASL (PASL), 136 Purcell, E. M., 150 Q QALY, see Quality adjusted live years (QALY) QOL of familial caregivers for brain cancer patients

373

at the acute survivorship phase, 324 caregivers distress, characteristics/related factors, 324 caregivers, impact on their physical health, 324 location of the tumor, concern, 324 neurologic/cognitive functioning, management, 324–325 neuropsychiatric symptoms, management, 324 psychological distress/depression on caregivers, 324 spiritual aspect, 324 at the bereavement phase, 326–327 impact on health behaviors, 327 impact on psychological/physical health of caregivers, 326–327 at the mid– to long–term survivorship phase, 325–326 cancer recurrence and end-of-life care, 326 changes in caregiver’s physical/behavioral aspects, 325 spiritual concerns, 325 QOL, see Quality of life (QOL) Quality adjusted live years (QALY), 299 Quality of life (QOL) in brain tumors, 296–297 depressive symptoms/anxiety, impact on QoL, 297 cross-cultural domains of QoL, 295 definition (WHO), 295 glioma patients/lung cancer patients, healthy controls in, 297 health, definition (WHO), 295 health related QoL, 295–296 in high-grade malignant gliomas, Italian study, 297–298 FLIC, 297–298 HRQoL assessment by EORTC QLQ-C30, 298 QoL in glioma patients, affecting factors, 298 in low-grade glioma high/low-dose radiotherapy in glioma patients, study, 297 impact of radiotherapy on QoL, 297 QoL, effect on cognitive performance, 297 measurement of Brain Cancer Module, 299 Clinical Anxiety Scale, 299 EORTC QLQ-BN-20, 299 FACT, 299 Hamilton rating scale for depression, 299

374

Quality of life (QOL) (cont.) ILS, 299 LASA, 299 QoL and KPS, relation, 299–300 SF – 36, 299 SQLI/QALY, instruments, 299 in oncologic patients HRQoL, importance, 296 influence of psychosocial factors, 296 patient’s coping strategy, consideration, 296 QoL, importance in case of tumor recurrence, 296 therapeutic goals, 296 in patients with GBM, study, 297 Quiney, N., 334 R Rabban, J. T., 63 Rachinger, W., 123 Rades, D., 224 Radiation-sensitive detection probes, 240–243 Radiation Therapy Oncology Group (RTOG), 48, 57–58, 218, 221–225, 281 Radiation treatment, 30, 199, 288, 289f, 290, 308, 321 Radiochemotherapy, 123f, 124, 303 Radioguided surgery of brain tumors clinical experiences with beta probes, 246–247, 247f with gamma probes, 244–246 concept/clinical applications, 240 detection of brain tumors using 32 P-phosphate, limitation, 240 detection probes for intraoperative tumor localization beta-sensitive intraoperative probes, 242–243 gamma-sensitive intraoperative probes, 241–242 optimal lesion detectability, 240 radioguided detection, steps followed, 240 suppression of background noise, importance, 240–241 weight/size/shape of probe, critical features, 241 issues to be addressed in future, 248–249 radiation exposure during surgery, 247–248 brain phantom experiments with positron intraoperative probe, 247–248 dependent features, 247 radiopharmaceutical agents for tumorectomy, 238–239 beta-emitting radiotracers, 239–240

