Diseases of the Abdomen and Pelvis 2010-2013: Diagnostic Imaging and Interventional Techniques

Diseases of the Abdomen and Pelvis 2010-2013 Diagnostic Imaging and Interventional Techniques

J. Hodler • G.K. von Schulthess • Ch.L. Zollikofer (Eds)

DISEASES OF THE ABDOMEN AND PELVIS 2010-2013 DIAGNOSTIC IMAGING AND INTERVENTIONAL TECHNIQUES 42nd International Diagnostic Course in Davos (IDKD) Davos, March 21-26, 2010 including the Nuclear Medicine Satellite Course “Diamond” Davos, March 20-21, 2010 Pediatric Satellite Course “Kangaroo” Davos, March 20-21, 2010

IDKD in Greece

presented by the Foundation for the Advancement of Education in Medical Radiology, Zurich

Editors J. HODLER Radiology, University Hospital, Zurich, Switzerland

G. K. VON SCHULTHESS Nuclear Medicine, University Hospital, Zurich, Switzerland

CH. L. ZOLLIKOFER Kilchberg/Zurich, Switzerland

ISBN 978-88-470-1636-1

e-ISBN 978-88-470-1637-8

DOI 10.1007/978-88-470-1637-8 Springer Dordrecht Heidelberg London Milan New York Library of Congress Control Number: 2010922809 © Springer Verlag Italia 2010 This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting, reproduction on microfilms or in other ways, and storage in data banks. Duplication of this publication or parts thereof is only permitted under the provisions of the Italian Copyright Law in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the Italian Copyright Law. The use of general descriptive names, registered names, trademarks, etc., in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Product liability: The publisher cannot guarantee the accuracy of any information about dosage and application contained in this book. In every individual case the user must check such information by consulting the relevant literature. Cover design: Simona Colombo, Milan, Italy Typesetting: C & G di Cerri e Galassi, Cremona, Italy Printing and binding: Grafiche Porpora, Segrate (MI), Italy Printed in Italy Springer-Verlag Italia S.r.l., Via Decembrio 28, 20137 Milan Springer is a part of Springer Science+Business Media (www.springer.com)

IDKD 2010-2013

Preface

The International Diagnostic Course in Davos (IDKD) offers a unique learning experience for imaging specialists in training as well as for experienced radiologists and clinicians wishing to be updated on the current state of the art and the latest developments in the fields of imaging and image-guided interventions. This annual course is focused on organ systems and diseases rather than on modalities. This year’s program deals with diseases of the abdomen and pelvis. During the course, the topics are discussed in group workshops and in plenary sessions with lectures by world-renowned experts and teachers. While the workshops present state-of-the-art summaries, the lectures are oriented towards future developments. Accordingly, this Syllabus represents a condensed version of the contents presented under the 20 topics dealing with imaging and interventional therapies in abdominal and pelvic diseases. The topics encompass all the relevant imaging modalities including conventional X-rays, computed tomography, nuclear medicine, ultrasound and magnetic resonance angiography, as well as image-guided interventional techniques. The Syllabus is designed to be an “aide-mémoire” for the course participants so that they can fully concentrate on the lecture and participate in the discussions without the need of taking notes. Additional information can be found on the IDKD website: www.idkd.org J. Hodler G.K. von Schulthess Ch.L. Zollikofer

IDKD 2010-2013

Table of Contents

Workshops Emergency Radiology of the Abdomen: The Acute Abdomen . . . . . . . . . . . . . . . . . . . . . . .

3

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

Trauma of the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14

Philip J. Kenney, Stuart E. Mirvis

Diseases of the Esophagus and Stomach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

Marc S. Levine, Ahmed Ba-Ssalamah

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis

28

Marc J. Gollub

Diseases of the Small Bowel, Including the Duodenum – MRI . . . . . . . . . . . . . . . . . . . . . .

32

Karin A. Herrmann

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases . . . . . . . . . .

37

Jaap Stoker, Richard M. Gore

CT Colonography: Updated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

Daniel C. Johnson, Michael Macari

Imaging of Diffuse and Inflammatory Liver Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

50

Pablo R. Ros, Rendon C. Nelson

Focal Liver Lesions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

Wolfgang Schima, Richard Baron

Imaging Diseases of the Gallbladder and Bile Ducts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

75

Angela D. Levy, Celso Matos

Diseases of the Pancreas, I: Pancreatitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

81

Thomas Helmberger

Diseases of the Pancreas, II: Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

89

Ruedi F. Thoeni

Adrenal Imaging and Intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

96

William W. Mayo-Smith, Isaac R. Francis

Renal Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

99

Richard H. Cohan, Ronald J. Zagoria

Urinary Tract Obstruction and Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Parvati Ramchandani, Julia R. Fielding

VIII

Benign Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Caroline Reinhold, Rahel A. Kubik-Huch

Malignant Diseases of the Female Genital Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Evis Sala, Susan Ascher

Magnetic Resonance Imaging of Prostate Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Imaging of the Male Pelvis: The Scrotum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Brent J. Wagner

Spread of Metastatic Disease in the Abdomen and Pelvis . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 James A. Brink, Ali Shirkhoda

Abdominal Vascular Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 Johannes Lammer

Non-vascular Abdominal Disease: Diagnosis and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

An Approach to Imaging the Acute Abdomen in the Pediatric Population . . . . . . . . . . . 167 Alan Daneman, Simon G. Robben

Imaging Uronephropathies in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 Jeanne S. Chow, Fred E. Avni

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . . . . 183 Gerald Antoch

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging . . . . . . . . . . . . . . . . . 190 Thomas F. Hany

Nuclear Medicine Satellite Course “Diamond” Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals . . . . . 199 Angelika Bischof Delaloye

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 Ariane Boubaker

PET in Hepatobiliary-Pancreatic Tumors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215 Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri

PET in Tumors of the Digestive Tract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 Thomas F. Hany

Tumors of the Adrenergic System: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 Cornelis A. Hoefnagel

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy . . . . . . . . . . . . . . . . . . . . 231 Dik J. Kwekkeboom

Table of Contents

IX

Table of Contents

Pediatric Satellite Course “Kangaroo” Imaging Cystic Kidneys in Children . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Fred E. Avni

Understanding Duplication Anomalies of the Kidney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Jeanne S. Chow

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management . . . . . 247 Alan Daneman

Pediatric Intestinal Ultrasonography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 Simon G. Robben

IDKD 2010-2013

List of Contributors

Allegri V., 215 Antoch G., 183 Ascher S., 119 Avni F.E., 174, 239 Ba-Ssalamah A., 22 Barentsz J.O., 125 Baron R., 63 Bartolozzi C., 162 Battaglia V., 162 Bischof Delaloye A., 199 Boubaker A., 205 Bozzi E., 162 Brink J.A., 146 Bruel J.-M., 3 Chow J.S., 174, 243 Cohan R.H., 99 Daneman A., 167, 247 Fanti S., 215 Fielding J.R., 104 Francis I.R., 96 Futterer J.J., 125 Gollub M.J., 28 Gore R.M., 37 Hany T.F., 190, 219 Heiken J.P., 3 Heijmink S.W.T.P.J., 125 Helmberger T., 81 Herrmann K.A., 32

Hoefnagel C.A., 226 Hoeks C., 125 Hulsbergen-van der Kaa C., 125 Johnson D.C., 48 Kenney P.J., 14 Kubik-Huch R.A., 110 Kwekkeboom D.J., 231 Lammer J., 154 Levine M.S., 22 Levy A.D., 75 Macari M., 48 Maffione A.M., 215 Marincek B., 3 Matos C., 75 Mayo-Smith W.W., 96 Mirvis S.E., 14 Nelson R.C., 50 Ramchandani P., 104 Reinhold C., 110 Robben S.G., 167, 252 Ros P.R., 50 Sala E., 119 Schima W., 63 Shirkhoda A., 146 Stoker J., 37 Thoeni R.F., 89 Wagner B.J., 142 Zagoria R.J., 99

WORKSHOPS

IDKD 2010-2013

Emergency Radiology of the Abdomen: The Acute Abdomen Jean-Michel Bruel1, Borut Marincek2, Jay P. Heiken3 1 Medical

Imaging Department, Hôpital Saint-Eloi, CHRU de Montpellier, Montpellier, France of Diagnostic Radiology, University Hospital, Zurich, Switzerland 3 Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO, USA 2 Institute

Introduction The term “acute abdomen” defines a clinical syndrome characterized by a history of hitherto undiagnosed abdominal pain lasting less than one week. A large range of disorders, from benign, self-limited diseases to conditions that require immediate surgery, can cause acute abdominal pain. Eight conditions account for over 90% of patients who are referred to the hospital and who are seen on surgical wards with acute abdominal pain: acute appendicitis, acute cholecystitis, small bowel obstruction (SBO), urinary colic, perforated peptic ulcer, acute pancreatitis, acute diverticular disease, and non-specific, non-surgical abdominal pain (dyspepsia, constipation).

Imaging Techniques Clinical assessment of the acute abdomen is often difficult because the findings of the physical examination and laboratory investigations are often non-specific. Traditionally, plain abdominal radiographs have served as the initial imaging approach; however, because of their diagnostic limitations, plain radiographs now play only a limited role in this clinical setting. Currently, the major indications for plain radiography are to determine the presence of bowel obstruction, perforated viscus, urinary tract calculi, or a foreign body. The conventional radiographic examination consists of supine and either upright or left lateral decubitus images. Computed tomography (CT) now serves as the imaging test of choice for most adult patients with acute abdominal pain. It has been shown to be superior to plain radiography for diagnosing nearly all causes of acute abdominal pain. Several studies have demonstrated that the use of CT to evaluate patients with acute abdominal pain increases the accuracy of clinical diagnosis by >20% and results in management changes in up to 60% of patients. The major obstacles to replacing plain abdominal radiography with CT are its higher cost, more limited availability, and higher radiation dose. The use of CT in patients with an acute abdomen requires careful attention to the CT protocol.

1. The multidetector CT (MDCT) image data volume should be obtained from the dome of the diaphragm to the inferior aspect of the pubic symphysis (with thin collimation and a short acquisition time). CT data are reconstructed with thin, overlapped slices for multiplanar (coronal and sagittal planes) reformation and 3- to 5-mm thick contiguous axial slices. The image series is sent to a dedicated workstation and/or PACS. 2. Vascular enhancement by iodinated contrast medium (CM) is mandatory in most cases, unless contraindicated. Special attention should be paid to older patients and those with metabolic disorders (dehydration in SBO) in assessing the renal impact of CM administration. In general, the most helpful scanning phase is the late portal phase (70 s), but other scanning phases are useful in selected circumstances: arterial phase in bowel ischemia, bleeding, or visceral infarction; delayed phase (3 min) for assessing the lack of enhancement in a patient with suspected acute mesenteric ischemia. A pre-contrast CT scan allows demonstration of calcifications, lithiasis, and acute or subacute hemorrhage; multiphasic scanning should be used only for specific indications in order to limit radiation dose. 3. Changes in the CT protocol should be decided according to the clinical conditions and/or the preliminary results of the CT examination. In selected cases, colorectal opacification and/or image acquisition with the patient in the prone position may be helpful to clarify equivocal findings. In most cases, the systematic use of ingested oral contrast is not recommended. 4. The method of image evaluation is critical to optimize interpretation. Additional window level (M) and width (W) settings are useful to identify tiny bubbles of extraluminal gas (CT lung windows) or hyperattenuation from recent hemorrhage (narrow CT windows). The systematic use of multiplanar reformation (MPR), particularly in the coronal plane, is recommended, and 3D imaging may be helpful. Ultrasonography (US) is the initial imaging technique of choice for patients with suspected acute cholecystitis or acute gynecological abnormalities. It also is the primary

4

method for evaluating pregnant women and pediatric patients. Although less sensitive and specific than CT, US is an excellent imaging test for diagnosing acute appendicitis, when employed by experienced individuals. It can also be used to evaluate the presence or absence of the layered structure of the digestive tract wall or to assess the structure of a lesion identified at CT. Until recently, magnetic resonance imaging (MRI) has played a very limited role in patients with acute abdominal pain; however, it is now established in the imaging of pregnant women with abdominal pain who have had a negative or equivocal US examination. Recent studies assessing the use of MRI to evaluate all patients with acute lower abdominal pain have shown promising results. MRI may also have a role in patients with biliary diseases and/or pancreatitis. The differential diagnosis in a patient with an acute abdomen is influenced greatly by the nature and location of the pain. Therefore, the imaging strategies for acute pain localized to an abdominal quadrant should be discussed separately from those for acute pain that is diffuse or localized to the flank or epigastric region.

Acute Pain in an Abdominal Quadrant In many cases, acute abdominal pain can be localized to one either the right upper, left upper, right lower, or left lower abdominal quadrant.

Right Upper Quadrant Acute cholecystitis is by far the most common disease to involve the right upper quadrant. Other important diseases that can have a clinical presentation similar to that of acute cholecystitis are pyogenic or amebic liver abscess, spontaneous rupture of a hepatic neoplasm (usually hepatocellular adenoma or carcinoma), hepatitis, and myocardial infarction. The preferred imaging method for evaluating patients with acute right upper abdominal pain is US. It is a reliable technique for establishing the diagnosis of acute calculous cholecystitis. The imaging criteria include the detection of gallstones, the sonographic Murphy sign, gallbladder wall thickening ≥3 mm, and pericholecystic fluid. The association of three of these signs is highly suggestive of acute cholecystitis. Isolated gallbladder wall thickening may be secondary to other conditions, such as gallbladder adenomyomatosis, gallbladder carcinoma, HIV cholangitis, sclerosing cholangitis, acute hepatitis, cirrhosis, ascites, portal hypertension, hypoproteinemia, pancreatitis, and cardiac failure. In acute calculous cholecystitis, typically a calculus obstructs the cystic duct. The trapped concentrated bile irritates the gallbladder wall, causing increased secretion, which in turn leads to distention and edema of the wall. The rising intraluminal pressure compresses the vessels, resulting in thrombosis, ischemia, and subsequent necrosis and perforation of the

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

wall. Gallbladder perforation and complicating pericholecystic abscess typically occur adjacent to the gallbladder fundus because of the sparse blood supply. CT may be useful for confirmation of the sonographic diagnosis, but usually is not necessary. Emphysematous cholecystitis is a rare complication of acute cholecystitis that generally is associated with diabetes mellitus. US or CT demonstrates gas in the wall and/or lumen of the gallbladder, which implies underlying gangrenous changes (Fig. 1). Acalculous acute cholecystitis accounts for only approximately 5% of cases of acute cholecystitis but is especially common in patients in the intensive care unit. Prolonged bile stasis results in increased viscosity of the bile that ultimately leads to functional cystic duct obstruction. Both US and CT are accurate techniques for diagnosing liver abscesses. US usually demonstrates a round or oval hypoechoic mass with low-level internal echoes. Although the lesion may mimic a solid hepatic mass, the presence of through transmission is a clue to its cystic nature. Pyogenic liver abscesses most commonly are the result of seeding from appendicitis or diverticulitis or direct extension from cholecystitis or cholangitis. Amebic abscesses result from primary colonic involvement, with seeding through the portal vein. In most cases, the US appearances of pyogenic and amebic abscess are indistinguishable. The CT appearances of pyogenic and amebic abscesses also overlap substantially. Amebic abscesses are cystic masses of low attenuation. An enhancing wall and a peripheral zone of edema surrounding the abscess are common but not universally present. Extrahepatic extension of the amebic abscess with involvement of the chest wall, pleura, or adjacent viscera is a frequent finding. Whereas amebic abscesses usually are solitary and unilocular, pyogenic abscesses may be multiple or multi-

Fig. 1. Diabetic patient with emphysematous cholecystitis and gangrene of the gallbladder. CT shows air-fluid level in the gallbladder lumen and air in the gallbladder wall (arrows)

Emergency Radiology of the Abdomen: The Acute Abdomen

loculated and may demonstrate an irregular contour. Some pyogenic abscesses have a mixed cystic and solid appearance on US, CT, or MRI; rarely, they appear completely solid. A small percentage of hepatic abscesses, particularly those secondary to Klebsiella infection, are associated with portal vein thrombosis. Spontaneous rupture of a hepatocellular carcinoma with subsequent hemoperitoneum is a frequent complication in countries with a high incidence of this tumor, but is less commonly seen in Western countries. Subcapsular location and tumor necrosis have been implicated in the pathogenesis. US, and especially CT, are the most useful techniques for diagnosing a ruptured hepatocellular carcinoma, which appears as a peripheral or subcapsular mass. Transcatheter embolization of either the tumor or the bleeding hepatic artery is the treatment of choice. Spontaneous hemorrhage within a hepatocellular adenoma occurs most commonly in women taking oral contraceptives. Capsular rupture with subsequent hemoperitoneum is an uncommon complication. On CT, highdensity intraperitoneal fluid confirms the diagnosis of hemoperitoneum. Extravasation of CM, when present, is indicative of active bleeding.

Left Upper Quadrant Although infrequent, acute left upper quadrant pain is most often seen in splenic infarction, splenic abscess, gastritis, and gastric or duodenal ulcer. US is most frequently used for screening, while CT enables accurate further evaluation. The diagnosis of gastric pathology is established by endoscopy, with imaging playing a minor role. Common causes of splenic infarction include bacterial endocarditis, portal hypertension, and marked splenomegaly. Pancreatitis or tumors that extend into the splenic hilum can also result in infarction. Splenic infarction may be focal or global. Typical focal splenic infarcts appear as peripheral wedge-shaped defects, hypoechoic or isoechoic at US and hypoattenuating at CT. Most splenic abscesses are secondary to hematogenous dissemination of infection, e.g., bacterial endocarditis or tuberculosis. Intravenous drug abusers and immunocompromised individuals are predominantly affected. US and CT are sensitive, but the specificity of either one is low. On US, most abscesses appear as hypo- or anechoic, poorly defined lesions; on CT, they typically appear as rounded lesions of low attenuation and with rim enhancement. Spontaneous splenic rupture can occur in patients with hematological malignancy or secondary to rapid splenic enlargement from viral infections such as mononucleosis.

Right Lower Quadrant Acute appendicitis is not only the most frequent cause of acute right lower quadrant pain, it is also the most commonly encountered cause of an acute abdomen. Other

5

diseases that can present with acute right lower quadrant pain include acute terminal ileitis (Crohn’s disease), typhlitis, right-sided colonic diverticulitis and, in women, pelvic inflammatory disease, complications of ovarian cyst (hemorrhage, torsion, and leakage), endometriosis, or ectopic pregnancy. Less common causes of right lower quadrant pain include segmental infarction of the greater omentum, mesenteric adenitis, epiploic appendagitis, perforated cancer, and ileal or Meckel’s diverticulitis. The diagnosis of acute appendicitis is uncertain in up to one-third of patients. Thus, pre-operative imaging plays an important role in confirming or excluding the diagnosis. With the increasing use of medical imaging to evaluate patients with suspected acute appendicitis, the rate of both false-positive (unnecessary appendectomy) and false-negative (leading to complications from perforated appendicitis) diagnoses has decreased. The standard surgical teaching is that patients with typical clinical findings should undergo immediate appendectomy without pre-operative imaging. Nevertheless, at most medical centers pre-operative imaging is obtained even when the clinical presentation is typical. The most specific CT finding of acute appendicitis is a thickwalled appendix that contains an appendicolith (Fig. 2). The inflamed appendix often is dilated and fluid-filled. Additional helpful findings are stranding of the periappendiceal fat and thickening of the cecal apex. Findings that indicate appendiceal perforation include periappendiceal abscess, extraluminal gas, a right lower quadrant inflammatory mass, a defect in the appendiceal wall, and SBO. In the evaluation of suspected acute appendicitis in children, pregnant women, and women of reproductive age, US is an important imaging option. Demonstration of a swollen, noncompressible appendix >7 mm in diameter with a target configuration is the primary sonographic criterion (Fig. 3). Additional helpful US findings are “MacBurney’s sign” (maximum tenderness found with graded compression of the inflamed appendix) and demonstration of an appendicolith. These US signs may also be demonstrated by transvaginal highresolution US. The advantages of US include the lack of ionizing radiation, relatively low cost, and widespread availability. However, US requires considerable skill and is difficult to perform in obese patients, patients with severe pain, and patients likely to have a complicating periappendiceal abscess. When the sonographic findings are unclear, CT can provide a rapid and definitive diagnosis. Due to its exceptional accuracy, CT has emerged in many centers as the primary imaging test for patients with suspected acute appendicitis. In a small percentage of patients, diverticulitis manifests itself as a right-sided condition. Right-sided colonic diverticula are often congenital, solitary, and true diverticula, unlike sigmoid diverticula. In right-sided diverticulitis the normal appendix should be visible.

6

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

c

a

b

Fig. 2 a-c. Acute appendicitis. Axial (a, b) and sagittal (c) views of multidetector CT demonstrate a dilated appendix in retrocecal position, a calcified appendicolith at the base of the appendix (arrowheads), and inflammatory changes of the mesenteric fat (arrow)

a

b

c

Fig. 3 a-c. Ultrasonography of acute appendicitis in a 12-year-old girl. Oblique (a, b) and transverse (c) views show swollen appendix (diameter 10 mm, arrowheads) with a target configuration

Left Lower Quadrant Diverticulitis is the most common cause of acute left lower quadrant abdominal pain. The condition occurs in 10–20% of patients with diverticulosis and most commonly involves the sigmoid colon. CT is very sensitive and approaches 100% specificity and accuracy in the diagnosis or exclusion of diverticulitis; it has therefore largely replaced barium enema examinations. CT is also very useful in establishing the presence of pericolic complications. The CT diagnosis of acute diverticulitis is based on the identification of segmental colonic wall thickening and pericolic in-

flammatory changes, such as fat stranding, inflammatory mass, gas bubbles, or free fluid (Figs. 4, 5). Complications of acute diverticulitis include abscess, fistula (most commonly colovesical), SBO, peritonitis, septic thrombophlebitis, colonic obstruction, and ureteral obstruction. In patients with left lower quadrant pain, alternative diagnoses that should be considered are colitis (infectious/ inflammatory or ischemic), colonic carcinoma, epiploic appendagitis, neutropenic colitis, functional colonic disorders, and extragastrointestinal disorders (pyelonephritis or gynecological diseases). Epiploic appendagitis is a clinical condition mimicking acute colonic diverticulitis, with focal

7

Emergency Radiology of the Abdomen: The Acute Abdomen

Fig. 4. Sigmoid diverticulitis with pericolic abscesses. CT shows fine linear strands within pericolic fat, diverticula filled with air, barium, or fecal material, circumferential bowel thickening, and frank abscesses (arrows)

exquisite lower abdominal pain. It is diagnosed with CT (or US) by the demonstration of an ovoid lesion within the pericolonic fat, surrounded by inflammatory changes and abutting the colonic wall. As this disease resolves spontaneously within a few days, its correct diagnosis on CT images is important to avoid unnecessary surgery. Distinguishing sigmoid diverticulitis from carcinoma is a major differential diagnostic consideration but is often difficult on CT images despite CT findings suggestive of colon carcinoma, such as marked thickening of the colonic wall, focal thickening (length <5 cm), and the presence of pericolonic lymph nodes. The distinction between perforated colon carcinoma and diverticulitis may be especially difficult because many patients with colon carcinoma also have diverticulosis. Therefore, it is crucial to perform colonoscopy 3-4 weeks after the onset of acute diverticulitis to rule out a colonic carcinoma.

Gynecological Disorders a

b

Fig. 5 a, b. Perforated sigmoid diverticulitis. a CT demonstrates tiny bubbles of extraluminal gas within the left subphrenic space (arrows, 1) and hepatic pedicle (arrow, 2). The determination of a pneumoperitoneum from the perforated sigmoid diverticulitis (b) requires dedicated window settings. b Coronal reformat shows the patterns indicating the severity of the local extent of this case of sigmoid diverticulitis: thickened bowel wall with diverticula associated with dramatic changes of the pericolic fat and thickened root of the sigmoid mesocolon. Note the tiny bubble of extraluminal gas (arrow) within the hepatic pedicle

These are important causes of acute lower abdominal pain (right lower quadrant, left lower quadrant, or central pelvic) in female patients, particularly young women. Acute uterine and ovarian disorders, pelvic inflammatory disease, endometriosis, and ectopic pregnancy are diagnostic considerations in women of child-bearing age who present with an acute abdomen. Pelvic inflammatory disease (PID) is the most frequent gynecological cause of an acute abdomen. Its manifestations may include salpingitis, tubo-ovarian abscess, and peritonitis. The most common causative organisms are Neisseria gonorrhoeae and Chlamydia trachomatis. Since PID most commonly results from an ascending infection, it usually involves both adnexa. US is the imaging test of choice for patients with suspected PID; however, CT is helpful when the clinical diagnosis is equivocal. CT findings may include fluid within the cul-de-sac, endometrial thickening, fluid within the uterine cavity, and thickening and dilation of the fallopian tubes (pyosalpinx) (Fig. 6).

Fig. 6. Pelvic inflammatory disease. Transaxial CT image shows thickened, dilated fluid-filled fallopian tubes indicative of bilateral pyosalpinx

8

With progression to tubo-ovarian abscesses, unilateral or bilateral cystic adnexal masses may be seen, usually in association with pyosalpinx. In advanced cases of PID, the intrapelvic spread of purulent material may result in peritonitis. Patients with Fitz-Hugh-Curtis syndrome present with right upper quadrant pain due to perihepatic inflammation from intraperitoneal exudates stretching between the liver capsule and the peritoneum, which can mimic carcinomatosis on CT. In adult patients, ovarian torsion (adnexal torsion) usually occurs in association with an ovarian mass, which acts as a fulcrum to potentiate torsion. Teratoma is the most common cause of ovarian torsion. The typical presentation is non-specific, consisting of acute lower abdominal pain associated with nausea, vomiting, and leukocytosis. If ovarian torsion is suspected, Doppler US is the initial imaging test of choice; however, because the clinical presentation usually is non-specific, CT is often the first imaging test requested. A CT finding helpful in making the diagnosis of ovarian torsion is an enlarged ovary that is displaced from its normal location. Secondary signs include a thickened fallopian tube, a twisted vascular pedicle, hemoperitoneum, and deviation of the uterus toward the affected side. Complications of ovarian cysts such as hemorrhage and rupture also can cause acute lower abdominal pain. Hemorrhagic cysts contain fluid that is high in attenuation, sometimes with a fluid-fluid level. In a small percentage of patients, the cyst may rupture, resulting in hemoperitoneum. Correlation with β-human chorionic gonadotropin (hCG) levels is important as a ruptured ectopic pregnancy may present with similar clinical and imaging features. Endometriosis is characterized by the presence of functioning endometrial tissue outside of its normal intrauterine location. It presents as acute abdominal pain in only a small percentage of women and is usually caused by rupture or hemorrhage of an endometrioma or by torsion of an ovary that contains endometrial implants. On CT, endometriomas have a variable appearance, ranging from cystic to solid adnexal masses. Ectopic pregnancy remains the leading cause of death during the first trimester of pregnancy, with a mortality of 9-14%. The main risk factors for ectopic pregnancy include a history of ectopic pregnancy, tubal surgery, and PID. The initial evaluation of patients suspected of having an ectopic pregnancy requires quantitative measurement of serum β-hCG level and transvaginal US. The latter should be used to search for the presence of an adnexal mass (with or without highly specific signs such as adnexal gestational sac with a live embryo, suggested by the demonstration of cardiac activity and/or a “tubal ring sign”), hematosalpinx, pelvic free fluid (highly suggestive if heterogeneous), and hemoperitoneum, enlarged uterus (with a pseudo-gestational sac), and symmetrically enlarged ovaries. The diagnosis of ectopic pregnancy often is difficult, since transvaginal US may be normal in up to 25% of these patients and adnexal abnormalities

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

may be found in numerous alternative diagnoses. The diagnosis of ectopic pregnancy is based on an association of the β-hCG level with the transvaginal US findings: 1. a normal β-hCG level rules out an ectopic pregnancy; 2. an intrauterine gestational sac with a live embryo (cardiac activity) rules out an ectopic pregnancy (but in patients with assisted reproduction by ovulation induction the rate of “heterotopic pregnancy”, defined as the simultaneous occurrence of an intrauterine and an extrauterine pregnancy, has been reported to be 1–3%); 3. an elevated β-hCG level (>1000 IU/L) without an intrauterine gestational sac, associated with an abnormal adnexal pattern and/or heterogeneous pelvic fluid, indicates an ectopic pregnancy.

Acute Abdomen with Diffuse Pain Any disorder that irritates a large portion of the gastrointestinal (GI) tract and/or the peritoneum can cause diffuse abdominal pain. The most common disorder is gastroenterocolitis. Other important disorders are bowel obstruction, ischemic bowel disease, and GI tract perforation.

Bowel Obstruction Bowel obstruction is a frequent cause of abdominal pain and accounts for approximately 20% of surgical admissions for acute abdominal conditions. The small bowel is involved in 60-80% of cases. Frequent causes of SBO are postoperative adhesions, hernias, and neoplasms. Mechanical large bowel obstruction is most commonly due to colorectal carcinoma, but volvulus and diverticulitis are also important causes. Colonic volvulus most commonly involves the sigmoid region, followed by the cecum. The diagnosis of bowel obstruction is established on clinical grounds and usually confirmed with plain abdominal radiographs. Due to the diagnostic limitations of plain radiography, CT is increasingly used to establish the diagnosis, identify the site, level, and cause of obstruction, and determine the presence or absence of associated bowel ischemia. CT can be useful for differentiating between simple and closed-loop obstruction. Closed-loop obstruction is a form of mechanical bowel obstruction in which two points along the course of the bowel are obstructed at a single site. It is usually secondary to an adhesive band or a hernia. Since a closed loop tends to involve the mesentery and is prone to produce a volvulus, it represents the most common cause of strangulation. However, only colonic volvulus is associated with classic features on plain abdominal radiography. CT is particularly reliable in higher grades of bowel obstruction. It has proved useful in characterizing bowel obstruction from various causes, including adhesions, hernia, neoplasm, extrinsic compression, inflammatory bowel disease, radiation enteropathy, intussusception, gallstone ileus, or volvulus. The essential CT finding of

9

Emergency Radiology of the Abdomen: The Acute Abdomen

bowel obstruction is the delineation of a transition zone between the dilated and decompressed bowel. Careful inspection of the transition zone and luminal contents usually reveals the underlying cause of obstruction. However, the presumed point of transition from dilated to nondilated bowel can be difficult to determine in the transaxial plane. MDCT facilitates this task by providing the radiologist with a volumetric data set that can be viewed in the transaxial, sagittal, or coronal plane or any combination of the three. These MPR views centered on the anticipated transition point help to determine the site, level, and cause of obstruction. Mechanical obstruction of the small bowel (SBO) has to be differentiated from paralytic ileus, large bowel obstruction (LBO), and non-obstructive massive distention of the colon. Paralytic ileus is a common problem after abdominal surgery. It may be diffuse or localized and has numerous causes, e.g., secondary to ischemic conditions, inflammatory or infectious disease, abnormal electrolyte, metabolite, drug or hormonal levels, or innervation defects. In LBO, CT demonstrates distension of the large bowel to the point of obstruction, with collapse of the distal large bowel. The distal small bowel loops may also be distended if the ileocecal valve is incompetent. Luminal obstruction by a colonic carcinoma and colonic volvulus (Fig. 7) are the main causes of LBO. Perforation of the cecum, due to gross distention resulting in ischemia of the cecal wall, is the main complication of severe LBO. A massively dilated colon may be seen in toxic megacolon and Ogilvie’s syndrome. In toxic megacolon secondary to severe colitis, CT demonstrates a thickened wall with “thumbprinting” caused by wall edema and inflammation. Ogilvie’s syndrome is an acute pseudoobstruction with dilation of the colon in the absence of a colonic transition zone. It may occur in severely ill patients after surgery and/or with neurological disorders, serious infections, cardiorespiratory insufficiency, and metabolic disturbances. Drugs that disturb colonic motility (e.g., anticholinergics or opioid analgesics) contribute to the development of this condition.

a

Fig. 7 a, b. Cecal volvulus. Transaxial CT (a) and coronal volume-rendered (b) CT demonstrate a markedly dilated cecum (C) in the left side of the pelvis. The arrow points to the area of twist of the ascending colon. Note the dilated small bowel loops due to the proximal colonic obstruction

Ischemic Bowel Disease Predominant causes of bowel ischemia are arterial or venous occlusion and hypoperfusion. Occlusive disease involves the mesenteric arteries, most commonly the superior mesenteric artery, in the large majority of cases. Bowel ischemia secondary to venous thrombosis is much less common. The only direct sign of vascular impairment of the bowel is diminished bowel wall enhancement, which is due to inadequate arterial inflow to the bowel. Increased bowel wall enhancement may be seen in some cases secondary to reactive hyperemia or compromised venous outflow. Other CT findings are direct visualization of a thrombus in the superior mesenteric artery or vein. Bowel distention and bowel wall edema are nonspecific findings and may accompany inflammatory or infectious causes. Bowel distention reflects the interruption of peristaltic activity in ischemic segments. Nonocclusive acute mesenteric ischemia usually is due to hypoperfusion secondary to severe cardiac disease, but also occurs in patients with end-stage renal or hepatic disease. An important radiographic manifestation of nonocclusive acute mesenteric ischemia due to low arterial flow is mesenteric arterial vasoconstriction. A less common form of non-occlusive acute mesenteric ischemia is severe vasculitis, which often affects younger individuals. Mesenteric vasculitis usually results in bowel edema and mucosal hyperenhancement. The small and the large bowel often are both involved. The duodenum is involved in approximately one-quarter of patients. The most common mechanical cause of bowel ischemia is obstruction. A closed-loop SBO is more likely than other types of obstruction to result in vascular compromise (strangulation obstruction). Strangulation obstruction has a reported prevalence of 5-40% and is a predominantly venous disease. The most frequent abnormality seen on CT is bowel wall thickening. The thickened bowel wall sometimes is associated with a target sign, consisting of alternating layers of high and low attenuation within the thickened bowel wall, which results from submucosal edema and hemorrhage. The bowel segment proximal to an obstruction can become ischemic

b

10

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

due to severe distention. CT findings that suggest subsequent infarction are non-enhancement of the bowel wall, gas in the bowel wall, mesenteric or portal veins, edema/hemorrhage in the mesentery adjacent to thickened and/or dilated bowel loops, and ascites (Fig. 8).

Perforation of the Gastrointestinal Tract Gastrointestinal perforation usually causes localized pain initially, and culminates in diffuse pain if peritonitis develops. Gastroduodenal perforation associated with peptic ulcer disease or a necrotic neoplasm has become less frequent in recent decades due to earlier diagnosis and improved therapy. At the same time, the incidence of gastroduodenal perforation resulting from endoscopic instrumentation has increased. Perforation of the small bowel is relatively uncommon but may be secondary to a foreign body, small bowel diverticulitis, or trauma. Spontaneous rupture of the colon is more frequent and can occur when the colon becomes markedly dilated proximal to an obstructing lesion (tumor, volvulus) or when the bowel wall is friable (ischemic or ulcerative colitis, necrotic neoplasm). Fiberoptic colonoscopy with or without biopsy is another cause of colonic perforation. Pneumoperitoneum can be recognized by the presence of subdiaphragmatic gas on an upright chest radiograph or an upright or left lateral decubitus abdominal radiograph. A large pneumoperitoneum generally is indicative of colonic perforation, whereas moderate quantities of free gas are seen with gastric perforation. Small bowel perforation usually results in either a limited amount of peritoneal gas or none, because the small bowel usually does not contain gas. Detection of subtle pneumo-

a

peritoneum is often difficult. As CT is far more sensitive than conventional radiography in demonstrating a small pneumoperitoneum, it has become the imaging test of choice when the results of conventional radiography are equivocal. Viewing the CT images at “lung window” settings improves the demonstration of small amounts of extraluminal gas. Retroperitoneal perforations (duodenal loop beyond the bulbar segment, or involving the appendix; posterior aspect of the ascending and descending colon, or the rectum below the peritoneal reflection) tend to be contained locally and remain clinically silent for several hours or days. Retroperitoneal gas has a mottled appearance and may extend along the psoas muscles. In contrast to intraperitoneal gas, retroperitoneal gas does not move freely when the patient’s position is changed from supine to upright for plain abdominal radiographs.

Acute Abdomen with Flank or Epigastric Pain Acute flank or upper abdominal pain radiating to the back is commonly a manifestation of retroperitoneal pathology, especially urinary colic, acute pancreatitis, or leaking abdominal aortic aneurysm.

Urinary Colic For decades, intravenous urography was the primary imaging technique used to evaluate patients with suspected urinary colic. Plain abdominal radiography and US may be useful for patients with a contraindication to radiation or iodinated intravenous CM. However, because

c

b Fig. 8 a-c. Strangulating small bowel obstruction due to an adhesive band that developed after cholecystectomy and appendectomy. Axial (a, b) and sagittal (c) views of multidetector CT show bowel wall thickening, enhancing bowel wall with submucosal edema, and/or hemorrhage giving a target sign (long arrows), indicating ischemia. Non-enhancement of the bowel wall of the jejunum (short arrow) corresponds to segmental infarction. A Ascites

11

Emergency Radiology of the Abdomen: The Acute Abdomen

of the low sensitivity of abdominal radiographs and US in the detection of urinary tract calculi, the role of unenhanced CT has become well established over the past 15 years. On CT, virtually all ureteral stones are radiopaque, regardless of their chemical composition. Uric acid stones have attenuation values of 300-500 Hounsfield units (HU), and calcium-based stones >1000 HU. In addition to the direct demonstration of a ureteral stone, secondary signs of ureterolithiasis, including hydroureter, hydronephrosis, perinephric stranding, and renal enlargement, may be visible (Fig. 9). Perinephric stranding and edema result from reabsorbed urine infiltrating the perinephric space along the bridging septa of Kunin. The more extensive the perinephric edema shown on unenhanced CT, the higher the degree of urinary tract obstruction. Focal periureteral stranding resulting from a local inflammatory reaction or irritation and induced by the passage of a stone helps to localize subtle calculi. Occasionally, a repeat CT examination using intravenous CM may be required, particularly if infectious complications are suspected. For the diagnosis of such complications (pyelonephritis), CT is helpful as it reveals a “striated nephrogram” after CM administration, as well as global enlargement of the kidney, renal and/or perirenal abscesses, or emphysematous pyelonephritis.

When no stone is detected, an alternative diagnosis must be established. Non-calculus urinary tract abnormalities causing symptoms of colic include acute pyelonephritis, renal cell carcinoma, acute renal vein thrombosis, spontaneous dissection of the renal artery, and renal infarction. Extraurinary diseases, such as retrocecal appendicitis, diverticulitis, SBO, pancreatitis, gynecological disorders, and retroperitoneal hemorrhage, may also simulate acute urinary colic.

Acute Pancreatitis An important disease causing upper abdominal pain is acute pancreatitis. US may be helpful for the demonstration of choledocolithiasis as a cause of acute pancreatitis and for the follow-up of known fluid collections. Since the CT findings correlate well with the clinical severity of acute pancreatitis, CT has become the imaging test of choice to stage the extent of disease (CT severity index of Balthazar) and to detect complications. The initial CT should be performed 48-72 h after disease onset (a CT examination performed too early in the course of the disease may not demonstrate any abnormality). Pancreatic enlargement due to interstitial parenchymal edema may progress to pancreatic exudate collecting in

a

b

Fig. 9 a-c. Right-sided urinary colic. Axial (a), coronal (b), and oblique (c) views of multidetector CT show dilatation of the proximal urinary tract due to a ureteral stone (arrow). Note slight secondary periuretral and perinephric stranding

c

12

Jean-Michel Bruel, Borut Marincek, Jay P. Heiken

a

c

b Fig. 10 a-c. Severe acute pancreatitis. Axial (a, b) and coronal (c) views of MDCT demonstrate an enlarged pancreatic area. a Only small areas of enhanced pancreatic parenchyma (p) remain within b extensive pancreatic necrosis (n). Note the tiny hyperattenuating calculi within the gallbladder and the lower bile duct (arrows). c The large amount of pancreatic necrosis contrasts sharply with the minimal extrapancreatic changes at this phase

the abdominal ligaments and potential spaces surrounding the pancreas. The pancreatic parenchyma may undergo necrosis or hemorrhage (Fig. 10). Severe pancreatitis is often complicated by infection of the necrotic site and/or thrombosis of the splenic and portal vein. Identifying infectious complications in patients with pancreatitis may be difficult; the demonstration of gas bubbles within peripancreatic collections is neither sensitive nor specific. CT-guided fluid aspiration for bacteriological examination is often the best technique to establish the presence of infection. Acute pancreatic and peripancreatic fluid collections may evolve into pseudocysts, which exhibit defined walls. A pseudocyst can erode peripancreatic vessels, thus causing bleeding or the formation of a pseudoaneurysm (Fig. 11).

Leaking Abdominal Aortic Aneurysm One of the most life-threatening diagnoses in patients with acute flank pain is a leaking abdominal aortic or iliac artery aneurysm. When a patient with suspected rupture of an abdominal aortic aneurysm is hemodynamically unstable, US is the initial imaging technique. The examination can be performed rapidly with portable equipment in the emergency room. However, para-aortic hemorrhage is poorly diagnosed by US. Instead, in hemodynamically stable patients, non-contrast-enhanced CT is the initial imaging test of choice as it is almost always able to demonstrate a para-aortic hematoma, if present, and may show additional findings helpful in establishing

Fig. 11. Recurrent pancreatitis. CT shows a pseudocyst of the pancreatic tail (PC) and a pseudoaneurysm of the gastroduodenal artery (PA). The splenic vein is indicated by the arrowhead

the diagnosis, such as a high-attenuating crescent sign. If endoluminal stent graft repair of the aorta is planned, contrast-enhanced CT should be performed.

Conclusions The imaging evaluation of patients with an acute abdomen has changed dramatically in the past decade. Plain abdominal radiographs largely have been replaced with US and CT. MDCT permits a rapid examination with high

Emergency Radiology of the Abdomen: The Acute Abdomen

diagnostic accuracy. Close cooperation with the referring physician prior to imaging remains essential for rapid and accurate diagnosis, as the character and location of the patient’s abdominal pain strongly influences the differential diagnosis and the choice of initial imaging test.

Suggested Reading Ahn SH, Mayo-Smith WW, Murphy BL et al (2002) Acute nontraumatic abdominal pain in adult patients: abdominal radiography compared with CT evaluation. Radiology 225:159-64 Balthazar EJ, Robinson DL, Megibow AJ, Ranson JH (1990) Acute pancreatitis: value of CT in establishing prognosis. Radiology 174:331-336 Freeman AH (2001) CT and bowel disease. Br J Radiol 74:4-14 Gore RM, Miller FH, Pereles FS et al (2000) Helical CT in the evaluation of the acute abdomen. AJR 174:901-913 Mindelzun RE, Jeffrey RB (1997) Unenhanced helical CT for evaluating acute abdominal pain: a little more cost, a lot more information. Radiology 205:43-47

13

Novelline RA, Rhea JT, Rao PM, Stuk JL (1999) Helical CT in emergency radiology. Radiology 213:321-339 Paulson EK, Jaffe TA, Thomas J et al (2004) MDCT of patients with acute abdominal pain: a new perspective using coronal reformations from submillimeter isotropic voxels. AJR 183:899-906 Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic appendagitis and its mimics. Radiographics 25:1521-34 Smith RC, Varanelli M (2000) Diagnosis and management of acute ureterolithiasis. AJR 175:3-6 Stoker J, van Randen A, Lameris W, Boermeester MA (2009) Imaging Patients with acute abdominal pain. Radiology 253:31-46 Taourel P, Kessler N, Lesnik A et al (2003) Helical CT of large bowel obstruction. Abdom Imaging 28:267-275 Urban BA, Fishman EK (2000) Tailored helical CT evaluation of acute abdomen. Radiographics 20:725-749 Werner A, Diehl SJ, Farag-Soliman M, Düber C. (2003) Multi-slice spiral CT in routine diagnosis of suspected acute left-sided colonic diverticulitis: a prospective study of 120 patients. Eur Radiol 13:2596-2603 Wiesner W, Khurana B, Ji H, Ros PR (2003) CT of acute bowel ischemia. Radiology 226:635-650

IDKD 2010-2013

Trauma of the Abdomen and Pelvis Philip J. Kenney1, Stuart E. Mirvis2 1 University 2 University

of Arkansas for Medical Sciences, Little Rock, AR, USA of Maryland School of Medicine, Baltimore, MD, USA

Introduction Trauma is a major health problem in all age groups but it is especially true in the young, due to high-velocity transportation, altercations, including with weapons and resulting in penetrating injuries, as well as falls and sportsrelated injuries. In addition, both the elderly and pregnant women are vulnerable to trauma. Improvements in the management of trauma include more rapid rescue, better organization of trauma centers, and advances in treatment. Current trends include increased non-operative management of trauma-related injuries, accurate imagingbased diagnosis, and greater emphasis on the efficient but cost-effective use of imaging. One aspect of the trend to non-operative care is the desire to avoid non-therapeutic surgery; this is possible if imaging can identify those patients who require surgery. Another is the realization that non-operative care can result in better long-term outcome, such as splenic salvage.

Computed Tomography vs. Ultrasound Controversy exists about the appropriate use of computed tomography (CT) vs. ultrasound (US), although each modality has its advantages and disadvantages [1, 2]. In general, CT has the best statistical accuracy for detecting, characterizing, and excluding injuries. In modern high-volume trauma centers, the CT apparatus must be located in the trauma suite such that even unstable patients can be examined quickly without compromise. This allows for the efficient use of CT in rapid and accurate diagnostics and obviates the need for outmoded studies, such as diagnostic peritoneal lavage. CT is also more reliable at excluding injury, allowing the patient to be discharged home and avoiding the expense of observation in hospital. However CT may be overused; indeed, in one study only three of 100 patients had alterations of clinical management due to follow-up CT [3]. US can detect significant injury which can then be appropriately treated; conversely, low-risk patients with normal sonograms may be observed and possibly avoid CT [2]. However, patients

with abnormal US findings often require further evaluation with CT. In a large study, US had 86% sensitivity and 98% specificity but with 43 false-negative and 23 indeterminate studies, including six splenic, one liver, one renal, one pancreatic, and one bowel injury [2]. In traumatized pregnant women, US should be the first examination as it can evaluate the pregnancy, documenting fetal death or viability. US is nearly as accurate in detecting the abnormal presence of fluid in pregnant patient as in non-pregnant patients [4]. If US shows fluid or other injury, CT is justified for further evaluation (Fig. 1). The best outcome for the fetus is assured by best care of the mother. The radiation risk is reasonable if there is lifethreatening injury, such that prompt diagnosis and treatment are paramount [5].

Urinary Tract The nearly universal use of CT has altered the assessment of urinary tract trauma. While significant hematuria has been shown to be the best indicator of urinary tract injury, presently the decision to perform CT has little to do with the presence or absence of hematuria. CT is a primary investigation, after standard radiographs, in those with significant mechanisms of injury or any signs or symptoms of significant injury. Intravenous urography has been replaced by CT (Fig. 2). The ability of US to evaluate renal injury is limited [6] whereas CT has excellent negative predictive value for renal injury. CT also accurately indicates the presence and type of renal injury [7]. Renal contusion appears as an illdefined region of diminished enhancement. Segmental renal infarction is identified as a wedge-shaped, well-defined area of non-enhancement. Renal artery occlusion can be accurately diagnosed by its complete lack of either contrast enhancement or excretion by the kidney, usually with little to no associated hematoma. Angiography is thus not needed, and conservative therapy is most often used today. Most renal injuries are lacerations, with simple lacerations limited to the cortex and deep lacerations extending into the collecting system, which may show extravasation. Delay scans of 2-10 min aid in demonstrating or excluding

15

Trauma of the Abdomen and Pelvis

a

b

Fig. 1 a-c. A pregnant woman suffered a high-speed motor vehicle collision. a US demonstrates intrauterine fetus (heart motion documented) and free pelvic fluid. b US shows perisplenic fluid with hypoechoic defect. c CT confirms splenic laceration; note the higher density of the perisplenic hematoma compared to the rim of fluid about liver (sentinel clot sign)

c

a

b

Fig. 2 a, b. Hematuria and left upper quadrant pain after a footballrelated injury. a Intravenous urogram shows no abnormality. b Subsequent CT for persistent pain showed free fluid in the pelvis and extensive splenic laceration with extensive “blush”. Surgery confirmed a grade 4 splenic injury

extravasation (Fig. 3), although in most cases small amounts of extravasation will resolve with conservative therapy. Subcapsular hematoma is delimited by the renal cortex and may deform the renal surface; perinephric hematoma extends from the renal surface to fill Gerota’s space but does not deform the renal contour, although it may displace the kidney. CT is excellent at demonstrating the extent of hematoma and in evaluating enlargement on

follow-up scans [7]. Renal fracture indicates a single complete fracture plane, often extending through the collecting system; multiple planes of disruption are seen in a shattered kidney. CT can also diagnose avulsion of the ureteropelvic junction (UPJ) or ureteral injury, demonstrating lack of opacification of the ureter, retroperitoneal water attenuation collections adjacent to the pelvis or ureter, and possibly extravasation of contrast on delay scans (Fig. 4) [7].

16

a

Philip J. Kenney, Stuart E. Mirvis

a

b b

Fig. 3 a, b. Hematuria after fall from a power line. a Initial CT shows left renal laceration with perinephric hematoma. b Delay image shows no leak from the collecting system

Ureteral injuries, including UPJ avulsion, are uncommon. They can occur either with penetrating trauma or high-velocity blunt trauma and have no specific signs or symptoms, but can be detected with CT. Routine CT scans show subtle suggestive signs, such as perinephric and peripelvic stranding or fluid, that indicate the need for delay scans, if not routinely done, to demonstrate extravasation. In a study based on over 4,000 trauma patients, CT enabled the correct identification of seven of eight UPJ avulsions [8]. The AAST Organ Injury severity scale for the kidney includes lesions with different appearances in each category (1: contusion, small subcapsular hematoma; 2: <1 cm laceration without extravasation; 3: >1 cm laceration without extravasation; 4: deep laceration with extravasation or main renal artery or vein injury; 5 shattered kidney or UPJ avulsion) and has been shown to correlate with need for surgery and outcome [9]. Urethral injuries are predominantly seen in males. Anterior urethral ruptures most commonly occur due to straddle injury. Posterior urethral ruptures most often are

Fig. 4 a, b. Routine trauma CT image shows fluid and stranding about the right ureter. a Note that contrast filling of the ureter has not yet occurred. Delay image shows extravasation medial to the kidney typical, of UPJ avulsion. b Note the intact parenchyma

due to compressive force and resultant pubic bone fractures, although both anterior and posterior urethral injury can result from penetrating injury. Retrograde urethrography is the only accurate diagnostic imaging procedure. If a urethral injury is strongly suspected, a urethrogram should be performed before passage of a catheter (Fig. 5). However, in patients with moderate risk, a urethral catheter may be gently passed so that the patient may go on to CT. A pericatheter urethrogram may then be performed after any other injuries have been stabilized. Five types of urethral injuries are recognized. In type 1, the posterior urethra is stretched but intact; in type 2, there is a tear of the membranous urethra above the urogenital diaphragm; in type 3, the posterior urethral tear is above and below the urogenital diaphragm; type 4 is defined as a bladder-neck injury and type 5 as an anterior urethral injury [10]. Bladder injuries consist of contusions and ruptures; classically, they have been detected with standard radiographic cystography (Fig. 6). They may be extraperitoneal, most commonly, intraperitoneal, less commonly, or combined in about 5%. While CT with only intravenous contrast may fail to identify extravasation from

17

Trauma of the Abdomen and Pelvis

a

b

Fig. 5. Blunt trauma resulted in pubic rami fractures. Retrograde urethrogram reveals type 3 posterior urethral rupture

Fig. 7 a, b. Gross hematuria following motor-vehicle collision resulting in extensive pelvic fractures. a Standard CT shows pelvic fluid but no extravasation. b CT cystogram documents extraperitoneal bladder rupture (note the clot in the bladder)

Fig. 6. Gross hematuria after gunshot wound to the pelvis. Standard cystogram shows extraperitoneal rupture (X marks entry site, O exit)

a ruptured bladder, several studies have shown very high accuracy for CT-cystography (Fig. 7), which is now the standard in our institutions. In patients with suspected bladder rupture (primarily those with gross hematuria, over 25 RBC/hpf, with pelvic fractures or unexplained pelvic fluid), a standard CT with the

bladder catheter clamped is performed. If there is no extravasation, the bladder is drained and then re-filled with 300-500 mL of dilute contrast and the pelvis rescanned. Bladder ruptures are virtually always associated with fluid or hematoma in the pelvis, but such blood or fluid may be due to splenic or other injuries or to pelvic fracture. Extravasation confined to the lower pelvis and not outlining bowel loops (and which may extend up the retroperitoneum) indicates extraperitoneal rupture, which most often is managed conservatively. Extravasation high near the dome and outlining bowel loops or extending to the gutters or higher indicates intraperitoneal rupture, which is more often managed surgically [11]. In a study of 495 patients with potential pelvic injuries, CT-cystography detected 98% of the bladder injuries while standard cystography detected 95%. Of the patients with bladder injury (65% extraperitoneal, 35% intraperitoneal, 5 combined), 89% had gross hematuria and pelvic fracture, 9% had gross hematuria with no pelvic fracture, and one patient had microscopic hematuria and pelvic fracture [12]. CTcystography carried out with multidetector CT allows for reformatting, which can more clearly demonstrate the point of leakage and allows more accurate characterization of the type of injury [13].

18

Bowel and Mesenteric Injuries Bowel and mesenteric injuries are found in about 5% of patients undergoing surgery for trauma and are seen in 0.7% of all traumatized patients [1, 14]. The mechanism of injury is direct compressive force, including from seatbelts, although deceleration may play a role. Morbidity and mortality can occur, with peritonitis and abscess resulting if the injury is missed. Clinical signs and symptoms are non-specific. Although diagnosis by CT is not as straightforward as is the case for other abdominal organ injuries, CT is the most accurate diagnostic modality, with >90% sensitivity and specificity reported [14, 15]. The use of orally administered contrast is now somewhat controversial; while extravasation of oral contrast can be a very specific sign of bowel injury, it is rarely seen. Contrast administration delays performance of the scan, and most bowel injuries will be evident based on other signs. There is no one CT sign that is both sensitive and specific for bowel or mesenteric injury. Focal bowel wall thickening, mesenteric stranding, interloop fluid, and hematoma are common but less specific, particularly for surgically important injuries (Fig. 8). Active bleeding, vessel beading, abrupt termination of mesenteric vessels, and bowel wall defect are more specific but less sensitive signs [16]. Active bleeding is seen as a focal extraluminal collection with attenuation similar to that of the aorta at the same level and different from the adjacent organs. Free air is considered a good sign of perforated bowel, but in fact it has limited value. It is infrequently seen in those with bowel injury and may represent air tracking into the peritoneum from thoracic injuries. In a study reported in 2008, free air had a sensitivity of 24% albeit a specificity of 95%. There were three false-positives with intraperitoneal air instead resulting from supradiaphragmatic or bladder injuries [16]. If a single finding

Fig. 8. Motor-vehicle collision. Focal hematoma and thickening of cecum; at surgery, cecal laceration found

Philip J. Kenney, Stuart E. Mirvis

is noted, the likelihood of injury is low; a combination of findings, particularly free fluid without obvious source in combination with focal bowel wall thickening and/or mesenteric stranding, is very suggestive of bowel injury and such patients should be explored or followed very carefully [14-16]. In our practice, we have found that performing a repeat abdominal-pelvic CT 4-6 h after the admission scan can be helpful in patients with suspicious but non-diagnostic findings for full-thickness bowel injury, by demonstrating injury progression such as the development of free air, increasing intraperitoneal fluid, or stability of findings. Of course, management decisions are made in conjunction with any evolution of the clinical findings.

Splenic Injuries The spleen is the most frequently injured abdominal organ in blunt trauma. There may be signs of blood loss or left upper quadrant pain, but the diagnosis largely rests on imaging or surgical exploration. A trend to nonoperative management is supported by evidence that long-term health is better in those who have had splenic function preserved. This necessitates accurate noninvasive diagnosis and is aided by signs predictive of the success or failure of conservative management. Splenic injuries can cause free fluid, perisplenic or elsewhere, which can readily be detected by sonography. Splenic injury may alter echo-texture: lacerations may be anechoic if there is rapid bleeding, but more commonly are more echogenic than normal spleen [2]. With such findings on sonography, the decision whether to further evaluate with CT or to proceed to surgery can be made on clinical grounds. Splenic injuries may be missed by sonography, particularly if they are not associated with free fluid. In one large study, there were 43 false-negative sonograms, including six splenic ruptures that required surgery [2]. CT is quite sensitive in the detection of splenic injuries [17]. Subcapsular hematoma is seen as a crescentic, lowattenuation, peripheral rim; intraparenchymal hematoma as a rounded area within the spleen with low attenuation and no enhancement. Lacerations are common, appearing as linear or branching low-attenuation lesions that often extend to the surface; if so, they are often associated with perisplenic or free fluid. Hemoperitoneum tends to be of higher attenuation close to the source of bleeding; thus, when the spleen is the source, the collection adjacent to the spleen may be higher in attenuation than elsewhere, a finding referred to as the sentinel clot sign. Lacerations may involve the vasculature. There can be devascularization of the spleen by hilar injury, or active extravasation into the peritoneal cavity, or a confined area of extravasation (pseudoaneurysm) (Fig. 9). Both types of extravasation indicate that non-operative management may not succeed, although angiographic embolization may control the bleeding and allow splenic salvage [18].

19

Trauma of the Abdomen and Pelvis

a

b

Fig. 9 a, b. Blunt trauma. a Routine trauma CT image shows area of hyperdensity (arrowhead). b Delay image shows that the hyperintensity is no longer visible; the lesion has become isodense with blood pooling, indicating confined pseudoaneurysm rather than free active bleeding

A number of schemes have been devised to grade splenic injury on CT in an attempt to predict outcome, with variable correlation with the need for surgery [1]. One of the commonest is the AAST scoring system. In a large study, failure of non-operative management correlated with splenic injury grade: the failure rate was <10% with grades 1 or 2, while one-third of the grade 4 injuries and three-fourths of the grade 5 injuries required surgery [19]. Nevertheless, in occasional cases of low-grade injury, the patient suffered delayed rupture, while some high-grade injuries have been successfully managed conservatively. Attempts have been made to develop CTbased criteria that may be more predictive: one referred to three key features: devascularization, laceration of >50% of the parenchyma; contrast blush >1 cm; or large hemoperitoneum; however, further study showed that this approach also had limited predictive value with poor sensitivity although fair specificity [20]. The additional finding of traumatic pseudoaneurysm or active extravasation (which does not confer a specific stage in the AAST scoring system) increased the likelihood of failure of nonoperative management, regardless of grade [21]. Delayed images can help distinguish between active bleeding, which persists as a hyperdense area, and confined vascular injury (pseudoaneurysm), which washes out [22]. Patients with active bleeding are more likely to require surgery or other forms of intervention.

of the liver and the difficulty in clearly imaging all portions of the organ ultrasonographically. Injuries to the liver include contusion, seen on CT as an ill-defined area of low attenuation; subcapsular hematoma, a crescentic collection limited by the capsule; and intraparenchymal hematoma, a collection of blood within a liver laceration. Laceration is commonest, seen as linear or branching low-attenuation regions, sometimes with jagged margins, that can extend to the hepatic surface or to vessels. Superficial lacerations are <3 cm in size. Periportal low attenuation is usually edema, a distended inferior vena cava and renal veins, and subserosal edema of the gallbladder wall, but on occasion may represent blood tracking along the portal veins (Fig. 10). It is rare that periportal low attenuation is the only sign of liver injury, and patients with only this finding should be managed conservatively [24]. Liver injuries may require surgery but most can be managed non-operatively. The liver, with its dual blood supply, is relatively resistant to infarction and has considerable functional reserve. Grading systems

Hepatic Injuries The liver is the second most frequently injured abdominal organ, accounting for about 20% of abdominal injuries [1, 17]. The right lobe is more often affected than the left, with the posterior right lobe the most commonly injured segment. Hepatic injuries may be associated with intraperitoneal hemorrhage, but injury may be confined to the liver, or hemorrhage may be limited by an intact capsule. Lacerations involving the bare area may be associated with retroperitoneal hematoma. US may show liver lacerations, which appear similar to splenic injuries but this modality has limited sensitivity (67%, compared to 93% for CT) [23]. This is in part due to the large size

Fig. 10. Blunt trauma, shock, and aggressive resuscitation followed by trauma CT. Note the periportal low attenuation tracking throughout the liver with intact parenchyma and no perihepatic hematoma

20

(e.g., AAST) have less direct correlation with the need for intervention than is the case for the spleen. Although not included in the AAST scheme, active extravasation may predict the need for surgery or angioembolization. Sub-classification of extravasation can be useful: extravasation into the peritoneal cavity is highly correlated with the need for intervention; intraparenchymal extravasation with significant hemoperitoneum may also require intervention; extravasation limited within a hepatic hematoma without hemoperitoneum usually can be managed conservatively. Actually, the success rate of non-operative management is >80%, and clinical signs of hemodynamic instability dictate the need for intervention more than imaging features. However in one recent study of 214 patients with hepatic injury, all 14 who showed intraperitoneal contrast extravasation on CT required surgery [25]. Massive hemoperitoneum in six compartments also correlated independently with the need for surgical intervention. Gallbladder injuries occur in <2% of major blunt abdominal traumas and are usually seen with concurrent liver injury. Injuries to the gallbladder include wall contusion, intraluminal hemorrhage, laceration, and partial or complete avulsion. CT findings may consist of focal or diffuse wall thickening, pericholecystic fluid (blood and/or bile), ectopic position, intraluminal clot or mucosal flaps, and focal mass effect on the adjacent duodenum (Fig. 11). Full wall thickness tears result in a collapsed lumen [26].

Fig. 11. Gallbladder rupture. Coronal reformation in blunt trauma patient shows ectopic location and tear in gallbladder. Note enhancement of the gallbladder mucosa. Hematoma fills the gallbladder fossa

Philip J. Kenney, Stuart E. Mirvis

Adrenal Injuries Adrenal injuries are uncommon, seen in about 2% of patients with blunt abdominal trauma, and are rarely isolated, but their presence indicates a high-energy mechanism. In a review of 73 cases, 71% were right adrenal only, 15% left, and 8% bilateral [27]. The right adrenal is more prone to injury due to its location between the liver and spine, resulting in crush injury; also, the short right adrenal vein can transmit increased pressure. Acute adrenal hematomas are usually 2-4 cm in diameter, round or oval lesions with relatively high attenuation (40-60 HU, mean 55), often with stranding. Active bleeding may be seen and correlates with poor outcome. If there is any concern that a lesion represents an adrenal adenoma or pre-existing mass, such as adenoma, repeat exam with pre- and post-contrast technique can be done, or the patient simply followed-up at 6-8 weeks as hematomas will decrease in size and attenuation.

Pancreatic and Duodenal Injuries Pancreatic and duodenal injuries are also uncommon, together accounting for about 2% of all abdominal injuries [28]. The mechanism of injury is anteroposterior compression, with compression against the spine leading to the actual injury; thus, pancreatic injuries usually are at the mid-body. Like adrenal injuries, they are also often associated with other injuries, including hepatic, vascular, splenic, renal, and gastric. Morbidity and mortality are relatively high in part due to the multisystem trauma seen in these patients. Significant complications are common and include pancreatitis, pseudocysts, abscess, fistulas, and pneumonia. A delay in diagnosis is not unusual as the initial findings can be subtle but it contributes to high morbidity and mortality. Clinical signs include upper abdominal pain, with laboratory signs of leukocytosis and elevated amylase, but amylase also may be normal for 2-48 h in up to 40% of patients with pancreatic injuries [28]. Duodenal contusion may be present when there is hematoma or edema limited to the duodenal wall, perhaps with intramural gas and focal mural thickening. Duodenal perforation should be diagnosed when there is an extraluminal retroperitoneal collection of contrast, gas, or fluid, or loss of continuity of the wall. Stranding of the retroperitoneal fat can be seen with either condition. While use of oral contrast may aid in the diagnosis, it is not absolutely necessary; if needed, additional images can be obtained with oral contrast if it was not used initially. It has been reported that the pancreas appears normal on CT for the first 12 h after injury in some 40% of patients, but this conclusion was based on older CT technology and most trauma scans do not include a pancreatic parenchymal phase. However, a recent multicenter

Trauma of the Abdomen and Pelvis

Fig. 12. High-speed deceleration injury with blunt-force trauma. Note the discrete linear low attenuation defect crossing the entire diameter of the body of the pancreas, consistent with a pancreatic transection. Given the extent of the defect, involvement of the pancreatic duct is likely

study using 16- or 64-slice scanners showed a sensitivity of 60 and 47% and a specificity of 95 and 90%, respectively, in 206 subjects [29]. Fracture or laceration of the pancreatic parenchyma, active hemorrhage in the pancreas, and hematoma between the splenic vein and pancreas are specific findings. Lacerations appear as hypoattenuating linear lesions; if crossing more than half the diameter, pancreatic duct injury should be suspected (which correlates with increased complications) (Fig. 12). Hematoma and contusion are diffuse or localized, ill-defined, mixed-attenuation regions within the pancreas [28].

References 1. Novelline RA, Rhea JT, Bell T (1999) Helical CT of abdominal trauma. RCNA 37:591-612 2. Dolich MO, McKenney MG, Varela JE et al (2001) Ultrasounds for blunt abdominal trauma. J Trauma 50:108-112 3. Shapiro MJ, Krausz C, Durham RM et al (1999) Overuse of splenic scoring and computed tomographic scans. J Trauma 47:651-658 4. Goodwin H, Holmes JF, Wisner DH (2001) Abdominal ultrasound examination in pregnant blunt trauma patients. J Trauma 50:689-693 5. Lowdermilk C, Gavant ML, Qaisi W et al (1999) Screening helical CT for evaluation of blunt traumatic injury in the pregnant patient. Radiographics 19:S243-255. Kenney PJ Invited commentary S256 6. McGahan JP, Richards JR, Jones CD et al (1999) Use of ultrasonography in the patient with acute renal trauma. JUM 18:207-213 7. Kawashima A, Sandler CM, Corl FM et al (2001) Imaging of renal trauma: a comprehensive review. Radiographics 221: 557-574

21

8. Ortega SJ, Netto FS, Hamilton P et al (2008) CT scanning for diagnosing blunt ureteral and uretropelvic junction injuries. BMC Urology 8:3-7 9. Santucci RA, McAninch JW, Safir M et al (2001) Validation of the American Association for the Surgery of Trauma organ injury severity scale for the kidney. J Trauma 50:195-200 10. Goldman SM, Sandler CM, Corriere JN et al (1997) Blunt urethral trauma: a unified anatomical mechanical classification. J Urol 157:85-89 11. Morgan DM, Nallamalla LK, Kenney PJ et al (2000) CT cystography: radiographic and clinical predictors of bladder rupture. AJR 174:89-95 12. Quagliano PV, DeLair SM, Malhotra AK (2006) Diagnosis of blunt bladder injury: a prospective comparative study of CT Cystography and conventional retrograde cystography. J Trauma 61:410-422 13. Chan DPN, Abujudeh HH, Cushing GL et al (2006) CT Cystography with multiplanar reformation for suspected bladder rupture. AJR 187:1296-1302 14. Malhotra AK, Fabian TC, Katsis SB et al (2000) Blunt bowel and mesenteric injuries: the role of screening CT. J Trauma 48:991-998 15. Butela ST, Federle MP, Chang PJ et al (2001) Performance of CT in Detection of bowel injury. AJR 176:129-135 16. Atri M, Hanson JM, Grinblat L et al (2008) Surgically important bowel and or mesenteric injury in blunt trauma: accuracy of multidetector CT for evaluation. Radiology 249:524-533 17. West OC (2000) Intraperitoneal abdominal injuries. ARRS Categorical course syllabus. ARRS, Reston, VA, p 87 18. Davis KA, Fabian TC, Croce MA et al (1998) Improved success in nonoperative management of blunt splenic injuries: embolisation of splenic artery pseudoaneurysms. J Trauma 44:1008-1013 19. Peitsman AB, Heil B, Rivera L et al (2000) Blunt splenic injury in adults: multi-institutional study of the Eastern Association for the Surgery of Trauma. J Trauma 49:177-187 20. Cohn SM, Arango JL, Myers JG et al (2009) Computed tomography grading systems poorly predict the need for intervention after spleen and liver injuries. American Surgeon 75:133-139 21. Gavant ML, Schurr M, Flick PA et al (1997) Predicting clinical outcome of nonsurgical management of blunt splenic injury: using CT to reveal abnormalities of splenic vasculature. AJR 168:207 22. Anderson SW, Varghese JC, Lucey BC et al (2007) Blunt splenic trauma: delayed phase CT for differentiation of active hemorrhage from contained vascular injury in patients. Radiology 243:88-95 23. Richards JR, McGahan JP, Pali MJ et al (1999) Sonographic detection of blunt hepatic trauma. J Trauma 47:1092-1097 24. Yoon W, Jeong YY, Kim JK et al (2005) CT in blunt liver trauma. Radiographics 25:87-104 25. Fang J-F, Wong Y-C, Lin B-C et al (2006) CT risk factors for the need of operative treatment in initially hemodynamically stable patients after blunt hepatic trauma. J Trauma 61:547-554 26. Wittenburg A, Minotti AJ (2005) CT diagnosis of traumatic gallbladder injury. AJR 185:1573-1574 27. Sinelnikov AO, Abujudeh HH, Chan D, Novelline RA (2007) CT manifestations of adrenal trauma: experience with 73 cases. Emerg Radiol 13:313-318 28. Linsenmaier U, Wirth S, Reiser M et al (2008) Diagnosis and classification of pancreatic and duodenal injuries in emergency radiology. Radiographics 28:1591-1601 29. Phelan HA, Velmahos GC, Jurkovich J et al (2009) An evaluation of multidetector CT in detecting pancreatic injury: results of a multicenter AAST study. J Trauma 66:641-647

IDKD 2010-2013

Diseases of the Esophagus and Stomach Marc S. Levine1, Ahmed Ba-Ssalamah2 1 Department 2 Department

of Radiology, Penn Radiology at the Hospital of the University of Pennsylvania, Philadelphia, PA, USA of Radiology, Medical University of Vienna, General Hospital of Vienna, Wien, Austria

Introduction The esophagus and stomach are susceptible to a wide spectrum of diseases, including benign and malignant tumors, inflammatory diseases, and other conditions. For the diagnosis of this large variety of disorders, multimodality imaging is required. Barium studies, particularly double-contrast studies, continue to have a major role in the diagnostic work-up of inflammatory diseases and in post-operative follow-up, whereas cross-sectional imaging studies, particularly multidetector computed tomography CT (MDCT), is used in the pre-operative staging of oncological disorders. In this chapter, we review the most frequent diseases and describe the use of different imaging modalities for their diagnostic work-up.

Fig. 1. Candida esophagitis. Upright, left posterior-oblique spot image from doublecontrast esophagogram shows multiple, discrete, plaque-like lesions in the mid-esophagus. Note the linear configuration of the lesions and their separation by segments of normal, intervening mucosa. These findings are characteristic of fungal esophagitis

Gastroesophageal Reflux Disease Mild reflux esophagitis may be manifested on doublecontrast studies by small, shallow ulcers or granularity of the mucosa in the distal esophagus [1]. In advanced disease, the esophagus can have an irregular contour, with serrated margins and decreased distensibility from ulceration, edema, and spasm. Subsequent scarring can lead to the development of smooth, tapered, or ring-like peptic strictures in the distal esophagus, almost always above a hiatal hernia [2]. Barrett’s esophagus is a well-recognized complication of reflux esophagitis, and it is associated with an increased risk of esophageal adenocarcinoma. The classic radiological features of Barrett’s esophagus include a distinctive reticular pattern of the mucosa or a high esophageal stricture or ulcer occurring at a discrete distance from the gastroesophageal junction [3].

epidemic has led to a more fulminant form of candidiasis, characterized by a “shaggy” esophagus that has a grossly irregular contour due to multiple plaques and pseudomembranes [5]. In contrast, herpes esophagitis is manifested by small, shallow ulcers (Fig. 2) [6] whereas esophagitis due to cytomegalovirus (CMV) or human immunodeficiency virus (HIV) can be associated with the development of giant, flat ulcers [7]. Since CMV ulcers are treated with antiviral agents, and HIV ulcers with steroids, endoscopy is required to differentiate CMV- from HIV-mediated esophagitis before treatment in these patients is instituted.

Infectious Esophagitis

Drug-Induced Esophagitis

By far the most common cause of opportunistic esophageal infection is Candida albicans. Candida esophagitis is manifested on double-contrast studies by discrete, plaque-like lesions separated by normal mucosa (Fig. 1) [4]. The AIDS

Drug-induced esophagitis is caused by various oral medications, including tetracycline, doxycycline, potassium chloride, quinidine, alendronate sodium, aspirin, and other non-steroidal anti-inflammatory drugs (NSAIDs).

23

Diseases of the Esophagus and Stomach

Fig. 2. Herpes esophagitis. Upright, left posterior-oblique spot image from doublecontrast esophagogram shows multiple small ulcers surrounded by radiolucent mounds of edema in the midesophagus. These findings are characteristic of viral (especially herpes) esophagitis

These patients often have a history of ingesting the pills with little or no water immediately before retiring. As a result, the pills tend to lodge in the mid-esophagus where it is compressed by the aortic arch or left main bronchus. This can result in contact esophagitis, manifested by small, discrete ulcers in the mid-esophagus [8]. Affected individuals may present with odynophagia, but there is rapid clinical improvement after withdrawal of the offending agent.

Fig. 3. Adenocarcinoma of the esophagus. Prone, right anterioroblique spot image from single-contrast esophagogram shows a polypoid mass (arrows) in the distal esophagus in this patient with adenocarcinoma arising in Barrett’s esophagus

superficially spreading carcinomas may be manifested by poorly defined nodules, producing confluent nodularity of the mucosa [10]. In contrast, the lesions of advanced esophageal carcinomas are polypoid (Fig. 3), ulcerated, or infiltrating, with irregular luminal narrowing and shelf-like borders. Rarely, these tumors have a varicoid appearance due to submucosal spread of tumor; in such cases, they can be mistaken for esophageal varices.

Erosive Gastritis Idiopathic Eosinophilic Esophagitis Idiopathic eosinophilic esophagitis (IEE) usually occurs in adolescent or young men with a long history of compensated dysphagia and occasional food impactions. These patients often have an atopic history, asthma, or peripheral eosinophilia. Barium studies may reveal a “ringed esophagus”, with multiple distinctive ring-like indentations, or a “small-caliber esophagus”, with diffuse loss of distensibility of the esophagus in the absence of a discrete stricture [9]. The diagnosis can be confirmed by endoscopic biopsies showing >20 eosinophils per highpowered field. These patients usually have a dramatic positive response to oral steroids or inhaled steroid preparations.

Erosive gastritis is usually manifested on double-contrast studies by varioliform erosions with punctate or slit-like collections of barium surrounded by radiolucent mounds of edema. Varioliform erosions tend to be located in the gastric antrum and are often aligned on the crests of the folds. Aspirin and other NSAIDs are by far the most common cause of erosive gastritis. Occasionally, NSAID-induced erosive gastritis may also manifest as distinctive linear or serpiginous erosions clustered together on or near the greater curvature of the gastric body [11]. It has been postulated that these erosions result from localized mucosal injury, as the dissolving NSAID tablets collect by gravity in the most dependent portion of the stomach.

Esophageal Carcinoma

Helicobacter Pylori Gastritis

Early esophageal cancers classically appear on doublecontrast studies as small, protruded tumors [10]. They may be plaque-like or small polypoid lesions. Other,

Gastritis caused by the bacterium Helicobacter pylori can be diagnosed on barium studies by the presence of thickened folds in the antrum, body, or, less commonly, the

24

Marc S. Levine, Ahmed Ba-Ssalamah

change in its shape. Usually, ulcer healing produces a visible scar manifested by a central pit or depression, radiating folds, and/or retraction of the adjacent gastric wall [13].

Gastric Carcinoma

Fig. 4. Helicobacter pylori gastritis. Left posterior-oblique spot image from double-contrast upper gastrointestinal examination shows thickened, irregular folds in the gastric body due to chronic H. pylori gastritis

fundus of the stomach (Fig. 4) [12]. Other patients with H. pylori infection may have a polypoid form of gastritis, with grossly thickened, lobulated folds such that the lesions resemble those of Menetrier’s disease, lymphoma, or a submucosally infiltrating carcinoma [12]; thus, endoscopy and biopsy are required for a definitive diagnosis.

Gastric Ulcers Benign gastric ulcers classically appear en face as round or ovoid collections of barium, often surrounded by a smooth mound of edema or thin, straight folds radiating to the edge of the ulcer crater [13]. When viewed in profile, benign ulcers project beyond the contour of the adjacent gastric wall and are often associated with an ulcer mound or collar. In contrast, malignant gastric ulcers appear en face as irregular ulcer craters within a discrete mass, sometimes associated with nodularity or clubbing of adjoining folds due to tumor infiltration of the folds [13]. When viewed in profile, malignant ulcers project inside the lumen within a mass that forms acute angles with the gastric wall rather than the obtuse, gently sloping angles expected for a benign mound of edema. Most benign ulcers are located on the lesser curvature or posterior wall of the gastric antrum or body [13]. Occasionally, benign gastric ulcers may occur on the greater curvature of the distal stomach, in which case the vast majority are caused by aspirin or other NSAIDs [13]. As these NSAID-induced greater-curvature ulcers enlarge, they can penetrate inferiorly into the transverse colon, producing a gastrocolic fistula [14]. Ulcer healing may be seen as a decrease in the size of the ulcer or a

Advanced gastric carcinomas may appear on barium studies as polypoid, ulcerated, or infiltrating lesions. Other primary scirrhous carcinomas can have a “linitis plastica” appearance, with luminal narrowing, irregularly thickened folds, and nodularity of the mucosa [15]. Scirrhous carcinomas classically involve the gastric antrum, but 40% of these lesions are confined to the gastric body or fundus (Fig. 5) [15]. Early gastric cancers may be manifested by small polypoid or ulcerated lesions. However, in the western world, the vast majority of patients with gastric carcinoma already have advanced lesions at presentation. As a result, early gastric cancer is unlikely to be detected as long as barium studies are performed predominantly on symptomatic patients [16].

Gastric Lymphoma Chronic H. pylori gastritis can lead to the development of mucosa-associated lymphoid tissue (MALT) in the stomach. This lymphoid tissue is the precursor of low-grade, B-cell gastric MALT lymphomas, which, if untreated, may undergo blastic transformation to more high-grade lymphomas. Gastric MALT lymphomas may sometimes be recognized on double-contrast studies by variably sized, rounded, confluent nodules in the stomach [17]. In contrast, advanced gastric lymphomas may be manifested by thickened folds, multiple submucosal masses, ulcerated bull’s-eye lesions, or giant, cavitated lesions.

Fig. 5. Scirrhous adenocarcinoma of the stomach. Front spot image from double-contrast upper gastrointestinal examination shows irregular narrowing of the gastric body and fundus due to infiltration of the wall by tumor. Note transition (arrows) to uninvolved gastric antrum distally. About 40% of scirrhous carcinomas are confined to the body or fundus of the stomach with antral sparing

Diseases of the Esophagus and Stomach

Gastrointestinal Stromal Tumors Gastrointestinal stromal tumors (GISTs) are the most common mesenchymal tumors of the gastrointestinal tract. Approximately 70% of all GISTs are found in the stomach and only 2-5% originate from the esophagus. GISTs have a wide clinical spectrum, ranging from benign, incidentally detected nodules to large malignant tumors, and must be distinguished from other mesenchymal tumors [18]. The most frequent symptom related to gastric or esophageal tumor is dysphagia or heartburn; thus, most patients undergo endoscopy of the esophagus, stomach and duodenum, with simultaneous biopsy if tumor is seen [19]. If the histopathological results reveal a tumor, accurate staging is required. Endoscopic ultrasound (EUS) can depict the normal gastric and esophageal wall with its five-layered internal structures, thus allowing detailed evaluation of the depth of tumor penetration even in early-stage disease. EUS is useful in the diagnostic work-up of early cancer, as it can distinguish between T1a tumors, in which only mucosectomy is needed, and T1b tumors, in which a complete resection is indicated [20]. However, because ultrasound penetration is not deep enough when transducers with higher frequencies are used to visualize fine structures, evaluation of deep tumor infiltration may be difficult and assessment of metastases may be limited by the finite depth of penetration [21]. EUS also is examiner-dependent, time-consuming, and unable to pass stenotic tumors. These limitations can be overcome using multidetector CT technology (MDCT), with its ability to cover a large volume in a very short scan time with a single breathhold. Thin collimation and isotropic voxels allow imaging of the entire esophagus and stomach with high-quality multiplanar reformation and 3D reconstruction (Fig. 6)

Fig. 6. Hydro-MDCT of the esophagus in coronal reformation to follow the course of the esophagus, demonstrating normal wall thickening of the esophagus (≤3 mm) and homogeneous enhancement

25

[22, 23]. Furthermore, MDCT with water filling (hydroMDCT, HMDCT) provides information about esophageal and gastric wall infiltration, extramural extent of disease, lymph node involvement, and distant metastases (Fig. 7) [23, 24]. Inadequately distended hollow viscera on CT may hide large lesions and may even mimic pseudolesions. Thus, optimal distention of the esophagus and stomach is a necessary prerequisite for achieving good diagnostic imaging. When water is used as the oral contrast agent, subtle pathology is easier to visualize. Gas or CO2 resulting from the administration of effervescent granules can be used for hollow-organ distention alone or in combination with water. The use of negative rather than positive contrast media is preferred, especially if CT angiography images are needed. Subtle pathology is easier to visualize, especially when an adequate intravenous contrast material bolus is administered [23, 24]. Threedimensional reconstructions of CT data sets with multiplanar reconstruction, curved planar reformations, or other protocols are mandatory to exploit the full potential of MDCT. Three-dimensional virtual gastroscopy provides an endoluminal image similar to conventional fiberoptic gastroscopy. Therefore, HMDCT is a valuable tool for the complete staging of gastric and esophageal tumors and serves as an adjunct to endoscopy. However, EUS and MDCT are anatomically based diagnostic techniques with certain drawbacks. These include limited sensitivity with false-negative findings due

Fig. 7. Hydro-MDCT of the esophagus in coronal reformation shows a huge mass with inhomogeneous enhancement in the mediastinum arising from the esophageal wall, with infiltration of the left main bronchus (arrow) and the left diaphragm (arrowhead), with regard to the T4 tumor. Note the pleural effusion

26

to non-enlarged, tumor-involved lymph nodes, and limited specificity with false-positive findings due to enlarged lymph nodes not involved by tumor. Furthermore, after neo-adjuvant chemotherapy and re-evaluation, it is not possible with these techniques to distinguish between fibrotic changes or remaining vital tumor due to their morphological similarities [25].Thus, there is an urgent need for additional functional examination techniques. Positron emission tomography (PET) yields physiological information that provides a means to diagnose cancer based on altered tissue metabolism. PET takes advantage of the principle that biochemical changes often precede or are more specific than the structural changes associated with any given disease process [26]. Therefore, PET offers the potential to show early esophageal cancer or small lymph node metastases before any structural abnormality is detectable or to exclude the presence of tumor in an anatomically altered structure. For example, FDG-PET can detect metastatic lymph nodes that are not enlarged on CT, and can help differentiate pathological from non-specifically enlarged lymph nodes, which usually show no uptake of FDG. Tumor uptake of FDG, measured as the maximal standardized uptake value (SUVmax) in FDG-PET, even provides a quantitative estimate of tumor aggressiveness [27]. Recent studies demonstrated that FDG-PET can be used not only for pre-treatment staging, but also for assessment of treatment response, detection of recurrence, and prediction of survival in patients with adenocarcinoma of the esophagus or stomach [27]. However, for mucinous and signet-ring cell adenocarcinoma of the stomach, it is less useful. This may be due to low or absent FDG activity in these tumors, which is the result of a high content of metabolically inert mucus, leading to a reduced FDG concentration. Another reason could be the lack of expression of the glucose transporter Glut-1 on the cell membrane of most signet-ring cell and mucinous adenocarcinomas [28]. The spatial resolution of FDG-PET is lower than that of CT scans. When metastatic nodes exist around the primary tumor, it can be difficult to distinguish uptake in these nodes from the intense activity of the primary tumor. However, the advent of PET/CT imaging, enabling co-registration of both anatomical and functional information, has overcome this disadvantage and improved the localization of increased FDG uptake (Fig. 8) [29]. FDGPET/CT shows the extent of disease more accurately than other imaging methods, and this frequently leads to a radical change in patient management. Combined PET/CT imaging is therefore a valuable diagnostic tool for the primary diagnosis of GISTs or assessment of the response to therapy [30, 31]. PET/CT scans can also be used in staging patients with primary gastric lymphoma, as well as for monitoring these tumors after therapy [30]. However, the availability of FDG-PET, and FDG-PET/CT in particular, is still limited and their use expensive. Thus, at present, HMDCT plays a major role as a triage tool to aid in choosing the appropriate treatment

Marc S. Levine, Ahmed Ba-Ssalamah

a

b

Fig. 8 a, b. Hydro-MDCT PET scan of an early cervical esophageal carcinoma and enlarged right-sided lymph node. a On the axial HMDCT scan of the upper thorax, the small esophageal cancer is not clearly seen and the enlarged lymph appears suspicious. b Corresponding axial fused HMDCT-PET image shows increased activity in the region of the small primary tumor (arrow) and an enlarged metastatic lymph node (arrowhead)

for patients with esophageal and gastric tumors. HMDCT may help distinguish surgical candidates with limited disease from patients in need of pre-operative chemoradiation for down-sizing a tumor or from patients who need palliative therapy for advanced, unresectable tumor. When HMDCT shows definite advanced disease with extensive tumor spread, pre-surgical chemotherapy or radiochemotherapy is used to improve the prognosis. After completion of neo-adjuvant treatment, the tumor can then be restaged to determine whether surgical resection is indicated. Thus, pre-operative staging of esophageal and gastric tumors is by far the most important indication for HMDCT. This technique also plays an important role in the evaluation of postoperative complications such as fistulas when the findings on barium studies are equivocal.

References 1. Graziani L, Bearzi I, Romagnoli A et al (1985) Significance of diffuse granularity and nodularity of the esophageal mucosa at double-contrast radiography. Gastrointest Radiol 10:1-6 2. Gupta S, Levine MS, Rubesin SE et al (2003) Usefulness of barium studies for differentiating benign and malignant strictures of the esophagus. AJR 180:737-744 3. Levine MS, Kressel HY, Caroline DF et al (1983) Barrett esophagus: reticular pattern of the mucosa. Radiology 147:663-667

Diseases of the Esophagus and Stomach

4. Levine MS, Macones AJ, Laufer I (1985) Candida esophagitis: accuracy of radiographic diagnosis. Radiology 154:581-587 5. Levine MS, Woldenberg R, Herlinger H, Laufer I (1987) Opportunistic esophagitis in AIDS: radiographic diagnosis. Radiology 165:815-820 6. Levine MS, Loevner LA, Saul SH et al (1988) Herpes esophagitis: sensitivity of double-contrast esophagography. AJR 151:57-62 7. Sor S, Levine MS, Kowalski TE et al (1995) Giant ulcers of the esophagus in patients with human immunodeficiency virus: clinical, radiographic, and pathologic findings. Radiology 194:447-451 8. Bova JG, Dutton NE, Goldstein HM, Hoberman LJ (1987) Medication-induced esophagitis diagnosed by double-contrast esophagography. AJR 148:731-732 9. Zimmerman SL, Levine MS, Rubesin SE et al (2005) Idiopathic eosinophilic esophagitis in adults: the ringed esophagus. Radiology 236:159-165 10. Levine MS, Dillon EC, Saul SH, Laufer I (1986) Early esophageal cancer. AJR 146:507-512 11. Levine MS, Verstandig A, Laufer I (1986) Serpiginous gastric erosions caused by aspirin and other nonsteroidal antiinflammatory drugs. AJR 146:31-34 12. Sohn J, Levine MS, Furth EE et al (1995) Helicobacter pylori gastritis: radiographic findings. Radiology 195:763-767 13. Levine MS, Creteur V, Kressel HY et al (1987) Benign gastric ulcers: diagnosis and follow-up with double-contrast radiography. Radiology 164:9-13 14. Levine MS, Kelly MR, Laufer I et al (1993) Gastrocolic fistulas: the increasing role of aspirin. Radiology 187:359-361 15. Levine MS, Kong V, Rubesin SE et al (1990) Scirrhous carcinoma of the stomach: radiographic and endoscopic diagnosis. Radiology 175:151-154 16. White RM, Levine MS, Enterline HT, Laufer I (1985) Early gastric cancer: recent experience. Radiology 155:25-27 17. Yoo CC, Levine MS, Furth EE et al (1998) Gastric mucosa-associated lymphoid tissue lymphoma: radiographic findings in six patients. Radiology 208:239-243 18. Miettinen M, Sarlomo-Rikala M, Sobin LH, Lasota J (2000) Esophageal stromal tumors: a clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases and comparison with esophageal leiomyomas and leiomyosarcomas. Am J Surg Pathol 24:211-222

27

19. Galmiche JP, Clouse RE, Balint A et al (2006) Functional esophageal disorders. Gastroenterology 130:1459-1465 20. Jung M (2005) [Mucosectomy as sufficient therapy for early squamous cell]. Chirurg 76:1018-1024 21. Kutup A, Link BC, Schurr PG et al (2007) Quality control of endoscopic ultrasound in preoperative staging of esophageal cancer. Endoscopy 39:715-719 22. Prokop M. New challenges in MDCT (2005) Eur Radiol 15 Suppl 5:E35-45 23. Ba-Ssalamah A, Prokop M, Uffmann M et al (2003) Dedicated multidetector CT of the stomach: spectrum of diseases. Radiographics 23:625-644 24. Ba-Ssalamah A, Zacherl J, Noebauer-Huhmann IM et al (2009) Dedicated multi-detector CT of the esophagus: spectrum of diseases. Abdom Imaging 34:3-18 25. Westerterp M, van Westreenen HL, Reitsma JB et al (2005) Esophageal cancer: CT, endoscopic US and FDG PET for assessment of response to neoadjuvant therapy-systematic review. Radiology 236:841-851 26. Luketich JD, Schauer PR, Meltzer CC et al (1997) Role of positron emission tomography in staging esophageal cancer. Ann Thorac Surg 64:765-769 27. Cerfolio RJ, Bryant AS (2006) Maximum standardized uptake values on positron emission tomography of esophageal cancer predicts stage, tumor biology, and survival. Ann Thorac Surg 82:391-394; discussion 394-395 28. Chen J, Cheong JH, Yun MJ et al (2005) Improvement in preoperative staging of gastric adenocarcinoma with positron emission tomography. Cancer 103:2383-2390 29. Hsu WH, Hsu PK, Wang SJ et al (2009) Positron emission tomography-computed tomography in predicting locoregional invasion in esophageal squamous cell carcinoma. Ann Thorac Surg 87:1564-1568 30. Suga K, Yasuhiko K, Hiyama A et al (2009) F-18 FDG PET/CT findings in a patient with bilateral orbital and gastric mucosa-associated lymphoid tissue lymphomas. Clin Nucl Med 34:589-593 31. Antoch G, Kanja J, Bauer S et al (2004) Comparison of PET, CT, and dual-modality PET/CT imaging for monitoring of imatinib (STI571) therapy in patients with gastrointestinal stromal tumors. J Nucl Med 45:357-365

IDKD 2010-2013

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis Marc J. Gollub Memorial Sloan Kettering Cancer Center, New York, NY, USA

Computed Tomography Enterography Computed tomography enterography (CTEG) is a focused CT scan examination of the small intestine that combines the advantages of isotropic, thin-section multiplanar CT; the large volumes of neutral-density oral contrast; and rapid administration of intravenous contrast. Oral contrast agents such as 0.1% barium, polyethylene glycol (PEG) and methylcellulose contain additives that inhibit fluid reabsorption and allow maximal bowel distention. Intravenous contrast is injected at the “enteric phase” (about 45 s) to provide maximum wall enhancement against the neutral-density lumen (0-30 HU) [1]. Pharmacological manipulation to interrupt small-bowel spasm and encourage gastric emptying is commonly used, including glucagon and metoclopramide, respectively. The indications for CTEG include Crohn’s disease and other enteritides, obscure gastrointestinal bleeding (OGIB), detection of intestinal masses, and sprue. A prospective blinded comparison of CTEG with wireless capsule endoscopy (WCE) using clinical consensus as the gold-standard found similar sensitivities (82 vs. 83%) for active small-bowel Crohn’s disease, but CTEG was far more specific than WCE (89 vs. 53%). Although there is less experience in using CTEG for OGIB, one study found a bleeding source in 45% of 22 patients, in three of whom the source had been missed by initial WCE [2]. No comprehensive reports of mass detection have been published yet, but in our experience CTEG appears to represent a first-line test for suspected carcinoid. This chapter will discuss the pitfalls and limitations of CTEG and CT enteroclysis (CTEC).

Pitfalls of Computed Tomography Enterography General Even with the aid of neutral-density oral contrast to assist in mass conspicuity, mildly enhancing masses or mural inflammation may be subtle and overlooked without proper adjustment of the window and level. We recommend, in addition to standard abdominal settings, a

liver-type setting (W215, L135) and an “enterographictype” setting (W430, L155). Since CTEG does not create an enteral challenge, in contrast to CTEC, its use in the setting of low-grade small-bowel obstruction should be discouraged in favor of CTEC or magnetic resonance enteroclysis. However, even during a non-obstructive episode, CTEG may confer an advantage by delineating a subtle, underlying cause such as a mass or stricture. Fastidious technique and timing are basic requirements in CTEG. Inadequate oral intake, early or late scanning, poor intravenous enhancement (low rate, reduced dose), or underlying hypo- or hypermotility disorders can interfere with good distention and wall conspicuity. Some of these technical pitfalls can be overcome with low-kVp imaging, close monitoring of the drinking schedule, and pharmacological manipulation. Poor jejunal distention is a generally expected limitation compared with other examinations. Fortunately, many abnormalities are diagnosed at CTEG solely by virtue of mural hyperenhancement.

Crohn’s Disease Proof of disease, grading of severity, and assessment of penetrating disease are important tasks prior to surgical planning and even prior to medical treatment, since the therapeutic anti-tumor necrosis factor alpha (TNFα) antibodies (e.g., infliximab) are expensive and have side effects of infection and drug-induced immune disease [3]. The classic findings at CTEG in patients with active Crohn’s disease have been well described [2]. The American College of Radiology has deemed CTEG to be the most appropriate imaging test for Crohn’s disease [4]. Nonetheless, studies testing the accuracy of CTEG compared with other radiological and endoscopic tests have identified some clear limitations and pitfalls for CTEG in the setting of suspected Crohn’s disease, including: 1. non-specificity of certain findings, such as mural hyperenhancement without skip areas; 2. in known Crohn’s disease, signs of very early disease, such as aphthous ulcers, which will not be detected unless a mucosal study is done; 3. measurement of disease severity, including quantitation of ulcers or fistulas; and

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis

29

4. correlation with clinical findings. Wold et al. found CTEG to be superior to small-bowel follow-through for the detection of abscesses and fistulas, with a sensitivity and specificity of 78/83% vs. 62/90%, respectively [5]. Vogel et al. showed that, although CTEG is accurate for determining the presence or absence of strictures and fistulas (sensitivity 100, 92% respectively), it is less accurate in determining their number (sensitivity 67%, both). This may be clinically significant since uncorrected strictures or fistulas can result in postoperative symptoms [6]. Solem et al. found that the sensitivity of WCE and CTEG in the detection of active small-bowel Crohn’s disease was identical (both 83%), and there was no significant difference with ileocolonoscopy (74%) or small-bowel followthrough (65%) [7]. Hara et al., investigating disease activity, showed that findings at CTEG correlated with symptoms in 80% of patients, indicating CTEG’s excellence as a monitoring test [8]. In a study by Higgins et al., clinicians suspected only 84% of CTEG-identified strictures and CTEG excluded strictures in >50% of patients with a clinically suspected stricture [9]. Furthermore, a survey of clinicians showed that, following the use of CTEG, clinical management changes 50% of the time [10].

contrast, and 1800 mL of neutral-density contrast. With this approach, the majority of patients with a source of OGIB were detected compared with WCE, surgery, or follow up. Ten of 22 (43%) studies showed positive findings, including the detection of angioectasias [12]. Huprich et al. emphasized that, while three-phase CTEG can be done with the goal of detecting angioectasias and other arterialphase dominant lesions, the radiation incurred is substantial (effective dose 59 mSv per exam) such that the relative risks and benefits have to be weighed, including patient age, necessity of repeat examinations, and the known risk of radiation with multiple CT exams, especially in younger patients [12, 13]. A follow-up prospective study using three-reader comparisons, presented by Huprich at RSNA 2009, indicated that two out of three phases are often adequate, but conclusive results have not been published. In a similar study by Hara, using three-phase CTEG, 33% of lesions were detected (specificity 89%) and 52% were detectable in retrospect. Some of the missed lesions were in the stomach and colon, emphasizing the advantage of CTEG in depicting abnormalities throughout the GI tract. Currently, the conditions that appear to be undetectable by CTEG, as proven by other tests, include ulcers, vascular lesions, and non-bleeding lesions [14].

Obscure Gastrointestinal Bleeding

Neoplasms

In the search for sources of gastrointestinal (GI) bleeding, CTEG has found increasing use since it may be a more sensitive, non-invasive method – using intravenous contrast and CT angiography techniques – made possible by neutral-density oral contrast. However, its application is limited to being a diagnosis-only test, with no therapeutic capability. Abstracts from the 2009 Radiological Society of North America (RSNA) meeting in which the detection of bleeding rates using various techniques was compared suggested equal sensitivity with tagged RBC and catheter digital subtraction angiography, such that bleeds of as little as 0.3 mL/min were detected. The search for bleeding sources in the small bowel may be subsequent to an apparently negative upper- and lower-GI endoscopy. It should be kept in mind that missed gastric and colonic pathology may still be discovered at imaging if these areas are well-filled, since endoscopic examinations are performer-dependent, imperfect gold-standards. Conventional barium studies do not identify mucosal erosions or vascular ectasia and have yields of 6% for small-bowel follow-through and 10% for small-bowel enteroclysis [11]. Angioectasias are the most common cause of OGIB in patients over the age of 50. These lesions are typically small, may be multiple, and are sessile or slightly raised. They are typically only visible at endoscopy or potentially at arterial-phase catheter-based or CT angiography. A major limitation of CTEG is its relative insensitivity to these small, flat, vascular lesions. The most recent study used optimized thin-sections, three scanning phases (arterial, enteric, and delayed), rapid boluses of intravenous

The accuracy of CTEG in the detection of masses such as carcinoid, adenocarcinoma, lymphoma, gastrointestinal stromal tumor, and polyps is not known. Capsule endoscopy, an examination being used more frequently to examine the small bowel, has several documented limitations, such that CTEG plays an important ongoing role for small-bowel mass detection and is especially important for masses that may have a predominant extraluminal component [15]. A recent RSNA 2009 Abstract indicated that masses in the setting of OGIB were better detected at CTEG than at WCE. Hyperenhancing lesions (e.g., some melanoma metastases or carcinoid tumors) will be easily detected with careful observation and windowing; however, the detection of isoattenuated masses (some melanoma metastases, polyps, and even some primary adenocarcinomas) and smaller masses may require excellent luminal distention and perusal for secondary findings, such as increased wall thickening, increased luminal caliber, or complications (intussusception, obstruction). CTEG may underestimate the number of lesions due to their smaller size or the lower degree of vascularity of small tumors.

Computed Tomography Enteroclysis This fused test combines the advantages of enteral challenge from a catheter small-bowel examination (enteroclysis) with the isotropic, multiplanar, cross-sectional images obtained at helical CT. Indications for CTEC include small-bowel obstruction, small-bowel masses, OGIB, Crohn’s disease, and malabsorption.

30

Pitfalls of Computed Tomography Enteroclysis General Careful attention to technique, including the use of infusion rates that allow uniform bowel distention; catheter placement in the proper location, with correct balloon insufflation and tip placement; the appropriate administration of pharmacological agents; and proper use of multidetector CT with reformatted images, will obviate many sources of misinterpretation. Spasm and inadequate filling should be easily recognized and addressed by adjustment of the pump flow-rate and use of spasmolytics (glucagon or Buscopan). Variable window and level settings are advised to adjust for high-attenuation contrast or, in the interpretation of a neutral-density exam, to appreciate subtle differences in enhancement of the bowel wall.

Crohn’s Disease In the diagnosis of suspected or known Crohn’s disease, a comparison of WCE with neutral-density oral contrast CTEC (nCTEC) in 56 patients showed that 27 could not undergo WCE due to strictures (≤10 mm) for fear of capsule retention. In the other 41 patients, the limitations of nCTEC included its inability to detect very early, minimal inflammatory changes or small mucosal lesions such as villous denudation, aphthoid ulcerations, or erosions. These were far better detected by WCE than CTEC (p = 0.004) [16-18]. Since the detection of non-elevated or non-depressed lesions in Crohn’s disease and superficial erosions in NSAID enteropathy will be limited, these entities would probably be best investigated using others methods, such as push enteroscopy, single- or double-balloon endoscopy, WCE or “air” (C02) double-contrast fluoroscopic enteroclysis without CT [19, 20]. In addition, in late-stage Crohn’s disease, nCTEC may show fewer fistulae than positive oral contrast CTEC (pCTEC), as the higher-density contrast may fill the GI tract more conspicuously.

Masses In the workup of small-bowel masses or OGIB, several pitfalls may be encountered. False-positive masses may be seen in patients with Kerkring fold thickening or even transient intussusception [21]. Unfortunately, the findings may be so convincing as to necessitate surgery to prove the lack of a mass. In Pilleul’s study of 219 patients with possible small-bowel neoplasms, the overall accuracy was 84.7%. There were five false-positive masses (2.3%) ranging from 6 to 25 mm in size. In two of these, small-bowel fold thickening was found at surgery and in the others no mass could be detected [22]. CTEC may also fail to identify jejunal polyps <10 mm, angioectasias, and ectopic pancreas 3-5 mm in size. A false-negative rate of 4.1% was reported as well, as any lesion that is sessile,

Marc J. Gollub

minimally enhancing, or <5 mm in size may be difficult to detect. As such, the use of CTEC for patients with polyposis syndromes (familial adenomatous polyposis and Peutz-Jeghers) may be somewhat limited [22, 23]. Voderholzer et al. found that, compared with WCE, polyps, erosions, angioectasia, and lymphangiectasia were missed by CTEC but seen on WCE [17]. Careful perusal of thin-sections at CTEC might prevent overlooking of these lesions. Careful perusal of wall enhancement is necessary in patients with Crohn’s disease, to avoid missing more focal, concurrent masses (secondary malignancy), as these can be overlooked as well. In the diagnosis of OGIB, there are few reported series in which CTEC was used. In fact, in the workup of OGIB, Fillipone et al. reported that in patients over age 50 OGIB is more commonly caused by angioectasia than by small-bowel masses. Here too, WCE may be the more appropriate modality due to the nonelevated nature of these often subtle, small lesions [18].