Index

gamma-emitting radiotracers, 239 surgery for brain tumors aim, 237 diagnostic tools for intraoperative techniques, 238 extent of surgical resection, prognostic factor, 237 intraoperative image-guided surgery, 238 stereotactic image-guided surgery, removal of deep-seated tumors, 238 stereotactic navigation station, 237 Radioimmunotherapy, 123 Radio-inducible gene therapies, 288–289 Ad.EGR-TNF, antitumor effects, 289 application of apoptosis-inducible genes caspase cascade, induction of apoptosis, 290–291, 290f Egr-1/GFP transfected into glioma cells, study, 290 Egr-1/luciferase gene transfected into glioma cells, study, 290 xenograft tumor models, 291f Egr-1 promoter, expression of tumor-sensitive genes, 288, 289f external factor-mediated gene therapies, 288 pE9, radiosensitization of hypoxic tumor cells, 289 radiation treatment, 288 radioinducible iNOS gene therapy approach, 289 Radioinducible iNOS gene therapy approach, 289 Radiolabelled amino acids, 117–125 see also Brain tumors, diagnostic impact of PET using radiolabelled amino acids Radiopharmaceutical agents, 238 beta-emitting radiotracers, 239–240 32 P-labeled sodium phosphate, 239–240 positron emitters, 240 radiophosphorus, 239 gamma-emitting radiotracers 111 Indium-labeled somatostatin analogues, expression in meningiomas, 239 99m Tc-MIBI, application, 239 lesion detectability, ability/parameters, 239 physical half-life, limiting factor, 239 Radioresistant tumors, 56 See also Renal cell carcinoma (RCC) Radioresponsive gene therapy for malignant brain tumors apoptosis-inducible genes

Index

application in radio-inducible gene therapies, 289–291 in treatment of cancer, viral promoters used, 287 radio-inducible gene therapies, 288–289 Ad.EGR-TNF, antitumor effects, 289 Egr-1 promoter, expression of tumor-sensitive genes, 288, 289f external factor-mediated gene therapies, 288 pE9, radiosensitization of hypoxic tumor cells, 289 radiation treatment, 288 radioinducible iNOS gene therapy approach, 289 targeted gene therapies, 287–288 ability to increase transfection rates, challenge, 287 cell-specific promoter to target tumors, examples, 288 Radiosurgery, 48, 56–57, 135, 270, 275–278 See also Stereotactic radiosurgery (SRS) Radiosurgical dose, 281 Radiotherapy, 2–3, 44, 47–48, 49t, 50–51, 55–56, 59, 66t, 68–69, 85, 91, 94, 97–98, 103, 117, 120, 123–124, 128, 182f, 185, 232, 235, 237, 251, 270, 274–276, 280–281, 289, 297, 303, 307–308 See also Radiochemotherapy; Whole brain radiotherapy (WBRT) Radiotracers, 238–243, 247–248 Raghavan, R., 212 Rahimi, S. Y., 331–336 Rahman, R., 21–31, 135 Raichle, M. E., 138 Ramsay, J., 10 Ram, Z., 231 Ranasinghe, M. G., 75 Rao, S. A., 187 Rat brain tumor model, 111 Ratner, N., 128 Ravaud, A., 79 Raylman, R. R., 242–244, 248 Raza, S. M., 208 RCC, see Renal cell carcinoma (RCC) Reactive lesions, 63 Reasoner, D. K., 340 Recht, L., 299 Reckamp, K. L., 79 Recursive partitioning analysis (RPA), 51, 55, 218

375

Reijneveld, J. C., 297 Reinhardt, H., 239, 246 Relieff method, 6, 9, 11 Renal cell carcinoma (RCC), 36–37, 42, 44, 53–60, 78, 81, 217, 222, 244, 265, 283 Ren, Y., 188 Reynolds, B. A., 24 RGD peptide, 172 Ribom, D., 122 Ricci, R., 165 Richards, G. M., 98 Rich, J. N., 255 Riffaud, L., 271 Rigamonti, D., 144 Ring, C. J. A., 288 Rinker-Schaeffer, C. W., 40 Riva, M., 316–317 Rivera, P. P., 144–145 RNA, see MiRNAs in brain tumors, clinical role Roa, W., 270 Robinson, V. L., 88 Rocha, S., 288 Rogne, S. G., 269–271 Rols, M. P., 199 Ronning, P., 269–271 Rosai-Dorfman disease, 70 Rosati, A., 319 Rosenberg, G. A., 210 Rosenthal, N. R., 181, 183 Roser, F., 271 Rosner, D., 48 RPA, see Recursive partitioning analysis (RPA) Ryberg, M., 198 S Saito, E. Y., 50 Saito, R., 213–214 Sakamoto, G. T., 276–277 Salazar, O. M., 280 Salford, L. G., 202 Salo, J., 306 Salvati, M., 279–280 Samlowski, W. E., 53–60 Samnick, S., 2 Sampson, J., 214 Sampson, J. H., 251, 256–257 Sanai, N., 237 Sanders, B. M., 63