Obstruction In the context of small-bowel-obstruction, Walsh et al. concluded that CTEC has greater sensitivity and specificity (89 and 100%, respectively) than conventional CT (50 and 94%, respectively) [23]. Pitfalls in technique specific to patients with obstruction include: failure to perform suctioning prior to examination of the patient, failure to place the balloon in the jejunum (as opposed to the duodenum, for better suctioning and the prevention of back-flow into the stomach), and failure to properly adjust the infusion rate so as to elicit the transition point in low-grade obstructions. Overzealous infusion can cause spasm, which may be misinterpreted as a site of stricture and obstruction. This can be avoided to some extent with the administration of glucagon or Buscopan. Problemsolving regarding possible strictures can be accomplished with repeat limited CT slices through the area of interest; however, this can be a costly approach given the radiation-burden. The use of multidetector CT with 40 or more channels can reduce radiation by 10-66% because of more efficient detector configurations, automatic exposure controls, improved filters, and the availability of image post-processing algorithms [24].

Summary The workup of small intestinal abnormalities is changing rapidly due to technological advances in cross-sectional imaging and endoscopic techniques. Although improvements are still being made, certain early conclusions regarding the detection of small-bowel pathologies can be offered: 1. CTEG, the most rapidly investigated new technique, appears to be comparable if not superior to most endoscopic methods in determining disease presence, severity, extent, complications, and activity status, as well as response to treatment; however, pitfalls exist, including

Small-Bowel Imaging: Pitfalls in Computed Tomography Enterography/Enteroclysis

31

non-specificity of early findings, the enumeration of fistulous tracts, and the detection of aphthous ulcers. 2. Patients with OGIB who are stable, i.e., not requiring surgery or catheter embolization immediately, if at all, should be offered CTEG using two or more phases to search for a source of blood – with the known pitfalls of the technique’s inability to detect shallow gastric or small-bowel ulcers, small angioectasias, or flat lesions. 3. In the search for small-bowel masses, with or without OGIB, CTEG appears to perform quite well, but if a mass is not hyperenhancing, perfect bowel distention may be required, an achievement not always possible, especially in the proximal jejunum. 4. CTEC, performed in limited centers in the USA and perhaps more widely in Europe, is subject to mostly technique-related pitfalls but has the advantage of an enteral challenge, which is not available with other non-catheter/pump methods and which may define its predominant indication as the best test for low-grade, intermittent small-bowel obstruction not detectable by other means.

9. Higgins PDR, Caoli E, Zimmerman M et al (2007) Computed tomographic enterography adds information to clinical management in small bowel Crohn’s disease. Inflamm Bowel Dis 13:262-268 10. Bruining DH, Siddicki H, Fletcher JG (2008) Clinical benefit of CT enterography in suspected or established Crohn’s disease: impact on patient management and physician level of confidence. Gastroenterology S1211 11. Singh V, Alexander JA (2008) The evaluation and management of obscure and occult gastrointestinal bleeding. Abdom Imaging 34:311-319 12. Huprich JE, Fletcher JG, Alexander JA (2008) Obscure gastrointestinal bleeding: Evaluation with 64-section multiphase CT enterography – initial experience. Radiology 246:562-571 13. Brenner DJ, Hall EJ (2007) Computed tomography – an increasing source of radiation exposure. N Engl J Med 357:2277-2284 14. Hara AK, Walker FB, Silva AC, Leighton JA (2009) Preliminary estimate of triphasic CT enterography performance in hemodynamically stable patients with suspected gastrointestinal bleeding. AJR 193:1252-1260 15. Postgate A, Despott E, Burling D et al (2008) Significant small-bowel lesions detected by alternative diagnostic modalities after negative capsule endoscopy. Gastrointest Endosc 68:1209-1214 16. Voderholzer WA, Beinhoelzl J, Rogalla P et al (2005) Small bowel involvement in Crohn’s disease: a prospective comparison of wireless capsule endoscopy and computed tomography enteroclysis. Gut 54:369-373 17. Voderholzer WA, Ortner M, Rogalla P et al (2003) Diagnostic yield of wireless capsule enteroscopy in comparison with computed tomography enteroclysis. Endoscopy 35:1009-1014 18. Fillipone A, Cianci R, Milano A et al (2008) Obscure gastrointestinal bleeding and small bowel pathology) comparison between wireless capsule endoscopy and multidetector-row CT enteroclysis. Abdom Imaging 33:398-406 19. Romano S, De Lutio E, Rollandi GA et al (2005) Multidetector computed tomography enteroclysis (MDCT-E) with neutral enteral and IV contrast enhancement in tumor detection. Eur Radiol 15:1178-1183 20. Maglinte DDT, Lappas JC, Heitcamp DE et al (2003) Technical refinements in enteroclysis. Radiol Clin North Am 41: 213-229 21. Boudiaf M, Jaff A, Soyer P et al (2004) Small-bowel diseases: Prospective evaluation of multi-detector row helical CT enteroclysis in 107 consecutive patients. Radiology 233:338-344 22. Pilleul F, Penigaud M, Milot L et al (2006) Possible smallbowel neoplasms: contrast-enhanced and water-enhanced multidetector CT enteroclysis. Radiology 241:796-801 23. Walsh D, Bender G, Timmons H (1998) Comparison of computed tomography enteroclysis and traditional computed tomography in the setting of suspected partial small bowel obstruction. Emerg Radiol 5:29-37 24. Mannudeep K, Rizzo SMR, Novelline RA (2005) Technologic innovations in computer tomography dose reduction: implications in emergency settings. Emergency Radiology 11:127-128

References 1. Colombel JF, Solem CA, Sandborn WJ et al (2006) Quantitative measurement and visual assessment of ileal Crohn’s disease activity by computed tomography enterography: correlation with endoscopic severity and C reactive protein. Gut 55:1561-1567 2. Paulsen SR, Huprich JE, Hara AK (2007) CT enterography: Noninvasive evaluation of Crohn’s Disease and obscure gastrointestinal bleed. Radiol Clin N Am 45:303-315 3. Huprich JE, Fletcher JG (2009) CT enterography: Principles, technique and utility in Crohn’s disease. Eur J Radiol 69:393-397 4. Dave-Verma H, Moore S, Singh A et al (2008) Computed tomographic enterography and enteroclysis: pearls and pitfalls. Curr Probl Diagn Radiol 37:279-287 5. Wold PB, Fletcher JG, Johnson CD et al (2003) Assessment of small bowel Crohn disease: Noninvasive peroral CT enterography compared with other imaging methods and endoscopyfeasibility study. Radiology 229:275-281 6. Vogel J, Moreira A, Baker M (2007) CT enterography for Crohn’s disease: Accurate preoperative diagnostic imaging. Dis Colon Rectum 50:1761-1769 7. Solem CA, Loftus Jr EV, Fletcher JG et al (2008) Small-bowel imaging in Crohn’s disease: a prospective, blinded, 4-way comparison trial. Gastrointest Endosc 68:255-266 8. Hara AK, Alam S, Heigh RI et al (2008) Using CT enterography to monitor Crohn’s disease activity: a preliminary study. AJR 190:1512-1516

IDKD 2010-2013

Diseases of the Small Bowel, Including the Duodenum – MRI Karin A. Herrmann Institute of Clinical Radiology, University Hospitals Munich, Munich, Germany

Introduction For decades, barium fluoroscopy studies have been the standard of reference to investigate small bowel diseases. Since the small bowel was not accessible to endoscopic techniques, these studies represented the only non-invasive diagnostic approach to the intestine. Both bowel followthrough and small bowel enteroclysis yielded fairly good results, with sensitivities and specificities of, respectively, 98.3% and 99.3% for Crohn’s disease (CD) [1] and 61-95% for neoplastic disease [2], notably in the assessment of the intestinal mucosa due to the high spatial resolution obtained with these techniques. However, their limitations are that they provide almost exclusively intraluminal information and are associated with considerably high radiation exposure, up to 10-18 mSv. The technical advances in cross-sectional imaging achieved with computed tomography (CT) and magnetic resonance imaging (MRI) over the past 10 years have tremendously improved image quality in the abdomen, thus encouraging small bowel imaging. Thin-section multi-detector row CT with 3D multiplanar reformations and accelerated image acquisition, in addition to breath-hold techniques in MRI, has fostered imaging of the bowel with no or few limitations. Consequently, CT and MRI are nowadays considered as state-of-the-art imaging modalities. Both not only provide insight into the bowel lumen but also depict mural and extramural pathology, which is essential for a comprehensive diagnostic assessment and the staging of small bowel diseases. Moreover, CT and MRI convey far more information than obtained with conventional barium techniques [3]. MRI, in addition, has the advantage of obviating radiation exposure, and is therefore preferable in children, young individuals, in pregnancy, and if multiple follow-up studies are required such as in patients with inflammatory bowel disease. In order to appropriately assess the intestine, bowel distention is essential. This can be achieved with contrast agents that typically have osmotic effects and enhance the contrast between the bowel lumen and the bowel wall. Depending on the mode of application, the corresponding examination is called MR- or CT-enterography if the contrast medium is administered orally prior to the examination, and MR- or CT-enteroclysis if the contrast agent is injected

using a nasojejunal tube, aided by an automatic infusion pump. In CT-enteroclysis, scanning is typically performed just after completion of the filling process; this results in a static, high-resolution data set that allows 3D post processing. MR-enteroclysis has the advantage that it provides not only static but also functional information, since repetitive scanning can be carried out during the filling process. The diagnosis of small bowel diseases with CT and MRI is fundamentally based on morphological criteria. Distinct imaging findings allow disease detection, diagnosis, and staging. This chapter is designed to familiarize the reader with the typical imaging features indicating inflammatory, neoplastic, and ischemic small bowel diseases, with special focus on their MRI appearances.

The Duodenum Since the upper gastrointestinal tract is readily accessible to endoscopic techniques, imaging techniques are less important to establish a diagnosis. Yet, cross sectional imaging is helpful to assess the intra- and extraluminal extent of disease. Since the duodenum is likely to be involved in entities affecting surrounding organs, such as the pancreas, bile duct, ampulla, and papilla, these should always be considered in the differential diagnosis of duodenal pathology. Embryological anomalies include ectopic and annular pancreas, webs, and choledochoceles. Annular pancreas occurs in 1:20,000 births and is typically diagnosed in the early days after birth due to intestinal obstruction. During pregnancy, it can be discovered during evaluation of the polyhydramnion. Less frequently, it is found in young adults or incidentally later in life; in both cases it then has to be distinguished from neoplasm. Other benign non-neoplastic conditions in the duodenum include ulcerative and inflammatory disease (i.e., CD) and pseudoinflammatory polyps. These pathologies are typically worked up endoscopically.

Neoplasms of the Duodenum Lipoma, leiomyoma, neurofibroma, adenoma, Brunner gland adenoma, and polyps are benign neoplasms occur-

33

Diseases of the Small Bowel, Including the Duodenum – MRI

ring in the duodenum. Patients are typically asymptomatic until a late stage of disease. While MRI is helpful in characterizing fat-containing lesions, it is less useful in the final histopathological diagnosis of most of these neoplasms. Malignant neoplasms of the duodenum include adenocarcinoma, lymphoma, neuroendocrine tumors, gastrointestinal stromal tumors (GIST), and metastases. Duodenal adenocarcinoma accounts for only 0.4% of gastrointestinal tumors. By the time it is clinically apparent, there is often advanced disease. Clinical symptoms are indicative of high bowel obstruction, vomiting, chronic bleeding and, if close to the ampulla, obstructive jaundice. The duodenum is the most common site for adenocarcinoma, which may appear as a defined nodular mural mass or a diffusely infiltrating mass with spiculated borders. Another pattern is that of a diffuse or annular wall thickening causing constrictive narrowing of the lumen. On T1weighted imaging, the tumor may be homogeneously isointense or hypointense compared to the wall and slightly hyperintense or isointense in T2-weighted imaging. Contrast enhancement is moderate, mostly homogeneous. Lymphoma: in the duodenum, as in the small and large bowel, the most frequent form of lymphoma is nonHodgkin’s lymphoma (NHL). In 50% of nodal NHLs, there are concomitant intestinal manifestations. Typical imaging features of duodenal and intestinal lymphoma are marked asymmetrical circumferential wall thickening with mild pre-lesional luminal dilatation. Luminal stenosis is not always predominant. Lymphoma may also appear as multiple polypoid intraluminal protrusions or an exophytic extraluminal mass. Ulceration and fistula formation are not uncommon. An extraluminal desmoplastic reaction is unlikely. Associated lymphadenopathy is a helpful diagnostic hint but is not always present. Compared to normal bowel wall, intestinal lymphoma shows moderately increased and homogeneous signal intensity on T2-weighted images. On T1-weighted imaging, it is isointense or hypointense and exhibits mild and slightly inhomogeneous contrast enhancement after gadolinium administration [4]. Unlike neuroendocrine tumors (NET) of the small bowel, the role of CT and MRI in detecting duodenal NET is minor compared to endoscopic techniques, including endoscopic ultrasound. Along with the stomach, duodenum is a common location for NET. Typically, the lesions are small; they occasionally can be detected on CT and MRI as single or multiple hypervascular nodules in arterial-phase imaging. Luminal distention, e.g., with hydro-CT of the stomach, is recommended and may help to achieve a detection rate of up to 89%. GISTs are most commonly located in the stomach and small bowel and are described in the following section.

The Small Intestine Inflammatory Bowel Disease Currently, the most frequent clinical application of small bowel MRI is in patients with inflammatory bowel

disease, especially CD. CD is a chronic inflammatory autoimmune disorder that frequently involves the small bowel but may affect the entire gastrointestinal tract. It is most commonly located in the terminal ileum (40-80%) and the colon. Involvement of the proximal ileum and the jejunum is less frequent (22-40%). In imaging, CD manifestations have been classified by Maglinte et al. in three categories: acute inflammatory disease, fibrostenotic disease, and fistulizing disease [5]. Due to its chronically recurrent character, different stages of CD may coexist. Discontinuity of multiple disease manifestations (skip lesions) is a characteristic feature. The morphological spectrum ranges from early superficial mucosal disease with disruption, flattening, thickening, and distortion of the fold pattern, to mild or pronounced longitudinal or transverse fissures and ulcerations resulting in a cobblestone appearance, to transmural disease characterized by wall thickening, stenosis, and mesenteric hypervascularity. Eventually, extramural extension of the inflammation into the mesentery may occur, accompanied by the development of blind sinus tracts, fistulas, and micro- or macro abscesses. Crohn’s Disease: Acute Inflammatory Type A number of imaging findings in MRI have been reported to determine inflammatory activity in CD: increased small bowel wall thickening (typically >4 mm), increased mural contrast enhancement [6], submucosal edema with high signal intensity on T2-weighted images [7] (Fig. 1), the presence of deep mucosal ulcers and fissures [8], mesenteric hypervascularity (comb sign) [9], and contrastenhancing enlarged lymphadenopathy [8]. The most recent literature describes increased mural thickening, increased wall signal intensity on T2-weighted fat-saturated images, and layered mural enhancement, but not mural enhancement alone, as the strongest indicators for active disease. Layered mural enhancement, however, is also commonly associated with coexisting fibrostenosis and scar formation [10]. Early and mild stages of CD may present at MRI as a focal or regional disruption of the fold pattern, showing only subtle or no increased enhancement. Superficial erosions may not be detected at all. In contrast to overt wall thickening and stenosis, these lesions can easily be missed at MRI especially when distention is suboptimal. This is why MR-enterography or MR-enteroclysis (MRE), compared to capsule endoscopy or invasive double balloon endoscopy, reaches an overall sensitivity of only 75-80% [11, 12]. In subacute CD and under effective treatment, wall thickening may persist initially but decreases over time. The hyperintense signal intensity from submucosal edema disappears and becomes intermediate signal intensity. Likewise, the formerly increased contrast enhancement decreases. Crohn’s Disease: Fibrostenosing Type Fibrostenotic lesions are less conspicuous and more difficult to identify on MRI since they typically show no wall thickening or increased enhancement. The seemingly

34

a

Karin A. Herrmann

b

Fig. 1 a, b. MR-enteroclysis study with biphasic intraluminal contrast in a 32 year-old patient with Crohn’s disease showing typical signs of active inflammatory disease. a Wall thickening, luminal narrowing and submucosal edema (arrow) is seen on T2weighted single-shot fast spin-echo (SSFSE) (HASTE) imaging in a loop with prominent active inflammation. b Slightly further up, the loop shows increased transmural enhancement on contrast-enhanced T1-weighted imaging with fat suppression, mesenteric hypervascularity (arrow), and inflammatory extramural stranding of the mesentery

normal wall thickness and contrast enhancement in fibrostenosis mimics the normal bowel wall. Low signal intensity on T2-weighted images with and without fat suppression has been described [9]. Persistent focal and fixed circumscript intestinal stenosis and marked prestenotic dilatation are the most conspicuous and reliable indicators and indirect signs of fibrous strictures. Crohn’s Disease: Fistulizing Type Fistulas and peri- extramural abscesses represent the most severe complications and can be diagnosed accurately with MRI (accuracy 92%). Fistulas are tubular inflammatory pathways between bowel loops, bowel, and other organs or bowel and the abdominal wall. They are defined as internal fistulas (IF) when located between bowel loops (entero-enteric) or internal organs (e.g., entero-vesical). External fistulas (EF) are those that abut the skin (entero-cutaneous). A vast majority of IF remain asymptomatic or cause non-specific symptoms, especially if located in distal bowel segments. Complex IF typically involve multiple bowel loops [13] and bowel obstruction due to concomitant stenosis. Abnormal tubular tracts within the mesentery containing air or fluid are highly indicative of IF and are well seen on SSFP (steady-state free precession; e.g., TrueFISP) imaging. The outlines of IF are of intermediate signal intensity on T2-weighted images, similar to the intestinal wall, and may show moderate enhancement after intravenous gadolinium. A characteristic, somewhat stellate arrangement of the bowel loops is referred to as the “star-sign” and is reportedly a strong indicator for the presence of complex IF (Fig. 2) [14]. The inflamed bowel loops converge to a common center, which often is the origin of abscess formation. Urinary bladder involvement is likely when there is marked thickening of the bladder wall in the vicinity of an inflamed bowel segment [13]. A contiguous inflammatory tract between the inflamed bowel and the bladder wall may or may not be visualized.

Fig. 2. MR-enteroclysis study of a 45-year-old patient with negative intraluminal contrast (Lumirem®, Guerbet, France) shows an example of a typical stellate appearance of converging bowel loops in an entero-enteric fistula involving the cecum, terminal ileum, sigmoid, and the middle ileum. Note the pre-stenotic dilatation of the bowel loops proximal to it, indicating a mechanical stenosing effect of the fistula complex (T2-weighted SSFSE)

Differential Diagnosis of Crohn’s Disease Although less frequently a reason for imaging than CD, infectious enteritis is an important differential diagnosis to consider. Infectious enteritis is typically of bacterial origin, such as Enterococcus, Salmonella, Yersinia, and tuberculous bacillae. In immunocompromised patients, viral and fungal infections such as Cytomegalovirus (CMV) have to be considered. In patients who have undergone bone marrow transplant, CMV and graft-versus-host reactions are

35

Diseases of the Small Bowel, Including the Duodenum – MRI

likely. The appearance of infectious enteritis on CT or MRI is non-specific, the main feature being submucosal edema. For tuberculosis, mesenteric lymphadenopathy is a major finding and a helpful clue to the diagnosis. Ischemic bowel disease is to be considered if arteriosclerosis and vascular obstruction are observed additionally to altered bowel wall with submucosal edema.

Small Bowel Neoplasms Small bowel neoplasms are rare and account for less than 5% of all gastrointestinal tumors. Adenocarcinoma, carcinoid, lymphoma, GIST, and metastases constitute the malignant component. MR-enteroclysis has recently been reported to be an effective diagnostic tool for the detection of small bowel tumors, with a sensitivity, specificity, and accuracy of 86, 98, and 97%, respectively [15]. The incidence of small bowel cancer has risen in the past 30 years, with the greatest increase for carcinoid tumors. Adenocarcinoma is the most common among the intestinal malignant neoplasms, followed by carcinoid tumors. After the duodenum, the jejunum is the second most frequent location for adenocarcinoma typically involving the ileum. The imaging morphology for both adenocarcinoma and lymphoma (Fig. 3) is as previously described (see duodenal neoplasms). Both may occur as complications of CD. Neuroendocrine Tumors: Carcinoid Carcinoid tumors are neuroendocrine neoplasms and account for approximately 2% of all gastrointestinal tumors. They may be found along the entire gastrointestinal tract

(85%) as well as in the pancreas and lung (10%). The appendix (50%) and the ileum (~30%) are the most common primary locations. Up to 30% of intestinal carcinoids are multifocal. Carcinoids have a tendency to metastasize early to lymph nodes and the liver, even when they only are 1-2 cm in size [16]. Therefore, distention of the bowel lumen in MRE is required for better detection. Carcinoids are typically hypervascularized and therefore are best identified on contrast-enhanced T1-weighted fat saturated GRE (gradient-recalled echo) sequences (Fig. 4). SSFSE and SSFP sequence types depict these tumors less well as slightly hyperintense or isointense to muscle and bowel wall. Approximately half the tumors appear as a nodular intraluminal mass, one third as focal circumferential wall thickening, and 20% with both characteristics [17]. If only wall thickening is present, carcinoids may easily be confounded with inflammatory disease. A desmoplastic reaction in the adjacent mesentery occurs in up to 73% of the cases of small tumors and may cause vascular engorgement. Gastrointestinal Stromal Tumors Gastrointestinal stromal tumors represent 0.3-3% of all gastrointestinal tumors. They derive from the intestinal cells of Cajal and originate most often from the stomach (~70%) or the small bowel wall (20-30%). After appropriate immunohistochemical preparation, they can be shown to strongly express the KIT protein (CD 117), which is a characteristic feature of GISTs. These tumors are well delineated, non-infiltrative masses located in or arising from the intestinal wall, typically extraluminally at the serosal side [18]. When <5 cm in diameter, they are slightly heterogeneous and mildly hyperintense on T2. Larger masses are heterogeneous,with a soft-tissue rim encompassing a necrotic center. This rim is isointense or slightly hyperintense to muscle on T2-weighted images and hypointense on T1-weighted images, showing heterogeneous enhancement after contrast administration. In the majority of cases, GISTs do not show signs of infiltration, vessel encasement, or lymphatic spread. The primary sites of metastases from malignant GISTs are the liver and peritoneum. Treatment with tyrosine kinase inhibitors induces a loss of vascularization, and, consequently, reduced contrast enhancement of the tumor and its metastases [19], but no necessarily change in size.

Other Small Bowel Pathologies

Fig. 3. MR-enteroclysis study of a 63-year-old male with occult bleeding at fecal occult blood test. T2-weighted SSFSE image after biphasic intraluminal contrast administration shows an ileal segment of marked circumferential wall thickening with mass effect preserving the lumen (arrow). The intraluminal surface is irregular; contrast enhancement (not shown) was moderate. This example shows the typical features of intestinal lymphoma

For some small bowel pathologies, MRE has so far not proven to be useful. For superficial mucosal pathologies and angiodysplasia, MRI of the small bowel cannot be recommended. The detection of acute ischemia and occult gastrointestinal bleeding is the domain of CT, mainly because of its availability in an emergency setting, short examination time, and spatial resolution. Post-operative or post-inflammatory small bowel adhesions represent another important condition, formerly investigated with conventional enteroclysis using manual

36

a

Karin A. Herrmann

b

Fig. 4 a, b. T1-weighted contrast-enhanced coronal GRE image after MRE with biphasic intraluminal contrast depicts a 1.5-cm Ushaped intraluminal mass (arrow) in the ileum with increased enhancement (a; zoomed in b) which after surgical removal proved to be carcinoid. No mesenteric desmoplastic reaction is seen

compression and mobilization of the bowel loops. This manipulation is not practicable in MRI due to the limited access to the patient inside the magnet. To date, no literature is available to support the usefulness of MRE for small bowel adhesions.

References 1. Maglinte DD, Chernish SM, Kelvin FM et al (1992) Crohn disease of the small intestine: accuracy and relevance of enteroclysis. Radiology 184:541-545 2. Bessette JR, Maglinte DD, Kelvin FM, Chernish SM (1989) Primary malignant tumors in the small bowel: a comparison of the small-bowel enema and conventional follow-through examination. AJR Am J Roentgenol 153:741-744 3. Lee SS, Kim AY, Yang SK et al (2009) Crohn disease of the small bowel: comparison of CT enterography, MR enterography, and small-bowel follow-through as diagnostic techniques. Radiology 251:751-761 4. Kim KW, Ha HK (2004) MRI for small bowel diseases. Magn Reson Imaging Clin N Am 12:637-650 5. Maglinte DDT, Gourtsoyiannis N, Rex D et al (2003) Classification of small bowel Crohn’s subtypes based on multimodality imaging. Radiol Clin North Am 41:285-303 6. Sempere GAJ, Sanjuan VM, Chulia EM et al (2005) MRI evaluation of inflammatory activity in Crohn’s disease. AJR Am J Roentgenol 184:1829-1835 7. Maccioni F, Bruni A, Viscido A et al (2006) MR imaging in patients with Crohn disease: value of T2- versus T1-weighted gadolinium-enhanced MR sequences with use of an oral superparamagnetic contrast agent. Radiology 238:517-530 8. Gourtsoyiannis N, Papanikolaou N, Grammatikakis J et al (2004) Assessment of Crohn’s disease activity in the small bowel with MR and conventional enteroclysis: preliminary results. Eur Radiol 14:1017-1024

9. Prassopoulos P, Papanikolaou N, Grammatikakis J et al (2003) MR enteroclysis imaging of Crohn disease. Radiographics 21 Spec No:S161-172 10. Punwani S, Rodriguez-Justo M, Bainbridge A et al (2009) Mural inflammation in Crohn disease: location-matched histologic validation of MR imaging features. Radiology 252:712-720 11. Tillack C, Seiderer J, Brand S et al (2008) Correlation of magnetic resonance enteroclysis (MRE) and wireless capsule endoscopy (CE) in the diagnosis of small bowel lesions in Crohn’s disease. Inflamm Bowel Dis 14:1219-1228 12. Seiderer J, Herrmann K, Diepolder H et al (2007) Double-balloon enteroscopy versus magnetic resonance enteroclysis in diagnosing suspected small-bowel Crohn’s disease: results of a pilot study. Scand J Gastroenterol 42:1376-1385 13. Herrmann KA, Michaely HJ, Zech CJ et al (2006) Internal fistulas in Crohn disease: magnetic resonance enteroclysis. Abdom Imaging 31:675-687 14. Herrmann KA, Michaely HJ, Seiderer J et al (2006) The “starsign” in magnetic resonance enteroclysis: a characteristic finding of internal fistulae in Crohn’s disease. Scand J Gastroenterol 41:239-241 15. Masselli G, Polettini E, Casciani E et al (2009) Small-bowel neoplasms: prospective evaluation of MR enteroclysis. Radiology 251:743-50 16. Horton KM, Kamel I, Hofmann L, Fishman EK (2004) Carcinoid tumors of the small bowel: a multi-technique imaging approach. AJR Am J Roentgenol 182:559-567 17. Schmid-Tannwald C, Zech CJ, Panteleon A et al (2009). Characteristic imaging features of carcinoid tumors of the small bowel in MR enteroclysis. Radiologe 49:242-245 18. Burkill GJ, Badran M, Al-Muderis O et al (2003) Malignant Gastrointestinal Stromal Tumor: Distribution imaging features and pattern of metastatic spread. Radiology 226:527-532 19. Schlemmer M, Sourbron SP, Schinwald N et al (2009) Perfusion patterns of metastatic gastrointestinal stromal tumor lesions under specific molecular therapy. Eur J Radiol 27: 278-284

IDKD 2010-2013

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases Jaap Stoker1, Richard M. Gore2 1 Department 2 Department

of Radiology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands of Radiology, North Shore University Health System, Evanston Hospital, Evanston, IL, USA

Introduction Infectious and inflammatory diseases of the colon and rectum are common disorders that are becoming increasingly prevalent. Diverticulitis, appendicitis, and the inflammatory and infectious colitides are among the commonest causes of acute abdominal pain. These patients frequently present with non-specific symptoms. Since in most cases of acute abdominal pain clinical and laboratory assessments cannot confidently identify a specific etiology, cross-sectional imaging plays an important part in the work-up and management of these patients. Ultrasound (US) and computed tomography (CT) are the primary means of evaluating acute abdominal pain, as both modalities provide insight into the pathological changes in the colon wall, serosa, surrounding mesentery, and peritoneum. US is a readily available, real-time technique with reasonable accuracy that uses no ionizing radiation. CT provides rapid assessment with high accuracy, optimal field of view, and good reproducibility. Magnetic resonance imaging (MRI) in patients with acute abdominal pain offers great clinical potential by virtue of its global imaging perspective, exquisite soft-tissue resolution, and high accuracy without the need for ionizing radiation. In patients with perirectal and perianal inflammatory conditions, MRI is the preferred imaging technique, as the combination of high intrinsic contrast resolution and large field of view provides detailed information on the presence and extent of disease. This chapter describes the imaging features of these infectious and inflammatory disorders, provides differential diagnostic guidelines, and presents the advantages and disadvantages of the various imaging modalities in the context of optimizing patient management.

of ulcerative colitis are beneath the spatial resolution of CT. With progressive disease, submucosal edema producing a “target sign” may be seen. Severe mucosal ulceration can denude certain portions of the colonic wall, leading to inflammatory pseudopolyps (Fig. 1), which, when sufficiently large, can be visualized on CT. Mural thickening and luminal narrowing are the CT hallmarks of the subacute and chronic ulcerative colitides. Mural thinning, unsuspected perforations, and pneumatosis can be detected on CT in patients with toxic megacolon. In this regard, CT can be quite helpful in determining the urgency of surgery in patients with stable abdominal films yet a deteriorating clinical course [1-3]. In chronic ulcerative colitis, the muscularis mucosa becomes markedly hypertrophied, often by a factor of 40. Forceful contraction of this hypertrophied longitudinal muscle may pull the mucosa away from the submucosa,

Ulcerative Colitis Ulcerative colitis is characterized pathologically by extensive confluent and circumferential ulceration and by diffuse inflammation of the mucosa. The disease characteristically begins in the rectum and extends proximally in a contiguous fashion to involve part or all of the colon. The pathological changes found in the very early stages

Fig. 1. Acute ulcerative colitis. CT demonstrates deep ulcerations (arrows) of the fluid-filled rectosigmoid. Inflammatory pseudopolyps appear as residual islands of inflamed mucosa that protrude above the denuded colonic surface

38

producing diffuse or segmental narrowing of the lumen. The contraction also causes shortening of the colon. The submucosa becomes thickened due to the deposition of fat or, in acute and subacute cases, edema. Submucosal thickening further contributes to narrowing of the lumen. Additionally, in acute and in chronic ulcerative colitis, the lamina propria is thickened due to round-cell infiltration. On CT, these mural changes produce a “target” or “double-halo” appearance when axially imaged. The lumen is surrounded by a ring of soft-tissue density (mucosa, lamina propria, hypertrophied muscularis mucosae), then by a low-density ring (fatty infiltration of the submucosa), and in turn by a ring of soft-tissue density (muscularis propria). This mural stratification is not specific and can also be seen in Crohn’s disease, infectious enterocolitis, pseudomembranous colitis, ischemic and radiation enterocolitides, mesenteric venous thrombosis, bowel edema, and graft-versus-host disease [1-3]. There are certain CT findings that can help differentiate granulomatous from ulcerative colitis. Mural stratification, i.e., the ability to visualize individual layers of bowel wall, is seen in 61% of patients with chronic ulcerative colitis but only in 8% of patients with chronic granulomatous colitis. Also, mean colon wall thickness in chronic ulcerative colitis is 7.8 mm, significantly less than that observed in Crohn’s colitis (11 mm). Finally, the outer contour of the thickened colonic wall is smooth and regular in 95% of ulcerative colitis patients while serosal and outer mural irregularities are present in 80% of patients with granulomatous colitis [1-3]. Rectal narrowing and widening of the presacral space are hallmarks of chronic ulcerative colitis. CT depicts the anatomical alterations that underlie these rather dramatic morphological changes. The rectal lumen is narrowed due to the previously described mural thickening that attends chronic ulcerative colitis. As a result, the rectum has a target appearance on axial scans, which should not be mistaken for the external anal sphincter, mucosal prolapse, or the levator ani muscles. The increase in the presacral space is caused by proliferation of the perirectal fat, which on CT is characterized by an increased number of nodular and streaky soft-tissue densities and an abnormal attenuation value, 10-20 HU higher than the normal extraperitoneal or mesenteric fat. These fatty changes relate to a number of factors, including ex vacuo replacement by fat of the void produced by rectal lumen narrowing and lipodystrophy resulting from an influx of inflammatory cells and edema. Edematous adipose tissue and enlarged lymph nodes are often observed in the perirectal region at the time of abdominoperineal resections in patients with chronic ulcerative colitis [1-3].

Crohn’s Disease Crohn’s disease most commonly affects the terminal ileum and proximal colon. The acute, active phase of Crohn’s disease is characterized by focal inflammation,

Jaap Stoker, Richard M. Gore

aphthoid ulceration with adjacent cobblestoning, an often transmural inflammatory reaction with lymphoid aggregates and granuloma formation, fissures, fistulas, and sinus tracts. The chronic and resolving phase of this disorder is associated with fibrosis and stricture formation. The presence and extent of Crohn’s disease can be determined by US, CT, or MRI. The accuracy of each of these examinations is comparable but each technique has its strength and limitations [4]. When Crohn’s disease is limited to the mucosa, the CT scan is often normal. Although inflammatory and post-inflammatory pseudopolyps may be identified on CT, the assessment of the mucosa is best reserved for barium studies and colonoscopy, which are more direct and sensitive. Crohn’s disease is manifested on CT by bowel wall thickening of 12 cm. This thickening, which occurs in up to 83% of patients, is most frequently observed in the terminal ileum, but other portions of the small bowel, colon, duodenum, stomach, and esophagus may be similarly affected [1-3]. During the acute, non-cicatrizing phase of Crohn’s disease, the small bowel and colon maintain mural stratification and often have a target or double-halo appearance. As in ulcerative colitis, there is a soft-tissue density ring (corresponding to mucosa), which is surrounded by a low-density ring with an attenuation near that of water or fat (corresponding to submucosal edema or fat infiltration, respectively), which in turn is surrounded by a higher density ring (muscularis propria). Inflamed mucosa and serosa may show significant contrast enhancement following bolus intravenous contrast administration, and the intensity of enhancement correlates with the clinical activity of the disease [1-3]. CT demonstration of mural stratification, i.e., the ability to visualize distinct mucosal, submucosal, and muscularis propria layers, indicates that transmural fibrosis has not occurred and that medical therapy may be successful in ameliorating lumen compromise. Additionally, prior to the onset of fibrosis, the edema and inflammation of the bowel wall responsible for mural thickening and lumen obstruction are reversible to some extent. A modest decrease in wall thickness often produces a dramatic increase in lumen cross-sectional area as well as resolution of the patient’s obstructive symptoms. Loss of mural stratification is indicative of transmural fibrosis [1-3]. In the background of long-standing Crohn’s disease, this is typically visualized on CT as homogeneous attenuation in the affected bowel wall. If this finding occurs against a background of good levels of intravascular contrast medium and thin-section reconstructions, then the fibrosis is most likely irreversible. In these patients, anti-inflammatory agents may not provide a significant reduction in bowel wall thickness. If these segments become sufficiently narrow, surgery or stricturoplasty will be necessary to relieve the obstruction. In a patient with Crohn’s disease, the palpation of an abdominal mass or the separation of bowel loops as seen on a barium study evokes a large differential diagnosis: abscess, phlegmon, “creeping fat” or fibrofatty proliferation

39

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

of the mesentery, bowel wall thickening, and enlarged mesenteric lymph nodes. Each of these disorders has significantly different prognostic and therapeutic implications. This diagnostic dilemma is further complicated by the fact that many of these patients are receiving immunosuppressive therapy, which can mask other signs and symptoms. CT can readily differentiate the extraluminal manifestations of Crohn’s disease. Fibrofatty proliferation, also known as “creeping fat” of the mesentery, is the most common cause of bowel-loop separation seen on barium studies in patients with Crohn’s disease. On CT, the sharp interface between bowel and mesentery is lost and the attenuation value of the fat is elevated by 20-60 HU due to the influx of inflammatory cells and fluid. Mesenteric adenopathy, with lymph nodes ranging in size between 3 and 8 mm, may also be present. If these lymph nodes are larger than 1 cm, the presence of lymphoma or carcinoma, both of which occur with greater frequency in Crohn’s disease, must be excluded [1-3]. Contrast-enhanced CT scans often show hypervascularity of the involved mesentery, manifesting as vascular dilatation, tortuosity, prominence, and wide spacing of the vasa recta. These distinctive vascular changes have been called the “comb sign” and its identification should suggest active disease. In addition, it may be useful in differentiating Crohn’s disease from lymphoma or metastases, which tend to be hypovascular lesions. With the advent of novel biological therapeutic agents such as infliximab, CT enterography has become an important technique for evaluating small bowel disease activity in patients with Crohn’s disease, because of its accuracy and non-invasive nature. Since colonic involvement is common in patients with inflammatory bowel disease (IBD), the capability of CT enterography in evaluating colorectal involvement (Fig. 2) is being intensively studied, and the preliminary results are encouraging [5]. Since patients with Crohn’s disease are often young and typically will require multiple examinations, the cu-

a

Fig. 2. Acute Crohn’s colitis. Sagittal reformatted image from a CT enterography study reveals marked mural thickening of the descending colon with markedly prominent vasa rectae

mulative radiation dose of multiple CT examinations is substantial [6]. Accordingly, accurate technique without ionizing radiation is preferable. For small bowel imaging in IBD, there is a considerable body of evidence showing the comparable accuracy of US, CT, scintigraphy, and MRI [4]. The lack of ionizing radiation exposure thus favors the use of either US or MRI (Fig. 3). US gives realtime information on peristalsis in addition to the morphological features. While contrast-enhanced US correlates well with colonoscopic disease activity [7, 8], this technique is operator-dependent. The major advantages

b

Fig. 3 a, b. Acute Crohn’s colitis of the descending colon shows a thickened descending colon at (a) ultrasound and (b) MR enterography (coronal T2-weighted turbo spin-echo). The arrow indicates the thickened descending colon

40

of MRI over US are its unlimited field of view and good reproducibility; however, its lack of availability and high cost may be prohibitive. Optical colonoscopy can directly visualize the colon, but the impetus for cross-sectional imaging studies of the colon in patients with IBD has lagged behind imaging evaluation of the small bowel. Indeed, there is a large body of data concerning MR enterography and MR enteroclysis, but very few studies have evaluated colonic disease [9, 10]. In MR evaluations of the colon, lumen distention is mandatory [11, 12] to achieve optimal imaging results. However, the experience with dedicated examinations of the colon in IBD is limited and conflicting [13]. While use of the bright lumen technique is preferred, in a series of 15 healthy volunteers and 23 patients with known IBD dark lumen MR colonography resulted in high sensitivity (87%) for identifying segmental IBD changes [14]. However, no biopsies were obtained of endoscopically normal mucosa, which might have influenced the results. Several studies have evaluated the efficacy of MRI in determining small bowel disease activity in Crohn’s disease, with bowel wall thickening and bowel wall enhancement employed as markers of disease activity. Although severe disease activity is well depicted, limited disease activity and non-active disease are less reliably demonstrated [15]. Data regarding the assessment of colonic disease activity are sparse and partly conflicting. In one study, increased contrast enhancement of the bowel wall was significantly related to colonoscopically active disease in patients with IBD [16]. In another series, combining increased contrast enhancement of the bowel wall with other MRI features (bowel wall thickening, presence of mesenteric lymph nodes, loss of the normal haustral pattern), the sensitivity was poor (32%) but the specificity was good (88%) [17]. The combination of MR enterography with a water-based enema seems to be a valuable approach to assess disease activity in patients with established Crohn’s disease [18]. Data comparing MRI with other techniques are sparse, nor is it possible to recommend the routine use of water-based enema or other colon distention techniques for evaluating the colon in patients with IBD.

Jaap Stoker, Richard M. Gore

Fig. 4. Pseudomembranous colitis. Coronal reformatted CT image shows marked mural thickening and submucosal edema of the distal colon. Note the ascites

CT shows a pancolitis with mural thickening that may be irregular or polypoid and characterized by a shaggy endoluminal contour. The thickened wall (Fig. 4), which is usually 1.6-1.8 cm, is a result of submucosal edema. Mucosal and serosal enhancement is seen following intravenous contrast administration. The haustra are also thickened and edematous, producing the “accordion pattern” highly suggestive of pseudomembranous colitis. This pattern, which is the result of contrast trapped between thickened haustral folds aligned in a parallel fashion, can sometimes simulate deep ulcerations or fissures. Pericolic stranding, ascites, pleural effusions, and subcutaneous edema are other ancillary CT findings. Complications of untreated pseudomembranous colitis include toxic megacolon and intestinal perforation with subsequent peritonitis. CT is also useful in monitoring the response to medical therapy with oral vancomycin and metronidazole [20, 21].

AIDS-Related Colitis Pseudomembranous Colitis Pseudomembranous colitis is being encountered with increasing frequency as a nosocomial infection complicating antibiotic therapy [19]. This potentially life-threatening disorder is caused by overgrowth of Clostridium difficile. The bacteria release a cytotoxic enterotoxin that causes ulceration of the colonic mucosa and the formation of 2-3 mm pseudomembranes consisting of fibrin, mucus, sloughed epithelial cells, and leukocytes. Mild cases may demonstrate only mucosal irregularity and nodularity, with the formation of small plaques that cannot be detected radiologically whereas in advanced cases there is thickening of the haustral folds, a shaggy wall contour, and mucosal plaques.

Cytomegalovirus and cryptosporidiosis are common pathogens seen in patients with acquired immunodeficiency syndrome (AIDS). Patients with CD4 lymphocyte counts of ≤200 mm–3 are at greatest risk for these infections. Typically, the cecum and proximal ascending colon are involved by these organisms; however, a pancolitis with continuous lesions may occur as well. CT shows mural thickening of the involved segments of colon, with low attenuation in the region of the submucosa due to edema as well as pericolonic fluid and stranding of the adjacent fat. Pneumatosis and ascites have also been described [21]. AIDS-related colitis is being encountered with decreasing incidence in industrialized countries due to the efficacy of highly active anti-retroviral therapy (HAART).

41

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

Typhlitis Typhlitis (neutropenic colitis) is a potentially fatal infection of the cecum and ascending colon caused by enteric pathogens in patients with severe immunosuppression. It is most frequently seen in patients with acute leukemia receiving chemotherapy but also occurs in the setting of AIDS, aplastic anemia, multiple myeloma, and bone marrow transplantation. Bacteria, viruses, and fungi penetrate the damaged cecal mucosa and are able to proliferate due to the profound neutropenia. Edema and inflammation involve the cecum, ascending colon, and, occasionally, the ileum. Fever, abdominal pain, nausea, and diarrhea are presenting symptoms. Prompt diagnosis and supportive therapy with intensive antibiotics and fluids are required to prevent transmural necrosis and perforation. Surgical resection is indicated in patients with transmural necrosis, intramural perforation, abscess, or uncontrolled sepsis and gastrointestinal hemorrhage. Due to the inherent risks in these critically ill patients of bowel perforation during barium enema administration and colonoscopies, CT is the study of choice in typhlitis. CT demonstrates circumferential mural thickening (1-3 cm) of the cecum, low-density areas within the colonic wall secondary to edema, pericolonic inflammation and fluid, and, in severe cases, pneumatosis. Clinically, CT is used to monitor the decrease in mural thickness with therapy and to detect subtle pneumoperitoneum in cases of silent perforation or necrosis [22, 23].

Graft-Versus-Host Disease Colonic complications of graft-versus-host disease (GVHD) are common. In this disorder, mature donor lymphocytes attack recipient tissue in bone marrow transplantation patients. Acute GVHD may manifest with profuse diarrhea, nausea, vomiting, intestinal hemorrhage, and cramping abdominal pain. Transmural inflammation is common but frank perforation is unusual. Initially, CT shows diffuse mural thickening of the colon with submucosal edema and increased intraluminal secretions. In chronic GVHD, submucosal fat may be rapidly deposited and is readily identified on CT [20].

Differential Diagnosis of the Colitides The differentiation of granulomatous colitis and ulcerative colitis is important in terms of medical management, surgical options, and prognosis. This distinction can usually be made on the basis of colonoscopy, with biopsy histology, double-contrast barium enema, disease distribution, and clinical course. CT can occasionally help distinguish these disorders by demonstrating differences in mural thickness, wall density, distribution of colonic

involvement, and the presence or absence of small bowel disease, abscess, fistula, and fibrofatty mesenteric proliferation. Idiopathic IBD must also be differentiated from the infectious colitides. Although there is considerable overlap in the CT findings of these disorders, there are certain differentiating features [24]. For example, the presence of ascites is more suggestive of an acute rather than a chronic cause of colonic inflammation. Peritoneal fluid is commonly found in the acute colitides, particularly pseudomembranous, infectious, and ischemic colitis, but not in chronic IBD. Ascites is only infrequently seen in patients with acute IBD. Submucosal fat deposition, seen on CT, is primarily found in the subacute and chromic colitides, usually ulcerative colitis, but not in acute disease.

Appendicitis Acute appendicitis is the most common abdominal surgical emergency, affecting 250,000 individuals in the USA annually. The lifetime risk of developing acute appendicitis is 8.6% for men and 6.7% for women. Radiologists play a critical role in evaluating patients with suspected appendicitis and in minimizing its complications by confirming or excluding the diagnosis in atypical cases. They also can reduce the number of misdiagnoses and negative laparotomies, provide a correct alternate diagnosis, and manage appendiceal abscesses and postoperative complications. For example, one study reported a drop in the negative appendectomy rate from 24 to 3% after the widespread use of CT [25]. Of the cross-sectional imaging tests, CT has been the most extensively studied in terms of accuracy, impact on patient management, length of hospital stay, and cost savings. Contrast-enhanced helical CT has a sensitivity, specificity, and accuracy of >95% in the diagnosis of acute appendicitis [26-28]. US, when carried out with the graded compression technique, has been shown to be a valuable approach as well (Fig. 5) [29], although head-tohead comparative studies show that CT has a significant higher accuracy than US [30]. On CT, the abnormal appendix presents as a slightly distended, fluid-filled (Fig. 6) or collapsed structure approximately 0.5-2 cm in diameter. The fluid level within the lumen is >2.6 mm [31]. As a result of the inflammatory hyperemia, the wall of the diseased appendix shows hyperenhancement during the arterial phase of contrast administration. The wall is circumferentially and asymmetrically thickened (usually 1-3 mm). Periappendiceal inflammation, the hallmark of appendicitis, is characterized by increased hazy density or linear fat stranding in the adjacent mesoappendix, by fluid-containing abscesses, and by ill-defined, heterogeneous soft-tissue densities representing a phlegmon [26-28]. There may be secondary inflammatory and edematous changes – with thickening of the wall of the adjacent ileum and cecum – that may mimic primary ileocolic inflammatory disease.

42

Fig. 5 Acute appendicitis. Ultrasound shows a thickened, fluidfilled appendix, surrounding infiltration and free fluid

Jaap Stoker, Richard M. Gore

The combination of right lower quadrant inflammation, phlegmon, and an abscess adjacent to the cecum is suggestive but not diagnostic of appendicitis. Indeed, if an abnormal appendix or an appendicolith is not shown, the differential diagnosis also must include Crohn’s disease, cecal diverticulitis, ileal diverticulitis, perforated cecal or appendiceal carcinoma, and pelvic inflammatory disease. Abscesses may be found in locations distant from the cecum because of the length and position of the appendix and the patterns of fluid migration in the peritoneal cavity. It should be noted that (depending on referral patterns) the majority (60-70%) of patients with suspected appendicitis who are referred for cross-sectional imaging do not have this disease. Instead, most of these patients have benign, self-limited gastrointestinal disorders such as viral gastroenteritis. CT and US often can suggest a specific alternate diagnosis [26-28]. Adnexal cysts, masses, salpingitis, and tubo-ovarian abscesses are readily shown on US. Ureteral calculi and pyelonephritis can be detected on CT and US. Enlarged lymph nodes in the right lower quadrant suggest mesenteric adenitis or infectious ileitis; mural thickening of the terminal ileum can be seen in Crohn’s disease or infectious ileitis. Although CT has a higher accuracy than US, the latter technique can be used initially, to reduce radiation exposure. Only those patients with a negative or inconclusive examination should proceed to CT [34]. Another approach is to use MRI instead of CT, which combines the benefits of a lack of ionizing radiation exposure with high-contrast resolution cross-sectional imaging [35, 36]. Initial reports on MRI in diagnosing acute appendicitis are encouraging, particularly concerning pregnant women (Fig. 7) [37, 38].

Fig. 6. Acute appendicitis. Coronal reformatted CT image shows a thickened, fluid-filled appendix (arrow) along the lateral aspect of the right psoas muscle

CT after intravenous contrast medium administration is the preferred technique, as it facilitates identification of the inflamed appendix and suggests alternative diagnoses. In patients with (imminent) renal insufficiency, a non-contrast CT can be performed. The diagnosis of acute appendicitis on non-contrast CT scans requires the detection of a thickened appendix (diameter >6 mm) with associated inflammatory changes in the periappendiceal fat or abnormal thickening of the right lateroconal fascia, with or without a calcified appendicolith. The addition of coronal and sagittal reformatted images increases diagnostic confidence by virtue of the more reliable demonstration of the entire appendix, surrounding fat and lymph nodes, and periappendiceal infection and inflammation [32]. Confidence in image interpretation also improves with the use of thinner reconstruction sections [33].

Fig. 7. Acute appendicitis. Coronal HASTE (half-Fourier acquisition single-shot turbo spin-echo) in a 28-year-old woman at 18 weeks gestation, clinically suspected of having appendicitis but in whom ultrasound was non-diagnostic. MRI shows a thickened retrocecal appendix (arrow) with increased signal intensity and minimal infiltration of the surrounding fat. The MRI diagnosis of appendicitis was confirmed at surgery

43

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

Scan times are short, namely 10-15 min. At present, however, MRI is not widely used in the Emergency Department for the diagnostic work-up of patients with acute abdominal pain, due to a lack of availability and expertise and the uncertainly as to its cost-effectiveness. Further studies should be directed at better defining the role of MRI in acute abdominal pain, especially its role compared to US and CT. Nonetheless, in pregnant women with an inconclusive US, MRI is currently the preferred imaging technique.

Diverticulitis It is estimated that 10-25% of individuals with diverticulosis will suffer from episodes of peridiverticular inflammation during their lifetime. In the USA, this complication accounts for approximately 200,000 hospitalizations and a health-care expenditure of four billion dollars annually. Among the patients who are hospitalized, 10-20% require emergency surgery [39]. Clinical signs indicative of diverticulitis are inaccurate, although the combination of direct tenderness only in the left lower quadrant, the absence of vomiting, and an elevated C-reactive protein level is suggestive in 25% of these patients (Laméris et al., personal communication). Inflammatory change in the pericolic fat (Fig. 8) is the hallmark of diverticulitis on CT and is seen in 98% of patients with the disease. The extent of the inflammatory reaction is related to the size of the perforation, degree of bacterial contamination, and the host response. Mild cases may manifest as areas of slight in-

crease in the density of the fat adjacent to the involved colon or as fine linear stranding with small fluid collections or bubbles of extraluminal air. In sigmoid diverticulitis, the fluid is typically decompressed into the inferior interfascial plane. Due to the hypervascularity of the inflamed area, contrast-enhanced CT scans often reveal engorged mesenteric vessels in the involved pericolic fat. Pericolic heterogeneous soft-tissue densities representing phlegmons and partially loculated fluid collections indicating abscess are seen in more severe cases. The abscess cavities usually contain air bubbles or air-fluid levels. They develop within the sigmoid mesocolon or are sealed off by the sigmoid colon and adjacent small bowel loops. Less commonly, they may form in the groin, flank, thigh, psoas muscle, subphrenic space, or liver [40]. On CT, diverticula are seen at the site of perforation or adjacent to it in about 80% of cases. They appear as small outpouchings of air, contrast, or fecal material projecting through the colonic wall. Symmetrical mural thickening of the involved colon of approximately 4-10 mm is found in about 70% of cases; however, if there is marked muscular hypertrophy, the wall of the colon can measure up to 2-3 cm in thickness. CT can also demonstrate intramural abscesses and fistula, and is helpful in patients with suspected colovesical fistulas. In the latter case, a pericolic inflammatory mass involves the bladder wall; the presence of intraluminal gas confirms the diagnosis. CT has a reported sensitivity of up to 98% in the diagnosis of diverticulitis [40]. Additionally, it can demonstrate disease extent, such as abscess and peritonitis remote from the colon, and can guide percutaneous abscess drainage. The diagnosis of other pathological conditions that may clinically simulate diverticulitis can also be achieved with CT. Although the accuracy of US in most prospective studies is not inferior to that of CT [41], it has its limitations and a recent large cohort study demonstrated the superiority of CT. Moreover, CT is more accurate in indicating alternative diagnoses [41] and provides a better overview of disease extent, which is important for clinical management (Hinchey classification). An initial study demonstrated that MRI is also accurate in diagnosing diverticulitis [42], but further studies are needed to evaluate its role. The major advantage of MRI in middle-aged patients with diverticulitis is avoidance of intravenous contrast medium and thus of contrastinduced nephropathy.

Epiploic Appendagitis

Fig. 8. Acute diverticulitis. Coronal reformatted CT image shows mural thickening of the sigmoid colon, inflammatory changes, and a gas bubble (arrow) in the sigmoid mesocolon

Primary epiploic appendagitis is a relatively uncommon condition that results from acute ischemia and inflammation of the appendices epiploicae. This disorder is often associated with torsion and infarction of these appendices and can simulate diverticulitis if it occurs in the sigmoid

44

Fig. 9. Epiploic appendagitis at the junction of the descending colon and sigmoid colon. A fatty density mass is surrounded by increased attenuation. Thrombosed vessels (arrow) can be seen in the central portion of the epiploic appendage

and appendicitis if located in the proximal colon. A characteristic appearance of a small, round, or oval fatcontaining mass with an associated inflammatory reaction of the pericolic fat is seen on US and CT (Fig. 9) [43]. The thrombosed vessels within the affected appendage epiploica can sometimes be visualized. Epiploic appendagitis is a self-limited process with clinical resolution in a few days. Follow-up CT examination may show total resolution, shrinkage, and eventual calcification of the inflamed and infarcted epiploic appendix.

Acute Abdominal Pain Inflammatory conditions of the colon are the most frequent causes of acute abdominal pain requiring urgent treatment (e.g., appendicitis, diverticulitis). In the majority of patients with acute abdominal pain, the clinical diagnosis is unclear or incorrect such that imaging is mandatory [44]. In this setting, imaging tests give different results when used in a setting of a specific disease (e.g., acute appendicitis) than when referring to all patients with acute abdominal pain (a myriad of diagnoses). Thus, papers reporting on a specific diagnosis are less informative than those relatively scarce papers reporting on patients with acute abdominal pain in general. Plain abdominal radiographs have a low accuracy in this setting and result in management changes in only 4% of the patients [45, 46]. In the majority of patients, further imaging is warranted after plain radiography. US was shown to increase the

Jaap Stoker, Richard M. Gore

diagnostic accuracy of clinical evaluation (from 70 to 83%) and resulted in changes in management in 22% of patients [47-49]. CT has a significant impact on the diagnosis, as shown in one series in which the accuracy improved from 71% in the pre-CT clinical diagnosis to 93% in the post-CT diagnosis, with a change in management of 46% [50]. Also, the level of confidence increases significantly with CT [51]. In two randomized controlled trials from the same institution, CT was studied: (1) within 24 h versus routine workup by plain X-ray or (2) when considered necessary, US, CT, or fluoroscopy and CT within 1 h versus routine workup. The first study demonstrated significantly fewer deaths (0 vs. 7; p=0.014) [52]. Length of hospital stay was not significantly different but with CT there was a significant overstatement of serious diagnoses. In the second study, routine CT significantly improved diagnostic certainty, but there were no other significant differences to routine workup in patient management [53]. The overall inter-observer agreement of abdominal CT is good [54], and for specific urgent diagnoses it is very good (e.g., for appendicitis, diverticulitis, and bowel obstruction the values are 0.84, 0.90, and 0.81, respectively). A cohort study evaluated 11 diagnostic strategies of clinical diagnosis, plain radiography (supine abdominal X-ray and upright chest X-ray), US, and CT in 1021 patients with acute abdominal pain [34]. Clinical diagnosis (performed by surgical residents) had a sensitivity of 88% and a specificity of 41% for urgent diagnoses. Plain radiography did not contribute to a higher sensitivity or specificity for urgent cases of acute abdominal pain. Regarding the clinical diagnosis, US reduced the number of false-positive urgent diagnoses to 85%, but at the expense of a lower sensitivity (70%, thus missing 30% of the urgent conditions). CT had the highest sensitivity (89%) and specificity (77%) as a single imaging strategy. The best strategy for the detection of urgent diagnoses was an initial US, reserving CT for patients with negative or inconclusive US examinations [34]. This strategy led to a sensitivity for urgent diagnosis of 94%. Initial US reduced CT use and associated radiation burden by 51% compared to CT in all patients. Strategies driven by body mass index, age, or location of the pain had a lower sensitivity for urgent diagnoses than achieved with this conditional strategy. The use of MRI in acute abdominal pain primarily has been studied in acute appendicitis and diverticulitis [36].

Perianal Fistulas Perianal fistulas mostly occur either as the result of fistulous disease originating from the anal glands near the anal crypts (cryptoglandular hypothesis) or in patients with Crohn’s disease [55]. Infection of the anal glands may result in abscess formation. This is a relatively common condition, with a prevalence of approximately 0.01%, predominantly affecting young adults. The course

45

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

of the fistula track may be simple and superficial or complicated. The latter may be intersphincteric (through the internal anal sphincter and then downward through the intersphincteric space) or trans-sphincteric (transversing not only the internal sphincter but also the external sphincter or puborectis muscle) or have a supralevator extent or an extrasphincteric extension to the rectum, without involvement of the anal sphincter. These complicated tracks need detailed imaging for proper therapy, as inadequate treatment may lead to recurrent disease. The surgeon must be aware both of the presence and number of tracks and of their extent, the location of the internal opening, and the presence of abscesses. Pre-operative evaluation of perianal fistulas may include physical examination, examination under anesthesia (EUA), endoscopic ultrasound (EUS), or MRI. Physical examination has significant shortcomings, especially in patients with recurrent disease. While EUA can be used for determining disease extent, immediately followed by treatment, it has limitations and disadvantages, mostly related to probing. Firstly, not all fistulas have an external opening that can be probed, and probing may miss secondary tracks. It is well-recognized that missed extensions are the commonest cause of recurrence, which reaches 25% in some series. Forceful probing may lead to perforation of the levator plate, worsening the extent of the disease. Patients with recurrent disease are most likely to harbor missed disease but are also the most difficult

to assess. Digital palpation frequently cannot distinguish between scarring due to repeated surgery and induration due to an underlying extension. EUS, which can be enhanced by hydrogen peroxide instillation within the track, may be used to determine disease extent. Although initial reports were encouraging, later studies have been less sanguine especially when EUS was compared to MRI. This discrepancy probably relates to operator expertise since EUS is highly operator-dependent. Insufficient penetration beyond the external sphincter, especially with high-frequency transducers, limits the ability of EUS to resolve ischioanal and supralevator sepsis, with the result that it may miss extensions from the primary tract. This technique is impressive in demonstrating the internal opening whereas infection is distinguished only with difficulty from postoperative fibrosis; however, hydrogen peroxide instillation facilitates this differentiation. Low simple tracks presumably can be identified by EUS as accurately as by MRI, but the latter is definitely superior in cases involving complex or high tracks [56]. MRI has been proven to provide the most comprehensive assessment of patients with perianal fistulas, facilitating accurate identification of tracks and extensions as well as abscesses. MRI examination for perianal fistulas should include T2-weighted sequences in multiple planes, a fat-saturation sequence, and a contrastenhanced (fat-saturation) T1-weighted sequence [57]. Tracks (Figs. 10, 11) are identified on T2 as hyper-

a

b

a

b

Fig. 10 a, b. Crohn’s disease and perianal fistulas. a Axial T2-weighted turbo spin-echo demonstrates multiple tracks (arrowheads), intersphincteric and outside the anal sphincter. On the right, an abscess (A) is seen in the ischioanal space. b Coronal T2-weighted turbo spin-echo demonstrates a trans-sphincteric track in the ischioanal space (arrowhead) and extension of the abscess (A) in the levator ani at both sides and in the rectal wall (arrow)

Fig. 11 a, b. Cryptoglandular perianal fistula. a Coronal T2-weighted turbo spin-echo with endoanal coil shows a fibrous track (arrowhead) extending through the left external sphincter. Normal sphincter anatomy is seen on the right. E External sphincter, I internal sphincter, PR puborectal muscle, LA levator ani plate. b Axial T2-weighted turbo spinecho demonstrates the hyperintense track (arrow), surrounded by fibrous tissue (arrowhead), along the left external sphincter (E)

46

intense longitudinal structures, often with a hypointense, fibrous wall. Collateral inflammation is often appreciated on fat saturation sequences. After the administration of intravenous contrast medium, the lining of the wall will enhance. Non-enhancing fluid can be identified in the center of the track or the track can be completely obliterated by granulation tissue. In the latter case, there is complete enhancement of the part of the track that is hyperintense on T2 sequences (Fig. 11). Abscesses are readily appreciated on fat saturation sequences, although some small fluid collections may be more difficult to identify. The use of external phased array coils may result in limitations in detecting superficial extensions and difficulty in locating the precise level of the internal opening. In such cases, endoluminal MRI may provide more information (Fig. 11). A prospective triple-blinded comparison of the accuracy of anal endosonography (AES), pelvic MRI, and surgical EUA in perianal Crohn’s disease showed that AES correctly classified fistulas in 91% of the cases, compared with 87% for pelvic MRI and 91% for surgical evaluation [58]. A combination of any two of the three modalities increased the accuracy to 100%. Another study, in which MRI and AES were compared with surgical findings, showed MRI to be superior to AES in fistula classification, with sensitivities of 84 vs. 60% for the two modalities, and specificities of 68 and 21%, respectively. Several studies have indicated a positive effect of preoperative MRI on patient outcome. In one study, the therapeutic effect of MRI before EUA was 21.1% [59]. Additionally it was shown that disease recurrence after surgery could be reduced by about 75% if surgery was guided by the MRI findings. The differential diagnosis of perianal fistulas primarily concerns fistulas originating from skin appendages: acne conglobata, suppurative hidranitis, and pilonidal sinus. The first two disorders are easy to recognize clinically, but this is more difficult with pilonidal sinus. Imaging can be used to differentiate between perianal fistula and pilonidal sinus. In a study in seven patients with pilonidal sinus and 14 sex- and age-matched individuals with perianal fistulas, these conditions could be readily discriminated by the absence in the former of intersphincteric sepsis or an enteric opening [60]. Osteomyelitis of the pelvis or femur may give rise to abscesses and tracks that extend to the anorectal region, whereas osteomyelitis is a rare finding in perianal fistulas due to Crohn’s disease. Differentiating these two conditions is usually not difficult, as the predominant disease localization (either extensive bone marrow edema or extensive tracks with intersphincteric extension and internal opening) will establish the diagnosis.

References 1. Gore RM, Laufer I, Berlin JW (2008) Ulcerative and granulomatous colitis: idiopathic inflammatory bowel disease. In: Gore RM, Levine MS (eds) Textbook of gastrointestinal radiology, 3rd edn. Saunders, Philadelphia, pp 1071-1109

Jaap Stoker, Richard M. Gore

2. Markose G, Ng CS, Freeman AH (2003) The impact of helical computed tomography on the diagnosis of unsuspected inflammatory bowel disease in the large bowel. Eur Radiol 13:107-113 3. Furukawa A, Saotome T, Yamasaki M et al (2004) Cross-sectional imaging in Crohn disease. Radiographics 24:689-702 4. Horsthuis K, Bipat S, Bennink R, Stoker J (2008) Inflammatory bowel disease diagnosed with US, MR, scintigraphy, and CT: Meta analysis of prospective studies. Radiology 247:64-79 5. Johnson KT, Hara AK, Johnson CD (2009) Evaluation of colitis: usefulness of CT enterography technique. Emerg Radiol 16:277-282 6. Desmond AN, O’Regan K, Curran C et al (2008) Crohn’s disease: factors associated with exposure to high levels of diagnostic radiation. Gut 57:1524-1529 7. Ripolles T, Martinez MJ, Pardes JM et al (2009) Crohn disease: Correlation of findings at contrast-enhanced US with severity at endoscopy. Radiology 253:241-248 8. Migaleddu V, Scanu AM, Quaia E et al (2009) Contrast-enhanced ultrasonographic evaluation of inflammatory activity in Crohn’s disease. Gastroenterology 137:43-61 9. Ziech M, Stoker J (2010) MRI of the small bowel: enterography. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 117-134 10. Papanikolaou N, Gourtsoyianni S (2010) MRI of the small bowel: enteroclysis. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 135-148 11. Rimola J. Rodriguez S, Gracia-Bosch O et al (2009) Role of 3.0-T colonography in the evaluation of inflammatory bowel disease. Radiographics 29:701-719 12. Ajaj W, Lauenstein TC, Langhorst J et al (2005) Small bowel hydro-MR imaging for optimized ileocecal distension in Crohn’s disease: should an additional rectal enema filling be performed? J Magn Reson Imaging 22:92-100 13. Zijta F, Stoker J (2010) Magnetic resonance imaging of the colon (Colonography): Results. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 185-204 14. Ajaj WM, Lauenstein TC, Pelster G et al (2005) Magnetic resonance colonography for the detection of inflammatory diseases of the large bowel: quantifying the inflammatory activity. Gut 54:257-263 15. Horsthuis K, Bipat S, Stokkers PC, Stoker J (2009) Magnetic resonance imaging for evaluation of disease activity in Crohn’s disease: a systematic review. Eur Radiol 19:1450-1460 16. Röttgen R, Herzog H, Lopez-Häninnen E, Felix R (2006) Bowel wall enhancement in magnetic resonance colonography for assessing activity in Crohn’s disease. Clin Imaging 30:27-31 17. Langhorst J, Kühle CA, Ajaj W et al (2007) MR colonography without bowel purgation for the assessment of inflammatory bowel diseases: diagnostic accuracy and patient acceptance. Inflamm Bowel Dis 13:1001-1008 18. Rimola J, Rodríguez S, García Bosch O et al (2009) Magnetic resonance for assessment of disease activity and severity in Crohn disease. Gut 58:1113-1120 19. Kelly CP, LaMont JT (2008) Clostridium difficile – more difficult than ever. N Engl J Med 359:1932-1940 20. Horton KM, Corl FM, Fishman EK (2000) CT evaluation of the colon: inflammatory disease. Radiographics 20:399-418 21. Turner DR, Markose G, Arends MJ et al (2003) Unusual causes of colonic wall thickening on computed tomography. Clin Radiol 58:191-200 22. Gluecker TM, Williamson EE, Fletcher JG et al (2003) Diseases of the cecum: a CT pictorial review. Eur Radiol 13 Suppl 4:L51-61 23. Kirkpatrick ID, Greenberg HM (2003) Gastrointestinal complications in the neutropenic patient: characterization and differentiation with abdominal CT. Radiology 226:668-674 24. Thoeni RF, Cello JP (2006) CT imaging of colitis. Radiology 240:623-638

47

Imaging of the Colon and Rectum: Inflammatory and Infectious Diseases

25. Raman SS, Osuagwu FC, Kadell B et al (2008) Effect of CT on false positive diagnosis of appendicitis and perforation. N Engl J Med 358:972-973 26. Pinto Leite N, Pereira JM, Cunha R et al (2005) CT evaluation of appendicitis and its complications: imaging techniques and key diagnostic findings. AJR Am J Roentgenol 185:406-417 27. Daly CP, Cohan RH, Francis IR et al (2005) Incidence of acute appendicitis in patients with equivocal CT findings. AJR Am J Roentgenol 184:1813-1820 28. Paulson EK, Harris JP, Jaffe TA et al (2005) Acute appendicitis: added diagnostic value of coronal reformations from isotropic voxels at multi-detector row CT. Radiology 235: 879-885 29. Puylaert JB, Rutgers PH, Lalisang RI et al (1987) A prospective study of ultrasonography in the diagnosis of appendicitis. N Engl J Med 317:666-669 30. van Randen A, Bipat S, Zwinderman AH et al (2008) Acute appendicitis: meta-analysis of diagnostic performance of CT and graded compression US related to prevalence of disease. Radiology 249:97-106 31. Moteki T, Ohya N, Horikoshi H (2009) Prospective examination of patients suspected of having appendicitis using new computed tomography criteria including “maximum depth of intraluminal appendiceal fluid greater than 2.6 mm. J Comput Assist Tomogr 33:383-389 32. Kim YJ, Kim J-E, Kim HS, Hwang HY (2009) MDCT with coronal reconstruction: clinical benefit in evaluation of suspected acute appendicitis in pediatric patients. AJR Am J Roentgenol 192:150-152 33. Johnson PT, Horton KM, Kawamoto S et al (2009) MDCT for suspected appendicitis: effect of reconstruction section thickness on diagnostic accuracy, rat of appendiceal visualization, and reader confidence using axial images. AJR Am J Roentgenol 192:893-901 34. Laméris W, van Randen A, van Es HW et al (2009) Imaging strategies for detection of urgent conditions in patients with acute abdominal pain: diagnostic accuracy study. BMJ 338:b2431 35. Lee KS, Pedrosa I (2010) Magnetic resonance imaging of acute conditions of the gastrointestinal tract. In: Stoker J (ed) Magnetic resonance imaging of the gastrointestinal tract. Springer-Verlag, Berlin, Heidelberg, pp 283-314 36. Stoker J (2008) Magnetic resonance imaging and the acute abdomen. Br J Surg 95:1193-1194 37. Oto A, Ernst RD, Ghulmiyyah LM et al (2009) MR imaging in the triage of pregnant patients with acute abdominal and pelvic pain. Abdom Imaging 34:243-250 38. Cobben L, Groot I, Kingma L (2009) A simple MRI protocol in patients with clinically suspected appendicitis: results in 138 patients and effect on outcome of appendectomy. Eur Radiol 19:1175-1183 39. Humes DJ, Solyamani-Dodaran M, Fleming KM et al (2009) A population-based study of perforated diverticular disease incidence and associated mortality. Gastroenterology 136:11981205 40. Kircher MF, Rhea JT, Kihiczak D, Novelline RA (2002) Frequency, sensitivity, and specificity of individual signs of diverticulitis on thin-section helical CT with colonic contrast material: experience with 312 cases. AJR Am J Roentgenol 178:1313-1318 41. Laméris W, van Randen A, Bipat S et al (2008) Graded compression ultrasonography and computed tomography in acute

42.

43. 44. 45.

46. 47. 48. 49. 50. 51.

52.

53.

54. 55. 56. 57. 58.

59. 60.

colonic diverticulitis: meta-analysis of test accuracy. Eur Radiol 18:2498-2511 Heverhagen JT, Sitter H, Zielke A, Klose KJ (2008) Prospective evaluation of the value of magnetic resonance imaging in suspected acute sigmoid diverticulitis. Dis Colon Rectum 51:1810-1815 Singh AK, Gervais DA, Hahn PF et al (2005) Acute epiploic appendagitis and its mimics. Radiographics 25:1521-1534 Stoker J, van Randen A, Laméris W, Boermeester MA (2009) Imaging in acute abdominal pain. Radiology 253:31-46 MacKersie AB, Lane MJ, Gerhardt RT et al (2005) Nontraumatic acute abdominal pain: unenhanced helical CT compared with three view acute abdominal series. Radiology 237:114-122 Kellow ZS, MacInnes M, Kurzencwyg D et al (2008) The role of abdominal radiography in the evaluation of the nontrauma emergency patient. Radiology 248:887-893 Walsh PF, Crawford D, Crossling FT et al (1990) The value of immediate ultrasound in acute abdominal conditions: a critical appraisal. Clin Radiol 42:47-49 Allemann F, Cassina P, Rothlin M, Largiader F (1999) Ultrasound scans done by surgeons for patients with acute abdominal pain: a prospective study. Eur J Surg 165:966-970 Dhillon S, Halligan S, Goh V et al (2002) The therapeutic impact of abdominal ultrasound in patients with acute abdominal symptoms. Clin Radiol 57:268-271 Tsushima Y, Yamada S, Aoki J et al (2002) Effect of contrastenhanced computed tomography on diagnosis and management of acute abdomen in adults. Clin Radiol 57:507-513 Rosen MP, Sands DZ, Longmaid HE, 3rd et al (2000) Impact of abdominal CT on the management of patients presenting to the emergency department with acute abdominal pain. AJR Am J Roentgenol 174:1391-1396 Ng CS, Watson CJ, Palmer CR et al (2002) Evaluation of early abdominopelvic computed tomography in patients with acute abdominal pain of unknown cause: prospective randomised study. BMJ 325:1387 Sala E, Watson CJ, Beadsmoore C et al (2007) A randomized, controlled trial of routine early abdominal computed tomography in patients presenting with nonspecific acute abdominal pain. Clin Radiol 62:961-969 van Randen A, Lameris W, Nio CY et al (2009) Inter-observer agreement for abdominal CT in unselected patients with acute abdominal pain. Eur Radiol 19:1394-1407 Halligan S, Stoker J (2006) Imaging fistula-in-ano. Radiology 239:18-33 Ziech M, Felt-Bersma R, Stoker J (2009) Imaging of perianal fistulas. Clin Gastroenterol Hepatol 7:1037-1045 Horsthuis K, Stoker J (2004) MRI of perianal Crohn’s disease. AJR Am J Roentgenol 183:1309-1315 Schwartz DA, Wiersema MJ, Dudiak KM et al (2001) A comparison of endoscopic ultrasound, magnetic resonance imaging, and exam under anesthesia for evaluation of Crohn’s perianal fistulas. Gastroenterology 121:1064-1072 Buchanan G, Halligan S, Williams A et al (2002) Effect of MRI on clinical outcome of recurrent fistula-in-ano. Lancet 360:1661-1662 Taylor SA, Halligan S, Bartram CI (2003) Pilonidal sinus disease: MR imaging distinction from fistula in ano. Radiology 226:662-667

IDKD 2010-2013

CT Colonography: Updated Daniel C. Johnson1, Michael Macari2 1 Department 2 Department

of Radiology, Mayo Clinic, Scottsdale, AZ, USA of Radiology, New York University Langone School of Medicine, New York, NY, USA

Computed tomography (CT) colonography has been in development for more than a decade, with hundreds of articles now published on its performance and technical capabilities. With the conclusion and publication of the National CT Colonography trial [1] and endorsement of the technique for screening by a multi-society task force (including the American Cancer Society, American College of Radiology, US Multi-society Task Force on Colorectal Cancer) [2], the clinical validation of CT colonography in the prepared colon has been completed. This chapter highlights the most important current issues for CT colonography. Patient acceptance of routine colorectal screening, including CT colonography, remains a major barrier. In 2000, only 43% of US adults age 50 or older had undergone a sigmoidoscopy or colonoscopy within the previous 10 years or had used a fecal occult blood home test kit within the preceding year [3]. The major disincentive for patients considering CT colonography as a screening option is the laxative purgation (the same as that required for colonoscopy) [4]. Advantages include the lack of required sedation and intravenous line placement for CT colonography, a quick return to work following the examination, and no need to inconvenience others for transportation to and from the exam. The risk of perforation at CT colonography is considerably less than at colonoscopy. Furthermore, the examination only requires two breath-holds on the CT scanner (in the supine and prone positions), with the completion of most average examinations in 10 min, which may help reassure hesitant patients. Still, the reality of a full bowel preparation, an enema tip, and full (although brief) colonic insufflation is likely to delay the decision for screening for some people. The performance of CT colonography has undergone exhaustive testing. The Pickhardt trial demonstrated a sensitivity similar to colonoscopy [5], but concerns were raised that community practices might not be able to achieve these results. The National CT Colonography trial (ACRIN 6664) studied 2531 individuals nationally across 15 centers, including academic and private practices. The findings of this trial were similar to those of the Pickhardt trial and have reassured many groups [1]. Radiologist’s training and testing were required for the

ACRIN trial. Although some participants required more training than others, all of them received a passing score of 90% for easy and moderately difficult to detect lesions [6]. The ACRIN trial also insisted on strict adherence to protocol requirements, including stool tagging regimens, mechanical insufflation of the colon, and thin-section and low-dose CT techniques [7]. It is clear that meticulous attention to all aspects of the examination is required to achieve optimal results. Extracolonic abnormalities are common in patients of screening age [8-11]. A pragmatic approach to these findings is needed; for example, radiologists should recommend follow-up studies for those patients with findings most likely to be of clinical significance. Patients (and clinicians) will be grateful if additional testing is minimized; for those that need addition studies, the recommended optimal follow-up should be included in the report. It is unfortunate that the risk associated with the low radiation dose required for CT colonography has been misunderstood. The standard dose at CT colonography is about half of the dose used for a standard body CT examination. This results in an average dose of approximately 5 mSv. The real risk of this exposure is unknown, but the Health Physics Society has stated that for doses in this range the risks for the development of radiationinduced cancer are too small to measure or are nonexistent [12]. Even if a very small risk is assumed from radiation exposure at CT, it must be balanced against the risk of developing colon cancer and of other alternative procedures. The risk of perforation (1:1000) and death (1:17,000) at colonoscopy are real and can be measured [13], but as a society there is a consensus that these risks are outweighed by the risk of developing colon cancer (about 1 in 13 without screening) [14]. Maintaining high-quality interpretations is a responsibility that each individual, each practice, and our specialty should assume. The American College of Radiology has established a national CT colonography database within the National Radiology Data Registry (NRDR) [15]. Selected process and outcome metrics can be quickly entered on-line and compared to national benchmarks. These measures include process metrics related to the CT

CT Colonography: Updated

technique and the adequacy of patient preparation, and outcome metrics related to colon perforation, true-positive and false-positive rates for large (≥1 cm) polyps, and the prevalence of significant extracolonic findings. Practices seriously interested in providing the best care should be encouraged to participate in this data registry and manage their practice such that benchmark metrics are achieved. A spirit of cooperation between radiologists and gastroenterologists is needed for optimal patient care. Guidelines will need to be jointly developed for the proper use of colonography and colonoscopy, and for processes to efficiently transfer patients with polyps to colonoscopy. Those practices that are able to do this effectively will offer patients a service of high value – and will likely find themselves very busy. In summary, CT colonography has completed its clinical validation and is now ready for widespread clinical application. Radiologists committed to performing the examination to the highest quality must obtain the education and equipment needed. We must focus our efforts on the best in patient care, and ignore the political distractions that will come. We have an obligation to educate referring physicians on the correct use of the technique. Collaborations with gastroenterologists to ensure sameday polypectomy for selected patient will enhance patient care. Extracolonic findings must be vigilantly and properly reported so that only those patients with highly significant lesions are recommended for additional followup testing. Lastly, we should be committed to ongoing quality measures to both improve and maintain the highest standards of care. Radiology has another exciting opportunity to serve the public, and to potentially help reduce the incidence of a common cancer killer.

References 1. Johnson CD, Chen MH, Toledano A et al (2008) Accuracy of CT colonography for detection of large adenomas and cancer. N Engl J Med 359:1207-1217

49

2. Levin B, Lieberman DA, McFarland B et al (2008) Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. Gastroenterology 134:1570-1595 3. Colorectal (Colon) Cancer. http://cdcgov/cancer/colorectal/ statistics/screening_rateshtm 4. Beebe TJ, Johnson CD, Stoner S et al (2007) Assessing attitudes toward laxative preparation in colorectal cancer screening and effects on future testing: potential receptivity to computed tomographic colonography. Mayo Clinic Proceedings 82:666-671 5. Pickhardt PJ, Choi JR, Hwang I et al (2003) Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. New Eng J Med 349:2191-2200 6. Fletcher JG, Johnson CD, Toledano A et al (2005) ACRIN 6664: Lessions for CT colonography (CTC) training and certification. Radiological Society of North America Scientific Assembly and Annual Meeting Program, Chicago, IL 7. Johnson CD, Chen MH, Toledano A et al. The National CT Colonography Trial Protocol, ACRIN 6664. http://wwwacrinorg/ Portals/0/Protocols/6664/Protocol-ACRIN%206664%20 Amendment%201,%207706pdf 8. Gluecker TM, Johnson CD, Wilson LA et al (2003) Extracolonic findings at CT colonography: evaluation of prevalence and cost in a screening population. Gastroenterology 124:911916 9. Hara AK, Johnson CD, MacCarty RL, Welch TJ (2000) Incidental extracolonic findings at CT colonography. Radiology 215:353-357 10. Hellstrom M, Svensson MH, Lasson A (2004) Extracolonic and incidental findings on CT colonography (virtual colonoscopy). AJR Am J Roentgenol 182:631-638 11. Rajapaksa RC, Macari M, Bini EJ (2004) Prevalence and impact of extracolonic findings in patients undergoing CT colonography. Journal of Clin Gastroenterol 38:767-771 12. Radiation risk in perspective (2004) Position Statement of the Health Physics Society 13. Waye JD, Kahn O, Auerbach ME (1996) Complications of colonoscopy and flexible sigmoidoscopy. Gastrointest Endosc Clin N Am 6:342-377 14. Lifetime Probability of Developing or Dying From Cancer. http://wwwcancerorg/docroot/CRI/content/CRI_2_6x_Lifetime_Probability_of_Developing_or_Dying_From_ Cancerasp?sitearea=&level= 15. National Radiology Data Registry. https://nrdracrorg/portal/ Nrdr/Main/pageaspx

IDKD 2010-2013

Imaging of Diffuse and Inflammatory Liver Diseases Pablo R. Ros1, Rendon C. Nelson2 1 Department 2 Department

of Radiology, University Hospitals Health System, Case Western Reserve University, Cleveland, OH, USA of Radiology, Duke University, Durham, NC, USA

Introduction The category of diffuse liver diseases includes a variety of disorders that typically involve the liver in a non-focal fashion. It is important to note, however, that even within this group there may be focal abnormalities, representing unusual expression of a disease that typically has a diffuse manifestation. At the same time there may be focal neoplasms, some of which are benign and others malignant. With recent advances in cross sectional imaging, the detection, characterization, and follow-up of diffuse liver disease has been greatly facilitated. This chapter is divided along the traditional classification of cirrhosis, vascular disorders, congenital, metabolic and storage, and neoplastic diseases. In addition, diffuse and focal inflammatory/infectious diseases are discussed.

Fibrous Tissue Deposition (Cirrhosis) Background The preliminary events that lead to cirrhosis involve a mechanism of injury whereby there is repeated exposure of some noxious agent to the liver, resulting in hepatocyte injury and/or death. Etiological noxious agents include alcohol, viral infection (specifically hepatitis B and hepatitis C), non-alcoholic steatohepatitis, autoimmune disorders (such as primary sclerosing cholangitis and primary biliary cirrhosis), and toxic agents (such as aflatoxin and iron, specifically, primary hemochromatosis). Pathologically, the liver attempts to repair itself from the injury by regenerating hepatocytes and depositing collagen. With repeated exposure to the noxious agent, there is ongoing hepatocyte regeneration and collagen deposition, resulting in the formation of regenerative nodules and, eventually, the formation of fibrous scars, respectively. Note that fibrous tissue deposition, specifically in the form of collagen, can be treated and is reversible to a point. However, with more severe or profound degrees of collagen and fibrous tissue deposition, this process is no longer reversible and is henceforth referred to as cirrhosis. Con-

siderable research has been devoted to the non-invasive quantification of fibrosis using ultrasound (US) and magnetic resonance imaging (MRI). Some of these techniques use elastography, in which a mechanical impulse is used to push or distort the liver followed by measurement of the amount of liver movement. A stiffer, more fibrotic liver will move less than a soft, less fibrotic liver.

Regenerative Nodules Regenerative nodules are classified as micro-regenerative or macro-regenerative. Micro-regenerative nodules measure <3 mm in diameter while macro-regenerative nodules are those ≥3 mm. Macro-regenerative but not microregenerative nodules are large enough to be visualized with US, computed tomography (CT), or MRI. Furthermore, in 10-20% of patients (not 10-20% of regenerative nodules), iron deposition occurs within the cytoplasm of the regenerating hepatocytes, leading to the formation of so-called siderotic nodules. This is in contradistinction to transfusional hemosiderosis, in which excess iron is phagocytized by the reticuloendothelial cells or Kuppffer cells, not the hepatocytes. Although somewhat controversial, there is evidence that intracellular iron has a toxic effect that may increase the risk of hepatocellular carcinoma (HCC) development later in life. Since 7-10% of patients with cirrhosis develop HCC and 10-20% of patients with cirrhosis develop siderotic nodules, there may well be a correlation between the two entities.

Confluent Hepatic Fibrosis The coalescence of fibrous tissue leads to the formation of a large fibrous scar in the liver; this is referred to as confluent hepatic fibrosis. It typically occurs in the middle of the liver, specifically, the anterior segment of the right hepatic lobe (segments V and VIII) and the medial segments of the left hepatic lobe (segments IVa and IVb). The large fibrous scar emanates from the more central portion of the liver and is associated with peripheral capsular atrophy and retraction, which at times can be profound. The enhancement pattern of confluent hepatic fibrosis on CT is similar to that of fibrous tissue elsewhere;

51

Imaging of Diffuse and Inflammatory Liver Diseases

that is, the scar is usually hypoattenuating during the unenhanced state, hypo- to iso-enhancing during the late hepatic arterial phase, hypo- to iso-enhancing during the portal venous phase, and iso-to hyper-enhancing during the equilibrium phase (Fig. 1). The hyper-enhancement phenomenon noted during the equilibrium phase occurs as a result of slow accumulation and then slow wash-out of contrast material within fibrous tissue. On MRI, the appearance of confluent hepatic fibrosis may be different than that of fibrous tissue elsewhere in the body, such as ligaments and tendons. Due to the high water content within this fibrous scar, hepatic fibrosis is typically hypointense on T1-weighted images and hyper-intense on T2weighted images. Whereas the signal-intensity characteristics parallel those of hepatic neoplasms, the location, shape, and capsular retraction aid in distinguishing the two entities. Following Gd-chelate administration, the scan reveals slow wash-in and slow wash-out, similar to the pattern obtained with iodinated agents on CT.

ing the ratio between the size of the caudate lobe and right hepatic lobe. The transverse measurement of the caudate lobe medially to the bifurcation of the right portal vein, divided by the transverse dimension of the entire right hepatic and caudate lobes, shows low sensitivity but high specificity for cirrhosis when exceeding 0.60-0.65. Nodularity of the capsular surface of the liver is also a feature of cirrhosis. The appreciation of nodularity, however, is related to the size of the regenerative nodules. With smaller regenerative nodules, as may occur in alcoholic cirrhosis, it may be difficult to detect capsular nodularity. Among the disorders characterized by the development of large regenerative nodules, such as primary sclerosing cholangitis, there may be deep nodularity. Furthermore, well-defined clefts can be identified in the capsular surface, particularly on the gastric side of the lateral segment of the left hepatic lobe and the renal side of the posterior segment of the right hepatic lobe. Lastly, there tends to be prominence of the hepatic fissures, specifically, the gallbladder fossa and the fissure for the falciform ligament.

Hepatic Morphology in Cirrhosis Portal Hypertension The morphology and shape of the liver typically change in patients as they develop progressive fibrosis. Although there are variations from patient to patient and from disease to disease, the most typical pattern is atrophy of the right hepatic lobe and hypertrophy of the left hepatic and caudate lobes. As a result of right hepatic lobar atrophy, there is a decrease in the angle of the gallbladder fossa with the horizontal. Furthermore, caudate lobe hypertrophy can be quantified with US, CT, or MRI by measur-

Fig. 1 a-d. Confluent hepatic fibrosis in a 55-year-old man with cirrhosis. Axial images through the liver during the a unenhanced state, b late hepatic arterial phase, c portal venous phase, and d equilibrium phase. There is a wedgeshaped area in the supcapsular portion of the medial segment of the left hepatic lobe that demonstrates hypoattenuation pre-contrast and slow wash-in and slow wash-out post-contrast. There is also associated capsular retraction

Signs of portal hypertension in patients with cirrhosis include splenomegaly and/or porto-systemic shunts. A spleen with a volume exceeding 250 mL or with a longitudinal dimension >12 cm or an anteroposterior dimension >9 cm is considered enlarged. In some patients with splenomegaly, focal iron deposits can be appreciated within the parenchyma of the spleen. These are seen on MRI as hypointense foci on either T1-weighted gradient

a

b

c

d

52

Pablo R. Ros, Rendon C. Nelson

echo or T2-weighted pulse sequences and are referred to as Gamna-Gandy bodies. Porto-systemic shunts can occur either intrahepatically or extrahepatically. Intrahepatic porto-systemic shunting is a result of high-pressure, low-volume arterial blood that mixes at the level of the hepatic sinusoids with lowpressure, high-volume venous blood. In the cirrhotic liver, arterial flow is increased and portal venous flow is decreased, thereby increasing the prevalence of these shunts. During the arterial phase of imaging, there is typically a parenchymal blush that is associated with early portal venous enhancement. Extrahepatic porto-systemic shunts are also seen in the setting of cirrhosis. The most common is a spontaneous spleno-renal shunt whereby venous blood from the spleen is shunted to the left renal vein via the left inferior adrenal vein. In this setting, a prominent collateral vein is usually seen draining directly into an enlarged left renal vein. Porto-systemic varices are also common in patients with cirrhosis and portal hypertension. These include esophageal varices, splenic hilar varices, perigastric varices, and peripancreatic varices. In some patients, profound shunting of blood from the splenic vein retrograde into the inferior mesenteric vein precludes the development of splenomegaly.

Ancillary Findings in Cirrhosis Other findings that often occur in the setting of cirrhosis include ascites, mesenteric edema, and bowel wall edema, particularly in the small bowel. Some of these changes are due to portal hypertension although they may be exacerbated by hypoproteinemia.

a

c

Vascular Disorders Arterial-Portal Shunts Although some arterial-portal shunts appear to arise spontaneously, many of them are due to prior liver intervention, such as a biopsy, or in the setting of cirrhosis, as noted above. Furthermore, they are often associated with other vascular malformations, such as focal nodular hyperplasia and cavernous hemangiomas. On contrast-enhanced CT or MRI, findings include a focal or wedge-shaped parenchymal blush, a large hepatic artery, early portal venous enhancement, and a possible increase in the luminal diameter of the portal vein (Fig. 2). Most of these arterial portal shunts are subclinical and do not require intervention.

Perfusion Abnormalities Perfusion abnormalities are common following the administration of contrast agents. They are seen as focal, often web-shaped, areas of hyperenhancement that are most pronounced during the late hepatic arterial phase, hence the term transient hyperattenuation defect (THAD) for CT and transient hyperintensity defect (THID) for MRI. Although the etiology of these defects is not always apparent, many of them are due to alterations in portal venous flow. For example, they are often visualized in the setting of liver tumors (either primary or metastatic) that block flow from a branch of the portal vein. They can also be seen when there are aberrant veins draining directly into the liver, leaving parenchyma that receives venous blood from the

b

d

Fig. 2 a-d. Arterial-portal shunt in a 49-year-old man with hepatitis C. Axial images through the liver during the late hepatic arterial phase demonstrate a wedgeshape area of focal parenchymal hyperenhancement in the left hepatic lobe (a), a large hepatic artery (b) as well as early and vivid enhancement of a large periumbilical collateral vein (c, d). These findings are consistent with an intrahepatic arterial-portal shunt

53

Imaging of Diffuse and Inflammatory Liver Diseases

aberrant source only and not from the portal vein. These perfusion abnormalities typically occur: a) in the subcapsular area; b) in the periligamentous area (about both the fissure for the falciform ligament and the ligamentus venosum), due to aberrant internal mammary or periumbilical veins; c) about the porta hepatis, due to aberrant gastric veins (right more common than left); d) about the gallbladder fossa, due to aberrant cholecystic veins. On contrast-enhanced CT or MRI, they appear as transient areas of hyperenhancement during the late hepatic arterial phase and are often web-shaped and subcapsular. There is usually not an appreciable abnormality in this area on the unenhanced images or during the portal venous or equilibrium phases.

of the portal vein, they are uncommonly associated with portal vein thrombosis, either bland or malignant. On CT and MRI, occlusive malignant portal vein thrombosis typically enlarges the portal vein, the luminal diameter of which may exceed 23 mm. Furthermore, the thrombus itself may enhance and this enhancement can be quite vivid during the late hepatic arterial phase. With Doppler US, sampling of the thrombus itself reveals low-resistance arterial wave forms that often flow in the hepatofugal direction. In patients with cirrhosis and HCC who are anticipating liver transplantation, it is important to determine whether portal vein thrombosis is bland or malignant. If the thrombosis is indeed malignant, the patient cannot be considered a candidate for transplantation. The best way to confirm the diagnosis is to percutaneously biopsy the intraluminal thrombus itself under direct real-time US guidance.

Portal Vein Thrombosis

Budd-Chiari Syndrome

In portal vein thrombosis, the thrombus can be either bland or malignant and either occlusive or non-occlusive. Bland thrombosis typically occurs in the setting of trauma, cirrhosis, following an orthotopic liver transplant at the end-to-end anastomosis, in certain hypercoagulable states, and in 25% of patients with BuddChiari syndrome. If the thrombus is not occlusive, an eccentric intraluminal filling defect is identified, which often resolves with anti-coagulant therapy. In this case, there is no cavernous transformation (Fig. 3). With occlusive portal vein thrombosis, however, the imaging findings are different. Early on, the portal vein demonstrates a non-enhancing intraluminal filling defect that may distend the vein and increase the luminal diameter. Furthermore, there is often enhancement of the wall of the vein via the vaso vasorum. Over time, typically in the range of 3-6 weeks, the thrombus undergoes retraction, with the development of cavernous transformation, mainly via collaterals that develop in the vaso vasorum itself. Malignant portal vein thrombosis typically occurs in patients that have HCC, in the setting of either cirrhosis or, less commonly, non-cirrhosis. Interestingly, although metastases to the liver commonly obstruct small branches

Budd-Chiari syndrome may occur when there is obstruction to the outflow of blood from the hepatic veins. In Western countries, such as the USA, the majority of these cases (70%) are idiopathic. In the Orient, however, they are commonly due to congenital webs that occur within the hepatic veins themselves. Other disorders that are associated with Budd-Chiari syndrome include trauma, pregnancy, certain hypercoagulable states, and malignant hepatic vein thrombosis. The latter is commonly associated with primary tumors of the liver, right adrenal gland, and right kidney, specifically HCC, adrenal cortical carcinoma, and renal cell carcinoma, respectively. Budd-Chiari syndrome can occur with occlusion of one, two, or all three hepatic veins or with occlusion of the suprahepatic inferior vena cava. Furthermore, the imaging findings may differ depending upon whether the disease is acute or chronic. With acute Budd-Chiari syndrome, there may be one or more intraluminal thrombi, most commonly identified by Doppler US. With chronic Budd-Chiari syndrome, however, the hepatic veins are small and often difficult to identify, although tortuous intrahepatic collateral veins or shunts may be apparent. In this setting, shunts may develop from one hepatic vein

a

Fig. 3 a, b. Non-occlusive portal vein thrombosis in a 39-year-old woman with a long history of taking birth control pills. a Axial T2weighted and b post-contrast (late hepatic arterial phase) images through the liver demonstrate an eccentric intraluminal defect in the main portal vein

b

54

a

c

Pablo R. Ros, Rendon C. Nelson

b

Fig. 4 a-d. Axial images through the liver during the portal venous phase demonstrate intense enhancement of the central portion of the liver and hypoenhancement of the peripheral portion (especially on a and b). The hepatic veins are not visualized. There is a cleft in the medial aspect of both the right and left hepatic lobes (c, d) consistent with chronic Budd-Chiari syndrome and peripheral parenchymal atrophy

d

that is obstructed to another hepatic vein that is not. Shunting can also occur from a hepatic vein that is obstructed to the hepatic vein in the caudate lobe. Finally, shunting may occur from a hepatic vein that is obstructed to the portal vein, one of the reasons why 25% of patients with Budd-Chiari syndrome develop portal vein thrombosis. Furthermore, chronic occlusion of the hepatic vein can result in significant enlargement of the caudate lobe and atrophy of the peripheral portion of the right and left hepatic lobes. At times, large intrahepatic collateral veins can be seen shunting blood to the hepatic vein in the caudate lobe. Furthermore, on T1-weighted MRI, the caudate lobe is often hyperintense. Following contrast administration on either CT or MRI, there is often differential enhancement of the central and peripheral portions of the liver (Fig. 4); that is, early on, the central portion of the liver hyperenhances but the periphery does not. Later on, there is a flip-flop phenomenon in which the central portion washes out and the peripheral portion accumulates contrast media. Over time, profound atrophy of the peripheral parenchyma can be seen. Patients commonly have and first clinically present with ascites, which develops shortly after hepatic vein occlusion. In suprahepatic inferior vena caval obstruction, the hepatic vein in the caudate lobe cannot be used as a conduit to shunt blood from the hepatic veins to the inferior vena cava. As a result, there is neither hypertrophy of the caudate lobe nor development of large intrahepatic collateral veins. It is important to note that, in the setting of Budd-Chiari syndrome, the liver is often swollen, which

may narrow the lumen of the intrahepatic inferior vena cava. Although this is not the cause of Budd-Chiari syndrome, it certainly exacerbates the condition.

Passive Hepatic Congestion Right-sided heart failure can result in the delayed drainage of blood from the liver into the inferior vena cava and right atrium. Images through the lower chest typically demonstrate either cardiac enlargement or a large pericardial effusion, although in the setting of restrictive pericarditis the heart may be normal in size. The key finding with right-sided congestive heart failure or passive hepatic congestion is enlarged and distended hepatic veins and inferior vena cava. Secondary signs include reflux of contrast material from the superior vena cava into the hepatic veins, although this is occasionally seen in normal patients as well. In addition, there may be a mottled enhancement pattern in the liver that is more pronounced peripherally and more apparent during the late hepatic arterial phase. This pattern, referred to as the “nutmeg liver”, may not be apparent during the portal venous and/or equilibrium phases. Over time, the liver can become fibrotic, at which point it may have all of the manifestations of cirrhosis and portal hypertension.

Macro-regenerative Nodules In a small percentage of patients with outflow obstruction of the hepatic veins, large regenerative nodules develop

55

Imaging of Diffuse and Inflammatory Liver Diseases

that tend to hyperenhance during the late hepatic arterial phase. These nodules have been associated with both Budd-Chiari syndrome and passive hepatic congestion. They typically are iso-enhancing during the portal venous and equilibrium phases and demonstrate only slow, periodic growth over time.

Hepatic Veno-occlusive Disease Hepatic veno-occlusive disease (VOD) typically occurs in bone marrow transplant recipients who have undergone total body irradiation. The definition of VOD is progressive non-thrombotic occlusion of the hepatic venules. Although some reports have noted sporadic reversal or hepatofugal flow in the portal vein with VOD, there are no reliable imaging findings for this diagnosis. As a result, VOD can only be reliably diagnosed by microscopic examination of biopsy tissue.

Peliosis Hepatis Peliosis hepatis is a rare entity in which the hepatic sinusoids throughout the liver dilate, resulting in numerous blood-filled lacunar spaces ranging in size from 1 to 3 mm. Similar lacunar spaces can also occur in the spleen, lymph nodes, bone marrow, and lungs. Although the cause of peliosis is poorly understood, it is believed to be due to outflow obstruction of the sinusoid. It typically occurs in patients who use anabolic steroids, corticosteroids, tamoxifen, or birth control pills; following cardiac or renal transplantation; with chronic debilitating diseases, such as tuberculosis, malignancy or AIDS; in association with diabetes, sprue, or Hodgkin’s disease; or in patients exposed to arsenic or polyvinyl chloride. Imaging shows numerous small cystic lesions that demonstrate an enhancement pattern similar to that of the blood pool.

Hepatic Infarction Parenchymal infarction in the liver is relatively uncommon for two reasons: 1. the liver has a dual blood supply; 2. the hepatocytes are relatively insensitive to hypoxia. It has been seen, however, in patients with shock, sepsis, eclampsia, sickle cell disease or trait, or arteritis; in those who have taken birth control pills; and in those who have suffered arterial embolic events such as those due to endocarditis, rheumatic heart disease, trauma, intra-arterial chemotherapy, or iatrogenic tumor embolization. With imaging, infarcted parenchyma may be hypoechoic on US: anechoic bile lakes may be visualized as necrosis progresses. On contrast-enhanced CT, there is a wedge-shaped subcapsular region of hypoenhancement that later on may contain bile lakes and gas bubbles. On MRI, the edema of infarction is typically hypointense on T1-weighted images and hyperintense on T2-weighted images; as with CT, hypoenhancement and bile lakes are seen as well.

Metabolic and Storage Diseases Steatosis Hepatic steatosis results from a variety of abnormal processes, including the increased production or mobilization of fatty acids (e.g., obesity, steroid use) or the decreased hepatic clearance of fatty acids due to hepatocellular injury (e.g., alcoholic liver disease, viral hepatitis). Histopathologically, the hallmark of all forms of fatty liver is the accumulation of fat globules within hepatocytes. The distribution of steatosis can be variable, ranging from focal, to regional, to diffuse. Diffuse steatosis is common and estimated to occur in approximately 30% of obese patients. Patients with steatosis are usually asymptomatic although some may present with right upper quadrant pain or abnormal liver function parameters. The histopathological findings of non-alcohol-related liver steatosis, also known as non-alcoholic fatty liver disease (NAFLD), vary from steatosis alone to steatosis with inflammation, necrosis, and fibrosis. Non-alcoholic steatohepatitis (NASH), with or without cirrhosis, is positioned at the most severe end of the NAFLD spectrum. The histopathological findings of NASH include steatosis (predominately macrovesicular), mixed lobular inflammation, and hepatocellular ballooning. Unlike steatosis alone, NASH may progress to cirrhosis. Diffuse fatty change is easily identified on CT. The attenuation value of normal liver is usually ~8 HU greater than that of spleen on non-contrast CT images. In patients with fatty change, however, an abnormally decreased density will be demonstrated, typically 10 and 25 HU less than the spleen on non-contrast and contrast-enhanced CT images, respectively. The diagnosis of hepatic steatosis is more reliably made on non-contrast images. Undoubtedly, the most sensitive technique to detect fatty change of the liver is the use of in-phase and out-phase gradient echo MRI pulse sequences (Fig. 5). Hepatic fatty change is, however, not always uniform but can instead present as a focal area of steatosis in an otherwise normal liver (focal steatosis) or as subtotal fatty change with sparing of certain areas (focal sparing). On imaging, several features allow the correct identification of focal fatty change or focal spared areas: 1. the typical periligamentous and periportal location; 2. lack of mass effect; 3. sharply angulated boundaries of the involved area; 4. non-spherical shape; 5. absence of vascular displacement or distortion; 6. lobar or segmental distribution.

Iron Overload Iron overload states may arise from hemochromatosis, with the preferential accumulation of iron within hepatocytes, or hemosiderosis, in which iron is deposited in Kupffer cells.

56

a

Pablo R. Ros, Rendon C. Nelson

b

c

Fig. 5 a-c. Diffuse fatty liver in a 41-year-old female presenting with epigastric pain. a Axial CECT image demonstrates diffuse low attenuation of the liver without displacement of the hepatic vessels. b In- and c out-of-phase T1-weighted images show significant signal drop in the liver on the out-of-phase images

Primary Hemochromatosis Hereditary or primary hemochromatosis is an autosomal recessive disorder of iron metabolism that is characterized by the abnormal absorption of iron from the gut, with subsequent excessive deposition of iron into hepatocytes, pancreatic acinar cells, myocardium, joints, endocrine glands, and skin. In addition, cells of the reticuloendothelial system (RES) in patients with primary hemochromatosis are abnormal and unable to store processed iron effectively. Consequently, these patients do not accumulate iron into the RES. The clinical findings of cirrhosis and its complications (portal hypertension, development of HCC) predominate in patients with longlasting disease. On CT, excessive iron storage within hepatocytes will result in an overall increased density. However, this CT appearance of a hyperdense liver is non-specific, since similar features can be seen with gold deposition and in Wilson’s disease, type IV glycogen storage disease, and following amiodarone administration. The use of noncontrast CT in patients with suspected hemochromatosis is important because excessive iron cannot be detected in the setting of enhancing parenchyma. MRI is far more specific than any other imaging modality for the characterization of iron overload, due to the magnetic susceptibility effect of iron. The superparamagnetic effect of accumulated iron in the hepatocytes results in a significant reduction of signal intensity on T2-weighted images. Comparison of the signal intensity of liver with that of paraspinal muscles provides a useful internal control. HCC, seen in 35% of patients with advanced hemochromatosis, is usually easily detected on T1- and T2-weighted images due to the background of decreased signal intensity of the liver.

thalassemia major, sideroblastic anemia, pyruvate kinase deficiency, chronic liver disease), the excess iron is processed and accumulates in organs containing reticuloendothelial cells, including liver, spleen, and bone marrow. Diffuse, increased attenuation of the liver and spleen is seen on CT (Fig. 6). On MRI, the extrahepatic signal intensity changes in the spleen and bone marrow allow primary hemochromatosis to be distinguished from hemosiderosis. Although, in general, the clinical significance of transfusional iron overload states is negligible, patients with chronic hemosiderosis can develop symptoms similar to those of the primary form as well as cirrhosis and HCC. Wilson’s Disease Wilson’s disease, also known as hepatolenticular degeneration, is a rare autosomal recessive abnormality of copper metabolism that is characterized by the accumulation

Hemosiderosis In patients with hemosiderosis or siderosis, due to transfusional iron overload states or dyserythropoiesis (e.g.,

Fig. 6. Hemosiderosis in a 45-year-old female with long history of sickle cell anemia requiring multiple transfusions. Axial nonenhanced CT image demonstrates increased attenuation of the liver

57

Imaging of Diffuse and Inflammatory Liver Diseases

of toxic levels of copper in the brain, cornea (KayserFleischer rings), and liver, the latter due to impaired biliary excretion. Hepatic deposition of copper, predominantly seen in periportal areas and along the hepatic sinusoids, evokes an inflammatory reaction resulting in acute hepatitis with fatty change. Subsequently, chronic hepatitis may result in liver fibrosis and, eventually, macronodular cirrhosis. Due to the high atomic number of copper, a hyperdense liver may be seen on unenhanced CT scans. However, this finding is not universally present; instead, usually only non-specific signs, such as hepatomegaly, fatty change, and in advanced cases, cirrhosis, are observed. During the early stage of the disease, and due to the paramagnetism of ionic copper, MRI can be valuable as it demonstrates focal copper depositions, such as multiple nodular lesions, typically appearing hyperintense and hypointense on T1- and T2-weighted images, respectively.

Fig. 7. Diffuse metastatic breast cancer (pseudocirrhosis pattern) in a 43-year-old woman treated for metastatic breast cancer. Axial contrast-enhanced CT image demonstrate several small low-density lesions in the liver and a nodular contour of the liver due to hepatic capsular retraction

Amyloidosis Lymphoma In amyloidosis, fibrils of protein-mucopolysaccharide complexes are deposited throughout the body. The disease is classified based on the biochemical composition of the amyloid fibrils. Primary amyloidosis is due to the deposition of immunoglobin light chains and is associated with multiple myeloma and monoclonal gammopathy. Secondary amyloidosis results from the deposition of amyloid A protein and is associated with chronic infection, rheumatoid arthritis, and malignancies. Exceeded only by the spleen and kidney, the liver is the third most common solid organ prone to amyloid deposition. Hepatic amyloidosis has a non-specific imaging appearance, with the most common finding being diffuse hepatomegaly. CT sporadically demonstrates focal areas of low attenuation within the liver, corresponding to sites of amyloid deposition (amyloid pseudotumor).

Neoplastic Diseases Metastatic Disease Neoplastic infiltration due to diffuse metastatic disease can occur with many primary tumors. Melanoma, malignant neuroendocrine tumors, pancreatic adenocarcinoma, breast carcinoma, and colonic adenocarcinoma are some of the more commonly encountered causes of diffuse hepatic metastatic disease. The CT appearances of hepatic metastases depend on the vascularity of the lesions compared with the normal liver parenchyma. Diffuse metastatic involvement may produce only subtle imaging findings and be detectable only through indirect features, such as diffuse parenchymal heterogeneity, vascular and architectural distortion, or alterations of the liver contour. The latter, particularly seen in patients with treated breast cancer metastases, has been described as the “pseudocirrhosis” sign (Fig. 7).

Lymphoma can infiltrate the liver both primarily and secondarily. Primary lymphoma of the liver is extremely rare. Conversely, the liver is often secondarily involved in Hodgkin’s and in non-Hodgkin’s lymphoma. Typically, the liver parenchyma is diffusely infiltrated with microscopic nests of neoplastic cells, without significant architectural distortion. Consequently, lymphomatous involvement is difficult to detect by imaging alone. Associated abnormalities, such as splenomegaly and lymphadenopathy, may narrow the differential diagnosis.

Diffuse Infectious and Inflammatory Diseases Fungal Infections Hepatosplenic fungal infection is a clinical manifestation of disseminated fungal disease in patients with hematological malignancies or compromise of the immune system. The reported prevalence of fungal dissemination ranges from 20 to 40%. Most hepatic fungal microabscesses occur in leukemia patients and are caused by Candida albicans.

Candidiasis Candida albicans in the liver may evoke little or no inflammatory reaction, cause a superlative response, or occasionally produce granulomas. The typical histological pattern of hepatic candidiasis is characterized by microabscesses, with the yeast or pseudohyphal forms of the fungus in the center of the lesion and a surrounding area of necrosis and polymorphonuclear infiltrate. At contrast-enhanced CT, fungal microabscesses usually appear as multiple, round, discrete areas of low attenuation, generally ranging in size from 2 to 20 mm (Fig. 8). These microabscesses usually enhance centrally

58

Pablo R. Ros, Rendon C. Nelson

Hepatic contrast-enhanced CT may typically reveal multiple, diffuse, small, low-density areas in both the liver and spleen. MRI features of hepatic sarcoidosis are also non-specific and include organomegaly, multiple lesions of low signal intensity relative to background parenchyma with all sequences, increased periportal signal, irregularity of the portal and hepatic vein branches, and patchy areas of heterogeneous signal. Tuberculosis

Fig. 8. Hepatic candidiasis in a 66-year-old female with leukemia who presented with abnormal liver function tests. Axial contrastenhanced CT image demonstrates multiple, small, low-attenuation lesions distributed throughout the liver and spleen. Splenomegaly and bilateral pleural effusions are also seen

after intravenous administration of contrast medium, although peripheral enhancement may occur as well. At MRI, the untreated nodules are rounded lesions <1 cm in diameter that are minimally hypointense on T1weighted and gadolinium-enhanced images and markedly hyperintense on T2-weighted images. After treatment, the lesions appear mildly to moderately hyperintense on T1and T2-weighted images and demonstrate enhancement on gadolinium-enhanced images. A dark ring is usually seen around these lesions with all sequences. Completely treated lesions are minimally hypointense on T1-weighted images, isointense to mildly hyperintense on T2-weighted images, moderately hypointense on early gadoliniumenhanced images, and minimally hypointense on delayed gadolinium-enhanced images. MRI is superior to CT and US in the detection of these fungal foci.

Granulomatous Diseases Granulomatous hepatitis is associated with numerous conditions, most commonly, sarcoidosis, tuberculosis, and histoplasmosis. Hepatic granulomas usually appear as discrete, sharply defined nodules consisting of aggregates of epithelioid cells by a rim of mononuclear cells, predominantly lymphocytes.

Tuberculosis is one of the most common infectious diseases worldwide. Generally, tuberculosis of the liver presents as either a miliary form or a local form, which is further subdivided into nodular tuberculosis (i.e., tuberculous abscess and tuberculoma) and tubular or hepatobiliary tuberculosis (i.e., tuberculosis involving the intrahepatic ducts). Hepatic miliary tuberculosis is most common and is reported to occur in 50-80% of all patients with terminal pulmonary tuberculosis. Miliary tuberculosis is usually not detected at imaging. Hepatomegaly may be the only radiological abnormality. In the healing stage of tuberculosis, CT may show diffuse hepatic calcifications (approximately 50% of cases). Reported CT findings of nodular tuberculosis are non-specific and include hypoattenuating lesions both before and after intravenous administration. At MRI, the lesions are hypointense on T1-weighted images and hypo- to iso-intense on T2-weighted images. Tuberculosis lesion differently enhance after gadolinium administration. Given these rather non-specific findings with all imaging techniques, percutaneous liver biopsy is necessary. Histoplasmosis Histoplasmosis is the most common cause of fungal infection in the Ohio River Valley of the USA. Fortunately, 99% of patients exposed to histoplasmosis develop only subclinical infections. Liver involvement is common in disseminated histoplasmosis, which usually originates in the lungs. The most common hepatic findings include portal lymphohistiocytotic inflammation and discrete, well-delineated granulomas. In patients with healed histoplasmosis, the presence of small, punctate calcifications scattered throughout the liver and spleen is typical but a non-specific finding.

Sarcoidosis

Parasitic Infections

Sarcoidosis is a multi-system disorder of unknown pathogenesis, characterized by non-caseating granulomas. Although it may involve almost any organ in the body, pulmonary sarcoidosis is the most common. Sarcoidosis of the liver is also relatively frequently seen, but the granulomas are usually not macroscopically detectable and thus may not produce focal abnormalities on imaging studies. Classically, the granulomas develop in a periportal location, resulting in periportal fibrosis, cirrhosis, and eventually portal hypertension.

Schistosomiasis Schistosoma japonicum, S. hematobium, and S. mansoni are the three most important species that infect humans. The schistosomas live in the bowel lumen and lay eggs in the mesenteric veins. The eggs may then embolize to the portal vein. The eggs themselves do not survive and subsequently calcify. Chronic infections with either S. japonicum or S. mansoni result in the formation of cirrhosis and the risk of development of HCC. Histologically, schistosomiasis

59

Imaging of Diffuse and Inflammatory Liver Diseases

is characterized by white, pinhead-sized granulomas scattered throughout the liver. At the center of each granuloma is a schistosome egg. In severe infections, the surface of the liver shows granulomatous involvement and widespread fibrous portal enlargement (“pipe-stem” fibrosis). At CT, the most pathognomonic pattern is the presence of calcified septa, usually aligned perpendicular to the liver capsule (“tortoise shell” or “turtle back” appearance).

Viral Infections Viral Hepatitis Acute viral hepatitis is a systemic infection that affects the liver and is usually caused by one of five viral agents: hepatitis A virus, hepatitis B virus (HBV), hepatitis C virus, the HBV-associated delta agent or hepatitis D virus, and hepatitis E virus. A vast array of other viruses may also produce hepatitis, including herpes viruses, yellow fever virus, rubella virus, Coxsackie virus, and adenovirus. The diagnosis of acute hepatitis is usually based on serological, virological, and clinical findings. Probably the most important role of radiology in patients with acute hepatitis is to help rule out other diseases that produce similar clinical and biochemical abnormalities, such as extrahepatic cholestasis, diffuse metastatic disease, and cirrhosis. At CT and MRI, the findings in acute viral hepatitis are non-specific and include hepatomegaly and periportal edema. At CT, heterogeneous enhancement and well-defined regions of low attenuation may be present. At MRI, periportal edema appears as areas of high signal intensity on T2-weighted images. Involved areas may be normal or demonstrate decreased signal intensity on T1-weighted images and increased signal intensity on T2-weighted images. There is also impaired uptake of liver-specific agents. Extrahepatic findings in patients with severe acute hepatitis include thickening of the gallbladder wall due to edema and, infrequently, ascites. On CT and MRI, the features of chronic hepatitis resemble those of early-stage liver cirrhosis. Periportal lymphadenopathy may be the sole detectable abnormality in both acute and chronic hepatitis.

HIV Infection The liver and biliary tracts are frequent sites of involvement during the course of HIV infection. Co-infection with HBV and hepatitis C virus is particularly common due to the shared means of transmission of these viruses with HIV. AIDS-related cholangiopathy is the newest common manifestation. At CT, inflammation of the gallbladder or biliary tree manifests as mural thickening or abnormal contrast enhancement. Magnetic resonance cholangiopancreatography is more sensitive and specific than US or CT in depicting the mural irregularity of the extrahepatic ducts, which results from the exuberant periductal inflammation, focal mucosal ulcers, and the interstitial edema found in AIDS-related cholangitis.

Uncommon Hepatic Infections Cat-scratch Disease Cat-scratch disease is an infection that affects immunocompetent children or adolescents. It is caused by Bartonella henselae, a gram-negative bacillus that is usually introduced by the scratch of a cat. Cat-scratch disease takes many forms, from regional lymphadenitis to disseminated infection. The typical clinical manifestation is painful lymphadenopathy proximal to the site of inoculation. Disseminated infection is seen in 5-10% of cases. In the abdomen, multiple granulomas ranging from 3 to 30 mm may form in the liver and spleen, with or without hepatosplenmegaly. Histopathological findings include vascular proliferative lesions (peliosis hepatis) and necrotizing granulomatous lesions. At unenhanced CT, the lesions are hypoattenuating relative to normal parenchyma. Three different patterns at contrast-enhanced CT have been described: 1. persistent hypoattenuation relative to the liver; 2. isoattenuation relative to surrounding tissues; 3. marginal enhancement. There are only a few MRI studies of cat-scratch disease. The lesions appear as nodules of low signal intensity on T1-weighted images and high signal intensity on T2-weighted images. Peripheral enhancement may be seen on gadolinium-enhanced T1-weighted images.

Bacillary Angiomatosis Bacillary angiomatosis is also a manifestation of infection by Bartonella henselae, the same organism that causes cat-scratch disease, but in immunocompromised patients. The disease is characterized by localized areas of vascular proliferation that may affect the skin, airway, mucous membranes, visceral organs, bone, and brain. Contrast-enhanced CT may demonstrate multiple diffuse low- or high-attenuation lesions less than 1 cm that are scattered throughout the hepatic parenchyma. Ascites, mild periportal edema, and intrahepatic biliary ductal dilatation may occur. These imaging features are nonspecific and must be distinguished, especially in AIDS patients, from hepatic abscesses related to other bacterial, viral, or fungal infections, AIDS-related lymphoma, Kaposi sarcoma, and, less commonly, disseminated Pneumocystis carinii infection.

Focal Infections Bacterial (Pyogenic) Abscess Pyogenic abscess, although uncommon in the antibiotic era, is still challenging clinically since its presentation is quite variable, from profound septicemia to chronic, indolent symptoms.

60

Pablo R. Ros, Rendon C. Nelson

Enhanced CT can reliably diagnose over 90% of hepatic pyogenic abscesses, revealing two main patterns: multiple microabscesses (disseminated or clustered) and large macroabscesses. By virtue of its good spatial and contrast resolution, CT is the single best method for detecting hepatic abscess, with a sensitivity as high as 97%. On CT scans, abscesses appear as generally rounded masses that are hypodense on both contrast and noncontrast scans. Central gas, as air bubbles or an airfluid level, is a specific sign, but it is present in less than 20% of cases. A thick, enhancing, peripheral rim is also noted. At MRI, air within the abscess appears as a signal void, thus making it more difficult to differentiate from calcifications. However, the shape and location (air-fluid level) of the abscess should enable the correct diagnosis. After the administration of gadolinium-DTPA, abscesses typically show rim enhancement (the “double target” sign). Small lesions (<1 cm) may enhance homogeneously, mimicking hemangiomas. Percutaneous, image-guided aspiration followed by drainage is the method of choice for definitive diagnosis and treatment, with success achieved in >90% of cases.

Amebic Abscess Hepatic abscess is the most common extraintestinal manifestation of amebiasis, affecting approximately 10% of patients with the disease. Although rare in the continental USA, 10% of the world’s population is infected with Entamoeba histolytica. Clinically, patients with amebic abscess are more acutely ill than those with pyogenic abscess, with high fever and right upper quadrant pain. Diagnosis is made by positive serological amebic titers, although they have false-negative rates of almost 20%. Extrahepatic extensions to the chest wall, pleura, or adjacent viscera are well demonstrated on CT. Percutaneous catheter drainage of an amebic abscess is rarely necessary due to the effectiveness of amebicidal therapy. Occasionally, percutaneous drainage is needed in large, symptomatic abscesses with poor response to medical therapy, suspected bacterial superinfection, and threatening intrapericardial rupture. The CT appearance of amebic abscess is variable and non-specific. The lesions are a

b

usually peripheral, round or oval areas of low attenuation (10-20 HU). A peripheral rim of slightly higher attenuation can be seen on non-contrast scans and shows marked enhancement after the administration of contrast material. On MRI, amebic liver abscesses are spherical and usually solitary lesions with a hyperintense center on T2weighted images and a hypointense center on T1-weighted images. The abscess wall is thick; on gadolinium enhanced images, the enhancement pattern is similar to that of pyogenic abscess.

Echinococcal Disease Hydatid disease has two main forms affecting humans, resulting from infection with either Echinococcus granulosus or Echinococcus multilocularis or alveolaris. These infections have well-defined and different geographical distributions. The pathological and imaging findings differ dramatically between these parasites. On CT scans, E. granulosus-infected sites are seen as unilocular or multilocular, well-defined cysts with either thick or thin walls. Daughter cysts are usually detected as areas of lower attenuation than the mother cyst and are usually in the periphery of the lesion. Daughter cysts can also float free in the lumen of the mother cyst; altering the patient’s position may change the position of these cysts, confirming the diagnosis of echinococcal disease. Curvilinear ring-like calcification is also a common feature. On MRI studies, the appearance of the cyst component of echinococcal cysts is similar to that of other cysts, with long T1 and T2 relaxation times. However, MRI best demonstrates the pericyst, matrix and hydatid sand (debris consisting of freed scolices), and the daughter cysts. The pericyst usually has low signal intensity on T1- and T2-weighted images, because of its fibrous component. This rim and a multiloculated or multicystic appearance are distinctive features. The hydatid matrix appears hypointense on T1-weighted images and markedly hyperintense on T2-weighted images. When present, daughter cysts are hypointense relative to the matrix on both T1- and T2-weighted images (Fig. 9). Fig. 9 a, b. Echinococcus granulosus cyst in a 28-year-old man with right upper quadrant pain. a Axial T2-weighted image demonstrates a large cystic mass in the right lobe of the liver which is surrounded by a hypointense rim and contains more hyperintense smaller cysts in its periphery. b On the axial T1weighted image, the hypointense rim is well visualized and the peripheral cysts are hypointense relative to the center of the lesion

61

Imaging of Diffuse and Inflammatory Liver Diseases

Floating membranes have low signal intensities on T1and T2-weighted images. E. multiloculars appear as a solid large mass or masses, with minimal to no enhancement after intravenous administration of contrast material and possible small punctate calcification.

Suggested Reading General Boll DT, Merkle EM (2009) Diffuse liver disease: Strategies for hepatic CT and MR imaging. Radiographics 29:1591-1614 Cohen EI, Wilck EJ, Shapiro RS (2006) Hepatic imaging in the 21st century. Semin Liver Dis 26:363-372 Foley WD, Jochem RJ (1991) Computed tomography: Focal and diffuse liver disease. Radiol Clin North Am 29:1213-1233 Gourtsoyiannis NC, Ros PR (eds) (2005) Radiologic-pathologic correlations from head to toe: Understanding the manifestations of disease, Springer-Verlag, Berlin, Heidelberg Mergo PJ, Ros PR (1998) Imaging of diffuse liver disease. Radiol Clin North Amer 36:365-375 Mergo PJ, Ros PR, Buetow PC, Buck JL (1994) Diffuse diseases of the liver: Radiologic-pathologic correlation. Radiographics 14:1291-1307 Mitchell DG (1992) Focal manifestations of diffuse liver disease at MR imaging. Radiology 185:1-11 Mortele KJ, Ros PR (2001) Imaging of diffuse liver disease. Semin Liver Dis 21:195-212 Rofsky NM, Fleishaker H (1995) CT and MRI of diffuse liver disease. Semin Ultrasound, CT, and MRI 16:16-33 Ros PR (1997) MR imaging of the liver. In: Magn Reson Imag Clin North Amer, Vol 5, WB Saunders, Philadelphia, PA Ros PR (1998) Hepatic imaging. In: Radiol Clin North Amer, WB Saunders, Philadelphia, PA Ros PR (2002) Hepatic imaging and intervention. In: Clinics in Liver Disease, WB Saunders, Philadelphia, PA Ros PR, Mortele K (2006) CT/MRI of the Abdomen and Pelvis, Lippincott, Williams and Wilkins, Philadelphia, PA

Fibrous Tissue Deposition (Cirrhosis) Aguirre DA, Behling CA, Alpert E et al (2006) Liver fibrosis: Noninvasive diagnosis with double contrast material-enhanced MR imaging. Radiology 239:425-437 Awaya H, Mitchell DG, Kamishima T et al (2002) Cirrhosis: Modified caudate-right lobe ratio. Radiology 224:769-774 Blachar A, Federle MP, Brancatelli G (2001) Primary biliary cirrhosis: Clinical, pathologic, and helical CT findings in 53 patients. Radiology 220:329-336 Bryce TJ, Yeh BJ, Qayyum A et al (2003) CT signs of hepatofugal venous flow in patients with cirrhosis. AJR Am J Roentgenol 181:1629-1633 Cobbold JF, Wylezinska M, Cunningham C et al (2006) Non-invasive evaluation of hepatic fibrosis using magnetic resonance imaging and ultrasound techniques. Gut 55:1670-1672 Colli A, Fraquelli M, Andreoletti M et al (2003) Severe liver fibrosis or cirrhosis: Accuracy of US for detection: Analysis of 300 cases. Radiology 227:89-94 Day CP (2006) From fat to inflammation. Gastroenterology 130: 207-210 Dodd GD III, Baron RL, Oliver JH III et al (1999) End-stage primary sclerosing cholangitis: CT findings of hepatic morphology in 36 patients. Radiology 211:357-362 Dodd GD, Oliver JH, Federle MP et al (1993) Spectrum of imaging findings in hepatic cirrhosis: Pathologic correlation in 500 complete hepatectomy specimens. Radiology 189:421

Faria SC, Ganesan K, Mwangi I et al (2009) MR imaging of liver fibrosis: Current state of the art. Radiographics 29: 1615-1635 Friedman SL, Rockey DC, Bissell DM (2007) Hepatic fibrosis 2006: Report of the third AASLD single topic conference. Hepatology 45:242-249 Hussain S, Reinhold C, Mitchell DG (2009) Cirrhosis and lesion characterization at MR imaging. Radiographics 29:16371652 Ito K, Mitchell DG, Gabata T et al (1999) Expanded gallbladder fossa: Simple MR imaging sign of cirrhosis. Radiology 211: 723-726 Johnson KJ, Olliff JF, Olliff SP (1998) The presence and significance of lymphadenopathy detected by CT in primary sclerosing cholangitis. Br J Radiol 71:1279-1282 Kang HK, Jeong YY, Choi JH et al (2002) Three-dimensional multi-detector row CT portal venography in the evaluation of portosystemic collateral vessels in liver cirrhosis. Radiographics 22:1053-1061 Kim M-J, Mitchell DG, Ito K (2000) Portosystemic collaterals of the upper abdomen: Review of anatomy and demonstration on MR imaging. Abdom Imaging 25:462-470 Kim YJ, Raman SS, Yu NC et al (2007) Esophageal varices in cirrhotic patients: Evaluation with liver CT. AJR Am J Roentgenol 188:139-144 Lee V (2006) Can MR imaging replace liver biopsy for the diagnosis of early fibrosis? Radiology 239:309-310 Lipson JA, Qayyum A, Avrin DE et al (2005) CT and MRI of hepatic contour abnormalities. AJR Am J Roentgenol 184: 75-81 Mortele KJ, Ros PR (2002) MR Imaging in chronic hepatitis and cirrhosis. Semin in Ultrasound, CT, and MRI 23:79-100 Ohtomo K, Baron RL, Dodd GD et al (1993) Confluent hepatic fibrosis in advanced cirrhosis: Appearance at CT. Radiology 188:31-35 Revelon G, Rashid A, Kawamoto S et al (1999) Primary sclerosing cholangitis MR imaging findings with pathologic correlation. AJR Am J Roentgenol 173:1037-1042 Rouviere O, Yin M, Dresner MA et al (2006) MR elastography of the liver: Preliminary results. Radiology 240:440-673

Vascular Disorders Abu-Yousef MM, Milam SG, Farner RM (1990) Pulsatile portal vein flow: A sign of tricuspid regurgitation on duplex Doppler sonography. AJR Am J Roentgenol 155:785-788 Arai K, Matsui O, Takashima T et al (1988) Focal spared areas in fatty liver caused by regional decreased portal flow. AJR Am J Roentgenol 151:300-302 Bargallo X, Gilabert R, Nicolau C et al (2003) Sonography of the caudate vein: Value in diagnosing Budd-Chiari syndrome. AJR Am J Roentgenol 181:1641-1645 Bradbury MS, Kavanaugh PV, Bechtold RE et al (2002) Mesenteric venous thrombosis: Diagnosis and noninvasive imaging. Radiographics 22:527-541 Brancatelli G, Federle MP, Grazioli L et al (2002) Large regenerative nodules in Budd-Chiari syndrome and other vascular disorders of the liver: CT and MR findings with clinicopathologic correlation. AJR Am J Roentgenol 178:877-883 Camera L, Mainenti PP, Di Giacomo A et al (2006) Triphasic helical CT in Budd-Chiari syndrome: Patterns of enhancement in acute, subacute and chronic disease. Clin Radiol 61: 331-337 Cazals-Hatem D, Vilfrain V, Genin P et al (2003) Arterial and portal circulation and parenchymal changes in Budd-Chiari syndrome: A study in 17 explanted livers. Hepatology 37: 510-519 Colegrande S, Centi N, Galdiero R et al (2007) Transient hepatic intensity differences: Part 1. Those associated with focal lesions. AJR Am J Roentgenol 188:154-159

62

Colegrande S, Centi N, Galdiero R et al (2007) Transient hepatic intensity differences: Part 2. Those not associated with focal lesions. AJR Am J Roentgenol 188:160-166 Erden A (2007) Budd-Chiari syndrome: A review of imaging findings. Eur J Radiol 61:44-56 Erden A, Erden I, Krayalcin S et al (2002) Budd-Chiari syndrome: Evaluation with multiphase contrast-enhanced threedimensional MR angiography. AJR Am J Roentgenol 179: 1287-1292 Gallego C, Velasco M, Marcuello P et al (2002) Congenital and acquired anomalies of the portal venous system. Radiographics 22:141-159 Gore RM, Mathieu DG, White EM et al (1994) Passive hepatic congestion: Cross-sectional imaging features. AJR Am J Roentgenol 162:71-75 Harter LP, Gross BH, St Hilaire J et al (1982) CT and sonographic appearance of hepatic vein obstruction. AJR Am J Roentgenol 139:176-178 Iannaccone R, Federle MP, Brancatelli G et al (2006) Peliosis hepatis: Spectrum of imaging findings. AJR Am J Roentgenol 187:W43-W52 Kane R, Eustace S (1995) Diagnosis of Budd-Chiari syndrome: Comparison between sonography and MR angiography. Radiology 195:117-121 Kim HJ, Kim AY, Kim TK et al (2005) Transient hepatic attenuation differences in focal hepatic lesions: Dynamic CT features. AJR Am J Roentgenol 184:83-90 Mathieu D, Vasile N, Menu Y et al (1987) Budd-Chiari syndrome: Dynamic CT. Radiology 165:409-413 Moore EH, Russell LA, Klein JS et al (1995) Bacillary angiomatosis in patients with AIDS: multiorgan imaging findings. Radiology 197:67-72 Pliskin M (1975) Peliosis hepatis. Radiology 114:29-30 Rossi S, Rosa L, Ravetta V et al (2006) Contrast-enhanced versus conventional and color Doppler sonography for the detection of thrombosis of the portal and hepatic venous systems. AJR Am J Roentgenol 186:763-773 Sandrasegaran K, Hawes DR, Matthew G (2005) Hepatic peliosis (bacillary angiomatosis) in AIDS: CT findings. Abdom Imaging 30:738-740

Metabolic and Storage Diseases Kawamori Y, Matsui O, Takahashi S et al (1996) Focal hepatic fat infiltration in the posterior edge of the medial segment associated with aberrant gastric venous drainage: CT, US, and MR findings. J Comput Assist Tomogr 17:590-595 Ko S, Lee T, Ng S et al (1998) Unusual liver MR findings of Wilson’s disease in an asymptomatic 2-year-old girl. Abdom Imaging 23:56-59 Marmoloya G, Karlins NL, Petrelli M, McCullough A (1990) Unusual computed tomography findings in hepatic amyloidosis. Clin Imaging 14:248 Queiroz-Andrade M, Blasbalg R, Ortega CD et al (2009) MR imaging findings of iron overload. Radiographics 29: 1575-1589 Siegelman ES, Mitchell DG, Semelka RC (1996) Abdominal iron deposition: metabolism, MR findings, and clinical importance. Radiology 199:13-22

Pablo R. Ros, Rendon C. Nelson

Infectious Diseases Akhan O, Ozmen MN, Dincer A et al (1996) Liver hydatid disease: long-term results of percutaneous treatment. Radiology 198:259-264 Balci NC, Sirvanci M (2002) MR imaging of infective liver lesions. Magn Reson Imaging Clin N Am 10:121-135 Cheung H, Lai YM, Loke TK et al (1996) The imaging diagnosis of hepatic schistosomiasis japonicum sequelae. Clin Radiol 51:5155 Czermak BV, Unsinn KM, Gotwald T et al (2001) Echinococcus multilocularis revisited. AJR Am J Roentgenol 176:1207-1212 Danon O, Duval-Arnould M, Osman Z et al (2002) Hepatic and splenic involvement in cat-scratch disease: imaging features. Abdom Imaging 25:182-183 Doyle DJ, Hanbidge AE, O’Malley ME (2006) Imaging of hepatic infections. Clin Radiol 61(9):737-748 Giorgio A, Tarantino L, Mariniello N et al (1995) Pyogenic liver abscesses: 13 years of experience in percutaneous needle aspiration with US guidance. Radiology 195:122-124 Hickey N, McNulty JG, Osborne H, Finucane J (1999) Acute hepatobiliary tuberculosis: A report of two cases and a review of the literature. Eur Radiol 9:886-889 Jeffrey RB, Tolentino CS, Chang FC et al (1998) CT of small pyogenic hepatic abscesses: the cluster sign. AJR Am J Roentgenol 151:487-489 Kalovidouris A, Gouliamos A, Vlachos L et al (1994) MRI of abdominal hydatid disease. Abdom Imaging 19:489-494 Lee TY, Wan YL, Tsai CC (1994) Gas-containing liver abscess: Radiological findings and clinical significance. Abdom Imaging 19:46-52 Mendez RJ, Schiebler ML, Outwater EK et al (1994) Hepatic abscesses: MR imaging findings. Radiology 190:431-436 Monzawz S, Ohtomo K, Oba H et al (1994) Septa in the liver of patients with chronic hepatic schistosomiasis japonica: MR appearance. AJR Am J Roentgenol 162:1347-1351 Mortele KJ, Segatto E, Ros PR (2004) The infected liver: radiologic-pathologic correlation. Radiographics 24:937-955 Pastakia B, Shawker TH, Thaler M et al (1988) Hepatosplenic candidiasis: wheels within wheels. Radiology 166:417-444 Ralls PW (2002) Inflammatory disease of the liver. Clin Liver Dis 6:203-225 Ralls PW, Henley DS, Colletti PM et al (1987) Amebic liver abscess: MR imaging. Radiology 165:801-804 Rappaport DC, Cumming WA, Ros PR (1991) Disseminated hepatic and splenic lesions in cat-scratch disease: Imaging features. AJR Am J Roentgenol 156:1227-1228 Ros PR, Sobin LH (1994) Amyloidosis: The same cat, with different stripes. Radiology 190:14-15 Ros PR, Ertuk SM (2007) Focal hepatic infections. In: Gore RM, Levine MS (eds) Textbook of Gastrointestinal Radiology, 3rd edn. WB Saunders, Philadelphia, PA, pp 1663-1684 Sakai T, Maeda M, Takabatake M et al (1995) MR imaging of hepatosplenic sarcoidosis. Radiat Med 13:39-41 Taourel P, Marty-Ane B, Charasset S et al (1993) Hydatid cyst of the liver: comparison of CT and MRI. J Comput Assist Tomogr 17:80-85 Van Allen RJ, Katz MD, Johnson MB et al (1992) Uncomplicated amebic liver abscess: prospective evaluation of percutaneous therapeutic aspiration. Radiology 183:827-830

IDKD 2010-2013

Focal Liver Lesions Wolfgang Schima1, Richard Baron2 1 Department 2 Department

of Radiology, KH Göttlicher Heiland, Wien, Austria of Radiology, University of Chicago, Chicago, IL, USA

Introduction State-of-the-art multidetector computed tomography (MDCT) and magnetic resonance (MR) imaging technologies offer detailed insights into the liver’s anatomy and the pathophysiology of liver disease, such that imaging has become the pacemaker in the development of new therapeutic techniques. Understanding different imaging techniques and the diagnostic potential of different modalities, including contrast utilization, is essential to optimize patient diagnoses. In the current environment of cost containment, the most appropriate modality should be chosen to answer the clinical question. Ultrasonography (US) is widely available, non-invasive, and the least expensive, but it is limited by low sensitivity and specificity unless US contrast agents are used. Instead, contrastenhanced CT has emerged as the modality of choice for routine liver imaging while MR imaging is used primarily as a problem-solving technique for liver evaluation when CT or US results are equivocal or if high concern exists for malignancy in certain high-risk populations. This chapter highlights the imaging of hepatic focal liver lesions, focusing on MDCT and MR imaging, but including ultrasonography in selected cases. It aims to provide readers with an understanding of how to optimize their CT and MR imaging protocols and to familiarize them with the different CT and MR imaging appearances of focal liver lesions. The value of liver-specific MR contrast agents for non-invasive characterization of focal lesions is highlighted as well.

MDCT Imaging Techniques Organ imaging in multiple planes rather than in single sequential slices is the concept underlying MDCT. Thus, in systems consisting of 64 or more detectors the entire liver can be scanned within 1-4 s using a sub-millimeter detector configuration to achieve high-quality multiplanar reconstructions [1]. When viewed axially, reconstructed sections of 2.5-4 mm (overlap 0.5-1 mm) are preferred – because thinner slices do not improve lesion conspicuity – but at the expense of increased image noise [2] and

substantially decreased specificity [3]. The amount of contrast material administered may depend on the patient’s weight, but is typically 42-45 g of iodine (i.e., 120150 mL of contrast at 300 mg iodine/mL, equivalent to 110 mL of contrast at 400 mg/mL). The total amount of iodine given determines the quality of portal-venousphase imaging, with a goal of increasing liver attenuation through contrast enhancement by 50 HU [4]. To achieve good arterial-phase imaging, the flow rate of the contrast material is crucial; a rate of 4-5 mL/s is recommended [5]. Others have used a variable amount of contrast material (e.g., 2 mL/kg body weight) and a fixed injection duration of 30 s, which means that the injection rate varies according to the patient’s weight. Timing of image acquisition in relation to contrast material administration depends on whether one needs to image during the early arterial phase (for arterial anatomy only), late arterial phase (for hypervascular tumor detection/characterization), or portal-venous phase (in some cases of follow-up imaging and hypovascular tumor detection). Routinely, late-arterial-phase imaging (with a delay consisting of the aortic transit time plus 15-18 s) [6, 7] and a portal-venous scan (20-30 s interscan delay or with a fixed delay of ~60-70 s) are acquired, although the actual combination of imaging phases will depend on the individual indication [8]. Automated methods of measuring arterial (aortic) enhancement on CT (often termed bolus tracking), rather than fixed scan times, can be helpful [9].

Magnetic Resonance Imaging Technique Unenhanced T1- and T2-weighted pulse sequences and contrast-enhanced sequences are mandatory in MR examination of the liver. Specific pulse sequences vary by manufacturer, and their use depends on patient compliance and the clinical question being addressed. In- and opposedphase (or out-of-phase) T1-weighted imaging is recommended to allow for maximal tumor detection and in the characterization of fatty tumors and steatosis/non-steatosis. T2-weighted pulse sequences with fat suppression provide better lesion contrast than non-fat-suppressed sequences. Diffusion-weighted imaging, which is helpful for lesion

64

Wolfgang Schima, Richard Baron

detection, is now available on most scanners. In general, dynamic imaging with extracellular gadolinium-based contrast agents is required for lesion characterization, the detection of tumors in cirrhosis, evaluation of the tumor response to therapy, and the detection of marginal tumor recurrences following tumor ablation. Two types of liver-specific MR contrast agents are available. Following their intravenous injection, hepatocyte-specific agents (mangafodipir: Teslascan, GE Healthcare, Norway; gadobenate: MultiHance, Bracco, Italy; and gadoxetic acid: Primovist or Eovist, Bayer-Schering, Germany) are taken up by hepatocytes and provide T1 enhancement of liver tissue. These compounds are used to improve the detection of metastases and to characterize lesions [10-13]. The reticuloendothelial agent ferumoxide (Endorem, Guerbet), a superparamagnetic iron oxide (SPIO), is phagocytosed by Kupffer cells after intravenous infusion and leads to a drop in the T2 signal intensity of liver tissue but not of most parenchymal masses.

Benign Hepatic Lesions Cyst Cysts are common, usually asymptomatic liver lesions with an incidence in the population of 5-14%. At US, hepatic cysts are anechoic, with an imperceptible wall and increased acoustic enhancement behind the cyst. On CT scans, a hepatic cyst appears as a well-circumscribed, homogeneous mass with an attenuation value similar to that of water (<15 HU). Cysts lack any mural thickening or nodularity and do not show contrast enhancement after the administration of intravenous contrast material. Small lesions may appear to have higher attenuation measurements because of partial-volume averaging. Occasionally, unenhanced scans will suggest the diagnosis of small cysts if they are well visualized as hypodense lesions, whereas very small metastases are usually not discernible on unenhanced scans. In most instances, however, these

a

b

small lesions remain problematic for characterization at CT. On MR imaging, cysts are well-defined, homogeneous lesions that are hypointense on T1-weighted images and markedly hyperintense on T2-weighted images. The marked increased in T2 signal intensity of even very small cysts can be very helpful to confirm the benign nature of small lesions.

Hemangioma Hemangioma is the most common benign liver tumor. Its typical US characteristics include a sharply circumscribed, well-defined hyperechoic lesion with distal acoustic enhancement. Small hemangiomas are usually homogeneous in appearance whereas larger ones (>4 cm) are frequently heterogeneous and do not demonstrate all of the characteristic features of these lesions. On unenhanced CT images, hemangiomas are welldefined hypodense masses. On MR imaging, they are hypointense on T1-weighted sequences and markedly hyperintense on T2-weighted sequences. In addition, MR imaging is useful in differentiating hemangiomas from solid neoplasms based on the long T2 relaxation time (= hyperintensity) of hemangioma compared with other hepatic masses [14, 15]. A relatively long T2 echo time (>140 ms) will demonstrate the presence of a homogeneously “light-bulb-bright” lesion, which is characteristic of a benign lesion, either cyst or hemangioma. The exceptions to this would include cystic metastases, and gastrointestinal stromal tumor (GIST) and neuroendocrine tumor metastases. Hemangiomas show three distinctive patterns of enhancement at CT/MR imaging [16], with the common characteristic feature that areas of lesion enhancement closely follow the enhancement characteristics of blood pooling elsewhere [17]. Small lesions (up to ~2 cm) may show immediate and complete filling in the arterial phase, with sustained enhancement in the venous and delayed phases (type I, also termed flash filling) [18] (Fig. 1). The most common enhancement pattern is one of peripheral

c

Fig. 1 a-c. Hemangioma type 1. a Unenhanced CT shows a small hypodense lesion adjacent to the falciform ligament (arrow). b Contrastenhanced CT in the arterial phase shows rapid and complete enhancement of the hemangioma, which persists in the venous phase (c). The attenuation of the hemangioma in the enhanced phases is similar to that of the aorta

65

Focal Liver Lesions

a

b

c

d

Fig. 2 a-d. Hemangioma type III. a T2-weighted turbo spin-echo (TSE) MR image shows a very hyperintense lesion in the right lobe. b-d Dynamic gadolinium-enhanced T1-weighted gradient-recalled echo (GRE) images: b arterial, c venous, and d equilibrium phases show peripheral nodular enhancement with progressive centripetal fill-in. d In the equilibriumphase, after 5 min, there is no complete fill-in

nodular discontinuous enhancement that progresses with increased fill-in over time (type II). Larger lesions (>5 cm) or lesions with central thrombosis/fibrosis may lack central fill-in (type III) (Fig. 2). With SPIO agents, the blood pooling effect, with accumulation of contrast material in the sinuses, may lead to the prolonged enhancement of hemangiomas on T1 images and signal intensity loss on T2 images, despite the lack of Kupffer cells in these lesions. This imaging feature helps in the differentiation between hemangiomas and metastases [19]. Recent studies have shown that non-contrast diffusionweighted imaging may help to differentiate between hemangioma and solid lesions, as the apparent diffusion coefficient of hemangiomas is higher than that of solid lesions [20].

Focal Nodular Hyperplasia This benign lesion is usually of no clinical consequence other than the confusion it causes when incidentally detected during abdominal imaging examinations. The sonographic appearance of focal nodular hyperplasia (FNH) is non-specific; the lesion may be isoechoic, or slightly hypoechoic [21] to liver, while in patients with diffuse hepatic steatosis it is always hypoechoic. One characteristic feature is the presence of a central scar, seen in approximately two-thirds of large lesions but in only one-third of small lesions (<3 cm) [22]. The central scar is most often hyperintense on T2-weighted images, with a comma-shaped or spoke-wheel appearance. This is a key differentiating feature from fibrolamellar hepato-

cellular carcinoma, in which the central scar is predominately of low signal intensity on T2-weighted MR sequences. The use of color/power Doppler US may demonstrate blood vessels within the scar [23]. On unenhanced CT, FNH are isodense or minimally hypodense and are sometimes detectable only by the mass effect on adjacent vessels. On unenhanced MR images, FNH often has a signal intensity similar to that of hepatic parenchyma but usually slightly different on either T1- or T2-weighted images (Fig. 3). Due to the prominent arterial vascular supply, FNH undergoes marked homogeneous enhancement during the arterial phase of contrast-enhanced CT/MR imaging, with rapid wash-out of contrast to isodensity/isointensity on venousphase images [22]. The central scar often enhances on delayed scans [21], a feature typical of a fibrous component. One key aspect is that, other than the scar, these lesions tend to be very homogeneous in appearance. With liver-specific MR contrast agents, FNH shows enhancement on delayed images after the administration of hepatobiliary contrast agents (such as mangafodipir (Fig. 3) and signal loss after the administration of reticuloendothelial agents [24]. This difference is particularly helpful for the differentiation between hypervascular metastases (which do not accumulate liver-specific agents) or hepatic adenoma and incidentally encountered FNH [18]. The influence of oral contraceptives on the growth of FNH is still discussed controversially. Studies with serial imaging have shown FNH growth during follow-up to be rare (3-11%) [25, 26] and not stimulated by oral contraceptives [26].

66

a

Wolfgang Schima, Richard Baron

b

c

d

Fig. 3 a-d. Focal nodular hyperplasia (FNH). a The lesion (arrow) is isointense on T1-weighted imaging, with a small central scar. b On T2, the lesion is also isointense (arrow); the central T2-weighted bright scar is better discernible. Mangafodipir-enhanced T1-weighted images in the axial (c) and coronal (d) planes show homogeneous uptake of the liver-specific agent, typical for FNH (arrow). The central spoke-wheel scar is nicely depicted

Hepatocellular Adenoma This is an uncommon benign neoplasm of low malignant potential [26]. Unlike FNH, the relationship between oral contraceptives and the development of hepatocellular adenoma (HCA) is well established. Histologically, HCA is composed of cells resembling normal hepatocytes but lacking bile ducts. This is a key feature in histologically differentiating between adenoma and FNH [26]. Adenomas are typically hypervascular and, when large, often heterogeneous due to the presence of fat, necrosis, or hemorrhage [26, 27] (Fig. 4). These tumors often contain intratumoral fat, and, as T1-weighted chemical-shift MR imaging is sensitive for detecting fat, this technique can be very helpful in characterizing adenomas. The second typical feature of adenoma is its propensity for spontaneous hemorrhage. If no tumor capsule is present, such hemorrhage may lead to spontaneous rupture, with hypotension and even death. The CT and MR imaging appearances of HCA may be non-specific. On unenhanced CT images, the lesion may be hypodense due to the presence of fat or necrosis. On T1- and T2- weighted MR images, HCA are non-specific in appearance, being usually mildly hypointense on T1 and isointense or mildly hyperintense on T2. If intracellular lipid is present within the lesion, it will often appear hyperintense on T1 in-phase images, with a drop in signal intensity on opposed-phase images [26]. The presence of intratumoral fat helps to narrow the differential diagnosis, as hemangioma can be excluded and metastases and FNH only very rarely contain fat. On dynamic con-

trast-enhanced CT or MR, adenomas usually show marked arterial-phase enhancement with a rapid transition to either isoattenuating or hypoattenuating/intense to hepatic parenchyma on portal-venous-phase imaging. Hepatic adenomas may show signal loss after the administration of reticuloendothelial agents and delayed enhancement after hepatobiliary MR contrast agents. However, they will not show sustained enhancement on delayed imaging to degrees greater than normal liver parenchyma after gadobenate administration, which helps to differentiate FNH from adenoma [28].

Biliary Hamartomas (von Meyenburg Complex) Bile duct hamartomas are congenital malformations of the ductal plate but they have no connections with the bile ducts. These lesions are of no clinical significance but are incidentally encountered in patients undergoing abdominal imaging examinations. They appear as small cystic lesions of round, oval, or irregular shape, found either in the periportal region or diffusely spread throughout the liver (Fig. 5). They are usually <10 mm in diameter and lack contrast enhancement, but occasional peripheral rim enhancement may simulate small hypovascular metastases [27].

Hepatic Abscess and Hydatid Cyst Abscess appearances can vary depending on etiology. Peribiliary abscesses tend to be small and scattered adjacent to the biliary tree; hematogenous distribution via the

67

Focal Liver Lesions

a

b

c

d

Fig. 4 a-d. Hepatocellular adenoma. a T1-weighted in-phase GRE image demonstrates a very large mass in a young woman. The mass is inhomogeneous and shows bright spots, suggestive of hemorrhage (asterisk). b There is a typical drop in signal intensity on the opposed-phase image, indicative of intratumoral fat (arrows), whereas the hemorrhage (methemoglobin, asterisk) does not lose signal intensity. c T2-weighted TSE sequence confirms the presence of hemorrhage (asterisk). Intratumoral fat and hemorrhage are typical for adenoma. d The gadolinium-enhanced image shows moderate and inhomogeneous enhancement. In a large, very inhomogeneous adenoma, malignant degeneration cannot be ruled out radiologically. However, at surgery this lesion was shown to be an adenoma with central hemorrhage

a

b

Fig. 5 a, b. Biliary hamartomas (von Meyenburg complex). Coronal (a) and axial (b) T2-weighted images show multiple small cystic lesions of different sizes and shapes (“starry sky” appearance), typical for biliary hamartomas

hepatic artery or portal vein in appendicitis or diverticulitis tends to lead to larger lesions. US reveals a cystic lesion with internal echoes. On CT, hepatic abscess appears as a hypodense lesion with a capsule that may show enhancement. The cluster sign may be noted when multiple abscesses are present as focal clusters of lesions [29]. The CT appearance of hepatic abscess is non-specific and can be mimicked by cystic or necrotic metastases. Thus, clinical information and laboratory values play a key role in guiding the radiological diagnosis. Although seen in only a small minority of patients, the presence of central gas is highly specific for abscess (Fig. 6). On T1weighted MR images, hepatic abscesses are hypointense

relative to liver parenchyma whereas on T2-weighted sequences they are markedly hyperintense and often surrounded by a local area of slight T2 hyperintensity, representing perilesional edema (Fig. 6). Amoebic liver abscess has a non-specific appearance on CT, but is usually seen as a solitary, hypodense lesion with an enhancing wall that may be smooth or nodular, often associated with an incomplete rim of edema. With MR imaging, the lesions are hypointense on T1-weighted images and heterogeneously hyperintense on T2-weighted images [30]. On CT scan, involvement of the liver by Echinococcus granulosus (hydatid cyst) manifests as unilocular or

68

a

Wolfgang Schima, Richard Baron

b

c

Fig. 6 a-c. a Typical large subcapsular abscess with an air-fluid level and a pleural empyema. b In another patient, CT shows a small, thickwalled abscess after pancreatic surgery. c T2-weighted image of the same patient as in (b) shows the thick indistinct wall of the abscess and peripheral edema

multilocular cysts, with thin or thick walls and calcifications and usually accompanied by daughter cysts. The latter are seen as smaller cysts with septations at the margins of or inside the mother cyst. This appearance is therefore quite different from that of a “usual” multicystic tumor. On MR imaging, a hypointense rim on T1- and T2-weighted images and a multiloculated appearance are considered to be important diagnostic features.

Malignant Primary Tumors Hepatocellular Carcinoma Hepatocellular carcinoma (HCC) is the most common primary liver cancer worldwide and is particularly prominent in Asian and Mediterranean populations. In European countries, HCC occurs mostly in patients with chronic liver disease (hepatitis B or C, liver cirrhosis, or hemochromatosis). These tumors consist of abnormal hepatocytes arranged in a typical trabecular, sinusoidal pattern. They may be solitary, multifocal, or diffusely infiltrating. The imaging appearances of HCC can vary dramatically, but generally can be separated into those based on early versus late presentation. Early presentation is typical of patients with chronic liver disease who undergo

a

imaging-based screening for HCC or frequent liver imaging due to complications of their chronic disease. In these patients, HCC lesions are typically small (<3 cm) and homogeneous in appearance. In non-cirrhotic patients, the disease is usually asymptomatic; thus, by the time symptoms occur and imaging is necessarily performed, the lesion is very large and usually heterogeneous in appearance. The US presentation of HCC is quite variable, with isointensity, hypointensity, or hyperintensity. Smaller lesions are typically homogeneous and larger ones heterogeneous. A surrounding fibrous capsule is often present and relatively characteristic for HCC, appearing as a hypoechoic rim surrounding the lesion. On unenhanced CT images, most HCCs are hypodense. The presence of intratumoral fat can result in the lowered CT attenuation of these tumors; this finding is characteristic of primary hepatocellular tumors. Due to their predominant arterial supply, small HCCs enhance vividly in the arterial phase of hepatic contrast enhancement, becoming isoattenuating or hypoattenuating with hepatic parenchyma in the portal-venous phase of enhancement (so-called wash-out). On delayed images, most HCC lesions are hypodense with surrounding liver (Fig. 7). There have been several studies addressing the required phases of scanning for optimal HCC detection

b

Fig. 7 a, b. Detection of hepatocellular carcinoma with late-arterial-phase multidetector CT. a In the late-arterial phase, MDCT shows a small HCC that was not visible on unenhanced scan. b The delayed-phase scan reveals wash-out of the lesion, which is now slightly hypodense. The combination of arterial hypervascularity and wash-out is a very specific sign of malignancy

69

Focal Liver Lesions

a

b

c

Fig. 8 a-c. Large HCC with tumor thrombus in the inferior vena cava. a Arterial-phase MDCT shows a large HCC with peripheral irregular rim enhancement. The tumor has grown into the right hepatic vein and the inferior vena cava (IVC). b Arterial- and c venous-phase images in the coronal plane better demonstrate the extension of the tumor thrombus into the IVC above the diaphragm

and characterization. Arterial-phase imaging is the most sensitive for the detection of small lesions; as highest visibility is achieved in the late arterial phase, to allow time for contrast diffusion into the tumor parenchyma, there is no need for early-arterial-phase imaging [6, 31, 32]. A venous phase is always necessary for tumor detection and the assessment of venous structures (Fig. 8) as well as other abdominal organs. For HCC detection, the delayed phase can visualize a few lesions that would otherwise go unnoticed [33] and also is very helpful in differentiating HCC from benign enhancing lesions by demonstrating tumor wash-out greater than liver parenchyma [34]. Unenhanced images are important to document siderotic nodules as different from arterial enhancing lesions and to detect intratumoral fat. However, to reduce the radiation dose to the patient, these images should be obtained only intermittently during serial imaging examinations. Nonetheless, they are mandatory in the follow-up after chemoembolization or tumor ablation and when hemorrhage is suspected. In summary, a three- to four-phase MDCT protocol is recommended by most centers to optimally detect and characterize HCC. The presence of focal hypervascularity in the arterial phase may lead to false-positive results [35]: transient focal enhancement of liver parenchyma during arterial phase enhancement, often termed transient hepatic attenuation differences (THAD), can be caused by a multitude of factors. In cirrhotic patients, transient focal enhancement is most often due to arterial-portal shunting, resulting in inappropriately early focal areas of portal-venous distribution enhancement in the liver. These usually are peripheral, often wedge shaped, and not round. Subcapsular lesions that do not show a substantial mass effect or round nature should be evaluated carefully before a diagnosis of HCC is concluded. The combination of hyperdensity on arterial-phase images combined with wash-out to hypodensity on venous-phase or delayed-phase images, although not sensitive (33%), is a very specific (100%) feature for the presence of HCC [36] (Fig. 7). Diffusely infiltrating and small HCCs in

cirrhosis may be difficult to detect with CT. Larger HCC lesions typically have a different appearance and are visualized as a mosaic, due to hemorrhage and fibrosis. Also, about 10% of small HCCs can appear hypodense to liver; these are generally thought to be well-differentiated lesions. Typical MR imaging findings of larger HCC consist of a fibrous capsule, central scar, intratumoral septa, daughter nodules, and tumor thrombus [37]. In addition, there is often a somewhat organized internal, mosaiclike pattern that is seen on CT as well as on MR imaging [38]. While most large HCCs are hyperintense on T2-weighted sequences, small lesions (<2 cm) are often isointense but may also be hypointense. On T1-weighted sequences, HCC has variable signal intensity relative to hepatic parenchyma. A tumor capsule may be seen on T1-weighted and, less commonly, on T2-weighted images as hypointense (Fig. 9). Conventional gadolinium contrast imaging in HCC parallels that described for CT, with characteristic early peak contrast enhancement and delayed-phase tumor wash-out. These enhancement features are useful in differentiating HCC from hemangioma, which generally shows early peripheral enhancement, marked peak enhancement >2 min after contrast injection, and marked pooling of contrast on delayed images. Dynamic gadolinium-enhanced MR imaging has been found to be superior to MR with liverspecific contrast agent in terms of HCC detection [39, 40], because hypervascularity is the key feature marking the transition of dysplastic nodules into early HCC [41]. However, double-contrast MR imaging with sequential administration of gadolinium and reticuloendothelial contrast agents has been found to be the most sensitive and specific method to detect HCC and to differentiate between early HCC and dysplastic nodules [42, 43]. HCC may show enhancement on delayed images after the administration of hepatobiliary MR contrast agent (Fig. 9). However, such enhancement is not specific for HCC and can be seen with other primary hepatocellular tumors, such as dysplastic nodules, FNH, and adenoma.

70

a

Wolfgang Schima, Richard Baron

b

c

d

Fig. 9 a-d. Hepatocellular carcinoma. MR imaging with mangafodipir and gadolinium. a Axial T1-weighted GRE shows an isointense mass with a pseudocapsule and a small hyperintense hemorrhage. b The lesion is only minimally hyperintense on T2-weighted images. c The mangafodipir-enhanced T1-weighted GRE image shows enhancement of this welldifferentiated HCC. The enhancement does not help in the differential diagnosis of hepatocellular lesions. d The gadolinium-enhanced sequence in the arterial phase shows typical hypervascularity. The key features guiding the correct diagnosis in this hypervascular lesion are the presence of a pseudocapsule (very rare in FNH or adenoma) and cirrhosis (see enlarged left and caudate lobes)

Fibrolamellar HCC Fibrolamellar HCC (FL-HCC) is a less aggressive tumor with a better prognosis than HCC. It consists of malignant hepatocytes separated into cords by fibrous strands. On CT, FL-HCC appears as a large, welldefined vascular mass with a lobulated surface and often a central scar; calcifications are identified in up to 70% of patients [44, 45]. On MR imaging, FL-HCC is hypointense on T1-weighted images and hyperintense on T2-weighted images, with the central scar being hypointense on both sequences (Fig. 10). This is in contrast to the scar in FNH, which is most often hyperintense on T2-weighted images. The fibrous central zones of both FNH and FL-HCC will show delayed

a

retention of conventional CT and MR contrast agents. Contrast enhancement patterns in FL-HCC are almost always heterogeneous whereas in FNH they are almost always homogeneous.

Cholangiocellular Carcinoma Cholangiocellular carcinoma (CCC) is the second most common primary malignancy of the liver. Intrahepatic CCC originates from the intralobular bile ducts, unlike hilar CCC, which arises from a main hepatic duct or from the bifurcation. Intrahepatic CCC presents as a large mass, because the tumor does not cause symptoms in the early stages [46] (Fig. 11). According to its growth characteristics, CCC is classified as mass-forming, periductal-

b

Fig. 10 a, b. Fibrolamellar hepatocellular carcinoma. a CT during arterial phase shows a typical heterogeneously enhancing mass in the left lobe (arrows), with a low-attenuation central fibrous scar containing calcifications (arrowheads). b MR T2-weighted shows a large mass in the left lobe (arrows) with a heterogeneous appearance and mild to moderately increased signal intensity. The fibrous central scar is of very low signal intensity (arrowheads)

71

Focal Liver Lesions

a

b

Fig. 11 a, b. Cholangiocellular carcinoma. Contrast-enhanced CT in the arterial (a) and venous (b) phases demonstrates a large hypovascular mass with some calcifications. Capsular retraction is quite often seen in peripheral CCC

infiltrative, or intraductal-growing, with the mass-forming type being most common in intrahepatic CCC [46]. At CT and MR imaging, the lesions tend to be hypodense on unenhanced CT and hypointense on T1-weighted images, with peripheral enhancement at dynamic contrastenhanced studies [47]. Delayed-phase CT/MR imaging (after 5-15 min) may show enhancement homogeneously or in the center of the lesion due to the rich fibrous stroma, which is suggestive of the diagnosis of CCC [48]. Periductal-infiltrative CCC causes early segmental dilatation of the bile ducts at a stage when the tumor itself may be difficult to discern [47].

Rare Primary Liver Tumors Biliary cystadenoma/cystadenocarcinomas These tumors have a similar appearance and morphology as their mucinous counterparts in the pancreas and are seen predominantly in women. Even when benign, they have a propensity for malignant degeneration, and any such tumor should be considered malignant. These unilocular or multilocular cystic masses have a typical anechoic and hypoechoic US appearance; on CT, their contents have near-water attenuation, with peripheral soft-tissue nodularity and traversing septations (Fig. 12).

Fig. 12. Biliary cystadenoma. T2-weighted MR shows a mass of very high signal intensity (arrow) that is mostly homogeneous in appearance, with the exception of a few thin internal septations (arrowheads)

The greater the presence of papillary excrescences, softtissue nodularity, or septations, the more likely it is that the lesion is malignant [49]. However, this is a moot point as it has been shown that benign lesions can undergo malignant degeneration. The cystic areas at T1-weighted imaging are of variable signal intensity, including hyperintense to liver, presumably due to proteinaceous content. Coarse calcifications can be seen at US and CT in both cystadenoma and cystadenocarcinoma and therefore are not a helpful differentiating feature. Hepatic Angiosarcoma This rare tumor has a strong association with carcinogens such as vinyl chloride and Thorotrast and is also seen in patients with hemochromatosis. However, the majority of hepatic angiosarcoma patients have no known exposure to toxic agents. Pathologically, angiosarcoma may appear as a large solitary mass or with multiple tumor nodules of varying size. In addition, the tumors may contain vascular channels that create sinusoidal spaces; these can result in imaging findings in some ways simulating those of hemangiomas. The imaging appearances of angiosarcoma are most often nonspecific, with hypoattenuation on unenhanced CT, hypointensity on T1 MR, and mild hyperintensity on T2 (although if prominent sinusoidal vascular spaces are present pathologically, these can be of homogeneous very high T2 signal). Following iodinated or gadoliniumbased contrast administration, most lesions show nonspecific heterogeneous enhancement. Potentially problematic, however, are those tumors with prominent sinusoidal vascular spaces, in that, albeit rarely, the CT or MR contrast enhancement characteristics of some angiosarcomas can simulate those of benign hemangioma. The high MR T2 signal in such lesions further compounds this problem. In most such cases, however, careful observation will reveal that tumoral enhancement does not follow the characteristics of blood pooling at all phases or that there are other features, such as innumerable lesions, that make the diagnosis of hemangioma unlikely [50, 51].

72

Wolfgang Schima, Richard Baron

Epithelioid Hemangioendothelioma

Hepatic Lymphoma

An EHE is a rare tumor of vascular origin and is not to be confused with infantile hemangioendothelioma, which is a very different tumor. EHE is a primary liver tumor characterized by the presence of multiple, peripheral-based lesions that progressively become confluent masses (Fig. 13). In addition to its unusual peripheral liver distribution, a key and characteristic feature of EHE is the presence of overlying capsular retraction, attributed to fibrosis and scarring within the tumor [52]. The CT attenuation or MR signal intensity characteristics are non-specific and mimic those of other tumors, although tumoral calcifications may occasionally be seen. Contrast enhancement with CT or MR gadolinium chelates often shows a central zone of decreased enhancement with marked enhancement peripherally and, in some cases, concentric zones of marked enhancement (Fig. 13). The lesions often become confluent and may grow large enough to replace nearly the entire liver parenchyma.

This tumor is most often seen in patients with widespread non-Hodgkin’s lymphoma or, rarely, in those with Hodgkin’s disease. Although unusual, hepatic lymphoma constitutes a primary liver tumor; it is usually associated with an immunocompromised state, such as in AIDS or post-transplantation with immunosuppression therapy. The imaging appearances of these lesions are variable, without any unique characteristics. CT, MR, or US imaging shows focal lesions with an appearance similar to that of many other neoplastic lesions. Diffuse infiltrating forms can be difficult to detect regardless of the imaging modality, although hepatomegaly may be present.

Fig. 13. Epithelioid hemangioendothelioma. Contrast CT (portalvenous phase) shows multiple, predominantly peripherally based hypodense lesions, some of which have a laminated appearance (arrows). Early development of capsular retraction is present, with flattening of the capsule overlying some of the lesions (arrowheads)

a

Hepatic Metastases On US, metastases may appear hypoechoic, isoechoic, or hyperechoic. Dynamic contrast-enhanced CT visualizes most metastases as hypovascular and hypodense relative to liver parenchyma on portal-venous phase (Fig. 14). Hypervascular metastases are most commonly seen in patients with renal cell or neuroendocrine tumors, sarcomas, or breast tumors (Fig. 14). These tumors are best visualized in the arterial phase and may become isodense and difficult to detect during the redistribution phase of enhancement. At MR imaging, metastases are usually hypointense on T1-weighted images and hyperintense on T2weighted images [53] (Fig. 15). Peritumoral edema makes the lesions appear larger on the T2, a finding that is very suggestive of a malignant mass [54]. High signal intensity on T1-weighted sequences is typical for melanoma metastases due to the paramagnetic nature of melanin. Some lesions may have a central area of hyperintensity (target sign) on T2-weighted images, which corresponds to central necrosis. On dynamic contrast-enhanced MR imaging, metastases demonstrate enhancement characteristics similar to those described for these tumors on CT. Metastases may demonstrate a hypointense rim compared with the center of the lesion on delayed images (peripheral wash-out sign), which is very specific for malignancy.

b

Fig. 14 a, b. a Contrastenhanced MDCT in the arterial phase demonstrates several predominantly hypervascular liver metastases (arrows) of a neuroendocrine cancer of the pancreas. b Contrast-enhanced MDCT in the venous phase shows typical hypovascular colorectal metastases

73

Focal Liver Lesions

a

b

c

Fig. 15 a-c. Colorectal liver metastases at gadoxetate-enhanced MR imaging. a Unenhanced T1-weighted MR imaging shows two hypointense lesions in segments 6/7 and 4. b The T2-weighted TSE pulse sequence shows the lesions to be moderately hyperintense. c The gadoxetateenhanced T1-weighted GRE image in the hepatobiliary phase shows two additional small subcapsular metastases (arrows) that were not seen on unenhanced MR imaging or MDCT

The most recent studies have shown MR imaging to be more sensitive than contrast-enhanced CT for the detection of hepatic metastases [10, 12, 13], and especially for the detection of small lesions (Fig. 15).

Differential Diagnosis of Focal Liver Lesions The approach to characterizing a focal liver lesion seen on CT begins with a determination of its density. If the lesion is of near-water density, homogeneous in character, and has sharp margins, then a cyst should be considered and can be confirmed with an equilibrium-phase CT, or US or even MR imaging (T2 bright and non-enhancing post-gadolinium). If the lesion has some enhancement, then the aim of the next analysis is to determine whether the enhancement is peripheral and nodular, with the density of the enhancing portions of the lesions following the same general levels of blood vessels in the arterial, venous, and delayed phases. In this case, a hemangioma may be diagnosed with high certainty. Arterially enhancing lesions include FNH, HCA, HCC, and metastases from neuroendocrine tumors, melanoma, renal cell carcinoma, and breast cancer. In general, HCC is considered in the setting of cirrhosis. FNH is most likely in a young woman with a non-cirrhotic liver, if the lesion is homogeneous and near-isodense/isointense on unenhanced CT/MR imaging, with a central T2-weighted hyperintense scar. Thick irregular, heterogeneous enhancement or the presence of peripheral wash-out at delayed phase suggests a malignant mass, such as metastases, CCC, or even HCC.

References 1. Laghi A (2007) Multidetector CT (64 Slices) of the liver: examination techniques. Eur Radiol 17:675-683 2. Weg N, Scheer MR, Gabor MP (1998) Liver lesions: improved detection with dual-detector-array CT and routine 2.5-mm thin collimation. Radiology 209:417-426

3. Ichikawa T, Nakajima H, Nanbu A et al (2006) Effect of injection rate of contrast material on CT of hepatocellular carcinoma. AJR Am J Roentgenol 186:1413-1418 4. Foley WD, Hoffmann RG, Quiroz FA et al (1994) Hepatic helical CT: contrast material injection protocol. Radiology 192:367-371 5. Kim T, Murakami T, Takahashi S et al (1998) Effects of injection rates of contrast material on arterial phase hepatic CT. AJR Am J Roentgenol 171:429-432 6. Schima W, Hammerstingl R, Catalano C et al (2006) Quadruple-phase MDCT of the liver in patients with suspected hepatocellular carcinoma: effect of contrast material flow rate. AJR Am J Roentgenol 186:1571-1579 7. Sultana S, Awai K, Nakayama Y et al (2007) Hypervascular hepatocellular carcinomas: bolus tracking with a 40-detector CT scanner to time arterial phase imaging. Radiology 243:140-147 8. Oliver JH, Baron RL (1996) Helical biphasic contrast-enhanced CT of the liver: technique, indications, interpretations, and pitfalls. Radiology 201:1-14 9. Mehnert F, Pereira PL, Trubenbach J et al (2001) Biphasic spiral CT of the liver: automatic bolus tracking or time delay? Eur Radiol 11:427-431 10. Oudkerk M, Torres CG, Song B et al (2002) Characterization of liver lesions with mangafodipir trisodium-enhanced MR imaging: multicenter study comparing MR and dual-phase spiral CT. Radiology 223:517-524 11. Scharitzer M, Schima W, Schober E et al (2005) Characterization of hepatocellular tumors: value of mangafodipirenhanced magnetic resonance imaging. J Comput Assist Tomogr 29:181-190 12. Ward J, Robinson PJ, Guthrie JA et al (2005) Liver metastases in candidates for hepatic resection: comparison of helical CT and gadolinium- and SPIO-enhanced imaging. Radiology 237:170-180 13. Hammerstingl R, Huppertz A, Breuer J et al (2008) Diagnostic efficacy of gadoxetic acid (Primovist)-enhanced MRI and spiral CT for a therapeutic strategy: comparison with intraoperative and histopathologic findings in focal liver lesions. Eur Radiol 18:457-467 14. Schima W, Saini S, Echeverri JA et al (1997) T2-weighted MR imaging for characterization of focal liver lesions: conventional spin-echo vs fast spin-echo. Radiology 202:389-393 15. Farraher SW, Jara H, Chang KJ et al (2006) Differentiation of hepatocellular carcinoma and hepatic metastasis from cysts and hemangiomas with calculated T2 relaxation times and the T1/T2 relaxation times ratio. J Magn Reson Imaging 24:1333-1341 16. Semelka RC, Brown ED, Ascher SM et al (1994) Hepatic hemangiomas: a multi-institutional study of appearance on T2weighted and serial gadolinium-enhanced gradient-echo MR images. Radiology 192:401-406

74

17. Kim T, Federle MP, Baron RL et al (2001) Discrimination of small hepatic hemangiomas from hypervascular malignant tumors smaller than 3 cm with three-phase helical CT. Radiology 219:699-706 18. Ba-Ssalamah A, Uffmann M, Saini S et al (2009) Clinical value of MRI liver-specific contrast agents: a tailored examination for a confident non-invasive diagnosis of focal liver lesions. Eur Radiol 19:342-357 19. Grangier C, Tourniaire J, Mentha G et al (1994) Enhancement of liver hemangiomas on T1-weighted MR SE images by superparamagnetic iron oxide particles. J Comput Assist Tomogr 18:888-896 20. Vossen JA, Buijs M, Liapi E et al (2008) Receiver operating characteristic analysis of diffusion-weighted magnetic resonance imaging in differentiating hepatic hemangioma from other hypervascular liver lesions. J Comput Assist Tomogr 32:750-756 21. Kehagias D, Moulopoulos L, Antoniou A et al (2001) Focal nodular hyperplasia: imaging findings. Eur Radiol 11:202-212 22. Brancatelli G, Federle MP, Grazioli L et al (2001) Focal nodular hyperplasia: CT findings with emphasis on multiphasic helical CT in 78 patients. Radiology 219:61-68 23. Uggowitzer MM, Kugler C, Mischinger HJ et al (1999) Echoenhanced Doppler sonography of focal nodular hyperplasia of the liver. J Ultrasound Med 18:445-451; quiz 453-444 24. Ba-Ssalamah A, Schima W, Schmook MT et al (2002) Atypical focal nodular hyperplasia of the liver: imaging features of nonspecific and liver-specific MR contrast agents. AJR Am J Roentgenol 179:1447-1456 25. Leconte I, Van Beers BE, Lacrosse M et al (2000) Focal nodular hyperplasia: natural course observed with CT and MRI. J Comput Assist Tomogr 24:61-66 26. Mathieu D, Kobeiter H, Maison P et al (2000) Oral contraceptive use and focal nodular hyperplasia of the liver. Gastroenterology 118:560-564 27. Prasad SR, Sahani DV, Mino-Kenudson M et al (2008) Benign hepatic neoplasms: an update on cross-sectional imaging spectrum. J Computer Assist Tomogr 32:829-840 28. Grazioli L, Morana G, Kirchin MA, Schneider G (2005) Accurate differentiation of focal nodular hyperplasia from hepatic adenoma at gadobenate dimeglumine-enhanced MR imaging: prospective study. Radiology 236:166-177 29. Jeffrey RB, Jr, Tolentino CS, Chang FC, Federle MP (1988) CT of small pyogenic hepatic abscesses: the cluster sign. AJR Am J Roentgenol 151:487-489 30. Barreda R, Ros PR (1992) Diagnostic imaging of liver abscess. Crit Rev Diagn Imaging 33:29-58 31. Laghi A, Iannaccone R, Rossi P et al (2003) Hepatocellular carcinoma: detection with triple-phase multi-detector row helical CT in patients with chronic hepatitis. Radiology 226:543-549 32. Ichikawa T, Kitamura T, Nakajima H et al (2002) Hypervascular hepatocellular carcinoma: can double arterial phase imaging with multidetector CT improve tumor depiction in the cirrhotic liver? AJR Am J Roentgenol 179:751-758 33. Monzawa S, Ichikawa T, Nakajima H et al (2007) Dynamic CT for detecting small hepatocellular carcinoma: usefulness of delayed phase imaging. AJR Am J Roentgenol 188:147-153 34. Iannacone R, Laghi A, Catalano C et al (2005) Hepatocellular carcinoma: Role of unenhanced and delayed-phase multi-detector row helical CT in patients with cirrhosis. Radiology 234:460-467 35. Baron RL, Brancatelli G (2004) Computed tomographic imaging of hepatocellular carcinoma. Gastroenterology 127:S133-143 36. Forner A, Vilana R, Ayuso C et al (2008) Diagnosis of hepatic nodules 20 mm or smaller in cirrhosis: Prospective valida-

Wolfgang Schima, Richard Baron

37. 38. 39. 40.

41.

42. 43.

44. 45. 46. 47. 48. 49.

50.

51. 52. 53. 54.

tion of the noninvasive diagnostic criteria for hepatocellular carcinoma. Hepatology (Baltimore, Md) 47:97-104 Tublin ME, Dodd GD, Baron RL (1997) Benign and malignant portal vein thrombosis: differentiation by CT characteristics. AJR Am J Roentgenol 168:719-723 Stevens WR, Gulino SP, Batts KP et al (1996) Mosaic pattern of hepatocellular carcinoma: histologic basis for a characteristic CT appearance. J Comput Assist Tomogr 20:337-342 Pauleit D, Textor J, Bachmann R et al (2001) Hepatocellular carcinoma: detection with gadolinium- and ferumoxidesenhanced MR imaging of the liver. Radiology 222:73-80 Tang Y, Yamashita Y, Arakawa A et al (1999) Detection of hepatocellular carcinoma arising in cirrhotic livers: Comparison of gadolinium- and ferumoxides-enhanced MR imaging. AJR Am J Roentgenol 172:1547-1554 Kitao A, Zen Y, Matsui O et al (2009) Hepatocarcinogenesis: multistep changes of drainage vessels at CT during arterial portography and hepatic arteriography-radiologic-pathologic correlation. Radiology 252:605-614 Ward J, Guthrie JA, Scott DJ et al (2000) Hepatocellular carcinoma in the cirrhotic liver: double-contrast MR imaging for diagnosis. Radiology 216:154-162 Lee DH, Kim SH, Lee JM et al (2009) Diagnostic performance of multidetector row computed tomography, superparamagnetic iron oxide-enhanced magnetic resonance imaging, and dualcontrast magnetic resonance imaging in predicting the appropriateness of a transplant recipient based on milan criteria: correlation with histopathological findings. Invest Radiol 44:311-321 Ichikawa T, Federle MP, Grazioli L et al (1999) Fibrolamellar hepatocellular carcinoma: imaging and pathologic findings in 31 recent cases. Radiology 213:352-361 Ichikawa T, Federle MP, Grazioli L, Marsh W (2000) Fibrolamellar hepatocellular carcinoma: pre- and posttherapy evaluation with CT and MR imaging. Radiology 217:145-151 Lim JH (2003) Cholangiocarcinoma: morphologic classification according to growth pattern and imaging findings. AJR Am J Roentgenol 181:819-827 Han JK, Choi BI, Kim AY et al (2002) Cholangiocarcinoma: pictorial essay of CT and cholangiographic findings. Radiographics 22:173-187 Lee WJ, Lim HK, Jang KM et al (2001) Radiologic spectrum of cholangiocarcinoma: emphasis on unusual manifestations and differential diagnoses. Radiographics 21 Spec No:S97-S116 Buetow PC, Buck JL, Pantongrag-Brown L et al (1995) Biliary cystadenoma and cystadenocarcinoma: clinical-imagingpathologic correlations with emphasis on the importance of ovarian stroma. Radiology 196:805-810 Peterson MS, Baron RL, Rankin SC (2000) Hepatic angiosarcoma: findings on multiphasic contrast-enhanced helical CT do not mimic hepatic hemangioma. AJR Am J Roentgenol 175:165-170 Koyama T, Fletcher JG, Johnson CD et al (2002) Primary hepatic angiosarcoma: findings at CT and MR imaging. Radiology 222:667-673 Miller WJ, Dodd GD 3rd, Federle MP, Baron RL (1992) Epithelioid hemangioendothelioma of the liver: imaging findings with pathologic correlation. AJR Am J Roentgenol 159:53-57 Schima W, Kulinna C, Langenberger H, Ba-Ssalamah A (2005) Liver metastases of colorectal cancer: US, CT or MR? Cancer Imaging 5 Spec No A:S149-156 Lee MJ, Saini S, Compton CC, Malt RA (1991) MR demonstration of edema adjacent to a liver metastasis: pathologic correlation. AJR Am J Roentgenol 157:499-501

IDKD 2010-2013

Imaging Diseases of the Gallbladder and Bile Ducts Angela D. Levy1, Celso Matos2 1 Department 2 Department

of Radiology, Georgetown University Hospital, Washington, DC, USA of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium

Introduction Patients with gallbladder and biliary disease may present with complaints of right upper quadrant or mid-epigastric pain, fever, jaundice, pruritis, nausea, vomiting, or they may be asymptomatic, with only laboratory abnormalities. Ultrasound, multidetector computed tomography (MDCT), and magnetic resonance imaging (MRI) are used for the non-invasive evaluation of patients with signs and symptoms of gallbladder and biliary disease. New MRI techniques, such as functional magnetic resonance cholangiography (MRC), have improved visualization of the bile ducts. In many instances, non-invasive imaging will establish the diagnosis prior to endoscopic, radiological, surgical intervention. The differential diagnosis for these patients is broad and includes infectious, non-infectious inflammatory, neoplastic, and congenital disorders of the liver, gallbladder, and bile ducts. This chapter discusses functional MRI of the bile ducts, including the use of functional MRC, and the patterned approach to the differential diagnosis of gallbladder and bile-duct disease.

Functional Magnetic Resonance Imaging of the Bile Ducts The bile ducts are generally investigated at MRI by using a heavily T2-weighted sequence. This approach is completely non-invasive and is extremely sensitive for the diagnosis of anatomical abnormalities of the biliary tree. However, the assessment of biliary drainage remains challenging with T2-weighted imaging. In T2weighted MRC, dynamic views of the Vaterian sphincter complex can be obtained in addition to functional information when cholecystokinin stimulation is performed. Unfortunately, cholecystokinin is associated with important side effects (nausea, vomiting) and therefore is not widely used. An alternative approach that can be used to obtain functional information regarding the bile ducts relies on the intravenous administration of paramagnetic contrast agents. These compounds are taken up by the hepatocytes and then transported to the bile, where they cause T1 shortening as a result of their paramagnetic

properties. The functional capabilities of MRI such as those conferred by contrast-enhanced MRC have expanded the clinical applications of magnetic resonance cholangiography.

Technique, Indications, Results T2-weighted imaging The use of T2-weighted imaging on functional MRC allows evaluation of the Vaterian sphincter. These types of studies require a sequence with high temporal resolution, as is the case with single-shot single-slice turbo spin echo (TSE) T2-weighted sequences, which have an acquisition time of about 2-3 s. Serial MRC images (we usually obtain 20 consecutive dynamic views) are optimally obtained in the coronal plane to evaluate dynamic changes of the sphincteric segment of the pancreaticobiliary junction. This technique, also called kinematic magnetic resonance cholangiopancreatography (MRCP), allows normal physiological sphincter contractions to be distinguished from biliary stenosis in patients presenting with bile duct dilatation without an evident cause of obstruction. In the study of Kim et al., failure to visualize sphincteric relaxation had a sensitivity of 88% and a specificity of 100% in the diagnosis of periampullary obstruction. Schmidt et al. reported high accuracy (similar to that of endoscopic ultrasound) of the same technique for the detection of choledocholithiasis. Kinematic MRCP may also be used after the intravenous administration of cholecystokinin. This hormone, which is formed and secreted by the APUD cells of the duodenum, induces bile flow by contracting the gallbladder and imposing an inhibitory effect on the sphincter of Oddi. Therefore, a transient or persistent biliary dilatation may be observed when dilatation of the bile-duct without an evident cause is evaluated. Contrast-Enhanced Functional MRC This approach takes advantage of the biliary excretion of paramagnetic contrast agents, allowing images to be obtained with a high-resolution 3D gradient-recalled echo

76

Angela D. Levy, Celso Matos

(GRE) T1-weighted sequence. Three different hepatobiliary contrast agents are available for functional MRC: mangafodipir-trisodium, gadobenate dimeglumine, and gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA). Proportions of hepatobiliary excretion are higher for mangafodipir-trisodium and for GdEOB-DTPA (about 50%) than for gadobenate dimeglumine. Unlike mangafodipir-trisodium, gadobenate dimeglumine and Gd-EOB-DTPA allow evaluation of liver parenchyma and angiographic renderings in the same examination. For both reasons (high proportion of biliary excretion and vascular assessment), we currently use GdEOB-DTPA. It should be emphasized that the degree of biliary excretion of the contrast agent is dependent on liver function. Tschirch et al. reported decreased visualization or non-visualization of the biliary tree in a substantial percentage of patients with liver cirrhosis. In patients with normal liver function, good visualization of the biliary tree is observed 20-30 min after intravenous administration of 0.025 mmol Gd-EOB-DTPA/kg body weight. Due to the fact that contrast injection indications are generally based on the T2-weighted MRCP findings and because of possible signal extinction in the bile duct on T2weighted imaging and lack of visualization of the pancreatic duct, contrast-enhanced functional MRC is always obtained after T2-weighted MRCP has been acquired.

a

Therefore, after T2-weighted MRCP, high-resolution fatsuppressed 3D GRE T1-weighted sequences are obtained without contrast and then again after intravenous administration of the contrast agent. First, dynamic axial views are acquired, to evaluate the vessels and the liver parenchyma in the arterial and portal-venous phases. Subsequent scans are acquired in the coronal plane until visualization of the contrast agent in the gallbladder and duodenum is noted. If additional delayed imaging is needed, the transit time of the contrast agent into the biliary tree and gallbladder may be determined; this estimate is valuable in the diagnosis of partial obstruction of the bile duct. Contrast-enhanced functional MRC increases the diagnostic performance of conventional MRC in detecting bile-duct leaks (Fig. 1), in assessing the patency of biliary-enteric anastomoses (Fig. 2), and in diagnosing biliary complications of liver transplantation (Fig. 3). Bridges et al. reported that, in the evaluation of liver transplants, delayed excretion (>45 min) was associated with high-grade strictures. Moreover, neither the type of anastomosis nor the presence of edema or ascites had any influence on contrast-enhanced fMRC. When no excretion is observed in the absence of biliary dilatation, transplant rejection should be considered as a possible cause. Other potential clinical scenarios include sphincter of Oddi dysfunction and impaired patency of biliary stents.

b

Fig. 1 a, b. Bile duct leakage after cholecystectomy. a T2-weighted MRCP shows a fluidfilled collection adjacent to the extrahepatic bile duct (arrows). b Contrast-enhanced functional MRC confirms the leakage (arrows)

a

b

Fig. 2 a, b. Biliary-enteric anastomoses: assessing patency. a T2-weighted MRCP shows dilatation of the intrahepatic bile duct (arrow). The anastomoses is not depicted. b Contrast-enhanced functional MRC shows the anastomoses and progressive filling of the jejunal loop (arrow)

77

Imaging Diseases of the Gallbladder and Bile Ducts

a

b

Fig. 3 a, b. Liver transplantation: assessment of biliary anastomoses. a T2-weighted MRCP shows dilatation of the intrahepatic bile ducts upstream to the biliary-biliary anastomoses (arrow). b Contrast-enhanced functional MRC shows the patency of the anastomoses (arrow)

Conclusion Major indications of functional MRC concern the assessment of complications after bile-duct surgery (diagnosis of bile duct leakage and determination of the patency of biliary-enteric anastomoses). Contrast administration should be considered after the biliary morphology has been evaluated on T2-weighted MRC.

Patterned Approach to the Diagnosis of Gallbladder and Bile-Duct Diseases Gallbladder Wall Thickening Focal or diffuse thickening of the gallbladder wall may be caused by acute or chronic cholecystitis or non-inflammatory conditions such as heart failure, cirrhosis, hepatitis, hypoalbuminemia, renal failure, and human immunodeficiency virus (HIV) infection. Gallbladder carcinoma may also cause thickening of the gallbladder wall and should be suggested when there are findings of a focal mass, lymphadenopathy, extension of the process to adjacent organs, hepatic metastases, or biliary obstruction at the level of the porta hepatis. Xanthogranulomatous cholecystitis (XGC) is

a Fig. 4 a, b. A 48-year-old man evaluated for jaundice was found to have a benign stricture in the distal common bile duct and adenomyomatous hyperplasia of the gallbladder. a Longitudinal sonogram of the gallbladder shows mural thickening along the anterior gallbladder wall and echogenic foci in the gallbladder wall that produce ring-down reverberation artifact. b Single-shot fast spinecho T2-weighted MR image with fat saturation shows multiple hyperintense foci within the thickened gallbladder wall, consistent with adenomyomatous hyperplasia

an unusual pseudotumoral chronic inflammatory condition of the gallbladder that radiologically simulates gallbladder carcinoma. There is a significant overlap in the CT features of XGC and gallbladder carcinoma. Both entities may demonstrate wall thickening, infiltration of the surrounding fat, hepatic involvement, and lymphadenopathy. Adenomyomatous hyperplasia, also called adenomyomatosis, is a more common tumor-like lesion of the gallbladder that may produce focal, segmental, or diffuse mural thickening. Sonographically, adenomyomatous hyperplasia is characterized by focal or diffuse thickening of the gallbladder wall, with echogenic foci and ring-down artifact emanating from the wall. The echogenic foci represent bile salts, cholesterol crystals, or small stones in Rokitansky-Aschoff sinuses. On MRI, Rokitansky-Aschoff sinuses are best visualized on breath-hold T2-weighted sequences (Fig. 4). Consequently, MRI can be useful to distinguish adenomyomatous hyperplasia from gallbladder carcinoma.

Gallbladder Polypoid Mass Polypoid gallbladder masses are commonly demonstrated on ultrasound as incidental findings when the gallbladder is imaged sonographically. Polyps are estimated to be present in approximately 3% of gallbladders. The differential diagnosis for a gallbladder polyp includes cholesterol polyp,

b

78

Angela D. Levy, Celso Matos

adenoma, focal adenomyomatous hyperplasia, inflammatory polyp, heterotopia, neurofibroma, carcinoma, carcinoid tumor, lymphoma, and metastasis. The majority of gallbladder polyps are benign. Of these, the most common type is a cholesterol polyp, which accounts for ~50% of all polypoid lesions. Cholesterol polyps have no malignant potential. On sonography, they are typically brightly echogenic, round or slightly lobulated masses that do not produce acoustic shadowing. Larger cholesterol polyps are usually less echogenic and may contain an aggregation of echogenic foci. The management of gallbladder polyps is based on the risk of malignancy, which increases for polyps >10 mm in size and in patients over the age of 60. As the incidence of malignancy in polyps >10 mm ranges from 37 to 88%, patients with symptomatic polyps >10 mm are encouraged to undergo cholecystectomy while those with polyps <10 mm should be followed periodically by ultrasound. At sonography, careful attention should be paid to other features that suggest malignancy: thickening or nodularity of the gallbladder wall, evidence of hepatic invasion such as an indistinct margin between the liver and gallbladder, biliary duct dilatation, and peripancreatic hepatoduodenal ligament adenopathy. If there are any features suggestive of malignancy, MDCT or MRI should be considered for further evaluation of the lesion.

These cysts may have normal intrahepatic bile ducts or partial dilatation of the intrahepatic bile ducts in a nonobstructive pattern. It may be more difficult to differentiate a choledochal cyst with mild or fusiform dilatation of the extrahepatic duct from a duct that is dilated secondary to an obstructing lesion. In these cases, MRCP and/or endoscopic retrograde cholangiopancreatography (ERCP) are useful to exclude an obstructing lesion and to evaluate the pancreaticobiliary junction, an anomaly which is commonly observed in patients with choledochal cyst. Occasionally, pancreatic pseudocysts, echinococcal (hydatid) cysts, and cystic biliary neoplasms, such as biliary cystadenoma or biliary cystadenocarcinoma, may occur in or around the porta hepatis, simulating biliary dilatation and a choledochal cyst. The appearance of rim-like calcification and enhancing septations or mural nodules should help in establishing the diagnosis of a biliary cystadenoma or cystadenocarcinoma. Likewise, echinococcal (hydatid) cysts generally have evidence of inner membranes, daughter cysts, or rim-like peripheral calcification. T2-weighted MRI of hydatid cysts may show a fibrous capsule of low signal intensity as well as membranes.

Cystic Dilatation of the Extrahepatic Bile Duct

Similar to dilatation of the extrahepatic duct, mechanical biliary obstruction is the most common cause of intrahepatic bile duct dilatation. Intrahepatic biliary dilatation due to mechanical obstruction is generally tubular and lacks focal stricture formation. Caroli disease, recurrent pyogenic cholangitis, polycystic liver disease, primary sclerosing cholangitis, choledochal cyst, and peribiliary cysts should be included in the differential diagnosis of cystic intrahepatic biliary dilatation. Caroli disease is suggested by focal or diffuse biliary dilatation that is cystic or fusiform in character (Fig. 6). When diffuse involvement is present, the bile ducts converge toward the porta hepatis. Echogenic intraductal sludge or inflammatory debris may be seen, as well as echogenic stones with posterior acoustic shadowing. The most important differential diagnosis for patients with suspected Caroli disease is recurrent pyogenic cholangitis, which is characterized by biliary dilatation with intrahepatic stone formation. The left hepatic lobe is more commonly involved than the right in recurrent pyogenic cholangitis. Polycystic liver disease may also mimic Caroli disease. However, in most cases, the bile ducts in polycystic liver disease are intrinsically normal; only rarely will the cysts communicate with the bile ducts. Although intrahepatic bile duct dilatation is a feature of primary sclerosing cholangitis, the dilatation is typically fusiform and isolated. In primary sclerosing cholangitis, the degree and extent of duct dilatation is not as severe as that in obstructive biliary dilatation, Caroli disease, or recurrent pyogenic cholangitis; instead, fibrosis, stricture formation, and secondary cirrhosis are the major features.

Mechanical biliary obstruction is the most common cause of extrahepatic bile duct dilatation. Upon initial imaging, an obstructive lesion should always be sought when biliary dilatation is present. Once an obstructive lesion is excluded, congenital etiologies of bile duct dilatation should be considered. Choledochal cysts, unlike obstructive dilatation, generally have more focal extrahepatic bile duct dilatation or are typically more expansive than what is usually encountered in mechanical dilatation (Fig. 5).

Fig. 5. Choledochal cyst in a 3-year-old boy evaluated for a palpable right upper quadrant mass. MRCP shows marked extrahepatic bile duct dilatation with minimal dilatation of the central intrahepatic ducts

Cystic Dilatation of Intrahepatic Bile Ducts

79

Imaging Diseases of the Gallbladder and Bile Ducts

a

b

Fig. 6 a, b. Caroli disease in a 23-year-old man who complained of abdominal pain. a Sagittal sonogram of the liver shows multiple cystic spaces in the posterior right lobe. b MRCP shows high signal intensity within these cysts

Choledochal cyst should be considered in the differential diagnosis if there is both intrahepatic and extrahepatic duct dilatation. Generally, in patients with choledochal cyst the extrahepatic dilatation is more severe than the intrahepatic dilatation. Multiple peribiliary cysts in sequence may simulate bile duct dilatation characterized by a beaded or saccular appearance. Since the bile ducts adjacent to peribiliary cysts are normal, correct diagnosis depends upon the visualization of a normal bile duct. Peribiliary cysts are usually associated with hepatic diseases such as cirrhosis, polycystic liver disease, portal hypertension, portal vein obstruction, and metastatic disease.

(in or near the confluence of the right and left hepatic ducts) may be secondary to hilar cholangiocarcinoma (Klatskin tumor) (Fig. 7), inflammation, or vascular impressions. Strictures in the mid-portion of the extrahepatic bile duct are commonly related to diseases of the gallbladder, such as carcinoma, that have invaded the cystic duct and hepatoduodenal ligament or to inflammatory conditions, such as impaction of a stone in the cystic duct (Mirrizi syndrome). Distal extrahepatic strictures may be due to inflammatory or neoplastic diseases of the pancreas, primary carcinomas of the bile duct or ampulla, sphincter of Oddi dysfunction, or, less commonly, infectious papillitis such as seen in AIDS cholangiopathy.

Biliary Stricture

Conclusions Focal narrowing or strictures in the biliary ducts may be secondary to neoplasia, inflammation, trauma (iatrogenic or non-iatrogenic), or mass effect from adjacent processes. The location of the biliary stricture narrows the differential diagnosis. Strictures at the level of the porta hepatis

a

Fig. 7 a, b. Hilar cholangiocarcinoma in a 40year-old man who complained of abdominal pain. Coronal MRCP (a) and percutaneous transhepatic cholangiogram (b) show a hilar stricture and mild intrahepatic biliary dilatation

A patterned approach to differential diagnosis of gallbladder and biliary duct disease is useful and may be applied to the findings identified on all non-invasive imaging techniques.

b

80

Suggested Reading Functional MR imaging of the bile duct Bridges MD, May GR, Harnois DM (2004) Diagnosing biliary complications of orthotopic liver transplantation with mangafodipir trisodium-enhanced MR Cholangiography: comparison with conventional MR Cholangiography. AJR 182:1497-1504 Fayad LM, Holland GA, Bergin D et al (2003) Functional MR cholangiography of the gallbladder and biliary tree with contrast-enhanced magnetic resonance cholangiography. JMRI 18:449-460 Hottat N, Winant C, Metens T et al (2005) MR cholangiography with manganese dipyridoxyl diphosphate in the evaluation of biliary-enteric anastomoses: preliminary experience. AJR 184:1556-1562 Kim JH, Kim MJ, Park SI et al (2002) Using kinematic MRCP to evaluate biliary dilatation. AJR 178:909-914 Schmidt S, Chevalier P, Novellas S et al (2007) Choledocholithiasis: repetitive thick slab single-shot projection MRCP versus endoscopic ultrasonography. Eur Radiol 17:241-250 Tschirch FTC, Struwe A, Petrowsky et al (2008) Contrast-enhanced MR cholangiography with Gd-EOB-DTPA in patients with liver cirrhosis: visualization of the biliary ducts in comparison with patients with normal liver parenchyma. Eur Radiol 18:1577-1586 Van Hoe L, Gryspeerdt S, Vanbeckevoort D et al (1998) Normal Vaterian sphincter complex: evaluation of morphology and contractility with dynamic single-shot MRCP. AJR 170:14971500

Patterned Approach to the Diagnosis of Gallbladder and Bile Duct Diseases Albores-Saavedra, J, Hensen De, Klimsta DS (2000) Tumors of the gallbladder, extrahepatic bile ducts, and ampulla of vater: Atlas of tumor pathology. Fasc 27, ser 3. Armed Forces Institute of Pathology, Washington, DC

Angela D. Levy, Celso Matos

Baron RL, Campbell WL, Dodd GD 3rd (1994) Peribiliary cysts associated with severe liver disease: imaging-pathologic correlation. AJR Am J Roentgenol 162:631-636 Baron RL, Tublin ME, Peterson MS (2002) Imaging the spectrum of biliary tract disease. Radiol Clin North Am 40:1325-1354 Chun KA, Ha HK, Yu ES et al (1997) Xanthogranulomatous cholecystitis: CT features with emphasis on differentiation from gallbladder carcinoma [see comments]. Radiology 203:93-97 Goodman ZD, Ishak K (1981) Xanthogranulomatous cholecystitis. Am J Surg Pathol 5:653-659 Guy F, Cognet F, Dranssart M et al (2002) Caroli’s disease: magnetic resonance imaging features. Eur Radiol 12:2730-2736 Haradome H, Ichikawa T, Sou H et al (2003) The pearl necklace sign: an imaging sign of adenomyomatosis of the gallbladder at MR cholangiopancreatography. Radiology 227:80-88 Ishikawa O, Ohhigashi H, Imaoka S et al (1989) The difference in malignancy between pedunculated and sessile polypoid lesions of the gallbladder. Am J Gastroenterol 84:1386-1390 Levy AD, Murakata LA, Rohrmann Jr CA (2001) Gallbladder carcinoma: radiologic-pathologic correlation. Radiographics 21:295-314 Levy AD, Rohrmann Jr CA (2003) Biliary cystic disease. Curr Probl Diagn Radiol 32(6):233-263 Levy AD, Rohrmann Jr CA, Murakata LA, Lonergan GJ (2002) Caroli’s disease: radiologic spectrum with pathologic correlation. AJR Am J Roentgenol 179:1053-1057 Raghavendra, BN, Subramanyam BR, Balthazar EJ et al (1983) Sonography of adenomyomatosis of the gallbladder: radiologic-pathologic correlation. Radiology 146:747-752 Schulte SJ, Baron RL, Teefey SA et al (1990) CT of the extrahepatic bile ducts: wall thickness and contrast enhancement in normal and abnormal ducts. AJR Am J Roentgenol 154:79-85 Sugiyama M, Atomi Y, Kuroda A et al (1995) Large cholesterol polyps of the gallbladder: diagnosis by means of US and endoscopic US. Radiology 196:493-497 Yoshimitsu K, Honda H, Jimi M et al (1999) MR diagnosis of adenomyomatosis of the gallbladder and differentiation from gallbladder carcinoma: importance of showing RokitanskyAschoff sinuses. AJR Am J Roentgenol 172:1535-1540

IDKD 2010-2013

Diseases of the Pancreas, I: Pancreatitis Thomas Helmberger Institute of Diagnostic and Interventional Radiology and Nuclear Medicine, Klinikum Bogenhausen, Munich, Germany

Introduction The incidence of pancreatic inflammatory diseases in the Western world is 110-240 cases/1 million individuals. In about 80% of these cases, the underlying cause is an undetected gall stone disease or alcohol abuse. The former is more prevalent in women and the latter in men, especially those between 30 and 50 years of age. Nevertheless, there are many other factors that can cause pancreatitis (Table 1). The manifestations of pancreatitis range from mild, self-limiting disease to severe, lethal forms in acute pancreatitis or permanent loss of exocrine and/or endocrine function in chronic pancreatitis [1-5]. Clinically, it can be difficult to differentiate cystic changes of the pancreas from cystic tumors, and chronic pancreatitis (CP) from pancreatic cancer due to similarities in imaging presentation. Nevertheless, the differential diagnosis is of ample

importance because of significant differences in prognosis and management. The differential diagnosis in terms of solid and cystic tumors is discussed in the chapter by Thoeni (Diseases of the Pancreas II: Tumors).

Acute Pancreatitis Acute pancreatitis (AP) is defined as an inflammatory process involving the gland that was normal prior to the attack and is normal again once the derangements that precipitated the attack have been corrected. AP is a common problem, with a rising incidence in the Western world. In the USA, it currently accounts for about 200,000 admissions/year and about 10,000 deaths/year. The disease is most frequently seen in individuals in the fifth to sixth decade of life, and the most common causes are

Table 1. Potential causes of pancreatitis Metabolic

Mechanical/obstructive

Infectious Vascular Other

Alcohol Drugs (e.g., glucocorticosteroids, azathioprine, hydrochlorothiazide, furosemide, sulfonamides, estrogens, pentamidine, didanosine (DDI), valproic acid) Hypertriglyceridemia Hyperalimentation with lipids Hypercalcemia Hyperparathyroidism Gall stones Tumors (carcinoma of the pancreas or common bile duct, tumor in the ampulla and duodenum, metastases) Sphincter of Oddi dysfunction After upper gastrointestinal tract surgery or endoscopy Following abdominal trauma Duodenal obstruction/scar (e.g., after ulcer disease) Mumps Coxsackie virus infection Ascaris infection, with a mechanical component Panarteriitis nodosa After cardiac or pulmonary surgery Severe arteriosclerosis Hereditary, idiopathic Anomalies (pancreas divisum, annulare) Autoimmune Scorpion sting/gila monster bite Unknown (microlithiasis??)