376

Sandwich generation, 327 Sanghavi, S. N., 221 Sankhyan, N., 63–71 Santa Cruz Biotechnologies, 149 Sapra, P., 171 Sarcomas, 63, 127, 129, 131 Sasao, A., 141 Sasayama, T., 187 Saulis, G., 199 Savage, J. J., 279–284 Sawaya, R., 55 Sawaya, R. E., 75 Sawyer, A. J., 208 Schackert, G., 218 Schiff, D., 44 Schiffelers, R. M., 172 Schlamann, M., 149 Schmidt, H., 130 Schneider, J. P., 238 Schoeggl, A., 282–283 Schoggl, A., 224–225 Schouten, 75 Schulz, R., 326 Schwannoma, 36, 37t, 127–128, 130, 132, 179, 183 SCLC, see Small cell lung cancer (SCLC) Scopinaro, F., 242 Scott, L., 63 Seckin, H., 36 Seed and soil theory (Paget), 38–39, 45, 86 Sehlen, S., 309 Seibert, K., 334, 336 Seizures in brain tumours, 314–315 Selverstone, B., 238, 242, 246 Seng, S., 111–112, 115 Serizawa, T., 56 Serrano, J., 244 Seunguk, O., 213 Severinghaus, J. W., 342 SGZ, see Subgranular zone (SGZ) Shahidi, H., 44 Shah, M. D., 68 Shamji, M., 316 Sharma, S., 63–71 Shaw, E., 222–223, 281 Shay, J. W., 27 Sheehan, J. M., 75–81 Sheehan, J. P., 80, 273–278, 279–284 Shehata, M. K., 222 Shelden, C. H., 263

Index

Shenkar, R., 143–153 Sherbourne, C. D., 299 Sherman, J. W., 288 Sherwood, P., 324 Shiau, C. Y., 55, 281 Shi, L., 187–188 Shimizu, K., 169–176 Shin, S. I., 114 Shintani, K., 132 Shirasaki, F., 87–88 Shlim, D. R., 342 Short-Form Health Survey (SF – 36), 299 Short TE spectra, 7, 9, 13–14, 17 Shuch, B., 54, 81 Shuto, T., 56 Siddiqui, A., 317 Siebert, R., 249 Siesjo, P., 202 Signal-to-noise ratio (SNR), 7, 140, 242 Silber, J., 187 Silverman, B., 10 Silverman, D. H., 240 Simonetti, A. W., 11, 155–166 Singhal, T., 118–124 Singh, J. M., 340 Singh, S. K., 23, 27 Single brain metastasis, 57, 91, 218–220, 223–225 Single pass sequencing technology, 110 Single/solitary metastases, 36, 44 siRNA, see Small interfering RNAs (siRNA) Sizoo, E. M., 303–310 Slamon, D. J., 80 Sleeman, J., 44–45 Small cell lung cancer (SCLC), 91, 104, 221 Small interfering RNAs (siRNA), 188, 254 Smith, I., 98 Sneed, P. K., 221–225 Snee, M. P., 50 Sneeuw, K. C., 304 Soffietti, R., 280–281 Sogayar, M. C., 109–115 Solid phase transition (SophT), 173, 174f Solitary brain metastasis, 91 Somerville, M., 264 Sonic Hedgehog (SHH) signaling pathway, 28–29, 189 SophT, see Solid phase transition (SophT) Sorafenib, 54, 59, 79 Sordillo, P. P., 130

Index

Sorensen, S. A., 128 Sorokin, L., 78 Soussain, C., 281 Sperduto, P. W., 198 Spitzer quality of life index (SQLI), 299 Spitzer scale, 304 sPNET, see Supratentorial primitive neuroectodermal tumors (sPNET) SQLI, see Spitzer quality of life index (SQLI) SRS for cerebral metastases of digestive tract tumors cerebral metastases, 279–280 clinical symptoms, 279–280 hematogenous dissemination, 280 prognostic factors, 280 gastrointestinal cancer, 280–281 cancer types in G.I. tract, 280 cerebral metastasis from gastric carcinoma, 280 colorectal cancer, 280 digestive system malignancies, 280 esophageal carcinoma, risk factors, 281 global incidence, 280 hepatocellular carcinoma, 280 median survival time, 280 pancreatic cancer, 280 G.I. tract origin, 282–284 efficacy of SRS, assessment, 282 local tumor control, MRI study, 283, 284f median survival times, 283 patient characteristics for study analysis, 282 post-SRS survival, prognostic factors, 283 preoperative symptoms/post-operative complications, 283 single/multiple cerebral metastases from colorectal carcinoma, study, 282 principles of SRS, 281–282 acute/chronic complications, examples, 281–282 dose planning, RTOG study, 281 principal objectives, 281 steep radiation dosing, 281 treatment of radionecrosis, approaches, 281 SRS, indications/patient counseling impact of various histologic subtypes on melanoma/RCC (radioresitant), 222 on patients who underwent GK radiosurgery, 222 patient selection, 218–219

377

clinical outcomes of SRS, inluencing factors, 218 RTOG partitioning analysis, criteria, 218 SRS/WBRT/SRS and WBRT, recommendations, 218–219 radiosurgery with WBRT for multiple brain metastasis, 220–221 radiosurgery with WBRT for single brain metastasis, 220 surgical resection or radiosurgery alone, 221–222 retrospective study of patients with SRS and WBRT/SRS alone, 221 risk of local and distant tumor control, 221 study of patients with SRS and WBRT/SRS alone/WBRT alone, 221–222 surgical resection with WBRT for single brain metastasis, 219–220 randomized study of patients with WBRT alone/WBRT plus surgery, 219–220 WBRT for single/multiple brain metastases acute side-effects, 219 altered dose/customary dosing, outcomes, 219 long-term side effects, 219 SRS, see Stereotactic radiosurgery (SRS) SRS treatment of brain metastases, clinical outcomes indications and patient counseling impact of various histologic subtypes, 222 patient selection, 218–219 radiosurgery with WBRT for multiple brain metastasis, 220–221 radiosurgery with WBRT for single brain metastasis, 220 surgical resection or radiosurgery alone, 221–222 surgical resection with WBRT for single brain metastasis, 219–220 WBRT for single/multiple brain metastases, 219 therapeutic devices used, 218 treatment outcomes impact of treatment on survival, 223–224 local and regional tumor control, 223 neurological acute and late toxicity, 224 treatment planning and methods dose limitations of adjacent structures, 223 dose selection, 222–223 types of radiosurgery, 222 Staba, M. J., 288 Stadlbauer, A., 166 Standardized uptake values (SUV), 132–133, 133f

378

Stark, A. M., 85–94 Steeg, P. S., 44–45, 88 Stemmler, H. J., 50, 80, 98–99, 103 Stepwise algorithm, 6, 9, 13t Stereotactic radiosurgery (SRS), 48, 56–57, 90, 198, 217–225, 234, 274–275, 275t, 279–284 Strauss, J. S., 38 Streptomyces verticillus, 200 Stupack, D. G., 113 Stupp, R., 207, 251, 270, 303, 308 Subgranular zone (SGZ), 22, 25f, 28f Subventricular zone (SVZ), 22, 25f, 28f Sugiyama, K., 63 Suh, J. H., 217–218, 220, 222, 224–225, 261 Suicide gene therapies, see Apoptosis-inducible genes Sukhatme, V. P., 288–290 Su, L. D., 64 Sunitinib, 54, 59, 79, 104 Supratentorial primitive neuroectodermal tumors (sPNET), 21 Surgery for intracranial tumors in elderly patients age and treatment, association, 270–271 brain metastases, treatment modalities, 270 EORTC trial 26062, 270 geriatric glioblastoma patients limited treatment for, 270 surgical outcomes, 271 methodology, 269 “primum non noncere”, 269 radiation and chemotherapy in old patients, study, 270 radio-surgery, old asymptomatic patients, 271 results, 269–270 astrocytomas/metastases and reduced hazard ratio, 270 Cox model, 270 ECOG scores, 270 overall survival rates, 270 tumor categories/age groups selected for surgery, 269–270 selection bias towards older patients, 270 transsphenoidal surgery for pituitary adenomas, 271 Susceptibility weighted imaging (SWI), 147 Sutherland, S., 97–104 SUV, see Standardized uptake values (SUV) Su, W., 68 Suykens, J. A. K., 155–166 SVZ, see Subventricular zone (SVZ)