82

cholelithiasis (75%) and alcohol abuse (15%). Other causes are listed in Table 1. With respect to pathophysiology, the most likely factors for the sudden onset of AP are pancreatic hypersecretion, intra- and extravasation of pancreatic secretions, and premature activation of pancreatic enzymes, followed by autodigestion and necrosis of the pancreatic gland and peripancreatic tissues. AP can be divided clinically into mild and severe forms, which are almost paralleled by the pathophysiological findings of interstitial (edematous) and necrotizing forms [6, 7]. Mild AP (~50% of cases) is characterized by mild symptoms and transitory elevation of amylase levels that recover rapidly without complications. In general, the gland may be enlarged due to a moderate

Fig. 1. Computed tomography image of acute biliary pancreatitis. Edematous enlargement of the pancreatic head and an exudation surrounding the duodenum and along the pararenal fascia are seen. The cause was a biliary stone trapped in the ampullary region (arrow)

a

Thomas Helmberger

edema and peripancreatic fluid collections can be present. Nonetheless, in 30% of the cases, no morphological changes can be appreciated. In cholelithiasis, segmental pancreatitis, mainly of the pancreatic head, is seen in up to 20% of cases (Fig. 1). Ultrasound (US) may reveal a normal to mildly enlarged gland with homogeneous (hypoechoic) echogenicity, but sufficient visualization by US is possible only in 60-70% of cases. In contrast-enhanced computed tomography (CT) and, while generally not necessary, magnetic resonance imaging (MRI), the gland is diffusely enlarged and a small amount of fluid is seen outlining the gland. Imaging is needed to rule out other underlying conditions that can be accompanied by hyperamylasemia, such as bowel obstruction, bowel infarction, gangrenous cholecystitis, and perforated ulcers. In mild to moderate progressing forms of AP, the contour of the gland becomes shaggy. The appearance of the parenchyma on CT and MRI may be heterogeneous, and small intraglandular and/or retroperitoneal fluid collections adjacent to the organ may develop (Fig. 2). The severe forms of pancreatitis are determined by a delayed/absent response to conservative therapy or even deterioration under therapy. Mortality at the latter stage can be as high as 100%. Typical findings in severe (necrotizing) AP are varying degrees of parenchymal necrosis accompanied by progressive exudation, superinfection of necrotic tissue, hemorrhage, abscess formation, phlegmon (~inflammatory pannus), and vascular erosion (Fig. 3). Depending on the particular pathomorphological condition, the pancreas and its surroundings present a rather wide spectrum of imaging findings. In severe cases, US imaging is often compromised by overlying gas, peripancreatic exudation, and phlegmonous changes. In necrosis,

b

Fig. 2 a, b. Magnetic resonance imaging of acute pancreatitis after cholecystectomy and reconstruction of the extrahepatic common bile duct. a T1-weighted gradient-echo image shows the slightly dilated side branches of the pancreatic duct (arrow). A small fluid rim on the anterior renal fascia is seen. b The heavily T2-weighted image reveals fluid outlining the pancreatic gland and the slightly dilated pancreatic duct. Note the moderate stenosis (arrow) of the common bile duct after surgical reconstruction

83

Diseases of the Pancreas, I: Pancreatitis

a

b

Fig. 3 a, b. Acute severe pancreatitis. a Contrast-enhanced CT obtained at admission shows a fuzzy contour of the pancreatic gland together with a peripancreatic exudation (arrows) during the venous phase. Note the hypo- and hyperdense hepatic lesions (arrowheads). b Control scan 10 days later revealed an almost normal gland with resorption of the peripancreatic fluid. However, an area with a lack of enhancement, representing focal necrosis (large arrow), was observed. In the liver, one lesion turned out to be a hemangioma (arrow), while the other two lesions were small abscesses

the pancreatic appearance becomes increasingly hypoechoic, without differentiation of vital from necrotic tissue. Therefore, US is generally used for second-line, complementary imaging during patient follow-up in order to detect fluid formations such as pseudocysts. Parenchymal necrosis is best displayed on contrast-enhanced CT during at least the portal-venous phase. Characteristic findings are patchy areas showing a lack of enhancement, (pseudo) fragmentation, and liquid necroses. Additionally, increasing peripancreatic exudations dissecting along retroperitoneal fascia planes into the mesocolon and the small bowel mesentery as well as peripancreatic inflammatory tissue (phlegmon) and infected areas are frequently seen. In <10% of cases, small amounts of intraperitoneal fluid (ascites) are present, whereas large volumes of intraperitoneal fluids are very rare. According to the literature, there is no significant superiority of MRI over CT in the diagnosis of AP and its related complications. The superior tissue resolution and higher sensitivity to slightly edematous or necrotic changes and to hemorrhage or fluid dissection of fat planes favor MRI. However, these advantages are often hampered by the impaired study conditions in severely ill patients that may degrade the image quality. In patients with severe AP who are administered iodinated contrast agents, pancreatic flow can be reduced followed by an increased rate of necrosis and mortality. While this would seem to favor MRI as a staging tool in AP, this potential complication has not been proved for the non-ionic contrast agents that are nowadays used almost exclusively in CT [8-10]. The local inflammatory conditions are often complicated by regional and systemic involvement induced by autodigestion and activation of systemic inflammatory mediators [11-17]. A rapid change in the local pancreatic and overall abdominal situation demands an adequate diagnostic and therapeutic regime to avoid a disastrous

Table 2. CT grading in acute pancreatitis (from [26, 27]) Grade

CT findings

A B

Normal Focal (~20%), diffuse enlargement of the gland, irregular contour, inhomogeneous density Grade B + inflammation of the peripancreatic fat Small, mostly occasional fluid collections or phlegmon Two or more fluid collections, gas within the pancreas or retroperitoneum

C D E

outcome. Several clinical and laboratory scoring systems have been established to stage and predict the clinical course of severe AP (Ranson’s score, APACHE II). However, these scoring systems are mainly dependent on systemic alterations and are quite non-specific, as they fail to address the local condition of the pancreas [18-25]. Balthazar et al. [26, 27] showed that contrast-enhanced CT is the most helpful diagnostic modality to detect complications that may necessitate medical, surgical, or interventional management, and to predict outcome depending on the local condition of the pancreas. The proposed 5-grade scoring system (Tables 2, 3), by estimating the presence and degree of pancreatic and peripancreatic inflammation and fluid accumulation and by detecting the presence and extent of pancreatic necrosis together with estimation of the lack of gland enhancement (<30, 30-50, >50%), can be translated into a CT-severity index (Table 3) that allows estimation of the complications (morbidity) and of mortality (Fig. 4). If >50% of the pancreatic volume is necrotic, morbidity rises to almost 100%. Recently, a modified and simplified CT severity index was proposed by Mortele et al. [28] that more closely correlates with patient outcome measures than is the case with the currently accepted CT severity index (Table 3).

84

Thomas Helmberger

Table 3. “Classic” and modified CT severity index CT severity index

Points

Pancreatic inflammation Normal pancreas Focal or diffuse enlargement of the pancreas Intrinsic pancreatic abnormalities with or without inflammatory changes in peripancreatic fat Single, ill-defined fluid collection or phlegmon Two or more poorly defined collections or presence of gas in or adjacent to the pancreas Pancreatic necrosis None ≤30% >30-50% >50%

100 90 80 70 60 50 40 30 20 10 0

0 1 2 3 4

0 2 4 6

%

Mortality Complications

0-3

4-6

7 - 10

Fig. 4. CT severity index related to the degree of necrosis of the pancreatic parenchyma. Grade A = 0, B = 1, C = 2, D = 3, E = 4; no necrosis = 0, 30% necrosis = 2, <50% necrosis = 4, >50% necrosis = 6 (from [26,27])

About a fifth of the patients without necrotic changes of the pancreatic gland will also develop local complications. Fluid collections are seen in up to 50% of patients with AP. In about half of these patients, the collections will resolve spontaneously within several weeks. In the rest, however, the fluid collections will persist, eventually followed by encapsulation, superinfection (abscess), or pseudocyst formation.

Modified CT severity index Pancreatic inflammation Normal pancreas Intrinsic pancreatic abnormalities with inflammatory changes in peripancreatic fat Pancreatic or peripancreatic fluid collection or peripancreatic fat necrosis Pancreatic necrosis None ≤30% >30% Extrapancreatic complications (one or more of pleural effusion, ascites, vascular complications, parenchymal complications, or gastrointestinal tract involvement)

Points 0 2 4

0 2 4 2

In a septic patient, not-water-like fluid collections and rim enhancement on contrast-enhanced CT or MRI studies should be considered as abscesses until proven otherwise. Gas, a characteristic sign of an infected fluid collection, is detected in only 20% of patients with pancreatic abscesses. Percutaneous aspiration or drain placement is the proper treatment. In contrast to abscess formations, superinfected necrotic areas of the pancreas are much more difficult to treat. Due to the more solid consistency of the infected necrosis, percutaneous drainage therapy is mostly frustrating; however, biopsy is often needed to prove the diagnosis. In most cases, either percutaneous, endoscopyguided necrosectomy, or surgical intervention has to be considered. Pseudoaneurysm formation and hemorrhage may result from the extravasated pancreatic enzymes that cause vascular injury. They are typically late complications that occur after several episodes of severe AP. While pseudoaneurysms are generally easily detected by any kind of imaging modality, retroperitoneal hemorrhage is best depicted by contrast-enhanced CT or unenhanced MRI. Angiography with arterial embolization is the treatment of choice and in general is superior to surgical therapy [29].

Groove Pancreatitis Complications Pseudocysts are fluid collections with a noticeable capsule that typically develop 4-5 weeks after the onset of AP. On US, CT, and MRI, they have a cyst-like appearance, usually without septations. Since large cysts are prone to complications (e.g., rupture, infection, hemorrhage, biliary obstruction, or fistulization to the gastrointestinal (GI) tract), cysts >5-7 cm in diameter should be treated by percutaneous drainage or operative marsupialization.

First described by Becker in 1973, groove pancreatitis is a rare, late complication occurring after several attacks of AP. It is defined as an inflammatory reaction and fluid collection located in the groove between the head of the pancreas, the duodenum, and the common bile duct. The anterior anlage of the pancreas seems to be mainly affected, with duodenal stenosis and/or strictures of the common duct in about 50% of the cases. Therefore, this disease may mimic cancer of the pancreatic head, necessitating surgi-

85

Diseases of the Pancreas, I: Pancreatitis

cal exploration. Dynamic CT and MRI with delayed enhancement of collagen fibrous tissue during the late postequilibrium phase may reveal a potential soft-tissue mass to be fibrosis and thus, in the absence of complications, obviate the need for surgical exploration [30-33].

Autoimmune Pancreatitis Autoimmune pancreatitis (AIP) is a relatively new syndrome of clinical and histological findings that was first described by Yoshida in 1995 [34]. The condition has also been described as lymphoplasmocytic sclerosing pancreatitis with cholangitis, non-alcoholic duct-destructive chronic pancreatitis, and chronic sclerosing pancreatitis. The features of AIP include hypergammaglobulinemia, elevation of serum IgG4, IgG4-containing immune complexes, and a number of other antibodies as antinuclear antibodies, as well as antibodies against lactoferrin, carbonic anhydrase type II, and rheumatoid factors. Histologically, there is fibrosis and a lymphoplasmacytic infiltration of the interlobular ducts. The majority of lymphocytes are CD8+ and CD4+, while B lymphocytes are less frequent. In general, the diagnosis of AIP is established by clinical signs, together with laboratory and morphological findings. An association with other autoimmune diseases, such as Sjögren-syndrome, primary biliary cirrhosis, primary sclerosing cholangitis, Crohn’s disease or ulcerating colitis, systemic lupus erythematosus, and retroperitoneal fibrosis is found in a third of the cases. At imaging, a focal (“mass-forming”) or diffuse (“sausage-like”) enlargement of the pancreas may be present. In contrast-enhanced studies, peripancreatic nodular or rim-like enhancement can be appreciated. Focal AIP of the pancreatic head that involves the pancreatic and distal common bile duct must be differentiated from pancreatic carcinoma, necessitating biopsy proof [35, 36]. In most patients, the symptoms as well as the laboratory and morphological abnormalities appear to respond to steroid treatment [34, 37-42].

Chronic Pancreatitis The hallmarks of chronic pancreatitis (CP) are a continuing (aseptic) inflammation of the gland accompanied by irreversible morphological and functional damage. The most common reasons are chronic alcohol abuse (70%) and cholelithiasis (20%), with rare cases arising from cystic fibrosis or idiopathically. Patients are typically in their 3rd to 4th decade of life and present with a history of epigastric pain (95%), weight loss (95%), and signs of endocrine/ exocrine deficiency (diabetes mellitus 58%, malabsorption syndrome and steatorrhea 80%). Acute exacerbations of CP are accompanied by episodes of pain attacks that may mimic an acute abdomen. With progressive destruction of the gland, CP may be painless after several years. In 1.5-12% of cases it is complicated by pancreatic cancer.

Tumor markers such as CA 19-9 and CA-50 may be elevated transiently and are non-specific. Laboratory tests of secretin-creozyme and secretin-caerulein have a high diagnostic accuracy except in early stages of the disease but are invasive and cumbersome for the patient. However, these tests are of particular importance in the diagnostically challenging, newly defined small-duct CP, in which chronic inflammation occurs without ductal abnormalities. In CP, the most characteristic findings are dilatation of the pancreatic main duct and of the ductal side branches (70-90%), small cystic changes, scattered glandular and ductal calcifications (40-50%), and ductal protein plugs. The grade and shape of the ductal dilatation may help to differentiate chronic (benign) obstructions from malignant occlusions: in CP, the contour of the pancreatic duct and its side branches is commonly irregular (73%) while this is true only in 15% of pancreatic malignancies. Additionally, the duct usually accounts for <50% of the pancreatic anterior-posterior diameter in CP while the opposite is true in pancreatic cancers (due to obstructive atrophy). In some cases, additional secretin-enhanced magnetic resonance cholangiopancreatography (MRCP) can be helpful as it provides an improved display of the duct system and allows assessment of the excretory capacity of the pancreatic gland [43]. In CP, the gland may have a normal appearance in 1520% of patients, but most common is a diffuse (50%) or focal (25%) enlargement that may arouse suspicion of a neoplasm. With time, atrophy of the organ will occur in 10-50% of these patients. The variable appearance of CP explains the shortcomings in establishing the diagnosis. In the absence of gross morphological changes it is very difficult to identify incipient forms of CP. Moreover, morphological changes correlate very poorly with the functional exocrine and endocrine deficits. Consequently, endoscopically guided (endoscopic US) or percutaneous biopsy may be necessary for the diagnosis.

Complications These include consolidated cystic degeneration involving dilatation of the cystic and side branches of the pancreatic duct and peripancreatic pseudocyst formation (Fig. 5). In addition, there may be obstruction of the common bile duct secondary to fibrosis, splenic vein thrombosis (note: upper GI tract bleeding may result from gastric varices in the absence of esophageal varices!), and pancreatic fistulae (communications between the pancreatic duct and abdominal organs or the skin).

Diagnostic Challenge In general, the diagnosis of inflammatory and tumorous pancreatic conditions relies on imaging (US, CT, MRI) and, more invasively, endoscopic retrograde cholangiopancreatography (ERCP). Imaging findings are determined by the macro-structural changes of the organ and

86

a

Thomas Helmberger

b

Fig. 5 a, b. Chronic pancreatitis. Cystic degeneration of the pancreatic head (a) together with irregular dilatation of the pancreatic main duct (b) is seen on MRI (fast spin-echo T2)

its surroundings. On plain films and US, the calcifications in CP are readily depicted. Additionally, US is able to display ductal dilatation, micro- and macrocystic changes, and the gland itself. Contrast-enhanced multidetector CT (MDCT) is well established in the assessment of ductal changes, calcifications, and alterations in the form and shape of the pancreatic gland, as well as potential concomitant conditions such as pseudocysts. In addition, multiplanar, curved reconstructions yield high-resolution display of the entire gland and the anatomical course of the duct. Depending on the fibrotic changes in CP, contrast enhancement can be variable, whereas most ductal carcinomas show no or only minor enhancement during arterial-dominant- and parenchymal-phase imaging. However, late enhancement can be seen on delayed imaging without substantial additional information [44]. In addition to MRI’s superior tissue resolution in the differentiation of varying “qualities” of pancreatic parenchyma, using unenhanced and Gd-DTPA-enhanced T1-weighted (±fat suppression) and heavily T2-weighted sequences, it optimally displays the pancreatic gland, the pancreatic duct including the first-degree side branches, and even small stones. Nevertheless, initial, minor ductal changes are best seen on ERCP (Table 3), which in this respect is superior to all other imaging modalities. However, the clinical significance of these slight changes remains contentious, further compromised by potentially “non-physiological” distention of the ducts due to the injected contrast material. Pancreatic cancer is the most serious complication of CP and is the major diagnostic challenge because the focal enlargement of the gland induced by a fibrotic inflammatory pseudotumor may be indistinguishable from carcinoma. A comparison using state-of-the-art MDCT and MRI showed no difference in the detection rate of pancreatic carcinoma, according to the recent literature. Nevertheless, the potential tumor-like appearance of CP accounts for the fact that it is still the major reason for a

Table 4. Ranson score based on clinical and laboratory signs at admission and at 48-h follow-up (each sign = 1 point) At admission

48-h follow-up

Age >55 years WBC >16,000

Hematocrit decrease >10% Blood urea nitrogen increase >5 mg/dL Ca (serum) <8 mg/dL PO2 <60 mmHg Base deficit >4 meq/L Estimated fluid sequestration >600 mL Mortality (%) <10 10-20 >50

Blood glucose >200 mg/dL Serum LDH >350 IU/L SGOT (AST) >250 U/L

Score 0-2 3-5 >5

missed diagnosis of carcinoma. However, if local or regional lymph node enlargement, vascular encasement, or remote metastases is displayed, the differential is ruled by these secondary signs of malignancy, in which case the tumor must be staged correctly for further treatment stratification (Tables 3, 4). In ambiguous cases, biopsy or even surgical exploration may be necessary. CP can cause a focal pancreatic mass indicative of a neoplasm. Moreover, it represents a major risk factor for pancreatic cancer, with a 26-fold increased risk of developing cancer, according to an international, multicenter cohort study [45]. Therefore, the differential between CP and pancreatic cancer remains challenging and underlines the need for multiple diagnostic approaches. In one study, US, CT, MRI, and positron emission tomography (PET)/CT for pancreatic cancer were shown to have a sensitivity of 76-83% and a specificity of 91-93% [46]. Nevertheless, the rate of incorrect diagnoses is as high as 25%. The use of various differential criteria (Table 5) may help to improve the overall diagnostic accuracy beyond that achieved based solely on image interpretation.

87

Diseases of the Pancreas, I: Pancreatitis

Table 5. Differential criteria for chronic pancreatitis (CP) versus pancreatic cancer (PCa)

History Duct Duct/parenchyma Calcification Enhancement Cysts Lymph nodes Metastases

CP

PCa

+++ Irregular <0.5 +++ Diffuse +++ (+) –

– Smooth >0.5 – Focal (+) ++ +++

Recently, the use of new methods and techniques, such as oxygen insensitivity testing, have been described in conjunction with conventional pathology studies of brush cytology in facilitating discrimination between CP and pancreatic cancer [47, 48]. Nevertheless, to date, biopsy is the most reliable diagnostic tool in ambiguous cases of pancreatic masses, with no significant difference whether the biopsy was performed endoscopically or percutaneously, while laparoscopic procedures are compromised by an increased potential for tumor seeding and adverse events.

Pancreatitis in Children Fortunately, pancreatitis in childhood is rare. The best known causes for pediatric pancreatitis are traumas to the pancreas (typically, a bicycle accident), the genetically transmitted hereditary pancreatitis, and cystic fibrosis. Nevertheless, in most of the cases the reason is unknown (idiopathic pancreatitis), although microcalculi and protein plugs have been suggested. In general, the diagnosis can be established based on the history, clinical findings, and imaging findings, which are not different from those described in adults [49-51].

References 1. Mitchell RM, Byrne MF, Baillie J (2003) Pancreatitis. Lancet 361:1447-1455 2. Glasbrenner B, Kahl S, Malfertheiner P (2002) Modern diagnostics of chronic pancreatitis. Eur J Gastroenterol Hepatol 14:935-941 3. Etemad B, Whitcomb DC (2001) Chronic pancreatitis: diagnosis, classification, and new genetic developments. Gastroenterology 120:682-707 4. Halonen KI, Leppaniemi AK, Puolakkainen PA et al (2000) Severe acute pancreatitis: prognostic factors in 270 consecutive patients. Pancreas 21:266-271 5. Bank S, Indaram A (1999) Causes of acute and recurrent pancreatitis. Clinical considerations and clues to diagnosis. Gastroenterol Clin North Am 28:571-89, viii 6. Cavallini G, Frulloni L, Bassi C et al (2004) Prospective multicentre survey on acute pancreatitis in Italy (ProInf-AISP): results on 1005 patients. Dig Liver Dis 36:205-211 7. Losanoff JE, Asparouhov OK, Jones JW (2001) Multiple factor scoring system for risk assessment of acute pancreatitis. J Surg Res 101:73-78

8. Plock JA, Schmidt J, Anderson SE et al (2005) Contrast-enhanced computed tomography in acute pancreatitis: does contrast medium worsen its course due to impaired microcirculation? Langenbecks Arch Surg 390:156-163 9. Werner J, Schmidt J, Warshaw AL et al (1998) The relative safety of MRI contrast agent in acute necrotizing pancreatitis. Ann Surg 227:105-111 10. Robinson PJ, Sheridan MB (2000) Pancreatitis: computed tomography and magnetic resonance imaging. Eur Radiol 10:401-408 11. Pamuklar E, Semelka RC (2005) MR imaging of the pancreas. Magn Reson Imaging Clin N Am 13:313-330 12. Laurens B, Leroy C, André A et al (2005) [Imaging of acute pancreatitis]. J Radiol 86:733-46; quiz 746-747 13. Vaishali MD, Agarwal AK, Upadhyaya DN et al (2004) Magnetic resonance cholangiopancreatography in obstructive jaundice. J Clin Gastroenterol 38:887-890 14. Arvanitakis M, Delhaye M, De Maertelaere V et al (2004) Computed tomography and magnetic resonance imaging in the assessment of acute pancreatitis. Gastroenterology 126:715-723 15. Akahane T, Kuriyama S, Matsumoto M et al (2003) Pancreatic pleural effusion with a pancreaticopleural fistula diagnosed by magnetic resonance cholangiopancreatography and cured by somatostatin analogue treatment. Abdom Imaging 28:92-95 16. Sica GT, Miller FH, Rodriguez G et al (2002) Magnetic resonance imaging in patients with pancreatitis: evaluation of signal intensity and enhancement changes. J Magn Reson Imaging 15:275-284 17. Okai T, Fujii T, Ida M et al (2002) EUS and ERCP features of nonalcoholic duct-destructive, mass-forming pancreatitis before and after treatment with prednisolone. Abdom Imaging 27:74-76 18. Taylor SL, Morgan DL, Denson KD et al (2005) A comparison of the Ranson, Glasgow, and APACHE II scoring systems to a multiple organ system score in predicting patient outcome in pancreatitis. Am J Surg 189:219-222 19. Mentula P, Kylänpää ML, Kemppainen E et al (2005) Early prediction of organ failure by combined markers in patients with acute pancreatitis. Br J Surg 92:68-75 20. Leung TK, Lee CM, Lin SY et al (2005) Balthazar computed tomography severity index is superior to Ranson criteria and APACHE II scoring system in predicting acute pancreatitis outcome. World J Gastroenterol 11:6049-6052 21. Johnson CD, Toh SK, Campbell MJ (2004) Combination of APACHE-II score and an obesity score (APACHE-O) for the prediction of severe acute pancreatitis. Pancreatology 4:1-6 22. Gerlach H (2004) Risk management in patients with severe acute pancreatitis. Crit Care 8:430-432 23. King NK, Powell JJ, Redhead D, Siriwardena AK (2003) A simplified method for computed tomographic estimation of prognosis in acute pancreatitis. Scand J Gastroenterol 38:433-436 24. Triester SL, Kowdley KV (2002) Prognostic factors in acute pancreatitis. J Clin Gastroenterol 34:167-176 25. Sandberg AA, Borgstrom A (2002) Early prediction of severity in acute pancreatitis. Is this possible? Jop 3:116-125 26. Balthazar EJ (2002) Staging of acute pancreatitis. Radiol Clin North Am 40:1199-1209 27. Balthazar EJ (2002) Acute pancreatitis: assessment of severity with clinical and CT evaluation. Radiology 223:603-513 28. Mortele KJ, Wiesner W, Intriere L et al (2004) A modified CT severity index for evaluating acute pancreatitis: improved correlation with patient outcome. AJR Am J Roentgenol 183: 1261-1265 29. Mofidi R, Patil PV, Suttie SA, Parks RW (2009) Risk assessment in acute pancreatitis. Br J Surg 96:137-150 30. Irie H, Honda H, Kuroiwa T et al (1998) MRI of groove pancreatitis. J Comput Assist Tomogr 22:651-5 31. Gabata T, Kadoya M, Terayama N et al (2003) Groove pancreatic carcinomas: radiological and pathological findings. Eur Radiol 13:1679-1684

88

32. Shanbhogue AK, Fasih N, Surabhi VR et al (2009) A clinical and radiologic review of uncommon types and causes of pancreatitis. Radiographics 29:1003-1026 33. Triantopoulou C, Dervenis C, Giannakou N et al (2009) Groove pancreatitis: a diagnostic challenge. Eur Radiol 19:1736-1743 34. Okazaki K (2005) Autoimmune pancreatitis: etiology, pathogenesis, clinical findings and treatment. The Japanese experience. Jop 6(1 Suppl):89-96 35. Takahashi N, Fletcher JG, Hough DM et al (2009) Autoimmune pancreatitis: differentiation from pancreatic carcinoma and normal pancreas on the basis of enhancement characteristics at dual-phase CT. AJR Am J Roentgenol 193:479-484 36. Weili L, Jiaguo W (2009) Education and imaging: Hepatobiliary and pancreatic: autoimmune pancreatitis. J Gastroenterol Hepatol 24:1574 37. Sahani DV, Kalva SP, Farrell J et al (2004) Autoimmune pancreatitis: imaging features. Radiology 233:345-352 38. Farrell JJ, Garber J, Sahani D, Brugge WR (2004) EUS findings in patients with autoimmune pancreatitis. Gastrointest Endosc 60:927-936 39. Wakabayashi T, Kawaura Y, Satomura Y et al (2003) Clinical and imaging features of autoimmune pancreatitis with focal pancreatic swelling or mass formation: comparison with socalled tumor-forming pancreatitis and pancreatic carcinoma. Am J Gastroenterol 98:2679-2687 40. Ito K, Koike S, Matsunaga N (2001) MR imaging of pancreatic diseases. Eur J Radiol 38:78-93 41. Irie H, Honda H, Baba S et al (1998) Autoimmune pancreatitis: CT and MR characteristics. AJR Am J Roentgenol 170:1323-1327

Thomas Helmberger

42. Sahani DV, Sainani NI, Deshpande V et al (2009) Autoimmune pancreatitis: disease evolution, staging, response assessment, and CT features that predict response to corticosteroid therapy. Radiology 250:118-129 43. Sainani NI, Conwell DL (2009) Secretin-enhanced MRCP: proceed with cautious optimism. Am J Gastroenterol 104:1787-1789 44. Yamada Y, Mori H, Matsumoto S et al (2009) Pancreatic adenocarcinoma versus chronic pancreatitis: differentiation with triple-phase helical CT. Abdom Imaging 22 45. Lowenfels AB, Maisonneuve P, Cavallini G et al (1994) Prognosis of chronic pancreatitis: an international multicenter study. International Pancreatitis Study Group. Am J Gastroenterol 89:1467-1471 46. Tang S, Huang G, Liu J et al (2009) Usefulness of (18)F-FDG PET, combined FDG-PET/CT and EUS in diagnosing primary pancreatic carcinoma: A meta-analysis. Eur J Radiol (in press) 47. Cho SG, Lee DH, Lee KY et al (2005) Differentiation of chronic focal pancreatitis from pancreatic carcinoma by in vivo proton magnetic resonance spectroscopy. J Comput Assist Tomogr 29:163-9 48. van Kouwen MC, Jansen JB, van Goor H et al (2005) FDGPET is able to detect pancreatic carcinoma in chronic pancreatitis. Eur J Nucl Med Mol Imaging 32:399-404 49. Manfredi R, Lucidi V, Gui B et al (2002) Idiopathic chronic pancreatitis in children: MR cholangiopancreatography after secretin administration. Radiology 224:675-682 50. DeBanto JR, Goday PS, Pedroso MR et al (2002) Acute pancreatitis in children. Am J Gastroenterol 97:1726-1731 51. Levy MJ, Geenen JE (2001) Idiopathic acute recurrent pancreatitis. Am J Gastroenterol 96:2540-2555

IDKD 2010-2013

Diseases of the Pancreas, II: Tumors Ruedi F. Thoeni Department of Radiology and Biomedical Imaging, University of California, San Francisco, CA, USA

Introduction In the imaging of pancreatic disease and in assessment of the etiology of jaundice, abdominal ultrasound (US) and computed tomography (CT) traditionally have been employed [1]. These two methods are widely available and have the advantages of their familiarity to radiologists and clinicians and their non-invasiveness. With the introduction of magnetic resonance imaging (MRI), magnetic resonance cholangiopancreatography (MRCP), and endoscopic ultrasound (EUS), visualization of the pancreatic and biliary ducts improved, allowing tumors to be more accurately staged and safely sampled [2-12]. This led to a diminishing role for endoscopic retrograde cholangiography in the diagnostic arena but its therapeutic use has remained unchallenged. In recent years, technological advances with multidetector row CT (MDCT) imaging have improved the ability of CT to detect even small lesions in the pancreas and to stage pancreatic tumors more accurately [13]. Microbubble contrast enhancement and secretin stimulation have increased the diagnostic acumen of US and MRI, respectively, and may widen the utility of these techniques [14, 15]. Nevertheless, MDCT remains the primary tool in assessing patients with suspected pancreatic disease, while EUS and MRI are used as problem-solving modalities to confirm suspected lesions not identified with CT, to find additional lesions, and to obtain a definitive tissue diagnosis with EUS-guided tissue sampling. In recent years, position emission tomography/CT (PET/CT) has been increasingly employed in the assessment of patients with suspected pancreatic tumors but its ultimate role still needs further definition [16-21]. Also, somatostatin receptor scintigraphy has gained popularity in recent years for neuroendocrine tumors [22, 23]. This discussion will focus on diagnosing and staging the various pancreatic neoplasms with CT and MRI, mentioning EUS, PET/CT, and somatostatin receptor scintigraphy where appropriate.

Ductal Adenocarcinoma of the Pancreas About 90% of all neoplasms of the pancreas are ductal adenocarcinomas. Pancreatic adenocarcinoma is one of

the leading causes of cancer death in the western world, and the overall relative 5-year survival rate of only 5.5% is dismal [24]. Late clinical presentation with advanced disease and the aggressiveness of the tumor result in a low rate of surgical intervention and overall poor outcome. It is estimated that 42,470 men and women will be diagnosed with cancer of the pancreas in 2009 and that over 80% of these patients will die of the disease [25]. The tumor serum marker CA 19-9 is sensitive, although not specific for the diagnosis of adenocarcinoma of the pancreas. The treatment approach is based on whether the tumor can or cannot be resected at presentation. Therefore, imaging plays a crucial role in disease management. The initial diagnosis of pancreatic tumor, particularly if the patient presents with jaundice and the tumor is located in the head of the pancreas, may be made by US. The ultrasonographic signs of pancreatic carcinoma include a focal or diffuse pancreatic mass that is hypoechoic relative to normal gland parenchyma and dilatation of the pancreatic duct without or with biliary duct distention (double-duct sign). The accuracy of US for detecting the level of bile duct obstruction varies greatly, and ultrasonographic staging of pancreatic carcinoma is inferior to that of CT. Ultrasonography often fails to provide an adequate examination of the entire gland, resulting in an overall decrease in the sensitivity of this technique. Some of these limitations are overcome by endosonography, but tumors in the tail of the pancreas are not always accessible by EUS. For optimal evaluation of pancreatic neoplasms, MDCT is the modality of choice. A triple-phase protocol is recommended that includes thin sections (0.625 or 1.25 mm) through the abdomen, initially without intravenous contrast followed by a rapidly delivered bolus of contrast material (we use bolus tracking and 150 mL at 5 mL/s chased by 50 mL of saline). It is best to use a neutral oral contrast agent (water or VoLumen, Bracco Diagnostics, USA) because it permits optimal assessment of tumor extension to the stomach and/or duodenum and does not interfere with the determination of vascular invasion. We recommend a scan delay of 40-45 s (10-s delay from peak aortic enhancement) for the late arterial or pancreatic phase and a scan delay of 80 s for the hepatic phase. Rarely, an arterial phase

90

Ruedi F. Thoeni

at 20-25 s is performed if requested by surgery [26]. Arterial involvement and tumor mass are best detected in the pancreatic phase whereas the hepatic phase enables optimal visualization of the liver, veins, and the entire abdomen in the search for liver metastases and peritoneal seeding. One study demonstrated that a single-phase thinslice MDCT technique is sufficient for accurately assessing the resectability of pancreatic adenocarcinoma [27]. Pancreatic adenocarcinoma arises from the pancreatic duct. On MDCT, the tumor usually appears as a low-density mass, often associated with poorly defined margins (Fig. 1) and pancreatic and/or bile duct dilatation. The low-density central zone represents either hypovascular, scirrhous tumor surrounded by normal parenchyma or inflammatory tissue caused by obstructive pancreatitis. Occasionally, cystic degeneration is seen within the tumor [28]. Neoplastic pancreatic duct obstruction frequently produces a dilated duct as well as atrophy of the pancreatic parenchyma proximal to the neoplasm. Tumor obstruction of the main pancreatic duct can lead to rupture of the side branches, resulting in the formation of pseudocysts. Occasionally, a low-density mass cannot be identified because the tumor is isodense to the surrounding normal parenchyma. In these cases, a dilated duct with abrupt cut-off is often seen proximal to a small imperceptible tumor mass. Ancillary findings are local tumor extension, including direct invasion of neighboring organs such as the liver and the stomach, the arteries (loss of fat planes surrounding celiac axis, superior mesenteric artery, etc., vascular “cuffing”) and veins (tear-drop sign, flattening, irregularity of margins, etc., of the portal vein, superior mesenteric vein and its branches), and metastatic disease to local lymph nodes, as well as spread to the

liver, peritoneum (often associated with ascites), and more distant sites. The double-duct sign (dilatation of the biliary and pancreatic ducts) occurs in <5% of patients with pancreatic carcinoma. Biductal obstruction is a nonspecific sign and may also be seen in bile duct or ampullary carcinoma, pancreatitis, and ampullary stenosis. For MRI, T1-weighted fat-suppressed sequences and dynamic gadolinium-enhanced spoiled gradient echo (SPGR) sequences are superior to T2-weighted sequences, as most pancreatic carcinomas have a significant desmoplastic reaction that renders the tumor less conspicuous on T2-weighted images. T1-weighted fatsuppressed images using an early (arterial) gadoliniumenhanced 3D vascular time-of-flight SPGR sequence provide optimal delineation of the tumor, particularly if it is small and does not change the contour of the pancreas. Diffusion-weighted imaging appears promising, especially for metastases to the liver. MRCP sequences consisting of thin and thick axial and coronal sequences with heavy T2-weighting and breath-holding are often added to better assess the pancreatic and biliary ducts. The CT imaging results for pancreatic carcinoma vary widely, but with the current generation of scanners and state-of-the-art scanning techniques a sensitivity of >90% for detecting pancreatic carcinoma can be achieved [1, 27]. Small metastatic implants on the liver and peritoneum are the lesions most likely to be missed by MDCT. MDCT generally provides accurate information on vascular involvement as long as a pancreatic protocol is observed; for resectability, sensitivities of >80% have been obtained [1]. The positive predictive values for predicting unresectability are much better than those for predicting resectability. Presently, most studies show a slight advantage of MDCT

a

b

Fig. 1 a, b. a Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the pancreatic phase (~40 s). A low-attenuation mass is apparent in the head of pancreas near the uncinate process (white arrows), with encasement of the replaced right hepatic artery originating from the superior mesenteric artery (arrowhead). Retroperitoneal lymphadenopathy also is seen (black arrow). The mass is easily distinguished from adjacent normal pancreas. b Thin-section (1.25 mm) axial MDCT of a pancreas carcinoma in the hepatic phase (~80 s) in the same patient. Note the teardrop-shaped superior mesenteric vein (black arrowhead) as a sign of venous encasement. The low-attenuation pancreatic neoplasm (arrows) is less well seen

91

Diseases of the Pancreas, II: Tumors

over MR for detecting and staging pancreatic adenocarcinomas. A meta-analysis comparing CT, MRI, and US demonstrated a sensitivity and specificity of 91 and 85% for helical CT and a sensitivity and specificity of 84 and 82% for MRI whereas the results for resectability were similar [1]. For US, the sensitivity for diagnosing pancreas carcinoma and the specificity for determining resectability are much lower. The advantage of MRI is thought to be in the area of small tumors that do not alter the contour of the gland [10] and in detecting hepatic metastases. At present, MRI appears to be a problem-solving modality. It should be considered in patients with suspected pancreatic neoplasms in the presence of: (1) allergy to iodine contrast or other contraindications to iodine contrast administration, (2) a MDCT scan showing focal enlargement of the pancreas but no definable mass, (3) a clinical history suggesting malignancy and MDCT images that are equivocal or difficult to interpret, and (4) when distinction between chronic pancreatitis with focal enlargement and pancreatic cancer is needed. When choosing an imaging modality, one has to take into account that, today, MDCT of the pancreas requires a small fraction of the time needed for a complete MRI study of the pancreas. False-positive MDCT diagnoses of pancreatic cancer can occur, especially in patients with chronic pancreatitis; therefore, percutaneous aspiration biopsies are needed if non-operative treatment is planned. Fine-needle aspiration biopsy of pancreatic cancer under CT guidance is a frequently performed procedure and is associated with severe pancreatitis in <3%. The sensitivity of percutaneous CT biopsies reaches 79%, with a positive predictive value of 100% and a negative predictive value of 47% [7]. However, because of possible tumor seeding in the needle tract, patients with potentially resectable tumors (only 10% of all cases) who are acceptable candidates for surgery should undergo exploratory surgery [7]. While EUS excels in detecting even small pancreatic adenocarcinomas, reaching sensitivities as high as 97% [9], and can be used in the differential diagnosis of pancreatic tumors [29], it demonstrates poor sensitivity and specificity for diagnosing vascular involvement by the tumor [30]. The technique suffers from limited depth penetration. Today, endoscopic biopsies often replace percutaneous CT biopsies of the pancreas. They have a sensitivity of 80% with a positive predictive value of 99% and a negative predictive value of 73%. They are particularly indicated when CT is equivocal or negative despite a strong clinical suspicion for tumor and when the lesion is <3 cm in size [7]. PET, and particularly PET/CT, has emerged as an important modality for effectively managing patients with suspected pancreatic cancer. Nevertheless, more studies are needed to demonstrate its true value and cost-effectiveness since at least one study found no benefit over CT alone [16]. It was also reported that if helical CT was positive for pancreas carcinoma, PET had a sensitivity of 92% and a specificity of 68%; if CT was negative, the sensitivity of PET was 73% and the specificity 86% [19]. PET/CT allows hot tracer spots to be to more precisely localized

and has been shown to improve patient management before possible resection. In one PET/CT study, management was changed in 16% of patients with pancreatic cancers that were initially staged as being resectable [17]. In suspected tumor recurrence, PET reliably detected local recurrence and was advantageous in diagnosing distant disease [18].

Neuroendocrine Neoplasms of the Pancreas Hyperfunctioning Neuroendocrine Neoplasms Among hyperfunctioning or syndromic neuroendocrine neoplams (NEN, formerly called islet cell tumors) of the pancreas, insulinoma is the most common followed by gastrinoma, glucagonoma, VIPoma, and other rarely encountered secretory neoplasms. In functioning pancreatic adenomas, the clinical diagnosis is based on clinical data and laboratory tests that usually permit an accurate diagnosis, with cross-sectional imaging used only for localizing the pancreatic neoplasm [31]. Insulinomas, and especially extrapancreatic NENs, that are small and located in the duodenal or gastric wall (Fig. 2) may be difficult to detect pre-operatively by any

a

b

Fig. 2 a, b. a Thin-section (1.25 mm) axial MDCT of an ectopic neuroendocrine tumor of the pancreas (insulinoma) in the pancreatic phase. A small hypervascular mass (arrow) is seen at the junction of the second to third duodenum. b Thin-section (1.25 mm) coronal MDCT of an ectopic neuroendocrine tumor of the pancreas (insulinoma) in the pancreatic phase, which clearly demonstrates the mass in the duodenal wall (arrow)

92

of the radiographic techniques, and even intra-operative US fails to detect some of these lesions. However, MDCT with 0.5-0.625 mm sections has improved the results. These ectopic lesions are more likely to occur in patients with multiple endocrine adenomatosis (MEA) or multiple endocrine neoplasia (MEN). A combination of intraoperative palpation and intra-operative US was found to achieve the best results during surgery. Intra-operative US is particular important in patients with multiple lesions and MEN. On MDCT and MRI, functioning NENs generally show intense enhancement in the arterial phase with rapid washout in the portal venous phase. The most common NEN, the insulinoma, usually is small (≤2 cm in diameter) and seldom metastasizes (5-10% of cases). All other NENs tend to be large and frequently have metastases (60-65% of cases). The appearance of liver metastases in patients with functioning NEN is similar to that of the primary tumor. The reported sensitivity of conventional CT for detecting an insulinoma ranges from 28 to 79% with a mean of 38%. It is slightly higher for gastrinomas, due primarily to their larger size. A dual-phase MDCT protocol with thin sections improves the detection rate to 94% and reaches 100% if combined with EUS [32]. The latter modality provides excellent results in the head of the pancreas, but the results are less convincing for the tail of the pancreas because of its distance from the stomach. EUS usually allows the detection of even small NENs and their precise location. Ectopic gastrinomas may be missed by EUS but combining this technique with somatostatin receptor scintigraphy (SSR or Octreoscan) increases the overall sensitivity for gastrinomas [33]. The sensitivity of transabdominal US for detecting insulinomas is low (mean of 46%) and therefore should not be used for this purpose. Functioning NENs of the pancreas are of low signal intensity on T1-weighted images and of high signal intensity on T2-weighted images [5]. Occasionally, an insulinoma is of dark signal intensity on T2-weighted sequences due to a fibrous stroma. In our study, the MRI sensitivity for detecting functioning NENs of ≤2 cm in diameter reached 85%, which is similar to the sensitivity achieved by invasive procedures [5]. For gastrinomas, a MRI sensitivity of up to 62% has been reported [34]. With present techniques, MRI should detect lesions >2 cm with a sensitivity of over 85%. Therefore, MRI with state-of-the-art equipment and optimal imaging techniques appears to be a useful strategy for diagnosing small pancreatic NENs. Nevertheless, contrast-enhanced MDCT, with its superior spatial resolution and very thin sections, currently surpasses MRI in diagnosing these small neoplasms. SSR with various derivatives of long-acting somatostatin analogues [22] can be used for small gastrinomas, somatostatinoma, glucagonoma, carcinoid, and VIPoma, but insulinomas may be missed due to reduced receptor expression [22, 33]. SSR with 111In-octreotide can

Ruedi F. Thoeni

often diagnose small lesions only suspected on CT or MRI and detect metastases not diagnosed with other modalities.

Non-hyperfunctioning Neuroendocrine Neoplasms Non-hyperfunctioning or non-syndromic NENs are less frequently encountered than insulinomas or gastrinomas, representing 15-25% of these neoplasms [31]. While they arise from the alpha or beta cells of the pancreas, these neoplasms are hormonally quiescent (probably very minimal secretion) and often present as a mass with or without jaundice or gastric outlet obstruction. These tumors are usually located in the pancreatic head and can measure up to 20 cm in diameter. There may be solid and necrotic components, with coarse calcifications present in up to 25%. The mass is hypervascular with a late capillary stain. The tumor does not encase vessels but in 80-100% there is malignant transformation, with liver metastases and adenopathy. The cumulative 5-year survival is 52-58% [35]. The key features of non-functioning NENs are their large size, hypervascularity, and the absence of vascular encasement. Results with CT and MRI are similar.

Cystic Neoplasms of the Pancreas Serous and Mucinous Cystic Neoplasms of the Pancreas Cystic neoplasms of the pancreas are uncommon tumors and account for <5% of pancreatic neoplasms. Pancreatic cystic neoplasms are classified into two categories: (1) serous cystic (usually microcystic, occasionally macrocystic: unilocular or oligocystic) neoplasms that are benign; and (2) mucinous cystic (macrocystic) neoplasms that are potentially malignant or already malignant at the time of diagnosis. The rare serous macrocystic variant is benign and exhibits radiological features similar to those of mucinous cystadenoma. Serous and mucinous cystic neoplasms, except for intraductal papillary mucinous tumors (IPMT), do not communicate with the pancreatic duct. Serous cystic neoplasms of the pancreas are observed in middle-aged and elderly women. This type of tumor may not require surgical treatment whereas mucinous cystic tumors should be resected because of their malignant potential. Nevertheless, some surgeons prefer to resect the serous type as well. In general, the patient’s age, symptoms, and overall condition, the lesion’s location, and its growth over time are factors that help to decide whether surgery is needed [36]. Often, any cyst that increases in size over time, any symptomatic cyst, and cysts in older fit patients are selected for surgery. CT can accomplish pre-operative differentiation of the two types in many cases. In serous cystic tumors, traditionally the diagnosis is made if the number of cysts within the tumor is more than six and the diameters of the cysts <2 cm. A

Diseases of the Pancreas, II: Tumors

newer nomenclature prefers to call cysts ≤1 cm definitely serous, those >1-2 cm equivocal, and those >2 cm definitely mucinous. Grossly, these serous tumors appear either as solid tumors with innumerable tiny cysts or as honeycombed cystic tumors, depending on the amount of connective tissue. They have a lobulated margin (Fig. 3). At times, it is difficult to visualize the cystic areas. Calcifications in serous tumors are central in location. A central enhancing scar may be present and is characteristic of a serous tumor [28]. Mucinous cystic neoplasms of the pancreas (also called cystadenomas and cystadenocarcinomas according to the old nomenclature) have six or fewer cysts, the diameters of the cysts measure >2 cm, a central enhancing scar is rarely seen, and calcifications are peripheral [28]. The margins usually are smooth and metastatic disease may be present at the time of diagnosis. Based on the above-mentioned criteria, a correct diagnosis of a serous cystic pancreatic tumor can be made in 62% of patients by CT, in 74% by sonography, and in 84% using both modalities [37]. Overall, the results for mucinous cystic tumors are inferior. Pancreatic pseudocysts and cystic forms of islet cell tumors, ductal carcinomas, solid and papillary tumors, and lymphangioma of the pancreas can be indistinguishable on CT from cystic neoplasms. Thus, EUS needle biopsies of the lesions often are necessary [38]. Better definition of the internal architecture of these cystic neoplasms is frequently obtained with MRI rather than CT. MRI also demonstrates the presence of mucin, seen as an area of increased signal intensity within the cysts on T1-weighted sequences. Septa and wall thickness of the lesions are well demonstrated by MRI but this is not always true for calcifications. MRI is of great help in distinguishing these cystic neoplasms from pseudo-

93

cysts of the pancreas, particularly if they are multiple. Both MRCP and MDCT with curved planar reconstruction can demonstrate the absence of a connection to the main pancreatic duct.

Intraductal Papillary Mucinous Neoplasm of the Pancreas A rare tumor that is considered a subtype of the mucinous cystic neoplasms of the pancreas is the intraductal papillary mucinous tumor of the pancreas (IPMN, formerly also called ductectatic cystadenoma or ductectatic cystadenocarcinoma). IPMN can be classified as main duct, branch duct (side-branch), or mixed type depending on the site and extent of involvement [39]. The cystic changes always demonstrate a connection to the pancreatic duct, which is a diagnostic feature that can be seen on MDCT and MRI. The branch duct tumor consists of cystic dilation of the side branches of the pancreatic duct, usually in the uncinate process. These ducts are lined with atypical, hyperplastic, or clearly malignant epithelium. In the late stages, the tumor nodules of the ducts produce copious mucinous secretions that fill the entire duct. Since extension into the parenchyma and beyond occurs relatively late in branch duct IPMN and overall malignant degeneration is rare, the overall prognosis is good. In 2544% of resected specimens of the other two types, malignancy is present. Resection is therefore the treatment of choice in these patients. CT shows markedly dilated ducts and cystic-appearing structures filled with mucinous material, which has a slightly higher attenuation than that of water. MRI seems to have a slight advantage over CT because it can visualize mucin within the cysts as well as the internal architecture of the lesion, including a solid mass and mural nodules (which are signs of malignancy) slightly better than CT. EUS also is well suited to detect mural nodules.

Solid Pseudopapillary Epithelial Neoplasm of the Pancreas

Fig. 3. Thin-section (1.25 mm) axial MDCT of a serous cystic neoplasm of the pancreas in the pancreatic phase. A lobulated and septated cystic mass is present in the tail of the pancreas (arrows). The individual cysts are small and the septa barely perceptible

Solid pseudopapillary epithelial neoplasms, previously called solid and cystic tumors of the pancreas, are rare tumors. They are seen almost exclusively in young women and are located mostly in the tail of the pancreas. This neoplasm is characterized by a solid peripheral area of tumor and a central zone of degeneration consisting of hemorrhage and cystic spaces filled with necrotic debris that can be visualized by CT (Fig. 4) and MRI. On imaging, these tumors appear as sharply defined, heterogeneous, large cystic pancreatic masses with solid components. They usually are benign but in older women they may be malignant [40]. Calcifications may be present in the capsule or in the inner portion of the mass. EUS also can be helpful in visualizing the nodules and the internal architecture of these masses.

94

Fig. 4. Thin-section (1.25 mm) axial MDCT of a solid pseudopapillary epithelial neoplasm of the pancreas in the pancreatic phase. A large enhancing mass (arrows) with partial necrosis is seen in the tail of the pancreas. The origin from the pancreas is clearly visualized on this image

Follow-Up for Cystic Neoplasms of the Pancreas Small lesions (≤3 cm) that are asymptomatic, show no signs of malignancy, and have negative fine-needle aspiration biopsy results can be followed every 6 months for one year and then yearly for a total of 4 years. If the cyst becomes symptomatic, increases in size during observation, shows malignant features including a thick wall, multiple irregular septation, and nodules, and/or has increased CEA or CA 19.9, positive cytology, or mucin in the aspirate, it should be surgically removed. Often, cysts detected in the elderly and fit patient are removed regardless of the features because of the increased incidence of malignancy in these lesions.

References 1. Bipat S, Phoa SS, van Delden OM et al (2005) Ultrasonography, computed tomography and magnetic resonance imaging for diagnosis and determining resectability of pancreatic adenocarcinoma: a meta-analysis. J Comput Assist Tomogr 29: 438-445 2. Midwinter MJ, Beveridge CJ, Wilsdon JB et al (1999) Correlation between spiral computed tomography, endoscopic ultrasonography and findings at operation in pancreatic and ampullary tumours. British J Surg 86:189-193 3. McLean AM, Fairclough PD (2005) Endoscopic ultrasound in the localisation of pancreatic islet cell tumours. Best Pract Res Clin Endocrinol Metab 19:177-193 4. Canto MI, Goggins M, Yeo CJ et al (2004) Screening for pancreatic neoplasia in high-risk individuals: an EUS-based approach. Clin Gastroenterol Hepatol 2:606-621 5. Thoeni RF, Mueller-Lisse UG, Chan R et al (2000) Detection of small, functional islet cell tumors in the pancreas: Selection of MRI Imaging sequences for optimal sensitivity. Radiology 214:483-490

Ruedi F. Thoeni

6. Semelka RC, Kroeker MA, Shoenut JP et al (1991) Pancreatic disease: prospective comparison of CT, ERCP, and 1.5 T MR imaging with dynamic Gadolinium enhancement and fat suppression. Radiology 181:785-791 7. Volmar KE, Vollmer RT, Jowell PS et al (2005) Pancreatic FNA in 1000 cases: a comparison of imaging modalities. Gastrointest Endosc 61:854-861 8. Ardengh JC, de Paulo GA, Ferrari AP (2004) EUS-guided FNA in the diagnosis of pancreatic neuroendocrine tumors before surgery. Gastrointest Endosc 60:378-384 9. Maguchi H (2004) The roles of endoscopic ultrasonography in the diagnosis of pancreatic tumors. J Hepatobiliary Pancreat Surg 11:1-3 10. Vachiranubhap B, Kim YH, Balci NC et al (2009) Magnetic resonance imaging of adenocarcinoma of the pancreas. Top Magn Reson Imaging 20:3-9 11. Manfredi R, Graziani R, Motton M et al (2009) Main pancreatic duct intraductal papillary mucinous neoplasms: accuracy of MR imaging in differentiation between benign and malignant tumors compared with histopathologic analysis. Radiology 253:106-115 12. Ku YM, Shin SS, Lee CH et al (2009) Magnetic resonance imaging of cystic and endocrine pancreatic neoplasms. Top Magn Reson Imaging 20:11-18 13. Brennan DD, Zamboni GA, Raptopoulos VD et al (2007) Comprehensive preoperative assessment of pancreatic adenocarcinoma with 64-section volumetric CT. Radiographics 27:1653-1666. 14. D’Onofrio M, Zamboni G, Faccioli N et al (2007) Ultrasonography of the pancreas. 4. Contrast-enhanced imaging. Abdom Imaging 32:171-181 15. Akisik MF, Sandrasegaran K, Aisen AA et al (2006) Dynamic secretin-enhanced MR cholangiopancreatography. Radiographics 26:665-677 16. Lytras D, Connor S, Bosonnet L et al (2005) Positron emission tomography does not add to computed tomography for the diagnosis and staging of pancreatic cancer. Dig Surg 22:55-61 17. Heinrich S, Goerres GW, Schafer M et al (2005) Positron emission tomography/computed tomography influences on the management of resectable pancreatic cancer and its cost-effectiveness. Ann Surg 242:235-243 18. Ruf J, Lopez Hanninen E, Oettle H et al (2005) Detection of recurrent pancreatic cancer: comparison of FDG-PET with CT/MRI. Pancreatology 5:266-272 19. Orlando LA, Kulasingam SL, Matchar DB (2004) Meta-analysis: the detection of pancreatic malignancy with positron emission tomography. Aliment Pharmacol Ther 20:1063-1070 20. Lee TY, Kim MH, Park do H et al (2009) Utility of 18F-FDG PET/CT for differentiation of autoimmune pancreatitis with atypical pancreatic imaging findings from pancreatic cancer. AJR Am J Roentgenol 193:343-348 21. Farma JM, Santillan AA, Melis M et al (2008) PET/CT fusion scan enhances CT staging in patients with pancreatic neoplasms. Ann Surg Oncol 15:2465-2471 22. Virgolini I, Traub-Weidinger T, Decristoforo C (2005) Nuclear medicine in the detection and management of pancreatic isletcell tumours. Best Pract Res Clin Endocrinol Metab 19:213-227 23. Buchmann I, Henze M, Engelbrecht S et al (2007) Comparison of 68Ga-DOTATOC PET and 111In-DTPAOC (Octreoscan) SPECT in patients with neuroendocrine tumours. Eur J Nucl Med Mol Imaging 34:1617-1726 24. Horner MJ, Ries LAG, Krapcho M et al (eds) posted to the SEER web site (2009) SEER Cancer Statistics Review, 1975-2006, National Cancer Institute. Bethesda, MD, based on November 2008 SEER data submission, http://seer.cancer.gov/csr/1975_2006 25. Cancer Facts & Figures (2009) American Cancer Society (ACS), Atlanta, Georgia 26. Fletcher JG, Wiersema MJ, Farrell MA et al (2003) Pancreatic malignancy: value of arterial, pancreatic, and hepatic phase imaging with multi-detector row CT. Radiology 229:81-90

Diseases of the Pancreas, II: Tumors

27. Imbriaco M, Megibow AJ, Ragozzino A et al (2005) Value of the single-phase technique in MDCT assessment of pancreatic tumors. AJR Am J Roentgenol 184:1111-1117 28. Sahani DV, Kadavigere R, Saokar A et al (2005) Cystic pancreatic lesions: a simple imaging-based classification system for guiding management. Radiographics 25:1471-1484 29. Morris-Stiff G, Webster P, Frost B et al (2009) Endoscopic ultrasound reliably identifies chronic pancreatitis when other imaging modalities have been non-diagnostic. JOP 10: 280-283 30. Aslanian H, Salem R, Lee J et al (2005) EUS diagnosis of vascular invasion in pancreatic cancer: surgical and histologic correlates. Am J Gastroenterol 100:1381-1385 31. Thoeni RF (2009) Imaging of endocrine tumors. In: Heiken JP (ed) Pancreatic cancer. Contemporary issues in cancer imaging. Cambridge University Press, Cambridge, pp 104-129 32. Gouya H, Vignaux O, Augui J et al (2003) CT, EUS combined protocol for preoperative evaluation of pancreatic insulinoma. AJR Am J Roentgenol 181:987-992 33. T. Zimmer, U. Stolzel, M. Bader et al (1996) Endoscopic ultrasonography and somatostatin receptor scintigraphy in the preoperative localisation of insulinomas and gastrinomas. Gut 39:562-568

95

34. Pisegna JR, Doppman JL, Norton JA et al (1993) Prospective comparative study of the ability of MR imaging and other imaging modalities to localize tumors in patients with Zollinger-Ellison syndrome. Dig Dis Sci 38:1318-1320 35. H. Liang H, Wang P, Wang XN et al (2004) Management of nonfunctioning islet cell tumors. World J Gastroenterol 10:1806-1809 36. Spinelli KS, Fromwiller TE, Daniel RA et al (2004) Cystic pancreatic neoplasms: Observe or operate. Ann Surg 239:651-659 37. Procacci C, Graziani R, Bicego E et al (1997) Serous cystadenoma of the pancreas: report of 30 cases with emphasis on the imaging findings. J Comput Assist Tomogr 21:373-382 38. Belsley NA, Pitman MB, Lauwers GY et al (2008) Serous cystadenoma of the pancreas: limitations and pitfalls of endoscopic ultrasound-guided fine-needle aspiration biopsy. Cancer 114:102-110 39. Ogawa H, Itoh S, Ikeda M et al (2008) Intraductal papillary mucinous neoplasm of the pancreas: assessment of the likelihood of invasiveness with multisection CT. Radiology 248:876-886 40. Lee JH, Yu JS, Kim H et al (2008) Solid pseudopapillary carcinoma of the pancreas: differentiation from benign solid pseudopapillary tumour using CT and MRI. Clin Radiol 63:1006-1014

IDKD 2010-2013

Adrenal Imaging and Intervention William W. Mayo-Smith1, Isaac R. Francis2 1 Department 2 Department

of Radiology, Warren Alpert School of Medicine Brown University, Rhode Island Hospital, Providence, RI, USA of Radiology, University of Michigan, Ann Arbor, MI, USA

Introduction The objectives of this chapter are: (1) to describe the different work-ups for adrenal masses, depending on clinical scenario, (2) define adrenal incidentaloma, (3) discuss the relative risk factors for benign and malignant adrenal masses, (4) describe the imaging techniques to differentiate benign from malignant adrenal masses, and (5) discuss the recommended medical work-up of an adrenal mass. The adrenal gland is a complex organ that is made up of the catecholamine-producing medulla and the steroidproducing cortex. It is a common site of primary tumors (functional and non-functional) and of metastases. The optimal work-up for an adrenal mass depends on the patient’s clinical scenario and whether detection or characterization is the primary concern. In general, it is useful to separate adrenal work-ups into one of three algorithms: 1. Detection of an adrenal tumor in a patient with a known biochemical abnormality. 2. Staging of a patient with a known primary neoplasm. 3. Characterization of an incidental adrenal mass detected on cross-sectional imaging. This is a relevant topic as there has been an over 20fold increase in medical literature reports on this subject in the past two decades.

Detection of Biochemically Active Adrenal Tumor Biochemically active adrenal neoplasms originate in the adrenal cortex, in which case they produce an excess of either glucocorticoids, aldosterone, or androgens, or in the adrenal medulla, thus producing an excess of catecholamines. Cushing’s syndrome results from an overproduction of cortisol by the adrenal cortex and approximately 80% of these cases are due to stimulation of the adrenal glands by a pituitary adenoma. A primary adrenal cortical tumor is seen in 20% of patients with Cushing’s syndrome and <1% have ectopic production of ACTH by a non-pituitary neoplasm. The work-up of patients presenting with Cushing’s syndrome involves a dexamethasone suppression test, pituitary magnetic resonance imaging (MRI) to look for a pituitary adenoma, and an adrenal

computed tomography (CT) study depending on the suspected source of ACTH production. If pituitary and adrenal neoplasms are ruled out and an ectopic source of hormone secretion is suspected, then a chest and abdominal CT should be performed to search for it. Conn’s syndrome can result from either adrenal hyperplasia or an adrenal cortical tumor producing elevated levels of aldosterone, leading to increased sodium retention, hypertension, and potassium wasting. The diagnosis is suspected in a hypertensive patient with low serum potassium and is confirmed by measuring the ratio of serum aldosterone to renin levels. When the diagnosis is suspected based on biochemical assays, CT scans using 3-mm collimation targeted to the adrenals is useful to differentiate a small adrenal neoplasm from bilateral hyperplasia. If findings are equivocal on CT, then adrenal venous sampling to localize and lateralize the site of elevated aldosterone production should be performed. Pheochromocytomas originate from the adrenal medulla and produce an excess of catecholamines, causing hypertension. These tumors are solitary and occur sporadically in 90% of cases. If the diagnosis is suspected, the most appropriate first-line test is the measurement of plasma metanephrines. If these are equivocal, urinary metanephrines can be measured. Once the diagnosis has been established biochemically, the primary role of the radiologist is to determine the site of origin of the pheochromocytoma. Over 95% of pheochromocytomas originate in the adrenals; therefore, a non-contrast-enhanced CT examination of the adrenals is usually sufficient. MRI of pheochromocytoma typically demonstrates a T2 hyperintense mass, although the finding is non-specific, as pheochromocytomas can also have intermediate signal intensity on T2-weighted images, thus simulating adrenal cortical carcinoma; also, other adrenal lesions can be T2 hyperintense (adrenal cysts and lymphangiomas). MRI can also detect extra-adrenal paragangliomas along the sympathetic chain. While metaiodobenzyl-guanidine (MIBG) scintigraphy has high specificity (>95%) for the diagnosis of pheochromocytoma, its sensitivity is only 77-90%. Recent studies have suggested that MIBG scintigraphy should be used selectively and only in patients with familial or hereditary disorders, in the detection of metastatic

Adrenal Imaging and Intervention

disease, and in patients with biochemical evidence for pheochromocytoma and negative CT or MRI. These studies also concluded that MIBG scintigraphy does not offer any added advantage in patients with biochemical evidence for a pheochromocytoma, no hereditary or familial diseases, and a unilateral adrenal mass detected on CT or MRI [1, 2]. The standard treatment of a biochemically active adrenal tumor is open or laparoscopic resection. More recently, non-invasive techniques have been described, including selective arterial embolization, percutaneous injection of acetic acid, and radiofrequency ablation.

Staging Patients with Known Carcinoma Evaluation of the adrenal gland in the oncology patient is complicated because the gland is a frequent site of metastases, but benign adrenal adenomas are also common (detected in 2-5% of autopsy series). Thus, the presence of an adrenal mass does not necessarily implicate metastases. The role of cross-sectional imaging in the oncology patient is to detect enlargement of the adrenal gland and characterize the enlargement as either benign or malignant. More recently, PET imaging has facilitated the staging of neoplasms because adrenal metastases tend to demonstrate increased activity, having a greater uptake relative to the liver, while most benign adenomas do not. More recent studies have confirmed the high sensitivity of PET/CT in detecting malignant lesions but the specificity is lower (87-97%). This loss of specificity is attributable to a small number of adenomas and other benign lesions that mimic malignant lesions [3, 4]. Depending on the primary tumor, CT or PET/CT is a useful first-line exam to stage a known neoplasm. If the patient demonstrates multiple sites of metastatic disease, then evaluation of an adrenal mass is not important. If the adrenal mass is the only abnormality, further evaluation is required to differentiate an adenoma from a metastatic focus. Currently, there are two main criteria (anatomical and physiological) used to differentiate benign adenomas from malignant adrenal masses: (1) the intracellular lipid content of the adrenal mass, which represents the anatomical difference between adenomas and metastases, and (2) differences in vascular enhancement patterns, which represent the physiological difference. Approximately 80% of benign adenomas have abundant intracytoplasmic lipid in the adrenal cortex and thus are of low density on unenhanced CT or show signal drop-off on out-of-phase chemical shift MRI (CSMRI). Conversely, most metastases have little intracytoplasmic lipid and thus do not have a low density on non-contrast CT. At a threshold of 10 HU, CT has a 71% sensitivity and 98% specificity for characterizing adrenal adenomas. While a low HU is useful to characterize lipid-rich adenomas, it is estimated that up to 20% of adenomas do not contain sufficient lipid to be of low density on unenhanced CT [5-7]. More recently, Bae et al. showed that a histogram analysis of adrenal masses

97

(evaluating microscopic levels of lipid on a pixel by pixel basis) is useful to differentiate adenomas from metastases on non-contrast- and contrast-enhanced CT [8]. According to their findings, if an adrenal mass has >10% negative pixels, it is diagnostic of an adenoma (on either noncontrast- or contrast-enhanced CT). However, due to differing results in more recent studies, this approach is still considered as “research” and its use in clinical practice remains limited. The physiological difference in perfusion between adenomas and metastases can be used to differentiate these entities. Adenomas enhance rapidly with intravenous contrast (iodinated CT contrast or MR gadolinium chelates) and also have rapid washout. Metastases also enhance vigorously with dynamic contrast but the washout of contrast is more prolonged than in adenomas. This difference in contrast washout has been exploited to further differentiate benign from malignant adrenal lesions by comparing pre-contrast HU values with dynamic and 15-min delayed HU values [9, 10]. Absolute percent washout (APW) values are calculated by the formula: (HU at dynamic CT – HU at 15-min delayed CT)/(HU at dynamic CT – HU at non-contrast CT) × 100. A value ≥60% is diagnostic of an adenoma. Relative percent washout (RPW) is used when a non-contrast CT value is not available and the dynamic enhanced values are compared to 15-min delayed scans. RPW is calculated by the formula: (HU dynamic CT – HU 15-min delayed CT)/HU dynamic CT × 100, and a value >40% is diagnostic of adenoma. Adenomas can be differentiated from metastases using CSMRI if the patient has a non-diagnostic CT, is allergic to iodinated contrast, or in young patients, in whom radiation exposure is an issue [11, 12]. Most adrenal adenomas contain sufficient intracellular lipid and lose signal on the out-of-phase image compared to the spleen. Visual analysis is adequate in most cases to make this observation, but quantitative methods, such as the signal intensity index, may also be useful [13, 14]. If the CT, MRI, or PET findings are equivocal, adrenal biopsy using CT guidance should be performed, particularly to stage a lung carcinoma in patient who has no other sites of metastatic disease, as this may determine whether surgical resection is a therapeutic option. The role of adrenal biopsy has evolved in the last few years; in addition to the above indication of an indeterminate adrenal mass, adrenal biopsy can also be used to confirm metastatic disease to the adrenal glands in patients with suspected solitary adrenal metastasis. CT-guided biopsy has been shown to be safe, with a diagnostic yield of 8396% and a 3% complication rate [15].

Evaluation of an Incidentally Discovered Adrenal Mass As the indications for abdominal imaging (particularly CT) continue to increase, so does the detection of the incidental adrenal mass-given the high prevalence of

98

adenomas in the general population (3-7%) [16-18]. In general, the overwhelming majority of incidentally discovered adrenal masses (incidentalomas) are benign in a patient with no known malignancy [19]. An adrenal incidentaloma is defined as “an unsuspected and asymptomatic mass (≥1 cm) detected on imaging exams obtained for other purposes”. Risk factors for an incidental adrenal mass being malignant include lesion size, change in size, and their occurrence in a patient with a history of malignancy. For patients with no history of malignancy, most small (<4 cm) incidentally discovered adrenal masses are benign, and an extensive and costly imaging workup is usually not justified. An additional biochemical workup to exclude functioning adrenal cortical or medullary lesions is recommended by endocrinologists; however, these functioning lesions present very rarely as “incidental” adrenal masses, such that this expensive workup is not done routinely [20]. If a mass of any size has the typical features of a benign lesion, such as a cyst or myelolipoma, no additional workup or follow-up imaging is needed. For adrenal lesions <4 cm with non-diagnostic imaging features, if prior imaging is available and the lesion has been stable for at least 1 year, then it can be deemed benign and there is no need for additional imaging follow-up. But if the lesion is enlarging, then it may be prudent to proceed to adrenal biopsy or resection. If there are no prior comparison CT or MRI exams and patient has no hystory of malignancy, and if the lesion has benign imaging features (homogeneous, smooth margins), a diagnosis of a benign lesion may be made and one can consider a follow-up with an unenhanced CT or CSMRI exam in 12 months. However, if there are suspicious imaging features on contrast-enhanced CT, then one should proceed with an unenhanced CT or CSMRI; if these exams do not confirm that the lesion is a lipid-rich adenoma, an adrenal protocol CT with washout calculations is recommended. If the lesion does not have the imaging and washout features of a benign lesion, then a biopsy may be appropriate. In patients with a prior history of cancer and an adrenal masses of any size, if imaging features on contrastenhanced are not diagnostic for a benign lesion and there is no prior imaging, one can consider an unenhanced CT or CSMRI or PET imaging. If the lesion does not behave like a typical adenoma, then one should proceed to adrenal CT with washout. If the lesion does not show either washout features of an adenoma or the findings of an adenoma on PET imaging, then a biopsy should be considered. In patients with no history of cancer and an adrenal mass >4 cm in size, resection is an option; but if there is a history of prior cancer, then a PET scan or a biopsy is recommended [21].

William W. Mayo-Smith, Isaac R. Francis

References 1. Miskulin J, Shulkin BL, Doherty GM et al (2003) Ispreoperative iodine 123 metaiodobenzylguanidine scintigraphy routinely necessary before initial adrenalectomy for pheochromocytoma? Surgery 134:918-923 2. Greenblatt DY, Shenker Y, Chen H (2008) The utility of metaiodobenzylguanidine (MIBG) scintigraphy in patients with pheochromocytoma. Ann Surg Oncol 15:900-905 3. Boland GW, Blake MA, Holalkere NS, Hahn PF (2009) PET/CT for the characterization of adrenal masses in patients with cancer: qualitative versus quantitative accuracy in 150 consecutive patients. AJR Am J Roentgenol 192:956-962 4. Groussin L, Bonardel G, Silvéra S et al (2009) 18F-Fluorodeoxyglucose positron emission tomography for the diagnosis of adrenocortical tumors: a prospective study in 77 operated patients. J Clin Endocrinol Metab 94:1713-1722 5. Lee MJ, Hahn PF, Papanicolaou N et al (1991) Benign and malignant adrenal masses: CT distinction with attenuation coefficients, size, and observer analysis. Radiology 179:415-418 6. Korobkin, M, Brodeur FJ, Francis IR et al (1998) CT time-attenuation washout curves of adrenal adenomas and nonadenomas. AJR Am J Roentgenol 170:747-752 7. Boland GW, Lee MJ, Gazelle GS et al (1998) Characterization of adrenal masses using unenhanced CT: an analysis of the CT literature. AJR Am J Roentgenol 171:201-204 8. Bae KT, Fuangtharnthip P, Prasad SR et al (2003) Adrenal masses: CT characterization with histogram analysis method. Radiology 228:735-742 9. Korobkin M, Giordano TJ, Brodeur FJ (1996) Adrenal adenomas: relationship between histologic lipid and CT and MR findings. Radiology 200:743-747 10. Caoili EM, Korobkin M, Francis IR et al (2002) Adrenal masses: characterization with combined unenhanced and delayed enhanced CT. Radiology 222:629-633 11. Tsushima Y, Ishizaka H, Matsumoto M (1993) Adrenal masses: differentiation with chemical shift, fast low-angle shot MR imaging. Radiology 186:705 12. Israel GM, Korobkin M, Wang C et al (2004) Comparison of unenhanced CT and chemical shift MRI in evaluating lipidrich adrenal adenomas. AJR Am J Roentgenol 183:215-219 13. Fujiyoshi F, Nakajo M, Fukukura Y, Tsuchimochi S (2003) Characterization of adrenal tumors by chemical shift fast lowangle shot MR imaging: comparison of four methods of quantitative evaluation. AJR Am J Roentgenol 180:1649-1657 14. Mayo-Smith WW, Lee MJ, McNicholas MM et al (1995) Characterization of adrenal masses (5 cm) by use of chemical shift MR imaging: observer performance versus quantitative measures. AJR Am J Roentgenol 165:91-95 15. Silverman SG, Mueller PR, Pinkney LP et al (1993) Predictive value of image-guided adrenal biopsy: analysis of results of 101 biopsies. Radiology 187:715-718 16. Grumbach MM, Biller BM, Braunstein GD et al (2003) Management of the clinically inapparent adrenal mass (“incidentaloma”). Ann Intern Med 138:424-429 17. Young WF (2007) Clinical practice. The incidentally discovered adrenal mass. N Engl J Med 356:601-610 18. Choyke PL (2006) ACR Appropriateness Criteria on incidentally discovered adrenal mass. J Am Coll Radiol 3:498-504 19. Song JH, Chaudhry FS, Mayo-Smith WW (2008) The incidental adrenal mass on CT: Prevalence of adrenal disease in 1049 consecutive adrenal masses in patients with no known malignancy. AJR Am J Roentgenol 190:1163-1168 20. NIH state-of-the-science statement on management of the clinically inapparent adrenal mass (“incidentaloma”) (2002) NIH Consens State Sci Statements 19:1-25 21. Boland GW, Blake MA, Hahn PF, Mayo-Smith WW (2008) Imaging characterization of adrenal incidentalomas: principles, techniques and algorithms. Radiology 249:756-775

IDKD 2010-2013

Renal Tumors Richard H. Cohan1, Ronald J. Zagoria2 1 University 2 Wake

of Michigan, Ann Arbor, MI, USA Forest University, Winston-Salem, NC, USA

Introduction This chapter reviews recent advances in the imaging of renal masses, primarily using computed tomography (CT) and magnetic resonance imaging (MRI). The focus is on developments in the use of imaging to differentiate benign from malignant renal lesions and to assess patients following treatment for a renal mass.

Ultrasonography Although non-contrast ultrasound evaluates the internal morphology of cystic lesions with more detail than CT, it is not as sensitive in detecting renal masses as CT or MRI. Further, ultrasound is limited in its ability to characterize detected masses. Most investigators consider ultrasound to be diagnostically definitive only when it identifies a renal mass as a simple cyst. The majority of radiologists recommend that complex cysts and solid masses detected on ultrasound be evaluated further with CT or MRI.

CT and MRI Techniques Adequate CT evaluation of a patient with a known or suspected renal mass requires that at least two series be obtained: unenhanced and delayed-enhanced images, with the latter obtained at least 100 s after initiation of contrast material injection [1, 2]. Similarly, MRI examinations should include T1-weighted sequences obtained before and after gadolinium administration [3]. Additional MRI sequences or techniques that are often helpful include T2weighted, fat suppression, chemical shift, and diffusionweighted imaging.

Small Renal Masses Most renal masses detected with CT and some masses detected with MRI cannot be characterized due to their small size. For CT, this is felt to be true for renal masses

<1-1.5 cm in diameter. Given that small renal masses are very common, further imaging assessment of all of these lesions is not feasible, even though a few will be cancers. Recently, Patel and colleagues reported that the subjective impressions of radiologists are frequently quite accurate in identifying which small renal masses are cysts and therefore which lesions may not require follow-up [4]. Accordingly, we recommend that follow-up imaging of small renal masses be performed in only a few circumstances: (a) when the imaging suggests to a radiologist that a lesion may not be a simple cyst (often due to lesion heterogeneity), (b) when a lesion is detected in a young patient (<40 years), or (c) when a new small mass is seen in a patient at risk for renal malignancy (such as in patients with von Hippel Lindau, hereditary papillary renal cell cancer, or hereditary leiomyomatosis-renal cancer syndromes) [5]. In these instances, further evaluation with immediate MRI can be performed. Often, long-term follow-up (for up to 5 years) will be necessary. Followup imaging can be performed safely in most patients and does not usually result in a change in tumor stage or prognosis, even if the mass being followed is later found to be a small cancer. Most small renal cancers grow very slowly, with a mean growth rate of 0.4-0.5 cm/year [6, 7]. Therefore, on a 6-month follow-up examination, the typical growth of a renal cancer is near the range of CT measurement error (~2-3 mm). While the only generally known exceptions are the small aggressive renal type II papillary cancers seen in patients with hereditary leiomyomatosisrenal cancer syndrome [5], a recent study has demonstrated that some other small (<4 cm) renal cancers can invade the renal capsule or even metastasize distantly [8].

Cystic Renal Masses In 1986, Bosniak presented a CT cyst classification system [9] that categorized cysts based on their likelihood of their being malignant. This system has undergone a number of revisions, with the most recent version published in 2005 [10]. According to Bosniak, lesions with no or minimal complexity on CT, classified as category I or II lesions, are nearly always benign and do not require follow-up.

100

Only a tiny percentage of these contains malignant cells [11]. Category II lesions require imaging follow-up with immediate MRI, and/or with surveillance imaging, since a few of these are malignant. When these follow-up studies are interpreted, it must be remembered that both benign and malignant lesions can enlarge over time [11]. To further confuse the issue, some benign and malignant lesions can decrease in size over time [7]. Therefore, it is important to assess cystic lesions for increasing complexity, such as the development of new wall thickening or nodularity on follow-up studies [11]. These features are the best predictors of malignancy. More complex category III and IV cysts should be treated (in appropriate candidates), since many of these are cancers and imaging distinction between benign and malignant category III and IV lesions is not possible. When the Bosniak system is used with MRI, about 80% of cystic masses demonstrate similar complexity as seen on CT, but one in five appears more complex, demonstrating additional septations, wall or septal thickening, or subtle enhancement [12]. This is due, in part, to the ability of MRI to detect some internal cyst features not visible with CT, but also to the fact that MRI is more prone to artifacts.

Angiomyolipomas While nearly all angiomyolipomas (AMLs) are echogenic on ultrasound examinations, so are many small renal cancers. In a few series, acoustic shadowing has been identified posterior to some AMLs but not posterior to renal cancers; however, at many institutions this finding alone is not accepted as conclusive evidence of an AML [13]. Thus, echogenic masses generally are evaluated further with CT or MRI to determine whether macroscopic fat is present in the mass. If macroscopic fat is present, then the mass can be diagnosed as an AML (with case reportable exceptions, see [14]). Otherwise, the mass is highly likely to be a renal cell carcinoma. On CT, the presence within a renal mass of even small areas measuring –10 HU or less is considered diagnostic of macroscopic fat and of an AML [15]. On MRI, such fat typically is of high T1 and T2 signal intensity and loses signal with fat suppression. On opposed-phase chemicalshift imaging, there is a characteristic “India ink” artifact at fat-water interfaces within the AML and between the AML and adjacent tissue [16]. On occasion, AMLs can be very exophytic and difficult to differentiate from perinephric liposarcomas. This is an important distinction, given the very different treatments and prognoses for these two neoplasms. While percutaneous biopsy is often helpful in making the distinction, imaging features have also been identified that can be used to facilitate differentiation. Fatty perinephric masses are more likely to be exophytic AMLs if they contain large vessels or vessels that can be seen to extend to the renal cortex, or if they are associated with a renal

Richard H. Cohan, Ronald J. Zagoria

parenchymal defect [17, 18]. A mass that contains calcification is more likely to be a liposarcoma than an AML. Some AMLs do not contain identifiable macroscopic fat, including up to one-third of such tumors in patients with tuberous sclerosis. Although these tumors cannot be differentiated from other renal neoplasms on gross inspection, a number of studies have assessed the ability of imaging to differentiate minimal-fat-containing AMLs from other solid renal masses. A few studies have attempted to identify small foci of fat in minimal-fat-containing AMLs in the hope that this will permit these lesions to be correctly identified. The authors of these series have studied unenhanced CT attenuation [19], analyzed CT histograms [20], counted fat attenuation pixels [21], and quantitatively assessed fat on MRI [22]. Unfortunately, the results have been mixed. In some series, minimal-fat-containing AMLs have contained more measurable fat than renal cancers, while in other studies this has not been confirmed. A series by Catalano et al. [20] comparing AMLs to clear cell carcinomas is particularly concerning. The authors found that many clear cell renal cancers actually contained more measurable fat than did minimal-fat-containing AMLs. Additional MRI features independent of fat detection have also been evaluated. It has been suggested that a diagnosis of minimal-fat-containing AML should be considered, albeit not definitively, if a solid mass demonstrates low signal intensity on T2-weighted sequences and if it has only mild enhancement after gadolinium-based contrast material injection. These MRI characteristics are not seen in the majority of renal cancers.

Differentiating Non-Fat-Containing Renal Tumors Renal masses that do not contain recognizable fat cannot be distinguished consistently from each other. In the past, this was not felt to be an important problem, as it was believed that the vast majority of such lesions were renal cancers. Accordingly, all patients with non-fat-containing solid renal masses were referred for treatment. It has become apparent, however, that a sizeable minority of small solid renal masses are benign, a fact that has led some to recommend pre-treatment biopsy for all small renal tumors. Recent studies suggested that about one in five solid renal masses measuring <4 cm is benign [23, 24]. The frequency of benign tumors is even greater for renal lesions measuring <1 cm, with nearly half being benign [24].

Differentiating Benign from Malignant Renal Neoplasms Benign solid renal masses that are encountered routinely include the previously discussed minimal-fat-containing AMLs and oncocytomas. While some oncocytomas contain central scars that can be detected on imaging studies and oncocytomas also may demonstrate a spoke-wheel vascular pattern at CT or MR arteriography, these features are not diagnostic. Necrosis in renal cancers and

Renal Tumors

scars in oncocytomas, when they are present, are indistinguishable from one another. Furthermore, most oncocytomas evaluated on CT do not contain identifiable central scars [25]. Many renal cancers demonstrate a spokewheel arterial pattern and since renal cancer is much more common, most tumors that have this appearance are malignant. One cannot rely on an assessment of growth rate to distinguish oncocytomas from malignant renal neoplasms, because small renal cancers and many oncocytomas enlarge slowly and to a similar degree over time. In a meta-analysis reported by Chawla et al., there was no significant difference in the growth rates between these two different neoplasms [26]. Traditionally, percutaneous biopsy differentiation of oncocytomas from some renal cancers based upon histological appearance was difficult, if not impossible. However, several stains have recently been used in conjunction with an assessment of tumor morphology to facilitate this distinction, such as Hale colloidal iron, cytokeratin-7, and vimentin immunohistochemical stains. While several researchers have suggested that in many cases these stains permit the distinction of these two types of tumors [27, 28], the accuracy of biopsy for this distinction remains controversial. There are a variety of additional benign renal neoplasms that likewise have a non-specific appearance. These include the common papillary adenomas and renomedullary interstitial cell tumors (generally measuring <1 cm and encountered in 40-50% of adults in autopsy series) as well as very rare lesions, such as metanephric neoplasms, hemangiomas and lymphangiomas, leiomyomas, juxtaglomerular tumors, mixed epithelial and stromal tumors, and cystic nephromas [29].

Differentiating among Renal Cancer Subtypes As is the case with benign versus malignant renal lesions, one also cannot rely on differences in growth rates to distinguish among cell types of renal cancers. In a recent series, neither the initial size of a detected renal cancer nor the cell type of that cancer predicted the likelihood that the cancer would be more or less likely to grow quickly [7]. Nonetheless, there are some morphological differences among the different types of renal cancers. Papillary tumors can often be differentiated from other renal cancer cell types on MRI, as many papillary cancers demonstrate characteristic T2 signal hypointensity [30, 31]. In contrast, only a small minority of clear cell carcinomas are T2 hypointense [31]. Additionally, on contrastenhanced CT or MRI, both papillary subtypes, i.e., the better-prognosis type 1 tumors and the more aggressive type II tumors, tend to demonstrate more well-defined margins, more homogeneity, and less enhancement than do other renal cancer cell types [30, 32]. In one recent study, for example, the mean percent increase of T1 signal intensity after gadolinium-based contrast material administration during the corticomedullary phase was

101

over 200 for 75 clear cell, but only 110 for 10 chromophobe, and 32 for 28 papillary cancers [32]. A number of investigators have used diffusion-weighted MRI in the evaluation of renal masses [33]. Preliminary studies have demonstrated differences in diffusion for different types of renal masses. Not surprisingly, diffusion is least restricted in simple renal cysts and most restricted in the more cellular renal neoplasms (such as papillary cancers and angiomyolipomas) [33]. Diffusionweighted MRI may be able to play a role in renal mass characterization, particularly in patients who cannot receive gadolinium-based contrast material [33]. However, diffusion-weighted imaging has a number of problems. Firstly, there is overlap in the degree of restricted diffusion identified in benign and malignant renal lesions. Secondly, there are many technical issues, with measured apparent diffusion coefficients varying among scanner brands and protocols.

Pre-operative Imaging and Treatment of Renal Cancers CT and MRI are very accurate in their ability to localize and stage renal cancers prior to treatment. Pre-treatment imaging studies should be used to assess renal cancer size, location (with respect to the renal poles, renal sinus, and collecting system), evidence of gross perinephric extension, involvement of the ipsilateral renal vein or inferior vena cava (including a description of the extent of involvement), the presence or absence of enlarged lymph nodes, and of any distant metastases. Either technique can also be used for anatomical definition prior to anticipated surgery (including location and number of renal vessels). Treatment choice depends upon the pre-operative imaging characteristics of each tumor and the patient’s condition. Although percutaneous thermal ablation techniques are being performed with increasing frequency, they are still most often reserved for patients who are not good operative candidates, who have renal insufficiency, or who have multiple renal neoplasms. Based on early studies, the cure rate for percutaneous thermal ablation is comparable to that for surgery, but studies evaluating long-term oncological efficacy have yet to be completed. Masses most amenable to percutaneous ablation include smaller lesions (<4 cm) and lesions for which there is no radiological evidence of advanced disease (N0, M0) [34]. Masses in almost any location in the kidney now can be treated with equal efficacy with ablation. Precautions to protect adjacent organs, such as the bowel, should be used with ablation when the mass to be treated is closer than 2 cm from these structures. Currently, many urologists have shifted away from use of open total and partial nephrectomy to laparoscopic procedures, because, as is the case with percutaneous thermal ablation, laparoscopic nephrectomy is associated with considerably less patient perioperative morbidity and a shorter length of hospital stay [34].

102

Imaging after Renal Mass Treatment Both CT and MRI, often performed using unenhanced arterial phase imaging, and delayed enhanced imaging, are used widely to image patients after treatment of renal cancers. Recent studies have described the imaging appearance of masses treated with radiofrequency ablation [35, 36]. It is important to be aware of the normal imaging appearance of the ablated mass following successful ablation in order to avoid confusion with tumor recurrence. After successful renal mass ablation, there is initial expansion of the ablation site. The ablation bed then typically decreases in size, but rarely disappears entirely. Other typical “normal” findings include fat invagination between the ablation bed and normal renal parenchyma and a perilesional halo, such that the ablation cavity mimics an AML. Recurrent or residual tumor should be suspected when the ablation bed increases in size or when areas of enhancement are identified. The latter are usually nodular and crescentic and are often located at the interface between the ablation bed and adjacent renal parenchyma [37, 38]. Close-interval follow-up (for example, at 1, 3, 6, and 12 months) should be performed after ablation, since most tumor recurrences are detected within the first 3 months of the procedure [38]. The imaging appearance of patients treated with open and laparoscopic partial nephrectomy has also been described. Hemostatic devices such as Surgicel (Johnson & Johnson, USA) and renal bolsters, when inserted after partial nephrectomy, can be mistaken for areas of infection or recurrent tumor. While recurrent tumors often typically occur in the surgical bed, in adjacent lymph nodes, or with distant metastatic disease, patients treated laparoscopically may demonstrate a different pattern of recurrence, with implants growing along the laparoscopic port sites or elsewhere in the mesentery or peritoneum [39].

References 1. Cohan RH, Sherman LS, Korobkin M et al (1995) Renal masses: assessment of corticomedullary-phase and nephrographicphase CT scans. Radiology 196:445-451 2. Birnbaum BA, Jacobs JE (1996) Multiphasic renal CT: comparison of renal mass enhancement during the corticomedullary and nephrographic phases. Radiology 200:753-758 3. Hecht EM, Israel GM, Krinsky GA et al (2004) Renal masses: quantitative analysis of enhancement with signal intensity measurements versus qualitative analysis of enhancement with image subtraction for diagnosing malignancy at MR imaging. Radiology 232:373-378 4. Patel NS, Poder L, Wang ZJ et al (2009) The characterization of small hypoattenuating renal masses on contrast enhanced CT. Clin Imag 33:295-300 5. Choyke PL (2003) Imaging of hereditary renal cancer. Radiol Clin North Am 41:1037-1051 6. Bosniak MA, Birnbaum BA, Krinsky GA, Waisman J (1995) Small renal parenchymal neoplasms: further observations on growth. Radiology 197:589-597 7. Zhang J, Kang SK, Wang L et al (2009) Distribution of renal tumor growth rates determined by using serial volumetric CT measurements. Radiology 250:137-144

Richard H. Cohan, Ronald J. Zagoria

8. Pahernik S, Ziegler S, Roos F et al (2007) Small renal tumors: correlation of clinical and pathological features with tumor size. J Urol 178:414-417 9. Bosniak MA (1986) The current radiological approach to renal cysts. Radiology 158:1-10 10. Israel GM, Bosniak MA (2005) How I do it: evaluating renal masses. Radiology 236:441-450 11. Gabr AH, Gdor Y, Roberts WW, Wolf JS (2008) Radiographic surveillance of minimally and moderately complex renal cysts. Br J Urol Int 103:1116-1119 12. Israel GM, Hindman N, Bosniak MA (2004) Evaluation of cystic renal masses: comparison of CT and MR imaging by using the Bosniak classification system. Radiology 231:365-371 13. Farrelly C, Delaney H, McDermott R, Malone D (2008) Do all non-calcified echogenic renal lesions found on ultrasound need further evaluation with CT? Abdominal Imaging 33:44-47 14. Helenon O, Chretien Y, Paraf F et al (1993) Renal cell carcinoma containing fat: demonstration with CT. Radiology 188:429-430 15. Simpson E, Patel U (2006) Diagnosis of angiomyolipoma using computed tomography-region of interest ≤10 HU or 4 adjacent pixels ≤10 HU are recommended as the diagnostic thresholds. Clin Radiol 61:410-416 16. Israel GM, Hindman N, Hecht E, Krinsky G (2005) The use of opposed-phase chemical shift MRI in the diagnosis of renal angiomyolipomas. AJR Am J Roentgenol 194:1868-1872 17 Israel GM, Bosniak MA, Slywotzky CM et al (2002) CT differentiation of large exophytic renal angiomyolipomas and perirenal liposarcomas. AJR Am J Roentgenol 179:769-773 18. Ellingson JJ, Coakley FV, Joe BN et al (2008) Computed tomographic distinction of perirenal liposarcoma from exophytic angiomyolipoma: a feature analysis study. J Comput Assist Tomogr 32:548-552 19. Jinzaki M, Silverman SG, Tanimoto A et al (2005) Angiomyolipomas that do not contain fat attenuation at unenhanced CT. Radiology 234:311 20. Catalano OA, Samir AE, Sahani DV, Hahn PF (2008) Pixel distribution analysis: can it be used to distinguish clear cell carcinomas from angiomyolipomas with minimal fat? Radiology 247:738-746 21. Simpfendorfer C, Herts BR, Motta-Ramirez GA et al (2009) Angiomyolipoma with minimal fat on MDCT: can counts of negative attenuation pixels aid diagnosis? AJR Am J Roentgenol 192:438-443 22. Kim JK, Kim HS, Jang YJ et al (2006) Renal angiomyolipoma with minimal fat: differentiation from other neoplasms at double-echo chemical shift FLASH MR imaging. Radiology 239:274-280 23. Tuncali K, van Sonnenberg E, Shankar S et al (2004) Evaluation of patients referred for percutaneous ablation of renal tumors: importance of a preprocedural diagnosis. AJR Am J Roentgenol 183:575-582 24. Frank I, Blute ML, Cheville JC et al (2003) Solid renal tumors: an analysis of pathological features related to tumor size. J Urol 2217-2220 25. Choudhary S, Rajesh A, Mayer NJ et al (2009) Renal oncocytoma: CT features cannot reliably distinguish oncocytoma from other renal neoplasms. Clin Radiol 64:517-522 26. Chawla SN, Crispen PL, Hanlon AL et al (2006) The natural history of observed enhancing renal masses: meta-analysis and review of the world literature. J Urol 175:425-431 27. Shah RB, Bakshi N, Hafez KS et al (2005) Image-guided biopsy in the evaluation of renal mass lesions in contemporary urological practice: indications, adequacy, clinical impact, and limitations of the pathological diagnosis. Hum Pathol 36:1309-1315 28. Hes O, Michal M, Kuroda N et al (2007) Vimentin reactivity in renal oncocytoma: immunohistochemical study in 234 cases. Arch Pathol Lab Med 131:1782-1788 29. Prasad S, Surabhi VR, Menias CO et al (2007) Benign renal neoplasms in adults: cross-sectional imaging findings. AJR Am J Roentgenol 190:158-164

Renal Tumors

30. Yamada T, Endo M, Tsuboi M et al (2008) Differentiation of pathologic subtypes of papillary renal cell carcinoma on CT. AJR Am J Roentgenol 191:1559-1563 31. Oliva MR, Glickman JN, Zou KH et al (2009) Renal cell carcinoma: T1 and T2 signal intensity characteristics of papillary and clear cell types correlated with pathology. AJR Am J Roentgenol 192:1524-1530 32. Sun MRM, Ngo L, Genega EM et al (2009)) Renal cell carcinoma: dynamic contrast-enhanced MR imaging for differentiation of tumor subtypes- correlation with pathologic findings. Radiology 250:793-802 33. Taouli B, Thakur RK, Mannelli L et al (2009) Renal lesions: characterization with diffusion-weighted imaging versus contrast-enhanced MR imaging. Radiology 251:398-407 34. Ng CS, Wood CG, Silverman PM et al (2008) Renal cell carcinoma: diagnosis, staging, and surveillance. AJR Am J Roentgenol 191:1220-1232

103

35. Wile GE, Leyendecker JR, Krehbiel KA et al (2007) CT and MR imaging after imaging-guided thermal ablation of renal neoplasms. Radiographics 27:325-339 36. Davenport M, Caoili EM, Cohan RH et al (2009) MR and CT characteristics of successfully ablated renal masses status-post radiofrequency ablation. AJR Am J Roentgenol 192:1571-1578 37. Rutherford EE, Cast JEI, Breen DJ (2008) Immediate and long-term CT appearances following radiofrequency ablation of renal tumours. Clin Radiol 63:220-230 38. Kawamoto S, Solomon SB, Bluemke DA, Fishman EK (2009) Computed tomography and magnetic resonance imaging appearance of renal neoplasms after radiofrequency ablation and cryoablation. Semin Ultrasound CT MRI 30:67-77 39. Masterson TA, Russo P (2008) A case of port-site recurrence and locoregional metastasis after laparoscopic partial nephrectomy. Nat Clin Prac 5:345-349

IDKD 2010-2013

Urinary Tract Obstruction and Infection Parvati Ramchandani1, Julia R. Fielding2 1 Department 2 Department

of Radiology, University of Pennsylvania Medical Center, Philadelphia, PA, USA of Radiology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Urinary Tract Obstruction

as discussed below; however, some controversy exists as to which imaging studies are best for investigating suspected ureteral obstruction. Nonetheless, regardless of composition, virtually all ureteral stones will have high attenuation values, making them readily detectable with CT. Non-mineralized matrix stones and some drug-related stones (protease inhibitors) may not be visible on CT images but these are rarely encountered in routine clinical practice. A history of protease inhibitor therapy use is required to deduce that the drugs are the cause of acute flank pain and of the imaging findings of obstruction [10].

Imaging plays an important role in the evaluation of patients with acute and chronic urinary tract obstruction. The cause of obstruction varies greatly with the patient population and the geographic locale. In a series reporting on percutaneous nephrostomy drainage, urinary obstruction was related to calculus disease in 26% of patients and to malignancy in 61%. Carcinoma of the bladder, cervix, and colon were the most common primaries in patients with malignancies and urinary obstruction [1]. Acute ureteral obstruction is usually secondary to urolithiasis, and patients commonly present to an acute care facility with flank pain or acute renal colic. Noncontrast helical computed tomography (CT) is often the first study performed in suspected acute obstruction, as it is a safe and rapidly performed examination and the accuracy for detecting ureteral stones, the most common cause of ureteral obstruction, exceeds that of other imaging studies. Other causes of acute abdominal pain, such as appendicitis, leaking aortic aneurysm, and diverticulitis, can also be readily diagnosed and occur in 13-19% of cases (Table 1) [2-4]. Non-contrast helical CT has an overall accuracy of 97% for diagnosing ureteral stone disease, with a reported sensitivity of 97-100% and specificity of 94-96% [5-9]. This far exceeds the accuracy of intravenous urography (IVU) or ultrasound (US) (Table 2),

Table 1. CT findings in acute renal colic Hoppe (2006) [2] N Urinary stone Additional or alternative diagnosis Alternate diagnosis only Normal Pyelonephritis Renal cell carcinoma/ renal mass Cholecystitis Adnexal mass

Katz (2000) [3]

Ahmad (2003) [4]

1500 69% 71%

1000 62% 10%

233 68% 12%

24% 7% 3% 2%

7.50% 28% 1% 0.40%

X X 1% 1%

0.30% X

0.30% 2%

0.40% 2%

Table 2. Diagnostic performance for CT, US, and IVU in detection of ureteral stones Lead author

Year of publication

N

Stone +

Catalano

2002

181

82

Boulay Sheley Sourtzis

1999 1999 1999

51 180 36

49 87 36

Yilmaz

1998

97

64

Smith

1996

210

100

N, number of patients; Stone +, number of patients with ureteral stone

Test CT US/plain radiography CT CT CT IVU CT US IVU CT

Sensitivity

Specificity

0.92 0.77 1.0 0.86 1.0 0.66 0.94 0.19 0.52 0.97

0.96 0.96 0.96 0.91 1.0 1.0 0.97 0.97 0.94 0.96

Urinary Tract Obstruction and Infection

The proper technique for performing non-contrast helical CT to detect ureteral stone disease using a helical scanner is detailed below. Imaging should be performed from the top of the kidneys to the base of the bladder without intravenous or oral contrast material and scans should be obtained during a single breath-hold. A 16- to 64-slice MDCT with 1.5-mm collimation and a review of 5-mm contiguous images is appropriate. To reduce the radiation dose, a variable mA is used for each slice based on beam attenuation. In a recent report, the artificial addition of noise to a renal colic CT was used to show that tube current could be diminished 75% without changing the detection rate of stones >3 mm in diameter [11]. For the obese patient, a fixed mA equal to his/her weight in pounds will usually suffice. The expected dose is 30-40 mSv. Review of the images in cine mode on a workstation facilitates continuous identification of the ureter and workflow. While 3D reconstructions are usually not necessary, they may be occasionally useful to distinguish intraureteral from extraurinary calcifications. Due to the accuracy of CT and greater familiarity with the examination by referring physicians, especially within the emergency room, the number of renal colic CT scans has dramatically increased. This alone is not worrisome. More concerning is the number of patients who undergo repeat CT scanning. In a study by Katz et al. published in 2004, 4% of patients were found to have undergone three or more CT scans for renal stones during a 6-year period [12]. However, other reports indicate that there has been no decrease in the rate of positive diagnosis of obstructing urinary stones on CT, validating the increased frequency with which CT scans are obtained in the setting of acute flank pain [13, 14]. The overall rate of stone positivity on CT scanning in patients with suspected renal colic was reported in several studies to be 60-66% [13, 14]. Additionally, as discussed above, CT is the diagnostic test of choice to evaluate many acute abdominal problems, and a decrease in the stone positivity rate may not necessarily reflect overuse of CT [15]. Nevertheless, it seems reasonable that a patient with known history of stones and the appropriate clinical findings can be treated conservatively and without imaging beyond an abdominal radiograph. In addition to direct visualization of the ureteral stone, secondary signs of ureteral obstruction on non-contrast CT scans include unilateral nephromegaly, perinephric stranding, hydronephrosis, and periureteral stranding. The combination of perinephric stranding and unilateral hydronephrosis has a positive predictive value of 96% for the presence of stone disease. The absence of both of these signs has a negative predictive value of 93% for excluding stone disease. CT also gives information that determines therapy. Stones that are f5 mm in size, of smooth shape, and located within the distal one-third of the ureter are likely to pass spontaneously [7]. The major pitfall in non-contrast helical CT evaluation of the urinary tract for stone disease is difficulty in distinguishing pelvic phleboliths from ureteral calculi. The

105

presence of a tissue “rim” sign usually indicates that the calcification is a stone rather than a phlebolith. Alternatively, absence of the tissue rim sign or presence of a “comet tail” sign strongly suggests that the calcification is a phlebolith rather than a stone. In practice, the presence of two or more secondary signs of obstruction even without clear visualization of a calcification within the ureter indicates obstruction. If there is no history of recent stone passage and the CT scan demonstrates findings suggestive of obstruction, a contrast-enhanced study of the upper tracts may be needed to exclude a urothelial neoplasm, with additional cystoscopy to optimally evaluate the bladder. Intravenous urography is an alternative technique for the detection of urinary tract obstruction. It was once the imaging modality of choice in a patient who presented with acute flank pain. However, the necessity for intravenous radiographic contrast administration and the delay in obtaining relevant information about the site of obstruction makes IVU a less desirable study than unenhanced CT in an acute pain setting. Extraurinary causes of acute abdominal pain are not usually detectable with IVU. The technique was also once considered to have a role in the evaluation of pregnant patients with acute flank pain, when the results of an ultrasonographic examination were negative or equivocal. However, even in this patient population, CT may be the modality of choice, as it provides a definitive and rapid diagnosis. These advantages outweigh the slightly greater radiation exposure [16, 17]. Acute colic in pregnant patients is discussed below. It is safe to state that IVU in the evaluation of acute colic has been relegated to historical significance in most patients. Ultrasound, usually combined with an abdominal radiograph, is an alternative method for evaluating the obstructed or dilated urinary tract and is often the first imaging procedure in patients who should avoid radiation, such as pregnant women and children. Renal calculi as small as 0.5 mm may be detectable under optimal imaging conditions. For stones >5 mm in diameter, US has been shown to have a sensitivity and specificity of 96% and nearly 100%, respectively [18]. US is excellent for evaluation of the renal parenchyma and the collecting system to the ureteropelvic junction but it is limited in its evaluation of the ureter and of soft-tissue lesions within the collecting system. The use of renal US in the evaluation of suspected acute ureteral obstruction is also restricted because dilatation often does not develop for hours or even days. In these cases, US findings are normal in up to 50% of patients. US, either alone or combined with conventional radiography, has been compared with unenhanced CT. It has a much lower sensitivity: 24-77% [19-22] vs. 92-96% for CT. In Sheafor’s study, in which CT and US were compared [19], CT depicted 22 of 23 ureteral calculi (sensitivity 96%) whereas US depicted 14 of 23 (sensitivity 61%). Differences in sensitivity were statistically significant (p = 0.02). The specificity for each technique was 100%. CT can give a rapid and

106

definitive diagnosis of urinary calculus disease as well as other abdominal disorders with the same presentation. US identification of ureteral jets within the urinary bladder lumen is helpful for assessing the presence of obstruction. One study showed an absent ureteral jet in 11 of 12 patients with high-grade obstruction and in 3 of 11 patients with low-grade obstruction [23]. The identification of jets at the ureterovesical junctions indicates that obstruction is incomplete, a finding that may be used to guide therapy. In diuresis renography, radionuclides are injected to evaluate the urinary tract for obstruction. Since considerably less anatomical detail is available with this test than with other radiographic examinations, it is less useful in the acute setting than for follow-up or the evaluation of chronic urinary tract obstruction. Diuresis renography does have the advantage of yielding objective data regarding the physiological significance of hydronephrosis detected on imaging studies, and also allows evaluation of the function of each kidney. Administration of a diuretic, usually furosemide, augments the standard renogram and is useful in evaluating whether dilated urinary systems are functionally obstructed or not. Magnetic resonance urography (MRU) using rapid scanning techniques, such as half acquisition turbo spin echo or single-shot fast spin echo and 3D gradient echo contrast-enhanced sequences, is used for evaluation of the urinary tract. Following the administration of 250 mL of normal saline and 10 mg of furosemide, the kidneys and dilated ureters are very bright on T2-weighted images and their stable position allows for clear imaging of the level of obstruction (Table 3). Unfortunately, stones appear as signal voids and can be difficult to identify and measure on MRU. Small calculi (which account for the majority of symptomatic stones) are also difficult to detect with this method. Also, urothelial lesions, blood clots, and debris can mimic calculi. In patients with renal impairment due to ureteral obstruction [24, 25], non-contrast CT was found to be the best imaging modality to identify calculus causes of obstruction, while MRU was superior for identifying noncalculus causes. In patients with normal renal function, contrast-enhanced CT can identify the presence and cause of hydronephrosis in nearly all cases [26]. MRU is particularly helpful in delineating the anatomy in patients with urinary diversion to bowel conduits [27]. An abdominal radiograph is a reasonable initial test in patients who have a history of radiopaque calculi and

Table 3. Computed tomography urography protocol No oral contrast, patient supine, 10 mg IV furosemide Axial non-contrast enhanced abdomen Inject intravenous contrast agent (100 mL of 350 mgI/mL) Axial abdomen and pelvis at 75 s post-injection Axial abdomen and pelvis at 5 min post-injection Coronal reformatted images of 5-min delay exam Review using bone and soft tissue windows

Parvati Ramchandani, Julia R. Fielding

Table 4. Magnetic resonance urography protocol 250 mL IV saline: begin 15 min before imaging Patient supine, give 10 mg IV furosemide Coronal HASTE/SSFSE scout Axial abdomen HASTE/SSFSE Axial IP/OOP gradient echo image Coronal 3D TSE, HASTE or other fluid-bright sequence If GFR <3.0 or serum Cr >3.0: STOP If GFR >3.0 and suspicion for transitional cell carcinoma is high: coronal 3D gradient echo image such as LAVA, VIBE with fat saturation; 2-, and 5-min delay

acute flank pain that is similar to that of previous episodes. In the absence of such a history, abdominal radiography may not be of value as an initial examination [9]. If a ureteral calculus is present on CT but not clearly identifiable on the CT scout view, a conventional abdominal radiograph may be useful – especially if the stone is >4-5 mm in size and over 300 HU in density [28] – to follow the progress of the stone for management purposes. For radio-opaque calculi, confirmation of stone location during conservative therapy is best performed using plain films [29] (Table 4). In dealing with the pregnant patient with flank pain, fetal age and estimated radiation dose are of paramount importance. Right hydronephrosis is commonly encountered, as the enlarging uterus turns slightly to the right thus compressing the ureter. When an obstructing stone is suspected in pregnant patients, US should be performed first. Based on clinical findings, some urologists will place a stent in a pregnant patient with severe hydronephrosis. If more imaging information is needed from a patient in the first trimester, a limited IVU using a plain scout film followed by a 10-min post-infusion delayed film yields the least radiation. After 20-24 weeks, IVU becomes difficult to interpret because of the enlarging uterus and the developing fetal skeleton, such that CT should be considered [17]. The expected fetal dose is approximately 16 mSv, well below that expected to cause developmental anomalies.