Index

Swain, R. S., 64 Sweet, W. H., 244 SWI, see Susceptibility weighted imaging (SWI) Synchronous/metachronous tumors, 36, 44 Szabo de Edelenyi, F., 155, 165 T Takahashi, T., 288 Takeda, A., 172 Takei, H., 36–38, 41, 44 Tally, P. W., 37, 40 Tanabe, H., 280 Tang, J., 334 Tang, L., 63 Tang, T. T., 63–66 Tanter, M., 229 Taphoorn, M. J. B., 297, 299, 303–310 Taratuto, A. L., 178 Targeted gene therapy, 287–288 Targeted therapy agents, 55 Target sign, 131 Tate, A. R., 12 Tatter, S. T., 208 Tavora, F., 64 Taylor, M. D., 22–23, 27 Teissie, J., 199 Tekkök, I. H., 65 Telepathology, 181–182 Telomerase, 22–31, 29f, 288 Telomere, 22–30 Temozolomide (TMZ), 99, 104, 188, 251, 257, 298, 308–309, 319–321 Temsirolimus, 81 TE, see Time echo (TE) Testicular cancer, 201 Tham, Y. L., 80 Therapeutic agents, 68, 76, 99, 110, 170, 188, 208–209, 213–214, 233–235, 288, 313, 319–320 Therapeutic issues in epileptic patients antitumour treatments, 320–321 chemotherapy, 320–321 cranial radiation, 321 neurosurgery, 320 choice of optimal AEDs, 318–320 AEDs to be avoided, 319 first-line antiepileptic medication, 319–320 influencing factors, 318–320

Index

second-line antiepileptic medication, 320 side effects of AEDs, 318 prophylactic treatment, 317–318 general guidelines, 317–318 perioperative period, 318 reality, 318 Therapies for CNS tumors 5-aminolevulinic acid photodynamic therapy, 2 antisense treatment, 2 boron neutron capture, 2 gene therepy, 2 ligand-toxin conjugate administration, 2 locoregional redioimmunotherapy, 2 Therapy Oncology Group, 48, 218, 281 Thomas, J. L., 230 Tie-1, 87 Time echo (TE), 6, 149 Tipnis, S. V., 242–243, 248 Topiramate, 319–320 Tornai, M. P., 242, 248 Tortajada, S., 14 Tounekti, O., 201–202 Towner, R. A., 153 Toxicity, see Neurological acute and late toxicity Trackless surgery, 227 Tramadol, 331, 336 Transcranial MR-image guided focused ultrasound surgery (tcMRgFUS), 231–232, 233f, 234 Transgene, 112, 288–289 Transmembrane potential, 199–200 Transsphenoidal surgery, 271 Trastuzumab, 49t, 50, 80, 99–104, 102f Treat, L. H., 234 Treatment of brain metastases chemotherapy, 48–50 efficacy of, reports, 49t corticosteroids, 47 radiosurgery, 48 surgical resection, 48 trastuzumab use, 50 efficacy of, reports, 49t HER-2 overexpression, risk factor, 50 WBRT, 47–48 patients with metatatic lung and breast cancer, 48 Treatment of brain metastases by SRS indications and patient counseling