Urinary Tract Infection Acute Pyelonephritis This is usually an ascending infection from the bladder, seen predominantly in females. Rarely, the source of infection is hematogenous bacteremia. Diagnosis is usually made on clinical grounds and with urine analysis. Imaging may be needed to detect complications or the sequela of pyelonephritis. When clinical pyelonephritis persists for more than 3 days after suitable antibiotic therapy has been initiated, imaging is recommended. In such cases, CT is the imaging technique of choice to evaluate the kidneys for possible complications of pyelonephritis, e.g., the development of an abscess. CT is also the most sensitive and

107

Urinary Tract Obstruction and Infection

specific test for detecting the changes of acute pyelonephritis and its complications. Typical CT findings of pyelonephritis include unilateral nephromegaly, renal striations, wedge-shaped defects, and perinephric inflammatory changes; detection of the latter usually requires contrast-enhanced images. Areas of liquefaction within the renal parenchyma indicate the development of a renal abscess. CT is more sensitive than US in detecting the development of a renal abscess and in assessing its extent. In males with a urinary tract infection (UTI) and/or suspected pyelonephritis, clinical and imaging evaluation for causes of UTI, such as epididymitis, orchitis, and prostatitis, may be helpful in management. Patients with a neurogenic bladder secondary to a spinal cord injury pose a difficult problem as the urine is usually colonized with bacteria. Development of systemic symptoms should prompt rapid imaging as these patients may not be sensate to pain and a devastating abscess can develop quickly [30]. Finally, in order to diminish radiation dose to pregnant patients, US with power Doppler may be attempted prior to CT to detect areas of aberrant blood flow. This approach also has been shown to be useful in children [31, 32]. Vesicoureteral reflux in children can lead to reflux nephropathy, which may not be detected until they are adults. The reflux of urine through the ducts of Bellini results in broad-based renal parenchymal scars that are centered over clubbed and blunted calyces. Changes of reflux nephropathy occur preferentially in the poles of the kidneys, with the upper pole being affected most frequently. The interpolar regions are almost always spared of scarring in these patients.

Emphysematous Pyelonephritis This life-threatening infection with a gas-producing organism may have a mortality of up to 90% without prompt treatment, and nephrectomy is often required. In diabetic patients, this infection is usually caused by a strain of Escherichia coli. The diagnosis of emphysematous pyelonephritis is made when gas is seen in the renal parenchyma. CT is the most accurate technique for diagnosing emphysematous pyelonephritis and for differentiating this entity from emphysematous pyelitis or perinephric emphysematous infections. CT is also the most accurate approach to differentiate localized gas-producing infections from diffuse emphysematous pyelonephritis; the former can be successfully treated with percutaneous drainage in combination with systemic antibiotic management.

Granulomatous Renal Infections Tuberculosis, xanthogranulomatous pyelonephritis (XGP), malacoplakia, and fungal infections can all affect the urinary tract. Renal tuberculosis is usually spread hematogenously from the lungs, seeding the kidneys. Symptomatic renal tuberculosis results from

secondary, reactivation tuberculosis. Symptoms typically include hematuria and sterile pyuria. The earliest signs of renal tuberculosis are, among others, focal papillary necrosis and inflammation of the calyces. With progression, areas of fibrosis and calcification may develop. Long-standing tuberculosis may result in numerous fibrotic strictures, ureteral wall thickening, hydronephrosis, and autonephrectomy. Pyelonephritis related to Bacillus-Calmette-Guerin (BCG) therapy for urothelial carcinoma has also been reported. XGP, an inflammatory condition with a marked female predominance, is associated with recurrent UTIs caused by proteases or by E. coli infection. An infection-based stone is seen in the majority of cases. The classic radiographic triad consists of reniform enlargement of the kidney, a renal stone, and markedly decreased or absent renal function in the affected kidney. Localized XGP occurs in 20% of patients and can mimic renal neoplasms on imaging studies. Both malacoplakia and fungal infections have nonspecific appearances on imaging, with diagnosis only being established by histological examination to exclude neoplasm. Malacoplakia constitutes congregations of histiocytes and is more commonly seen in the bladder and ureter than in the kidney. The microscopic hallmark of malacoplakia is the Michaelis-Gutman inclusion body, which is seen within the abnormal histiocytes. When malacoplakia involves the ureter or bladder, multiple submucosal masses are usually identified. Imaging findings are non-specific and tissue is required for definitive diagnosis. Fungal infections are usually seen in immunocompromised patients, including diabetics. Debris, often present within the renal collecting system, forms a “hand-in-glove” filling defect of the contrast-opacified calyces.

AIDS Nephropathy Autoimmune deficiency syndrome (HIV/AIDS) nephropathy comprises a variety of renal pathologies. Findings are generally non-specific but patients with HIV infection, renal failure, and hyperechoic nephromegaly likely have AIDS nephropathy. These sonographic findings in an AIDS patient usually indicate that the patient will develop irreversible renal failure.

BK Virus Nephropathy During the last decade, strong anti-rejection agents have been used to improve renal allograft survival. The resulting immunosuppression has allowed a generally benign virus, known as BK, to become a deadly pathogen. BK is present in the urine of 50% of transplant recipients 3 months post-operatively and may cause graft failure in 10%. The method of action of this small polyoma virus is unknown. Patients with transplanted kidneys who are BK-positive are treated with steroids. There is no known cure [33].

108

Pyonephrosis Pyonephrosis is a bacterial infection of the urine associated with ureteral obstruction. It is a medical emergency as it has high morbidity if untreated. Pyonephrosis is best diagnosed with US. Any febrile patient with hydronephrosis should be suspected of also having pyonephrosis. Findings that suggest pyonephrosis include echogenic urine and the presence of debris within the hydronephrotic calyces. Prompt drainage of the urinary tract accompanied by systemic antibiotic administration is essential in these patients. When obstruction and infection are due to ureteral calculi, retrograde ureteral catheterization and percutaneous nephrostomy are equally effective in relieving the obstruction and infection; neither technique is superior to the other in promoting rapid drainage or clinical defervescence [34].

Schistosomiasis Schistosomiasis of the urinary tract is caused by infection with Schistosoma hematobium. Infection is endemic in Egypt but also occurs in Southern and sub-Saharan Africa, South America, the Middle East, Southern China, and in Southeast Asia. With the global nature of today’s world, patients may be seen in developed western countries as well. This infection usually arises in the bladder but can spread to the ureters and kidneys. Dystrophic calcifications in the wall of the bladder and ureter are typical findings and are caused by calcification of the dead parasitic ova. Typical radiographic findings include mural calcifications, ureteral strictures, and vesicoureteral reflux. These patients have a markedly increased risk for the development of squamous cell carcinoma as well as transitional cell carcinoma of the urinary tract.

Post-therapeutic Changes One of the most common causes of urinary obstruction is pelvic cancer and its complications. Many patients with gynecological and prostatic malignancies are treated with surgery and radiation therapy, which may cause stricturing of the ureter. The ureter may be transected during hysterectomy or other pelvic surgery, leading to hydronephrosis and urinoma. Large lymphoceles are uncommon with these surgeries but if present may compress the ureter or bladder. Ureteric obstruction secondary to stricture or fibrosis may occur, as well as fistulae involving the bowel and urinary tract [35].

References 1. Farrell TA, Hicks MS (1997) A review of radiologically guided percutaneous nephrostomies in 303 patients. J Vasc Interv Radiol 8:769-774 2. Hoppe H, Studer R, Kessler TM et al (2006) Alternate or additional findings to stone disease on unenhanced computerized tomography for acute flank pain can impact management. J Urol 175:1725-1730; discussion 1730

Parvati Ramchandani, Julia R. Fielding

3. Katz DS, Scheer M, Lumerman JH et al (2000) Alternative or additional diagnoses on unenhanced helical computed tomography for suspected renal colic: experience with 1000 consecutive examinations. Urology 56:53-57 4. Ahmad NA, Ather MH, Rees J (2003) Incidental diagnosis of diseases on un-enhanced helical computed tomography performed for ureteric colic. BMC Urology 17:2 5. Sourtzis S, Thibeau JF, Damry N et al (1999) Radiologic investigation of renal colic: unenhanced helical CT compared with excretory urography. AJR Am J Roentgenol 172:1491-1494 6. Yilmaz S, Sindel T, Arslan G et al (1998) Renal colic: comparison of spiral CT, US and IVU in the detection ureteral calculi. Eur Radiol 8:212-217 7. Fielding JR, Silverman SG, Samuel S et al (1998) Unenhanced helical CT of ureteral stones: a replacement for excretory urography in planning treatment. AJR Am J Roentgenol 171:1051-1053 8. Catalano O, Nunziata A, Altei F et al (2002) Suspected ureteral colic: primary helical CT versus selective helical CT after unenhanced radiography and sonography. AJR Am J Roentgenol 178:379-387 9. Levine JA, Neitlich J, Verga M et al (1997) Ureteral calculi in patients with flank pain: correlation of plain radiography with unenhanced helical CT. Radiology 204:27-31 10. Chan-Tack KM, Truffa MM, Struble KA, Birnkrant DB (2007) Atazanavir-associated nephrolithiasis: cases from the US Food and Drug Administration’s Adverse Event Reporting System. AIDS 21:1215-1218 11. Ciaschini MW, Remer EM, Baker ME et al (2009) Urinary calculi: Radiation dose reduction of 50% and 75% at CT - effect on sensitivity. Radiology 251:105-111 12. Katz SI, Saluja S, Brink JA et al (2006) Dose associated with unenhanced CT for suspected renal colic: impact of repetitive studies. AJR Am J Roentgenol 186:1120-1124 13. Kirpalani A, Khalili K, Lee S, Haider MA (2005) Renal colic: comparison of use and outcomes of unenhanced helical CT for emergency investigation in 1998 and 2002. Radiology 236: 554-558 14. Ha M, MacDonald RD (2004) Impact of CT scan in patients with first episode of suspected nephrolithiasis. J Emerg Med 27:225-231 15. Kenney PJ (2003) CT evaluation of urinary lithiasis. Radiol Clin N Am 41: 979-999 16. LeRoy AJ (2006) Imaging of acute maternal diseases in pregnancy. Categorical course in diagnostic radiology at 92nd scientific assembly and annual meeting, Chicago, Illinois, RSNA. Genitourinary radiology syllabus 2006:271-279 17. Fielding JR, Washburn D (2005) Imaging the pregnant patient: a uniform approach. J Women’s Imaging 7:16-21 18. Middleton WD, Dodds WJ, Lawson TL, Foley WD (1988) Renal calculi: sensitivity for detection with US. Radiology 167:239-244 19. Sheafor DH, Hertzberg BS, Freed KS (2000) Nonenhanced Helical CT and US in the Emergency Evaluation of Patients with Renal Colic: Prospective Comparison. Radiology 217:792-797 20. Catalano O, Nunziata A, Alei F et al (2002) Suspected ureteral colic: primary helical CT versus selective helical CT after unenhanced radiography and sonography. AJR Am J Roentgenol 178:379-387 21. Fowler KAB, Locken JA, Duchesne JH, Williamson MR (2002) US for detecting renal calculi with nonenhanced CT as a reference standard. Radiology 222:109-113 22. Tublin ME, Dodd GD, Verdile VP (1994) Acute renal colic: diagnosis with duplex Doppler US. Radiology 193:697-701 23. Burge HJ, Middleton WD, McClennan BL, Hildebolt CF (1991) Ureteral jets in healthy subjects and in patients with unilateral ureteral calculi: comparison with color Doppler US. Radiology 180:437-442 24. Shokeir AA, El-Diasty T, Eassa W et al (2004) Diagnosis of ureteral obstruction in patients with compromised renal function: the role of noninvasive imaging modalities. J Urol 171:2303-2306

Urinary Tract Obstruction and Infection

25. Shokeir AA, El-Diasty T, Eassa W et al (2004) Diagnosis of noncalcareous hydronephrosis: role of magnetic resonance urography and noncontrast computed tomography. Urology 63:225-229 26. El-Ghar MEA, Shokheir AA, El-Diasty T et al (2004) Contrast enhanced spiral computerized tomography in patients with chronic obstructive uropathy and normal serum creatinine: a single session for anatomical and functional assessment. J Urol 172:985-988 27. Zielonko J, Studniarek M, Markuszewski M (2003) MR Urography of obstructive uropathy: diagnostic value of the method in selected clinical groups. Eur Radiol 13:802-809 28. Zagoria RJ, Khatod EG, Chen MYM (2001) Abdominal radiography after CT reveals urinary calculi: a method to predict usefulness of abdominal radiography on the basis of size and CT attenuation of calculi. AJR Am J Roentgenol 176:1117-1122 29. Assi Z, Platt JF, Francis IR et al (2000) Sensitivity of CT scout radiography and abdominal radiography for revealing ureteral calculi on helical CT: implications for radiologic follow-up. AJR Am J Roentgenol 175:333-337

109

30. Rubenstein JN, Schaeffer AJ (2003) Managing complicated urinary tract infections. The urologic view. Infect Dis Clin N Am 17:333-351 31. Dacher J, Pfister C, Monroc M et al (1996) Power Doppler sonographic pattern of acute pyelonephritis in children: comparison with CT. AJR Am J Roentgenol 166:1451-1455 32. Majd M, Nussbaum Blask AR, Markle BM et al (2001) Acute pyelonephritis: comparison with Tc99m-DMSA SPECT, spiral CT, MR imaging, and power Doppler US in an experimental pig model. Radiology 218:101-108 33. Wiseman AC (2009) Polyomavirus nephropathy: A current perspective and clinical considerations. Am J Kidney Dis 54:131-142 34. Pearle MS, Pierce HL, Miller GL et al (1996) Optimal method of urgent decompression of the collecting system for obstruction and infection due to ureteral calculi. J Urol 160:12601264 35. Yablon CM, Banner MP, Ramchandani P et al (2004) Complications of prostate cancer treatment: spectrum of imaging findings. Radiographics 24:S181-S194

IDKD 2010-2013

Benign Diseases of the Female Genital Tract Caroline Reinhold1, Rahel A. Kubik-Huch2 1 McGill

University Health Center, Montreal, Quebec, Canada of Radiology, Department of Medical Services, Kantonsspital Baden, Baden, Switzerland

2 Institute

Introduction Endovaginal sonography (EVS) remains the procedure of choice for the initial evaluation of benign diseases of the female genital tract. When EVS findings are indeterminate, further evaluation is typically performed with magnetic resonance imaging (MRI), due to its excellent softtissue differentiation, multiplanar capabilities, and absence of ionizing radiation. MRI is thus well suited for imaging women of reproductive age, particularly during pregnancy. Accordingly, the technique has come to play an increasing role in pelvimetry and, more recently, as an adjunct to sonography for fetal imaging. MRI is used in the pre-operative characterization of adnexal masses and as a problem-solving tool in benign uterine disease (for example, uterine malformations), adenomyosis, and to select appropriate candidates for therapies such as myomectomy and uterine embolization. The role of computed tomography (CT) is limited in the evaluation of benign disease of the female pelvis and is usually employed in an emergency situation, such as in an acute abdomen caused by ovarian torsion or pelvic inflammatory disease.

Anatomy of the Female Genital Organs In women of reproductive age, the uterus is approximately 6-9 cm in length and varies in appearance according to the menstrual cycle. The uterine zonal anatomy is best depicted using sagittal T2-weighted images (Fig. 1). In the pre-menopausal woman, three distinct zones are recognized: 1. the high-signal-intensity endometrium of varying thickness, depending on the menstrual cycle; 2. the hypointense junctional zone, anatomically corresponding to the innermost layer of the myometrium; 3. the outer layer of the myometrium, which is of intermediate signal intensity. Four zones are distinguished in the cervix by highresolution MRI: 1. the hyperintense mucous within the endocervical canal; 2. the cervical mucosa of intermediate to high signal intensity;

Fig. 1. Normal MR anatomy (sagittal T2-weighted image) of the uterus and cervix. Three zones are recognized in the uterus: the hyperintense endometrium, the hypointense junctional zone (arrow), and the outer layer of the myometrium of intermediate signal intensity. Four zones are distinguished in the cervix: the hyperintense mucous within the endocervical canal, the cervical mucosa, the hypointense cervical stroma, and an additional layer of smooth muscle. Scarring from C-section (arrowhead)

3. the hypointense cervical stroma surrounding the mucosa; 4. an additional layer of intermediate signal intensity in continuity with the uterine myometrium, representing smooth muscle. In post-menopausal patients, the uterine corpus, but not the cervix, regresses and decreases in size. In pre-menopausal women, the mean ovarian volume is 10 ± 6 mL, with an upper limit of 22 mL [1]. In postmenopausal women, the ovary atrophies and the follicles disappear over subsequent years. As a result of atrophy, the ovaries are reduced in volume (1-6 mL) and may be difficult to visualize sonographically [1, 2]. An ovary is abnormal if the volume is >8 mL or it is more than twice the volume of the contralateral ovary. Small (<3 cm) anechoic cysts can be seen in up to 15% of menopausal women [3].

111

Benign Diseases of the Female Genital Tract

Unlike sonography, MRI can detect the ovaries in almost 100% of patients [3]. On T2-weighted images, the cortex and stroma of a pre-menopausal ovary is of low signal intensity while the medulla is of higher signal intensity. In pre-menopausal women, gadolinium enhancement of the ovaries is less than that of the myometrium while in postmenopausal women, it is equivalent [4, 5].

Magnetic Resonance Imaging Technique If possible, patients should be scheduled for MRI in the second half of the menstrual cycle, since the thickness of the endometrial stripe increases during the follicular and secretory phases, allowing better appreciation of the normal zonal anatomy of the uterus. The objective of patient preparation is to obtain the best possible image quality, while making the examination as comfortable as possible for the patient. To minimize motion artifact induced by bowel peristalsis, patients are advised to fast for 6-8 h before the procedure. Unless contraindicated, the intravenous or intramuscular injection of peristaltic inhibitors, i.e., glucagon or butyl-scopolamine, is recommended to further decrease peristalsis artifacts. An empty urinary bladder minimizes ghosting artifacts from patient motion. In addition, it maintains the uterus in a more caudal position in the pelvis, away from small bowel loops, and assures that the normally visible fat plane between the uterus and urinary bladder – an important criteria to exclude tumor invasion of the bladder wall in oncological patients – will not be obliterated. Examinations are usually performed with the patient in supine position and using a body-flex phased-array MRI surface coil; pregnant patients in the third trimester may also be placed in an oblique or lateral decubitus position for greater comfort during imaging. Depending on the clinical questions to be answered, one or several fast sequences of the upper abdomen should be performed, e.g., a cine steady-state free precession sequence (TrueFisp) in the coronal and axial planes and a T1-weighted gradient echo sequence to exclude ascites, enlarged retroperitoneal lymph nodes, hydronephrosis, or renal agenesis in patients with congenital uterine malformations. T1-weighted as well as axial STIR (short tau inversion recovery) sequences are standard techniques in the assessment of the small pelvis. The presence of a bright mass on T1-weighted imaging requires an additional fat suppressed T1-weighted sequence using chemical presaturation to distinguish fat (e.g., in a mature teratoma of the ovary, which shows signal loss) from blood, mucin, or other proteinaceous material that remains of high signal intensity. A pre-contrast fat-saturated T2-weighted sequence in the axial or sagittal plane is also helpful to detect small endometriosis implants, e.g., in the cul-de-sac. T2-weighted images are most important in the assessment of the uterus, with sagittal sections best-suited to image the uterus. Oblique coronal and axial slice orientations

that are in accordance with the long and short axes of the endometrial cavity should be applied for the assessment of uterine and cervical pathologies. It is recommended that these oblique sequences are acquired under medical supervision, as the technician might not have sufficient anatomical knowledge to clearly identify the long axis of the uterine body. The short-axis coronal oblique sequence (perpendicular to the long-axis of the endometrial cavity) is particularly valuable for assessing localized endometrial pathology as well as the thickness of the junctional zone. It is also valuable for determining the extent of a uterine septum. A T2weighted sequence performed parallel to the long axis of the endometrial cavity is critical to characterize the external uterine contour in patients with mullerian duct anomalies. Since this series is so important in the classification of uterine anomalies, it is best performed early in the examination, prior to filling of the bladder, which with increasing distention often displaces the uterus. If the fundal contour is inadequately characterized on T2, then T1-weighted images parallel to the long axis can facilitate characterization of the external contour due to increased contrast between the myometrial fundal contour and the overlying fat. The uterus should be imaged using the smallest possible field of view (20-24 cm), with thin sections of 4-5 mm and the largest possible matrix size appropriate to each individual sequence. These imaging parameters provide important anatomical detail, which becomes critical when uterine anomalies and endometrial pathology are imaged. In the imaging of large leiomyomas, the section thickness and field-of-view (FOV) may need to be adjusted accordingly. However, when the myometrial origin of a subserosal leiomyoma must be established, thin sections at the level of the pedicle are frequently helpful. We have recently added an axial echo-planar diffusionweighted MRI sequence of the small pelvis to our clinical routine protocol. Diffusion-weighted imaging (DWI), well established for intracranial imaging, is a functional imaging technique whose contrast derives from the random motion of water molecules within the extracellular tissue. DWI is useful in pelvic female tumors, i.e., in the differentiation of benign from malignant lesions, and has a high sensitivity in the detection of iliac lymph nodes. Gadolinium-enhanced imaging is not needed for most benign conditions but can be useful in selected pathologies. We routinely perform contrast-enhanced imaging in patients with symptomatic leiomyomas scheduled for laparoscopic surgery or uterine artery embolization, in the latter case with additional magnetic resonance angiography of the iliac arteries. Vascularity can help to predict response to treatment. In addition, the enhancement pattern is one criterion that can be used to distinguish a benign fibroid from a malignant leiomyosarcoma.

Congenital Uterine Anomalies Buttram and Gibbons proposed a classification system in 1979 that was based on the degree of failure of normal de-

112

Caroline Reinhold, Rahel A. Kubik-Huch

I Hypoplasia/agenesis

(a) Vaginal

(c) Fundal

(b) Cervical

(d) Tubal (e) Combined

II Unicornuate

(a) Communicating

(c) No cavity

(b) Non Communicating

(d) No horn

IV Bicornuate

(a) Complete

VI Arcuate

V Septate

(a) Complete

III Didelphus

(b) Partial

VII DES drug related

(b) Partial

Fig. 2. Classification system of mullerian duct anomalies according to the American Fertility Society [7]

velopment while taking into account similar clinical features, reproductive outcomes, and management [6]. Modified by the American Fertility Society (now the American Society of Reproductive Medicine) in 1988, the classification system is the most widely accepted framework of reporting uterovaginal anomalies [7] (Fig. 2). Uterine malformations can be associated with subfertility, pregnancy wastage, and menstrual disorders.

communicating and communicating rudimentary horns are usually surgically resected. Hematometra and endometriosis, as well as the potential for pregnancy within the hypoplastic horn, may complicate the patient’s course in this setting. The appearance of the rudimentary horn is variable and depends on the presence of functioning endometrial tissue and whether or not the horn is obstructed (Fig. 3).

Class I: Mullerian agenesis and hypoplasia. In approximately 10% of uterine congenital anomalies, there is some degree of early failure to form the mullerian ducts prior to fusion [8]. Complete vaginal agenesis, or MayerRokitansky-Kuster-Hauser (MRKH) syndrome, is the most common presentation of the class I anomalies. In the group of patients with vaginal agenesis, 90% have associated uterine agenesis, while 10% present with a rudimentary uterus. The ovaries are normal in appearance and function, although they may be situated more cranially, outside of the pelvis. Vaginal agenesis is best diagnosed in the transverse plane with fatty and connective tissue in the expected location of the vagina, between the urethra and rectum.

Class III: uterus didelphys. In uterus didelphys, two separate uterine horns (often widely divergent) and two separate cervices are visualized with MRI. The normal zonal anatomy is maintained within each hemiuterus, and the two horns remain symmetrical in size.

Class II: unicornuate uterus. The unicornuate uterus appears elongated, curved, and is typically deviated to one side [8]. The endometrium frequently takes on a “banana” or “bullet” shape. Unicornuate uteri without rudimentary horns or those with non-cavitary rudimentary horns require no treatment. Those with both non-

Class IV: bicornuate uterus. In the bicornuate uterus, the horns are symmetrical in size, with an intervening cleft that extends to the internal cervical os in the “complete” bicornuate uterus and which terminates more proximally in the “incomplete” bicornuate uterus. The endocervical canal may be solitary (bicornuate unicollis) or duplicated (bicornuate bicollis). Communication of the endometrial and/or endocervical segments must be present to diagnose this class of anomaly. Communication between the segments allows differentiation of a bicornuate bicollis uterus from a uterus didelphys. The uterus didelphys has no communication between the endometrial segments. In the bicornuate uterus, the fundal cleft on MRI is greater than 1.0 cm. This feature distinguishes the bicornuate from the septate uterus [8-11].

113

Benign Diseases of the Female Genital Tract

a

b

Fig. 3 a, b. Uterine malformation: a axial and b coronal T2-weighted MR images. A unicornuate uterus with a rudimentary left horn (arrows) is seen

Class V: septate uterus. In the septate uterus, the external uterine contour can be convex, flat, or minimally concave, with the fundal cleft always <1.0 cm [8-12]. The fundal segment of the septum is isointense to myometrium in both partial and complete septa. In complete septa, the inferior segment of the septum is usually of low signal intensity on T2-weighted images, corresponding to the more fibrous component. This low-signal-intensity band is frequently absent in partial septa, however, which tend to be uniformly isointense to myometrium. It is important not to use the signal intensity of the septum to classify this anomaly since septal composition can overlap between the septate and bicornuate uteri.

Benign Conditions of the Vagina, Cervix, and Uterus Bartholin’s cysts are caused by retained secretions within the vulvo-vaginal glands, mostly as a result of chronic inflammation or trauma. They are located in the posterolateral parts of the lower vagina and vulva, whereas Nabothian cysts are retention cysts of the cervical glands and clefts. Gartner duct cysts represent remnants of the caudal end of the Wolffian or mesonephric ducts and are

a

Fig. 4 a, b. Multiple uterine leiomyomas. a Sagittal T2-weighted and b contrastenhanced T1-weighted images. The multiple lesions are sharply demarcated and present as hypointense lesions on the T2-weighted sequence. Contrast-enhancement is less pronounced than in the adjacent normal myometrium. A signal void is indicative of calcification (arrow)

typically located in the anterolateral wall of the vagina, above the level of the symphysis pubis. Endometrial polyps are among the most common pathological lesions of the uterine corpus. Patients with postmenopausal bleeding and endometrial polyps usually undergo endometrial sampling and polyp removal [13]. Endometrial polyps have a variable appearance at EVS but are typically echogenic, with an intact overlying endometrium or subendometrial halo. A vascular pedicle is usually identified at color/power Doppler imaging. On T2-weighted images, polyps present as masses that are slightly hypointense or isointense relative to the normal endometrium. Large polyps are frequently heterogeneous in signal intensity [14, 15]. On T2-weighted sequences, the fibrous core is seen as a hypointense area within the polyp. Endometrial polyps show a variable degree of enhancement after gadolinium administration, and the addition of gadoliniumenhanced sequences significantly improves the detection rate of endometrial polyps. Small polyps enhance early and are well delineated against the hypointense endometrial complex on early dynamic scans. In addition, a vascular stalk can frequently be identified during the arterial phase. Magnetic resonance imaging is useful for distinguishing leiomyomas (Fig. 4) from other myometrial pathology and

b

114

Caroline Reinhold, Rahel A. Kubik-Huch

solid pelvic masses, especially in patients with nondiagnostic or equivocal ultrasound findings. A mass of intermediate signal intensity on T1-weighted images, low signal intensity on T2-weighted images, and splaying the uterine serosa or myometrium allows the diagnosis of leiomyoma to be made with confidence. The presence of feeding vessels originating in the myometrium further supports the uterine origin of the mass (Fig. 5). However, if a mass is adjacent to the uterus and is of intermediate or high signal intensity relative to the myometrium on T2-weighted images, the differential diagnosis includes degenerated leiomyoma and extrauterine tumors (benign and malignant). In these patients, the diagnosis of leiomyoma should be reserved only for cases in which the uterine origin of the mass is firmly established. Occasionally, it may be difficult to distinguish a pedunculated subserosal leiomyoma from an ovarian fibroma, since both lesions may be hypointense on T2-weighted images. This distinction is likely not significant as the latter is rarely malignant [16]. Submucosal leiomyomas are usually distinguished from endometrial polyps by identifying their myometrial origin and by their low signal intensity on T2-weighted images.

a

Leiomyomas must also be differentiated from uterine adenomyosis (Fig. 6), although these conditions frequently coexist. Differentiating the two entities may be critical because uterine-conserving therapy is established for leiomyomas; whereas hysterectomy remains the definitive treatment for debilitating adenomyosis. While MRI is extremely accurate in making this distinction, especially in patients with diffuse adenomyosis, the imaging features of focal adenomyosis or adenomyomas can overlap with those of leiomyomas [17-19]. Imaging characteristics that favor the diagnosis of focal adenomyosis include: 1. a lesion with poorly defined margins; 2. a lesion that is elliptical in shape extending along the endometrium; 3. a lesion that has little mass effect upon the endometrium relative to its size; 4. a lesion with high signal intensity striations radiating from the endometrium into the myometrium [17-19]. Cystic adenomyosis needs to be differentiated from a leiomyoma with central hemorrhagic degeneration or a bicornuate uterus with an obstructed horn. Another entity that may mimic a leiomyoma or adenomyosis is a myometrial contraction. Myometrial contractions

b

Fig. 5 a, b. Subserosal leiomyoma: a axial T2weighted and b sagittal contrast-enhanced T1-weighted sequences. A large, relatively hyperintense mass is seen adjacent to the uterus. The presence of feeding vessels (arrow) originating in the myometrium supports the uterine origin of the mass. A smaller hypointense intramural leiomyoma is seen in the uterine fundus

a

b

c

Fig. 6 a-c. Focal adenomyosis: a sagittal T2-weighted and b, c axial T2-weighted STIR images. Enlarged uterus with diffusely thickened junctional zone with multiple hyperintense foci indicative of adenomyoma. A small leiomyoma is also present (arrow)

115

Benign Diseases of the Female Genital Tract

are transient and usually change or resolve over the course of the exam. Contractions image as hypoechoic or low signal intensity lesions within the myometrium that deform the endometrium while sparing the outer uterine contour. Malignant degeneration of a leiomyoma is a rare occurrence. Unfortunately, echogenicity or signal characteristics do not reliably distinguish a benign leiomyoma from a leiomyosarcoma. However, if a leiomyoma suddenly enlarges, especially after menopause, and/or has an irregular or indistinct border, the possibility of sarcomatous transformation should be raised.

Non-neoplastic Ovarian/Adnexal Masses Follicular cysts are the most common benign ovarian masses. Other non-neoplastic adnexal lesions include hydrosalpinx, peritoneal inclusion cyst, endometrioma, adnexal torsion, and tubo-ovarian abscess. These nonneoplastic masses should be distinguished from benign and malignant ovarian neoplasms. Paraovarian cysts, also known as paratubal cysts, account for 10-20% of adnexal masses. They are found within the broad ligament or paraovarium and arise from mesonephric (wolffian)/paramesonephric (mullerian) structures or mesothelial inclusions.The hydatid cyst of Morgani is the most common paramesonephric cyst and arises from the fimbrial end of the fallopian tube [20]. On imaging, paraovarian cysts have the typical appearance of a cyst, although they can vary in size and contain complex contents secondary to hemorrhage. In pre-menopausal women, peritoneal fluid is produced by normally functioning ovaries and is reabsorbed by the mesothelial cells of the peritoneal cavity. In patients with adhesions composed of mesothelial and

a

fibrous strands, either from previous surgery or secondary to pelvic inflammatory disease, endometriosis, or trauma, the peritoneal fluid may not be reabsorbed; Instead, it accumulates and entraps the ovary, forming a peritoneal inclusion cyst [21]. On imaging, peritoneal inclusion cysts are multiloculated cystic adnexal masses. The diagnostic finding is the presence of an intact ovary amid septations and fluid. Hydrosalpinx is the result of obstruction of the fimbriated end of the fallopian tube and dilatation of its ampullary and infundibular portions. The cause of obstruction includes pelvic inflammatory disease, endometriosis, adjacent tumors, and adhesions from prior surgery. On imaging, a hydrosalpinx appears as a fluid-filled tubular structure with a somewhat folded configuration and welldefined walls. It may contain folds that simulate an incomplete thick septum. The prevalence of endometriosis in all women is estimated to be 5-10%. It is defined as the presence of functioning endometrial glands and stroma outside the uterus. The most common sites of involvement, in order of decreasing incidence, are: the ovaries, cul-de-sac, posterior uterine wall, uterosacral ligaments, anterior uterine wall, and bladder dome. Other sites include the sigmoid colon, fallopian tubes, and distal ureters. Endometrioma is usually a term reserved for ovarian involvement. On EVS, endometriomas typically appear as well-defined unilocular or multilocular cystic masses with diffuse, homogeneously dispersed low-level internal echoes. A fluid-fluid level and a thickened wall with calcifications are sometimes present. On MRI, due to their hemorrhagic contents, endometriomas are typically high in signal intensity on T1-weighted images with and without fat saturation. They demonstrate “shading” or a gradient of low signal intensity on T2-weighted images, reflecting the shortening of T2 due to blood (Fig. 7). Based on the

b

Fig. 7 a-c. Endometrioma: axial T1-weighted image with (a) and without (b) fat saturation; c sagittal T2-weighted image. There is a large lesion of the right ovary with bright signal on T1-weighted images with and without fat saturation and low signal intensity on T2-weighted imaging. Small hyperintense follicular cysts adjacent to this lesion are seen on the sagittal image

c

116

Caroline Reinhold, Rahel A. Kubik-Huch

criteria of multiple adnexal, masses of very high signal intensity on T1-weighted images or any mass with very high signal intensity on T1-weighted images and low signal intensity (shading) on T2-weighted images, the overall sensitivity, specificity, and accuracy for diagnosing endometriomas with MRI is 90, 98, and 96%, respectively [22]. The MRI appearance of solid fibrotic endometriosis has also been described: masses of intermediate signal intensity studded with high-signal-intensity foci on T1-weighted images, low-signal-intensity masses on T2-weighted images, and enhancement following intravenous contrast [23]. Some fibrotic solid masses have small foci of hyperintensity on T2-weighted images, reflecting embedded endometrial glands.

Neoplastic Ovarian Masses Epithelial tumors are the most common histological type, comprising 60% of all ovarian neoplasms and >85-90% of ovarian malignancies. Serous tumors are the most common ovarian epithelial tumors. On imaging, serous cystadenomas are unilocular, thin-walled, large cystic masses that may contain thin septations and occasionally papillary projections [24]. Mucinous tumors are the second most common epithelial tumor. On imaging, mucinous cystadenomas typically present as large cystic masses (up to 30 cm) with multiple thin septations and lowlevel echoes/T1-T2 shortening in the dependent portions due to mucoid material. This results in the typical “stained glass” appearance on MRI [25]. Papillary projections are less frequently seen than in serous cystadenomas. Rupture of the tumor capsule can result in pseudomyxoma peritonei. Mature cystic teratomas contain derivatives of at least two of the three germ layers; ectoderm, mesoderm, and endoderm. Ectodermal elements, however, tend to predominate, thus the term “dermoid cyst” was adopted. In 10% of the cases, mature cystic teratomas are bilateral. They are usually asymptomatic except when complications, usually torsion, occur. Rupture is uncommon but its

a

b

occurrence can cause chemical peritonitis. In 1-2% of patients, usually in older women, there is malignant transformation to squamous carcinoma. On EVS, mature cystic teratomas can have variable appearances depending on the tumor contents; nonetheless, several specific features have been described. A “dermoid plug” or “Rokitansky” protuberance can be seen as an echogenic mural nodule in a predominantly cystic mass. It is composed of hair, teeth, and fat and causes acoustic shadowing. The “dermoid mesh” is another specific sign; it refers to hair fibers that appear as linear hyperechogenic interfaces floating in the cyst [26]. On MRI, T1-weighted imaging with and without fat saturation techniques can establish the presence of fat in a cystic teratoma. The adipose tissue will have high signal intensity on T1 without fat suppression and will show signal loss on fat-suppressed images, thus excluding the diagnosis of an endometrioma or hemorrhagic cyst (Fig. 8). Chemical shift imaging using in- and out-of-phase imaging is helpful in identifying tumors that have only a tiny amount of fat [27]. Fibrothecomas account for only 1% of all ovarian neoplasms; however, they are the most common solid benign tumors affecting the ovary. They are derived from stromal cells; because the histological appearance of fibromas and thecomas overlap, the term “fibrothecoma” is often applied to this spectrum of tumors. Like granulose cell tumors, they are hormonally active. Pure thecomas are composed predominantly of theca cells and are most common in peri-menopausal and post-menopausal women. In 15% of patients, there is co-existing endometrial hyperplasia, and up to 30% of patients have endometrial carcinoma. In contrast, pure fibromas are composed predominantly of fibroblasts and are most common in women under age 50, most of whom are asymptomatic. Meigs syndrome, a combination of ovarian fibroma, ascites, and right pleural effusion, is rare. On EVS, fibrothecomas present as hypoechoic solid masses with marked posterior sound attenuation as a result of homogeneous fibrous tissue. The MRI appearance of fibromas and thecomas are similar: intermediate signal intensity on T1-weighted images and very low

c

Fig. 8 a-c. Ovarian dermoid: a axial T2-weighted, b T1-weighted, and c T1-weighted fat suppressed images. A right adnexal mass with a fluid center and a rim that is hyperintense on the T1-weighted image; the mass shows signal loss on the fat suppressed image, indicating the presence of fat

117

Benign Diseases of the Female Genital Tract

a

Table 1. Reference values for magnetic resonance pelvimetry based on 100 women undergoing spontaneous vaginal delivery (from [30]) Reference values ± SD (cm) Obstetric conjugate Sagittal outlet Interspinous diameter Intertuberous diameter Transverse diameter

b

Fig. 9 a, b. Ovarian fibrothecoma: a sagittal T2-weighted image and b macroscopic specimen. The hypointense adnexal mass ventral to the uterus was histologically confirmed as ovarian fibrothecoma. A small submucosal and a larger subserosal leiomyoma are seen in the uterus

signal intensity on T2-weighted images [28] (Fig. 9). Due to these signal characteristics, fibrothecomas may be difficult to distinguish from pedunculated fibroids or Brenner tumors. Multiple imaging planes combined with the observation of small follicular cysts surrounding the mass should establish the diagnosis of a solid ovarian neoplasm. The presence of a thickened endometrium secondary to estrogen production favors the diagnosis. Larger fibrothecomas may have high signal intensity secondary to intratumoral edema, myxomatous change, or cyst formation. Following intravenous gadolinium, fibrothecomas show little enhancement.

Magnetic Resonance Imaging in Obstetrics There is no clinical or experimental evidence of teratogenic or other adverse fetal effects from the use of MRI in pregnancy. The technique is thus well suited in the imaging of pregnant women for maternal, e.g., for the

12.2 ± 0.9 11.6 ± 1.0 11.2 ± 0.8 12.1 ± 1.1 13.0 ± 0.9

work-up of colicky abdominal pain or suspected appendicitis, or fetal reasons. Magnetic resonance pelvimetry is indicated in pregnant women with a history of pelvic trauma, previous cesarean section, or in women who desire a trial of labor when the fetus is in breech presentation. The groundwork in pelvimetry was conducted using conventional radiography, with measurement of the various parameters on lateral and anterior-posterior views using various techniques to correct the distortion due to differing distances from the film. In recent years, X-ray pelvimetry has been steadily replaced by MRI, which offers the benefit of accurate measurement of the maternal pelvic dimensions without exposure of the patient to ionizing radiation. Magnetic resonance pelvimetric reference values in a large study population, stratified by delivery modality, have been established by a member of the author’s own group [29, 30]. The results are shown in Table 1. In addition, it was demonstrated that the pelvimetric parameters associated with the largest intra- and interobserver error and intraindividual variability are the intertuberous distance and the sagittal outlet – a finding that must be taken into careful account by obstetric decision-makers. In conclusion, technical advances in ultrafast MRI have revolutionized our ability to image the fetus. MRI is most likely to soon become an important tool for intrauterine fetal imaging as an operator-independent supplement to prenatal ultrasound [31, 32].

References 1. Cohen HL, Tice HM, Mandel FS (1990) Ovarian volumes measured by US: bigger than we think. Radiology 177:189-192 2. Goswamy RK, Campbell S, Royston JP et al (1988) Ovarian size in postmenopausal women. Br J Obstet Gynaecol 95:795-801 3. Levine D, Gosink BB, Wolf SI et al (1992) Simple adnexal cysts: the natural history in postmenopausal women. Radiology 184:653-659 4. Outwater EK, Talerman A, Duncon C (1996) Normal adnexa uteri specimens: anatomic basis of MR imaging features. Radiology 201:751-755 5. Outwater EK, Mitchell DG (1996) Normal ovaries and functional cysts: MR appearance. Radiology 198:397-402 6. Buttram VC, Gibbons, WE (1979) Mullerian anomalies: a proposed classification (an analysis of 144 cases). Fertil Steril 32:40-46 7. The American Fertility Society (1988) The American Fertility Society classifications of adnexal adhesions, distal tubal ob-

118

8. 9. 10. 11.

12. 13. 14. 15. 16. 17. 18. 19.

Caroline Reinhold, Rahel A. Kubik-Huch

struction, tubal occlusion secondary to tubal ligation, tubal pregnancies, mullerian anomalies and intrauterine adhesions. Fertil Steril 49:944-955 Trojano RN (2003) Magnetic resonance imaging of mullerian duct anomalies of the uterus. Top Magn Reson Imaging 14: 269-279 Carrington BM, Hricak H, Nuruddin RN et al (1990) Mullerian duct anomalies: MR imaging evaluation. Radiology 176:715-720 Pellerito JS, McCarthy SM, Doyle MB et al (1992) Diagnosis of uterine anomalies: relative accuracy of MR imaging, endovaginal ultrasound, and hysterosalpingography. Radiology 183:795-800 Junqueira BLP, Allen LM, Spitzer RF et al (2009) Mullerian duct anomalies and mimics in children and adolescents: Correlative intraoperative assessment with clinical imaging. Radiographics 29:1085-1103 Homer HA, Li TC, Cooke ID (2000) The septate uterus: a review of management and reproductive outcome. Fertil Steril 73:1-14 Sheth S, Hamper UM, Kurman RJ (1993) Thickened endometrium in the postmenopausal woman: sonographic-pathologic correlation. Radiology 187:135-139 Pertl B, Lahousen M, Pieber D et al (1996) Stellenwert der Sonograhie bei der Früherkennung des Endometriumkarzinoms. Gynäkologisch-geburtshilfliche Rundschau 36:14-20 Grasel RP, Outwater EK, Siegelman ES et al (2000) Endometrial polyps: MR imaging features and distinction from endometrial carcinoma. Radiology 214:47-52 Hricak H, Tscholakoff D, Heinrichs L et al (1986) Uterine leiomyomas: correlation of MR histopathologic findings, and symptoms. Radiology 158:385-391 Reinhold C, McCarthy S, Bret PM et al (1996) Diffuse adenomyosis: Comparison of endovaginal US and MR imaging with histopathologic correlation. Radiology 199:151-158 Togashi K, Ozasa H, Konishi I et al (1989) Enlarged uterus: Differentiation between adenomyosis and leiomyoma with MR imaging. Radiology 171:531-534 Mark AS, Hricak H, Heinrichs LW et al (1987) Adenomyosis and leiomyoma: Differential diagnosis with MR imaging. Radiology 163:527-529

20. Kurman RJ (1987) Blaustein’s pathology of the female genital tract. 3rd edn. Springer-Verlag, New York 21. Levy AD, Arnaiz J, Shaw JC, Sobin LH (2008) From the archives of the AFIP: Primary peritoneal tumors: Imaging features with pathologic correlation. Radiographics 28:583-607 22. Togashi K, Nishimura K, Kimura I et al (1991) Endometrial cysts: diagnosis with MR imaging. Radiology 180:73-78 23. Hottat N, Larrousse C, Anaf V et al (2009) Endometriosis: Contribution of 3.0-T pelvic MR imaging in preoperative assessment – initial results. Radiology 253:126-134 24. Thomassin-Naggara I, Bazot M, Darai E et al (2008) Epithelial ovarian tumors: Value of dynamic contrast-enhanced MR imaging and correlation with tumor angiogenesis. Radiology 248:148-159 25. Togashi K (2003) MR imaging of the ovaries: normal appearance and benign disease. Radiol Clin N Am 41:799-811 26. Malde HM, Kedar RP, Chadha D et al (1992) Dermoid mesh: a sonographic sign of ovarian teratoma. Letter Am J Roentgenol 159:1349-1350 27. Outwater EK, Siegelman ES, Hunt JL (2001) Ovarian teratomas: tumor types and imaging characteristics. Radiographics 21:475-490 28. Troiano RN, Lazzarini KM, Scoutt LM et al (1997) Fibroma and fibrothecoma of the ovary: MR imaging findings. Radiology 204:795-798 29. Michel SC, Rake A, Treiber K et al (2002) MR obstetric pelvimetry: effect of birthing position on pelvic bony dimensions. AJR Am J Roentgenol 79:1063-1067 (Abstract and editorial comment in Obstetrical&Gynecological Survey 2003; 54:238-239) 30. Keller TM, Rake A, Michel SCA et al (2003) Obstetric MR pelvimetry: Reference values and evaluation of inter- and intraobserver error and intraindividual variability. Radiology 227:37-43 31. Keller TM, Rake A, Michel SCA et al (2004) MR assessment of fetal lung development using lung volumes and signal intensities. Eur Radiol 14:984-989 32. Huisman TA, Wisser J, Martin E et al (2002) Fetal magnetic resonance imaging of the central nervous system. Eur Radiol 12:1952-1961

IDKD 2010-2013

Malignant Diseases of the Female Genital Tract Evis Sala1, Susan Ascher2 1 Department 2 Department

of Radiology, University of Cambridge and Addenbrookes Hospital, Hills Road, Cambridge, United Kingdom of Radiology, Georgetown University Hospital, Washington, DC, USA

Introduction

Patient Preparation, Positioning, and Coil Selection

Advances in cross-sectional imaging have led to an increasingly important role for radiology in the management of malignant gynecological conditions. A number of imaging modalities can be used to evaluate malignant diseases of the female pelvis, including ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography/computed tomography (PET/CT). These modalities have different roles in screening, diagnosis, staging, treatment selection and follow-up. The aim of this chapter is to review the role of different techniques in the imaging of malignant gynecological conditions. The emphasis is on the use of MRI in the staging of endometrial and cervical cancer following the revised FIGO (International Federation of Gynecology and Obstetrics) criteria, implemented beginning June 1, 2009 [1].

Patient preparation and positioning are very important to obtain optimal imaging results. Patients are usually instructed to fast for 4-6 h before the MRI examination in order to limit artifacts due to small-bowel peristalsis. An anti-peristaltic agent (hyoscine butylbromide or glucagon) may be administered before imaging as an alternative. Ideally, the patient is asked to empty the bladder prior to examination, as a full bladder may degrade images due to ghosting and motion artifacts. Patients are imaged in the supine position using a pelvic surfacearray multichannel coil; a cardiac coil usually offers the best image quality [2].

Ultrasound The primary imaging modality in the initial assessment of suspected gynecological pathology is US. It is used to evaluate a suspected pelvic mass, characterize adnexal lesions, and identify endometrial abnormalities in the postmenopausal patient. Transabdominal and transvaginal US can assist in image-guided fine needle aspiration cytology or biopsy and can also be used to guide placement of brachytherapy seeds in the treatment of cervical and endometrial cancer.

Magnetic Resonance Imaging This is the imaging modality of choice in the staging of uterine and cervical cancer and in the characterization of adnexal lesions when the US findings are indeterminate. The advantages of MRI include superb spatial and tissue contrast resolution, the absence of ionizing radiation, its multiplanar capability, and its fast techniques. However, the optimization of MRI sequences and clinical protocols, as outlined below, is crucial to ensure best results.

Choice of Sequences and Imaging Planes The basic imaging protocol for gynecological MRI consists of T1-weighted images in the axial plane and T2weighted images in the axial, sagittal, and coronal planes. T1-weighted axial images with a large field of view to evaluate the entire pelvis and upper abdomen for lymphadenopathy as well as bone marrow changes are essential in staging gynecological malignancies. High-resolution axial oblique T2-weighted fast spin-echo (FSE) images taken parallel to the short axis of the uterine corpus are favored for the evaluation of primary tumor and depth of myometrial invasion [3] in endometrial carcinoma, whereas axial oblique T2-weighted FSE (parallel to the short axis of the cervix) is crucial in assessing parametrial invasion in patients with cervical cancer. Dynamic multiphase contrast-enhanced 3D T1-weighted sequences through the uterus in the sagittal and axial (parallel to the short axis of the uterine corpus) planes are routinely used to improve staging accuracy in endometrial cancer [4, 5]. They are also useful for the evaluation of complex adnexal lesions, as they may help differentiate solid components or papillary projections from clots and debris. Dynamic imaging is not necessary. Diffusionweighted imaging (DWI) may be useful in differentiating benign from malignant endometrial lesions [6] and can provide valuable information for pre-operative staging in patient with endometrial and cervical carcinoma [7].

120

Evis Sala, Susan Ascher

DWI can help in the evaluation of tumor response to radiotherapy in patients with cervical cancer [8] and is useful in the detection of peritoneal implants and metastatic lymph nodes in patients with gynecological malignancies [9]. DWI is also useful in detecting peritoneal implants in patients with gynecological malignancies [10]. Ultrasmall particles of iron oxide (USPIO) have been shown to improve the detection of lymph node metastases independent of node size in patients with endometrial and cervical cancer [10].

Computed Tomography The role of CT in the imaging of malignant uterine conditions is limited due to its poor soft-tissue contrast. The main role of CT is in staging, treatment planning, and follow-up of patients with ovarian cancer. However, CT is important in the evaluation of other gynecological malignancies by identifying enlarged lymph nodes and distant metastases and detecting recurrent pelvic tumors.

Positron Emission Tomography/Computed Tomography Patients with malignant gynecological conditions are increasingly being evaluated with PET/CT. This modality is very valuable in the detection of metastatic lymph nodes as it has better sensitivity and specificity than MRI; it can also differentiate tumor recurrence from radiation fibrosis. PET/CT is also very useful in evaluating recurrent tumor prior to salvage therapy. Maximum standard uptake values (SUV) at staging can predict survival in patients with cervical carcinoma. However, it must be remembered that in pre-menopausal patients physiological uptake can be seen in the uterus, ovarian follicles, and corpus luteum cysts. The uptake of 2-deoxy-2-[fluorine18]fluoro-d-glucose (FDG) can also be seen in certain benign ovarian and uterine tumors as well as in inflammatory and infectious processes.

a

Endometrial Carcinoma On US, endometrial carcinoma is seen as a thickened endometrium (>5 mm in post-menopausal patients). On sonohysterography, endometrial carcinoma may present as an intrauterine polypoid mass or as an asymmetrical thickening of the endometrium. It is, however, impossible to distinguish between benign endometrial polyps, endometrial hyperplasia, and endometrial carcinoma confined to the endometrium using US alone. Therefore, endometrial carcinomas are typically diagnosed at endometrial biopsy or dilatation and curettage, with MRI being reserved to evaluate the extent of disease [11]. Imaging criteria for staging of endometrial cancer are based on the TNM/FIGO (International Federation of Obstetrics and Gynecology) classification. The overall staging accuracy of MRI has been reported to be between 85 and 93% [4, 5, 12]. Routine use of dynamic intravenous contrast enhancement is necessary for state-ofthe-art MRI evaluation of endometrial carcinoma [4, 5]. Stage IA tumors involve <50% of the myometrial thickness (Fig. 1). The presence of low-signal-intensity tumor (equilibrium and later phases of enhancement) within the outer myometrium or beyond indicates deep myometrial invasion and thus stage IB disease. Erroneous MRI assessment of the depth of myometrial invasion may occur due to an indistinct zonal anatomy, the presence of coexistent benign pathology (leiomyomas, adenomyosis), tumor extension into the uterine cornu, or a polypoid tumor distending the uterus so that rather than deep myometrial infiltration there is a thin rim of myometrium stretched over the tumor [2]. In stage II disease, the fibrocervical stroma is disrupted by high-signal-intensity tumor on T2-weighted images, together with the disruption of normal enhancement of the cervical mucosa by low-signal-intensity tumor on late dynamic contrast-enhanced MRI. In stage III disease, tumor extends outside the uterus but not outside the true pelvis. Stage IIIA is marked by

b

Fig. 1 a, b. Stage IA endometrial carcinoma. Sagittal T2-weighted fast spin echo (FSE) (a) and sagittal gadolinium-enhanced 3D T1weighted gradient-recalled echo (GRE) (b) images show an endometrial carcinoma invading the inner half of the myometrium. The depth of myometrial invasion is better appreciated on the gadolinium-enhanced 3D T1weighted GRE MR image (arrow)

Malignant Diseases of the Female Genital Tract

parametrial involvement, in which there is disruption of the serosa with direct tumor extension into the surrounding parametrial fat. In stage IIIB disease, tumor extends into the vagina and there is segmental loss of the lowsignal-intensity vaginal wall. In stage IIIC disease, lymphadenopathy is present. Tumor that extends beyond the true pelvis or invades the bladder or rectum constitutes stage IV disease. A loss of the low signal intensity of the bladder or rectal wall indicates stage IVA disease whereas stage IVB disease includes distant metastasis, malignant ascites, or peritoneal deposits. CT is valuable in detecting upper abdominal lymphadenopathy and distant metastases in patients with advanced endometrial carcinoma. PET/CT is useful in assessing nodal disease and distant metastases and has a role in monitoring treatment response in these patients.

Cervical Carcinoma The main role for imaging is staging a biopsy-proven cervical carcinoma. MRI is the best single imaging investigation and can accurately determine tumor location (exophytic or endocervical) and size, the depth of stromal invasion, and extension into the lower uterine segment [13, 14]. On T1-weighted images, tumors are usually isointense with the normal cervix and may not be visible. On T2-weighted images, cervical cancer appears as a mass of intermediate signal intensity and is easily distinguishable from the low signal intensity of the cervical stroma. MRI is recommended in evaluating cervical carcinoma patients with clinical stage IB disease or greater when the primary lesion is >2 cm, because of the relatively high likelihood of parametrial invasion and/or lymph node metastases [13, 15]. The staging accuracy of MRI ranges from 75 to 96%. The reported sensitivity of MRI in the evaluation of parametrial invasion is 69% and the specificity 93% [13, 15]. Although the FIGO system does not include radiology in the staging of cervical cancer, the revised FIGO staging criteria for cervical carcinoma [1], implemented as of June 1, 2009, encourage the use of imaging techniques, if available, to assess important prognostic factors such as parametrial and pelvic side wall invasion, tumor size, and lymph node metastases. Imaging is therefore complimentary to the clinical assessment. The FIGO committee made three changes that impact on radiology vs MRI of the pelvis. 1. The use of diagnostic imaging techniques to assess the size of the primary tumor is now encouraged by FIGO but is still not mandatory. MRI is highly accurate in measuring tumor size, which can affect prognosis and treatment. 2. Stage IIA has been subdivided: Stage IIA1 consists of tumors without parametrial invasion that are ≤4 cm in diameter and involve less than the upper two-thirds of the vagina. Stage IIA2 tumors are without parametrial invasion, >4 cm in diameter, and involve less than the upper two-thirds of the vagina.

121

3. Examination under anesthesia (EUA), which includes a cystoscopy and proctoscopy, is now recommended rather than mandatory, as previously stated. MRI has excellent imaging capabilities in assessing tumor involvement of the bladder and of the rectosigmoid, therefore avoiding an additional invasive, timely, and costly EUA procedure. This is especially helpful in clinical stage I and II disease, when the potential for invasion and hence upgrading of the tumor staging after EUA is low. The most important issue in the staging of cervical cancer is to distinguish early disease (stage IIA1 and below), which is treated with primary surgery, from advanced disease, which is treated with radiation, either alone or in combination with chemotherapy. The exception is stage IB2, in which the lesion is >4 cm in diameter and is treated as for advanced disease. Stage IA and its subdivisions are defined as microinvasive tumors that cannot be reliably demonstrated on T2weighted images. However, dynamic contrast-enhanced MRI may detect microinvasive disease as a focally enhanced area seen in the early dynamic phase. In distinguishing deep invasion (>3 mm) from superficial disease, the accuracy of T2-weighted MRI, dynamic MRI, and contrast-enhanced T1-weighted imaging is 76, 98, and 63%, respectively [16]. Stage IB is defined as clinically visible lesions limited to the cervix uteri and is subdivided into stage IB1, in which lesions are <4 cm in their greatest dimension, or stage IB2, in which lesions are >4 cm in their greatest dimension. The carcinoma appears as a mass of high signal intensity in contrast to the low signal intensity of the cervical stroma on T2-weighted images. Young women with stage IA or small IB (<2 cm) tumors who wish to retain their fertility may be considered for trachelectomy, in which the cervix is excised but the uterine body and hence fertility is preserved. MRI is highly accurate in predicting myometrial invasion, with a sensitivity and specificity of 100 and 99% respectively. For internal os involvement, the sensitivity of MRI is 90% and the specificity 98%. Stage IIA is defined as a tumor that invades the upper two-thirds of the vagina without parametrial invasion. Segmental disruption of the hypointense vaginal wall is demonstrated on T2-weighted images. When the tumor extends beyond the uterus, with parametrial invasion, it is defined as stage IIB. Spiculated irregular tumor/ parametrial interface, soft-tissue extension into the parametria, or encasement of the peri-uterine vessels is required to confidently diagnose parametrial invasion (Fig. 2). MRI has a specificity and negative predictive value of 97 and 100%, respectively, in evaluating parametrial invasion [3]. An important pitfall is the overestimation of parametrial invasion on T2-weighted images of large tumors (accuracy of 70%) compared to smaller tumors (accuracy 96%). Large tumors can cause stromal edema by tumor compression or inflammation – a fact that must be considered when making treatment decisions in these patients.

122

a

Evis Sala, Susan Ascher

b

Fig. 2 a, b. a Stage IIB cervical carcinoma. Sagittal T2-weighted image demonstrates an intermediate signal intensity mass centred on the posterior fornix (white arrow). b Axial oblique T2-weighted image demonstrates parametrial invasion with spiculated irregular tumor to parametrium interface on the right (black arrows)

In stage IIIA, the tumor involves the lower third of the vagina without extending to the pelvic side wall (>3 mm from pelvic side wall). When the tumor extends to the pelvic side wall (pelvic musculature or iliac vessels) or causes hydronephrosis, it is defined as stage IIIB. If the tumor invades the bladder or rectal mucosa it is stage IVA. There is segmental disruption of the low signal intensity of the bladder or rectal wall or segmental thickening of the rectal wall. Prominent strands between the tumor and the rectal wall may also indicate rectal invasion. MRI can confidently exclude bladder or rectal involvement, with a negative predictive value of 100% [17]. Distant metastases define stage IVB disease. Although pelvic lymph node metastases do not change the FIGO stage, para-aortic or inguinal lymph node metastases are also defined as stage IVB. CT has a limited role in staging cervical cancer due to its low accuracy in the detection of early parametrial extension [18]. However, CT has a diagnostic accuracy of approximately 90% in staging advanced cervical carcinoma and is very useful in evaluating the presence of distant metastases [19]. PET/CT allows the identification of involved nodes when CT findings are negative, resulting in a change in management in up to 25% of patients. In the detection of recurrent cervical cancer, it has a reported sensitivity, specificity, and accuracy of 90.3, 81.0, and 86.5, respectively. This is especially valuable to exclude the presence of distant disease prior to pelvic exenteresis [20].

Ovarian Carcinoma Ultrasound enables the detection and characterization of adnexal masses but has no role in staging. It can guide biopsy of adnexal or peritoneal masses in patients deemed unsuitable for primary surgery. CT is currently the modality of choice in staging ovarian cancer and can also be used to guide the biopsy of peritoneal or adnexal disease. CT provides information on the primary tumor,

the site and size of peritoneal deposits, and the presence of enlarged lymph nodes and ascites (Fig. 3). This information stratifies those patients with non-resectable disease, for whom neoadjuvant chemotherapy would be beneficial, from those patients who should undergo primary cytoreductive surgery. The primary ovarian tumor may be seen as mixed solid/cystic tumors, which are often bilateral, or as multilocular cystic lesions with thick internal septations and solid mural or septal components. Assessment can often be made as to whether the tumor invades the pelvic side wall or rectosigmoid colon or bladder, and associated complications, such as hydronephrosis and bowel obstruction, can be identified. Peritoneal deposits can be clearly identified; they are usually seen as discrete enhancing soft-tissue nodules. Liver, lung, and renal metastases and malignant pleural effusion indicate stage IV disease. PET/CT is of value in cases of suspected recurrence in which there is an increase in the level of the tumor marker CA-125 but indeterminate findings on CT or MRI [20]. Currently, the main role of MRI is in the characterization of ovarian masses rather than the staging of histologically proven ovarian cancer. MRI is very sensitive (95%) in the detection of peritoneal metastases, which show delayed enhancement on contrast-enhanced MRI [21]. Gadolinium-enhanced MRI is comparable to laparotomy but superior to serum CA-125 levels in the detection of residual or recurrent peritoneal and serosal implants in women who have been treated for ovarian cancer [21, 22]. MRI plays a crucial role in the detection of recurrent disease. It is important to realize that secondlook surgery is no longer routine and imaging diagnosis of recurrence may obviate a second-look laparotomy since secondary cytoreduction is only justified if resection is possible with no residual tumor. Imaging findings that indicate non-resectable recurrent tumor are invasion of the pelvic side wall, which should be suspected when the primary tumor lies within 3 mm of the pelvic side wall or when the iliac vessels are surrounded or distorted by tumor.

123

Malignant Diseases of the Female Genital Tract

a

b

c

d

Fig. 3 a-d. Advanced ovarian carcinoma. Contrast-enhanced CT images show (a) a left pleural effusion and a subcapsular liver deposit (arrow) and (b) peritoneal deposits along the surface of the liver and gastrosplenic ligament (arrows). There are also (c) a large omental cake and serosal deposits in the right paracolic gutter (arrows), (d) bilateral solid cystic ovarian masses and a large peritoneal deposit in the Pouch of Douglas (arrows)

References 1. Pecorelli S (2009) Revised FIGO staging for carcinoma of the vulva, cervix, and endometrium. Int J Gynaecol Obstet 105:103-104 2. Sala E, Wakely S, Senior E, Lomas D (2007) MRI of malignant neoplasms of the uterine corpus and cervix. AJR Am J Roentgenol 188:1577-1587 3. Shibutani O, Joja I, Shiraiwa M et al (1999) Endometrial carcinoma: efficacy of thin-section oblique axial MR images for evaluating cervical invasion. Abdom Imaging 24:520-526 4. Manfredi R, Mirk P, Maresca G et al (2004) Local-regional staging of endometrial carcinoma: role of MR imaging in surgical planning. Radiology 231:372-378 5. Sala E, Crawford R, Senior E et al (2009) Added value of dynamic contrast-enhanced magnetic resonance imaging in predicting advanced stage disease in patients with endometrial carcinoma. Int J Gynecol Cancer 19:141-146 6. Fujii S, Matsusue E, Kigawa J et al (2008) Diagnostic accuracy of the apparent diffusion coefficient in differentiating benign from malignant uterine endometrial cavity lesions: initial results. Eur Radiol 18:384-389 7. Shen SH, Chiou YY, Wang JH et al (2008) Diffusion-weighted single-shot echo-planar imaging with parallel technique in assessment of endometrial cancer. AJR Am J Roentgenol 190:481-488 8. Harry VN, Semple SI, Gilbert FJ, Parkin DE (2008) Diffusionweighted magnetic resonance imaging in the early detection of response to chemoradiation in cervical cancer. Gynecol Oncol 111:213-220

9. Fujii S, Matsusue E, Kanasaki Y et al (2008) Detection of peritoneal dissemination in gynecological malignancy: evaluation by diffusion-weighted MR imaging. Eur Radiol 18:18-23 10. Rockall AG, Sohaib SA, Harisinghani MG et al (2005) Diagnostic performance of nanoparticle-enhanced magnetic resonance imaging in the diagnosis of lymph node metastases in patients with endometrial and cervical cancer. J Clin Oncol 23:2813-2821 11. Ascher SM, Reinhold C (2002) Imaging of cancer of the endometrium. Radiol Clin North Am 40:563-576 12. Hricak H, Rubinstein LV, Gherman GM, Karstaedt N (1991) MR imaging evaluation of endometrial carcinoma: results of an NCI cooperative study. Radiology 179:829-832 13. Nicolet V, Carignan L, Bourdon F, Prosmanne O (2000) MR imaging of cervical carcinoma: a practical staging approach. Radiographics 20:1539-1549 14. Okamoto Y, Tanaka YO, Nishida M et al (2003) MR imaging of the uterine cervix: imaging-pathologic correlation. Radiographics 23:425-445 15. Scheidler J, Heuck AF (2002) Imaging of cancer of the cervix. Radiol Clin North Am 40:577-590,vii 16. Seki H, Takano T, Sakai K (2000) Value of dynamic MR imaging in assessing endometrial carcinoma involvement of the cervix. AJR Am J Roentgenol 175:171-176 17. Rockall AG, Ghosh S, Alexander-Sefre F et al (2006) Can MRI rule out bladder and rectal invasion in cervical cancer to help select patients for limited EUA? Gynecol Oncol 101:244-249 18. Hricak H, Gatsonis C, Chi DS et al (2005) Role of imaging in pretreatment evaluation of early invasive cervical cancer:

124

results of the intergroup study American College of Radiology Imaging Network 6651-Gynecologic Oncology Group 183. J Clin Oncol 23:9329-9337 19. Mitchell DG, Snyder B, Coakley F et al (2006) Early invasive cervical cancer: tumor delineation by magnetic resonance imaging, computed tomography, and clinical examination, verified by pathologic results, in the ACRIN 6651/GOG 183 Intergroup Study. J Clin Oncol 24:5687-5694 20. Subhas N, Patel PV, Pannu HK et al (2005) Imaging of pelvic malignancies with in-line FDG PET-CT: case exam-

Evis Sala, Susan Ascher

ples and common pitfalls of FDG PET. Radiographics 25:1031-1043 21. Ricke J, Sehouli J, Hach C et al (2003) Prospective evaluation of contrast-enhanced MRI in the depiction of peritoneal spread in primary or recurrent ovarian cancer. Eur Radiol 13:943-949 22. Low RN, Duggan B, Barone RM et al (2005) Treated ovarian cancer: MR imaging, laparotomy reassessment, and serum CA-125 values compared with clinical outcome at 1 year. Radiology 235:918-926

IDKD 2010-2013

Magnetic Resonance Imaging of Prostate Cancer Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer Department of Radiology, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands

Introduction With a total of 192,280 new cases predicted for 2009, prostate cancer (PC) now accounts for 25% of all new male cancers diagnosed in the USA [1]. Furthermore, in their lifetime, one in six men will be clinically diagnosed with PC, although many more will be found to have histological evidence of PC at autopsy [2-4]. Presently, approximately one in ten men will die of PC [5, 6]. The ever-aging population and more widespread use of the blood prostate-specific antigen (PSA) test [7, 8], as well as the tendency to apply lower cut-off levels for this test [9], will further increase the diagnosis of PC [10]. An elevated PSA level, abnormal changes in PSA level and dynamics (such as PSA velocity or doubling time), or an abnormal digital rectal examination are biological indicators signaling an increased risk of PC. With the improvement and wider range of curative therapies, the detection and subsequent exact localization of PC have become increasingly important because of their influence on treatment strategy [11, 12], in particular, laparoscopic (robotic) radical prostatectomy and intensity-modulated radiation therapy (IMRT) [13]. The urologist’s inability to palpate the operating field during laparoscopic surgery makes it even more crucial to precisely localize the cancer. Moreover, the urologist must determine whether the cancer is near a neurovascular bundle since this affects the decision of whether or not to perform nerve-sparing prostatectomy [14]. IMRT also necessitates accurate PC localization. While the prostate receives a standard dose of radiation, a higher (boost) dose can be given to dominant intraprostatic lesion(s) since it is those lesions that regularly appear to be the sites of recurrent disease [15]. Furthermore, precision radiation dosimetry will decrease radiation complications, particularly rectal wall toxicity [16], thereby likely diminishing the development of postradiation rectal cancer [17]. In order to determine the optimal treatment for each patient, it is necessary to thoroughly evaluate him and to determine the cancer’s characteristics. In this regard, laboratory values (PSA level and dynamics), the results of the digital rectal examination (clinical staging), and histopathological prostatic biopsy findings (Gleason

score) are important aspects. Additionally, however, magnetic resonance imaging (MRI) can play an important role in detecting and localizing those areas most reflective of the actual aggressiveness of the cancer. This directly influences patient assessment and may lead to important changes in treatment strategy, which can mean the difference between treatment success and failure. In the mid-1980s, the first prostate MRI examinations were performed. Since that time MRI has evolved from a promising technique into a mature imaging modality for PC imaging [18, 19]. Beside anatomical information, MRI provides functional tissue-characteristic information. Multiparametric MRI consists of a combination of anatomical T2-weighted imaging and functional MRI techniques such as dynamic contrastenhanced MRI (DCE-MRI), diffusion-weighted imaging (DWI), and 1H MR-spectroscopic imaging (MRSI). Within a multiparametric MRI examination the relative value of its component techniques differ. In addition to T2-weighted MRI, which mainly assesses anatomy, MRSI [20] can add specificity for PC detection, while DCE-MRI [21] and DWI [22] are both very sensitive and very specific. The clinical challenges in the work-up of patients with either suspected or proven PC include detection, localization, TNM staging, determination of cancer aggressiveness, follow-up of patients in active surveillance protocols, and determination of the site and extent of cancer recurrence after therapy. This chapter reviews the MRI anatomy of the prostate and the basic MRI techniques that can be applied in PC, and discusses the clinical role of this imaging modality. At the end of this chapter, three clinically applicable protocols are provided.

Magnetic Resonance Imaging: Anatomy In order to effectively apply the various MRI techniques, it is important to first understand the prostate’s normal anatomy and its intrinsic age-related changes. The superior part of the prostate is called the base while its most caudal part is referred to as the apex, analogous to the

126

a

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

b c

d

e

Coronal

Sagittal Axial Fig. 1 a-e. Normal prostate with signs of benign prostatic hyperplasia (BPH) as seen on sagittal (a, d), coronal (b), and axial (c, e) highresolution T2-weighted images. The peripheral zone is white, BPH is blue, the urethra is yellow, and the seminal vesicles are green

anatomy of the heart. The prostate consists of three zones: (1) the peripheral zone, located posteriorly and caudally at its middle portion; (2) the transition zone, located interiorly, around the urethra; and (3) the central zone, which is posterior and superior to the transition zone. Ventral to the prostate is the anterior fibromuscular stroma. In aging, an important frequent change in prostate zonal anatomy occurs, namely, the transition zone becomes hypertrophic (as in benign prostatic hyperplasia), thus compressing the central gland. Consequently, most men who are imaged for prostate cancer have only two identifiable compartments in the prostate, the hyperplastic transition zone surrounded by the peripheral zone (Fig. 1). Up to 70-80% of PCs are located in the peripheral zone [23], with an overall analysis of these cancers showing that they are homogeneously distributed across the entire zone [24]. Additionally, over half of the prostates examined contained two or more distinct cancer foci [25]. Nevertheless, while up to 20-52% of all PCs originate in the transition zone, only a small percentage (3.6-25%) of these cancers occur solely in that zone [24, 26], and many such patients will have foci of concurrent peripheral-zone cancer [23, 27, 28]. Thus, a solitary transition zone cancer is rare in the general PC population.

Magnetic Resonance Techniques and Their Role in Detection and Localization For evaluation of the prostate, anatomical (high-resolution) MRI can be combined with functional and metabolic information. DWI, dynamic MRI, and MRSI provide information about the motion of free water molecules and thus about cellular density (neo-)vascularization and metabolism, respectively. These different types of information can be combined into a multiparametric MRI examination.

T2-Weighted Imaging Compared to CT computed tomography (CT) scanning, MRI has a high soft-tissue contrast resolution (Fig. 2). The use of a disposable endorectal coil combined with other external coils at 1.5 Tesla (T) increases the softtissue contrast significantly and is now the accepted clinical standard for MRI of the prostate, when information about submillimeter extracapsular penetration is of clinical importance [29]. A drawback is the extra time required for inserting and checking the position of the endorectal coil as well as the substantial expense and patient discomfort.

127

Magnetic Resonance Imaging of Prostate Cancer

a

b

Fig. 2 a, b. A 60-year-old patient with a urinary catheter, and stage 2b prostate cancer (PC) in the left peripheral zone. a Axial MDCT does not well-delineate the prostate and the tumor is not visible. b T2-weighted axial MRI shows good prostate delineation and tumor in the left peripheral zone (arrow) without capsular penetration

a

Fig. 3 a, b. a On axial T1-weighted MRI, the post-biopsy hematoma (H (H) is white. b On the T2-weighted image, both the hematoma (H (H), benign prostatic hyperplasia ((B), and tumor (*) are of low signal intensity

On MRI, PC typically appears as an area of low signal intensity within the brighter, healthy peripheral zone, as seen using a T2-dominated sequence [30-32] (Figs. 2b, 3). In the central gland, PC is not as clearly discernible because the transition zone generally has lower signal intensity than the peripheral zone and is more inhomogeneous due to the architectural changes induced by benign prostatic hyperplasia, which may mimic PC. A recent study showed that a homogeneously low T2 signal intensity and lenticular shape were significantly associated with the presence of transitionzone [4, 33]. It was reported that, relative to muscle, cancers with higher Gleason scores had lower signal intensities than cancers with low Gleason scores [32]. However, the number of patients in that study was limited. In a comparison of T2-weighted MRI with prostatectomy specimens, MRI attained 52-83% sensitivities in PC localization, while specificities were somewhat lower (46-88%). A study that directly compared endorectal MRI with digital rectal examination and transrectal ultrasound (TRUS)-guided biopsy localization revealed significant

incremental value from MRI [34]. In patients subjected to multiple prior negative TRUS-guided biopsies, anatomical MRI by means of T2-dominated acquisition plays an important role. In this patient population, a sensitivity of 83% and positive predictive value of 50% for MRI have been established [35]. Post-biopsy hemorrhage causes areas of low signal intensity on T2-dominated sequences, thereby making PC detection more difficult. However, it was shown recently that the amount of hemorrhage was significantly lower in areas of cancer than in healthy tissue [36].

Diffusion-Weighted Imaging This non-invasive technique measures the fractional anisotropy of water molecules within the prostate, expressed in apparent diffusion coefficient (ADC) mapping. Thereby, the movement of water molecules in cancer tissue has been shown to be more restricted, thus producing lower ADC values [37, 38]. In a recent study of 38 patients who underwent DWI at 1.5 T with an endorectal coil, the mean ADC values of regions of interest

128

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Dynamic Contrast-Enhanced MRI

Fig. 4. Axial T2-weighted image obtained from a 70-year-old male with PSA = 35 and six negative TRUS biopsies shows a homogeneous area of low signal intensity in the right ventral prostate (arrows)

The use of intravenous contrast agents further enhances the localization accuracy of MRI. Endorectal DCEMRI, in which the contrast agent concentration is followed over time [46], discriminates between healthy prostate tissue and PC [47]. Early contrast enhancement and high (relative) peak enhancement are the most accurate predictors of PC of the peripheral zone, while fast washout of contrast agent and high permeability of the blood vessels (Fig. 6) are most sensitive for transition-zone PC [48, 49]. A recent study showed that the area under the receiver operating curve (AUC) for localizing PC increased significantly, from 0.68 with regular anatomical MRI to 0.91 with contrast-enhanced MRI. However, the limitations associated with the use of contrast agents are the lack of a uniform threshold, the low sensitivity for cancer, higher costs, and possible adverse reactions, of which the most serious is nephrogenic systemic fibrosis [50, 51]. 1H Magnetic Resonance Spectroscopic Imaging

Additionally, MRSI (Fig. 7) can be added to the imaging protocol to provide metabolic information based on the tissue citrate, choline, and creatine concentrations, and their relative ratios within the prostate. This is highly informative since the ratio between choline and citrate is markedly altered during the transformation from healthy to malignant prostate cells [52, 53] and an increasing choline+creatine/citrate ratio has been correlated with higher Gleason scores [54]. Presently, threedimensional (3D) MRSI of the entire prostate can be performed [55], thereby aiding in the diagnosis of

Fig. 5. Same patient as in Fig. 4. The ADC map (color coded) shows restriction in the area suspicious for tumor on this T2-weighted image (arrows)

within PC tissue were significantly lower than those within healthy prostate tissue [39]. In preliminary studies, the combination of this technique with MRSI [39] or T2weighted imaging [40] significantly improved localization accuracy (Figs. 4, 5). This was confirmed in a more recent study in 37 patients, in which sensitivity increased from 51% for T2-weighted imaging to 71% when the latter was combined with DWI [41]. In a recent multiparametric analysis, DWI was the best-performing parameter in localization [42]. Preliminary studies at 3 T have shown promising results [43-45].

Fig. 6. Same patient as in Fig. 4. Ktrans-map obtained with dynamic contrast-enhanced MRI shows pathological enhancement in area suspicious for tumor on this T2-weighted image (arrows)

129

Magnetic Resonance Imaging of Prostate Cancer

a

b

c

Fig. 7 a-c. Images obtained from a 65-year-old male with stage T3a PC in the left peripheral zone. The T2-weighted image (a) shows the tumor. MRSI of the right peripheral zone (b) shows low the choline and high citrate peak, whereas tumor in the left peripheral zone (c) shows high choline

central-gland PC (Fig. 7). The addition of 3D MRSI to MRI increases localization accuracy, by raising the specificity to as high as 91% [56]. However, a limitation of MRSI is its low spatial resolution and cumbersome post-processing. Compared to systematic biopsy, PC localization by means of MRI and MRSI was found to be more sensitive (67 and 76% vs. 50%) but less specific (69 and 57% vs. 82%) than systematic biopsy [57]. With whole-mount-section histopathology as the standard of reference, 3D MRSI had a significantly larger AUC (0.80) in localizing cancer than obtained with T2weighted MRI (0.68) [58]. The combination of T2weighted imaging and MRSI information to clinical data yielded the highest accuracy (AUC 0.85) in predicting the probability that a patient has insignificant PC [59], which was significantly higher than that obtained with clinical nomograms. A recent multi-institutional American College of Radiology Imaging Network study raised doubts on the additive value of MRSI over T2weighted imaging alone [60]. However, potential factors resulting in this conclusion were the selected prostatectomy population, the small size of the average cancer focus, and the inclusion of health centers without any previous MRSI experience.

sonable detection rates of 25-55% [62, 63]. Moreover, direct MRI-GB within the MRI scanner is technically feasible and can be performed on a routine basis. In patients with one previous negative TRUS biopsy, transrectal MRIGB performed at 1.5 T has produced promising cancer detection rates of 38-56% [64-66]. Lesions >10 mm can successfully be biopsied using this approach [66]. A multiparametric MRI approach consisting of T2weighted MRI, DWI, and DCE-MRI performed at 3 T has a median MRI-guided biopsy time of just 35 min and can generate an average of four biopsy cores per patient, as recently reported by Hambrock et al. (Fig. 8) [67]. Those

MRI-Guided Biopsies Random systematic TRUS-guided biopsy has relatively low detection rates and is prone to sampling error [61]. MRI-guided prostate biopsy (MRI-GB) has the potential to increase PC detection, as multiparametric MRI can target biopsies towards regions previously determined to be suspicious for cancer. Indeed, MRI findings have been used to direct biopsies under TRUS guidance, with rea-

Fig. 8. Same patients as in Fig. 4. MRI-guided biopsy of the tumorsuspicious region (red) showed PC with a Gleason score of 4+3. The biopsy needle is highlighted in white

130

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

authors showed that a cancer detection rate of 59% can be achieved in a large cohort of patients with more than two previous negative TRUS biopsies [68]. In addition, 93% of the cancers found were clinically significant, thus not contributing to the over-diagnosis of insignificant cancers. A limitation of MRI-GB is that multiparametric MRI for tumor localization and MRI-guided biopsy need to be performed in two different sessions, as image postprocessing and exact localization of the cancer demand time. Another disadvantage is movement of the prostate during the biopsy procedure [69]. A reduction of the MRI-GB intervention time remains an important challenge, perhaps solvable by robotics. In the future, MRIbased guidance might also be used in the focal treatment of PC, such as in the form of brachytherapy or cryotherapy.

Prediction of Prostate Cancer Aggressiveness Prostate cancer aggressiveness is pathologically graded by the Gleason score, which consists of a combination of the two most prevalent Gleason grades (range 1-5) based on the architectural characteristics of PC tissue [70]. Biopsy specimens obtained from random TRUS guidedbiopsy are subject to sampling error in approximately 64% of the cases [71]; this results in incorrect Gleason scores and thus incorrect patient risk stratification, which in turn leads to under- or overtreatment [72]. Apart from a relationship between muscle-normalized signal intensity on T2-weighted MRI and cancer Gleason scores [73], a correlation between cancer visibility on T2-weighted images and aggressiveness has been suggested, with low-grade cancers being detected in 43% and high-grade cancers in 79% of such cases [74]. Moreover, (choline+creatine)/citrate ratios, as determined by MRSI, have been shown to correlate with the Gleason

a

b

score [75-77]. Preliminary results in the evaluation of ADC as a marker of cancer aggressiveness are promising; ADC values were found to negatively correlate (ρ = –0.497, p<0.0001) with peripheral-zone cancers and Gleason scores. In another study, lesions in patients with less aggressive cancers (PSA <10, T1 or T2a, Gleason score <6) had significantly higher ADC values than lesions in patients with intermediate- to high-risk cancers (PSA >10, T2b, Gleason score ≥7) (Fig. 9) [77-79].

Local (T) Staging The application of MRI to determine whether PC is locally advanced remains controversial due to varying results across institutions. The most reliable MRI signs of extracapsular extension are bulging of the prostate into the periprostatic fat, obliteration of the recto-prostatic angle, and asymmetry of the neurovascular bundles (Fig. 10) [80]. Seminal vesicle invasion is usually easily detectable as areas of low signal intensity in the brighter seminal fluid (Fig. 10). Two meta-analyses on local staging by MRI at 1.5 T reported combined maximum sensitivities and specificities of 71-74%, while sensitivity was 62-69% at a specificity of 80% [80-82]. Imaging in more than one plane as well as the use of an endorectal coil resulted in a significantly better staging performance. A large study of 336 patients conducted by Cornud et al. found an overall sensitivity, specificity, and positive and negative predictive values of 40, 95, 79, and 76%, respectively [83]. Highspecificity MRI (in which only definite locally advanced cases are excluded from curative therapy) is now the optimal local staging method [84, 85]. The addition of MRI with an endorectal coil to clinical data such as PSA, biopsy Gleason score, and Kattan nomogram resulted in a significantly increased accuracy of predicting disease stage, extracapsular extension, and

c

Fig. 9 a-c. Images from a 62-year-old patient with PSA = 9 and PC with a Gleason score of 3 but with a local component with a score of 4. a Axial T2-weighted MRI shows low signal in almost the entire peripheral zone. b On the ADC map, the tumor can be delineated based on its lower value (uninterrupted line) and on the presence of local areas with very low value (stippled areas). c Prostatectomy confirmed the lower signal to be a Gleason 3 tumor and the very low signal component to be Gleason 4

131

Magnetic Resonance Imaging of Prostate Cancer

d

b

Fig. 10 a-d. T2-weighted images from a 49-year-old male with stage T3B PC. a Coronal image at basis shows seminal vesicle infiltration (*); b axial image at the mid-part shows bulging (arrows); c axial image at the apex shows obliteration of the recto-prostatic angle (arrow) and, for comparison, the normal recto-prostatic angle (drawn line); d schematic drawing, as part of the structured report

seminal vesicle invasion [79, 86, 87]. Recently, it was shown that the presence and the degree of pre-radiation therapy extracapsular extension predicted by MRI was a predictor of post-therapy outcomes [88]. The addition of DWI to T2-weighted imaging was recently shown to be significantly better in establishing urinary bladder wall invasion as well as seminal vesicle invasion [89, 90]. Also, the addition of three-dimensional MRSI to MRI improved staging accuracies, particularly for less-experienced readers, and increased interobserver agreement [91]. A drawback is the longer duration (by approximately 15 min) of the examination. Experience was found to be an important factor in disease staging [92]. However, the accuracy of a less-experienced reader could be increased by contrast-enhanced examinations [93]. Likewise, MRI interpretation using multiplanar cross-referencing significantly improved staging accuracy compared with interpretation without cross-referencing [94]. Interactive tutorials with direct feedback were also shown to significantly increase the accuracy of staging by less-experienced readers [95]. Imaging at higher magnetic field strengths (e.g., 3 T) can achieve better image resolution. Although not yet widely available for clinical work, local staging at 3 T was shown in two studies to improve the sensitivities and specificities of experienced readers to 80-88% and 94100%, respectively [96, 97]. In the current PSA era, this higher resolution is mandatory as PC is detected at earlier stages. Likewise, if extracapsular extension is present, it will most often be minimal.

Metastatic Disease (NM Staging) Nodes The prognosis of patients with PC is poorer if lymph node metastases are present. The risk of lymph node metastasis is currently determined (albeit inaccurately) using nomograms [98, 99]. In patients with an elevated risk for metastasis, additional examinations are required. Today, the most commonly used imaging techniques for detecting lymph node metastasis are multi-detector CT scan (MDCT) and conventional MRI, with image interpretation essentially based on lymph node size and shape criteria. Although the criteria vary slightly [100], lymph nodes with a short-axis diameter >8 mm for round lymph nodes and >10 mm for oval ones are generally considered to be malignant [101, 102]. Both MDCT [103] and MRI [104] have a low sensitivity (36 and 39%, respectively) for diagnosing PC lymph node metastases using these size and shape criteria. In studies that have employed thresholds as small as 6 mm [105], the specificity was very high (95-100%) but the sensitivity was too low (0-25%) to be useful in regular clinical practice for the evaluation of metastatic lymph node disease [106]. Some authors advocate restricting the application of these techniques to high-risk patients (e.g., with PSA levels >20 ng/mL) in order for them to be cost-effective [107, 108]. Thus, supplementary, invasive diagnostic examinations in the form of surgical pelvic lymph node dissection (PLND) are still commonly performed.

132

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

Magnetic resonance lymphography (MRL) using a lymph-node-specific contrast agent (Combidex/Sinerem) [109, 110] is an experimental technique that, compared to PLND, has a high negative predictive value (>96%) for the detection of lymph node metastasis in extended areas. Importantly, its use can render PLND unnecessary in negative cases [111]. Bone Most metastatic bone lesions are sclerotic [112]; a 50% change in bone mineral density is needed for metastatic bone lesions to be visible on X-ray images [113]. The most commonly used first-line diagnostic test to detect or exclude bone metastases is technetium-99m-diphosphonate bone scintigraphy. However, this approach lacks specificity, such that primary skeletal diseases may generate false-positive findings. Conventional Xray examinations can be used to exclude false-positive

a

findings on bone scintigraphy that have arisen due to conditions such as trauma, degenerative joint disease, and other chronic diseases. However, conventional Xray is too insensitive for the detection of metastatic bone lesions. Lecouvet et al. evaluated the accuracy of bone scintigraphy, targeted X-rays, and MRI in 66 patients with prostate cancer, 41 of whom had bone metastases (Fig. 11) [114]. Sensitivities were 46% for bone scintigraphy alone, 63% for bone scintigraphy and targeted X-rays, and 100% for MRI; the corresponding specificities were 32, 64, and 88%, respectively. Thus MRI was significantly more sensitive than any other approach ( <0.001); furthermore, MRI limited to the pelvis and (p axial skeleton was shown to be sufficient, as the probability of finding metastases outside these locations in the absence of metastases in the axial skeleton is negligible. This is particularly the case in PC, which predominantly metastasizes to the spine and pelvis due to the venous drainage routes [115, 116].

b

c

Fig. 11 a-c. Patient with bone metastasis from PC. a T1-weighted image shows a small, low signal intensity lesion (arrow) with high intensity on DWI (b), consistent with metastasis. The bone scan (c) and plain film were negative

133

Magnetic Resonance Imaging of Prostate Cancer

Recurrence Currently, a rise in PSA level is the only major indicator for PC recurrence after radical prostatectomy or radiotherapy [117, 118]. The recurrence of PC is indicated by PSA values >0.4 ng/mL following radical prostatectomy and by a PSA value 2 ng/mL above the nadir value (after the PSA bounce) following radiation therapy [119, 120]. Whenever such an elevation of PSA occurs, the main objective is to determine whether it is due exclusively to recurrent or residual disease in or outside the prostate, since locally recurrent disease might still be cured with adjuvant radiotherapy. The use of DRE and TRUS for recurrence detection is compromised because of the difficulty in distinguishing between fibrotic changes and recurrent disease. TRUSguided biopsy has a low sensitivity and specificity with respect to recurrence following prostatectomy or radiotherapy [121, 122] whereas MRI can help in the detection of recurrent disease. One of the major advantages of MRI is that it can be used to evaluate the recurrence of PC at either local or distant sites. After external-beam radiation therapy, prostatic tissue demonstrates diffusely low signal intensity on T2-weighted images, with indistinct zonal anatomy [123]. The contrast between benign, irradiated tissues and recurrent cancer is

therefore decreased, which makes the detection of recurrence more difficult. In one of the rare studies that used radical prostatectomy specimens after radiotherapy as the standard of reference, retrospectively assessed endorectal MRI had a low to fair accuracy for the detection of local post-radiotherapy recurrence (AUC 0.61-0.75), the prediction of extracapsular extension (0.76-0.87), and evidence of seminal vesicle invasion (0.70-0.76) [124]. Functional MRI techniques are of additional value in the detection of disease recurrence following radiation therapy (Fig. 12). On MRSI, the presence of three or more suspicious voxels in a prostate half had a sensitivity of 89% and a specificity of 82% in the detection of local post-radiotherapy recurrence (n=23) [125]. Furthermore, post-radiotherapy DCE-MRI was shown to have a sensitivity of 70-74% and a specificity of 73-85% for recurrence detection [126, 127]. The addition of DWI to T2-weighted imaging (AUC 0.61) also improved the detection of post-radiotherapy disease recurrence by 27% (AUC 0.88) [128]. In a retrospective study evaluating the value of anatomical T2-weighted MRI in the detection of post-prostatectomy disease recurrence, the sensitivity was reported to be 95% and the specificity 100% [129]. Thus, also in the evaluation of post-prostatectomy disease recurrence functional MRI techniques are of additional value (Fig. 13). In

a

b

a

b

Fig. 12 a, b. A 69-year-old patient 2.5 years after external radiation beam therapy; his PSA had risen to 2.1. a T2-weighted MRI shows entirely low signal gland; no tumor can be discriminated. b DCE-MRI shows enhancement of the right half of the prostate (arrow). Biopsy revealed Gleason 5+4 PC

Fig. 13 a, b. Patient with PSA recurrence after prostatectomy. a T1-weighted image shows no lesion whereas (b) DWI shows enhancement (red circle), which was due to recurrence of the tumor (Gleason 4+3)

134

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

this regard, DCE-MRI combined with anatomical T2weighted MRI was shown to improve sensitivity and specificity [130, 131]. According to Sciarra et al. [132], DCE-MRI is even better when used in combination with MRSI, resulting in AUC values of 0.94-0.96 for the detection of local recurrence compared to values of 0.810.94 for either DCE-MRI or MRSI alone. One of the limitations in the reported studies was the use of TRUSguided biopsies as the standard of reference [131].

Protocols The European Society of Urogenital Radiology and the Royal College of Surgeons (UK) are currently working on a set of guidelines, with the first version to be published at the end of 2011 in European Radiology. Currently, three protocols can be recommended: one for detection/localization and recurrence, one for staging, and one for the assessment of nodal size and bone marrow. Unfortunately, despite the enormous clinical potential of Combidex/Sinerem has, due to the inability of the pharmaceutical companies to provide convincing data to the FDA and EMEA, this contrast agent was not approved for nodal imaging and further development has been discontinued.

Detection, Localization, Recurrence This is a fast (<30 min) protocol that does not involve the use of an endorectal coil. The minimum requirements should include images covering the entire prostate: 1. T1-weighted axial images, to detect post-biopsy hematomas. 2. T2-weighted axial images and images in one other plane: 4 mm at 1.5 T, 3 mm at 3 T. The resolution should be at least 0.5 × 0.5 mm-0.7 × 0.7 mm at both 1.5 and 3 T. 3. DWI axial: 5 mm at 1.5 T, 4 mm at 3 T. The resolution should be at least 1.5 × 1.5 mm-2.0 × 2.0 mm at 1.5 T and 1.0 × 1.0 mm-1.5 × 1.5 mm at 3 T, with a b0 image plus multiple b images allowing quantification.

a

b

The highest b value should be 1000 if ADC is calculated, otherwise it should be 1400 (with adequate signal to noise ratio). 4. Contrast-enhanced imaging, axial: 4 mm at 1.5 and 3 T. The resolution should be at least 1.0 × 1.0 mm at 1.5 T and 0.7 × 0.7 mm at 3 T. Quantitative or semi-quantitative DCE-MRI analysis does not comprise minimal practice, but if available should be done as optimal practice. The maximum temporal resolution should be 10 s following a single dose of contrast, with an injection rate of 3 mL/s. For DCE-MRI, imaging acquisition should be continued for 5 min to detect washout. Imaging can adequately be performed at 1.5 T. A pelvic coil should always be used; bowel relaxants (Buscopan, glucagon) provide optimal imaging with fewer motion artifacts.

Staging This is a longer (45 min) protocol that allows determination of (minimal) capsular penetration (Fig. 14). Preferably, this exam should be done with an endorectal coil. Minimum requirements should include images covering the entire prostate: 1. T1-weighted axial, to detect hematomas. 2. T2-weighted axial and two other planes: 3 mm at 1.5 and 3 T. The resolution should be at least 0.3 × 0.3 mm 0.7 × 0.7 mm at 1.5 T and 0.3 × 0.3 mm - 0.5 × 0.5 mm at 3 T. 3. DWI axial: 5 mm at 1.5 T, 4 mm at 3 T. The resolution should be at least 1.5 × 1.5 mm - 2.0 × 2.0 mm at 1.5 T and 1.0 × 1.0 mm - 1.5 × 1.5 mm at 3 T, with a b0 image plus multiple b images allowing quantification. The highest b value should be 1000 if ADC is calculated, otherwise it should be 1400 (with adequate signal to noise ratio). 4. Contrast-enhanced imaging, axial: 3-4 mm at 1.5 T and 3 mm at 3 T. The resolution should be at least 1.0 × 1.0 mm at 1.5 T and 0.7 × 0.7 mm at 3 T. Quantitative or semi-quantitative DCE-MRI analysis does not comprise minimal practice, but if available should be done as optimal practice. The maximum temporal

c

Fig. 14 a-c. This 45-year-old patient requested erectile-function-preserving surgery. a High-resolution T2-weighted image obtained with an endorectal coil (3T) shows minimal capsular penetration close to the neurovascular bundle (detail in b, arrows). Prostatectomy was done, saving the right and sacrificing the left neurovascular bundle. c Histopathology revealed submillimeter capsular extension, but negative resection margins. T Tumor, ECE extracapsular extension

135

Magnetic Resonance Imaging of Prostate Cancer

Interpretation, Structured Reporting

resolution should be 10 s following a single dose of contrast, with an injection rate of 3 mL/s. For DCEMRI, imaging acquisition should be continued for 5 min to detect washout. 5. MRSI (optional).

Due to a lack of standardization, the combined interpretation of anatomical and functional images is not possible. Optimal results are obtained using an integrated computer program that combines T2-weighted anatomical images with DWI, DCE-MRI, and MRSI on one screen (Fig. 15). Additionally, there is a dire need for a uniform scoring system; for example, in which every modality allows the assignment of points on a 5-point scale regarding the impression of a lesion’s malignancy, analogous to the BIRADS classification system in breast cancer imaging. Thus, a lesion score of 20/20 or 15/15 definitely indicates an intermediate or highly aggressive tumor (Fig. 15). If the score is <10/15 or <14/20, there may be a non-aggressive tumor but other pathologies, such as prostatitis, are a possibility (Fig. 16). This type of standardized scoring system allows the quantification of lesion development and reliable comparison with a follow-up MRI examination (Fig. 16).

Nodes and Bone This 30-min protocol is used to assess nodal size and bone marrow metastases. 1. T1 (SE or f/T SE)-weighted coronal of the lower lumbar spine plus pelvis: 3-mm slices. 2. 3D f/T SE T2-weighted coronal of the lower lumbar spine plus pelvis; 1-mm isometric voxels. 3. DWI (b value 0 plus 600) coronal of the lower lumbar spine plus pelvis; 2.5-3 mm isometric voxels. 4. T1 (SE or f/T SE)-weighted sagittal images of the cervical and thoracic spine. 5. STIR or DWI sagittal images of the cervical and thoracic spine.

a

b

c

d

e

f

Fig. 15 a-f. A 59-year-old patient with PSA = 12 and a Gleason 8 tumor of the right peripheral zone, stage T3a at prostatectomy. Screenshot from computer monitor display. a DCE-MRI; b ADC-map (color coded); c choline image; d coronal T2-weighted image; e axial T2-weighted image; f time concentration curve and H-spectrum of cursor. This patient’s score for all modalities was 5 points (20/20). Scale: 1 no tumor, 5 definitely tumor

136

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

a

b

c

d

e

f

g Fig. 16 a-g. This 55-year-old patient with PSA = 5, PC of stage T1 at digital rectal examination and Gleason 3+3 <5% on TRUS biopsy chose active surveillance. a T2-weighted image shows no tumor: score 1/5; b DCE-MRI shows some asymmetrical enhancement with curve 2: score 3/5; c ADC map shows only a small reduction: score 2/5. The total score of 6/15 argues for no or non-aggressive tumor. Follow-up MRI 1 year later: d T2-weighted image at apex shows homogeneous, asymmetrical low signal: score 5/5; e DCE-MRI shows pathological asymmetrical enhancement: score 5/5; f DWI shows an area with very low value: score 5/5. Based on these results, MRI-guided biopsy was performed. A Gleason 5+3 specimen was obtained on prostatectomy (g), which confirmed a Gleason 5+3 tumor

Conclusion The use of MRI has led to improved localization of PC compared to the results obtained with transrectal ultrasound. This has resulted in increased cancer detection by targeting biopsy under MRI guidance. Furthermore, radiotherapy planning and the guidance of minimally invasive local therapies have become possible as a result of adequate PC localization with MRI. Multiparametric MRI using functional imaging has been shown to improve the prediction of tumor aggression. Also, compared to staging nomograms, high-resolution (endorectal coil) MRI considerably improves the determination of local tumor extension, thereby supporting the clini-

cian in choosing the appropriate therapy based on PC stage. In case of a biochemical recurrence after radical treatment, MRI can detect disease location with greater accuracy than obtained with DRE and TRUS. Finally, MRI offers advantages over bone scintigraphy. Nonetheless, among the different acquisition methods, protocols, magnetic field strengths, and multimodality techniques that are used, a consensus is needed regarding dedicated MRI protocols for different specific clinical indications. Implementation of protocols that on a larger scale are highly similar if not identical will promote the maturation of different multimodality techniques and improve further development of supportive techniques such

137

Magnetic Resonance Imaging of Prostate Cancer

as computer-aided diagnosis. Furthermore, the use of consensus protocols enables larger studies, the aims of which include estimation of the true value of multiparametric MRI on a broader, more representative scale.

Take-Home Messages • Currently, MRI is the most accurate imaging modality in localizing and staging PC. • Prostate cancer in the central gland of the prostate is more difficult to detect and localize than PC in the peripheral zone of the gland. • Multiparametric MRI comprises a combination of T2weighted imaging, DWI, MRSI, and DCE-MRI. • The typical PC focus is of low signal intensity on T2weighted imaging and has a low apparent diffusion coefficient value, a high choline+creatine/citrate ratio, and high contrast agent permeability and fast washout. • As not all PC foci will exhibit all these features, multiparametric MRI is needed to obtain the highest localization accuracy. • The apparent diffusion coefficient and choline+creatine/citrate ratio are associated with the aggressiveness of the cancer. • Direct MRI-guided biopsy of a suspicious lesion in patients whose previous TRUS biopsies were negative is feasible, highly accurate, and can be performed within 30 min by an experienced radiologist. • For staging PC at 1.5 T, the use of an endorectal coil is necessary; at 3 T, the endorectal coil is only needed if knowledge of minimal capsular penetration is important for therapeutic decision-making. • Conventional nodal staging with MRI is still based on lymph node size and shape criteria; novel, more specific techniques still have to be approved by the FDA and EMEA. • For bone staging, MRI of the axial skeleton outperforms bone scintigraphy in detecting metastatic foci. • Dynamic contrast-enhanced MRI may play an important role in improving the detection of PC recurrence after surgery or radiation therapy. Acknowledgements The authors would like to acknowledge Mrs. Louise Dickinson from the Royal College of Surgeons and her guideline group for her input in formulating minimal and maximal requirements. The final guidelines will appear by the end of 2010.

Suggested Reading Anatomy Coakley FV, Hricak H (2000) Radiological anatomy of the prostate gland: a clinical approach. Radiol Clin N Amer 38:15-30

Hricak H, Dooms GC, McNeal et al (1987) MR imaging of the prostate gland 148:51-58 McNeal JE (1988) Normal histology of the prostate. Amer J Surg Pathol 12:619-633 Sommer FG, McNeal JE, Carrol CL (1986) MR depiction of zonal anatomy of the prostate at 1.5T. JCAT 1096:983-989 Villiers G, De Meereleer GO (2007) MRI anatomy of the prostate and application of MRI in radiotherapy planning. Eur J Radiol 63:361-368

Prostate Cancer Futterer JJ, Heijmink SWTP, Scheenen TWJ et al (2006) Prostate cancer localization with dynamic contrast enhanced MR imaging and proton MR spectroscopic Imaging. Radiology 241:449-458 Hambrock T, Somford DM, Hoeks C et al (2010) Magnetic resonance imaging guided prostate biopsy in men with repeat negative biopsies and increased prostate specific antigen. J Urol 183:520-528 Heesakkers RAM, Hoevels AM, Jager GJ et al (2008) MRI with a lymph-node-specific contrast agent as an alternative to CT scan and lymph-node dissection in patients with prostate cancer: a prospective multicohort study. Lancet Oncol 9:850-856 Hrick H, Choyke, Eberhart SC et al (2007) Imaging prostate cancer: a multidisciplinary perspective. Radiology 243:28-53 Hoevels AM, Heesakkers RAM, Adang EM et al (2008) The diagnostic accuracy of CT and MRI in the staging of pelvic lymph nodes in patients with prostate cancer: a meta-analysis. Clin Radiol 63:387-395 Jager GJ, Severens JL, Thornbury JR et al (2000) Prostate cancer staging: should MR imaging be used? A decision analytic approach. Radiology 215:445-451

References 1. Jemal A, Siegel R, Ward E et al (2009) Cancer statistics. CA Cancer J Clin 59:225-249 2. Carter HB, Piantadosi S, Isaacs JT (1990) Clinical evidence for and implications of the multistep development of prostate cancer. J Urol 143:742-746 3. Parkin DM, Bray FI, Devesa SS (2001) Cancer burden in the year 2000. The global picture. Eur J Cancer 37 Suppl 8:S4-S66 4. Konety BR, Bird VY, Deorah S, Dahmoush L (2005) Comparison of the incidence of latent prostate cancer detected at autopsy before and after the prostate specific antigen era. J Urol 174:1785-1788 5. Crawford ED (2003) Epidemiology of prostate cancer. Urology 62(6 Suppl 1):3-12 6. Stewart SL, King JB, Thompson TD et al (2004) Cancer mortality surveillance – United States, 1990-2000. MMWR Surveill Summ 53:1-108 7. Catalona WJ, Loeb S, Han M (2006) Viewpoint: expanding prostate cancer screening. Ann Intern Med 144:441-443 8. Hoffman RM (2006) Viewpoint: limiting prostate cancer screening. Ann Intern Med 144:438-440 9. Graif T, Loeb S, Roehl KA et al (2007) Under Diagnosis and Over Diagnosis of Prostate Cancer. J Urol 178:88-92 10. Max W, Rice DP, Sung HY et al (2002) The economic burden of prostate cancer, California, 1998. Cancer 94:2906-2913 11. Mangar SA, Huddart RA, Parker CC et al (2005) Technological advances in radiotherapy for the treatment of localised prostate cancer. European Journal of Cancer 41:908-921 12. Meraney AM, Haese A, Palisaar J et al (2005) Surgical management of prostate cancer: Advances based on a rational approach to the data. European Journal of Cancer 41:888-907 13. Bucci MK, Bevan A, Roach M, III (2005) Advances in radiation therapy: Conventional to 3D, to IMRT, to 4D, and Beyond. CA Cancer J Clin 55:117-134

138

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

14. Hricak H, Wang L, Wei DC et al (2004) The role of preoperative endorectal magnetic resonance imaging in the decision regarding whether to preserve or resect neurovascular bundles during radical retropubic prostatectomy. Cancer 100:26552663 15. Cellini N, Morganti AG, Mattiucci GC et al (2002) Analysis of intraprostatic failures in patients treated with hormonal therapy and radiotherapy: implications for conformal therapy planning. Int J Radiat Oncol Biol Phys 53:595-599 16. Roach M, III (2004) Reducing the toxicity associated with the use of radiotherapy in men with localized prostate cancer. Urol Clin North Am 31:353-366 17. Baxter NN, Tepper JE, Durham SB et al (2005) Increased risk of rectal cancer after prostate radiation: A population-based study. Gastroenterology 128:819-824 18. Steyn JH, Smith FW (1982) Nuclear magnetic resonance imaging of the prostate. Br J Urol 54:726-728 19. Hricak H, Williams RD, Spring DB et al (1983) Anatomy and pathology of the male pelvis by magnetic resonance imaging. AJR Am J Roentgenol 141:1101-1110 20. Villeirs GM, Oosterlinck W, Vanherreweghe E, De Meerleer GO (2008) A qualitative approach to combined magnetic resonance imaging and spectroscopy in the diagnosis of prostate cancer. Eur J Radiol Dec 10 21. Tanimoto A, Nakashima J, Kohno H et al (2007) Prostate cancer screening: the clinical value of diffusion-weighted imaging and dynamic MR imaging in combination with T2-weighted imaging. J Magn Reson Imaging 25:146-152 22. Scheenen TW, Klomp DW, Roll SA et al (2004) Fast acquisition-weighted three-dimensional proton MR spectroscopic imaging of the human prostate. Magn Reson Med 52:80-88 23. Chen ME, Johnston DA, Tang K et al (2000) Detailed mapping of prostate carcinoma foci: biopsy strategy implications. Cancer 89:1800-1809 24. McNeal JE, Redwine EA, Freiha FS, Stamey TA (1988) Zonal distribution of prostatic adenocarcinoma. Correlation with histologic pattern and direction of spread. Am J Surg Pathol 12:897-906 25. Miller GJ, Cygan JM (1994) Morphology of prostate cancer: the effects of multifocality on histological grade, tumor volume and capsule penetration. J Urol 152:1709-1713 26. Brossner C, Winterholer A, Roehlich M et al (2003) Distribution of prostate carcinoma foci within the peripheral zone: analysis of 8,062 prostate biopsy cores. World J Urol 21: 163-166 27. Horninger W, Reissigl A, Rogatsch H et al (2000) Prostate cancer screening in the Tyrol, Austria: experience and results. Eur J Cancer 36:1322-1335 28. Ohori M, Kattan M, Scardino PT, Wheeler TM (2004) Radical prostatectomy for carcinoma of the prostate. Mod Pathol 17:349-359 29. Tempany CM, Zhou X, Zerhouni EA et al (1994) Staging of prostate cancer: results of Radiology Diagnostic Oncology Group project comparison of three MR imaging techniques. Radiology 192:47-54 30. Cruz M, Tsuda K, Narumi Y et al (2002) Characterization of low-intensity lesions in the peripheral zone of prostate on prebiopsy endorectal coil MR imaging. Eur Radiol 12:357-365 31. Claus FG, Hricak H, Hattery RR (2004) Pretreatment evaluation of prostate cancer: role of MR imaging and 1H MR spectroscopy. Radiographics 24:S167-S180 32. Wang L, Mazaheri Y, Zhang J et al (2008) Assessment of biologic aggressiveness of prostate cancer: correlation of MR signal intensity with Gleason grade after radical prostatectomy. Radiology 246:168-176 33. Akin O, Sala E, Moskowitz CS et al (2006) Transition zone prostate cancers: features, detection, localization, and staging at endorectal MR imaging. Radiology 239:784-792 34. Mullerad M, Hricak H, Kuroiwa K et al (2005) Comparison of endorectal magnetic resonance imaging, guided prostate

35.

36. 37.

38. 39.

40. 41.

42.

43. 44. 45. 46. 47.

48.

49.

50. 51. 52.

53. 54.

biopsy and digital rectal examination in the preoperative anatomical localization of prostate cancer. J Urol 174:2158-2163 Beyersdorff D, Taupitz M, Winkelmann B et al (2002) Patients with a history of elevated prostate-specific antigen levels and negative transrectal US-guided quadrant or sextant biopsy results: value of MR imaging. Radiology 224:701-706 Tamada T, Sone T, Jo Y et al (2008) Prostate cancer: relationships between postbiopsy hemorrhage and tumor detectability at MR diagnosis. Radiology 248:531-539 Sato C, Naganawa S, Nakamura T et al (2005) Differentiation of noncancerous tissue and cancer lesions by apparent diffusion coefficient values in transition and peripheral zones of the prostate. J Magn Reson Imaging 21:258-262 Zelhof B, Pickles M, Liney G et al (2009) Correlation of diffusion-weighted magnetic resonance data with cellularity in prostate cancer. BJU Int 103:883-888 Mazaheri Y, Shukla-Dave A, Hricak H et al (2008) Prostate cancer: identification with combined diffusion-weighted MR imaging and 3D 1H MR spectroscopic imaging – correlation with pathologic findings. Radiology 246:480-488 Haider MA, Van Der Kwast TH, Tanguay J et al (2007) Combined T2-weighted and diffusion-weighted MRI for localization of prostate cancer. AJR Am J Roentgenol 189:323-328 Yoshimitsu K, Kiyoshima K, Irie H et al (2008) Usefulness of apparent diffusion coefficient map in diagnosing prostate carcinoma: correlation with stepwise histopathology. J Magn Reson Imaging 27:132-139 Langer DL, Van Der Kwast TH, Evans AJ et al (2009) Prostate cancer detection with multi-parametric MRI: logistic regression analysis of quantitative T2, diffusion-weighted imaging, and dynamic contrast-enhanced MRI. J Magn Reson Imaging 30:327-334 Gibbs P, Pickles MD, Turnbull LW (2006) Diffusion imaging of the prostate at 3.0 tesla. Invest Radiol 41:185-188 Pickles MD, Gibbs P, Sreenivas M, Turnbull LW (2006) Diffusion-weighted imaging of normal and malignant prostate tissue at 3.0T. J Magn Reson Imaging 23:130-134 Miao H, Fukatsu H, Ishigaki T (2007) Prostate cancer detection with 3-T MRI: comparison of diffusion-weighted and T2weighted imaging. Eur J Radiol 61:297-302 Barentsz JO, Engelbrecht M, Jager GJ et al (1999) Fast dynamic gadolinium-enhanced MR imaging of urinary bladder and prostate cancer. J Magn Reson Imaging 10:295-304 Padhani AR, Gapinski CJ, Macvicar DA et al (2000) Dynamic contrast enhanced MRI of prostate cancer: correlation with morphology and tumour stage, histological grade and PSA. Clin Radiol 55:99-109 Engelbrecht MR, Huisman HJ, Laheij RJ et al (2003) Discrimination of prostate cancer from normal peripheral zone and central gland tissue by using dynamic contrast-enhanced MR imaging. Radiology 229:248-254 Van Dorsten FA, Van Der Graaf M, Engelbrecht MR et al (2004) Combined quantitative dynamic contrast-enhanced MR imaging and (1)H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imaging 20:279-287 Runge VM (2000) Safety of approved MR contrast media for intravenous injection. J Magn Reson Imaging 12:205-213 Lin SP, Brown JJ (2007) MR contrast agents: physical and pharmacologic basics. J Magn Reson Imaging 25:884-899 Kurhanewicz J, Vigneron DB, Hricak H et al (1996) Three-dimensional H-1 MR spectroscopic imaging of the in situ human prostate with high (0.24-0.7-cm3) spatial resolution. Radiology 198:795-805 Coakley FV, Qayyum A, Kurhanewicz J (2003) Magnetic resonance imaging and spectroscopic imaging of prostate cancer. J Urol 170:S69-S75 Zakian KL, Sircar K, Hricak H et al (2005) Correlation of proton MR spectroscopic imaging with gleason score based on step-section pathologic analysis after radical prostatectomy. Radiology 234:804-814

Magnetic Resonance Imaging of Prostate Cancer

55. Scheenen TW, Klomp DW, Roll SA et al (2004) Fast acquisition-weighted three-dimensional proton MR spectroscopic imaging of the human prostate. Magn Reson Med 52:80-88 56. Scheidler J, Hricak H, Vigneron DB et al (1999) Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging – clinicopathologic study. Radiology 213:473-480 57. Wefer AE, Hricak H, Vigneron DB et al (2000) Sextant localization of prostate cancer: comparison of sextant biopsy, magnetic resonance imaging and magnetic resonance spectroscopic imaging with step section histology. J Urol 164: 400-404 58. Fütterer JJ, Heijmink SWTPJ, Scheenen TWJ et al (2006) Prostate cancer localization with dynamic contrast-enhanced MR imaging and proton MR spectroscopic imaging. Radiology 241:449-458 59. Shukla-Dave A, Hricak H, Kattan MW et al (2007) The utility of magnetic resonance imaging and spectroscopy for predicting insignificant prostate cancer: an initial analysis. BJU Int 99:786-793 60. Weinreb JC, Blume JD, Coakley FV et al (2009) Prostate cancer: sextant localization at MR imaging and MR spectroscopic imaging before prostatectomy – results of ACRIN prospective multi-institutional clinicopathologic study. Radiology 251:122-133 61. Noguchi M, Stamey TA, McNeal JE, Yemoto CM (2001) Relationship between systematic biopsies and histological features of 222 radical prostatectomy specimens: lack of prediction of tumor significance for men with nonpalpable prostate cancer. J Urol 166:104-109 62. Kumar V, Jagannathan NR, Kumar R et al (2007) Transrectal ultrasound-guided biopsy of prostate voxels identified as suspicious of malignancy on three-dimensional (1)H MR spectroscopic imaging in patients with abnormal digital rectal examination or raised prostate specific antigen level of 4-10 ng/ml. NMR Biomed 20:11-20 63. Prando A, Kurhanewicz J, Borges AP et al (2005) Prostatic biopsy directed with endorectal MR spectroscopic imaging findings in patients with elevated prostate specific antigen levels and prior negative biopsy findings: early experience. Radiology 236:903-910 64. Anastasiadis AG, Lichy MP, Nagele U et al (2006) MRI-guided biopsy of the prostate increases diagnostic performance in men with elevated or increasing PSA levels after previous negative TRUS biopsies. Eur Urol Oct 50:738-748 65. Beyersdorff D, Winkel A, Hamm B et al (2005) MR imagingguided prostate biopsy with a closed MR unit at 1.5 T: initial results. Radiology 234:576-581 66. Engelhard K, Hollenbach HP, Kiefer B et al (2006) Prostate biopsy in the supine position in a standard 1.5-T scanner under real time MR-imaging control using a MR-compatible endorectal biopsy device. Eur Radiol 16:1237-1243 67. Hambrock T, Futterer JJ, Huisman HJ et al (2008) Thirty-twochannel coil 3T magnetic resonance-guided biopsies of prostate tumor suspicious regions identified on multimodality 3T magnetic resonance imaging: technique and feasibility. Invest Radiol;43:686-694 68. Hambrock T (2009) The Value of 3 Tesla Magnetic Resonance Imaging Guided Prostate Biopsies in Men with repeptitive Negative Biopsies and Elevated PSA 69. Pondman KM, Futterer JJ, ten HB et al (2008) MR-guided biopsy of the prostate: an overview of techniques and a systematic review. Eur Urol 54:517-527 70. Gleason DF, Mellinger GT (1974) Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 111(1):58-64 71. Noguchi M, Stamey TA, McNeal JE, Yemoto CM (2001) Relationship between systematic biopsies and histological features of 222 radical prostatectomy specimens: lack of prediction of tumor significance for men with nonpalpable prostate cancer. J Urol 166:104-109

139

72. Ruijter E, van LG, Miller G et al (2000) Errors in histological grading by prostatic needle biopsy specimens: frequency and predisposing factors. J Pathol 192:229-233 73. Wang L, Mazaheri Y, Zhang J et al (2008) Assessment of biologic aggressiveness of prostate cancer: correlation of MR signal intensity with Gleason grade after radical prostatectomy. Radiology 246:168-176 74. Ikonen S, Karkkainen P, Kivisaari L (2000) Magnetic resonance imaging of prostatic cancer: does detection vary between high and low gleason score tumors? Prostate 43:43-48 75. Zakian KL, Sircar K, Hricak H et al (2005) Correlation of proton MR spectroscopic imaging with gleason score based on step-section pathologic analysis after radical prostatectomy. Radiology 234:804-814 76. deSouza NM, Riches SF, Vanas NJ et al (2008) Diffusionweighted magnetic resonance imaging: a potential non-invasive marker of tumour aggressiveness in localized prostate cancer. Clin Radiol 63:774-782 77. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB (2002) Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer. J Magn Reson Imaging 16:451-463 78. Tamada T, Sone T, Jo Y (2008) Apparent diffusion coefficient values in peripheral and transition zones of the prostate: comparison between normal and malignant prostatic tissues and correlation with histologic grade. J Magn Reson Imaging 28: 720-726 79. Hosseinzadeh K, Schwarz SD (2004) Endorectal diffusionweighted imaging in prostate cancer to differentiate malignant and benign peripheral zone tissue. J Magn Reson Imaging 20:654-661 80. Wang L, Mullerad M, Chen HN et al (2004) Prostate cancer: incremental value of endorectal MR imaging findings for prediction of extracapsular extension. Radiology 232:133-139 81. Sonnad SS, Langlotz CP, Schwartz JS (2001) Accuracy of MR imaging for staging prostate cancer: a meta-analysis to examine the effect of technologic change. Acad Radiol 8:149-157 82. Engelbrecht MR, Jager GJ, Laheij RJ et al (2002) Local staging of prostate cancer using magnetic resonance imaging: a meta-analysis. Eur Radiol 12:2294-2302 83. Cornud F, Flam T, Chauveinc L et al (2002) Extraprostatic spread of clinically localized prostate cancer: factors predictive of pT3 tumor and of positive endorectal MR imaging examination results. Radiology 224:203-210 84. Langlotz CP, Schnall MD, Malkowicz SB, Schwartz JS (1996) Cost-effectiveness of endorectal magnetic resonance imaging for the staging of prostate cancer. Acad Radiol 1:S24-S27 85. Jager GJ, Severens JL, Thornbury JR et al (2000) Prostate cancer staging: should MR imaging be used? – A decision analytic approach. Radiology 215:445-451 86. Wang L, Hricak H, Kattan MW et al (2006) Prediction of organ-confined prostate cancer: incremental value of MR Imaging and MR spectroscopic imaging to staging nomograms. Radiology 238:597-603 87. Wang L, Hricak H, Kattan MW et al (2007) Prediction of seminal vesicle invasion in prostate cancer: incremental value of adding endorectal MR imaging to the Kattan nomogram. Radiology 242:182-188 88. McKenna DA, Coakley FV, Westphalen AC et al (2008) Prostate Cancer: Role of Pretreatment MR in Predicting Outcome after External-Beam Radiation Therapy – Initial Experience. Radiology 247:141-146 89. Ren J, Huan Y, Li F et al (2009) Combined T2-weighted and diffusion-weighted MRI for diagnosis of urinary bladder invasion in patients with prostate carcinoma. J Magn Reson Imaging 30:351-356 90. Ren J, Huan Y, Wang H et al (2009) Seminal vesicle invasion in prostate cancer: prediction with combined T2-weighted and diffusion-weighted MR imaging. Eur Radiol 19:2481-2486

140

Jelle O. Barentsz, Stijn W.T.P.J. Heijmink, Christina Hulsbergen-van der Kaa, Caroline Hoeks, Jurgen J. Futterer

91. Yu KK, Scheidler J, Hricak H et al (1999) Prostate cancer: prediction of extracapsular extension with endorectal MR imaging and three-dimensional proton MR spectroscopic imaging. Radiology 213:481-488 92. Mullerad M, Hricak H, Wang L et al (2004) Prostate cancer: detection of extracapsular extension by genitourinary and general body radiologists at MR imaging. Radiology 232:140-146 93. Fütterer JJ, Engelbrecht MR, Huisman HJ et al (2005) Staging Prostate Cancer with Dynamic Contrast-enhanced Endorectal MR Imaging prior to Radical Prostatectomy: Experienced versus Less Experienced Readers. Radiology 237:541-549 94. Wang L, Zhang J, Schwartz LH et al (2007) Incremental value of multiplanar cross-referencing for prostate cancer staging with endorectal MRI. AJR Am J Roentgenol 188:99-104 95. Akin O, Riedl CC, Ishill NM (2009 Interactive dedicated training curriculum improves accuracy in the interpretation of MR imaging of prostate cancer. Eur Radiol Nov 17 96. Fütterer JJ, Heijmink SW, Scheenen TW et al (2006) Prostate cancer: local staging at 3-T endorectal MR imaging – early experience. Radiology 238:184-191 97. Heijmink SWTPJ, Futterer JJ, Hambrock T et al (2007) Prostate Cancer: Body-Array versus Endorectal Coil MR Imaging at 3 T-Comparison of Image Quality, Localization, and Staging Performance. Radiology 244:184-195 98. Partin AW, Mangold LA, Lamm DM et al (2001) Contemporary update of prostate cancer staging nomograms (Partin Tables) for the new millennium. Urology 2001 58:843-848 99. Narayan P, Gajendran V, Taylor SP et al (1995) The role of transrectal ultrasound-guided biopsy-based staging, preoperative serum prostate-specific antigen, and biopsy Gleason score in prediction of final pathologic diagnosis in prostate cancer. Urology 46:205-212 100. Salo JO, Kivisaari L, Rannikko S, Lehtonen T (1986) The value of CT in detecting pelvic lymph node metastases in cases of bladder and prostate carcinoma. Scand J Urol Nephrol 20:261-265 101. Jager GJ, Barentsz JO, Oosterhof GO et al (1996) Pelvic adenopathy in prostatic and urinary bladder carcinoma: MR imaging with a three-dimensional TI-weighted magnetization-prepared-rapid gradient-echo sequence. AJR Am J Roentgenol 167:1503-1507 102. Heesakkers RA, Hovels AM, Jager GJ et al (2008) MRI with a lymph-node-specific contrast agent as an alternative to CT scan and lymph-node dissection in patients with prostate cancer: a prospective multicohort study. Lancet Oncol 9:850-856 103. Wolf JS, Jr., Cher M, Dall’era M et al (1995) The use and accuracy of cross-sectional imaging and fine needle aspiration cytology for detection of pelvic lymph node metastases before radical prostatectomy. J Urol Mar 153:993-999 104. Hovels AM, Heesakkers RA, Adang EM et al (2008)The diagnostic accuracy of CT and MRI in the staging of pelvic lymph nodes in patients with prostate cancer: a meta-analysis. Clin Radiol 63:387-395 105. Oyen RH, Van Poppel HP, Ameye FE et al (1994) Lymph node staging of localized prostatic carcinoma with CT and CT-guided fine-needle aspiration biopsy: prospective study of 285 patients. Radiology 190:315-322 106. Tiguert R, Gheiler EL, Tefilli MV et al (1999) Lymph node size does not correlate with the presence of prostate cancer metastasis. Urology 53:367-371 107. Levran Z, Gonzalez JA, Diokno AC et al (1995) Are pelvic computed tomography, bone scan and pelvic lymphadenectomy necessary in the staging of prostatic cancer? Br J Urol 75:778-781 108. Wolf JS Jr, Cher M, Dall’era M et al (1995) The use and accuracy of cross-sectional imaging and fine needle aspiration cytology for detection of pelvic lymph node metastases before radical prostatectomy. J Urol 153:993-999

109. Harisinghani MG, Barentsz J, Hahn PF et al (2003) Noninvasive detection of clinically occult lymph-node metastases in prostate cancer. N Engl J Med 2003 348:2491-2499 110. Heesakkers RA, Jager GJ, Hövels AM et al (2009) Prostate cancer: detection of lymph node metastases outside the routine surgical area with ferumoxtran-10-enhanced MR imaging. Radiology 251:408-414 111. Will O, Purkayastha S, Chan C et al (2005) Diagnostic precision of nanoparticle-enhanced MRI for lymph-node metastases: a meta-analysis. Lancet Oncol 7:52-60 112. O’Donoghue EP, P Constable AR, Sherwood T et al (1978) Bone scanning and plasma phosphatases in carcinoma of the prostate. Br J Urol 50:172-177 113. Rybak LD, Rosenthal DI (2001) Radiological imaging for the diagnosis of bone metastases. Quarterly Journal of Nuclear Medicine 45:53-64 114. Lecouvet FE, Geukens D, Stainier A et al (2007) Magnetic resonance imaging of the axial skeleton for detecting bone metastases in patients with high-risk prostate cancer: diagnostic and cost-effectiveness and comparison with current detection strategies. J Clin Oncol 25:3281-3287 115. Tombal B, Rezazadeh A, Therasse P et al (2005) Magnetic resonance imaging of the axial skeleton enables objective measurement of tumor response on prostate cancer bone metastases. Prostate 65:178-187 116. Cumming J, Hacking N, Fairhurst J et al (1990) Distribution of bony metastases in prostatic carcinoma. Br J Urol 66:411414 117. Pound CR, Brawer MK, Partin AW (2001) Evaluation and treatment of men with biochemical prostate-specific antigen recurrence following definitive therapy for clinically localized prostate cancer. Rev Urol 3:72-84 118. Horwitz EM, Vicini FA, Ziaja EL et al (1998) The correlation between the ASTRO Consensus Panel definition of biochemical failure and clinical outcome for patients with prostate cancer treated with external beam irradiation. American Society of Therapeutic Radiology and Oncology. Int J Radiat Oncol Biol Phys 41:267-272 119. Heidenreich A, Aus G, Bolla M et al (2008) EAU guidelines on prostate cancer. Eur Urol 53:68-80 120. Roach M, III, Hanks G, Thames H et al (2006) Defining biochemical failure following radiotherapy with or without hormonal therapy in men with clinically localized prostate cancer: recommendations of the RTOG-ASTRO Phoenix Consensus Conference. Int J Radiat Oncol Biol Phys 65:965-974 121. Crook J, Robertson S, Collin G et al (1993) Clinical relevance of trans-rectal ultrasound, biopsy, and serum prostate-specific antigen following external beam radiotherapy for carcinoma of the prostate. Int J Radiat Oncol Biol Phys 27:31-37 122. Leventis AK, Shariat SF, Slawin KM (2001) Local recurrence after radical prostatectomy: correlation of US features with prostatic fossa biopsy findings. Radiology 219:432-439 123. Chan TW, Kressel HY (1991) Prostate and seminal vesicles after irradiation: MR appearance. J Magn Reson Imaging 1:503-511 124. Sala E, Eberhardt SC, Akin O et al (2006) Endorectal MR imaging before salvage prostatectomy: tumor localization and staging. Radiology 238:176-183 125. Coakley FV, Teh HS, Qayyum A et al (2004) Endorectal MR imaging and MR spectroscopic imaging for locally recurrent prostate cancer after external beam radiation therapy: preliminary experience. Radiology 233:441-448 126. Haider MA, Chung P, Sweet J et al (2008) Dynamic contrastenhanced magnetic resonance imaging for localization of recurrent prostate cancer after external beam radiotherapy. Int J Radiat Oncol Biol Phys 70:425-430 127. Rouviere O, Valette O, Grivolat S et al (2004) Recurrent prostate cancer after external beam radiotherapy: value of contrast-enhanced dynamic MRI in localizing intraprostatic tumor – correlation with biopsy findings. Urology 63:922-927

Magnetic Resonance Imaging of Prostate Cancer

128. Kim CK, Park BK, Lee HM (2009) Prediction of locally recurrent prostate cancer after radiation therapy: Incremental value of 3T diffusion-weighted MRI. J Magn Reson Imaging 29:391-397 129. Sella T, Schwartz LH, Swindle PW et al (2004) Suspected local recurrence after radical prostatectomy: endorectal coil MR imaging. Radiology 231:379-385 130. Casciani E, Polettini E, Carmenini E et al (2008) Endorectal and dynamic contrast-enhanced MRI for detection of local recurrence after radical prostatectomy. AJR Am J Roentgenol 190:1187-1192

141

131. Cirillo S, Petracchini M, Scotti L et al (2009) Endorectal magnetic resonance imaging at 1.5 Tesla to assess local recurrence following radical prostatectomy using T2-weighted and contrast-enhanced imaging. Eur Radiol 19:761-769 132. Sciarra A, Panebianco V, Salciccia S et al (2008) Role of dynamic contrast-enhanced magnetic resonance (MR) imaging and proton MR spectroscopic imaging in the detection of local recurrence after radical prostatectomy for prostate cancer. Eur Urol 54:589-600

IDKD 2010-2013

Imaging of the Male Pelvis: The Scrotum Brent J. Wagner Department of Radiology, The Reading Hospital and Medical Center, West Reading, PA, USA

Introduction Although there are many potential ways one might compartmentalize the imaging of scrotal disease (including pattern recognition, etiologies, and clinical presentation), this chapter focuses on the requisite entities that, based on their imaging and clinical features, constitute the core knowledge needed for the radiologist when faced with an abnormal scrotal sonogram. Hence, while not comprehensive, this review addresses diseases that are of interest because of a distinctive combination of clinical relevance and characteristic imaging features. With the continued evolution of high-resolution linear transducers over the past 15 years, clinicians, patients, and imaging technicians are able to rely on ultrasound as a tool with near 100% sensitivity for significant intrascrotal pathology. Not only is the technique noninvasive and relatively inexpensive, it is also highly effective in both the detection and the characterization of a wide variety of disorders involving the scrotum.

Fig. 1. Seminoma. A palpable nodule in a 33-year-old male. Sonogram demonstrates a small, homogeneous, well-marginated intratesticular mass

Intratesticular Disorders Neoplastic Most neoplasms of the testis are germ cell tumors (GCTs), and the vast majority of these are malignant. Among the various histological types, seminoma is the most common to occur as a pure tumor, accounting for approximately 40% of all testicular neoplasms. Most of the remainder are mixed tumors, containing two or more histological types. Lesions typically present as a palpable mass, although some aggressive tumors may present with metastatic foci to lung or bone or as nodal masses. The majority of GCTs are hypoechoic relative to the homogeneous medium-high echogenicity background of the testis (Fig. 1). Calcifications are seen in at least onethird of cases, especially in non-seminomatous tumors, which also tend to exhibit more heterogeneity. The management of GCTs nearly always involves orchiectomy for the definitive diagnosis and treatment of the local disease. The well-established radiosensitivity of

seminomas is the most important distinguishing clinical feature of the two groups of tumors (seminomatous vs. non-seminomatous). This feature accounts for the previous, widespread use of prophylactic retroperitoneal radiotherapy for seminomas, even when imaging studies suggested that the lesion was confined to the testis. Over the last decade, however, there has been a trend toward surveillance (using imaging and serum tumor markers) in the 75% of patients with seminomas who are diagnosed as having stage I disease [1, 2]. For non-seminomatous tumors, computed tomography (CT) is often used to determine whether patients require retroperitoneal lymph node dissection. Patients with (1) nodal involvement (either enlarged on CT, or confirmed on pathological assessment) and/or (2) hematogenous metastasis are generally treated with chemotherapy. The prognosis for patients with non-seminomatous tumors is more guarded than that for patients with seminoma, although patients with stage I disease (limited to the scrotum) have 5-year survival rates approaching 90%. While

Imaging of the Male Pelvis: The Scrotum

attempts have been made to use imaging to differentiate GCTs [3], the initial management of both types is surgical (orchiectomy) and the exercise of pre-operatively differentiating seminoma vs. non-seminoma is typically not relevant [1]. Gonadal stromal tumors are less common (<10% of lesions) and are often incidental findings. Both Leydig cell and Sertoli tumors are occasionally the cause of gynecomastia. The vast majority of gonadal stromal tumors are benign but they have gray-scale sonographic features similar to the more sinister group of GCTs. Patients with lymphoma of the testis are typically older than those with GCTs. Lymphomas are rarely demonstrated in patients without a known diagnosis of systemic lymphoma. Hypoechoic, multifocal (or geographic), bilateral lesions, which often have increased vascularity compared with GCT’s, are characteristically seen [4]. While not strictly neoplastic, five entities deserve brief discussion in the context of neoplastic intratesticular tumors: 1. Simple cysts are typically peripheral, multiple, and contiguous with the tunica albuginea (more central cysts are usually a manifestation of tubular ectasia). These have no malignant potential and are rarely of clinical significance, although they may be palpable and clinically mimic a GCT. Rarely, they may be complicated by hemorrhage; however, most cases are easily diagnosed as simple cysts. Features that raise concern for malignancy (and indicate the need for prompt surgical removal) include solitary lesions complicated by internal soft tissue, calcification, or a thick irregular wall. 2. Tubular ectasia represents a dilatation of the rete testis as it converges along the mediastinum testis. Patients may have a history of prior epididymitis or vasectomy, but most have no specific symptoms. An ovoid or linear area of decreased echogenicity, contiguous with the epididymis, can usually be differentiated from a neoplasm based on the pattern of branching, anechoic channels (Fig. 2). In occasional cases that may mimic neoplasm, T2-weighted magnetic resonance imaging will show a hyperintense focus (in contrast to the hypointensity that is characteristic of most testicular neoplasms) [5]. 3. Epidermoid cyst is not considered to be a neoplasm by most authorities. Instead, it is an inclusion cyst, lined by squamous epithelium and filled with keratinized debris. Although it accounts for <10% of intratesticular masses, the differentiation of epidermoid cyst from germ cell neoplasm is important in order to avoid radical surgery for this lesion, which can, instead, be effectively diagnosed and treated with a more limited procedure (enucleation, sparing the testis) [5, 6]. Epidermoid cyst can be recognized in many instances by its “onion-skin” appearance, i.e., alternating concentric rings of hypoand hyperechogenicity [5, 7]. A minority of cysts will show wall calcification. The absence of Doppler flow may be of some help (the presence of central flow excludes the diagnosis), but one should remember that some malignant neoplasms, especially when small, will

143

Fig. 2. Tubular ectasia. The somewhat focal decreased echogenicity resembles a mass but the non-spherical shape and the elongate anechoic channels are characteristic and do not suggest the need for tissue diagnosis

not have flow that can be detected sonographically. Additionally, the onion-skin appearance is neither sensitive (it is seen in approximately half the cases) nor entirely specific (some neoplasms, including teratoma, may rarely have the same feature) [7]. 4. Congenital adrenal rest tumors are seen in patients with poorly controlled congenital adrenal hyperplasia (CAH). Aberrant adrenal cortical cells migrate with gonadal tissue during fetal development and may hypertrophy at some point during the disease. The sonographic appearance is extremely variable but is typically distinguished from that of malignant GCT by the multiplicity of the lesions, their eccentricity, and their contiguity with the mediastinum testis. Treatment is glucocorticoid replacement or, rarely, surgery (partial orchiectomy) [8]. 5. Testicular microlithiasis (TML) describes the presence of numerous small (1-2 mm) calcifications scattered throughout the testicular parenchyma. This idiopathic condition is associated with oligospermia in a minority of patients but in most it is merely an incidental finding [5]. Several published reports concluded that there is a small but real increased risk of testicular GCT in patients with TML (Fig. 3) [9]. This prompted multiple authors to suggest that patients should be carefully screened for coexistent tumors and followed sonographically at annual intervals for the development of malignancy. Recently, there has been a shift away from this recommendation, based on the high cost of serial sonographic exams in a large number of men who have an increased but still very low risk of developing a germ cell tumor [9, 10]. Nonetheless, the need for sonographic follow-up remains contentious within the field.

144

Fig. 3. Mixed germ cell tumor arising in a background of testicular microlithiasis

Brent J. Wagner

paratesticular tissues may show reactive hyperemia, despite the absence (or significant decrease) in Doppler flow to the testis itself. Traumatic injury to the testis is frequently an indication for sonography. The differentiation of testicular rupture from a less serious injury (for example, merely a hematoma with an intact tunica albuginea) generally represents the most important role of the imager. Findings of a heterogeneous echo-texture within the testis, testicular contour abnormality, and disruption of the tunica albuginea are considered very sensitive and specific for the diagnosis of testicular rupture [14]. Sarcoidosis is an unusual cause of a scrotal mass and the scrotum may rarely be the presenting site of this disease in some patients. The pattern of multiple, hypoechoic, intratesticular masses associated with epididymal enlargement and heterogeneity is characteristic, although it may also be seen in lymphoma. A useful differentiating feature is the relative prominence of testicular disease in lymphoma (sarcoidosis tends to affect the epididymis more than the testis, as discussed below) [5].

Non-neoplastic

Extratesticular Disorders Infection typically begins in the epididymis. Epididymoorchitis is usually a clinical diagnosis, with sonography sometimes used to rule out torsion or abscess. Sonographic findings are often absent, although some patients will show a hydrocele (with or without complicating elements indicating pyocele). There may also be increased Doppler flow to the epididymis. Progression to involvement of the testis; which occurs in a minority of cases, may result in abscess formation or infarction. Ischemia/infarction may result from torsion or, less commonly, from a variety of other causes, including vasculitis, diabetes, or orchitis, which may be segmental [11]. Patients with torsion typically have an acute clinical presentation that includes severe unilateral scrotal pain, often following minor trauma or physical exertion. The typical finding of ischemia/infarction is an asymmetrical decrease in the color or amplitude (power) of the Doppler signal on the symptomatic side. However, subtle variations of arterial spectral Doppler waveforms may be seen early, including the absence of the dicrotic notch and/or increased resistance (decreased or absent diastolic flow) [12]. The latter finding may also be seen in the early stages of torsion, when venous flow is altered but arterial flow is still observed using color Doppler. Emergency surgery is indicated to detorse and save the testis; however, when gray-scale findings are present, including heterogeneity and decreased echogenicity, the ischemia has almost always progressed to infarction. At this point, testicular salvage is not possible [13]. Important pitfalls in Doppler evaluation must be recognized in order to avoid misdiagnosis. The examiner must be careful not to alter the Doppler settings (gain, scale, etc.) when comparing normal (asymptomatic) testis to the painful side. One must also remember that

Neoplastic In adults, malignant extratesticular neoplasms are rare and have a non-specific appearance [15]. Mesothelioma is an uncommon neoplasm that usually presents as a hydrocele, with soft-tissue nodules of the tunica vaginalis. Alternatively, it may present as a large heterogeneous mass that may be difficult to separate from the testis. Mesotheliomas tend to occur in individuals who are decades older than those typically diagnosed with testicular GCTs. Lymphoma may occasionally involve the epididymis, although in the majority of patients this does not lead to a diagnostic dilemma as: (1) the patient will be known to have lymphoma and (2) there will be coexistent involvement of the testis itself. Very rarely, solid tumors may metastasize to the epididymis. Affected patients almost always have advanced metastatic disease elsewhere throughout the body. The most common extratesticular intrascrotal neoplasm is lipoma, which arises from the spermatic cord and can often be diagnosed clinically based on palpation. Adenomatoid tumors are nearly as common as lipomas, accounting for about one-third of extratesticular masses. These are benign, but may be surgically removed either to establish the diagnosis or because of local pain or tenderness. They are solid, well-marginated lesions that are typically <20 mm in size. They most frequently arise from the epididymis. Papillary cystadenomas of the epididymis are seen in about one-quarter of patients with von Hippel-Lindau disease (the lesions are otherwise extremely rare). They are typically solid, measure between 1 and 5 cm, and may be indistinguishable from adenomatoid tumors [15].

Imaging of the Male Pelvis: The Scrotum

Sarcoidosis is more likely to affect the epididymis than the testis. More than one-third of patients will have bilateral disease. Although discrete nodules are occasionally seen, the appearance is more commonly one of heterogeneous enlargement. A diagnostic pattern that may be of use in a previously undiagnosed patient with hilar adenopathy – which could be either lymphoma or sarcoid – is to compare the testicular and the epididymal involvement: in sarcoidosis, the degree of epididymal disease typically exceeds that of testis involvement, whereas in lymphoma the converse is expected.

References 1. Sohaib SA, Koh DM, Husband JE (2008) The role of imaging in the diagnosis, staging, and management of testicular cancer. AJR Am J Roentgen 191:387-395 2. Krohmer SJ, McNulty NJ, Schned AR (2009) Best cases from the AFIP: testicular seminoma with lymph node metastases. Radiographics 29:2177-2183 3. Tsili AC, Tsampoulas C, Giannakopoulos X et al (2007) MRI in the histologic characterization of testicular neoplasms. AJR Am J Roentgen 189:W331-W337 4. Mazzu D, Jeffrey RB, Ralls PW (1995) Lymphoma and leukemia involving the testicles: findings on gray-scale and color Doppler sonography. AJR Am J Roentgen 164: 645-647

145

5. Woodward PJ, Sohaey R, O’Donoghue MJ, Green DE (2002) Tumors and tumorlike lesions of the testis: radiologicpathologic correlation. Radiographics 22:189-216 6. Cho J-H, Chang J-C, Park B-H et al (2002) Sonographic and MR imaging findings of testicular epidermoid cysts. AJR Am J Roentgen 178:743-748 7. Maizlin ZV, Belenky A, Baniel J et al (2005) Epidermoid cyst and teratoma of the testis: sonographic and histologic similarities. J Ultrasound Med 24:1403-1409 8. Nagamine WH, Mehta SV, Vade A (2005) Testicular adrenal rest tumors in a patient with congenital adrenal hyperplasia: sonographic and magnetic resonance imaging findings. J Ultrasound Med 24:1717-1720 9. Lam DL, Gerscovich EO, Kuo MC, McGahan JP (2007) Testicular microlithiasis: our experience of 10 years. J Ultrasound Med 26:867-873 10. Costabile RA (2007) How worrisome is testicular microlithiasis? Curr Opin Urol 17:419-423 11. Fernández-Pérez GC, Tardáguila FM, Velasco M et al (2005) Radiologic findings of segmental testicular infarction. AJR Am J Roentgen 184:1587-1593 12. Dogra VS, Rubens DJ, Gottlieb RH, Bhatt S (2004) Torsion and beyond: new twists in spectral Doppler evaluation of the scrotum. J Ultrasound Med 23:1077-1085 13. Middleton WD, Middleton MA, Dierks M et al (1997) Sonographic prediction of viability in testicular torsion: preliminary observations. J Ultrasound Med 16:23-27 14. Bhatt S, Dogra VS (2008) Role of US in testicular and scrotal trauma. Radiographics 28:1617-1629 15. Woodward PJ, Schwab CM, Sesterhenn IA (2003) Extratesticular scrotal masses: radiologic-pathologic correlation. Radiographics 23:215-240

IDKD 2010-2013

Spread of Metastatic Disease in the Abdomen and Pelvis James A. Brink1, Ali Shirkhoda2 1 Department 2 Diagnostic

of Diagnostic Radiology, Yale University School of Medicine, New Haven, CT, USA Radiology, William Beaumont Hospital, Royal Oak, MI, USA

Introduction Basic knowledge of the normal intra-abdominal anatomy and of the anatomical variants is essential to understanding the spread of pathology within the peritoneum. Of special importance are constant landmarks, i.e., the anatomical relationships maintained and bounded by peritoneal and fascial attachments as well as by the abdominal adipose tissue. The peritoneal and extraperitoneal spaces and their fascial planes create complex three-dimensional structures with unique radiological characteristics. Intraperitoneal and extraperitoneal adipose tissue provides contrast interfaces between the organs and visceral structures. The intra-abdominal adipose also yields clues as to the spread and localization of many pathological conditions. The four different pathways for the spread of neoplastic diseases within the abdomen and pelvis are bloodborne metastasis, lymphatic extension, direct invasion, and intraperitoneal spread and seeding. Direct invasion may occur from contiguous primary tumors and usually implies that a locally aggressive tumor has broken through fascial planes. Direct invasion from non-contiguous primary tumors typically occurs via spread along the peritoneal ligaments and mesenteries. Intraperitoneal spread of malignancy occurs first by seeding of the peritoneal cavity with metastatic cells. Tumor spread occurs via the natural flow of ascitic fluid within the peritoneal spaces, which are defined by the peritoneal ligaments and mesenteries [1]. The peritoneum is the largest and the most complexly arranged serous membrane in the body. The potential space between the parietal peritoneum lining the abdominal wall and the visceral peritoneum enveloping the abdominal organs is called the peritoneal cavity. It consists of a main region, termed the greater sac, and a diverticulum, called the omental bursa or lesser sac, situated behind the stomach [2-4]. These two areas communicate via the epiploic foramen (foramen of Winslow) and normally contain only a small amount of fluid; therefore, they are normally not visible on cross-sectional imaging, except when there is ascites (Fig. 1) or they are filled inadvertently by contrast (Fig. 2). Fluid dynamics, respiratory

Fig. 1. The superior recess of the lesser sac is filled with ascetic fluid and seen here as it extends superiorly (long arrow) to the lower mediastinum, displacing the esophagus (short arrow) to the left side

Fig. 2. In this patient with a gastrostomy (PEG) tube, diluted watersoluble contrast material was injected through the tube before an abdominal CT scan. However, unbeknownst to the nurse, the tip of the tube was out of the stomach such that contrast was injected into the peritoneal cavity. The CT scan reveals opacification of the peritoneal cavity, with the exception of the bare area of the liver (double arrows). The falciform ligament (single arrow) is outlined by contrast material

147

Spread of Metastatic Disease in the Abdomen and Pelvis

motion, gravity, and anatomical barriers dictate the spread of disease processes within the peritoneal cavity and their appearance on cross-sectional imaging. Tumors that arise within the abdominal cavity have the propensity to spread (1) via the peritoneal ligaments and mesenteries that suspend their organs of origin, either by lymphatic extension or direct invasion, or (2) via seeding of the peritoneal spaces in which they reside [5]. A working knowledge of the anatomy of the peritoneal ligaments, mesenteries, and spaces, although complex, permits one to narrow the diagnostic possibilities when metastatic disease is recognized. The peritoneal reflections in the abdomen comprise four mesenteries, eight ligaments, and two omenta organized as follows: 1. four mesenteries: the small bowel mesentery, the transverse mesocolon, the sigmoid mesocolon, and the mesoappendix; 2. eight ligaments: right coronary, left coronary, falciform, hepatoduodenal, duodenocolic, gastrosplenic, splenorenal, and phrenicocolic ligaments; 3. two omenta: lesser omentum and greater omentum. The peritoneal ligaments suspend and support the intraperitoneal organs and subdivide the peritoneal cavity into interconnected compartments that dictate the flow of fluid and location of disease (Figs. 3, 4) [2, 6]. These reflections are generally recognizable on computed tomography (CT) scans as fat-containing structures, either by their typical location and organ relationships or by the landmarks provided by their major constituent vessels. The ligaments can serve as conduits or as barriers to the spread of disease [7].

Coronary Ligament

Transverse Mesocolon Small Bowel Mesentery Sigmoid Mesocolon

Fig. 3. Posterior peritoneal reflections and recesses. Intraperitoneal fluid flows naturally from the pelvis to the upper abdomen, preferentially through the right rather than the left paracolic gutter owing to the broader diameter of the former. In addition, flow in the left paracolic gutter is cut off from reaching the left subphrenic space by the phrenicocolic ligament. The transverse mesocolon divides the abdomen into supra- and inframesocolic spaces. In the right inframesocolic space, fluid is impeded from draining into the pelvis via the small bowel mesentery. Owing to natural holdup of fluid at the root of the small bowel mesentery and sigmoid mesocolon, these structures are naturally predisposed to involvement with serosal-based metastases in the setting of peritoneal carcinomatosis (Reprinted with permission from [6])

Left lobe of liver

Metastatic Spread via the Peritoneal Ligaments The major attachments of the upper abdomen include the lesser omentum, the gastrosplenic ligament, and the splenorenal ligament. The lesser omentum is subdivided into the gastrohepatic ligament and the hepatoduodenal ligament. In the embryo, the gastrosplenic ligament gives rise to the gastrocolic ligament (also known as the greater omentum) and the transverse mesocolon.

Stomach

GSL

Gastrohepatic Ligament

Lesser SAC

P

The gastrohepatic ligament can be recognized on CT scans as a fatty plane that joins the lesser curvature of the stomach to the liver. It extends from the fissure for the ligamentum venosum to the porta hepatis and contains the left gastric artery, coronary vein, and associated lymphatics. The size defining nodal enlargement in this region is somewhat smaller than elsewhere in the abdomen; nodes in the gastrohepatic ligament are generally considered abnormal when they exceed 8 mm in diameter [3]. On occasion, pathology in the gastrohepatic ligament may be mimicked by the absence of opacification of the bowel loops, the pancreatic neck, or

LK

Spleen Splenic artery SRL

Fig. 4. Peritoneal components in the left upper quadrant. The gastrosplenic ligament (GSL) and splenorenal ligament (SRL) are clearly seen (From [2])

148

James A. Brink, Ali Shirkhoda

Fig. 5. Intraperitoneal seeding of numerous hydatid cysts. This patient had a hydatid cyst in the liver that ruptured during resection, resulting in extensive peritoneal implants, including in the lesser omentum, which is seen here between the stomach, liver, and spleen

Fig. 6. Invasive pancreatic carcinoma arising from the pancreatic tail, with numerous hematogenous metastases to the liver. A portocaval lymph node at the base of the hepatoduodenal ligament (arrow) is not enlarged according to size criteria but contains metastatic disease, as evidenced by its central necrosis

the papillary process of the caudate lobe of the liver projecting into the expected plane of the gastrohepatic ligament [8, 9]. The gastrohepatic ligament provides an important conduit of disease from the stomach to the liver in that the subperitoneal areolar tissue within the ligament is continuous with the Glisson capsule (the perivascular fibrous capsule within the liver). Thus, gastric malignancy can spread directly into the left lobe of the liver and vice versa via this pathway. Neoplastic or infectious conditions that may spread throughout the entire peritoneum also involve this ligament. On CT, the abnormality is seen between the stomach and the liver (Fig. 5). Common neoplasms spreading via the gastrohepatic ligament include nodal metastases from gastric, esophageal, breast, pancreatic, and lung cancer as well as nodal involvement of lymphoma. Gastric and esophageal cancer can directly invade the gastrohepatic ligament and spread into the left hepatic lobe [3].

central necrosis suggests the presence of tumor within these nodes (Fig. 6) [11, 12]. A broad range of tumors may spread via the hepatoduodenal ligament. Liver or biliary cancer, whether primary or metastatic, may spread in an antegrade fashion through lymphatics in the hepatoduodenal ligament to deposit in periduodenal or peripancreatic lymph nodes. Similarly, malignant disease in the nodes about the superior mesenteric artery (commonly involved in pancreatic and colon cancer) can spread in a retrograde fashion up the lymphatics in the hepatoduodenal ligament. Lymphoma can involve these nodes as well. Primary gastric cancer arising in the lesser curvature of the stomach can directly spread through the gastrohepatic ligament to the hepatoduodenal ligament and then to peripancreatic and periduodenal nodes. Vascular complications related to the portal vein and hepatic artery can result; portal venous thrombosis and hepatic arterial pseudoaneurysms may occur in advanced cases owing to their coexistence in the hepatoduodenal ligament [2, 10].

Hepatoduodenal Ligament The hepatoduodenal ligament is the free edge of the gastrohepatic ligament along its rightward aspect. It contains important structures of the porta hepatis, including the common bile duct, hepatic artery, and portal vein. The hepatoduodenal ligament extends from the flexure between the first and second duodenum to the porta hepatis; the foramen of Winslow is immediately posterior to this ligament, permitting communication between the greater and lesser sacs [10]. The nodes of the foramen of Winslow, or portocaval space, have an unusual morphology such that their transverse dimension is greater than their anteroposterior dimension. Generally, the upper limit of normal for the latter is 1.0-1.3 cm, whereas the former can be up to 2.0 cm in width. Size criteria are somewhat less helpful than in other lymph nodes. In the absence of frank enlargement, a more spherical shape or

Gastrosplenic and Splenorenal Ligaments In the embryo, the gastrosplenic ligament is a long ligamentous attachment between the stomach and the retroperitoneum. It gives rise to the gastrocolic ligament (greater omentum) and the transverse mesocolon. In the adult, the gastrosplenic ligament is a thin ligamentous attachment between the greater curvature of the stomach and the splenic hilus (Fig. 4). It contains the left gastroepiploic and short gastric vessels as well as associated lymphatics. The gastrosplenic ligament is continuous with the gastrocolic ligament inferiorly and medially and with the splenorenal ligament posteriorly and medially [13, 14]. As such, it provides an important pathway of communication between the stomach, spleen, and retroperitoneum. Gastric malignancies commonly spread through the gastrosplenic

149

Spread of Metastatic Disease in the Abdomen and Pelvis

b

a

Fig. 7 a, b. Gastric adenocarcinoma invading the spleen via the gastrosplenic ligament. a Initial contrast-enhanced CT scan reveals circumferential tumor involving the gastric fundus. b A repeat CT scan 6 months later shows invasion and dissection of the spleen secondary to tumor spread via the gastrosplenic ligament

ligament (Fig. 7), thereby involving the spleen and ultimately resulting in disease about the tail of the pancreas. Conversely, pancreatic neoplasms may spread via the splenorenal ligament to the gastrosplenic ligament and involve the greater curvature of the stomach [2].

Gastrocolic Ligament The gastrocolic ligament (or greater omentum) joins the greater curvature of the stomach to the transverse colon. On the left, it is continuous with the gastrosplenic ligament; on the right, it ends at the gastroduodenal junction, near the hepatoduodenal ligament. Since, developmentally, it results from fusion of the anterior and posterior leaves of the gastrosplenic ligament, it contains the four layers of peritoneum that invest the stomach and has a potential space within it (Fig. 8) [15].

The gastrocolic ligament contains the gastroepiploic vessels and associated lymphatics. It provides an important conduit of malignant disease from the greater curvature of the stomach to the transverse colon and vice versa. When viewed in concert with the transverse mesocolon, a conduit exists between the greater curvature of the stomach and the retroperitoneum. In addition to allowing direct spread of disease between the stomach, transverse colon, and pancreas, the gastrocolic ligament serves as an important nidus for peritoneal metastases – as commonly occur with ovarian, gastric, colon, and pancreatic cancers [16, 17]. Finally, dilated veins within this ligament may represent gastroepiploic collaterals resulting from splenic venous compromise, such as might occur in the setting of invasive pancreatic tumors or in the presence of intraperitoneal tumors, which spread to the retroperitoneum via the transverse mesocolon (Fig. 9).

Transverse Mesocolon

RPS G P

The transverse mesocolon serves as a broad conduit of disease across the mid-abdomen; bare areas link the pancreas to the transverse colon, spleen, and small bowel. On the right, the transverse mesocolon is continuous with the duodenocolic ligament; in the middle, it is

D

TC

GCL J

Fig. 8. The gastrocolic ligament (GCL) joins the greater curvature of the stomach (G) to the transverse colon (TC). In concert with the transverse mesocolon, a pathway of disease is formed between retroperitoneal structures, such as the pancreas (P) and the duodenum (D), and the anterior aspect of the intraperitoneal cavity (Modified from [15])

Fig. 9. A pancreatic islet-cell tumor arising in the pancreatic tail has resulted in splenic vein thrombosis, with secondary short gastric venous collaterals in the gastrosplenic and splenorenal ligaments as well as gastroepiploic venous collaterals (arrow) in the gastrocolic ligament

150

James A. Brink, Ali Shirkhoda

continuous with the small bowel mesentery; and on the left, it is continuous with the phrenicocolic and splenorenal ligaments (Fig. 10). It contains the middle colic vessels and associated lymphatics. On CT, the transverse mesocolon may be recognized as a fatty plane at the level of the uncinate process. Pancreatic tumors often spread ventrally into the transverse mesocolon to involve

SRL

TM

PCL

SBM

Fig. 10. The transverse mesocolon (TM) provides an important conduit for the spread of disease across the mid-abdomen. It is continuous with the splenorenal ligament (SRL) and phrenicocolic ligament (PCL) on the left and with the duodenocolic ligament on the right. In its mid-portion, it is continuous with the small bowel mesentery (SBM) (Reprinted with permission from [6])

a

the transverse colon and have the propensity to continue through the gastrocolic ligament to involve the stomach (Fig. 11). Alternatively, they may spread through the transverse mesocolon to involve the proximal jejunum, just beyond the ligament of Treitz (Fig. 12). Like the gastrocolic ligament, a potential space exists within the transverse mesocolon due to embryological fusion of the gastrosplenic ligament with the embryological transverse mesocolon [13]. A less common but important route of spread also exists between the right colon and the periduodenal and peripancreatic nodes via the rightward aspect of the transverse mesocolon (duodenocolic ligament). This is important because lymphadenopathy in the periduodenal and peripancreatic regions may herald a right colon cancer when other, more common causes of lymphadenopathy in this region are excluded [2]. Thus, three routes of spread between the intraperitoneal viscera and retroperitoneum are provided by three pairs of ligaments. The gastrohepatic and hepatoduodenal ligaments link the liver and lesser curvature of the stomach to the retroperitoneum; the gastrosplenic and splenorenal ligaments link the superior greater curvature of the stomach and spleen to the retroperitoneum; and the gastrocolic and transverse mesocolon link the inferior greater curvature of the stomach and transverse colon to the retroperitoneum. The ligamentous pair in which metastatic disease is recognized can therefore suggest the organ of origin and, in the case of gastric cancer, the location of the primary tumor within the stomach.

b Fig. 11 a, b. Invasive pancreatic carcinoma invading the retroperitoneum, transverse mesocolon, and greater omentum. Transaxial image (a) and parasagittal reformation (b). M mass, P normal pancreas, S stomach, TC transverse colon. The transverse mesocolon (black arrows) provides a pathway for the spread of metastatic disease to the greater omentum (white arrows). Also noted is encasement of the left renal artery and ovarian vein (double arrows)

a

b

Fig. 12 a, b. An incidental pancreatic adenocarcinoma involving the splenic vein (a) was thought to be resectable owing to the lack of extrapancreatic involvement. At surgery, upon elevation of the transverse colon, the tumor was found to have penetrated the base of the transverse mesocolon and involved the proximal jejunum just beyond the ligament of Treitz (seen in retrospect in b, arrows)

151

Spread of Metastatic Disease in the Abdomen and Pelvis

Metastatic Spread via the Peritoneal Spaces Diseases such as ovarian carcinoma (Fig. 13 a) and lymphoma (Fig. 13 b) can diffusely extend through the peritoneum, with propensity for involvement of the greater omentum. However, the initiation and growth of seeded metastases on the peritoneal surfaces usually depend on the natural flow of ascites through the peritoneal spaces. Primary abdominal malignancies and secondary nodal metastases can break through the visceral peritoneum and shed cells into the peritoneal cavity (Fig. 13 c). Once intraperitoneal, such cells propagate through the peritoneal spaces along predicable routes. A thorough understanding of the anatomy of the peritoneal spaces may help refine differential diagnoses for the source of intra-abdominal metastases [18, 19]. By recognizing that a process is intraperitoneal, one may better predict its organ of origin and likely routes of spread (Fig. 13). The most dependent portion of the peritoneal cavity is in the pelvis. The cavity is anatomically continuous with the paracolic gutters and is subdivided into the midline pouch of Douglas (rectovaginal pouch in women, rec-

a

tovesical pouch in men), the lateral paravesical recesses, and the medial and lateral inguinal fossae. Due to the deep and dependent nature of the pelvic peritoneal cavity, many infections and half of all seeded metastases will involve the pouch of Douglas. Fluid flows preferentially to the right lower quadrant and from there to the inferior portion of the small bowel mesentery and right paracolic gutter.

Left Peritoneal Space The left peritoneal space can be subdivided into four compartments. Although these freely communicate with each other, the inflammatory nature of exudative fluid collections within them favors the development of fibrous adhesions, which may seal off one or more portions of the left peritoneal space from the others. The left anterior perihepatic space is limited on the right by the falciform ligament and on the left by the anterior wall of the stomach. It follows the posterior curve of the diaphragm and is limited posteriorly by the left coronary ligament (Fig. 14).

b

c

Fig. 13 a-c. Peritoneal malignancies. a Diffuse omental metastasis from a primary ovarian carcinoma associated with ascites. b Extensive involvement of the greater omentum by Burkitt’s lymphoma. c Seeding of the inferior aspect of the peritoneum (cul de sac) by gastric carcinoma (arrows)

a Fig. 14 a, b. The left (a) and right (b) perihepatic spaces are bounded posteriorly by the coronary ligaments. The reflections of the coronary ligaments mark the site of the non-peritonealized “bare area” of the liver. LL left lobe of the liver, LK left kidney, S stomach, TC transverse colon, P pancreas, D duodenum, Lu lung, L liver (right lobe), A adrenal, K kidney, C colon (Reprinted with permission from [6])

Left Subphrenic Space

Left Coronary Ligament

b

Rt. Coronary Lig. Superior reflection Inferior reflection L

A

LL

Lesser Omentum S

LK P

TC

D

Lesser Sac

D C

K

Rt. Posterior Perihepatic Space (Morison’s Pouch) Rt. Anterior Perihepatic Space

152

James A. Brink, Ali Shirkhoda

The left posterior perihepatic space (gastrohepatic recess) is limited on the left by the lateral wall of the stomach. This space follows the posterior margin of the left hepatic lobe deep into the fissure for the ligamentum venosum to form the posterior margin of the left hepatic lobe. Thus, it is in close proximity to the lesser curve of the stomach, the anterior wall of the duodenal bulb, and the anterior wall of the gallbladder [13]. Although the gastrohepatic recess is close to the lesser sac (divided from it by the lesser omentum), it is a portion of the left peritoneal space, while the lesser sac is a portion of the right peritoneal space (Fig. 15). This distinction is important in that lesser sac collections are very difficult to approach percutaneously whereas gastrohepatic recess collections are usually accessible by guiding a catheter along the inferior margin of the left hepatic lobe. The anterior left subphrenic space is in direct continuity with the left anterior perihepatic space, which forms its right boundary. Far to the left, on the anterolateral surface of the stomach, this space is limited by the greater omentum. This is a common site for fluid loculation in the setting of malignant ascites (Fig. 16).

Caudate lobe

Lesser omentum

Fissure for ligamentum venosum

Liver

IVC Ao

Diaphragm

Upper recess

The posterior left subphrenic (perisplenic) space is the posterior continuation of the anterior subphrenic space and generally surrounds the lateral and superior margins of the spleen. The “bare areas” of the spleen are reliably observed in perisplenic fluid collections [20-22]. Superiorly, the perisplenic space is entirely subphrenic and surrounds the top of the spleen (Fig. 16) [23].

Right Peritoneal Space There are three major subdivisions of the right peritoneal space: the right subphrenic space, the hepatorenal recess, and the lesser sac (Figs. 14, 15). The right subphrenic space occupies the smoothly contoured area between the superolateral margin of the liver and the right hemidiaphragm. The medial extension of this compartment is limited by the right coronary ligament, which is simply the right lateral margin of the liver’s bare area [24]. The hepatorenal recess (or Morison’s pouch) is the posteromedial extension of the subphrenic space, inferior to the coronary ligament. As its name implies, it extends between the right hepatic lobe and the anterior border of the right kidney. The lesser sac has two major components [14, 25]: a small superior recess is immediately posterior to the hepatoduodenal ligament. The caudate lobe of the liver is enveloped by this peritoneal reflection (Fig. 15). The larger inferior recess occupies the space behind the stomach, anterior to the transverse mesocolon and medial to the gastrosplenic ligament. As both portions of the lesser sac are surrounded by abdominal viscera, percutaneous drainage of collections within this space is difficult. Inferiorly, the superior recess communicates with the right perihepatic space through the foramen of Winslow. Caudally, behind the duodenum and pancreatic head, there may be an extension that is responsible for peritoneal fluid collections behind the pancreatic head [7].

Esophagus

Fig. 15. The boundaries of the superior recess of the lesser sac may be recognized when fluid engulfs the caudate lobe. The lesser omentum separates this fluid from fluid in the fissure for the ligamentum venosum, which is in continuity with the left posterior perihepatic space (gastrohepatic recess). IVC inferior vena cava, Ao aorta (Reprinted with permission from [6])

Fig. 16. Cystic mesothelioma. Cystic and solid lesions are seen filling the left anterior and posterior subphrenic spaces

References 1. Levy AD, Shaw JC, Sobin LH (2009) Secondary tumors and tumorlike lesions of the peritoneal cavity: imaging features with pathologic correlation. RadioGraphics 29:347-373 2. Meyers MA, Oliphant M, Berne AS, Feldberg MAM (1987) The peritoneal ligaments and mesenteries: pathways of intraabdominal spread of disease. Radiology 163:593-604 3. Balfe DM, Mauro MA, Koehler RE et al (1984) Gastrohepatic ligament: normal and pathologic CT anatomy. Radiology 150:485-490 4. Dodds WJ, Foley WD, Lawson TL et al (1985) Anatomy and imaging of the lesser peritoneal sac. Am J Roentgenol 144:567-575 5. Low RN (2007) MR imaging of the peritoneal spread of malignancy. Abdom Imaging 32:267-283 6. Meyers MA (1994) Dynamic radiology of the abdomen: normal and pathologic anatomy. 4th edn. Springer-Verlag, New York 7. Sivit CJ (1996) CT of mesentery-omentum peritoneum. Radiol Clin North Am 34:863-884 8. Auh YH, Rosen A, Rubenstein WA et al (1984) CT of the papillary process of the caudate lobe of the liver. AJR Am J Roentgenol 142:535-538

Spread of Metastatic Disease in the Abdomen and Pelvis

9. Donoso L, Martinez-Noguera A, Zidan A, Lora F (1989) Papillary process of the caudate lobe of the liver: sonographic appearance. Radiology 173:631-633 10. Weinstein JB, Heiken JP, Lee JKT et al (1986) High resolution CT of the porta hepatis and hepatoduodenal ligament. RadioGraphics 6:55-74 11. Zirinsky K, Auh YH, Rubenstein WA et al (1985) The portacaval space: CT with MR correlation. Radiology 156:453-460 12. Ito K, Choji T, Fujita T et al (1993) Imaging of the portacaval space. AJR Am J Roentgenol 161:329-334 13. Vincent LM, Mauro MA, Mittelstaedt CA (1984) The lesser sac and gastrohepatic recess: sonographic appearance and differentiation of fluid collections. Radiology 150:515-519 14. Dodds WJ, Foley WD, Lawson TL et al (1985) Anatomy and imaging of the lesser peritoneal sac. AJR Am J Roentgenol 144:567-575 15. Langman J (1971) Medical Embryology. Part two: Special embryology. Chap. 13 Digestive system. WB Saunders, Philadelphia, PA, p 29 16. Cooper C, Jeffrey RB, Silverman PM et al (1986) Computed tomography of omental pathology. J Comput Assist Tomogr 10:62-66 17. Rubesin SE, Levine MS, Glick SN (1986) Gastric involvement by omental cakes: radiographic findings. Gastrointest Radiol 11:223-228

153

18. DeMeo JH, Fulcher AS, Austin RF (1995) Anatomic CT demonstration of the peritoneal spaces, ligaments, and mesenteries: normal and pathologic processes. RadioGraphics 15:755-770 19. Pai RK, Longacre TA (2007) Pseudomyxoma peritonei syndrome: classification of appendiceal mucinous tumours. In: Ceelen WP (ed) Peritoneal carcinomatosis: a multidisciplinary approach. Springer-Verlag, New York, pp 71-107 20. Vibhakar SD, Bellon EM (1984) The bare area of the spleen: a constant CT feature of the ascitic abdomen. AJR Am J Roentgenol 141:953-955 21. Rubenstein WA, Auh YH, Zirinsky K et al (1985) Posterior peritoneal recesses: assessment using CT. Radiology 156: 461-468 22. Love L, Demos TC, Posniak H (1985) CT of retrorenal fluid collections. AJR Am J Roentgenol 145:87-91 23. Crass JR, Maile CW, Frick MP (1985) Catheter drainage of the left posterior subphrenic space: a reliable percutaneous approach. Gastrointest Radiol 10:397-398 24. Rubenstein WA, Auh TH, Whalen JP, Kazem E (1983) The perihepatic spaces: computed tomographic and ultrasound imaging. Radiology 149:231-239 25. Jeffrey RB, Federle MP, Goodman PC (1981) Computed tomography of the lesser peritoneal sac. Radiology 141: 117-122

IDKD 2010-2013

Abdominal Vascular Disease: Diagnosis and Therapy Johannes Lammer Cardiovascular and Interventional Radiology, AKH-Universitätsklinik, Wien, Austria

Acute Arterial Occlusion If acute occlusion of the abdominal aorta or branch vessels occurs, two causes have to be considered: (1) embolization and (2) dissection.

Embolization Embolization is most common in elderly patients with atrial fibrilation, a history of myocardial infarction, and aortic aneurysm. Embolization from the heart or thoracic aorta may cause acute subtotal or total occlusion of the celiac artery, superior mesenteric artery (SMA), inferior mesenteric artery (IMA), or the renal arteries. The most common causes of embolization are: • atrial fibrillation; • thoracic aortic aneurysm; • myocardial infarction with mural thrombus formation; • advanced aortic arteriosclerosis; • perforating arteriosclerotic ulcer; • hypercoagulability syndrome. The symptoms of acute ischemia are dependent on the involved vascular territory. Embolization to the liver or spleen may cause acute right or left upper abdominal pain. Acute mesenteric ischemia typically causes severe abdominal pain and bowel paralysis. Embolization to the kidney results in flank pain, hematuria, and hypertension. In any case, severe elevation of serum lactate dehydrogenase levels points to the ischemic nature of the acute pain. In case of complete ischemia without a sufficient collateral circulation, the warm ischemic time tolerated by the abdominal organs is <6 h. Therefore acute diagnosis and therapy are mandatory. The primary diagnosis is made by CT with contrast enhancement injected at a concentration of 300-400 mg iodine/mL and at an injection rate of 4 mL/s, with a total bolus volume of 80-120 mL. A bolus care technique with a delay of 20-40 s is used. The CT settings are 2-mm collimation, pitch = 2, with a reconstruction interval of 1 mm. In the arterial phase, the obstructing embolus and the ischemic territory can be visualized. In the delayed phase, residual perfusion through the collateral arteries may be demonstrated.

Interventional treatment with a thrombectomy device and/or local intra-arterial fibrinolysis with recombinant tissue plasminogen activator (loading dose 10 mg, infusion dose 5 mg/h) or urokinase (loading dose 250,000 IU, infusion dose 100,000 IU/h) together with a glycoprotein IIb/IIIa antagonist (Aciximab: loading dose 0.25 mg/kg, infusion dose 0.125 mg/kg/h) is one option. The other option is surgery, which may be faster, enables inspection, and, if necessary, resection of the ischemic organs.

Dissection Patients with chest trauma, chronic severe hypertension, or a connective tissue disease such as Marfan syndrome or Ehlers-Danlos syndrome are vulnerable to aortic dissection. Acute type A and B aortic dissection may cause dynamic compression of the true lumen of the aorta by the pressurized false lumen. This can result in acute ischemia of the liver and spleen, the bowel, and one or both kidneys. The dissection plane may also run into one of the arteries perfusing the organ, thereby causing obstruction of the true lumen. There are several options in the interventional treatment of dissection. First, in case of dynamic compression of the true aortic lumen, occlusion of the proximal entry into the false lumen with an aortic stent graft will decompress the false lumen and result in re-opening of both the true lumen of the aorta and the side branches (Fig. 1). Second, in case of static compression due to a sidebranch dissection, stent placement in the true lumen of the organ artery will reconstitute organ perfusion. Third, in case of organ perfusion through the false lumen, balloon fenestration of the intimal flap will re-establish flow into the malperfused territory.

Chronic Arterial Occlusive Disease In young patients, potential causes of chronic arterial occlusion are fibromuscular disease, Takayasu arteritis, and Recklinghausen neurofibromatosis. In elderly patients, the primary cause of chronic arterial occlusive disease is arteriosclerosis.

155

Abdominal Vascular Disease: Diagnosis and Therapy

a

d

b

c

e

f

Fig. 1 a-f. Complicated type B aortic dissection. CT demonstrating (a) primary entry tear of type B aortic dissection and (b) compression of true lumen of the aorta. c Aortic angiogram showing true and false lumen of type B dissection. d Aortography showing compression of the true lumen, with malperfusion of the superior mesenteric and renal arteries (“floating visceral sign”). Aortography (e) after stentgraft implantation and closure of primary entry tear, and (f) following stent graft implantation, which demonstrates spontaneous revascularization of the visceral arteries

Mesenteric Artery Stenosis Between the three large mesenteric arteries (celiac artery, SMA, IMA) there are two main collateral pathways: (1) the pancreatico-duodenal arteries and the arc of Buehler, between the celiac artery and SMA; (2) the arc of Riolan and the marginal artery of Drummond, between the SMA and IMA. Therefore, an obstruction of at least two mesenteric arteries is necessary to cause ischemic symptoms. The typical clinical symptom is abdominal angina, with abdominal pain in 94% of patients,

post-prandial cramps in 86%, weight loss in 74%, abdominal bruit in 70%, and diarrhea. The primary diagnosis is made by computed tomography angiography (CTA), magnetic resonance angiography (MRA), or by intra-arterial catheter angiography of the abdominal aorta in a lateral projection. Interventional treatment consists of percutaneous transluminal angioplasty (PTA) with or without secondary stent placement in at least one of the obstructed arteries. The causes and symptoms of chronic arterial occlusion depend on the obstructed artery and its location. In

156

celiac trunk stenosis, chronic obstruction may remain asymptomatic because of the collateral pathways through the gastroduodenal and pancreatic arteries from the SMA. The causes are an arteriosclerotic plaque, compression by the arcuate ligament, or carcinoma of the pancreas. Superior mesenteric artery stenosis will result in post-prandial abdominal pain (abdominal angina) only if two or all three gastrointestinal arteries are obstructed. The causes for SMA obstruction are arteriosclerosis, fibromuscular disease, Takayasu arteritis, pancreatic carcinoma, or chronic pancreatitis. Inferior mesenteric artery stenosis, with obstruction of the IMA, is most commonly observed in patients with advanced atheromatosis or a partially thrombosed abdominal aortic aneurysm. Due to the collateral circulation through the arc of Riolan and the marginal artery, IMA obstruction normally remains asymptomatic.

Renal Artery Stenosis Hypertension and/or renal insufficiency are the most frequent consequences of renal artery stenosis. Acute onset of the clinical symptoms and repeated flash pulmonary edema are suggestive. The most common etiology in patients over age 50 years is arteriosclerosis (65-75% of all patients), with males more often affected than females. In the majority of cases, the proximal 2 cm of the renal artery are involved, accompanied by atherosclerotic changes in the aorta. In 30% of patients, the stenosis is bilateral. In patients under the age of 50, renal artery stenosis is most often due to fibromuscular disease (20-30% of all patients). In this case, females are five times more likely than males ratio to be affected. Most commonly, the middle to distal renal artery, including its branches, is obstructed, with bilateral involvement in 50-70% of patients. Imaging shows the typical “string of pearls” appearance of the stenosis and, possibly, aneurysms and dissections. However, importantly, there is no aortic disease. Other causes of renal artery stenosis include Takayasu arteritis, mid-aortic syndrome, Recklinghausen neurofibromatosis, and as a consequence of radiation therapy. However, an algorithm for the diagnosis of a renal artery stenosis has yet to be established. Color duplex ultrasound is a non-invasive but complex examination that requires operator experience. The diagnostic criteria for renal artery stenosis are: increased peak systolic velocity >250 cm/s, a renal-to-aortic ratio of peak systolic velocity >3.5, intrastenotic turbulence, and a flattened pulse wave in the periphery (pulsus tardus). The sensitivity of color duplex sonography for detection of a renal artery stenosis >70% is 72-92%. Color duplex ultrasound with an angiotensin-converting enzyme (ACE) inhibitor provides a positive predictive value of 67-95% for cure or improvement after revascularization. A nuclear scan, specifically, renal scintigraphy with technetium-99m mercaptoacetyltriglycine (MAG3) or Tc-99m diethylenetriaminepentaacetic acid (DTPA) in

Johannes Lammer

patients previously administered an ACE inhibitor (captopril 25 mg) shows a delayed wash-out of the tracer within the post-stenotic kidney. In bilateral disease and in chronic ischemic nephropathy, however, lateralization of the tracer is less evident. In a selected population at clinically high risk for renal artery stenosis, the sensitivity for detection of a unilateral stenosis >70% is 51-96% (mean 82%). Its positive predictive value for a renal artery stenosis with improvement of hypertension after revascularization is 51-100% (mean 85%). However, scintigraphy is much less sensitive in unselected patients and in those with bilateral disease, impaired renal function, urinary obstruction, and chronic ACE inhibitor intake. Newer tests are gadolinium-enhanced MRA and spiral CTA. For state-of-the-art MRA, high-field-strength systems with high performance gradients are necessary to obtain breath-hold 3D T1-weighted spoiled gradient-echo imaging with short TR and TE. Intravenous administration of gadolinium contrast material (0.1 mmol/kg; flow rate 2 mL/s), a central k-space readout, and background subtraction are additional techniques to improve signalto-noise ratio and spatial resolution. The sensitivity of MRA to detect a renal artery stenosis >50% is over 95% (Fig. 2). The main limitations of renal MRA are its lack of accuracy in the evaluation of small accessory renal arteries and branch vessels, artifacts due to the presence of stents, and a tendency of the techniques to overestimate moderate stenoses. In a double-blind randomized study, contrast-enhanced MRA with the blood-pool contrast agent gadofosveset was not superior to gadobenate dimeglumine. The sensitivity of CTA to detect stenosis of the renal artery and its accessory arteries is >95%. For highquality opacification of the renal arteries and to avoid renal vein overlap, correct bolus planning is mandatory: density measurement during bolus rise, flow 4 mL/s, total volume 80-120 mL (multidetector scanners need less contrast). A short breath-hold acquisition, 1- to 2-mm collimation, pitch =1.5-6 (depending on single or multidetector technology), and an overlap of reconstruction of 0.5-0.75 are important parameters to obtain good spatial resolution of the study. Curved planar reconstruction (most useful for stents), volume rendering, and maximum intensity projection (MIP) are used for 3D imaging (Fig. 2). Nonetheless, intra-arterial catheter arteriography together with pressure gradient measurement is still the gold standard for the evaluation of a renal artery stenosis. The revascularization technique of choice is renal PTA, without or with stent placement (Fig. 2). Aorto-renal bypass surgery is indicated only if PTA fails. In a recently published meta-analysis, renal arterial stent placement proved to be technically superior and clinically comparable to renal PTA alone. The technical success rate of stent vs. PTA was 98 vs. 77%, and the re-stenosis rate 17 vs. 26% (p <0.001). In hypertension, the cure rate of PTA vs. stent was 10 vs. 20%, the rate of improvement 53 vs. 49 %. In renal insufficiency, the rate of improvement was 38 vs. 30%, that of stabilization 41 vs. 38%.

157

Abdominal Vascular Disease: Diagnosis and Therapy

a

b

c

d

Fig. 2 a-d. Imaging and intervention in a patient with hypertension due to renal artery stenosis. a MRA, b CTA, and c arteriography show the stenosis of the right renal artery. d Arteriography of right renal artery after stent placement

The complication rate was 11-13% (95% CI 6-19%), the in-hospital mortality rate 1%. In a randomized study comparing stents vs. PTA in ostial stenoses, the technical success rate was 88 vs. 57%, and the 6-month primary patency rate 75 vs. 29%. Surprisingly, randomized trials comparing the effect of PTA and drug therapy on renal hypertension did not reveal a significant benefit of PTA and stenting over continuous drug therapy. However, in a Dutch study, PTA patients required only 2.1 vs. 3.2 daily drug doses (p <0.001), and 22 of 53 patients in the drug group had to be switched to the PTA group because of persistent hypertension or deterioration of renal function.

new, emerging technique that may replace open surgery in the future. Since the first clinical implant of a tube stent graft, in 1990, many different stent graft designs have been developed and tested in feasibility studies. Most recently, randomized studies (EVAR 1, Dream) compared the results of open vs. endovascular repair. In the EVAR trial, the 30-day mortality in the EVAR group was 1.7% (9/531) vs. 4.7% (24/516) in the open repair group (p = 0.009). Four years after randomization, allcause mortality was similar in the two groups (about 28%; p = 0.46), although there was a persistent reduction in aneurysm-related deaths in the EVAR group (4 vs. 7%; p = 0.04). Indications

Aneurysms Abdominal Aortic Aneurysm The incidence of abdominal aortic aneurysm in European adults 60 years and older is 2.5%. Up to 10% of patients with symptomatic peripheral arterial disease die from rupture of the aneurysm. Currently, the standard treatment is open surgery. However, endovascular implantation of stent grafts is a

The indications for endovascular treatment of abdominal aortic aneurysm are currently the same as for open surgery: (1) diameter of the aneurysm >5 cm (Fig. 3), (2) documented growth >0.5 mm/year, (3) symptomatic aneurysm (i.e., embolization, pain, ureteral compression), and (4) rupture. The specific clinical indications for the endovascular approach are typically: patients >75 years old, ASA class 3 and 4, “hostile abdomen”, and inflammatory aneurysm or horse-shoe kidney.

158

a

Johannes Lammer

b

c

Fig. 3 a-c. Patient with abdominal aortic aneurysm and renal artery stenosis. a CTA with MIP reconstruction; aortography (b) before and (c) after stent graft placement

The anatomic indications for stent graft treatment are: • infrarenal neck >15 mm in length; • infrarenal neck without thrombus or severe calcification; • angulation of the infrarenal neck <65°; • patent celiac trunk and SMA; • stent graft diameter 10% more than neck diameter; • iliac artery angulation <90°; • iliac artery without thrombus or severe calcification; • overlap of >15 mm within the iliac arteries. Endovascular implantation of stent grafts can be performed under general, epidural, or local anesthesia. The use of epidural anesthesia is a major advantage in elderly and high-risk patients. Stent Graft Designs Stent grafts have a self-expandable stent structure covered by an ultrathin polyester or ePTFE fabric. Currently, only bifurcated stentgrafts are used for the treatment of abdominal aortic aneurysm. Imaging before Stent Graft Implantation Contrast-enhanced spiral CT with multiplanar reconstruction (MPR) or MIP reconstruction is the most important examination before stent graft implantation (Fig. 3). The diameter of the landing zones (infrarenal neck, iliac arteries), the maximum diameter of the aneurysm, the extent of the thrombus, and calcifications are well depicted on CT. Complications The most frequent complication is incomplete exclusion of the aneurysm, with remaining pressurization of the

aneurysm sack through an endoleak. White and May proposed a classification of primary (<30 days) and secondary (>30 days) endoleaks. Type 1 endoleaks are characterized by direct perfusion through the proximal (infrarenal) or distal (iliac) anastomosis. In type 2, there is retrograde perfusion through branch vessels (lumbar arteries, IMA, accessory renal artery). Type 3 consists of mid-graft leak due to disintegration of the stent graft (disconnection of the second iliac limb, fabric erosion). In type 4, there is fabric porosity, while type 5 is characterized by endotension.

Visceral Artery Aneurysm Aneurysms of the celiac trunk, splenic artery, hepatic artery, gastroduodenal artery, and SMA are caused by arteriosclerosis, arteritis, periarterial inflammation (such as pancreatitis), trauma, and soft-tissue diseases (such as Marfan and Ehlers-Danlos syndromes). An aneurysm >2.5 cm in diameter should be considered for treatment to prevent rupture. Meticulous imaging, including selective catheter angiography and 3D imaging with CTA or MRA, is necessary before surgery or endovascular treatment. The endovascular options are embolization and exclusion with a stent graft.

Renal Artery Aneurysm The causes are arteriosclerosis, systemic vasculitis (such as polyarteritis nodosa or lupus erythematosus), fibromuscular disease, soft-tissue disorders, and trauma. Arteriosclerotic and large aneurysms are usually calcified. The risk of rupture and chronic embolization are indications for treatment. Bypass surgery, coil embolization, and stent graft implantation are the therapeutic options.

159

Abdominal Vascular Disease: Diagnosis and Therapy

Therapeutic Embolization of Gastrointestinal Bleeding Bleedings in the upper gastrointestinal (GI) tract are usually treated by endoscopic coagulation therapy. However, if endoscopic therapy fails or the bleeding source is inaccessible to endoscopy, angiographic embolization is indicated. Patients after gastric resection involving bilioenteric anastomoses and patients bleeding from hepatobiliary and pancreatic pathologies (pseudoaneurysms) should be treated primarily by angiographic techniques. Due to the extensive collateral circulation in the upper abdomen, a detailed angiographic evaluation followed by embolization of all feeding arteries is required. The most common causes of upper GI bleeding are: • gastroduodenal ulcer; • proximal jejunal tumors (endocrine carcinomas, angiofibroma, melanoma metastasis, arteriovenous malformation etc.); • pseudoaneurysm of the gastroduodenal, pancreaticoduodenal, intrapancreatic, and splenic arteries after pancreatitis or trauma, bleeding through the Wirsung pancreatic duct; • pseudoaneurysm of the hepatic artery, bleeding through the common bile duct (triad of hemobilia: abdominal colic followed by jaundice and hematemesis or melena). In lower GI bleeding, diagnosis and treatment are preferentially done by colonoscopy. However, in selected cases embolization is required. For the primary diagnosis, a contrast-enhanced CT of the abdomen (400 mg iodine/mL, flow 4 mL, volume 100 mL) in the arterial and delayed phases usually demonstrates the area of bleeding nicely. This helps to find the bleeding source

a

b

during catheter angiography. Superselective catheterization and bowel paralysis with Buscopan (Boehringer Ingelheim, Germany) are mandatory for a bleeding angiogram. The causes of lower GI bleeding are: • small bowel tumours (endocrine carcinomas, angiofibroma, melanoma metastasis, arteriovenous malformation, etc.); • Meckels diverticulum; • large bowel diverticulum bleeding (most common cause in elderly patients); • hemangiomatosis and arteriovenous malformation of the colon; • colon cancer (rectal bleeding due to hemorrhoidal disease has to be ruled out primarily). GI bleedings are preferentially embolized by coils.

Treatment for Bleeding Complications after Surgery, Trauma, and Post-partum Bleeding in the abdomen may occur from iatrogenic causes, particularly in the kidneys and the liver after percutaneous interventions. It may also be due to trauma or tumor. Frequent and typical locations and causes are renal arteriovenous fistulas due to nephrostomy (Fig. 4) or biopsy, laceration of the hepatic arteries by percutaneous manipulations, psoas and pelvic bleeding due to traumatic arterial injury, and uterine artery bleeding post-partum. Temporary occlusion of the uterine artery can be a valid alternative to emergency hysterectomy in patients with intractable bleeding resulting from an atonic uterus. In other anatomical locations, the type, source, and location of the bleeding determine the method used to safely interrupt the extravasations.

c

Fig. 4 a-c. Hematuria and shock due to renal bleeding after nephrostomy. a CT showing large perirenal hematoma and active bleeding. b Selective angiography demonstrates the bleeding site. c Control angiography after selective coil embolization

160

Johannes Lammer

Therapeutic Embolization of Tumors Tumor embolization techniques range from palliative embolization in bleeding genitoureteral tumors, to chemoembolization techniques in hypervascularized hepatic neoplasms, in particular hepatocellular carcinoma (HCC), to the definitive treatment of benign lesions such as uterine fibroids. In the liver, tumor embolization is mainly used in patients with inoperable HCC but preserved liver function (Child-Pugh A and B) and in those with neuroendocrine tumor metastases. Classical treatment is chemoembolization with doxorubicin mixed with Lipiodol (Guerbet, France), sometimes in combination with a temporary blockade of the hepatic artery by Gelfoam or other embolization particles. In randomized trials, chemoembolization of unresectable HCC has shown to be superior to supportive treatment only. New techniques include doxorubicin-loaded particles as an alternative embolization agent for HCC (Fig. 5). In the Precision V trial, more than 200 patients were randomly assigned to either transarterial chemoembolization (TACE) with drug-eluting beads or to conventional chemoembolization with Lipiodol and doxorubicin. The former group had a higher rate of objective response and disease control within 6 months of follow-up. In addition, the use of drug-eluting beads resulted in a significantly

a

higher efficacy in high-risk patients (defined by ChildPugh B cirrhosis, ECOG 1 tumor symptoms, bilobar and recurrent disease), a lower liver toxicity, but also systemic doxorubicin-related toxicity. Regional chemotherapy with irinorecan drug-eluting beads in patients with colorectal liver metastases is under evaluation. Intrahepatic radiation by radioactive β-emitting particles directly injected into the hepatic arteries is another treatment modality for primary and metastatic liver tumors.

Venous Interventions TIPS Transjugular intrahepatic portocaval shunt (TIPS) to depressurize the portal venous system was introduced into clinical medicine at the end of the 1980s. Since then, a standardized technique has been developed that allows safe implantation of this artificial connection between the portal and hepatic veins. The introduction of stent grafts instead of bare stents has also led to improved patency of the shunt tract. TIPS is indicated in patients symptomatic from portal hypertension with acute or chronic bleeding of esophageal or gastric varices and in those with intractable ascites.

b

c d

Fig. 5 a-d. Hepatocellular carcinoma before and after transarterial chemoembolization (TACE) with drug-eluting beads. a Contrastenhanced CT shows HCC in right hepatic lobe. b Arteriography shows hypervascular HCC. c Control arteriography after TACE of HCC. d Contrast-enhanced CT shows complete necrosis of the HCC

Abdominal Vascular Disease: Diagnosis and Therapy

Endoscopic techniques to treat varices are competitive in bleeders whereas in some patients with ascites there are few alternatives. Randomized trials, meta-analyses, and Cochrane data review analyses have shown that TIPS is superior to endoscopic therapy in the prevention of rebleeding and is superior to paracenteses to remove ascites. In patients with acute or subacute Budd-Chiari syndrome, TIPS can be a life-saving procedure and help to overcome the acute phase, but the approach is burdened by a relatively high re-thrombosis rate. The risks after TIPS procedure are liver failure from shunted blood volume and encephalopathy.

Embolization of the Portal Veins An intervention of increasing importance is pre-operative embolization of the right or left portal vein in order to induce hypertrophy of the contralateral hepatic lobe prior to extended hemi-hepatectomy.

Suggested Reading ASTRAL Investigators, Wheatley K, Ives N, Gray R et al (2001) Revascularization versus medical therapy for renal artery stenosis. N Engl J Med 361:1953-1962 Bax L, Woittiez AJ, Kouwenberg HJ et al (2009) Stent placement in patients with atherosclerotic renal artery stenosis and impaired renal function: a randomized trial. Ann Intern Med 150:840-848 Blankensteijn JD, de Jong SE, Prinssen M et al (2005) Dutch Randomized Endovascular Aneurysm Management (DREAM) Trial Group. Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysms. N Engl J Med 352:2398-2405 Blum U, Voshage G, Lammer J et al (1997) Endoluminal stentgrafts for infrarenal abdominal aortic aneurysms. N Engl J Med 336:13-20 Covey AM, Tuorto S, Brody LA et al (2005) Safety and efficacy of preoperative portal vein embolization with polyvinyl alcohol in 58 patients with liver metastases. AJR Am J Roentgenol 185:1620-1626 D’Amico G, Luca A, Morabito A et al (2005) Uncovered transjugular intrahepatic portosystemic shunt for refractory ascites: a metaanalysis. Gastroenterology 129:1282-1293 EVAR trial participants (2005) Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1): randomised controlled trial. Lancet 365:2179-2186 Greenhalgh RM, Brown LC, Kwong GP et al, EVAR trial participants (2004) Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: randomised controlled trial. Lancet 364:843-848 Kaatee R, Beek FJ, de Lange EE et al (1997) Renal artery stenosis: detection and quantification with spiral CT angiography versus optimized digital subtraction angiography. Radiology 205:121-127

161

Kasirajan K, O’Hara PJ, Gray BH et al (2001) Chronic mesenteric ischemia: open surgery versus percutaneous angioplasty and stenting. J Vasc Surg 33:63-71 Khan S, Tudur Smith C, Williamson P, Sutton R (2005) Portosystemic shunts versus endoscopic therapy for variceal rebleeding in patients with cirrhosis. Cochrane Database Syst Rev 18:CD000553 Khuroo MS, Al-Suhabani H, Al-Sebayel M et al (2005) BuddChiari syndrome: long-term effect on outcome with transjugular intrahepatic portosystemic shunt. J Gastroenterol Hepatol 20:1494-1502 Lammer J, Malagari K, Vogl T et al, on behalf of the PRECISION V investigators (2009) Prospective randomized study of doxorubicin-eluting-bead embolization in the treatment of HCC: results of the PRECISION V study. Cardiovasc Intervent Radiol 12 [Epub ahead of print] Leertouwer TC, Gussenhoven EJ, Bosch JL et al (2000) Stent placement for renal artery stenosis: where do we stand? A metaanalysis. Radiology 216:78-85 Mann SJ, Pickering TG (1992) Detection of renovascular hypertension: state of the art 1992. Ann Intern Med 117:845-853 Perler AB, Becker GJ (1998) Vascular intervention – a clinical approach. Visceral vascular disease. Thieme, New York, Stuttgart, pp 517-637 Prinssen M, Verhoeven EL, Buth J et al (2004) Dutch Randomized Endovascular Aneurysm Management (DREAM) Trial Group. A randomized trial comparing conventional and endovascular repair of abdominal aortic aneurysms. N Engl J Med 351:1607-1618 Rose SC, Quigley TM, Raker EJ (1995) Revascularization for chronic mesenteric ischemia: comparison of operative arterial bypass grafting and percutaneous transluminal angioplasty. J Vasc Interv Radiol 6:339-349 Saab S, Nieto JM, Lewis SK, Runyon BA (2006) TIPS versus paracentesis for cirrhotic patients with refractory ascites. Cochrane Database Syst Rev 18:CD004889 Salerno F, Cammà C, Enea M et al (2007) TIPS for refractory ascites: a meta-analysis of individual patient data. Gastroenterology 133:825-834 Soulez G, Oliva VL, Turpin S et al (2000) Imaging of renovascular hypertension: respective values of renal scintigraphy, renal Doppler US, and MR angiography. Radiographics 20:13551368 van de Ven PJ, Kaatee R, Beutler JJ et al (1999) Arterial stenting and balloon angioplasty in ostial arteriosclerotic renovascular disease: a randomized trial. Lancet 353:282-286 Van Jaarsveld BC, Krijen P, Pieterman H et al (2000) The effect of balloon angioplasty on hypertension in atherosclerotic renalartery stenosis. Dutch Renal Artery Stenosis Intervention Cooperative Study Group. N Engl J Med 342:1007-1014 Webster J, Marshall F, Abdalla M et al (1998) Randomized comparison of percutaneous angioplasty vs continued medical therapy for hypertensive patients with atheromatous renal artery stenosis. Scottish and Newcastle Renal Artery Stenosis Collaborative Group. J Hum Hypertens 12:329-335 Williams DM, Lee DY (1997) Dissected aorta, parts I-III. Radiology 203:23-44 Zheng M, Chen Y, Bai J et al (2008) TIPS vs. endoscopic therapy in the secondary prophylaxis of variceal rebleeding in cirrhotic patients: meta-analysis update. J Clin Gastroenterol 42:507-516

IDKD 2010-2013

Non-vascular Abdominal Disease: Diagnosis and Therapy Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi Division of Diagnostic and Interventional Radiology, University of Pisa, Pisa, Italy

Introduction Recent technological advances have given rise to a wide range of diagnostic and interventional imaging-guided procedures for an increasing number of abdominal parenchymal diseases, such as those involving the kidneys, adrenals, or pancreas. However, the liver, and the cirrhotic liver in particular, still represents the main field of application of abdominal imaging modalities. Imaging plays a key role in the diagnostic work-up of the nodular lesion, especially in its initial detection and characterization. In addition, the results of imaging assessment guide the therapeutic options (surgical, interventional, or palliative) and provide the standard of reference in the evaluation of tumor response.

Hepatocellular Carcinoma: Pathology The diagnosis of hepatocellular carcinoma (HCC) is based on imaging examinations in combination with clinical and laboratory findings. However, imaging of the cirrhotic liver remains challenging since regenerative nodules and pre-neoplastic hepatocellular lesions, such as dysplastic nodules, can mimic small HCCs. New imaging technologies have improved investigations into the multistep process that takes place during carcinogenesis, from regeneration towards dysplasia and full malignancy. The most important pathological alteration in the development of HCC is the derangement in the liver’s vascular supply due to progressive capillarization of the sinusoids, associated with the formation of an increasing number of unpaired arterioles. In addition, there are progressive, intracellular histological changes, including loss of the biliary polarization of hepatocytes, a derangement of their microscopic secretory structure, and the progressive depletion of Kupffer cells within nodular lesions [1]. Due to the wide spectrum of morphological changes seen in cirrhotic parenchyma, all cross-sectional imaging modalities should be applied in the detection and characterization of nodular lesions. Ultrasound (US) is the modality most commonly applied in the surveillance of cirrhotic patients and is thus considered as the first-line approach in their diagnostic workup. The high spatial

resolution of US enables the demonstration even of very small lesions (<1 cm in maximum diameter). In nodules >1 cm in maximum diameter, the vascular supply should be assessed as well. Moreover, the introduction of microbubble contrast agents and the development of contrast-specific scanning techniques have expanded the capabilities of US in examinations of the cirrhotic liver. The advent of second-generation contrast agents and lowmechanical-index real-time scanning techniques has been fundamental in improving the ease and reproducibility of US examinations [2]. Consequently, contrast-enhanced US has been introduced into the diagnostic flow chart as one of the imaging modalities able to demonstrate the vascular pattern of nodules >1 cm in diameter [3]. Dynamic multidetector computed tomography (MDCT) and magnetic resonance imaging (MRI) are further imaging modalities used to detect and characterize nodular vascular changes. They provide a detailed view of the hepatic parenchyma and allow a confident diagnosis of neoplasm. Importantly, these imaging evaluations are a fundamental prerequisite for guiding the therapeutic approach. The added value of MRI over other crosssectional imaging modalities is its ability to assess the components of a hepatic lesion in baseline image acquisitions. The use of tissue-specific MRI contrast media (hepatobiliary and reticuloendothelial agents) can reveal metabolic disorders that occur as a result of cellular dedifferentiation. Current cross-sectional techniques demonstrate the pathological changes associated with carcinogenesis, such that needle biopsy is no longer considered as the standard reference procedure for the definitive diagnosis of HCC [3].

Pre-neoplastic Lesions: Regeneration Regeneration is the result of the architectural rearrangement of the hepatic parenchyma in response to chronic injury. Generally, regenerative nodules are small (around 1 cm) but in some cases may have a maximum diameter v3 cm. At US examination, a typical regenerative nodule may appear as a hypoechoic or hyperechoic small nodule that enhances relative to background because of its thin hyperechoic rim, which represents perinodular fibrous

163

Non-vascular Abdominal Disease: Diagnosis and Therapy

Arterial

a

Late

b

Fig. 1 a-c. a Baseline US examination shows the presence of a large (up to 3 cm) hyperechoic nodule of the VI segment, partially exophytic, in a background of cirrhosis (arrow). b At dynamic MRI study, the nodule does not show signs of hypervascularization (appearing as hypointense on arterial phase); instead, the portal supply is still well evident, while the nodule appears slightly isointense during the late phase. c On this baseline T2-weighted image, the nodule appears as hypointense, due to its intranodular iron content, which increases signal intensity compared to the surrounding parenchyma (arrow)

tissue and which may mimic a pseudocapsule. Generally, no further examination is required. In cases of larger nodules, dynamic evaluation is required to exclude neoplastic dedifferentiation (Fig. 1 a). At dynamic contrast studies (contrast-enahnced US, MDCT, or MRI), there is no evidence of a vascular supply different than that of the surrounding parenchyma, due to the conspicuous and predominant feeding of these nodules by portal vessels (Fig. 1 b) [4]. MRI can depict some peculiarities that may differentiate regenerative nodules from other focal lesions. On baseline examination, regenerative nodules are isointense on T1- and T2weighted images due to the presence of normal liver cells. The frequent intranodular content of iron may decrease the relaxation time on T2, thus reducing the signal intensity of the nodules on these sequences (Fig. 1 c). Nodular signal intensity after the administration of tissue-specific contrast agent (hepatobiliary or reticuloendothelial) is not modified due to the preserved metabolic activity of hepatocytes and Kupffer cells.