379

impact of various histologic subtypes, 222 patient selection, 218–219 radiosurgery with WBRT for multiple brain metastasis, 220–221 radiosurgery with WBRT for single brain metastasis, 220 surgical resection or radiosurgery alone, 221–222 surgical resection with WBRT for single brain metastasis, 219–220 WBRT for single/multiple brain metastases, 219 therapeutic devices used, 218 treatment outcomes impact of treatment on survival, 223–224 local and regional tumor control, 223 neurological acute and late toxicity, 224 treatment planning and methods dose limitations of adjacent structures, 223 dose selection, 222–223 types of radiosurgery, 222 Treatment of CNS disease, HER2 targeted strategies capecitabine monotherapy, 99 EGF100151 study capecitabine monotherapy vs. combination of capecitabine and lapatinib, 100 lapatinib monotherapy, phase II study, 99–100 LEAP/EGF100151 study, clinical efficacy, 100 LEAP study, 100 Treatment outcomes/planning and methods, SRS dose limitations of adjacent structures, 223 dose selection, 222–223 impact of treatment on survival, 223–224 local and regional tumor control RTOG study, 223 neurological acute and late toxicity, 224 types of radiosurgery cyclotron-based proton beam, 222 GK device, 222 LINAC device, 222 Treatment of postoperative pain using analgesics cox-2 inhibitors, 331, 335 increased risk of cardiovascular disease, 335 use of Vioxx, APPROVe study, 335 newer analgesics, benefits, 331–332 acetaminophen, 336 tramadol, 336 NSAIDs, 331, 334 opioids, 331, 334 steroids

380

Treatment of postoperative pain using (cont.) corticosteroid therapy, 336 prostaglandin synthesis, inhibition of, 336 Treatment of radionecrosis, approaches, 281 Treatments for epilepsy, 320–321 chemotherapy temozolomide/nitrosoureas, effects, 320–321 cranial radiation, 321 neurosurgery, 320 Tremont-Lukats, I., 318 Trevorrow, T., 342 Trojan, A., 63 TSG, see Tumor suppressor gene (TSG) Tsukada, Y., 79, 97 Tsurushima, H., 287–291 TTM, see Tumor-to-tumor metastasis (TTM) Tumor dissemination, see Ewing’s theory; Seed and soil theory (Paget) Tumorectomy, 238–240 radiopharmaceutical agents beta-emitting radiotracers, 239–240 gamma-emitting radiotracers, 239 lesion detectability, ability/parameters, 239 physical half-life, limiting factor, 239 Tumor grading, 2–3, 122 Tumor, see individual entries Tumor suppressor gene (TSG), 81, 109–115, 128, 131 Tumor-to-tumor metastasis (TTM) benign/malignant tumors, examples, 36 cases of metastases to intracranial tumors, 37t clinical and epidemiologic correlates, 40–41 brain metastases, incidence rates, 40 gender bias, 40 intracranial metastases, signs/symptoms, 40 suprasellar TTM, case study, 41, 41f and collision tumor, 36 diagnosis, 41–42 esophageal carcinoma metastatic to intracranial paraganglioma, case study, 42, 43f intraoperative identification of recipient tumors, 42 pre-operative imaging, case, 41–42 diagnostic criteria extracranial to intracranial TTM, 35 donor tumors, 36 pathogenesis, 36–40 evolution of metastatic niche, stages, 39 metastases to brain, influencing factors, 39–40 metastases, influencing factors, 36–37

Index

metastasis genes, classes, 40 metastatic process, 39 models of metastasis, 39 neoplasia in rabbits, study, 38 preferential targeting of hemangioblastoma, 37–38 preferential targeting of meningiomas, 38 preferential targeting of pituitary adenomas, 38 tumor dissemination, see Ewing’s theory; Seed and soil theory (Paget) single/solitary metastases, 36 synchronous/metachronous tumors, 36 treatment and prognosis, 42–45 malignant tumor treatment, 44–45 metastases to intracranial meningiomas, case study, 44 Tumour heterogeneity, 23 Tumour initiation and cellular immortality brain tumour stem cell paradigm, 27–29 brain tumour stem cell, 29 cancer stem cell, 29 cancer stem cell hypothesis, 28–29 CD133-glioblastoma cells, role, 27–28 VEGF secretion, 29 telomeres and telomerase, role, 26–27, 28f hTERT reactivation/overexpression, 27 telomere length, comprehensive analysis, 27 Tumour and peritumoural factors, 316 Tumour tissue, 320, 340–342 Type B cells, see Glial fibrillary acidic protein (GFAP) Tyrosinase promoter, 288 Tyrosine kinase inhibitors, 254–255 erlotinib, 255 gefitinib, 254 LY294002, 254 U Uemura, S., 275 Ujifuku, K., 188 Umezawa, H., 200 University of Classification rules in standardized partition spaces, 18 University Medical Center Nijmegen (UMCN), 156 University of Texas, 280 University of Virginia, 276 Unsgaard, G., 238