Pre-neoplastic Lesions: Dysplasia Dysplastic nodules are classified as low-grade (LGDN) and high grade (HGDN), depending on the degree of cellular atypia and on changes in architectural structure.

c

HGDN are considered to be pre-malignant lesions, and the subsequent development of HCC from a HGDN within a period of years or even a few months has been documented [5]. Plain US and CT examinations are generally useless in the characterization of dysplastic nodules, due to the lack of specific imaging finding (Fig. 2 a). At dynamic study, the post-contrast enhancement patterns of HGDN may be highly variable. In fact, even if the main blood supply is still provided by branches of the portal vein, the presence of a vascular supply on arterial phase may be seen due to the development of sporadic unpaired arteries. However, this finding is not associated with wash-out, which instead represents the diagnostic finding for HCC (Fig. 2 b). Thus, in addition to displaying the degree of vascular supply, MRI demonstrates parenchymal alterations, thereby playing an important role in the differential diagnosis between pre-neoplastic and neoplastic lesions. In fact, at baseline MRI examination, dysplastic nodules typically show a characteristic hyperintensity on T1-weighted sequences due to the intranodular presence of glycogen or lipids, while on T2weighted sequences they may be slightly hyperintense, isointense, or even hypointense. In the hepatobiliary phase, after the administration of hepatospecific contrast agents, DNs are generally isointense or hyperintense compared to the surrounding liver

164

Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

Arterial

Late

a

b

c

Fig. 2 a-c. a At baseline US, it is possible to appreciate the presence of a suspected isoechoic nodule surrounded by a thin hypoechoic rim. b The US post-contrast examination does not show any arterial supply within the nodule, which remains hypoechoic in all post-contrast acquisitions. c On hepatobiliary phase at MRI examination, the nodule is isointense with the surrounding parenchyma, due to the preserved biliary function

parenchyma, reflecting the maintenance of biliary function but also cholestasis, which occurs in so-called green nodules (Fig. 2 c). Sometimes, the progression from dysplasia to HCC is detected in a very early phase, when it is possible to identify a focus of HCC within a pre-malignant lesion, known as “a nodule within a nodule”. On T2-weighted images, the typical appearance is a focus of high signal intensity located within a low-signal-intensity nodule and also showing the post-contrast signal behavior of HCC.

Hepatocellular Carcinoma: Imaging Findings As noted above, one of the key pathological features diagnostic of HCC is the vascular supply of the nodule. The progression from regeneration to overt HCC is characterized by neoangiogenesis, that is, the concomitant development of feeding arteries and efficient arteriovenous shunts. This pathological blood supply is well demonstrated on contrast-enhanced dynamic studies by the typical findings of wash-in during the arterial phase and subsequent wash-out (Fig. 3 a). However, a typical vascular behavior may not be present at dynamic imaging. In these cases, MRI both at baseline and following the administration of hepatospecific contrast media may lead to a definitive diagnosis of HCC. In addition, early or moderately differentiated HCC may show peculiar signal intensity on T1- and T2weighted baseline acquisitions (Fig. 3 b). On baseline T1-weighted images, HCC usually appears as a hypointense nodule because of its increased cellularity, and thus its higher amount of intracellular water; however, small, well-differentiated HCCs may show different

signal intensities, appearing as either hyperintense or isointense. Hyperintensity, as in the case of HGDN, may be related to the intracellular presence of glycogen and fat, which accumulate because of the loss of normal cellular metabolic activity. On baseline T2-weighted images, HCC usually shows mild signal hyperintensity, while small and well-differentiated tumors may be isointense to the surrounding parenchyma. In a comparison of histological data obtained from explanted cirrhotic livers with the MRI signal intensity of corresponding lesions, a relationship was found between lesion malignancy and nodular intensity on T2-weighted images [6]. It also has been shown that nodular signal intensity on T2-weighted images is significantly associated with the intranodular blood supply; in fact, signal intensity increases as the intranodular portal venous blood supply decreases [7]. Although their application has not yet been introduced into diagnostic guidelines, the use of tissue-specific contrast medium may give additional information, such as the atypical baseline or vascular pattern at dynamic study, as in the case of borderline lesions (dysplastic nodules) or well-differentiated HCCs. Regarding hepatobiliary contrast agents, the lack of contrast uptake is strongly related to overt HCC, due to the loss of normal metabolic function whereas the uptake is preserved in early HCCs, resembling that of HGDNs [8]. In daily practice, the advantages of the most recent generation of MRI contrast media can be exploited, as they illustrate the nodule’s characteristic vascular and hepatospecific phases, i.e., neoangiogenesis and lack of hepatobiliary function, respectively, which allow a highly confident diagnosis of HCC (Fig. 3 c). Specifically, reticuloendothelial system (RES) agents allow the carcinogenetic

165

Non-vascular Abdominal Disease: Diagnosis and Therapy

Arterial

Fig. 3 a-c. a MR dynamic examination reveals a nodule of the VIII segment that is hypervascular on arterial phase and shows clear cut washout in the late phase. b At baseline MRI examination, the nodules has a typical signal intensity, i.e., hypointense on T1weighted sequences and hyperintense on T2weighted sequences. c On hepatobiliary phase, the nodule is hypointense to the surrounding parenchyma, due to the complete loss of biliary function

Late

a

b

T1 w.i

pathway to be followed; for example, the progressive increase in sinusoid capillarization provides a hostile environment for reticuloendothelial cells, such that their progressive loss explains the very high signal intensity of HCC [9]. Nonetheless, due to the less consistent vascular phases, the application of RES agents is strongly limited and has largely been discontinued.

Therapy and Follow-up Nowadays, the therapeutic approach to HCC is based on surgical (transplantation and resection) and non-surgical approaches, mini-invasive modalities (percutaneous and intra-arterial therapies), and palliative approaches. The decision is based upon clinical and functional data as well as on the imaging findings, i.e., number of lesions and their size, location, degree of vascularization, and relationships with vascular and biliary structures. For example, a candidate for liver transplantation should fulfill imaging criteria, which include the presence of a single lesion measuring <5 cm or of up to three lesions, each with a greatest dimension f3 cm. Moreover, postprocessing of native images is fundamental in candidates for surgical or intra-arterial therapies, in order to provide an overall representation of the vascular anatomy or of the feeding vessels of the lesions. Imaging modalities are also very important as guidance for percutaneous ablation. In these cases, US represents the most appropriate technique, as its real-time

T2 w.i

c

visualization of the target allows continuous assessment of the ongoing changes that occur during the ablation procedure. In addition, the inclusion of periprocedural contrast enhancement provides immediate evaluation of the presence of residual viable tumor, which may be retreated during the same ablation session. Contrast-enhanced CT and MRI play a major role in follow-up, permitting assessment of the tumor’s response in terms of necrosis or relapse as well as the detection of new lesions in the surrounding parenchyma [3]. MRI, when performed with hepatospecific contrast agents, can provide additional information about postablation tissue components. This may be useful in questionable cases, in which periablation hyperemia (especially after radiofrequency ablation) or arteriovenous shunts/thrombosis (especially after ethanol ablation) must be differentiated from tumoral persistence or recurrence [10].

Conclusions In conclusion, HCC in a cirrhotic liver represents one of the most important fields of application of imaging modalities, based on their key role in the detection and characterization of nodular lesions. Moreover, any therapeutic decision is strongly related to the imaging results, as they also guide the use of mini-invasive procedures and allow evaluation of therapeutic success.

166

References 1. International Consensus Group for Hepatocellular Neoplasia. The International Consensus Group for Hepatocellular Neoplasia (2009) Pathologic diagnosis of early hepatocellular carcinoma: a report of the international consensus group for hepatocellular neoplasia. Hepatology 49:658-664 2. Cosgrove D (2006) Ultrasound contrast agents. An overview. Eur J Radiol 60:324-330 3. Bruix J, Sherman M (2005) Practice Guidelines Committee, American Association for the Study of Liver Diseases. Management of hepatocellular carcinoma. Hepatology 42:1208-1236 4. Roncalli M, Roz E, Coggi G et al (1999) The vascular profile of regenerative and dysplastic nodules of the cirrhotic liver: implications for diagnosis and classification. Hepatology 30:1174-1178 5. Theise, Park YN, Kojiro M (2002) Dysplastic nodules and hepatocarcinogenesis. Clin Liver Dis 6:497-512

Carlo Bartolozzi, Valentina Battaglia, Elena Bozzi

6. Bartolozzi C, Cioni D, Donati F et al (2001) Focal liver lesions: MR imaging-pathologic correlation. Eur Radiol 11: 1374-1388 7. Shinmura R, Matsui O, Kobayashi S et al (2006) Cirrhotic nodules: association between MR imaging signal intensity and intranodular blood supply. Radiology 237:512-519 8. Bartolozzi C, Crocetti L, Lencioni R et al (2007) Biliary and reticuloendhotelial impairment in hepatocarcinogenesis: the diagnostic role of tissue specific MR contrast media. Eur Radiol 17:2519-2530 9. Chen RC, Lii JM, Chou CT et al (2008) T2-weighted and T1weighted dynamic superparamagnetic iron oxide (ferucarbotran) enhanced MRI of hepatocellular carcinoma and hyperplastic nodules. J Formos Med Assoc 107:798-805 10. Cioni D, Lencioni R, Bartolozzi C. (2001) Percutaneous ablation of liver malignancies: imaging evaluation of treatment response. Eur J Ultrasound 13:73-93

IDKD 2010-2013

An Approach to Imaging the Acute Abdomen in the Pediatric Population Alan Daneman1, Simon G. Robben2 1 Department 2 Department

of Radiology, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada of Radiology, Maastricht University Medical Centre, Maastricht, The Netherlands

Introduction The acute abdomen is a common and often challenging emergency in the pediatric population. This chapter provides an approach to the imaging evaluation of children, highlighting briefly the more common causes of abdominal pain that may require surgery.

Causes of Acute Abdomen There is a wide spectrum of pathologies that may give rise to acute abdominal pain. These include congenital and acquired lesions that may present in the immediate neonatal period or in older infants and children. In neonates, the acute abdomen is not an uncommon event in infants in neonatal intensive care units and may be the result of several causes, with congenital bowel obstruction, complications of mid-gut malrotation, and necrotizing enterocolitis being the three most significant conditions. In older infants and children, the common causes of acute abdominal pain that may require surgery include acute appendicitis, complications of malrotation, intussusception, and Meckel diverticulum. Inflammatory bowel diseases may also present with acute abdominal symptomatology. Although the common, above-mentioned causes of acute abdominal pain are due to lesions involving the gastrointestinal tract, the acute abdomen may also be due to abnormalities of other viscera, e.g., gynecological abnormalities such as ovarian torsion, omental lesions such as omental infarcts, obstructions of the biliary and urinary tracts due to stones and inflammatory processes such as pancreatitis or hepatitis. It has to be emphasized that abdominal pain may also be due to referred pain from extra-abdominal pathologies, such as pneumonia or pleural effusion (Fig. 1). Medical diseases, e.g., sickle cell disease, HenochSchonlein purpura, and hemolytic uremic syndrome, may also be the cause of significant acute abdominal pain. Imaging of the child with acute abdominal pain should include a search for evidence of these other pathologies. Abdominal trauma may also be the cause of an acute abdomen. Occasionally, mild abdominal trauma may

Fig. 1. A boy with right lower quadrant pain and fever. The abdominal radiograph shows a triangular consolidation (arrows) in a basal segment of the right lower lobe: basal pneumonia. Dilated air-filled ascending colon was considered secondary to referred pain. Uneventful recovery after antibiotic therapy

cause abdominal pain out of proportion to the degree of trauma. Imaging in these children may reveal an underlying abnormality such as a neoplasm or an anomaly such as uretero-pelvic obstruction. Furthermore, one should consider non-accidental injury or abuse when certain traumatic lesions are found, especially when they appear in combination with hematomas of the left lobe of the liver and duodenum and pancreatitis.

Modalities Ultrasonography (US) has come to play an ever-increasing role in the management of children with acute abdominal pain and has replaced the plain abdominal radiograph as the modality of initial choice in many clinical situations.

168

Alan Daneman, Simon G. Robben

The major advantages of US are that it does not use ionizing radiation, it is relatively inexpensive, and the abdominal viscera, including the bowel, are well delineated in children. Therefore, many pathological entities can be easily confirmed or excluded. Plain abdominal radiograph (AXR) remains a standard method for evaluation of the acute abdomen in some clinical situations. It is essential when peritonitis is present and perforation is suspected. All neonates with an acute abdomen are evaluated with AXR. This modality is essential for the detection of conditions such as necrotizing enterocolitis and congenital bowel obstruction. In the former, AXR may be diagnostic; in the latter, the findings guide the choice of subsequent contrast examinations of the gastrointestinal tract. Views with a horizontal beam are essential to exclude the presence of free air due to bowel perforation and can be performed with the neonate in the dorsal or lateral decubitus position. In older children, the diagnosis of intestinal obstruction can often be made based on the supine film alone. A search for airfluid levels on the upright view does not always add extra information and a search for free air in the abdomen is often more easily achieved with a lower radiation dose and a single upright view of the chest, which will also serve to exclude lung pathology. However, in both the neonate and the older child, AXR findings are often nonspecific, which limits the role of this modality. Contrast studies of the gastrointestinal (GI) tract are essential in certain conditions, such as suspected mid-gut malrotation and congenital bowel obstruction. In the latter situation, contrast enema may be important in some patients for diagnosis and in others for therapy as well. Computed tomography (CT) may be reserved for more complicated imaging situations, when US may not provide a

all the information required, e.g., appendicitis when gas obscures the right lower quadrant, or in older, obese children and those children with abscesses. CT without contrast injection is also extremely helpful in delineating urinary stones when these are not well shown by US. Magnetic resonance imaging (MRI) has a very limited role in children with acute abdominal pain. However, it can depict the anatomy exceptionally well in certain conditions, in which case it may be used to complement findings on US. These include biliary and pancreatic duct anomalies and gynecological disorders, such as complex anomalies associated with hydrocolpos.

Acute Appendicitis Acute appendicitis is a common clinical entity in pediatrics. In many patients, it is easy to make the diagnosis clinically with certainty and no imaging is required prior to appendicectomy. However, imaging is extremely important in those children with non-specific symptoms or signs of acute appendicitis. In such cases, we have used US as the modality of initial choice, reserving CT for those patients in whom the US examination is inconclusive or when abscesses are present in order to better define their extent prior to drainage by the interventional radiology team. The diagnosis of appendicitis is made on US when the appendix is >6 mm in diameter and is non-compressible. These features should not be considered absolute and others should be taken into account, including edema of the mesentery, hyperemia of the wall of the appendix on color or power Doppler examination, the presence of an appendicolith and local fluid collections, or abscess formation (Figs. 2, 3). There are other conditions that may

b Fig. 2 a, b. Abortive appendicitis in a 9-yearold boy who had local peritoneal tenderness for one day, exactly at the site of the appendix, and normal temperature. a US shows borderline appendix diameter (6 mm) and slight edema of the apendiceal mesentery (echogenic triangle, arrow) but no surrounding edema. b Color Doppler US demonstrated marked hyperemia. Complaints subsided within 2 days without therapy

a

b Fig. 3 a, b. Rapidly progressing appendicitis in an 8-year-old boy who had local peritoneal tenderness exactly at the site of the appendix for 1 day and low grade fever. a US shows thickened appendix (8 mm) with edema of the appendicular mesentery (echogenic triangle, arrow) and extensive surrounding edema. b Color Doppler US demonstrated moderate hyperemia. At surgery, an inflamed appendix was removed

169

An Approach to Imaging the Acute Abdomen in the Pediatric Population

cause the appendix to become thick-walled and dilated; these include cystic fibrosis, Henoch-Schonlein purpura, and inflammatory bowel diseases.

Malrotation See the chapter “Malrotation: Techniques, Spectrum of Appearance, Pitfalls, and Management”, in the “Kangaroo” section of this syllabus for a review of this important cause of acute abdomen in the pediatric population.

Intussusception Intussusception is not an uncommon phenomenon in children and is one of the commoner causes of the acute abdomen in those between 6 months and 5 years of age. The vast majority of intussusceptions arise in the ileum and are either ileocolic or ileo-ileocolic. They are thought to occur because of hyperplasia of the lymphoid tissue in the ileum, possibly as a result of a viral infection. There are other types of intussusceptions that may relate to pathological lead points or gastro-jejunostomy tubes, or that are seen in the post-operative period. These are discussed separately at the end of this chapter. Another type of intussusception is the benign small bowel intussusception. This is often an incidental finding and does not present as an acute abdominal emergency. This entity is discussed separately in the chapter “Pediatric Intestinal Ultrasonography”, in the “Kangaroo” section of this syllabus.

intussusception promptly and accurately. The diagnosis can be made by US, AXR, or contrast studies of the colon. Ultrasonography has been shown in many series to be 100% accurate in depicting the presence or absence of the common types of ileocolic or ileo-ileocolic intussusceptions in children. These lesions have a characteristic sonographic appearance and are usually found just under the abdominal wall, most commonly on the right side of the abdomen (Fig. 4). Since it is a non-invasive procedure and because of its accuracy, sonography is the modality of choice for the evaluation of patients suspected of having an intussusception. The sonographic appearance of intussusception was excellently reviewed in a 1996 article by del-Pozo et al. (see “Suggested Reading”). Some of the characteristic signs of an intussusception can be seen on AXR, including the meniscus sign, target sign, and, less commonly, a soft-tissue mass (Fig. 5).

Ileocolic and Ileo-ileocolic Intussusceptions Diagnosis Children presenting with ileocolic or ileo-ileocolic intussusception commonly do not present with the classical clinical triad of abdominal pain, red currant jelly stool, and a palpable abdominal mass. Instead, the presentation may be non-specific. For this reason, the clinician often has to rely on imaging procedures to diagnose or exclude

a Fig. 4 a, b. Sonograms of children with ileocolic intussusception. a Transverse scan through an intussusception shows the typical target sign, which can be easily detected just deep to the abdominal wall anteriorly. The characteristic of this target sign is multiple concentric rings representing alternating mucosal and muscular layers. b Transverse scan through an intussusception shows the presence of a lymph node (arrow) within the mesentery which has been drawn into the intussusception between the layers of the intussusceptum

Fig. 5. Child with ileocolic intussusception that extended into the transverse colon. Abdominal radiograph shows a gasless right upper abdomen due to the presence of the intussusception on the right. The arrows indicate a soft-tissue mass that represents the intussusception, which is in the transverse colon

b

170

Alan Daneman, Simon G. Robben

However, in our institutions, we have relied on the plain radiograph only in those instances in which there is a clinical consideration of peritonitis. In this clinical setting, AXR is essential to exclude perforation, which is the major contraindication for attempted enema reduction. Contrast enema represents a more invasive diagnostic procedure, requiring radiation. Moreover, in a study by Daneman and Navarro, in 2003, in which patients were administered contrast enema for diagnosis, 50% of the enemas were negative for intussusceptions (see “Suggested Reading”). Using sonography as the diagnostic modality enables one to avoid performing unnecessary contrast enemas in those patients without intussusception. Reduction According to the many series in the recent literature, the reduction rate of intussusception ranges from 80% to as high as 95%. These series have used either fluoroscopic or sonographic guidance for reducing the intussusception and either hydrostatic (barium, water-soluble contrast, saline) or pneumatic reduction. The fact that different techniques have been used with similar success rates suggests that it is not important which technique is used. Non-operative reduction of an intussusception should only be attempted after the surgical team has evaluated the patient and the patient is clinically stable, well-hydrated, has no evidence of peritonitis, and has an intravenous line in place. The major contraindications to administering the enema are the clinical findings of peritonitis or shock or signs of perforation on an abdominal radiograph. In order to improve the reduction rate, delayed, repeated reduction attempts can be used as long as the intussusception moves in response to the initial attempted reduction and the child becomes asymptomatic and maintains stable vital signs. It has been shown that this approach is safe and effective, with a good success rate. Navarro et al., in a study published in 2004 (see “Suggested Reading”), used this approach in approximately 15% of patients with intussusceptions, achieving successful reduction in 50% of those intussusceptions not reduced on the first attempt. There does not appear to be a fixed optimal timing between attempts, and delayed second or third attempts can be made several hours after the first. Pathological Lead Points Pathological lead points are found in about 5-7% of all intussusceptions. The commonest are Meckel diverticulum, polyps, Henoch-Schonlein purpura, and cystic fibrosis. Less common causes are lymphoma, duplication cyst, and various inflammatory lesions of the bowel. Management of these patients remains a challenge. Contrast or air enema techniques are not always diagnostic in documenting the presence of a pathological lead point. Sonography is extremely useful in this regard as it may depict two-thirds of pathological lead points, providing a specific diagnosis in one-third of these cases (Fig. 6).

Fig. 6. A 3-month-old child with an intussusception due to a duplication cyst. In the transverse scan, the duplication cyst (C) is easily identified as a pathological lead point

However, it remains a diagnostic challenge as to how to search for pathological lead points in those patients in whom there is a high index of suspicion for this lesion and in whom the sonogram is negative. In such cases, the choice of which other imaging modalities to use will depend on the clinical situation in each particular patient. We recommend attempted enema reduction in all patients with a lead point if there is no contraindication to non-operative reduction. Postoperative Intussusception and Intussusception with Gastroenterostomy Tubes Intussusception may occur as a complication in <1% of laparotomies. These are usually more difficult to diagnose on sonography than the usual ileocolic intussusceptions, as they are small bowel intussusceptions that are often surrounded by large, dilated loops of obstructed bowel. They frequently require surgical reduction. Intussusceptions may complicate the presence of a gastrojejunostomy (GJ) tube. Most of the patients presenting with this complication are usually clinically stable and do not require urgent reduction of the intussusception. The majority can be managed by replacing the tube with a standard or a shortened GJ tube or with a gastrostomy tube. However, manipulation of the GJ tube and flushing with air or saline may also be helpful. Surgery is rarely required for reduction.

Necrotizing Enterocolitis Necrotizing enterocolitis (NEC) usually presents in infants in the neonatal intensive care unit, and more commonly in premature neonates. The classical presentation includes abdominal distention and blood in the stool. The radiologist plays an important role at the time of

171

An Approach to Imaging the Acute Abdomen in the Pediatric Population

a

b

Fig. 7 a, b. a Necrotizing enterocolitis in a premature infant. Bowel loops show distention with gas and thickened bowel walls. Large bubbles of intramural air are seen at the descending colon; a bubbly appearance in the right flank also suggests the presence of intramural air in the ascending colon and distal ileal loops. b Detail shows the intramural gas (arrow) to better advantage. Intramural gas is most commonly seen in newborns with necrotizing enterocolitis

diagnosis of NEC, during follow-up, and in the detection of later complications such as strictures. At the time of diagnosis, there are three abnormalities that may be seen on AXR: bowel dilatation, intramural gas, and portal venous gas. Bowel dilatation is present in almost 100% of the patients with NEC and the degree of distention of the bowel usually correlates well with the clinical severity. Follow-up AXR may show asymmetrical dilatation and fixed loops in those infants who deteriorate. Intramural gas is not present in 100% of patients and the amount of intramural gas does not always correlate well with the degree of clinical severity (Fig. 7). Portal venous gas usually is present in those with severe NEC. Disappearance of the intramural gas and portal venous gas does not necessarily correlate with clinical improvement, as in either case the gas will eventually disappear even in those children who deteriorate clinically. Ultrasound is an extremely useful modality in the investigation of patients with NEC as it can provide information regarding the presence of intraperitoneal fluid, bowel wall thickness, and bowel perfusion (using color or power Doppler sonography). US is much more accurate than AXR in documenting the presence of free and focal intra-peritoneal fluid and can also define the character of this fluid. It is well known that not all patients with NEC will show free air on AXR following perforation; instead, they may only present with free fluid. In the early phases of NEC, we have found that the bowel wall will be quite thickened; in patients who are more severely affected, however, the mucosa and submucosa of the bowel sloughs into the lumen of the bowel, leaving a markedly thinned bowel wall, which is much more prone to perforation. Thinning of the bowel wall can be documented with sonography. In NEC, the bowel (particularly, the thickened bowel) becomes markedly hyperemic, which indicates the presence of viable bowel. However, the absence of bowel perfusion in single or multiple loops of

bowel (especially when the bowel wall is thinned) indicates the presence of necrosis – a condition that may warrant surgical intervention even if free air is not seen on AXR. Sonography comes to play a more major role in the follow-up of both those patients who are not responding to medical management and those who deteriorate clinically. In such cases, US may provide information that is not depicted with AXR.

Meckel Diverticulum Meckel diverticulum most commonly presents as painless rectal bleeding due to ulceration because of the presence of ectopic gastric mucosa. These patients are usually adequately diagnosed and managed following a radionuclide scan. In <50% of the children presenting with Meckel diverticulum, the clinical findings will be more complex, with a combination of abdominal pain, vomiting, and, occasionally, rectal bleeding. In these children with acute pain, the diagnosis is often difficult and non-specific. US can be used successfully to document the presence of an inflamed or hemorrhagic Meckel diverticulum, which in this situation has a variable appearance and may simulate the presence of an inflamed duplication cyst, appendicitis, or sometimes a small intussusception. The finding of this somewhat atypical appearance on US is diagnostically suggestive of a complicated Meckel diverticulum rather than the other pathologies it simulates.

Congenital Bowel Obstruction in the Neonate Obstruction due to congenital lesions may occur at all levels of the GI tract and are, from a practical point of view, usefully divided into those lesions that are termed

172

Alan Daneman, Simon G. Robben

might suggest each of the above conditions, the radiographic findings in all of them are often non-specific. Differentiation of these conditions will therefore depend on other factors, such as the clinical history and the results of the physical examination. Invariably, to increase diagnostic confidence, these conditions require the administration of a water-soluble contrast enema to define the distal colon and ileum. The contrast enema may also prove therapeutic in patients with meconium ileus or meconium plugging the colon.

Suggested Reading

Fig. 8. High gastrointestinal obstruction in a newborn with trisomy 21 (Down syndrome). A double-bubble appearance is present due to distention of the stomach and duodenum by gas. This finding suggests duodenal atresia. However, small amounts of gas are noted distally, indicating a partial obstruction, such as a duodenal web, duodenal stenosis, or malrotations with Ladd’s bands. In this patient, an annular pancreas was found at surgery. Patients with Down syndrome can have a variety of congenital duodenal abnormalities

“high” and those that are termed “low”. High obstructions denote lesions of the esophagus, stomach, duodenum, and upper small bowel (Fig. 8). Low obstructions include lesions of the lower small bowel and large bowel, and anorectal malformations. The distribution of dilated bowel loops on plain radiographs usually enables one to differentiate high from low obstruction relatively easily. This is simply done by evaluating the number of gas-filled loops that are visible. Fluid-filled loops may be difficult to visualize on AXR, possibly masquerading as free fluid or masses; thus, they can occasionally confuse the picture. It should be emphasized that the differentiation of dilated gas-filled small from large bowel loops may be impossible in neonates. Free air is not usually evident in these patients unless the diagnosis is delayed. Intramural air (and even portal venous gas) may be seen proximal to a high grade obstruction but it is much more commonly seen in NEC. Calcification may be present in the peritoneum (meconium peritonitis) due to prenatal perforation, in the wall of the bowel proximal due to an atresia, or in the bowel content within the lumen, occasionally proximal to a low obstruction. Most complete high obstructions are easily diagnosed on AXR. If the diagnosis is in doubt, air can be injected slowly via a feeding tube into the lumen of the GI tract to confirm or exclude an obstruction. In incomplete obstructions (e.g., malrotation and stenoses), positive water-soluble contrast agents are required to confirm the level and nature of the obstruction. The low obstructions include ileal and colonic atresias, meconium ileus, functional immaturity of the large bowel, Hirschsprung’s disease, and anorectal malformations. Although there are some features on AXR that

Ang A, Chong NK, Daneman A (2001) Pediatric Emergency Care 17:334-340 Baldisserotto M, Marchiori E (2000) Accuracy of Noncompressive Sonography of Children with Appendicitis According to the Potential Positions of the Appendix. AJR Am J Roentgenol 175:1387-1392 Bombelburg T, Von Lengerke HJ (1992) Sonographic findings in infants with suspected necrotizing enterocolitis. European J Radiol 15:149-153 Buonomo C (1999) The radiology of necrotizing enterocolitis. RCNA 37:1187-1198 Couture A, Baud C, Ferran JL et al (2008) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Heidelberg Berlin New York Daneman A, Alton DJ, Ein S et al (1995) Perforation during attempted intussusception reduction in children – a comparison of perforation with barium and air. Pediatr Radiol 25:81-88 Daneman A, Alton DJ, Lobo E et al (1998) Patterns of recurrence of intussusception in children: a 17 year review. Pediatr Radiol 28:913-919 Daneman A, Lobo E, Alton DJ, Shuckett B (1998) The value of sonography, CT and air enema for detection of complicated Meckel diverticulum in children with nonspecific clinical presentation. Pediatr Radiol 28:928-932 Daneman A, Myers M, Shuckett B, Alton DJ (1997) Sonographic appearances of inverted Meckel diverticulum with intussusception. Pediatr Radiol 27:295-298 Daneman A, Navarro O (2003) Intussusception Part 1: A review of diagnostic approaches. Pediatr Radiol 33:79-85 Daneman A, Navarro O (2004) Intussusception Part 2: An update on the evolution of management. Pediatr Radiol 34:97-108 Daneman A, Navarro O (2005) Intussusception: the debate endures. Pediatr Radiol 35:95-96 del-Pozo G, Abillos JC, Tejedor D (1996) Intussusception: US findings with pathologic correlation – the crescent-in-doughnut sign. Radiology 199:688-692 del-Pozo G, Albillos JC, Tejedor D et al (1999) Intussusception in children: current concepts in diagnosis and enema reduction. Radiographics 19:299-319 Doria AS, Amernic H, Dick P et al (2005) Cost effectiveness analysis of weekday and weeknight shifts for assessment of appendicitis. Pediatr Radiol 35:1186-1195 Elsayes KM, Menias CO, Harvin HJ, Francis IR (2007) Imaging manifestations of Meckel’s diverticulum. AJR Am J Roentgenol 189:81-88 Faingold R, Daneman A, Tomlinson G et al (2005) Bowel viability assessment by colour doppler sonography in necrotizing enterocolitis. Radiology 235:587-594 Fotter R, Sorantin E (1994) Diagnostic imaging in necrotizing enterocolitis. Acta Paediatr Supp 396:41-44 Frush DP, Frush KS, Oldham KT (2009) Imaging of acute appendicitis in children: EU versus U.S… or US versus CT? A North American perspective. Pediatr Radiol 39:500-505

An Approach to Imaging the Acute Abdomen in the Pediatric Population

Fujii Y, Hata J, Futagami K et al (2000) Ultrasonography improves diagnostic accuracy of acute appendicitis and provides cost savings to hospitals in Japan. J Ultrasound Med 19:409-414 Gu L, Alton DJ, Daneman A et al (1988) Intussusception reduction in children by rectal insufflation of air. AJR Am J Roentgenol 150:1345-1348 Holscher HC, Heij HA (2009) Imaging of acute appendicitis in children: EU versus U.S… or US versus CT? A European perspective. Pediatr Radiol 39:497-499 Hughes UM, Connolly BL, Chait PG, Muraca S (2000) Further report of small-bowel intussusceptions related to gastrojejunostomy tubes. Pediatr Radiol 30:614-617 Kim G, Daneman A, Alton DJ et al (1997) The appearance of inverted Meckel diverticulum with intussusception on air enema. Pediatr Radiol 27:647-650 Kornecki A, Daneman A, Navarro O et al (2000) Spontaneous reduction of intussusception: clinical spectrum, management and outcome. Pediatr Radiol 30:58-63 Kosloske AM, Love CL, Rohrer JE et al (2004) The diagnosis of appendicitis in children: outcomes of a strategy based on pediatric surgical evaluation. Pediatrics 113:29-34 Levy AD, Hobbs CM (2004) From the Archives of the AFIP. Meckel Diverticulum: Radiologic Features with Pathologic Correlation. Radiographics 24:565-587 McDermott VGM (1994) Childhood intussusception and approaches to treatment: A historical review. Pediatr Radiol 24:153-155 Meradji M, Hussain SM, Robben SG, Hop WC (1994) Plain film diagnosis in intussusception. Br J Radiol 67:147-149 Miller SF, Seibert JJ, Kinder DL, Wilson AR (1993) Use of ultrasound in detection of occult bowel perforation in neonates. JU Med 12:531-535 Navarro O, Daneman A (2004) Intussusception. Part 3: Diagnosis and management of those with an identifiable or predisposing cause and those that reduce spontaneously. Pediatr Radiol 34:305-312 Navarro OM, Daneman A, Chae A (2004) Intussusception: The use of delayed, repeated reduction attempts and the management of intussusceptions due to pathologic lead points in pediatric patients. AJR Am J Roentgenol 182:1169-1176 Navarro O, Dugougeat F, Kornecki A et al (2000) The impact of imaging in the management of intussusception owing to pathologic lead points in children. A review of 43 cases. Pediatr Radiol 30:594-603 O’Hara SM (1996) Pediatric gastrointestinal nuclear medicine. Radiol Clin North Am 34:845-862

173

Park NH, Park SI, Park CS et al (2007) Ultrasonographic findings of small bowel intussusception, focusing on differentiation from ileocolic intussusception. Br J Radiol 80:798-802 Pena BMG, Taylor GA, Fishman SJ, Mandl KD (2002) Effect of an imaging protocol on clinical outcomes among pediatric patients with appendicitis. Pediatrics 110:1088-1093 Poortman P, Lohle PNM, Schoemaker CMC et al (2003) Comparison of CT and sonography in the diagnosis of acute appendicitis: a blinded prospective study. AJR Am J Roentgenol 181: 1355-1359 Puylaert JB (1986) Acute appendicitis: US evaluation using graded compression. Radiology 158:355-360 Rohrschneider W, Troger J, Betsch B (1994) The post reduction donut sign. Pediatr Radiol 24:166-170 Rohrschneider WK, Troger J (1995) Hydrostatic reduction of intussusception under US guidance. Pediatr Radiol 25:530-534 Sargent MA, Babyn P, Alton DJ (1994) Plain abdominal radiography in suspected intussusception: A reassessment. Pediatr Radiol 24:17-20 Shanbhogue RL, Hussain SM, Meradji M et al (1994) Ultrasonography is accurate enough for the diagnosis of intussusception. J Pediatr Surg 29:324-327 Silva CT, Daneman A, Navarro OM et al (2007). Correlation of sonographic findings and outcome in necrotizing enterocolitis. Pediatr Radiol 37:274-282 Sivit CJ (2004) Imaging the child with right lower quadrant pain and suspected appendicitis: current concepts. Pediatr Radiol 34:447-453 Taylor GA (2004) Suspected appendicitis in children: In search of the single best diagnostic test. Radiology 231:293-295 Thurley PD, Halliday KE, Somers JM et al (2009) Radiological features of Meckel’s diverticulum and its complications. Clin Radiol 64:109-118 Todani T, Sato Y, Watanabe Y et al (1990) Air reduction for intussusception in infancy and childhood: Ultrasonographic diagnosis and management without X-ray exposure. Z Kinderchir 45:222-226 Verschelden P, Filiatraut D, Garel L et al (1992) Intussusception in children: reliability of US in diagnosis – a prospective study. Radiology 184:741-744 Wang G, Liu S (1988) Enema reduction of intussusception by hydrostatic pressure under ultrasound guidance. A report of 377 cases. J Pediatr Surg 23:814-818 Wiersma F, Allema JH, Holscher HC (2006) Ileoileal intussusception in children: ultrasonographic differentiation from ileocolic intussusception. Pediatr Radiol 36:1177-1181

IDKD 2010-2013

Imaging Uronephropathies in Children Jeanne S. Chow1, Fred E. Avni2 1 Departments 2 Department

of Urology and Radiology, Children’s Hospital, Boston, MA, USA of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium

Introduction and Techniques

Management of In Utero Renal Pelvis Dilatation

The principle that guides imaging of the pediatric genitourinary tract is the same that guides imaging elsewhere in the body of the child: judicious use of imaging while minimizing ionizing radiation exposure [1] and unnecessary sedation. Ultrasound (US) is the main imaging modality of the genitourinary tract, providing excellent images of the kidneys and bladder. Also, at the level of these organs, an abnormally dilated ureter is easily imaged. Doppler provides additional information regarding vascular flow and is especially helpful in evaluating the main renal artery and veins as well as the arcuate vessels. US contrast agents are used to further evaluate the vascular flow of the kidneys, intrarenal masses, and vesicoureteral reflux [2]. The main limitation of US is that it provides little functional imaging, instead mainly revealing the physical appearance of the urinary tract. If further imaging is necessary, contrast-enhanced computed tomography (CT) and magnetic resonance (MR) provide exquisite, detailed images of the urinary tract in addition to more functional information. MR urography has become very popular because it emits no ionizing radiation and yields both anatomical and functional data. The latter includes uptake and excretion rates, which together with relative function and indirect measurements of glomerular filtration rates provide more information than obtained from a MAG-3/Lasix renogram [3]. However, because MR is a long procedure that often requires sedation, Tc-99m(V) dimercaptosuccinic acid (DMSA) and MAG-3 are more commonly used to provide functional imaging of the genitourinary tract. In addition, functional information cannot be obtained without gadolinium, but this approach places children with chronic renal failure and end-stage renal disease at risk for developing nephrogenic systemic fibrosis [4]. Although nephroureterolithiasis may be commonly studied by CT in adults, US is the first imaging modality in children [4, 5].

Dilatation of the renal pelvis is a common finding on obstetric US, with a frequency of around 1-4% of all pregnancies. Yet, every dilatation does not have the same clinical relevance; furthermore, the antenatal and postnatal evolution is variable. This has led to controversy in the literature about the best work-up and follow-up after birth [6-8].

Key Point – Ultrasound is the main imaging modality of the genitourinary tract in children.

Definition The best criterion for objectivating urinary tract (UT) dilatation is the anteroposterior diameter of the renal pelvis, which is measured on a transverse scan of the fetal abdomen. The upper limit for normal should be 4 mm in the second trimester and 7 mm in the third trimester. Other US evidence of a UT abnormality includes the visibility of the fetal ureter, dilatation of the renal calyces, abnormal echogenicity of the renal parenchyma, and the demonstration of an enlarged bladder (Table 1).

Findings on Obstetric US Abnormalities of the UT can be found at any time during pregnancy. Once a dilatation has been detected in utero, subsequent evaluations should address three issues: the origin of the dilatation, the co-existence of associated anomalies, and the prognosis. The most common causes of dilatation in the fetus are related to obstructions of the ureteropelvic and ureterovesical junctions and to complicated duplex kidneys. UT Table 1. Characteristics suggesting an abnormality of the urinary tract on neonatal ultrasound Pelvic dilatation >7 mm Calyceal dilatation Parenchymal thinning Lack of cortico-medullary differentiation Small kidney Thickening of the pelvic wall Thickening of the ureteral wall Ureteral dilatation >3 mm Enlarged bladder

175

Imaging Uronephropathies in Children

dilatation can also result from vesicoureteric reflux (VUR) and intravesical obstruction, especially in male fetuses. In most cases, the US evaluation will be able to differentiate between etiologies. In some patients, especially those with bilateral and complex uropathies, fetal MR imaging will provide additional information. Other organ malformations also can be associated with UT dilatation; therefore, the US survey should be as meticulous and complete as possible. Chromosomal analysis may be indicated in selected patients. The prognosis of a uropathy will depend upon the type and extent of the anomalies. Amniotic fluid volume is important to the prognosis as well; oligohydramnios, thought to be related to decreased urine production, is a poor prognostic indicator. It is of utmost importance that any relevant information is correctly transmitted to the postnatal team that will be in charge of caring for the newborn.

Postnatal Management of Fetal Pelvis Dilatation Certain conditions require immediate postnatal confirmation and therapeutic maneuvers, such as obstructive posterior urethral valves and prolapsed ectopic ureterocele into the urethra. In those cases, US and micturating cystourethrography (MCU) should be performed directly after birth. In all other cases, the work-up can be planned without urgency. Patients with ureterovesical junction obstruction (UVJO) and complex uropathies should be put on prophylactic chemotherapy at least until the final diagnosis is made. An algorithm based on US examination is presently applied by most teams (Fig. 1). Micturating voiding urethrography is only applied if US displays a significant anomaly (Table 1). MCU is used to detect high-grade VUR and urethral anomalies. If VUR is not present, complementary imaging is necessary to determine the precise

origin of the dilatation. Renal function is assessed through isotopic studies, while complicated UT malformations are best evaluated by MR imaging. The latter is particularly helpful for the assessment of a very dilated UT and complicated duplex-kidney systems. The type of treatment (conservative or surgical) will depend upon the diagnosis, renal function on follow-up, and complications. Today, the trend is increasingly toward a conservative approach. The length of follow-up must be adapted to the type of anomaly as well as to clinical and imaging follow-up. Key Points – Fetal renal dilatation is a common finding during obstetric US. – Thresholds of 4 and 7 mm during the 2nd and 3rd trimester, respectively, are widely accepted. – Postnatally, these patients must be further evaluated by US; voiding cystourethrography is performed only if an anomaly is found at birth or at the age of one month. – The trend is toward a more conservative approach to treatment, based on clinical and imaging follow-up.

Imaging Cystic Kidneys in Children Introduction Renal cystic diseases may be discovered or suspected at any stage during fetal life or at any age in childhood. They encompass a large number of conditions that can be separated into those with or without hereditary transmission. Imaging, mainly US, plays an important role in differentiating between the various types of cystic diseases as it shows the features of renal involvement as well as associated anomalies.

US : 1st US around day 5

abnormal: pelvis ≥7 mm + dilated calices, or other anomalies VCUG

normal

normal US at 1 mo

abnormal

US at 3 mo

abnormal pelvis ≥ 10 mm other malformation, “extended criteria” normal

pelvis ≥10 mm

Stop follow-up

pelvis >10 (15) mm

further morphological & functional evaluation: scintigraphy, IVU, MRU …

Stop follow-up

Fig. 1. Algorithm in fetal hydronephrosis (HN). Antenatal diagnosis of mild to moderate renal pelvis dilatation [6]. VCUG, Voiding Cystourethrography

176

Cystic Kidneys in the Fetus In the fetus (and during the perinatal period), cystic renal disease should be suspected whenever bilateral hyperechoic kidneys or cysts (uni- or bilateral) are discovered during an obstetric US examination. The imaging approach to the diagnosis should be based on a detailed sonographic analysis that includes measurement of renal length, the presence or absence of cortico-medullary differentiation (CMD), and the presence, number, size, and location of the cysts. This evaluation should be completed through an analysis of the entire fetus, looking for associated anomalies. The timing of detection and the amount of amniotic fluid are the most important prognostic factors. Furthermore, a detailed clinical and familial inquiry is essential in disease evaluation. Autosomal recessive polycystic kidney disease (ARPKD) is the main diagnosis to consider in case of bilateral, markedly enlarged, hyperechoic kidneys without CMD. The prognosis is usually poor if the amniotic fluid volume is markedly decreased. In case of moderately enlarged hyperechoic kidneys, three diagnoses have to be considered: 1. nephropathy due to a mutation in the TCF2 gene, 2. a milder form of ARPKD, and 3. autosomal dominant polycystic kidney disease (ADPKD). In TCF2-mediated nephropathy, CMD is absent or present and whenever cysts are visible they are typically subcortical. In the milder form of ARPKD, cysts may be observed in the medullary area. The main sonographic feature of ADPKD is a striking cortical hyperechogenicity associated with increased CMD. Whenever cysts are the main US finding in the fetal kidneys, their number and location are the main criteria for the differential diagnosis. The diagnosis of unilateral multiple cysts suggests a multicystic dysplastic kidney. Bilateral multiple cysts can be visualized in a large number of renal or syndromic diseases, the most common diagnoses being bilateral multicystic dysplastic kidney, ADPKD, bilateral obstructive dysplasia, and glomerulocystic kidneys.

Renal Cystic Diseases in Children In children, renal cystic diseases are usually discovered during US examination performed in the follow-up of a known perinatally diagnosed disease, during the work-up of syndromes diagnosed after birth, during screening in an at-risk family, or as an incidental finding. The US approach is the same as that described for the fetus. The role of imaging in the diagnosis and follow-up of renal involvement is to search for complications such as hemorrhage or urolithiasis. Specifically, an important role for US is the detection of hepatic-biliary complications. Key Points – Renal cystic diseases can be diagnosed in the perinatal period or in later childhood.

Jeanne S. Chow, Fred E. Avni

– Hyperechogenicity or cysts are cardinal findings. – A familial history, detailed clinical inquiry, and associated findings help in establishing the diagnosis.

Renal Ectopia and Duplications One of the most interesting areas of pediatric uroradiology is studying and understanding the multitude of congenital abnormalities of the urinary tract. During normal renal development, the kidneys ascend from the renal pelvis while rotating medially. If the kidneys do not ascend or ascend past their normal location in the renal fossae, they are ectopic. In some cases they are as low as the pelvis and in others as high as the thoracic cavity. If the kidneys fuse during ascent, pelvic cake kidneys, midline horseshoe kidneys, or left- or right-sided cross-fused ectopic kidneys form. Since the embryological origin of the kidneys (metanephros) is separate from that of the ureters (ureteric buds), the site of ureteral insertion is normal even if the kidney is ectopic. However, the renal blood supply from the aorta will vary depending on the level of ectopia. The ureteric bud must meet the metanephros in order for the kidney to form. Without this interaction, kidney formation is not induced. If two ureteric buds meet at the metanephric blastema, then the kidney becomes “duplex”. Ureteral duplication may be complete or, more commonly, incomplete. In incomplete ureteral duplication, a single ureteric bud bifurcates and meets the metanephros during approximately the 5th to 6th week of gestation. The two branches of the ureter may join at the level of the renal pelvis (bifid pelvis) or at the proximal, mid-, or distal ureter (bifid ureter) and terminate in a single distal ureter that inserts orthotopically into the bladder. Since the two moieties of the kidney share a common distal ureter, they behave similarly and usually appear normal. Rarely, one of the ureteral buds may be blind-ending and never appear to “reach” the kidney (blind-ending ureteral duplication). The associated kidney has a single collecting system. In complete ureteral duplication, two separate ureteric buds arise from the Wolffian duct. The lower-pole ureter is considered the analogue to the normal single-system ureter. Thus, the lower pole of the kidney has all of the same abnormalities that can affect a single-system kidney, including VUR, ureteropelvic junction obstruction (UPJO), and UVJO. The upper-pole ureter is “abnormal” and ectopic (Weigert-Meyer rule). The ectopic ureter inserts medially and inferiorly to the normal ureteral orifice, usually in the bladder. In girls, the ectopic ureter may insert below the bladder base, into the urethra or the vagina. A vaginal ectopic ureter can cause constant urinary dribbling in girls and incontinence [9]. In boys, ectopic ureters never terminate below the urinary sphincter and thus never result in incontinence; however, the ureter can terminate in Wolffian duct derivatives, including the

177

Imaging Uronephropathies in Children

seminal vesicals and vas deferens. Very rarely, three completely or incompletely separated ureters form, resulting in ureteral triplication [10]. Ectopic ureters are often obstructed but rarely reflux. If the ectopic ureter inserts into the urethra at the level of the urinary sphincter, urinary flow is obstructed or refluxes depending whether the sphincter is closed or open [11]. The more distal the ureteral insertion, the more dysplastic and dysfunctional the associated renal parenchyma. Ectopic ureters, and all the associated abnormalities, can also occur in single-system kidneys (single ectopic ureter) [12]. A ureterocele is the dilated submucosal terminal segment of the ureter. It is associated with varying degrees of ureteral obstruction and subsequent dilatation of the renal pelvis and calyces. In girls, ureteroceles are most commonly seen in association with ectopic upperpole ureters. In boys, they are most commonly associated with single-system kidneys and are orthotopic. Although ureteroceles protrude into the bladder, when the intravesical pressure equals that of the ureterocele, the ureterocele can flatten and become imperceptible (efface). When the intravesical pressure exceeds that of the ureterocele, the latter everts or intussuscepts into its ureter. Ectopic bladder-neck ureteroceles or large simple ureteroceles can prolapse into the urethra and cause bladder outlet obstruction. Key Points

sent with intermittent pain from intermittent obstruction, with hydronephrosis only evident during obstruction. To be correctly diagnosed, these children must be imaged at the time of their painful episodes [13]. The conundrum of UPJO is that we are still unable to predict whether the degree of obstruction and thus its eventual effects on renal function will improve or worsen over time. US is routinely used to describe the degree of obstruction and the appearance of the renal parenchyma. However, functional imaging studies, primarily MAG-3 studies with Lasix (MAG-3/Lasix renogram) and MR urography, are used to help quantify the degree of obstruction and the contributing function of each kidney. An obstruction of the distal ureter as it enters into the bladder results in UVJO. Most such cases are primary and due to a ureteral obstruction, although secondary UVJO can occur with an abnormally thickened bladder. The insertion of the obstructed ureter may be orthotopic (primary mega-ureter) or ectopic. An orthotopic or ectopic ureterocele may also be associated with obstruction. Primary mega-ureter accounts for the majority of the cases of UVJO. In most patients with this condition, the degree of dilatation improves over time [14] such that surgical repair is required only for a minority of affected patients. Surgery is indicated if the degree of dilatation worsens, renal function is impaired, or the obstruction is thought to be contributing to stasis and UT infections.

– Renal ectopia is due to abnormalities in the normal ascent of the kidney. – Ureteral duplication may be incomplete (more common) or complete. – The Weigert-Meyer rule states that the upper-pole ureter of a duplex kidney inserts ectopically, medially, and inferiorly to the orthotopic location. – The lower-pole ureter is the analogue of the single-system kidney.

Key Points

Urinary Tract Obstruction

Children with lower-spine abnormalities may have detrusorsphincter dysynergia, in which the bladder contracts but the urinary sphincter does not relax during voiding, resulting in chronic obstruction of the bladder outlet. The uncoordinated voiding causes urinary retention, increased bladder pressures, secondary VUR, and secondary UVJO. Similarly, children with voiding dysfunction without a neurogenic cause can also develop secondary reflux. Urethral obstruction can occur in the posterior or anterior urethra in boys whereas the urethra is rarely obstructed in girls. Posterior urethral valve obstruction is the most common congenital urethral obstruction and is caused by an obstructing membrane just below the level of the verumontanum. Anterior urethral obstruction is most commonly due to traumatic strictures and mostly located in the bulbar urethra. Anterior urethral valves or diverticula are rare. Depending on the severity of the ob-

Urinary tract obstructions occur at three main areas: the ureteropelvic junction, the ureterovesical junction, and the bladder outlet (i.e., the urethra). Rarely, the midureter can be obstructed by webs, fibrosis, or compression from the inferior vena cava, or there may be obstruction at the level of the infundibula in the kidney. On US, the normal hypoechoic medullary pyramids seen routinely in infancy and childhood should not be confused with dilated calyces or a sign of obstruction. The most common congenital obstruction of the kidney is UPJO, which is due to a stenosis at the junction of the renal pelvis and proximal ureter. Since most children are now diagnosed prenatally and followed postnatally, they rarely present with symptoms and signs of obstruction, such as infection, pain, or renal stones. Some children have UPJO due to a crossing renal artery and pre-

– Ureteropelvic junction obstruction is the most common cause of urinary tract obstruction. – Most cases of ureterovesical junction obstruction improve with time.

Voiding Abnormalities and Secondary Vesicoureteric Reflux

178

struction, the portion of the urethra proximal to the obstruction may be dilated, the bladder wall may be hypertrophied, and secondary reflux and UVJO may occur. In circumcised boys, meatal stenosis is another cause of urethral obstruction. Retrograde urethrogram under fluoroscopy or ultrasound is the best way of studying the anterior urethra whereas the posterior urethra can only be studied during voiding. Key Points – Voiding dysfunction can lead to vesicoureteral reflux. – Most obstructions of the bladder outlet in boys are congenital (e.g., posterior and anterior urethral valves) or post-traumatic (bulbar urethral stricture).

Renal Masses in Children Once a mass is established to be intrarenal, its histology can be predicted based on its appearance and on the patient’s age. Most intrarenal masses occurring in the newborn period are benign. Although rare, the most common solid intrarenal mass seen in newborns is a mesoblastic nephroma [15]. These large, solid, enhancing masses are benign, although the cellular subtype is the most aggressive and can cause paraneoplastic syndromes. These must be removed but the prognosis is excellent. In the newborn, the most frequently occurring cystic abnormality of the kidney is multicystic dysplasic kidney, which can involve the entire kidney or be segmental. The condition is due to a congenital abnormality of the kidney in which the collecting system forms as cysts, and the renal parenchyma is dysplastic and nonfunctional. Multicystic dysplasic kidney is now commonly diagnosed in utero. Over time, the cyst fluid resorbs and a tiny nub of tissue remains. These are typically treated non-surgically. The most common renal mass in toddlers is Wilms’ tumor. Children with aniridia, WAGR (Wilms’ tumor, aniridia, genitourinary abnormalities, and mental retardation), Deny-Drash syndrome, Beckwith-Weidemann syndrome, hemihypertrophy, or nephroblastomatosis are predisposed to developing this tumor. Wilms’ tumor is a solid, cystic, and often hemorrhagic mass and is far more common but radiographically indistinguishable from either clear cell sarcoma or malignant rhabdoid tumor. However, if a tumor has a large subcapsular hematoma, and if there are brain metastases, malignant rhabdoid tumor should be considered [16]. Centrally located multilocular masses of the kidney may be a multilocular cystic nephroma, which is more common in boys in childhood and in women in adult life [17]. In children over 11 years of age, renal cell carcinoma becomes more common than Wilms’ tumor, although the likelihood of either tumor is extremely rare [18]. It is crucial to confirm that the child has no clinical indicators of UT infection, because focal pyelonephritis or lobar

Jeanne S. Chow, Fred E. Avni

nephronia mimics a tumor in appearance and thus must always be considered in the differential diagnosis of a renal mass. If the renal mass is bilateral, the appearance and clinical presentation are extremely helpful in predicting the histology. If there are multiple large masses and the kidneys are also enlarged, bilateral nephrogenic rests due to nephroblastomatosis are most likely. Nephrogenic rests are remnant fetal renal tissue that never fully matured. As they have a high propensity to develop into Wilms’ tumors, these masses need frequent surveillance. If the masses are partially echogenic, angiomyolipoma should be considered, especially if the patient has tuberous sclerosis. Wilms’ tumors, lymphoma, and infections may also be bilateral. Solitary simple cysts are much less commonly seen in children than in adults. Calyceal diverticula may appear as simple cysts but they actually communicate with the adjacent calyx and can become superinfected. Delayed intravenous pyelogram, CT, or MR imaging, which show contrast within the cyst, is able to distinguish calyceal diverticula from simple cysts. If there are multiple simple cysts, especially in enlarged kidneys, ADPKD should be considered. Key Points – Most newborn renal masses are benign. – The most common renal malignancy in toddlers is Wilms’ tumor. – Focal pyelonephritis mimics renal tumors.

Imaging Renal Failure in Children Introduction Ultrasound plays a central role in pediatric imaging, particularly in pediatric nephrology, in which it helps to differentiate between the etiologies of renal failure. For some diseases, the US pattern will be specific, while for others there will be little or no parenchymal changes. The US evaluation should therefore be very meticulous and correlated to the biological and clinical data [18-20].

Sonographic Technique Renal US has to be carried out with the highest-resolution transducers, depending on the patient’s size. The use of both curved and linear transducers is essential. US studies include measurements of the kidneys and of any dilatation as well as the evaluation of renal echogenicity (cysts? calcifications?) and CMD. Doppler analysis also must be performed. In case of UT dilatation, the cause and level of obstruction must be determined, including the bladder, within the field of investigation. It might be of interest in some patients to evaluate the liver, spleen, and biliary tract, too.

179

Imaging Uronephropathies in Children

Acute Renal Failure Acute renal failure (ARF) is defined as urine production <1 mg/kg/day. Its causes can be pre-renal, renal, or postrenal in origin. In the case of ARF of renal origin, US is often diagnostic. Hemolytic Uremic Syndrome Hemolytic microangiopathic anemia, thrombocytopenia, and ARF occurring together constitute hemolytic uremic syndrome, which is the commonest cause of ARF in the United States and in several European countries, especially in young infants. During the acute phase of the disease, the US appearance of the renal cortex is markedly hyperechoic, with increased CMD. On Doppler analysis, there is no diastolic flow, which correlates well with the lack of urine production. The return of the diastolic wave indicates a return to normal diuresis. Other organs may be involved as well, including the gallbladder and digestive tract. Medullary or Cortical Necrosis, Shock Kidneys Medullary and cortical necrosis in the neonate results from a lack of renal perfusion. On US, the cortex in cortical necrosis first appears hyperechoic, then shrinks, and finally calcifies. In medullary necrosis, calcifications develop within the medulla.

after birth. The most common form of CNS is the Finnish type. Proteinuria starts in utero and the placenta is thickened. On US, at birth, the kidneys are swollen and hyperechoic; CMD is present but the pyramids are irregular and within weeks they will “disappear”. Other causes of CNS include diffuse mesangial sclerosis (DMS), which can be part of the Drash syndrome (DMS, genital anomalies, and a risk for Wilms’ tumor). In some patients, CNSs evolve toward end-stage renal disease and necessitate renal transplantation. Syndromes Affecting the Tubules and Metabolic Diseases Renal diseases that include a primary and secondary tubulopathy are numerous. Hypercalciuria is a constant finding and may lead to nephrocalcinosis, which is easily detected by US. For instance, type 1 primitive hyperoxaluria is seen as strikingly hyperechoic kidneys already at birth and results in urolithiasis. Key Points – US is the key imaging examination in children with acute or chronic renal failure. – The etiologies of renal failure are numerous. – The US patterns will orient the diagnosis.

Urinary Tract Infection

Renal Vein Thrombosis

Introduction

While largely a neonatal disease, renal vein thrombosis may occur already in utero. When both renal veins are involved, the condition is associated with ARF. On US, the kidneys appear enlarged, CMD is absent, and hyperechoic streaks are demonstrated in the interlobar areas.

Urinary tract infection is one of the commonest bacterial diseases in children: 5% of girls and 0.5% of boys will suffer at least one episode. Despite numerous studies and publications, the role of imaging is controversial. The main challenges of imaging are to identify patients with complicated UT infection, those with an underlying cause, and those at risk for recurrence [21-23]. Imaging may be done at the time of diagnosis of an acute episode, or during treatment, or in a late assessment. No single imaging technique allows complete evaluation of the UT; instead, the use of each one must be optimized to obtain the maximum amount of information in association with the lowest morbidity.

Obstructive Uropathies Anuria and ARF may follow ureterocele prolapse within the urethra, obstruction of a single renal system, tumoral entrapment of the ureters, or bilateral obstructive urolithiasis.

Chronic Renal Failure Chronic renal failure (CRF) is defined as a glomerular filtration rate <50 mL/min/1.73 m2/kidney. One of the most common causes of CRF is renal hypodysplasia. On US, the kidneys are small, CMD absent, and small cysts can be visualized. Renal Cystic Diseases (see above) Congenital Nephrotic Syndromes Congenital nephrotic syndromes (CNSs) encompass diseases in which there is massive proteinuria occurring

Imaging Acute Pyelonephritis The main role of conventional US is to diagnose any associated malformative uropathy and/or the complications (abscess) of a UT infection. Color Doppler US may provide significant additional information regarding the degree of renal insult. The gold standard examination for the diagnosis of renal lesions during an acute episode (as well as for the detection of scars) is DMSA scintigraphy. CT scan with contrast enhancement is as efficient as a DMSA scan for the demonstration of acute renal lesions. Its drawbacks are the radiation hazard and the need for

180

Jeanne S. Chow, Fred E. Avni

contrast injection. MR imaging has great potential to detect renal inflammatory lesions and provides information on renal parenchyma status as well as on UT morphology. In addition, it may clarify ultrasound findings. MCU is an essential complementary examination for patients with UT infections, as there may be associated reflux in as many as 40% of cases. Therefore, the technique is generally advocated in case of UT infections to look for VUR and voiding dysfunction. However, MCU is an invasive technique and irradiation is a drawback. Proposed Work-Up Based on these considerations, it is important to identify among patients with clinical suspicion of acute pyelonephritis those who need an imaging work-up. Clearly, newborn boys and school-aged girls are groups at risk. Still, it is impossible to define why certain patients will develop renal involvement and in which of these patients the disease will recur. Therefore, every patient with a UT infection should undergo imaging evaluation albeit tailored and optimized to his/her condition. A work-up protocol is summarized in Fig. 2.

infection of the urine in a dilated urinary tract. Suspicion of this diagnosis should prompt a diagnostic and therapeutic nephrostomy.

Late Complications of UT Infection The development of renal scars is the long-term risk of untreated acute pyelonephritis. Patients with renal scars are at risk for developing renal hypertension, complications during pregnancy, and renal failure. Presently, DMSA scintigraphy is the gold standard method for the diagnosis of scars, if performed at a time sufficiently removed from the acute episode. Xanthogranulomatous pyelonephritis is an atypical chronic infectious renal lesion. It may appear as a tumor or diffusely. CT is the best imaging modality. Key Points – Imaging of UT infections is controversial. – The aim is to detect groups at risk, renal involvement, and risks for recurrence. – US Doppler, DMSA scintigraphy, and MCU form the basis of the imaging work-up.

Complications of Acute Pyelonephritis Renal abscess is the typical complication of delayed or non-adapted treatment. Lesions are demonstrated by US, although further, cross-sectional imaging (CT scan or MR) is occasionally useful. Pyonephrosis corresponds to

DMSA

Normal

Abnormal

APN

Abnormal

Introduction Urolithiasis is less common in children than in adults (1 child for every 20,000 adults). It is more frequent in certain geographic areas. Hematuria and pain are the presenting symptoms in about half of the patients whereas the disease is asymptomatic and an incidental finding in 20% of patients. A familial history is frequent and accounts for over 50% of patients. Metabolic disorders and UT infections are common etiologies [24, 25]. The work-up of a patient with urolithiasis includes urinalysis, blood tests, and genetic analysis, looking for favoring diseases or infections. Imaging must be applied systematically, starting with US and using a tailored decisionmaking process for additional imaging modalities [25].

US + Doppler

Normal

Imaging Urolithiasis in Children

VCUG

Stop Fig. 2. Proposed imaging workflow for a child with a urinary tract infection. US, Ultrasound; APN, Acute Pyelonephritis; DMSA, Tc-99m(V) Dimercaptosuccinic acid; VCUG, Voiding Cystourethrography

Imaging Ultrasound, with its entire spectrum of optimized techniques, is the ideal primary imaging modality in children (Fig. 3). It allows reliable demonstration of the kidneys and urinary tract and is useful in diagnosis as well as in post-therapeutic follow-up. On US, urolithiasis manifests as either nephrocalcinosis or lithiasis. In nephrocalcinosis, echogenic deposits in the pyramids lead to a reversed CMD. Nephrocalcinosis should be differentiated from other causes of echogenic pyramids. Lithiasis is easily demonstrated within the dilated renal upper cavities and at the ureterovesical junction. It has the same appearance as in adults.

181

Imaging Uronephropathies in Children

US (gray scale + DDS/CDS)

negative + low clinical suspicion

positive

negative + high clinical suspicion

US inconclusive or non-diagnostic mismatch US & clinics

only secondary signs + no stone visible, or before intervention, if therapeutically necessary

Fig. 3. Imaging algorithm for infants and children with suspected urolithiasis [7]. CDS, Color Doppler Sonography (power Doppler); ce-VUS, contrast-enhanced Voiding Urosonography; (uro-)CT, (urinary tract) CT; DD, Differential Diagnosis; DDS, Duplex Doppler Sonography; KUB, Kidney-UreterBladder film; IVU, Intravenous Urography; MRU, Magnetic Resonance Urography; US, Ultrasound

Stop follow-up after/during treatment potentially + KUB for confirmation, for treatment needs …

uro-CT x1 (ultra-)low dose + unenhanced x2 + KUB (± IVU) before intervention / lithotripsy

x1

or KUB (+ adapted IVU, particularly if low-dose CT unavailable)

x2

potentially contrast-enhanced CT-urography, if other DD or complication; MRU in selected cases

Some kidney diseases have a specific, characteristic pattern, such as primary hyperoxaluria type 1. A plain film of the abdomen may be necessary for the proper demonstration of the stone prior to treatment. Tailored intravenous urography is still used by some clinicians to demonstrate the morphology of the UT, especially if CT is not available. Furthermore, CT is much less frequently used than in adults. It serves as a complementary tool in case of a non-diagnostic US examination or prior to treatment. The radiation dose must be minimized and optimally adapted to the child’s size [5, 26].

Treatment The treatment of urolithiasis should aim to avoid or correct conditions that have led to the disease. If the lithiasis cannot be medically eliminated, more interventional treatments will be needed. For example, extracorporeal lithotripsy, the primary treatment approach in adults, has been adapted to children as well. However, in selected cases surgery will be unavoidable. Key Points – Urolithiasis is less frequent in children. – An etiology is more often detected than in adults. – US is the main imaging modality.

References 1. Brenner DJ, Elliston CD, Hall EJ, Berdon WE (2001) Estimated risks of radiation-induced fatal cancer from pediatric CT. AJR Am J Roentgenol 176:289-296 2. Ripolles T, Puig J (2009) Actualización del uso de contrastes en ecografía. Revisión de las guias clinicas de la Federación Europea de Ecografía, Radiología 51:362-375

3. Grattan-Smith JD, Jones RA (2008) MR urography: technique and results for the evaluation of urinary obstruction in the pediatric population, Magn Reson Imaging Clin N Am 16: 643-660 4. Thomsen HS (2007) ESUR guideline: gadolinium-based contrast media and nephrogenic systemic fibrosis Eur Radiol 17:692-2696 5. Passerotti C, Chow JS, Silva A et al (2009) Ultrasound versus computed tomography for evaluating urolithaisis. J Urol 182:1829-1841 6. Riccabona M, Avni FE, Blickman JG et al (2008) Imaging recommandations in pediatric uroradiology (Part 1). Pediatr Radiol 38:138-145 7. Riccabona M, Avni FE, Blickman JG et al (2009) Imaging recommendations in pediatric uroradiology (Part 2). Pediatr Radiol 39:891-898 8. De Bruyn R, Marks SD (2008) Post-natal investigations of fetal renal disease. Semin Fetal Neonatal Med 13:133-141 9. Carrico C, Lebowitz RL (1998) Incontinence due to an infrasphincteric ectopic ureter: why the delay in diagnosis and what the radiologist can do about it, Pediatric Radiology 28:942-949 10. Gill RD (1952) Triplication of the ureter and renal pelvis. J Urol 140:147 11. Wyly JB, Lebowitz RL (1984) Refluxing urethral ectopic ureters: recognition by the cyclic voidng cystourethrogram. AJR 142:1263-1267 12. Prewitt LH, Lebowitz RL (1976) The single ectopic ureter. AJR 127:941-948 13. Rooks VJ, Lebowitz RL (2001) Extrinsic ureteropelvic junction obstruction from a crossing renal vessel: demography and imaging. Pediatr Radiol 31:120-124 14. Shukla AR, Cooper J, Patel RP et al (2005) Prenatally detected primary meagureter: a role for extended follow-up. J Urol 173:1353-1356 15. Chaudry G, Perez-Ataude AR, Ngan BY et al (2009) Imaging of congenital mesoblastic nephroma wath pathological correlation. Pediatr Radiol 39:1080-1086 16. Eftekhari F, Erly WK, Jaffe N (1990) Malignant rhabdoid tumor of the kidney: imaging features in two cases. Pediatric Radiol 21:39-42 17. Boggs LK, Kimmelstiel P (1956) Benign multilocular cystic nephroma: report of two cases of so-called multilocular cyst of the kidney. J Urol 76:530-541

182

18. Estrada CR, Sthar AM, Eaton SH et al (2005) Renal cell carcinoma: Childrens Hospital Boston experience. Urology 66: 1296-1300 19. Wedekin M, Ehrich JH, Offner G, Pape L (2008) Aetiology and outcomes of acute and chronic renal failure in infants. Nephrol Dial Transplant 23:1575-1580 20. Ardissimo G, Dacco V, Testa S et al (2003) Epidemiology of CRF in children: data of the Ital Kid Project. Pediatrics 111:382-387 21. Mercado-Duane MG, Benson JE, John SD (2002) US of renal insufficiency in neonates. RadioGraphics 22:1429-1438 22. Lim R (2009) Vesico-ureteral reflux and UTI: evolving practices and current controversies in pediatric imaging. AJR Am J Roentgen 192:1197-1208

Jeanne S. Chow, Fred E. Avni

23. Lee M, Lin C, Huang F et al (2009) Screening young children with a first febrile UTI for high grade VUR with renal US scanning and Technecium-99m-labelled DMSA. J Pediatr 154:797-802 24. Ajdinovic B, Jaukovic L, Krstic Z, Dopuda M (2008) Impact of MCU and DMSA renal scintigraphy on the investigation schema in children with UTI. Ann Nucl Med 22:661-665 25. Hoppe B, Kemper MJ (2008) Diagnostic examination of the child with urolithiasis or nephrocalcinosis. Pediatr Nephrol DOI:10.1007/s00467-008-1073-x 26. Karmazyn B, Frush DP, Applegate KE et al (2009) CT with computed simulated dose reduction technique for detection of pediatric nephrourolithiasis. AJR 192:143-149

IDKD 2010-2013

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging Gerald Antoch Department of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, University at Duisburg-Essen, Essen, Germany

Introduction Malignant tumors are the second most common cause of death in the western world [1]. Based on the assumption that patients’ prognoses can be improved by the adoption of stage-adapted therapy, accurate clinical and radiological tumor staging must be considered as essential when assessing primary tumors and recurrent disease. In addition, different imaging procedures are used for therapy response assessment in patients undergoing treatment for malignant disease. For both, tumor staging and therapy response assessment, morphological and functional imaging procedures are available; however, these have well-known limitations in their diagnostic accuracy. Computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound (US) provide mainly morphological information on the tumor and its potential metastases, but the lack of functional data has been shown to hamper accurate assessment of local lymph node involvement [2, 3]. Therapy response assessment is mainly based on lesion size, which nonetheless has been shown to be an insensitive indicator of response, at least in the initial phase of tumor treatment. The functional data provided by positron emission tomography (PET) are known to be more sensitive and specific than morphological data and to complement the latter in the assessment of local lymph node involvement by malignant tumors [4-6]. Similarly, when PET and PET/CT are used to assess therapy response, they have been found to offer earlier and more reliable response assessment than achieved based on morphology alone. This review summarizes the applications of PET and PET/CT imaging in different genitourinary tumors. The indications and diagnostic accuracies of integrated PET/CT are discussed and the different radionuclides used in this type of imaging are addressed.

Prostate Cancer Prostate cancer is the most common malignancy in men [7]. The disease is characterized by a widely variable biological behavior which ranges from clinically silent tu-

mors to highly aggressive tumors with rapid growth and distant metastases. Although different radionuclides have been proposed, none has evolved as “the” radionuclide for the imaging of prostate cancer. The uptake of one such agent, fluorodeoxyglucose (FDG), was shown in different studies to be increased in poorly differentiated prostate cancers associated with high Gleason scores or high prostate-specific antigen (PSA) levels. However, well-differentiated prostate cancers frequently do not demonstrate an increase in glucolysis, with negative FDG-PET scans as a consequence. Therefore, a positive FDG-PET scan may add relevant information to morphology alone, when assessing the pelvis for potential metastases, but a negative FDG-PET scan has to be interpreted with caution. In addition, even in FDG-PET-positive tumors, urinary excretion of the radionuclide may limit the diagnostic accuracy with regard to assessment of the primary tumor and its potential infiltration into adjacent organs, such as the seminal vesicles. Based on its high soft-tissue contrast MRI has become the modality of choice for imaging the primary tumor. Its use is further supported by the fact that poorly differentiated prostate cancers, benign hyperplasia, and prostatitis show substantial overlap in FDG uptake. The limitations of FDG in imaging prostate cancer have led to the introduction of alternative radionuclides for prostate imaging, shifting the focus away from glucolysis to cell-membrane turnover. Accordingly, choline has become the most commonly used radionuclide in prostate imaging. Choline is involved in transmembrane signaling and phospholipid synthesis in cell membranes. Different malignant tumors show substantial increases in choline kinase activity, resulting in an increase in membrane phospholipids. Thus, tracers such as 11C-choline or 18F-choline can estimate the proliferative potential of a tumor by detecting membrane lipid synthesis [8]. An advantage of 18F-choline over 11C-choline is its longer half-life and better image quality due to an increase in spatial resolution, which is a consequence of the shorter positron range of 18F. However, the urinary excretion of 18F-choline is higher than that of 11C-choline, which has been attributed to incomplete tubular reabsorption of the former. Regardless, compared with FDG, both tracers

184

Gerald Antoch

have the advantage of reduced urinary excretion with less interpretive issues caused by radioactive urine. There have been reports on the use of 11C-choline in primary tumor assessment and in localization of the primary tumor within the prostate in patients with negative or unclear findings on biopsy. Some studies have even reported a higher accuracy for lesion localization than MRI using an

a

endorectal coil [9]. However, its high soft-tissue contrast and the availability of MR-spectroscopy favor MRI as the imaging method of choice for lesion localization within the prostate gland. The main indication for 11C-choline or 18F-choline is detection of tumor recurrence in patients with rising PSA levels and negative findings on CT or MRI after treatment of prostate cancer (Figs. 1, 2).

b

c

Fig. 1 a-c. CT, 18F-choline-PET, and 18F-choline-PET/CT of a retroperitoneal lymph node in a patient with an increase in the PSA after radical prostatectomy. Unremarkable lymph node on CT (a) diagnosed as metastasis on PET (b) and PET/CT (c) based on increased choline uptake

a

b

c

Fig. 2 a-c. a CT, b 18F-choline-PET, and c 18Fcholine-PET/CT of a patient with prostate cancer and a bone metastasis in the sacrum

185

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging

In selected patients undergoing diagnosis of prostate cancer, both radionuclides may be used for initial tumor staging. For tumor recurrence, the detection rate of 11C-choline is reportedly as high as 50% in patients with negative findings on morphological imaging procedures [10]. However, many authors have reported that 11C-choline detection rates depend upon the PSA level of the patient, with rising PSA levels paralleling an increase in tumor detection. Most institutions schedule PET at PSA levels of at least >1 ng/mL. Other radionuclides include 11C-acetate, 18F-tes tosterone, and 11C-methionine. The mechanism of 11C-acetate uptake by malignant cells is lipid synthesis. Thus, acetate accumulation has been attributed to an increase in fatty acid synthesis in prostate cancer cells. Like 11C-choline, acetate is not a tumor-specific radionuclide, since uptake is increased also in benign hyperplasia and prostatitis. As with 11C-choline, 11C-acetate has been found to be of higher diagnostic accuracy than FDG in the detection of prostate cancer foci [11]. 18F-dihydrotestosterone (18FHDT) and 11C-methionine are further tracers that have been used occasionally in imaging prostate cancer. 18FHDT has affinity for the androgen receptor and can be used to test for resistance to androgen ablation therapy. 11C-methionine is a nonspecific radionuclide for amino acid transport and protein synthesis; it has been used in prostate cancer imaging only infrequently.

Testicular Cancer In testicular cancer, FDG-PET and FDG-PET/CT can be used for initial tumor staging, treatment monitoring, and the detection of recurrent disease. The advantage of these modalities in tumor staging has been demonstrated in patients in whom CT alone indicated stage II disease (lymph node metastases) based on the detection of morphologically enlarged lymph nodes. Integration of FDG-PET into the staging algorithm may prevent falsepositive CT findings in this setting by demonstrating PET-negativity within such lymph nodes [12]. Nonetheless, upstaging of not enlarged, CT-negative lymph nodes based on FDG avidity within these nodes has been discussed controversially. FDG-PET and PET/CT have the potential to detect sub-centimeter metastases in non-enlarged nodes, but the limited spatial resolution of PET leads to a decrease in sensitivity with decreasing lymph node size (Fig. 3). In addition, in some histopathological types of germ cell tumors, such as mature teratoma, glycolysis is not increased. Accordingly, FDG-PET and PET/CT are currently not recommended for routine use in the staging of patients with testicular cancer [13, 14]. While CT is recommended for treatment monitoring in germ cell tumors, the addition of FDG-PET or PET/CT may be helpful in patients with residual lymph

a

b

c

Fig. 3 a-c. Retroperitoneal lymph node metastasis in a patient with seminoma. Unclear finding on CT (a) but increased FDG-uptake on PET (b) and PET/CT (c) indicates malignancy

186

node enlargement after therapy. In a prospective trial involving a series of patients with seminoma, FDG-PET was used to differentiate viable tumor tissue from scar/fibrosis in patients with residual masses >1 cm after chemotherapy [15]. There were no false-positives reported. All cases of residual disease seen on CT as masses >3 cm and 95% of those cases in which the residual masses were <3 cm were detected correctly with FDGPET. Based on these results, FDG-PET has been recommended in seminoma patients to assess residual masses seen on CT for potential residual disease. It must be emphasized, however, that other authors [16] have reported a substantial number of false-positive findings due to inflammation after chemotherapy of seminoma. In patients undergoing radiotherapy, the issue of inflammatory changes or tissue regeneration may take on additional importance. FDG-PET and PET/CT are currently not

Gerald Antoch

recommended for therapy response assessment in patients with non-seminomatous tumors, as mature differentiated teratoma typically has no or only mild FDG uptake and thus may not be reliably differentiated from scar tissue or fibrosis [17]. Figure 4 shows the currently recommended algorithm for the diagnosis and treatment of testicular germ cell tumors. FDG-PET can be used in patients with suspected recurrent disease as determined by tumor markers but with negative CT. In this population, FDG-PET or PET/CT can detect otherwise occult recurrent tumor sites, as discussed by Hain et al. [18]. In that study, although FDGPET was able to identify tumor recurrence more reliably than CT, FDG-PET was falsely negative in some of the patients later diagnosed with recurrence. Therefore, close imaging follow-up must be recommended in patients with clinical signs of recurrence but negative FDG-PET.

Fig. 4. Algorithm of diagnosis and treatment of testicular germ cell tumors modified according to the current recommendations. (From [13])

187

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging

Ovarian Cancer Ovarian cancer is the second most common gynecological malignancy and is frequently detected at advanced tumor stages, resulting in a poor patient prognosis. Even those patients diagnosed in an operable stage have a high risk of tumor recurrence postoperatively, requiring close patient follow-up, including the measurement of tumor markers (mainly 125Ca) and evaluation using different imaging procedures. While FDG-PET and PET/CT have been used to diagnose the primary tumor and in the differential diagnosis of ovarian cancer from other ovarian lesions, both imaging modalities are limited in menstruating women by false-positives due to follicular cysts, cystadenomas, schwannomas, endometriomas, or inflammation. In addition, low-grade malignancies and borderline tumors may be FDG-PET-negative, resulting in false-negative findings. Thus, the reported sensitivities and specificities of up to 87% [19] must be interpreted with caution. In post-menopausal women, however, the risk of false-positive findings is lower and any FDG uptake must initiate further work-up to exclude tumor growth. CT has been the standard of care in staging patients with ovarian cancer pre-operatively, whereas there are only limited data on a potential benefit of FDGPET/CT in this clinical setting and no recommendation can currently be made. The main focus of hybrid imag-

ing in ovarian cancer has been the diagnosis of recurrent ovarian cancer. CT has been the imaging modality of choice to localize tumor recurrence in patients with rising tumor markers, but may be supported with functional data in patients with equivocal findings on CT alone (Fig. 5). Picchio et al. [20] reported an increase in the sensitivity for detection of ovarian cancer recurrence from 70 to 83% and an increase in the specificity from 83 to 92% when FDG-PET was added to CT. Similar sensitivities and specificities were reported by Bristow et al. [21]. Based on the available literature, the role of FDG-PET and PET/CT in patients with clinically suspected ovarian cancer recurrence is currently limited to those patients without tumor detection on CT.

Cervical Cancer Cervical cancer is the third most common malignancy in women [7]. These tumors manifest as an aggressive local growth, and the infiltration of adjacent organs is not infrequent at the time of diagnosis. Metastases of cervical cancer are primarily found in locoregional lymph nodes while hematogenous metastases to the lungs, liver, or bone will be found in rather later stages of the disease. The role of FDG-PET or FDG-PET/CT in cervical cancer has yet to be defined clearly. Although both techniques have been used to identify the primary tumor

a

b

c

Fig. 5 a-c. a CT detected ascites, an indirect sign of peritoneal carcinomatosis, in this patient with rising tumor markers after therapy for ovarian cancer. b FDG-PET and c PET/CT show increased FDG uptake inter-intestinally, confirmed as peritoneal carcinomatosis on further follow-up

188

a

Gerald Antoch

b

c

Fig. 6 a-c. Detection of cervical cancer in a patient with an equivocal CT (a) with FDG-PET (b) and FDG-PET/CT (c)

(Fig. 6), diagnosis of the primary is most frequently achieved by Papanicolaou (PAP) test. It is generally accepted that MRI is the imaging procedure of choice when assessing a potential invasion of the primary tumor into adjacent organs. Thus, FDG-PET/CT has not been used frequently to assess primary cervical tumors. Tumor staging, however, has been considered an indication for FDG-PET and PET/CT in cervical cancer. The reported sensitivities and specificities vary widely. Sironi et al. [22] obtained promising results: a sensitivity of 72% in lymph nodes <5 mm and 100% in those >5 mm, both coupled with high specificities. Nevertheless, there are more data suggesting a rather low sensitivity for detection of lymph node metastases. For FDG-PET, Wright et al. [23] reported a sensitivity of 53% for pelvic nodes and an even lower sensitivity for retroperitoneal/paraaortic lymph nodes. Even though the addition of morphological data with PET/CT may be considered helpful over PET alone, Choi et al. [24] reported similar sensitivities for FDG-PET/CT in local lymph nodes (57%). Therefore, the role of FDG-PET and FDG-PET/CT in the staging of cervical cancer has not been established. Tumor recurrence, by contrast, can be detected reliably with either FDG-PET or PET/CT. Chung et al. [25] reported a sensitivity of 90% and a specificity of 81% for the detection of cervical cancer recurrence. Following PET/CT, patient management was altered from that chosen based on a conventional work-up in 24% of patients. In conclusion, currently available data support the limited use of FDG-PET and PET/CT in cervical cancer, mainly in the detection of tumor recurrence.

Endometrial Cancer The diagnosis of endometrial cancer is established from biopsy. Even though these tumors are typically FDG-avid, the role of FDG-PET and PET/CT in assessment of the primary tumor is limited due to false-positives caused by normal endometrial FDG uptake, which varies according to the endometrial cycle. If potential local invasion of the tumor into adjacent structures needs to be assessed preoperatively, this is typically done with MRI, because of its superior soft-tissue contrast. Despite reports on the use of FDG-PET and PET/CT for staging local lymph nodes and distant organs [26], CT currently remains the imaging procedure of choice for N-staging and M-staging. Further data are required to define the role of hybrid imaging procedures in endometrial cancer.

References 1. Jemal A, Thomas A, Murray T, Thun M (2002) Cancer statistics 2002. CA Cancer J Clin 52:23-47 2. Haberkorn U, Schoenberg SO (2001) Imaging of lung cancer with CT MRI and PET. Lung Cancer 34 Suppl 3:S13-S23 3. Toloza EM, Harpole L, McCrory DC (2003) Noninvasive staging of non-small cell lung cancer: a review of the current evidence. Chest 123(Suppl 1):S137-S146 4. Adams S, Baum RP, Stuchensen T et al (1998) Prospective comparison of 18F-FDG PET with conventional imaging modalities (CT MRI US) in lymph node staging of head and neck cancer. Eur J Nucl Med 25:1255-1260 5. Marom EM, McAdams HP, Erasmus JE et al (1999) Staging non-small cell lung cancer with whole-body PET. Radiology 212:803-809

Integrated Imaging in Genitourinary Oncology: PET/CT Imaging

6. van Tinteren H, Hoekstra O, Smit E et al (2002) Effectiveness of positron emission tomography in the preoperative assessment of patients with suspected non-small-cell lung cancer: the PLUS multicentre randomised trial. Lancet 359:1388-1393 7. Jemal A, Murray T, Samuels A et al (2003) Cancer statistics 2003. CA Cancer J Clin 53:5-26 8. Sutinen E, Nurmi M, Roivainen A et al (2004) Kinetics of [(11)C]choline uptake in prostate cancer: a PET study. Eur J Nucl Med Mol Imaging 31:317-324 9. Yamaguchi T, Lee J, Uemura H et al (2005) Prostate cancer: a comparative study of 11C-choline PET and MR imaging combined with proton MR spectroscopy. Eur J Nucl Med Mol Imaging 32:742-748 10. de Jong IJ, Pruim J, Elsinga PH et al (2003) Preoperative staging of pelvic lymph nodes in prostate cancer by 11C-choline PET. J Nucl Med 44:331-335 11. Fricke E, Machtens S, Hofmann M, Van Den J (2003) Positron emission tomography with 11C-acetate and 18F-FDG in prostate cancer patients. Eur J Nucl Med Mol Imaging 30:607-611 12. Albers P, Bender H, Yilmaz H, Schoeneich G (1999) Positron emission tomography in the clinical staging of patients with Stage I and II testicular germ cell tumors. Urology 53:808-811 13. Krege S, Beyer J, Souchon R et al (2008) European consensus conference on diagnosis and treatment of germ cell cancer: a report of the second meeting of the European Germ Cell Cancer Consensus group (EGCCCG): part I. Eur Urol 53:478-496 14. Krege S, Beyer J, Souchon R P et al (2008) European consensus conference on diagnosis and treatment of germ cell cancer: a report of the second meeting of the European Germ Cell Cancer Consensus Group (EGCCCG): part II. Eur Urol 53:497-513 15. De Santis M, Becherer A, Bokemeyer C et al (2004) 2-18fluoro-deoxy-D-glucose positron emission tomography is a reliable predictor for viable tumor in postchemotherapy seminoma: an update of the prospective multicentric SEMPET trial. J Clin Oncol 22:1034-1039 16. Lewis DA, Tann M, Kesler K et al (2006) Positron emission tomography scans in postchemotherapy seminoma patients with residual masses: a retrospective review from Indiana University Hospital. J Clin Oncol 24:e54-e55

189

17. Stephens AW, Gonin R, Hutchins GD et al (1996) Positron emission tomography evaluation of residual radiographic abnormalities in postchemotherapy germ cell tumor patients. J Clin Oncol 14:1637-1641 18. Hain SF, O’Doherty MJ, Timothy AR et al (2000) Fluorodeoxyglucose positron emission tomography in the evaluation of germ cell tumours at relapse. Br J Cancer 83:863-869 19. Kawahara K, Yoshida Y, Kurokawa T et al (2004) Evaluation of positron emission tomography with tracer 18-fluorodeoxyglucose in addition to magnetic resonance imaging in the diagnosis of ovarian cancer in selected women after ultrasonography. J Comput Assist Tomogr 28:505-516 20. Picchio M, Sironi S, Messa C et al (2003) Advanced ovarian carcinoma: usefulness of [(18)F]FDG-PET in combination with CT for lesion detection after primary treatment. Q J Nucl Med 47:77-84 21. Bristow RE, del Carmen MG, Pannu HK et al (2003) Clinically occult recurrent ovarian cancer: patient selection for secondary cytoreductive surgery using combined PET/CT. Gynecol Oncol 90:519-528 22. Sironi S, Buda A, Picchio M et al (2006) Lymph node metastasis in patients with clinical early-stage cervical cancer: detection with integrated FDG PET/CT. Radiology 238:272-279 23. Wright JD, Dehdashti F, Herzog T et al (2005) Preoperative lymph node staging of early-stage cervical carcinoma by [18F]-fluoro-2-deoxy-D-glucose-positron emission tomography. Cancer 104:2484-2491 24. Choi HJ, Roh JW, Seo SS et al (2006) Comparison of the accuracy of magnetic resonance imaging and positron emission tomography/computed tomography in the presurgical detection of lymph node metastases in patients with uterine cervical carcinoma: a prospective study. Cancer 106:914-922 25. Chung HH, Jo H, Kang WJ et al (2007) Clinical impact of integrated PET/CT on the management of suspected cervical cancer recurrence. Gynecol Oncol 104:529-534 26. Horowitz NS, Dehdashti F, Herzog TJ et al (2004) Prospective evaluation of FDG-PET for detecting pelvic and para-aortic lymph node metastasis in uterine corpus cancer. Gynecol Oncol 95:546-551

IDKD 2010-2013

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging Thomas F. Hany Department of Radiology, Clinic and Policlinic of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerland

Introduction The basic principle of positron emission tomography (PET) is the use of positron-emitting-isotope-labeled pharmaceuticals that are integrated into a metabolic pathway. Positron-emitting isotopes are characterized by a beta plus-decay, in which a positron is emitted. This positron collides with and annihilates any of the many shell electrons in the neighboring atoms, thereby producing two 511-keV gamma rays (photons) which are detected in coincidence by the PET scanner. The additional integration of a computed tomography scanner (PET/CT) allows the acquisition of PET and CT images of the patient in the same imaging session. The clinically and most widely evaluated positron-emitting isotope labeled pharmaceutical is fluorine-18 fluoro-2-deoxy-D-glucose (FDG). This glucose analogue is transported into the cell by specific transporters and phosphorylated by hexokinase to FDG-6-phosphate. As FDG-6-phosphate is inert to further metabolic processing or to transmembrane back-transport outside the cell, it accumulates within the cells. The physical half-life of FDG is around 110 min. The compound is used as a metabolic marker in oncology, cardiology, neurology and inflammation imaging. One of the most important advantages of this imaging technique is the extensive anatomical coverage. Thus, images from the head to the thighs are acquired routinely, allowing evaluation of the entire body. All of the currently available data indicate that PET/CT is more sensitive and specific than either of its constituent imaging methods. With PET/CT, the most relevant additional effect is that the CT data frequently add specificity to the FDGPET data [1].

Malignant Primary Liver Tumors Hepatocellular Carcinoma Hepatocellular carcinoma (HCC) is frequently multifocal such that at the time of exploration multiple sites in both liver lobes are involved. Pre-operative assessment should include the search for extrahepatic metastases, because

their presence will preclude a curative approach. Interestingly, extrahepatic metastasis of HCC occurs relatively late in the clinical course and therefore does not significantly shorten the survival time. Whenever the stage of disease allows, surgical treatment should be pursued. Torizuka et al. evaluated patients with known HCC by using dynamic and static FDG-PET data acquisition [2]. High degrees of correlation were found between the histological grade and the kinetic rate constants as well as hexokinase activity. There have been only a few studies involving PET/CT imaging. The largest of these, by Park et al., consisted of a prospective evaluation of FDG and 11C-acetate PET/CT in the detection of primary and metastatic hepatocellular carcinoma [3]. In an analysis based on biopsied lesions, the sensitivity for primary HCC of FDG-PET/CT was 64.4% while that of 11C-acetate PET/CT was 84.4%. The overall sensitivities of FDG, 11C-acetate, and dual-tracer PET/CT for 35 metastatic HCCs were 85.7, 77.0, and 85.7%, respectively. The addition of 11C-acetate to FDG-PET/CT increased the overall sensitivity for the detection of primary HCC but not for the detection of extrahepatic metastases. FDG, 11C-acetate, and dual-tracer PET/CT had a low sensitivity for the detection of small primary HCC, but FDG-PET/CT had a relatively high sensitivity for the detection of extrahepatic metastases of HCC. Thus, only in selected cases, in which there is known moderately to poorly differentiated HCC, is FDG-PET/CT useful in a pre-therapy setting to detect additional intra- or extrahepatic lesions. The same holds true for the evaluation of recurrent disease after therapy.

Cholangiocarcinoma There is limited knowledge on the use of FDG-PET or PET/CT in the evaluation of intrahepatic cholangiocarcinoma. This type of hepatic cancer is associated with primary sclerosing cholangitis. In the study by Kluge et al., both the sensitivity and the specificity of FDG-PET and PET/CT in the detection of primary lesions were >90% [4]; however, FDG-PET showed a poor performance in the detection of locoregional metastases. In the study by Fritscher-Ravens et al., there were several false-negatives, especially in patients with mucinous adenocarcinoma.

191

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging

Otherwise, the same results regarding locoregional and distant metastases were achieved with FDG-PET and PET/CT [5]. In the study by Petrowsky et al. [6], 61 patients with malignancies of the biliary tract, proven by histology or cytology, were evaluated by FDG-PET/CT and contrast-enhanced CT. PET/CT detected all gallbladder cancers (n = 14). Overall, 45 of 61 tumors were correctly identified with PET/CT (sensitivity 74%) and 40 by contrast-enhanced CT scan (sensitivity 66%). All 12 distant metastases were detected by PET/CT, but only 3 out of the 12 by CT (p <0.001) (Fig. 1). The detection of extrahepatic cholangiocarcinoma was significantly lower than that of intrahepatic cholangiocarcinoma for both techniques; also, regional lymph node metastases were detected by PET/CT in only 12% of the cases and by CT in only 24%. PET/CT findings resulted in a change of management in 17% of patients with tumors deemed resectable after the standard workup. Therefore, PET/CT seems particularly valuable in detecting unsuspected distant metastases, which are not diagnosed by standard imaging.

a

Fig. 1 a-g. A 56-year-old male patient with a histologically proven intrahepatic cholangiocarcinoma. a In the maximum intensity projection image of the FDG-PET/CT, the primary tumor is well demarcated, with multiple foci of pathological uptake. In the axial images at the level of the liver (from top to bottom: b PET contrastenhanced, c CT, d fused image), the intrahepatic tumor is highly FDG avid. In the mid-thoracic region (from top to bottom: e PET, f contrast-enhanced CT, g fused image), a mediastinal lymph node metastasis is easily demarcated paraesophageally

Pancreas Exocrine Pancreatic Carcinoma The staging of pancreatic cancer includes a histological workup and proof of the typical histological features of exocrine pancreatic cancers, mostly adenocarcinoma. However, T-staging essentially is only possible by evaluation of the arterial vasculature, since tumors that show invasion/encasement of the celiac trunk and superior mesenteric artery are unresectable (T4). The most important information from any imaging modality is, therefore, arterial vascular involvement of the primary tumor, the presence of peritoneal carcinomatosis, and the detection of distant metastases. Regarding the performance of PET/CT, the data are thus far rather sparse. In a study by Farma et al., 82 patients with assumed resectable pancreatic cancer underwent staging with non-contrast-enhanced PET/CT and contrast-enhanced CT of the chest and abdomen [7]. The sensitivity and specificity of PET/CT in

b

e

c

f

d

g

192

diagnosing pancreatic itself cancer were 89 and 88%, respectively. The sensitivity of detecting metastases was 61% for PET/CT, 57% for contrast-enhanced CT, and 87% for the two side by side. PET/CT findings influenced the clinical management of seven patients (11%), all of whom had distant metastases. As a major drawback of this study, the CT component of PET/CT was not performed using contrast-enhanced triple-phase CT. In a study by Strobel et al., 50 patients with biopsy-proven pancreatic cancer (adenocarcinoma) were evaluated by integrated triple-phase contrast-enhanced PET/CT regarding respectability and overall staging [8]. The criteria for irresectability were distant metastases, peritoneal carcinomatosis, arterial infiltration, or infiltration of neighboring organs other than the duodenum. Histology, intraoperative findings, and follow-up CT together with the clinical findings were used as the standard of reference. Accordingly, 27 patients had unresectable disease because of distant metastases (n = 17), peritoneal carcinomatosis (n = 5), or local infiltration (n = 5). In the assessment of resectability, PET alone had a sensitivity of 100%, a specificity of 44%, an accuracy of 70%, a positive predictive value of 61%, and a negative predictive value of 100%; unenhanced PET/CT had respective values of 100, 56, 76, 66, and 100%; the corresponding values for enhanced PET/CT were 96, 82, 88, 82, and 96%. In five patients, unresectability was missed by all imaging methods and was only diagnosed intraoperatively. Enhanced PET/CT was significantly superior to PET alone (p = 0.035), with a trend for the superiority of enhanced over unenhanced PET/CT (p = 0.070). Based on all of these data, it is clear that non-contrast-enhanced PET/CT is able to detect distant metastases but not local extent in a large proportion of cases. Furthermore, in a rather considerable proportion of patients, only intraoperative findings are able to reveal surgery-precluding factors, such as deep retroperitoneal infiltration, small liver metastases, and peritoneal involvement. It seems that the most favorable approach consists of the use of intravenous contrast-enhanced CT in the PET/CT protocol; however, this does not reflect clinical practice, since most patients with a suspected pancreatic lesion rapidly undergo contrast-enhanced CT or magnetic resonance imaging (MRI) evaluation.

Endocrine Pancreatic Cancer Neuroendocrine tumors (NETs) of the pancreas account for <5% of all malignant pancreatic tumors. PET imaging using FDG has limited value, since these tumors are often slow-growing, with an accordingly low metabolism [9]. In addition to morphological imaging modalities, including contrast-enhanced CT and MRI, somatostatin receptor scintigraphy (SRS) is used to localize and characterize NETs in general and such tumors in the pancreas in particular. Nonetheless, the interpretation of SRS findings can be challenging due to the difficulty of distinguishing tumor from intestinal structures and to the variable density of somatostatin receptors on the different tumors. Over-

Thomas F. Hany

all, a significant rate of false-negative results – with sensitivities ranging between 50 and 78% – in the detection of NET (depending on the localization) has been reported [10]. PET using the catecholamine precursor 6-(fluoride18)-fluoro-dopa (18F-DOPA) has been proposed as a valuable imaging option for NET [11]. This tracer highlights the tumor’s intracellular decarboxylase activity and in the context of PET imaging provides higher spatial resolution than obtained with SRS. A major drawback of 18F-DOPA is its physiological uptake by normal pancreas, which in certain cases obscures detection of the primary tumor. A rather newly developed PET tracer uses the basic principle of SRS, i.e., somatostatin labeling, but with 68Ga. In a study by Ambrosini et al., 18F-DOPA and 68Ga were compared in a rather small group of 13 patients [12]. 68GaDOTA-NOC (tetra-azycyclododecanetetra-acetic acid-[1Nal3]-octreotide) was found to be an accurate tracer for the assessment of NETs and was better than 18F-DOPA in the detection of primary NETs – especially those of the pancreas – and metastases. The authors concluded that since the pancreas is the most frequent site of NETs, the routine use of 68Ga-DOTA-NOC is more appropriate. Here, larger, comparative studies that include morphological imaging modalities are needed to determine the role of PET/CT in NETs of the pancreas.

Colorectal Carcinoma In the western world, colorectal carcinoma is the most important cause of death due to cancer, after bronchial carcinoma [13]. About 70% of patients have curable resectable tumor at initial diagnosis and are treated with curative intent. Approximately 50% of colon cancer patients will present with hepatic metastases, either at the time of initial diagnosis or as a result of recurrence [14]. From a diagnostic perspective, colon cancer and rectal cancer are often evaluated as a single group; however, especially deep rectal cancer has a clearly different pathway of locoregional and distant metastases (Fig. 2).

Initial Staging Two studies using FDG-PET alone for initial staging demonstrated the high sensitivity of this modality in the detection of both primary tumor (100 and 96%) and distant metastases (87 and 78%) and confirmed its low sensitivity (29 and 29%) in lymph node staging [15, 16]. In a study by Veit-Haibach et al., 47 patients underwent wholebody PET/CT colonography one day after colonoscopy [17]. Compared with optimized abdominal CT staging alone, PET/CT colonography was significantly more accurate in defining TNM stage (difference: 22%; 95% CI: 9-36%; p = 0.003), which was mainly based on a more accurate definition of the T-stage. Differences were not detected for defining N-stage between PET/CT colonography and CT alone with a threshold of 0.7 cm for malignant nodes but were detected with a threshold of 1 cm. There were no

193

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging

a

b

c

d

Fig. 2 a-d. A 63-year-old female patient with a histologically proven deep rectal cancer. Axial FDG-PET/CT images at the level of the primary tumor (a PET, b contrast-enhanced CT, c fused images) reveal the FDG-avid primary tumor as well as enlarged and FDGavid bilateral inguinal lymph node metastases. The lymph node spread is typical for deep rectal cancer invading the anal canal. d The same findings are seen in the contrastenhanced, T1-fat suppressed MRI images

differences in defining M-stage separately or when the accuracy of PET/CT colonography was compared to that of CT + PET. PET/CT colonography affected consecutive therapy decisions in four patients (9%; 95% CI: 2.4-20.4%) compared with conventional staging (CT alone and colonoscopy). Therefore, the combination of FDG PET/CT in conjunction with a dedicated contrastenhanced CT protocol may be of interest as a possible single-step staging procedure.