Index

Unterberg, A. W., 341 Urakami, T., 173–174 Utriainen, M., 124 Uziel, T., 189 V Vaccination, 253f, 256–257 Vadmal, M. S., 63 Valproic acid, 319–320 Van Breemen, M., 314, 316, 320 Van den Bent, M. J., 308 Vandermeulen, D., 155–166 Vandewalle, J., 161 Van Gestel, T., 161 Van Huffel, S., 155–166 Van Osch, M. J., 140 Varenika, V., 209, 211, 214 Vascular endothelial growth factor (VEGF), 29, 79–81, 169, 171, 253f, 281, 342 Vecht, C., 320 Vecht, C. J., 48, 56, 218–219, 309 Vedoy, C. G., 111–112 VEGF receptor (VEGFR), 170 VEGF, see Vascular endothelial growth factor (VEGF) Velikova, G., 310 Venkataraman, S., 189 Vermeeren, L., 242 VHL, see Von-Hippel-Lindau (VHL) Vicente, J., 13 Videtic, G. M., 198 Vilela Filho, O., 244, 246 Vile, R. G., 288 Viral (stong) promoters, 287 Visse, R., 87, 89 Vitaliano, P. P., 324 VM-26, 188 Vogelbaum, M. A., 208 Von Hippel, E., 81 Von-Hippel-Lindau (VHL), 37–38, 81 Voorhies, R. M., 279 Vuorinen, V., 273–274 W Wagner, G., 319 Walker, M., 306 Wall, P. D., 332 Wang, J., 27, 137, 140 Wang, L. G., 223

381

Wang, M. L., 99 Wang, M. Y., 253, 256 Ware, J. E., 299 Warfield, S. K., 155 Warmuth, C., 140 Watanabe, K., 38, 41 Watanabe, T., 38, 41 Water-suppressed MR spectrum, 157, 158f Wavelet, definition, 10 Wavelet (WAV) transformation, 8, 11 WBRT, see Whole brain radiotherapy (WBRT) Weber, F., 119 Weber, J., 37–38, 44 Weckesser, M., 117–125 Wegener’s granulomatosis, 68 Weichselbaum, R., 288 Weil, R. J., 79, 85, 90 Weinberg, J. S., 281 Weiss, S., 24 Weitzner, M. A., 298–299, 304, 324 Welch, D. R., 88–89 Wendler, T., 248 Wesseling, P., 40, 44 Whole brain radiotherapy (WBRT), 47–48, 50, 55–59, 98–104, 102f, 198–199, 204, 218–225, 281–283 The WHOQOL, 295 WHO, see World Health Organisation (WHO) Willems, P. W., 238 Wilms, E., 320 Wong, A., 256 Wong, W. W., 127, 129–130 World Health Organisation (WHO), 21–22, 109, 117–118, 120–124, 123f, 156, 163, 166, 178, 180f, 187, 270, 274, 295 Wright, L. S., 26 Wronski, M., 50, 55, 86, 91, 93 Wu, J., 189 Wu, S. X., 24 Wu, T., 161 Wyeth, 81 X Xia, H., 187 Xu, F., 342 Y Yabroff, K. R., 328 Yachnis, A. T., 179

382

Yamamoto, S., 242, 248 Yang, W., 213, 215 Yasargil, M. G., 266 Ye, F. Q., 140 Yin, D., 212 Yokosawa, M., 209, 213–215 Yoneda, T., 86 Yonezawa, S., 172 York, J. E., 280 Yoshimoto, K., 252 Youssef, N., 288

Index

Z Zabramski, J. M., 143, 145 Zaharchuk, G., 140 Zanzonico, P., 241 Zemansky, M., 341 Zhang, Y., 187 Zhao, C. S., 22 Zimm, 76 Zonisamide, 320 Zrinzo, L. U., 340, 342 Zubieta, J. K., 141