Recurrent Disease The standard patient workup for the detection of recurrence and metastases in colorectal cancer includes regular clinical examinations, CT scans, colonoscopy, and usually the measurement of tumor markers such as CEA. However, this approach lacks specificity and may result in diagnostic and therapeutic delays due to several pitfalls: 1. while serological tumor markers are useful, CEA levels have only a 60-70% sensitivity for the detection of colorectal cancer recurrence [18]; 2. the morphology based information provided by CT does not permit distinction between post-surgical changes and tumor recurrence nor can it detect tumor involvement of normal-sized lymph nodes [19]; 3. colonoscopy is only useful in the detection of local recurrence. The suitability of FDG-PET in detecting recurrence and metastases has been shown in several studies. While there is obviously a clear advantage of PET/CT over PET alone, a dedicated contrast-enhanced CT is required often by clinicians. The study by Soyka et al. showed that contrast-enhanced PET/CT, as a single-step examination, has the same diagnostic confidence and impact as a sequential approach consisting of contrast-enhanced CT followed by non-contrast-enhanced PET/CT [20]. Although the

lesion detection rate on contrast-enhanced CT images is high, evaluation solely by this approach can be challenging because of the likelihood of inconclusive results that require further diagnostic evaluation (56% of our patient population). The reason for this is predominantly related to the specificity of the structural abnormalities that may be identified by this modality. Thus, nowadays, patients with inconclusive contrast-enhanced CT findings are being frequently referred for further evaluation with 18F-FDG PET/CT. More importantly, the study of Soyka et al. showed that in 21% of the patients with apparently conclusive findings on contrast-enhanced CT, the addition of non-contrast-enhanced PET/CT information led to appropriate changes in therapy. In clinical routine, however, those patients in whom contrast-enhanced CT findings were regarded as conclusive would not routinely be referred for further evaluation with FDG-PET/CT. If contrast-enhanced PET/CT had served as the initial imaging modality, 65% of these patients would have had a clear benefit, including changes in management as well as in diagnostic confidence. Therefore, it could be argued that contrast-enhanced PET/CT should be performed as the first-line diagnostic tool in the re-staging of colorectal cancer. Conversely, in 35% of the patients the associated radiation exposure would be futile, in addition to the needless expense of the procedure. However, this analysis holds true only if contrast-enhanced CT and non-contrastenhanced PET/CT are performed within 2-4 weeks. In reality, surgeons generally insist on contrast-enhanced CT studies not older than 4 weeks before taking a patient into the operating room. Thus, another additional scan with contrast enhancement (CT or PET/CT) would be needed in the majority of patients. Post-surgical and radiotherapy changes in the small pelvis are the most challenging for morphological imaging studies in recurrent rectal cancer, since tumor

194

recurrence cannot be differentiated from benign scar tissue. In a study by Even-Sapir et al., PET/CT was used to distinguish benign and malignant pre-sacral abnormalities. The sensitivity, specificity, positive predictive value, and negative predictive value were 100, 96, 88, and 100%, respectively. PET/CT findings were clinically relevant in 47% of 62 patients [21]. However, there was no comparison to other, conventional imaging studies. In our own study, the diagnostic value of contrast-enhanced CT and non-enhanced PET/CT was prospectively evaluated and compared in 76 patients referred for pre-operative evaluation for liver resection for metastatic colorectal cancer [22]. Extrahepatic disease was missed by contrastenhanced CT in one-third of the cases (sensitivity 64%), while PET/CT failed to detect extrahepatic lesions in only 11% (sensitivity 89%; p = 0.02). New findings on PET/CT resulted in a change in the therapeutic strategy in 21% of the patients. This study also demonstrated the well known limitation in PET imaging’s spatial resolution at ~4-6 mm, since small tumors (<5 mm) were often not detected. Also, in patients who had received chemotherapy within the month prior to PET/CT there was a high incidence of false-negative results. This effect, however, might be used as a predictor of success in neoadjuvant chemotherapy before resection. In summary, the abovedescribed studies clearly illustrate the advantages of PET/CT imaging in colorectal cancer.

Stomach and Small Bowel Gastric Cancer In a review by Dassen et al. regarding the pre-operative diagnostic utility of FDG-PET/CT in gastric cancer, the authors concluded, that FDG-PET has no role in the primary detection of gastric cancer due to its low sensitivity [23]. FDG-PET, however, is slightly better than CT in the evaluation of lymph node metastases in gastric cancer and therefore might have a role in pre-operative staging. Improvements in accuracy could be achieved by using PET/CT or PET tracers other than FDG, but these strategies need further investigation. Nonetheless, FDG-PET is able to adequately detect therapy responders at an early stage following neoadjuvant chemotherapy.

Small Bowel Adenocarcinoma of the small bowel comprises a group of rare tumors, with an annual incidence of around 3.7 per million people. The pathological (development from adenomatous polyps), genetic, and epidemiological features are very similar to those of adenocarcinomas of the large bowel but the prognosis is poor, with reported 5-year-survival rates of 15-35%. Capsule endoscopy, CT enteroclysis, and contrast-enhanced CT are mainly used in the diagnosis and workup of the disease. While these tumors show clear FDG avidity and might therefore be

Thomas F. Hany

evaluated by FDG-PET/CT, there are no comprehensive data on the use of this modality. Carcinoid tumors, as malignant tumors of non-epithelial origin, have a similar annual incidence as adenocarcinomas (3.8 per million people). The different subtypes have a natural behavior ranging from benign to high-grade malignancies. Similar to NETs of the pancreas, the diagnostic workup consists of nuclear medicine studies including SRS combined with contrast-enhanced CT (SPECT/CT) or 68Ga-DOTA-NOC, since >80% of carcinoid tumors express somatostatin receptors. The detection rates are therefore similar to those of NET of the pancreas [12]. Gastrointestinal stromal tumors (GIST) are mesenchymal tumors that in approximately 90% of patients originate in the stomach and small intestine. Unlike contrastenhanced CT, FDG-PET is able to show early effects in patients undergoing treatment with imatinib mesylate (Glivec; Novartis, Switzerland) [24]. In two recent studies, it was shown that patients without FDG uptake after the start of treatment had a better prognosis than patients with residual activity not demonstrated with contrastenhanced CT [17, 25]. Furthermore, lesions were better defined on PET/CT than by PET and CT performed sideby-side. This is relevant information for clinical decisionmaking.

Conclusions In the work-up of several abdominal malignancies of the gastrointestinal tract, FDG-PET/CT imaging is becoming increasingly well established. Its main advantage lies in the comprehensive evaluation of the patient, including all body compartments, and therefore in the detection of pivotal, therapy-deciding lesions. In the evaluation of primary liver tumors (cholangiocarcinoma and poorly differentiated HCC), FDG-PET/CT offers high sensitivity in the detection of distant metastases. Secondary liver tumors, such as metastases from the gastrointestinal tract, are detected by FDG-PET/CT at a high rate, making this imaging technology a primary tool in the evaluation of patients with suspicion of recurrent colon cancer. Further, full integration of contrast-enhanced CT protocols improves diagnostic confidence and reduces the sometimes cumbersome diagnostic pathway for patients. New tracers such as 68Ga-DOTA-TATE or 18F-DOPA will bring significantly improved diagnostic confidence in the notoriously difficult evaluation of patients with neuroendocrine tumors.

References 1. von Schulthess GK, Steinert HC, Hany TF (2006) Integrated PET/CT: current applications and future directions. Radiology 238:405-422 2. Torizuka T, Tamaki N, Inokuma T et al (1995) In vivo assessment of glucose metabolism in hepatocellular carcinoma with FDG-PET. J Nucl Med 36:1811-1817

195

Integrated Imaging in Gastrointestinal Oncology: PET/CT Imaging

3. Park JW, Kim JH, Kim SK et al (2008) A prospective evaluation of 18F-FDG and 11C-acetate PET/CT for detection of primary and metastatic hepatocellular carcinoma. J Nucl Med 49:1912-1921 4. Kluge R, Schmidt F, Caca K et al (2001) Positron emission tomography with (18)F]fluoro-2-deoxy-D-glucose for diagnosis and staging of bile duct cancer. Hepatology 33:1029-1035 5. Fritscher-Ravens A, Bohuslavizki KH, Broering DC et al (2001) FDG PET in the diagnosis of hilar cholangiocarcinoma. Nucl Med Commun 22:1277-1285 6. Petrowsky H, Wildbrett P, Husarik DB et al (2006) Impact of integrated positron emission tomography and computed tomography on staging and management of gallbladder cancer and cholangiocarcinoma. J Hepatol 45:43-50 7. Farma JM, Santillan AA, Melis M et al (2008) PET/CT fusion scan enhances CT staging in patients with pancreatic neoplasms. Ann Surg Oncol 15:2465-2471 8. Strobel K, Heinrich S, Bhure U et al (2008) Contrast-enhanced 18F-FDG PET/CT: 1-stop-shop imaging for assessing the resectability of pancreatic cancer. J Nucl Med 49:1408-1413 9. Pasquali C, Rubello D, Sperti C et al (1998) Neuroendocrine tumor imaging: can 18F-fluorodeoxyglucose positron emission tomography detect tumors with poor prognosis and aggressive behavior? World J Surg 22:588-592 10. Gibril F, Reynolds JC, Doppman JL et al (1996) Somatostatin receptor scintigraphy: its sensitivity compared with that of other imaging methods in detecting primary and metastatic gastrinomas. A prospective study. Ann Intern Med 125:26-34 11. Koopmans KP, de Vries EG, Kema IP et al (2006) Staging of carcinoid tumours with 18F-DOPA PET: a prospective, diagnostic accuracy study. Lancet Oncol 7:728-734 12. Ambrosini V, Tomassetti P, Castellucci P et al (2008) Comparison between 68Ga-DOTA-NOC and 18F-DOPA PET for the detection of gastro-entero-pancreatic and lung neuro-endocrine tumours. Eur J Nucl Med Mol Imaging 35:1431-1438 13. Bade MA, Ohki T, Cynamon J, Veith FJ (2001) Hypogastric artery aneurysm rupture after endovascular graft exclusion with shrinkage of the aneurysm: significance of endotension from a “virtual”, or thrombosed type II endoleak. J Vasc Surg 33:1271-1274 14. Clarke MP, Kane RA, Steele G Jr et al (1989) Prospective comparison of preoperative imaging and intraoperative ultrasonography in the detection of liver tumors. Surgery 106:849-855 15. Abdel-Nabi H, Doerr RJ, Lamonica DM et al (1998) Staging of primary colorectal carcinomas with fluorine-18 fluoro-

16.

17. 18.

19.

20.

21. 22.

23. 24.

25.

deoxyglucose whole-body PET: correlation with histopathologic and CT findings. Radiology 206:755-760 Kantorova I, Lipska L, Belohlavek O et al (2003) Routine (18)F-FDG PET preoperative staging of colorectal cancer: comparison with conventional staging and its impact on treatment decision making. J Nucl Med 44:1784-1788 Veit-Haibach P, Kuehle CA, Beyer T et al (2006) Diagnostic accuracy of colorectal cancer staging with whole-body PET/CT colonography. JAMA 296:2590-2600 Zervos EE, Badgwell BD, Burak WE Jr et al (2001) Fluorodeoxyglucose positron emission tomography as an adjunct to carcinoembryonic antigen in the management of patients with presumed recurrent colorectal cancer and nondiagnostic radiologic workup. Surgery 130:636-643; discussion 643-644 Goldberg RM, Fleming TR, Tangen CM et al (1998) Surgery for recurrent colon cancer: strategies for identifying resectable recurrence and success rates after resection. Eastern Cooperative Oncology Group, the North Central Cancer Treatment Group, and the Southwest Oncology Group. Ann Intern Med 129:27-35 Soyka JD, Veit-Haibach P, Strobel K et al (2008) Staging pathways in recurrent colorectal carcinoma: is contrast-enhanced 18F-FDG PET/CT the diagnostic tool of choice? J Nucl Med 49:354-361 Even-Sapir E, Parag Y, Lerman H et al (2004) Detection of recurrence in patients with rectal cancer: PET/CT after abdominoperineal or anterior resection. Radiology 232:815-822 Selzner M, Hany TF, Wildbrett P et al (2004) Does the novel PET/CT imaging modality impact on the treatment of patients with metastatic colorectal cancer of the liver? Ann Surg 240:1027-1034; discussion 1035-1036 Dassen AE, Lips DJ, Hoekstra CJ et al (2009) FDG-PET has no definite role in preoperative imaging in gastric cancer. Eur J Surg Oncol 35:449-455 Joensuu H, Roberts PJ, Sarlomo-Rikala M et al (2001) Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N Engl J Med 344:1052-1056 Goerres GW, Stupp R, Barghouth G et al (2004) The value of PET, CT and in-line PET/CT in patients with gastrointestinal stromal tumours: long-term outcome of treatment with imatinib mesylate. Comparison of PET, CT, and dual-modality PET/CT imaging for monitoring of imatinib (STI571) therapy in patients with gastrointestinal stromal tumors. Eur J Nucl Med Mol Imaging 4:4

NUCLEAR MEDICINE SATELLITE COURSE “DIAMOND”

IDKD 2010-2013

Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals Angelika Bischof Delaloye Service de Médecine Nucléaire, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland

Introduction Lymphoma is a heterogeneous family of diseases of the lymphatic system. There are two main entities, Hodgkin’s disease (HD) and non-Hodgkin’s lymphoma (NHL), the latter emerging either from B cells or T cells. Positron emission tomography (PET) with 18F-fluorodeoxyglucose (FDG) assisted by in-line CT (PET/CT) plays a pivotal role in patient management. HD shows high FDG uptake in close to 100% of the patients whereas FDG avidity greatly varies among patients with NHL subtypes. Aggressive lymphomas, such as diffuse large B cell lymphoma (DLBCL), the most current form of NHL, and mantle cell lymphomas (MCL) as well as Burkitt lymphomas and even plasmocytomas take up FDG very avidly. Follicular lymphomas (FL) also show high uptake in more than 90% of the cases whereas extranodal marginal zone lymphomas (MZL), such as splenic or mucosaassociated lymphoid tissue (MALT) MZL, are positive in only 67 and 54%, respectively [1]. It has been stipulated that the intensity of FDG uptake is an expression of disease aggressiveness [2-4]. This is probably true, at least partially, but is not really important in clinical practice. In fact, a great number of FL, with no signs of transformation, show high FDG uptake despite being considered indolent. The intensity of FDG uptake can therefore not be reliably used for diagnosing transformation in FL. Due to possible inconsistencies of uptake intensity versus aggressiveness and the great diversity of NHL, thorough histopathological characterization is necessary to determine the most appropriate treatment scheme. This step cannot be replaced by FDG-PET studies.

Particularities of Abdominal Lymphoma Lymphomas of the abdomen and pelvis most often simply correspond to subdiaphragmatic nodal involvement by one of the most frequent lymphoma types (Hodgkin’s disease, DLBCL, FL), usually characterized by high FDG uptake. However, in the abdomen other types, such as nodal and extranodal MZL, including MALT lymphomas, with much less consistently increased uptake need to be

considered [5, 6]. Subtypes of MALT lymphomas in particular have different uptake patterns. Most patients with MALT in which there is plasmacytic differentiation show FDG uptake whereas uptake is much less consistent in typical MALT [7]. Besides histological subtype, location is another factor that determines uptake patterns. Accordingly, the sensitivity of FDG-PET is lower in gastric (25-39%) than in non-gastric MALT (46-75%) [8, 9]. Gastric and extragastric MALT lymphomas frequently present as multifocal disease that is associated with chromosomal translocations or trisomy 18. Dissemination beyond the gastrointestinal tract occurs in about 10% of patients with gastric MALT while significantly more patients (close to 50%) with extragastric MALT tend to present with involvement of other MALT organs [9]. Involvement of the gastrointestinal tract occurs in 10-30% of patients with NHL. The affected sites, in decreasing order of frequency, are the stomach, small bowel, large bowel, and esophagus. The appearance of these tumors varies – from polypoidal masses to wall thickening, diffuse or focal infiltrations, ulcerations, and sometimes solitary or multiple nodules or extension from mesenteric nodes – and therefore so do the FDG uptake patterns [4, 5]. The occurrence of normally increased uptake in the gastrointestinal tract or of non-lymphomatous pathological conditions further complicates the diagnosis with FDG-PET [5, 6, 10]. Primary colorectal lymphomas are rare, representing 10-20% of gastrointestinal lymphomas and 0.2-0.6% of all large bowel malignancies. They are usually FDG-avid B cell lymphomas, with the most common sites being the cecum and rectosigmoid [11]. Focal lesions in the liver are frequently observed in patients with NHL. Civardi et al. studied the prevalence, clinical significance, and role of hepatitis C virus (HCV) infection in a population of 414 consecutive patients with NHL [12] and 129 focal liver lesions, 76 detected at disease onset and 53 during follow-up. The prevalence of focal liver involvement by NHL at disease onset was 7%. Among the focal liver lesions detected at diagnosis, 30 (39%) were due to lymphoma. In addition, ten cases of focal liver disease due to other malignancies were found, seven hepatocellular carcinomas (HCC), and three

200

metastases of another tumor. Of the focal liver lesions, 58% were benign. Conversely, at follow-up, 74% of the liver lesions were due to NHL and 15% to another malignancy. Focal liver lesions were more often the result of underlying disease (26%) in patients with aggressive lymphoma than in those with indolent lymphoma (4%). The incidence of focal lesions was similar in HCV-negative and HCV-positive patients, whereas HCC only occurred in the latter group. Another particular form of abdominal lymphoma is enteropathy-associated T cell lymphoma, which is seen in patients with refractory celiac disease. FDG-PET has been shown to be superior to CT alone in detecting enteropathy-associated lymphoma as well as its distant spread [13]. Finally, post-transplant lymphoproliferative disorders (PTLD) need to be mentioned. Immunosuppression is an important risk factor in the development of PTLD. The WHO has classified PTLD into morphological categories: early lesions, polymorphic PTLD, or monomorphic PTLD (various types of B and T cell lymphomas), Hodgkin’s lymphoma, and Hodgkin’slymphoma-like PTLD. The risk that a transplant patient will develop lymphoma over a 10-year period is 11.8 times greater than in the general population. Early lesions are most often seen in children and young adults and occur within the first year of transplantation while the other types appear later, post-transplantation [14]. PTLD is the most common malignancy observed in children after transplantation. Overall, it has been observed in up to 2.3% of kidney transplants, 2.8% of liver transplants, 6.3% of heart transplants, and 20% of small bowel transplants. The clinical presentation is similar to that of nontransplant-related lymphomas (lymphadenopathy, fever, weight loss, abdominal pain, splenomegaly). PTLD usually involves extranodal sites, in particular the gastrointestinal tract, and the allograft itself. Treatment consists of modifying the immunosuppression regimen, chemotherapy, and immunotherapy, alone or in combination. Adoptive T cell immunotherapy is under investigation.

Indications of FDG-PET/CT FDG-PET is indicated in the staging, response assessment, and re-staging of lymphoma. At diagnosis, the assessment of FDG avidity is probably more important than the mere staging that is usually performed based on contrast enhanced CT. FDG-PET only occasionally leads to upstaging, by revealing distant sites of involvement in patients with stage I or II Hodgkin’s lymphoma or NHL, and therefore to management changes. In advanced disease, upstaging is less relevant, as the same first-line treatment is usually administered in stage III and IV disease.

Response Assessment The most important indication is assessment of the patient’s response to therapy already after a few treatment

Angelika Bischof Delaloye

courses (chemo- or immunochemotherapy). There are a consistent number of publications showing that early conversion to a negative FDG-PET scan, irrespective of the presence of residual masses on CT, indicates a much more favorable outcome than in patients showing persistent FDG uptake in their lymphomatous lesions [15-19]. This observation has led to the establishment of new response criteria (Table 1); these were originally meant to be used as surrogate markers for clinical trials only but are currently also applied in daily patient management [20-22]. This classification not only eliminates the CRu (complete response unconfirmed) category but these combined criteria provide a more accurate response classification than International Workshop Criteria (IWC) criteria alone. Indeed, in a multivariate analysis, it represented the only significant independent predictor of progression free survival (PFS) [22]. Interestingly, the predictive power of early FDG-PET (after 2-3 treatment courses) is higher than that of end of treatment PET. This suggests that the early disappearance of FDG uptake is related to the chemosensitivity of the tumor whereas end of treatment FDG-PET more likely indicates the presence or absence of viable tumor cells [23]. Figure 1 shows two FDG-PET scans (MIP, maximal intensity projection) of a patient with AIDS-related lymphoma; the first, acquired after the patient had completed chemotherapy, shows residual uptake in the stomach and spleen. The patient was treated by more aggressive chemotherapy and splenectomy, neither of which could hinder the fulgurating progression of the disease, as shown on the second scan. The IWC criteria include assessment by contrastenhanced CT, which is usually performed at diagnosis. The discussion about the necessity of a full diagnostic CT at response assessment is not entirely closed. Most authors tend to agree, however, that contrast-enhanced CT is not necessary as a routine procedure but may be helpful in special cases, such as in the characterization of small nodes when FDG uptake is difficult to estimate because of partial-volume effects [24, 25].

Table 1. Response criteria based on International Workshop Criteria (IWC) and FDG-PET IWC+PET response

Criteria

ACR

CR, CRu, PR, or SD (IWC) PET negative BM negative if positive prior therapy CR, CRu, or PR (IWC) PET positive SD (IWC) PET positive PD (IWC) PET positive corresponding to CT abnormality

PR SD PD

PET, Positron emission tomography; CR, complete response; CRu, complete response unconfirmed; PR, partial response; SD, sudden death; BM, bone marrow; PD, progressive disease; CT, computed tomography

201

Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals

a

b

Fig. 1 a, b. A 41-year-old patient after chemotherapy for AIDS-related diffuse large B cell lymphoma. a Partial response with residual uptake in the stomach and splenomegaly. b Progression 4 months later, after intensification of chemotherapy and splenectomy

A main difficulty, however, remains: the definition of FDG-PET negativity. An independent nuclear medicine evaluation of the Eastern Cooperative Oncology Group (ECOG) revealed only moderate reproducibility of the interpretation of interim FDG-PET scans [26]. Some faint diffuse uptake may persist in residual masses and is due merely to the presence of these masses, without necessarily being indicative of viable tumors. It may, however, be difficult to call such a site FDG-negative. For this reason, Mikhaeel et al. proposed an intermediate category, minimal residual uptake (MRU) [17]. In a series of 121 patients with aggressive lymphoma, a 5-year PFS of 89, 59, and 16% was observed for patients with no, minimal, and positive residual uptake, respectively, after therapy. However, this additional category is not widely used; instead, in most reports the results have been designated as negative or positive. An International Harmonization Project (IHP) was convened to discuss the standardization of clinical-trial parameters in lymphoma based on consensus recommendations derived from the published PET literature and the collective expertise of its members [27]. These experts recommended visual assessment for determining PET positivity/negativity. Others could show that a reduction of the maximum SUV (standardized uptake value) by 65.7% allowed better separation of patients with improved event-free survival (EFS) (2-year EFS: 21 vs. 79% in the groups with ≤65.7% >SUVmax decrease, respectively) than achieved with visual analysis (2-year EFS: 51 vs. 79% in the PET-positive and PET-negative groups, respectively), mainly by characterizing indeterminate uptake as negative [28].

IHP Criteria of Visual FDG-PET Interpretation • PET is considered positive in case of focal or diffuse FDG uptake above background in a region where no physiological uptake should occur. • Mild and diffusely increased FDG uptake at the site of a residual mass >2 cm in diameter regardless of localization, with an intensity less than mediastinal blood pool structures, is considered negative. • With respect to partial volume effect, uptake above surrounding background in residual masses <2 cm or in normal lymph nodes should be considered positive. • New lung nodules >1.5 cm in a patient with no evidence of lymphoma before therapy should be considered positive for lymphoma only if their uptake exceeds that of mediastinal blood pool, except in patients in whom a complete response is determined in all known lymphoma sites. In such patients, these “new lung lesions” most often correspond to infectious or inflammatory changes. • Residual hepatic or splenic lesions >1.5 cm should be considered positive if their uptake exceeds that of liver or spleen, respectively, and negative if their uptake is equal or lower than that of the surrounding organ. Diffusely increased splenic uptake (> normal liver) is compatible with splenic involvement except <10 days after treatment with cytokines, namely, granulocytecolony stimulating factors [29]. • Clearly increased (multi)focal bone marrow uptake should be considered positive for lymphoma but needs to be compared to baseline PET (multifocal involvement?). Diffusely increased bone marrow uptake is usually due to bone marrow stimulation and should not be considered as indicating lymphoma. These criteria were essentially developed for patients with Hodgkin’s lymphoma or aggressive NHL, in which a management change in patients who did not respond or only partially responded to first-line therapy might improve outcome. Persistence of FDG-uptake in most cases points to therapy-resistant disease, with a high recurrence rate and reduced PFS (10-50% at 1 year compared with 79-100% in FDG-PET responders) [30, 31]. However, it remains uncertain whether more aggressive treatment administered at an earlier time point (after 2-3 first-line treatment courses) will be able to improve outcome to a greater extent than achieved when such treatment is given at first clinical relapse or as consolidation at the end of first-line therapy. Several trials are currently under way to study this essential question but no definite response is available so far. On the contrary, the absence of residual FDG uptake has a high negative predictive value in patients with aggressive NHL as well as in those with Hodgkin’s disease [31, 32]. Therefore, it may well be possible to reduce the number of treatment courses and/or involved field irradiation in excellent responders in order to avoid late side effects, such as other organ tumors (increased incidence of unilateral and bilateral breast cancer in female patients previously treated for

202

mediastinal Hodgkin’s disease), or the increased incidence of heart, mainly coronary artery disease. In patients with primary gastric lymphoma, DLBCL, or MALT MZL, persistent FDG uptake is not necessarily a sign of residual lymphoma. In a study of 24 patients examined using FDG-PET and endoscopy at follow-up after treatment, 11 patients with ulcerative or mucosal lesions on endoscopy showed residual FDG uptake while no lymphoma cells were evidenced by histopathology [4].

Pitfalls in Interpretation In addition, to those already mentioned above (faint uptake in residual masses, inflammatory changes in treated lesions and lung infiltrates, bone marrow stimulation, normal stomach and bowel uptake, urinary tract activity) all other well known pitfalls have to be considered (brown adipose tissue uptake, thymic rebound, granulomatous disease, muscle/bone uptake, contaminations, etc.). A noteworthy situation may be found in patients treated with immuno-chemotherapy, in whom prolonged recruitment of inflammatory cells to the tumor may occur. In vaccination trials, increased FDG uptake in responding tumors was also observed and must be considered a favorable sign of response to treatment.

Angelika Bischof Delaloye

Pharma, Germany) to controls followed-up by watchful waiting, median PFS was 2 years longer in treated patients [41]. Consequently, the indication of consolidation has been added by most medical regulatory agencies (FDA, EMEA, Swissmedic). Figure 2 shows the wholebody scan obtained with 111In-ibritumomab tiuxetan (Zevalin) 3 days after injection in a patient who underwent first-line chemotherapy for FL and who was considered to be in partial remission according to IWC criteria. Radioimmunoscintigraphy clearly shows uptake in residual tumor sites of the abdomen. After radioimmunotherapy, this patient achieved CR, including molecular remission, which has been ongoing for more than 6 years. It is also interesting that a significant dose-efficiency relationship between the dose delivered to the whole body and bone marrow and PFS has been observed. This finding underlines the importance of further refining dosimetry to improve the overall success of radionuclide therapy [42]. Additional studies are necessary to evaluate the respective roles of consolidation with Zevalin and maintenance with non-labeled antibodies such as rituximab. Furthermore, available data suggest that radioimmunotherapy, associated with high-dose chemotherapy, can be successfully used in conditioning before autologous stem cell transplantation (ASCT). The data reported so far regarding reduced-intensity conditioning before allogeneic stem cell transplantation have also indicated

Radioimmunotherapy of Lymphomas Radiolabeled antibodies directed against the surface antigens CD20 (tositumomab, ibritumomab tiuxetan) and CD22 (epratuzomab) of B cells have successfully been used for the treatment of relapsing/resistant NHL, for firstline treatment as well as for consolidation after first-line chemotherapy [33-36]. In prior trials, the usual end point was defined as response to treatment (complete, CR, or partial, PR), with higher response rates, including CR, observed with radiolabeled than with unlabeled antibodies [36]. Higher response rates were observed in patients with FL but have also been reported for other, more aggressive types of lymphoma, such as DLBCL and MCL, as well as in transformed FL. Radioimmunotherapy is usually well tolerated, the only major side effects are related to bone marrow depression, which appears later than with conventional chemotherapy and is marked by a nadir 5-8 weeks after treatment. However bone marrow depression is most often easily managed. The incidence of myelodysplastic syndrome/acute myeloid leukemia (MDS/AML) is comparable to that observed with conventional chemotherapies, in particular those containing anthracyclines. In patients achieving CR, very long PFS times have been noted [37-39]. Thus, it can be expected that a subgroup of patients with advanced FL, a disease considered incurable today, might eventually be cured by a combined approach of chemo(immuno)therapy and radioimmunotherapy [40]. In a recent randomized trial comparing consolidation with 90Y-ibritumomab tiuxetan (Zevalin, Bayer Schering

Fig. 2. Anterior whole-body scan obtained 3 days after injection of 111Indium-tositumomab tiuxetan (Zevalin) in a patient who underwent first-line chemotherapy for follicular non-Hodgkin’s lymphoma. According to the International Workshop Criteria classification, this patient was in partial remission. The scan clearly shows uptake of the labeled antibody in the residual abdominal mass. Consolidation with 90Yttrium-ibritumomab tiuxetan (Zevalin) allowed her to achieve complete remission, including molecular remission, which has been ongoing for more than 6 years

Lymphoma: Diagnostic and Therapeutic Applications of Radiopharmaceuticals

positive results. There is also evidence that radioimmunotherapy combined with ASCT can improve clinical outcome (fewer relapses) without added toxicity; this therapeutic strategy therefore represents a very interesting treatment option for elderly patients, who account for the majority of the NHL population [43].

Conclusions Nuclear medicine plays a pivotal role in the management of patients with Hodgkin’s and NHL, particularly by allowing response assessment at an early stage of treatment (2-3 courses). The results have high prognostic implications, with complete metabolic responders having a highly significant better outcome than non-responders. Radioimmunotherapy using labeled anti-CD20 antibodies combined with chemo(immuno)therapy seems to represent an efficient treatment option, allowing prolonged PFS when administered as consolidation after first-line therapy. It is also of interest as a conditioning regimen before ASCT, and potentially in patients undergoing allogeneic stem cell transplantation or in those with relapsed/resistant NHL.

References 1. Weiler-Sagie M, Bushelev O, Epelbaum R et al (2010) 18FFDG avidity in lymphoma readdressed: A study of 766 patients. J Nucl Med 51:25-30 2. Hutchings M, Loft A, Hansen M et al (2006) Different histopathological subtypes of Hodgkin’s lymphoma show significantly different levels of FDG uptake. Hematol Oncol 24:146-150 3. Schoder H, Noy A, Gonen M et al (2005) Intensity of 18fluorodeoxyglucose uptake in positron emission tomography distinguishes between indolent and aggressive non-Hodgkin’s lymphoma. J Clin Oncol 23:4643-4651 4. Yi JH, Kim SJ, Choi JY (2009) (18)F-FDG uptake and its clinical relevance in primary gastric lymphoma. Hematol Oncol Epub 5. Lee WK, Lau EW, Duddalwar VA et al (2008) Abdominal manifestations of extranodal lymphoma: spectrum of imaging findings. AJR Am J Roentgenol 191:198-206 6. Tateishi U, Terauchi T, Inoue T, Tobinai K (2009) Nodal status of malignant lymphoma in pelvic and retroperitoneal lymphatic pathways: PET/CT. Abdom Imaging Epub 7. Hoffmann M, Wohrer S, Becherer A et al (2006) 18F-Fluorodeoxy-glucose positron emission tomography in lymphoma of mucosa-associated lymphoid tissue: histology makes the difference. Ann Oncol 17:1761-1765 8. Perry C, Herishanu Y, Metzer U et al (2007) Diagnostic accuracy of PET/CT in patients with extranodal marginal zone MALT lymphoma. Eur J Haematol 79:205-209 9. Raderer M, Wohrer S, Streubel B et al (2006) Assessment of disease dissemination in gastric compared with extragastric mucosa-associated lymphoid tissue lymphoma using extensive staging: a single-center experience. J Clin Oncol 24:31363141 10. Takahashi H, Ukawa K, Ohkawa N et al (2009) Significance of (18)F-2-deoxy-2-fluoro-glucose accumulation in the stomach on positron emission tomography. Ann Nucl Med 23:391-397 11. Wong MT, Eu KW (2006) Primary colorectal lymphomas. Colorectal Dis 8:586-591

203

12. Civardi G, Vallisa D, Berte R et al (2002) Focal liver lesions in non-Hodgkin’s lymphoma: investigation of their prevalence, clinical significance and the role of Hepatitis C virus infection. Eur J Cancer 38:2382-2387 13. Hadithi M, Mallant M, Oudejans J et al (2006) 18F-FDG PET versus CT for the detection of enteropathy-associated T-cell lymphoma in refractory celiac disease. J Nucl Med 47:1622-1677 14. Zafar SY, Howell DN, Gockerman JP (2008) Malignancy after solid organ transplantation: an overview. Oncologist 13:769-778 15. Hutchings M, Loft A, Hansen M et al (2006) FDG-PET after two cycles of chemotherapy predicts treatment failure and progression-free survival in Hodgkin’s lymphoma. Blood 107:52-59 16. Jerusalem G, Beguin Y, Fassotte MF et al (2000) Persistent tumor 18F-FDG uptake after a few cycles of polychemotherapy is predictive of treatment failure in non-Hodgkin’s lymphoma. Haematologica 85:613-618 17. Mikhaeel NG, Hutchings M, Fields PA et al (2005) FDG-PET after two to three cycles of chemotherapy predicts progressionfree and overall survival in high-grade non-Hodgkin’s lymphoma. Ann Oncol 16:1514-1523 18. Mikhaeel NG, Timothy AR, O’Doherty MJ et al (2000) 18FDG-PET as a prognostic indicator in the treatment of aggressive Non-Hodgkin’s Lymphoma-comparison with CT. Leuk Lymphoma 39:543-553 19. Spaepen K, Stroobants S, Dupont P et al (2002) Early restaging positron emission tomography with (18)F-fluorodeoxyglucose predicts outcome in patients with aggressive nonHodgkin’s lymphoma. Ann Oncol 13:1356-1363 20. Brepoels L, Stroobants S, De Wever W et al (2007) Aggressive and indolent non-Hodgkin’s lymphoma: response assessment by integrated international workshop criteria. Leuk Lymphoma 48:1522-1530 21. Cheson BD, Pfistner B, Juweid ME et al (2007) Revised response criteria for malignant lymphoma. J Clin Oncol 25: 579-586 22. Juweid ME, Wiseman GA, Vose JM et al (2005) Response assessment of aggressive non-Hodgkin’s lymphoma by integrated International Workshop Criteria and fluorine-18-fluorodeoxyglucose positron emission tomography. J Clin Oncol 23:4652-4661 23. Engles JM, Quarless SA, Mambo E et al (2006) Stunning and its effect on 3H-FDG uptake and key gene expression in breast cancer cells undergoing chemotherapy. J Nucl Med 47: 603-608 24. Rodriguez-Vigil B, Gomez-Leon N, Pinilla I et al (2006) PET/CT in lymphoma: prospective study of enhanced fulldose PET/CT versus unenhanced low-dose PET/CT. J Nucl Med 47:1643-1648 25. Schaefer NG, Hany TF, Taverna C et al (2004) Non-Hodgkin’s lymphoma and Hodgkin’s disease: coregistered FDG PET and CT at staging and restaging- do we need contrast-enhanced CT? Radiology 232:823-829 26. Horning SJ, Juweid ME, Schoder H et al (2009) Interim positron emission tomography (PET) scans in diffuse large B-cell lymphoma: an independent expert nuclear medicine evaluation of the Eastern Cooperative Oncology Group E3404 study. Blood Epub 27. Juweid ME, Stroobants S, Hoekstra OS et al (2007) Use of positron emission tomography for response assessment of lymphoma: consensus of the Imaging Subcommittee of International Harmonization Project in Lymphoma. J Clin Oncol 25:571-578 28. Lin C, Itti E, Haioun C et al (2007) Early 18F-FDG PET for prediction of prognosis in patients with diffuse large B-cell lymphoma: SUV-based assessment versus visual analysis. J Nucl Med 48:1626-1632 29. Sugawara Y, Zasadny KR, Kison PV et al (1999) Splenic fluorodeoxyglucose uptake increased by granulocyte colonystimulating factor therapy: PET imaging results. J Nucl Med 40:1456-1462

204

30. Brepoels L, Stroobants S, Verhoef G (2007) PET and PET/CT for response evaluation in lymphoma: current practice and developments. Leuk Lymphoma 48:270-282 31. Hutchings M, Barrington SF (2009) PET/CT for therapy response assessment in lymphoma. J Nucl Med 50 Suppl 1: 21S-30S 32. Kobe C, Dietlein M, Franklin J et al (2008) Positron emission tomography has a high negative predictive value for progression or early relapse for patients with residual disease after first-line chemotherapy in advanced-stage Hodgkin’s lymphoma. Blood 112:3989-3994 33. Kaminski MS, Tuck M, Estes J et al (2005) 131I-tositumomab therapy as initial treatment for follicular lymphoma. N Engl J Med 352:441-449 34. Morschhauser F, Illidge T, Huglo D et al (2007) Efficacy and safety of yttrium-90 ibritumomab tiuxetan in patients with relapsed or refractory diffuse large B-cell lymphoma not appropriate for autologous stem-cell transplantation. Blood 110:54-58 35. Sharkey RM, Brenner A, Burton J et al (2003) Radioimmunotherapy of non-Hodgkin’s lymphoma with 90Y-DOTA humanized anti-CD22 IgG (90Y-Epratuzumab) do tumor targeting and dosimetry predict therapeutic response? J Nucl Med 44:2000-2018 36. Witzig TE, Gordon LI, Cabanillas F et al (2002) Randomized controlled trial of yttrium-90-labeled ibritumomab tiuxetan radioimmunotherapy versus rituximab immunotherapy for patients with relapsed or refractory low-grade, follicular, or transformed B-cell non-Hodgkin’s lymphoma. J Clin Oncol 20:2453-2463

Angelika Bischof Delaloye

37. Buchegger F, Antonescu C, Bischof Delaloye A et al (2006) Long-term complete responses after 131I-tositumomab therapy for relapsed or refractory indolent non-Hodgkin’s lymphoma. Br J Cancer 94:1770-1776 38. Fisher RI, Kaminski MS, Wahl RL et al (2005) Tositumomab and iodine-131 tositumomab produces durable complete remissions in a subset of heavily pretreated patients with lowgrade and transformed non-Hodgkin’s lymphomas. J Clin Oncol 23:7565-7573 39. Witzig TE, Molina A, Gordon LI et al (2007) Long-term responses in patients with recurring or refractory B-cell nonHodgkin’s lymphoma treated with yttrium 90 ibritumomab tiuxetan. Cancer 109:1804-1810 40. Buchegger F, Press OW, Delaloye AB, Ketterer N (2008) Radiolabeled and native antibodies and the prospect of cure of follicular lymphoma. Oncologist 13:657-667 41. Morschhauser F, Radford J, Van Hoof A et al (2008) Phase III trial of consolidation therapy with yttrium-90-ibritumomab tiuxetan compared with no additional therapy after first remission in advanced follicular lymphoma. J Clin Oncol 26:5156-5164 42. Bischof Delaloye A, Antonescu C, Louton T et al (2009) Dosimetry of 90Y-ibritumomab tiuxetan as consolidation of first remission in advanced-stage follicular lymphoma: results from the international phase 3 first-line indolent trial. J Nucl Med 50:1837-1843 43. Gisselbrecht C, Vose J, Nademanee A et al (2009) Radioimmunotherapy for stem cell transplantation in non-Hodgkin’s lymphoma: in pursuit of a complete response. Oncologist 14 Suppl 2:41-51

IDKD 2010-2013

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders Ariane Boubaker Service de Médecine Nucléaire, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland

Introduction Conventional nuclear medicine investigations of the gastrointestinal (GI) tract are not commonly performed in routine clinical practice, although prevalence of disorders affecting the upper gastrointestinal tract is quite high ranging from 15% to 40% in European countries. Most diagnostic tests used to differentiate organic from nonorganic cause are invasive (endoscopy, manometry, pH monitoring) and may be not well tolerated by some patients. Radionuclide procedures are non-invasive and allow characterizing functional and motility abnormalities of the esophagus and stomach. They are easy to perform, widely available and deliver low radiation burden to the patient at low cost. Despite the lack of standardization, both esophageal transit studies and gastric emptying scintigraphy have shown excellent diagnostic results and are helpful complementary tools for the clinician at diagnosis and/or follow-up after surgery or conservative management. Conventional nuclear medicine investigations of the urinary system are part of the daily routine clinical practice and procedures are well established and standardized, being a common investigation in children with congenital abnormalities and/or recurrent urinary tract infections. In adults, dynamic renal scintigraphy with/without ACE inhibitor is useful to detect haemodynamically significant renal artery stenosis in patients with suspicion of renovascular hypertension. Preoperative assessment of renal function (hydronephrosis, live kidney donor, renal tumor) and follow-up of renal transplanted patients are also common indications.

Gastrointestinal Tract Salivary Glands The main indication to perform salivary-gland scintigraphy is xerostomia and suspicion of Sjögren disease. It may help the clinician in assessing parenchymal dysfunction due to inflammatory disease and/or fibrosis. A 30-minute dynamic acquisition is started after injection

of 150 to 200 MBq of Tc-99m pertechnetate with lemon juice given at 15 min, allowing studying both the extraction and secretion phase. Quantitative parameters (captation index, extraction ratio) are helpful to confirm low parenchymal extraction.

Esophageal Transit Esophageal transit scintigraphy is useful for the diagnosis and evaluation of response to treatment for achalasia, sclerodermia, esophageal spasm, or dysfunction due to gastroesophageal reflux disease (GERD) [1-3]. The patient can be examined in sitting or supine position: position and gravity will affect the transit-time, the upright position being more physiologic. Having the patient lying supine may increase the sensitivity of the procedure. The patient must have been fasting for at least 3 hours before examination. He should practice first with unlabeled liquid to avoid false-positive abnormal transit by moistening the esophageal mucosa, and get used to the test. Esophageal motility varies with viscosity and volume of the bolus: the ejection force of the pharynx is sufficient to propel a liquid from mouth to stomach, whereas a semisolid bolus requires more peristalsis from the two lower thirds of the esophagus, thus increasing the sensitivity of the test. A large field of view cameras is used to image the entire esophagus including mouth and stomach. If anterior and posterior projections cannot be acquired simultaneously, the anterior projection should be preferred. The typical procedure consists of a high-rate dynamic acquisition (0.5 s/frame for 2 min) started just before swallowing of 10-20 mL of orange juice containing 5-10 MBq of 99mTc-sulfur colloids, or any radiopharmaceutical not absorbed by the gastrointestinal tract. Both activity and volume have to be adjusted when examining children. Qualitative evaluation should be done in cine mode to identify abnormal pattern such as oral/pharyngeal retention, bolus fragmentation, tracheal aspiration and gastroesophageal reflux. Esophageal transit time is quantified by drawing ROIs encompassing the upper, middle and lower third of the esophagus and the stomach. Normal transit time (>10% activity has cleared from esophagus) is <15 s.

206

Gastroesophageal Reflux Disease Clinical presentation of GERD differs considerably in children when compared to adults [4-6]. In adults, the main symptom is heartburn. Common manifestations of reflux in children include respiratory symptoms and failure to thrive. Gastroesophageal reflux (GER) is physiologic in infants and resolves spontaneously at 79 months of age, a clinically important reflux being generally evident by 2 months of age. Serious complications of GERD may be esophagitis, bleeding, perforation, Barrett’s esophagus, cancer, stricture or recurrent pulmonary infections/asthma. The radionuclide method is the most sensitive non-invasive diagnostic test to detect GER. Because esophageal transit time is an important factor that may maintain or facilitate GER, the procedure should include an esophageal transit scintigraphy if possible. Half of the liquid (milk, water, or orange juice) is labeled with 1 MBq 99mTc-sulfur colloids/kg (minimum 3 MBq), the other half is used to wash radioactivity from mouth and esophagus and to complete the feeding. In case of dual phase study (liquid and solid), one third of the total activity to be administered is used for the liquid, and two thirds to label the solid (eggs). The child/patient is placed supine with the camera in the back with stomach and mouth in the field of view. A 30- to 60-min dynamic acquisition at a rate of 5-10 sec/frame is recommended. A high sensitivity collimator should be used to increase the sensitivity of the test. In older children and adults, GER can be provoked by an increased abdominal pressure (cough, Valsalva maneuver). Data should be analyzed in cine mode in order to describe the importance and frequency of the reflux. GER can be quantified as a fraction of total initial gastric activity and/or an overall reflux index. The study should be completed by 5-minute anteroposterior images of the thorax to look for possible aspiration. If negative, 10-15 min additional views of the thorax should be repeated at 2 and 24 h [7].

Gastric Emptying Scintigraphy Both rapid and delayed gastric emptying (GE) can cause the same symptoms [8]. Causes of rapid GE are postoperative (pyloroplasty, hemigastrectomy), functional dyspepsia, hyperthyroidism and Zollinger-Ellison syndrome. Gastroparesis is defined as a delayed gastric emptying without mechanical outlet obstruction. The most common causes are diabetes, post-surgical and idiopathic. The true prevalence of gastroparesis is unknown (around 5%), women being affected more frequently (4:1, females: males) [9]. Gastric emptying scintigraphy (GES) is considered as the reference method as it is simple, largely available and non-invasive and provides reproducible results of GE measurement [10]. Major efforts have been made to standardize the procedure because the GE rate is influenced by the volume and meal content, patient’s condition and position, as well as drug interference [11].

Ariane Boubaker

Once a standardized protocol has been defined, normal results should be determined by each centre. Solid GES is generally affected first, being more sensitive for the diagnosis than the liquid GES. Liquid emptying is rapid, begins immediately after ingestion without any lag phase, and the normal clearance curve pattern is exponential (normal half-emptying <20 minutes). A normal solid gastric emptying is characterized by an initial lag-phase with no emptying (normal 5-25 minutes), during which the solid material is converted into chime by acids and mechanical grinding, followed by a linear constant-rate clearance. 99mTc-sulfur colloids are used because they are not absorbed in the gastrointestinal system. The meal should be labeled by cooking the radiopharmaceutical with proteins (i.e., egg albumin) to avoid elution from the solid to the liquid phase which would result in erroneously shortened GE. After a fast of 4 hours, the patient must eat the meal within 10-15 min and be positioned immediately after the end of the meal. The use of simultaneous antero-posterior projections using a double-headed camera will allow correcting for movements and geometric variations in counting. The total duration of the study is 3 h. Images should be displayed to describe the progression of the solid within the different portions of the stomach and the time of activity appearance in the duodenum. Additional events like gastroesophageal reflux, retention in esophagus/hiatal hernia should be described (Fig. 1). A time activity curve is generated using ROIs and 50% emptying time is calculated.

Gastrointestinal Bleeding Lower GI bleeding accounts for one third of all acute bleeding events, is more frequent in men, and increases in incidence with age with a mortality of about 4% [12]. The most common cause of acute lower bleeding is diverticulosis, followed by angiodysplasia. Because GI bleeding is often intermittent and possibly occurs at a slow rate, scintigraphy is particularly suitable for the accurate diagnosis having the ability to detect bleeding flow rates <0.1 mL/min. 99mTc-sulfur colloids can be used to localize the source of bleeding if it occurs during the first minute after injection. Scintigraphy with 99mTc labeled red blood cells (RBCs) is the best method in patients with intermittent, low rate GI bleeding and has been proven to be superior to barium enema, angiography, computer tomography (CT) scan and even colonic endoscopy in this clinical setting [13]. It is recommended to use the in vitro labeled RBC to minimize elution of 99mTc from the RBCs, thus avoiding false-positive results due to secretion of free pertechnetate by the kidneys, and gastric and colonic mucosa [14]. Scintigraphy gives the possibility to perform successive acquisitions till 24 h after injection and continuous monitoring if bleeding does not occur during the initial acquisition period. Single photon emission CT (SPECT) or SPECT/CT may help to localize the bleeding site. The diagnosis is made by detection of an extravasation site of RBCs that increases with time. Both anterograde

207

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders

a

7’

8’

9’

10’

26’

27’

28’

29’

30’

b

1h

2h

3h

80 60 40 20

and retrograde movements may occur in the intestinal lumen (Fig. 2). Dynamic imaging is essential to correctly identify the active source of bleeding. Colonic bleeding is seen at the periphery of the abdomen whereas bleeding in the small intestine is more central with rapid visualisation of curvilinear progression of the activity.

Meckel’s Diverticulum and Heterotopic Gastric Mucosa 99mTc

1.5h

100

c

Gastric emptying (%)

Fig. 1 a-c. Solid gastric emptying scintigraphy performed in a 69-year-old woman presenting signs and symptoms of gastroparesis (early satiety, bloating and post-prandial abdominal fullness). a Consecutive 1-minute posterior projections showing gastroesophageal reflux at 6-7 and 28-30 min. b Anterior and posterior images at 1 h, 1.5 h, 2 h and 3 h after ingestion of the 99mTc labeled solid meal reveal retention in the fundus and body of the stomach. c The time-activity curve corrected for radioactive decay confirms a delayed gastric emptying with a dedifferentiation of the initial lagphase when compared to normal values (dashed lines representing ± 1 SD)

6’

pertechnetate (99mTcO4–) scintigraphy has been used since the 70s to diagnose heterotopic gastric mucosa [12-15]. Mucin-secreting cells of the stomach and proximal small bowel are responsible for the uptake and secretion of 99mTcO – via the NIS (sodium iodide symporter system). 4 Unexplained gastrointestinal bleeding and/or recurrent abdominal pain are the main indication to perform 99mTcO4– scintigraphy to look for Meckel’s diverticulum (MD) or heterotopic gastric mucosa (HGM). The prevalence of MD in the general population is estimated to be 1-4%, with HGM being present in about 50% of MD. Reported sensitivity of scintigraphy range from 50% to 92% and pharmacological preparation with pentagastrin, glucagon, or

60

120

180 Time (min)

histamine type-2 (H2) receptor antagonists (cimetidine, ranitidine) have been used to increase the sensitivity of the test [16]. Pentagastrin is administered subcutaneous 15-20 min prior to scintigraphy (6 μg/kg) to increase the uptake 99mTcO – by the gastric mucosa; H receptor blockers in4 2 crease the uptake in the mucosa by blocking the secretion from the cells to the lumen; ranitidine (2 mg/kg in children, 150 mg in adults) should be preferred to cimetidine because it has fewer side-effects. Glucagon relaxes the smooth muscle of the gastrointestinal system and decreases peristalsis (50-6 μg/kg intravenous 10 min after 99mTcO – injection). Omeprazole is a proton pump in4 hibitor which has been also reported to increase the sensitivity of the MD scintigraphy. Patient should be fasting for 3-4 hours, and barium enema studies should not be performed during the 3-4 days prior to scintigraphy. The camera is positioned in anterior projection with stomach and bladder in the field of view. A 60×60 sec/frame dynamic acquisition is started after injection of 2 MBq/kg (10-150 MBq) of 99mTcO4– and additional static views (post void, erect, oblique, SPECT) are obtained. A bladder full of non-radioactive urine prior to injection of the

208

Ariane Boubaker

a

15”

30”

45”

16’

45’

1h20’

1’

1’15”

1h45’

2h05’

b

Fig. 2 a, b. 99mTc labeled red blood cells (RBCs) bleeding scintigraphy performed in a 71-year-old woman presenting acute lower gastrointestinal bleeding. Colonic endoscopy performed two days before scintigraphy did not allow localizing the source of active bleeding. a Initial dynamic acquisition started immediately after injection of 920 MBq of 99mTc-RBCs shows a diffuse activity in the periphery of the abdomen (arrow) at 45 sec that increases with time. b Consecutive anterior static views show both anterograde and retrograde progression of the activity in the transverse and sigmoid colon confirming the left descending colon to be the source of active bleeding

radiopharmaceutical may delay the renal excretion of pertechnetate in the urinary tract and avoid false-positive results during the dynamic acquisition. The diagnostic criterion for HGM is the appearance of a focal uptake of 99mTcO – at the same time as the gastric mucosa (Fig. 3). 4 MD is usually seen as a focal increased activity in the peritoneal cavity most frequently in the right lower part of the abdomen. Pitfalls in interpretation and causes for falsepositive and false-negative results are listed in Table 1.

Protein-Losing Enteropathy An excessive loss of protein in the gastrointestinal system may be due to lymphatic obstruction (intestinal lymphangiectasia, cirrhosis), inflammatory bowel disease (Crohn’s disease, ulcerative colitis) gastrointestinal malignancy (gastric cancer, lymphoma) and increased permeability (celiac disease, infections). Serial abdominal images acquired after intravenous injection of 99mTc labeled human serum albumin may show tracer extravasation and accumulation in the intestine helping to localize the site of excessive loss of protein.

Hepatobiliary Scintigraphy Hepatobiliary scintigraphy (HBS) is an imaging technique performed using 99mTc labeled iminodiacetic acids

derivatives, the most commonly used being mebrofenin (Bridatec, GE Healthcare, The Netherlands) [17, 18]. After intravenous injection, the radiopharmaceutical is rapidly extracted at the vascular pole of the hepatocyte and secreted at the biliary pole of the cell. Although not frequently performed in clinical routine practice, HBS may be useful in a variety of clinical situations. In neonates past the age of 2 weeks with persistent jaundice and hyperbilirubinemia, HBS may be used to distinguish the treatable causes of hepatitis (hypothyroidism, sepsis, panhypopituitarism, galactosemia) from extrahepatic biliary atresia, as early surgery performed before the age of 2 months offers a better prognosis. The child has to fast for at least 2 hours before injection and should receive a phenobarbital pre-treatment (5 mg/kg/day) 3 to 5 days before HBS in order to stimulate excretion of the radiopharmaceutical and avoid false-negative results. A dynamic acquisition of 60×1 min/frame in anterior projection is started immediately after intravenous injection of 2-7 MBq 99mTc-mebrofenin/kg (minimal recommended activity 20 MBq), with consecutive static views performed at 2, 4, 6 and 24 hours. A normal HBS includes a rapid extraction in the liver with no significant cardiac blood-pool activity at 5-10 minutes post injection and the visualization of activity in the small bowel at 60 minutes. In the neonate the intrahepatic main bile ducts are usually not seen. The diagnostic value of HBS relies

209

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders

a

Fig. 3 a, b. Meckel’s diverticulum scintigraphy performed in a 2-year-old child 24 b hours after a first episode of acute gastrointestinal bleeding. a Initial dynamic acquisition demonstrates a focus of activity in the lower left quadrant of the abdomen that appears simultaneously to gastric mucosa (arrows). b Left and right lateral and antero-posterior static views obtained 1 hour after injection of 60 MBq of 99mTc-pertechnetate show persistent focal uptake highly suggestive of a Meckel’s diverticulum containing gastric mucosa as the source of acute gastrointestinal bleeding

1’

2’

3’

4’

5’

6’

7’

8’

Table 1. Pitfalls and sources of error in the interpretation of Meckel’s diverticulum scintigraphy False-negative results 1. Procedures Barium studies (3-4 days prior to scintigraphy) Administration of potassium perchlorate (thyroid blocade) 2. Anatomical causes Ischemia, necrosis, ulceration Lack of heteropic gastric mucosa Obscured by urinary tract or full stomach False-positive results 1. Procedures Endoscopy Laxative 2. Anatomical causes Urinary tract activity (hydronephrosis, ectopic kidney, bladder diverticulum) Small bowel obstruction (intussusception, volvulus) Neoplasm (carcinoid, lymphoma, colic adenocarcinoma) Inflammation (Crohn’s disease, appendicitis, peptic ulcer) Other site of heterotopic gastric mucosa (duplication, Barrett’s esophagus) Vascular abnormalities (angiodysplasia, hemangioma, aneurysm)

on its high-negative predictive value in case of free excretion in the small bowel, whereas an absence of intestinal activity at 24 hours is not specific for biliary atresia and may be due to other causes of hepatocellular dysfunction. Pitfalls in interpretation are mostly related to the urinary excretion of the tracer, posterior projection or SPECT (SPECT/CT) being helpful to avoid false-positive results. Reported sensitivity and specificity of HBS for the diagnosis of biliary atresia are 83 to 100%, 33 to 100%, respectively [19]. The gold standard is liver biopsy with a sensitivity of 89-99% and a specificity of 83-98%. Common indications for HBS in adults are acute cholecystitis, biliary leak/extravasation after surgery and follow-up after liver transplantation. A normal HBS is characterized by a diffuse activity in the liver beginning 6-8 s after spleen and kidneys (75% of blood supply to the liver supported by the portal vein), a rapid blood pool clearance (5-10 min), visualization of common bile duct and gallbladder 10-30 min after injection. Transit from the biliary ducts to the small bowel

210

Ariane Boubaker

occurs within 30-45 min. Non-visualization of the gallbladder is a diagnostic criteria for acute cholecystitis, but may be related to an insufficient (<2 h) or too long (>24 h) fasting period, bile duct obstruction and severe hepatocellular dysfunction. A partial biliary obstruction may be suspected in case of delayed biliary to bowel transit beyond 60 min: other causes are opiates drugs, chronic cholecystitis, dysfunction of the sphincter of Oddi. Some medications are known to decrease gallbladder contraction and should be interrupted before HBS: morphine, atropine, nifedipine, indomethacin, benzodiazepine, octreotide.

Unilateral hydronephrosis due to pelviureteric junction stenosis is the most frequent congenital malformation. It is a benign disease that will spontaneously regress in up to 70% of the cases, and surgery should be done in selected children with renal function deterioration and/or complications (recurrent abdominal pain, infections) [21]. Diuretic renography using tubular tracers (99mTcMAG3, 123I-hippuran) is up to now the only non-invasive diagnostic modality that gives crucial information on renal function and urinary flow during a single procedure, even in neonates and young infants. It is simple, reproducible and there is no need to insert an intravenous canula or a bladder catheter. In a well-hydrated child, the simultaneous injection of radiotracer and furosemide (1mg/kg) will enable to have the child voiding at least one time during the examination, which is mandatory in assessing urinary flow (Fig. 4). The procedure is well standardized and can be used from 3-4 weeks of life, and repeated at 1-month interval during the first 6 months when needed [22]. There is no consensus on the use of camera-based methods to measure absolute renal function, and the EANM recommendation is to use isotopic clearance methods to measure glomerular filtration rate (GFR) and/or effective renal plasma flow (ERPF) when absolute function has to be checked [23]. In adults, a unilateral hydronephrosis may be diagnosed by chance

Urinary Tract Urinary Tract Dilation With the increased use of prenatal ultrasound, the number of neonates diagnosed with unilateral or bilateral mild to moderate pelvic dilation has tremendously increased during the past years. The recommendation algorithm for post-natal examinations aimed first at confirming the dilation with US performed at 2 and 7 days of life, and to exclude other renal abnormalities such as urethral valves, vesicoureteric reflux, multicystic dysplasia, duplex kidney and primary mega-ureter [20].

b

c

2’

d

5’

10’

100

50

80

40

60

30

40

20

20

10

Count rate (Cts/s)

1’

a

0

0

e

200

400

600

800

1000 1200

0

f

0

200

20’

400

600

800

1000 1200

Time (s)

Fig. 4 a-f. 123I-hippuran F0 dynamic renography performed during follow-up in a 9 months old infant with congenital left pelviureteric junction stenosis (PUJS) that was treated conservatively. a Consecutive 1-minute posterior views obtained 1, 2, 5, 10 and 20 minutes after injection of 9 MBq of 123I-hippuran and 1 mg/kg of furosemide show a normal right kidney and a preserved parenchymal extraction by the left dilated kidney. Bladder activity is present at 5 minute. The clearance from the right kidney is normal, whereas there is clearly a delayed urinary output from the left kidney. Delayed posterior static views obtained at 20 min (b), after micturition (c), and at 1 hour after injection (d) show a progressive dilution of radioactive urine in the dilated renal pelvis that persists unchanged despite change of position and micturition. Left kidney time-activity curve (e) is cumulative and confirms the preserved renal function of the parenchyma. The time-activity curve of the normal contralateral kidney (f) demonstrates both rapid extraction and secretion of the radiotracer with a time-to-peak less than 3 minutes

211

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders

on ultrasound or CT-scan performed for other reason and diuretic renography is valuable to assess renal function and urinary flow.

Urinary Tract Infection and Vesicoureteric Reflux Urinary tract infection (UTI) is an important routine clinical problem in paediatrics: its prevalence in children with fever ≥38.5 ranges from 1% to 20% depending mainly on sex and age, being more frequent in girls and in children aged less than 2 years [24]. Symptoms and signs are often non specific, the major challenge being to differentiate cystitis from acute pyelonephritis (APN). The risk for permanent renal damage is related to the delay between the onset of infection and treatment. 99mTc-DMSA scintigraphy is the gold standard for cortical renal imaging. It is easy to perform, does not require sedation or special patient preparation and has a sensitivity ranging from 80% to 100% to detect parenchymal defects due to infection, but does not allow to distinguish between APN and renal scars: a normal 99mTc-DMSA scintigraphy during the acute phase has a high negative predictive value for late scarring, whereas an abnormal scan during the acute phase is not predictive of long-term outcome [25]. Vesicoureteric reflux (VUR) is a common cause of recurrent UTI and can be treated either conservatively in neonates and

a

1’

2’

young infants (spontaneous resolution in 45-70% of cases) or surgically. Voiding cystourethrography (VCUG) is the reference method for diagnosis and grading of VUR. Direct radionuclide cystography delivers a lower radiation dose and is more sensitive than VCUG because of continuous acquisition during filling and voiding phases [26]. The disadvantages are the lack of anatomical information and the invasiveness and non physiologic condition due to bladder catheterisation. Indirect radionuclide cystography is performed after a dynamic renography with 99mTc-MAG3 or 123I-hippuran [27]. It is less sensitive and specific than direct cystography, but is more physiologic and non-invasive.

Acute Renal Failure Dynamic renography may help the clinician in patients presenting acute renal failure (ARF) by assessing the potential for renal function recovery. ARF due to acute tubular necrosis (ATN) has an overall good prognosis when compared to cortical necrosis or loss of nephronic mass due to recurrent cholesterol embolism for example. In case of ATN, diuretic renography will demonstrate progressive and symmetric tracer uptake in both kidney and delayed parenchymal retention (Fig. 5). Post-renal cause of ARF may be difficult to exclude in anuric patients,

5’

10’

20’

220

Count rate (Cts/s)

200 180 160 140 120 100 80 60 40

b

c

d

20 0

0

200

400

600

800

1000

1200

Time (s)

Fig. 5 a-d. Acute renal failure in a 46-year-old man after acute rhabdomyolisis. 123I-hippuran dynamic renography was performed to evaluate the potential for renal functional recovery. a Consecutive 1-minute posterior views obtained 1, 2, 5, 10 and 20 minutes after injection of 46 MBq of 123I-hippuran show a symmetric and heterogeneous renal parenchyma with increase of activity in the renal cortex despite appearance of radioactive urine in the bladder at 5 min pi. There is no urine retention in the renal pelvis or ureters. Static views obtained at 20 min (b) and 8 hours (c) after injection show persistent retention of the radiotracer in the renal parenchyma, typical of acute tubular necrosis (ATN). Time-activity curves of both left and right kidneys (d) show preserved initial parenchymal extraction with delayed secretion consistent with a good potential for renal function recovery

212

because ultrasound may not show significant dilation. In such patients, dynamic renography can be a sensitive tool by allowing delayed acquisition until 24 h pi and possibly demonstrate urinary retention in renal pelvis and/or ureters even at very low rate of urine output.

Renovascular Hypertension Although the prevalence of renovascular hypertension (RVH) in non selected patients is less than 1%, 15-45% among patients referred to a specialty center, for refractory hypertension will have RVH. When a stenosis becomes hemodynamically significant, the glomerular filtration rate (GFR) of the affected kidney decreases and the reninangiotensin system is activated in order to maintain the GFR by vasoconstriction (angiotensin II) of the efferent arterioles. Giving an angiotensin converting enzyme inhibitor (ACEI) will block the compensatory mechanism in the affected kidney, and provoke a decrease of GFR. The ACE-inhibitor dynamic renography will show typically a shift of the time-to-peak, a delayed production of urine and parenchymal retention of the tracer [28]. Oral captopril given 1 hour prior to renography has been used for many years, but variable absorption may occur. Intravenous enalapril (40 μg/kg, maximum 2.5 mg) infused over 3-5 minutes is more reliable. The test accuracy can be improved by administration of 20 mg furosemide during renography. A normal ACEI renography should be reported as low-probability for RVH disease, and a baseline study is not mandatory. An equivocal result should be reported when the baseline renography is abnormal and there is no significant change under ACEI. A high probability study (significant change under ACEI when compared to baseline) is a strong indicator of potential improvement after angioplasty or surgery. The accuracy of the test is significantly decreased in patients with poor renal function and/or a small shrunken kidney.

Renal Transplantation Live Kidney Donor Live kidney donation has become more frequent over the last years to cover the increasing numbers of patients waiting for renal graft. The first objective is not to harm a previously healthy person, the preoperative assessment of the potential donor is mandatory including psychological and biological examinations [29]. Ablation of a kidney will usually lead to a loss of about 25% of renal function if both kidneys are equally participating to the global function. The role of conventional nuclear medicine procedure in the live kidney donor is to assess renal function and to select the best functioning kidney to be left in the donor. Isotopic clearance using 51Cr-EDTA for GFR and 99mTc-MAG3 or 123I-hippuran for tubular extraction rate/effective renal plasma flow (ERPF) measurement have not been used extensively so far, but may be used in case of non conclusive creatinin clearance

Ariane Boubaker

measurements or in patients with borderline renal function [30]. Dynamic renography with tubular tracers is useful to evaluate the renal parenchyma, to calculate relative renal function and to assess urinary flow. When ultrasound and angio-CT do not reveal any significant anatomical abnormality, the choice of the donated kidney is usually based on the results of the renography: the kidney with the best function will be left in the donor. Evaluation of the Renal Graft The most common causes of early and late complications after renal transplantation are listed in Table 2. Dynamic renography with tubular tracers can be performed during the first 24 to 48 hours after transplantation and repeated if necessary, as no contrast medium injection, potentially nephrotoxic, is required. It allows verification of parenchymal function: signs of acute tubular necrosis (cumulative curve due to parenchymal retention of the tracer) will be present in most patients. Renal function should be quantified either by concomitant plasma clearance (GFR, ERPF) or by quantification based on the renogram. The procedure should be standardized in order to allow comparison of the results during follow-up. Acute rejection may be difficult to distinguish from acute tubular necrosis in the first days after operation: a worsening of the vascular phase, decrease of renal parenchymal extraction and worsening of parenchymal retention will be easier to diagnose if a baseline study is available. Renography is a very sensitive method to diagnose a urinary leak even at very low urine flow. If 99mTc-MAG3 is used, delayed acquisitions may be difficult to interpret because of hepato-biliary excretion: bowel activity may mask urinary activity. Whereas arterial thrombosis is a rare complication (<1%), renal artery stenosis (RAS) has been reported in up to 23% of renal allografts [1, 24]. ACE-inhibitor scintigraphy is useful to determine if systemic hypertension is dependent on the renin-angiotensin system, thus allowing proper clinical management (Fig. 6). When obstruction is suspected, the use of diuretic renography may help in a similar manner as in native kidneys.

Table 2. Common complications after renal transplantation Early complications (delayed graft function or failure to improve) Acute tubular necrosis Acute rejection Vascular occlusion Urinary leak/urinoma Hematoma Drug toxicity Late complications (decrease of renal function) Chronic allograft nephropathy (chronic rejection) Drug toxicity Renovascular hypertension (renal artery stenosis) Urinary tract infection Vesicoureteric reflux Obstructive uropathy (intrinsic, lymphocoele) Bladder dysfunction (postvoiding residue, small bladder volume)

213

Conventional Nuclear Medicine in the Evaluation of Gastrointestinal and Genitourinary Tract Disorders

a

1’

2’

5’

10’

20’

b 120

500

Count rate (Cts/s)

100

400

80 300 60 200 40 100

20

0

0

0

200

400

600

800

Count rate (Cts/s)

c

1200

0

200

400

600

800

1000

Time (s) 280

240

240

200

200

160

160

120

120

80

80

40

40

0

200

400

600

800

1000

1200

Time (s)

1200

Time (s)

280

0

d

1000

0

0

200

400

600

800

1000

1200

Time (s)

Fig. 6 a-d. Baseline and ACE-inhibitor 123I-hippuran dynamic renography performed 3 weeks after renal transplantation in a 53-year-old woman presenting a severe systemic hypertension and episodes of acute left cardiac failure. a At baseline, the renal transplant shows preserved parenchymal extraction (1 min pi) with prompt urine output (5 min) and rapid washout (10 and 20 min). b Under ACE-inhibitor (2.5 mg of enalapril intravenously over 5 minutes) there is a clear delay in urinary output (faint bladder activity at 10 min) and significant retention of the tracer in the renal cortex. c Time-activity curves obtained at baseline study show normal pattern of the transplanted kidney (grey curve), and rapid bladder filling (dotted curve). d Under ACE-inhibitor, time-activity curve of the renal transplant is clearly abnormal with preserved initial extraction phase and clearly delay of secretion (no time-to-peak, lack of descending curve). The bladder timeactivity curve (dotted line) shows no significant activity till 10 min pi and low urine output. Renovascular hypertension was diagnosed and surgery confirmed a narrowing of renal artery at the site of anastomosis

214

Ariane Boubaker

References 1. Ziessman HA, O’Malley JP, Thrall JH (2006) Nuclear medicine: The requisite. Radiology, 3rd edn. Elsevier Mosby, Philadelphia, PA, USA 2. Mariani G, Boni G, Barreca M et al (2004) Radionuclide gastrooesophageal motor studies. J Nucl Med 45:1004-1028 3. Nakajima K, Inali A, Hiramatsu T et al (2009) Esophageal transit scintigraphy and structured questionnaire in patients with systemic sclerosis with endoscopically proven reflux esophagitis. Ann Nucl Med October 24 Epub ahead of print 4. Boz AB, Aydin F, Celmeli F et al (2009) Does gastroesophageal reflux scintigraphy correlate with clinical findings in children with chronic cough? Nulc Med Commun 30: 802-806 5. Morigeri C, Bhattacharya A, Mukhopadhyay K et al (2008) Radionuclide scintigraphy in the evaluation of gastrooesophageal reflux in symptomatic and asymptomatic pre-term infants. Eur J Nucl Med Mol Imaging 35:1659-1665 6. International Pediatric Endosurgery Group (IPEG) Standard and Safety Committee (2009) IPEG guidelines for the surgical treatment of pediatric gastrooesophageal reflux disease (GERD). J Laparoendosc Adv Surg Tech A 19:x-xiii 7. Ravelli AM, Panarotto MB, Verdoni L et al (2006) Pulmonary aspiration shown by scintigraphy in gastrooesophageal refluxrelated respiratory disease. Chest 130:1520-1526 8. Abell TL, Camilleri M, Donohoe K et al (2008) Consensus recommendations for gastric emptying scintigraphy: a joint report of the American Neurogastroenterology and Motility Society and the Society of Nuclear Medicine. Am J Gastroenterol 103:753-763 9. Waseem S, Moshiree B, Draganov PV (2009) Gastroparesis: current diagnostic challenges and management considerations. World J Gastroenterol 15:25-37 10. Szarka LA, Camilleri M (2009) Methods for measurement of gastric motility. Am J Physiol Gastrointest Liver Physiol 296:461-475 11. Donohoe KJ, Maurer AH, Ziessman HA et al (2009) Procedure guideline for adult solid-meal gastric-emptying study 3.0. J Nucl Med Technol 3:196-200 12. Mariani G, Pauwels EKJ, AlSharif A et al (2008) Radionuclide evaluation of the lower gastrointestinal tract. J Nucl Med 49:776-787 13. Howarth DM (2006) The role of nuclear medicine in the detection of acute gastrointestinal bleeding. Semin Nucl Med 36:133-146 14. Ford PV, Bartold SP, Fink-Bennett DM et al (1999) Society of Nuclear Medicine procedure guideline for gastrointestinal

15. 16.

17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30.

bleeding and Meckel’s diverticulum scintigraphy 1.0. J Nucl Med 40:1226-1232 Kiratli PO, Aksoy T, Bozkurt MF et al (2009) Detection of ectopic gastric mucosa using 99mTc pertechnetate: review of the litterature. Ann Nucl Med 23:97-105 Rerksuppaphol S, Hutson JM, Oliver MR (2004) Ranitidineenhanced 99m technetium pertechnetate imaging in children improves the sensitivity of identifying heterotopic gastric mucosa in Meckel’s diverticulum. Pediatr Surg Int 20:323-325 Balon HR, Brill DR, Fink-Bennett DM et al (2001) Society of Nuclear Medicine procedure guideline for hepatobiliary scintigraphy 3.0. J Nucl Med 38:1654-1657 Nadel HR (1996) Hepatobiliary scintigraphy in children. Semin Nucl Med 26:25-42 Yang JG, Ma DQ, Peng Y et al (2009) Comparison of different diagnostic methods for differentiating biliary atresia from idiopathic neonatal hepatitis. Clin Imaging 33:439-446 Riccabona M, Avni FE, Blickman JG et al (2009) Imaging recommendations in paediatric uroradiology. Pediatr Radiol 39:891-898 Koff SA (2008) Requirements for accurately diagnosing chronic partial upper urinary tract obstruction in children with hydronephrosis. Pediatr Radiol 38(suppl 1):S41-S48 Gordon I, Colarinha P, Fettich J et al (2001) Guidelines for standard and diuretic renogram in children. Eur J Nucl Med 28:BP21-BP30 Piepsz A, Colarinha P, Gordon I et al (2001) Guidelines for glomerular filtration rate determination in children. Eur J Nucl Med 28:BP31-BP36 Boubaker A, Prior JO, Meuwly JY et al (2006) Radionuclide investigations of the urinary tract in the era of multimodality imaging. J Nucl Med 47:1819-1836 Piepsz A, Colarinha P, Gordon I et al (2001) Guidelines on 99m Tc-DMSA scintigraphy in children. Eur J Nucl Med 28:BP37-41 Fettich J, Colarinha P, Fischer S et al (2003) Guidelines for direct radionuclide cystography in children. Eur J Nucl Med Mol Imaging 30:B39-44 Gordon I, Colarinha P, Fettich J et al (2001) Guidelines for indirect radionuclide cystography. Eur J Nucl Med 28:BP16-20 Taylor AT, Blaufox MD, Dubovsky EV et al (1998) Society of Nuclear Medicine procedure guideline for diagnosis of renovascular hypertension 3.0. J Nucl Med 39:1297-1302 A report of the Amsterdam Forum on the Care of the Live Kidney Donor (2005) Data and Medical Guidelines. Transplantation 79:S53-S66 Prigent A (2008) Monitoring renal function and limitations of renal function tests. Semin Nucl Med 38:32-46

IDKD 2010-2013

PET in Hepatobiliary-Pancreatic Tumors Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri Department of Nuclear Medicine, University of Bologna, Bologna, Italy

Overview Primary liver cancer is a relatively uncommon entity, accounting for only 1-2% of malignant tumors, while cholangiocarcinoma is even rarer. However, in parts of Africa and Asia these tumors may constitute up to 2030% of all malignancies. Other benign and malignant hepatobiliary-pancreatic tumors are very rare; they include hepatoblastoma in infancy and early childhood, hemangiomas (the most common benign tumors), and angiosarcomas (vascular tumors). Hepatic adenomas, although also rare, may occur particularly in women taking oral contraceptives. In the USA, pancreatic cancer has an incidence of 12 per 100,000 men and women per year, and cancer of the pancreas is the third most common malignancy of the gastrointestinal tract and the fifth most common cause of cancer-related mortality. Islet cell tumors constitute 5-10% of pancreatic tumors and are usually referred to as neuroendocrine tumors of the pancreas. It is noteworthy that metastatic malignant lesions of the liver are common in clinical practice, with an incidence at least 20 times greater than that of primary carcinoma. By contrast, true cancers of the liver may escape clinical recognition, as they often occur in patients with cirrhosis (60-75% of cases) and the symptoms may initially suggest a progression of the underlying liver disease. The course of hepatocellular carcinoma and of cholangiocarcinoma is fatal and usually rapid, with surgical resection offering the only chance of cure. Nonetheless, the 5-year survival is low and only very few patients have resectable tumors at the time of presentation. Pancreatic carcinoma also has a very poor prognosis, and early diagnosis is extremely difficult. The physical examination rarely confirms the presence of localized pancreatic cancer. The initial symptoms that prompt further investigations are painless jaundice (in case of bile duct obstruction) and weight loss, but they depend on the site of the tumor within the pancreas and the degree of organ involvement. In advanced disease, palpable lymph nodes, hepatomegaly (due to metastasis), splenomegaly, ascites, and peripheral edema (portal vein obstruction), or an abdominal mass may be seen.

Common laboratory findings in liver cancers are anemia as well as elevated alkaline phosphatase and α-fetoprotein (AFP); in particular, very high levels of AFP occur in 70% of patients with hepatocellular carcinoma. Laboratory findings in pancreatic cancer are non-specific, since prolonged biliary obstruction may alter liver enzymes and cause abnormalities in vitamin-K-dependent clotting factors (due to malabsorption of fat-soluble vitamins), while pancreatic duct obstruction may result in pancreatic atrophy and subsequent impaired glucose tolerance or frank diabetes. Carbohydrate antigen 19-9 (CA 19-9) is the most commonly employed serological marker in pancreatic cancer. Although it has a sensitivity of 80%, serum levels are increased in many benign and malignant gastrointestinal conditions, such that the specificity for detecting pancreatic cancer is low.

Current Role of Imaging A number of imaging procedures are used to detect liver tumors, including ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), and angiography. US has been suggested in the screening of high-risk populations. It is widely available, reasonably accurate, and does not involve radiation exposure. It therefore should be considered as the first imaging procedure to use when hepatocellular carcinoma or cholangiocarcinoma is suspected. Although the use of contrast-enhanced US is recommended, and contrast enhancement is mandatory for CT, MRI is used with increasing frequency and is usually the most accurate diagnostic approach. All these imaging modalities can be used for primary diagnosis, staging purposes, and to identify tumor recurrence. In pancreatic cancer, US, CT, and MRI are widely used as well, but they are supplemented by MR-cholangiopancreatography (MRCP) and endoscopic US (EUS).

Positron Emission Tomography Positron emission tomography (PET) imaging allows the in vivo study of tissue metabolism, and thus demonstrates

216

malignant tumors as hypermetabolic lesions based on their increased tracer uptake. Nowadays, fluorine-18 radiolabeled fluorodeoxyglucose (18F-FDG) is the tracer compound of choice in clinical practice. FDG is initially carried into the cell by glucose transporters (GLUT-1), just as normal glucose, then is rapidly phosphorylated and trapped in the cells. Cancer cells are known to have increased anaerobic glycolytic activity and to express higher numbers of glucose transporters. Despite the wide clinical application of FDG, not all tumors show significantly increased metabolic activity on FDG-PET imaging. In particular, prostate cancer and neuroendocrine tumors may be difficult to study with FDGPET, as the exam lacks sensitivity. Therefore, in addition to FDG, several other tracers have been proposed, some of which are already used for clinical applications. For example, 68Ga-DOTA-somatostatin analogues have been successfully applied in the study of neuroendocrine tumors, both pancreatic and extra-pancreatic, while 11C-acetate has been proposed for hepatocellular carcinoma PET. In the past, the main limitation of PET imaging was its failure to provide anatomical data; however, with the introduction of a PET-CT hybrid system, morphological and metabolic imaging can be performed in a single session, thereby reducing false-positive findings and inconclusive studies and increasing diagnostic accuracy.

Role of FDG-PET in Liver Cancer The success achieved with FDG-PET in studies of hepatocellular carcinoma has been limited by a false-negative rate of 40-50%, with poor reliability especially in evaluations of well-differentiated cancers. Of the few studies available in the literature, most are retrospective and all of them report an inadequate sensitivity of FDG-PET in the detection of primary hepatocellular carcinoma. The true-positive rate of FDG was better in poorly differentiated tumors, with an increase in FDG uptake correlating with lower survival. While FDG-PET might be useful in the evaluation of extra-hepatic metastases, data supporting this possibility are limited. Instead, alternative tracers have been proposed in conjunction with FDG; however, the use of FDG alone may provide important prognostic information, especially in patients who are candidates for liver transplantation. FDG-PET has also been proposed for patients with cholangiocarcinoma, mainly in disease diagnosis and tumor staging. In the former, PET seems to be helpful in discriminating between malignant and benign lesions. However, the accuracy of FDG-PET is dependent on the lesion’s anatomical location, growth pattern, and pathological characteristics. For this reason, its application is limited to the detection of extra-hepatic, infiltrating, and mucinous cholangiocarcinomas. Moreover, due to its low sensitivity, PET provides complementary rather than confirmative information in the diagnosis of regional lymph node metastasis. FDG-PET, however, has shown high ac-

Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri

curacy in detecting unsuspected distant metastases. Its role in detecting cancer recurrence, monitoring treatment response, and predicting prognosis is still controversial.

Role of Acetate PET in Liver Cancer 11C-acetate

is used by cells as a precursor of membrane fatty acids but it can also be transformed into acetyl-CoA, entering the tricarboxylic acid cycle. It is thus processed as an intermediate in glucose catabolism and in membrane synthesis. The original application of 11C-acetate was not in oncology but in cardiology, because accumulation of the labeled compound in the myocardium is proportional to the level of fatty acids oxidation and thus reflects cardiac energy metabolism. Initially, the oncological application of 11C-acetate was as a choline analogue for the detection of prostate cancer; later, its use was expanded to the evaluation of liver masses, together with 18F-FDG studies. Preliminary results showed that 11C-acetate has good sensitivity in the detection of low-grade but not high-grade hepatic cancer, while FDG has the opposite behavior. In our experience, acetate-based PET is the preferred approach to study most hepatocellular carcinomas, especially in the differential diagnosis of masses not identified by conventional imaging and not suitable for biopsy, and for suspected recurrence of a hepatocellular carcinoma that was previously treated surgically.

Role of PET with Other Tracers in Liver Cancer 11C-choline

is another tracer suggested for studies of hepatocellular carcinoma; the results have been similar to those achieved with acetate. As only a very few studies have been published, a role for 11C-choline remains to be confirmed, but as for acetate, it may be useful in the evaluation of low-grade tumors. Finally, use of the tracer 18F-fluorothymidine has been proposed to assess the proliferation of hepatocellular carcinoma and cholangiocarcinoma.

Role of FDG-PET in Pancreatic Cancer Most of the literature describing the use of PET in pancreatic cancer refers to the tracer FDG. Normal pancreas has low glucose utilization, whereas in pancreatic cancer GLUT-1 transporters are over-expressed compared with normal tissue; therefore, the tumor/background FDG uptake ratio is high. The most important step in the initial approach to a patient suspected to have pancreatic carcinoma is to decide whether the lesion is benign or malignant. The major limitation of morphological imaging techniques is their inability to confidently characterize small as well as cystic lesions. In this setting, PET/CT may be helpful due to its high sensitivity (85-100%) and moderate

PET in Hepatobiliary-Pancreatic Tumors

specificity (67-90%). Several studies have reported that FDG-PET is more accurate than CT (the average sensitivity and specificity for CT is 82 and 75%, respectively). Regarding characterization of the lesion, the principal cause of false-positive findings at PET is inflammation due to chronic pancreatitis. However, the distribution of areas of avid FDG uptake within the parenchyma can guide the diagnosis, as diffuse high uptake in the whole pancreas is more often due to inflammation while focal uptake is a feature of pancreatic cancer. Unfortunately, morphological analysis is not specific because pancreatic cancer is sometimes accompanied by pancreatitis, with FDG uptake in tumor tissue likely related to the presence of inflammatory cells. In such cases it is essentially impossible to clearly separate the two components. Furthermore, cancer cells can diffusely infiltrate the entire pancreas, with pancreatic cancer manifesting as diffuse, high FDG uptake throughout the organ. By contrast, benign lesions such as autoimmune pancreatitis can assume a pattern of focal FDG uptake. Semiquantitative approaches are not very helpful due to the wide overlap in standardized uptake values (SUVs) between inflammation and malignant pancreatic disease. False-positive findings can also occur due to recent surgery or endoscopy, tissue inflammation after irradiation, abscess, autoimmune pancreatitis, massive lymphocyte infiltration, retroperitoneal fibrosis, hemorrhage in pancreatic pseudocysts, inflammatory pseudotumors, pancreatic tuberculosis, and focal high-grade dysplasia. Most false-negative results occur in cases involving both tumors of small size and elevated serum glucose levels. It should be noted that many such patients suffer from pancreatic insufficiency and diabetes; consequently, the high serum glucose levels compete with FDG for glucose transporters sites, reducing the sensitivity of FDG-PET in the detection of malignant lesions. Poor tumor cellularity, characteristic of scirrhous-type and cystic-type tumors as well as those featuring a desmoplastic reaction, is also an important cause of false-negative findings. The paucity of cells in these not-rare forms of pancreatic cancer is seen even in fairly large tumors. False-negatives on FDGPET studies also arise from pancreatic tumors such as mucinous or neuroendocrine tumors (NETs), which do not have high glucose metabolism. Regarding the detection accuracy of FDG-PET/CT, a recent meta-analysis suggested that although the addition of FDG-PET to the diagnostic work-up may enhance the diagnosis of pancreatic malignancy, the usefulness of this combined approach will vary depending upon the pre-test probability of the tumor, the results of CT, and the provider’s testing thresholds. Average sensitivity and specificity shift from 92 and 68% after a positive CT report, to 73 and 86% after a negative CT report, and 100 and 68% after an indeterminate CT report. In conclusion, the greatest benefit of PET in the differentiation of benign from malignant lesions is the possibility of excluding cancer without the need for biopsy or surgery, either of which may increase morbidity.

217

Regarding the use of FDG-PET for staging, about 40% of pancreatic cancers determined to be resectable by preoperative imaging turned out to be non-resectable at the time of surgery. Therefore, correct staging is the principal aim of imaging in pancreatic malignancies, to determine the appropriate management and the prognosis of the disease. For T staging, the poor spatial resolution of FDG-PET limits its utilization: anatomical imaging modalities such as multidetector CT, EUS, and MRCP are better suited to demonstrate the relationship between the tumor and the adjacent organs or vascular structures. At present, there are no data to support the usefulness of the hybrid-modality PET/CT in local T staging. Lymph node metastasis is one of the most important aspects of clinical management, providing an independent prognostic indicator for patients with pancreatic cancer. Unfortunately, both CT and PET/CT are of low sensitivity (30-40%) for lymph node detection, perhaps due to the strong radioactive scatter from the main tumor to peripancreatic small lymph nodes and to the low number of cancers cells in small metastatic lymph nodes. The principal cause of false-positive lymph nodes at FDG-PET is the presence of reactive locoregional lymphadenopathies following biliary instrumentation. After the diagnosis of pancreatic cancer has been established and local resectability of the tumor confirmed, the main objective of staging a pancreatic cancer is to identify those patients with distant metastasis because this group is currently excluded from surgical treatment. In this regard, the capability of whole-body scanning with a single examination at a single session is evidently an advantage of FDG-PET over other imaging modalities. Whole-body FDG-PET detection of distant metastasis or unexpected lesions changes patient management, is costsaving, and improves the patient’s quality of life by avoiding unnecessary surgery. Several studies have reported better diagnostic accuracy in the detection of distant metastasis using whole-body FDG-PET rather than other modalities, such as CT or US. The sensitivity of PET was between 80 and 90% and thus better than CT; the positive predictive value is very high for both modalities. Regarding metastatic disease, liver is the commonest organ to be affected, followed by the lungs and bone marrow. Direct tumor spread into the peritoneum is also not uncommon and often missed on conventional anatomical imaging. The accuracy of PET in detecting liver metastasis is almost the same as obtained with conventional imaging (94 and 90%, respectively). False-positive findings can occur due to intrahepathic cholestasis or in some types of inflammation, such as abscess or intrahepatic bile duct infection, mainly due to percutaneous transhepatic cholangiodrainage. False-negative PET findings may occur with lesions of small size but also because of the heterogeneity of hepatic parenchymal FDG uptake and respiratory motion artifacts. PET has also been suggested in the evaluation of response to therapy, and it may allow an earlier therapeutic response assessment than is possible with conventional

218

imaging. However, only very preliminary data are available, and larger prospective studies are necessary to confirm the role of PET in treatment response and to assess the correct time between therapy and PET post-treatment examination. With respect to recurrence, an elevated CA 19-9 has a positive predictive value of only 69% for pancreatobiliary malignancy. This means that >30% of patients with elevated CA 19-9 may have another tumor originating in another organ, or they may have no tumor at all. False-positive results have been associated with other pancreatobiliary disorders, such as gallstones, pancreatitis, inflammatory bowel disease, other liver disorders, pulmonary diseases such as pneumonia, and hydronephrosis. Therefore, if CA 19-9 is elevated despite negative findings at CT, then FDG-PET may have a role in detecting sites of recurrence, either locally or as metastases in the liver, lungs, peritoneum, and distant lymph nodes. FDG-PET might also be appropriate for excluding the presence of recurrence in patients with indeterminate findings using other imaging modalities.

Suggested Reading Bang S, Chung HW, Park SW et al (2006) The clinical usefulness of 18-fluorodeoxyglucose positron emission tomography in the differential diagnosis, staging, and response evaluation after concurrent chemoradiotherapy for pancreatic cancer. J Clin Gastroenterol 40:923-929 Breitenstein S, Apestegui C, Clavien PA (2008) Positron emission tomography (PET) for cholangiocarcinoma. HPB (Oxford) 10:120-121

Stefano Fanti, Anna Margherita Maffione, Vincenzo Allegri

Dierckx R, Maes A, Peeters M, Van De Wiele C (2009) FDG PET for monitoring response to local and locoregional therapy in HCC and liver metastases. Q J Nucl Med Mol Imaging 53:336-342 Eckel F, Herrmann K, Schmidt S et al (2009) Imaging of proliferation in hepatocellular carcinoma with the in vivo marker 18Ffluorothymidine. J Nucl Med 50:1441-1447 Higashi T, Saga T, Nakamoto Y et al (2003) Diagnosis of pancreatic cancer using fluorine-18 fluorodeoxyglucose positron emission tomography (FDG PET) – usefulness and limitations in “clinical reality”. Ann Nucl Med 17:261-279 Kauhanen SP, Komar G, Seppänen MP et al (2009) A prospective diagnostic accuracy study of 18F-fluorodeoxyglucose positron emission tomography/computed tomography, multidetector row computed tomography, and magnetic resonance imaging in primary diagnosis and staging of pancreatic cancer. Ann Surg 250:957-963 Kornberg A, Küpper B, Thrum K et al (2009) Increased 18F-FDG uptake of hepatocellular carcinoma on positron emission tomography independently predicts tumor recurrence in liver transplant patients. Transplant Proc 41:2561-2563 Lee TY, Kim MH, Park do H et al (2009) Utility of 18F-FDG PET/CT for differentiation of autoimmune pancreatitis with atypical pancreatic imaging findings from pancreatic cancer. Am J Roentgenol 193:343-348 Pakzad F, Groves AM, Ell PJ et al (2006) The role of positron emission tomography in the management of pancreatic cancer. Semin Nucl Med 36:248-256 Salem N, Kuang Y, Wang F et al (2009) PET imaging of hepatocellular carcinoma with 2-deoxy-2[18F]fluoro-D-glucose, 6-deoxy-6[18F] fluoro-D-glucose, [1-11C]-acetate and [N-methyl-11C]-choline. Q J Nucl Med Mol Imaging 53: 144-156 Seo S, Hatano E, Higashi T et al (2008) Fluorine-18 fluorodeoxyglucose positron emission tomography predicts lymph node metastasis, P-glycoprotein expression, and recurrence after resection in mass-forming intrahepatic cholangiocarcinoma. Surgery 143:769-777

IDKD 2010-2013

PET in Tumors of the Digestive Tract Thomas F. Hany Department of Radiology, Clinic and Policlinic of Nuclear Medicine, University Hospital Zurich, Zurich, Switzerland

Introduction The basic principle of positron emission tomography (PET) is the use of pharmaceuticals labeled with positron-emitting isotopes. These agents are such that they can be integrated into one of the body’s metabolic pathways. Positron-emitting isotopes are characterized by a beta plus-decay, in which a positron is emitted. This positron collides with any of the numerous shell electrons of neighboring atoms and the resulting annihilation produces two 511-keV gamma rays. These two photons are detected in coincidence by the PET scanner. Additionally, the use of an integrated PET/computed tomography (CT) machine allows PET and CT images of the patient to be acquired in the same imaging session. The clinically and most widely evaluated of the labeled pharmaceuticals is fluorine-18 fluoro-2-deoxy-D-glucose (18F-FDG). This glucose analogue is transported into the cell by specific transporters and phosphorylated by hexokinase to 18F-FDG-6 phosphate. The latter is inert to further metabolic processing or to transmembrane back-transport outside the cell and therefore accumulates intracellularly. The long physical half-life of FDG, around 110 min, makes this compound amenable for use as a metabolic marker, with applications in oncology, cardiology, neurology, and inflammation imaging. One of the most important advantages of PET/CT imaging is its broad anatomical coverage, since images of the patient extending from the head to the thighs are acquired routinely. All of the currently available data indicate that PET/CT is more sensitive and specific than either of its constituent imaging methods. With PET/CT, the most relevant effect is that the CT data frequently add specificity to the FDGPET data [1].

Esophagus, Stomach, and Small Bowel Esophagus The incidence of esophageal cancer has been increasing worldwide, particularly for adenocarcinomas of the lower esophagus and gastroesophageal junction [2]. Squamous

cell carcinoma is associated with dietary factors whereas adenocarcinoma is related to reflux. This distinction explains why the former occurs more proximally in the esophagus and the latter more distally. The major goals of the pre-operative evaluation of patients with esophageal cancers are to exclude distant metastases and to decide whether complete resection of the primary tumor and its lymphatic drainage (R0 resection) can be achieved (Fig. 1). Evaluation of locoregional lymph nodes is important in the determination of prognosis, since patients with peritumoral lymph node metastases have a significantly reduced overall survival, but the presence of such lymph nodes has no influence on either tumor resectability or therapeutic strategy [3]. Endoscopic ultrasonography (EUS) is considered the most accurate modality for local staging of esophageal cancer. Most studies show that the accuracy of EUS is approximately 85% for T staging and 75% for N staging. CT and PET/CT scanning are not as accurate as EUS for local staging, because neither technique can assess the layers of the esophageal wall, nor can small lymph nodes be reliably distinguished from adjacent tumor. Further improvement of locoregional staging is achieved by EUS used in combination with fine-needle aspiration (EUS-FNA). EUSFNA is reported to be more sensitive (83 vs. 29%; p <0.001) and more accurate (87 vs. 51%; p <0.001) than CT or EUS (87 vs. 74%; p=0.012) for nodal staging [4]. For initial staging, a meta-analysis comparing EUS, CT, and FDG-PET demonstrated sensitivities for regional lymph node metastases of 80, 50, and 57%, respectively, and specificities of 70, 83, and 85%, respectively. Interestingly, diagnostic performance did not differ significantly across these tests. For distant metastases, sensitivity and specificity were, respectively, 71 and 93% for FDG-PET and 52 and 91% for CT. The diagnostic performance of FDG-PET for distant metastases was shown to be significantly higher than that of CT, which was not significantly affected by study type and patient characteristics. For the detection of regional lymph node metastases, EUS is the most sensitive modality, whereas CT and FDG-PET are more specific. For the evaluation of distant metastases, FDG-PET has probably a higher sensitivity than CT [5].

220

a

Thomas F. Hany

b

Fig. 1 a, b. Initial staging in a 49-year-old patient with an advanced distal esophageal cancer. In the maximum-intensity-projection image (a), FDG uptake is seen in the distal esophagus and in two sites in projection to the right upper abdomen. These two lesions can be localized easily to an enlarged celiac trunk lymph node metastasis and an adrenal metastasis (b)

Curative treatment can be achieved by radiochemotherapy and/or surgery. FDG-PET/CT has demonstrated a high negative predictive value. In a study by Lordick et al., the effect of neo-adjuvant chemotherapy prior to surgery after 2 weeks of chemotherapy was compared to the initial scanning. An improved median survival in socalled metabolic responders compared to non-responders was demonstrated, whereas response was defined as a ≥35% decrease in standard uptake values (SUV). Remarkably, major histological remissions (<10% residual tumor) were noted in 58% of metabolic responders, but no histological response was seen in metabolic nonresponders. Accordingly, FDG-PET has a relatively high negative predictive value after treatment and vital tumor tissue is very likely to be present in the unchanged FDG uptake of the primary tumor [6]. These facts have found their way into the clinical practice guidelines of the National Comprehensive Cancer Network (NCCN), which recommends PET/CT for initial staging and after neo-adjuvant treatment.

Stomach: Malignant Disease In a review by Dassen et al. regarding the pre-operative diagnostic utility of FDG-PET/CT in gastric cancer, the

authors concluded that the technique has no role in the primary detection of gastric cancer due to its low sensitivity [7]. FDG-PET shows, however, slightly better results than CT in the evaluation of lymph node metastases in gastric cancer and could therefore have a role in preoperative staging of the tumor. Improvements in the accuracy of FDG-PET could be achieved by using PET/CT or PET tracers other than FDG, but these approaches need further investigation. The role of FDG-PET/CT is likewise limited in the detection of gastric cancer recurrence after curative tumor resection. In a study by Sim et al., in which 52 patients underwent restaging by PET/CT and contrast-enhanced CT (ceCT), the sensitivity was 68.4% (26/38) for PET/CT and 89.4% (34/38) for ceCT (p=0.057). The specificity was 71.4% (10/14) and 64.2% (9/14), respectively (p=1.0). Contrast-enhanced CT was more sensitive than PET/CT (p=0.039) in the detection of peritoneal seeding. Additional PET/CT combined with ceCT showed no further increase of positive predictive value regardless of tumor site. PET/CT was as sensitive and specific as ceCT in detecting a recurrence of gastric cancer, except in the case of peritoneal seeding. However, additional PET/CT combined with ceCT did not increase diagnostic accuracy in the detection of recurrent gastric cancer [8]. Accordingly, further studies are warranted to

221

PET in Tumors of the Digestive Tract

validate the role of PET/CT in the detection of gastric cancer recurrence; however, this approach adequately detects therapy responders at an early stage following neoadjuvant chemotherapy. Regarding primary non-epithelial tumors, lymphoma is an important disease beside gastrointestinal stroma tumors (discussed below). Primary gastric lymphoma (PGL) is díagnostically challenging due to the physiological activity of FDG in the stomach and variability in the degree of the tracer’s uptake in tumors of different histological subtypes [9]. In a study by Radan et al., PET/CT studies of 62 newly diagnosed PGLs were reviewed: 24 of low-grade mucosa-associated lymphoid tissue (MALT) type and 38 consisting of aggressive non-Hodgkin’s lymphoma (AGNHL). FDG avidity was present in 89% of the PGLs, including all AGNHL, but only in 71% of MALTtype lymphomas. Especially in AGNHL-PGL, FDG uptake can be differentiated from physiological tracer activity by its intensity but not by its pattern.

were better defined on PET/CT than on either PET or CT, when compared side-by-side. Normally, Response Evaluation Criteria in Solid Tumors (RECIST) criteria are used in the evaluation of therapeutic response, based on the change in tumor size. In a modified RECIST analysis, carried out in time intervals of 2 months for up to 28 months, Choi et al. used ceCT studies to evaluate changes in tumor size and density [14]. When PET, RECIST, and modified RECIST criteria were used in the analysis, a decrease in tumor size >10% or a decrease in tumor density >15% on CT was demonstrated to have a sensitivity of 97% and a specificity of 100% in identifying PET responders vs. 52 and 100% by RECIST. Good responders on CT at 2 months had significantly longer time to progression than patients who did not respond (p = 0.01). Therefore, the search for small changes in tumor size or density on CT may be a sensitive and specific method to assess the response of GISTs.

Small Bowel: Malignant Disease

Colorectal Carcinoma

Adenocarcinomas of the small bowel are rare tumors, with an incidence of around 3.7 per million people each year. The pathological (development from adenomatous polyps), genetic, and epidemiological features are very similar to those of adenocarcinomas of the large bowel but the prognosis is poor (reported 5-year-survival of 15-35%). Capsule endoscopy, CT enteroclysis, and ceCT are mainly used in disease diagnosis and work-up. These tumors do show clear FDG avidity and might be better evaluated by FDG-PET/CT, although no comprehensive date on the use of FDG-PET/CT are yet available. Carcinoid tumors, as malignant tumors of nonepithelial origin, have a similar incidence as adenocarcinomas (3.8 per million people each year). The different subtypes have a natural behavior ranging from benign to high-grade malignancies. Similar to neuroendocrine tumors (NET) of the pancreas, the diagnostic work-up consists of nuclear medicine studies, including somatostatin receptor scintigraphy combined with contrastenhanced CT (SPECT/CT) or Ga-68 octreotide (68GaDOTA-NOC) labeling, since more than 80% of carcinoid tumors express somatostatin receptors. Detection rates are therefore similar to those of NET of the pancreas [10].

Colorectal carcinoma is the most important cause of death due to cancer in the western world, after bronchial carcinoma [15]. About 70% of patients have curable resectable tumor at initial diagnosis and are treated with curative intent. Approximately 50% of colon cancer patients will present with hepatic metastases, either at the time of initial diagnosis or as a result of recurrence [16]. From a diagnostic perspective, colon cancer is often evaluated together with rectal cancer as a single group; however, especially deep rectal cancer has a clearly different pathway of locoregional and distant metastases. In deep rectal cancer without invasion of the anal canal, lung metastases are more prevalent than metastases to the liver (Fig. 2). In very deep rectal cancer, lymphatic drainage is no longer towards para-rectal/meso-rectal lymph node stations but to inguinal sites, analogous to anal squamous cancer. In these cases, the inguinal lymph nodes and possibly, the external iliac region have to be evaluated, since clinical inspection, ultrasound, and fine-needle aspiration cost-effectively lead to correct staging.

Gastrointestinal Stromal Tumors Gastrointestinal stromal tumors (GIST) are mesenchymal tumors that in approximately 90% of patients originate in the stomach and small intestine. FDG-PET is able to show early effects in patients undergoing treatment with imatinib mesylate (Glivec; Novartis, Switzerland) [11]. In two recent short-term follow-up studies, patients without FDG uptake after the start of treatment had a better prognosis than those with residual activity not demonstrated with ceCT [12, 13]. Furthermore, lesions

Initial Staging Two studies in which FDG-PET alone was used for initial staging demonstrated the high sensitivity of this approach in the detection of primary tumor (100 and 96%) and distant metastases (87 and 78%) but a low sensitivity (29 and 29%) in lymph node staging [17,18]. In a study by Veit-Haibach et al., 47 patients underwent whole-body PET/CT colonography one day after colonoscopy [13]. Compared with optimized abdominal CT staging alone, PET/CT colonography was significantly more accurate in defining TNM stage (difference, 22%; 95% CI, 9-36%; p = 0.003), mainly based on a more accurate definition of the T stage. Differences in defining N stage were not detected between PET/CT colonography and CT alone

222

a

Thomas F. Hany

b

Fig. 2 a, b. A 64-year-old male patient with a histologically proven deep rectal cancer. In the maximum-intensity-projection image (a), focal uptake is seen just below the bladder. Unexpectedly, additional uptake is seen is in the right thorax. This lesion was identified as a singular lung metastasis, typical for deep rectal cancer without evidence of liver metastases (b)

when the threshold for malignant nodes was 0.7 cm but were detected at a threshold of 1 cm. Differences were not detected in defining M stage separately or when the accuracies of PET/CT colonography were compared with CT + PET. PET/CT colonography affected consecutive therapy decisions in 4 patients (9%; 95% CI, 2.4-20.4%) compared with conventional staging (CT alone and colonoscopy). The combination of FDG-PET/CT in conjunction with a dedicated ceCT protocol could be of interest as a single-step staging procedure.

Recurrent Disease Standard patient work-up for the detection of recurrence and metastases in colorectal cancer includes regular clinical examinations, CT scans, colonoscopy, and, usually, the measurement of tumor markers such as CEA (Fig. 3). However, this approach lacks specificity and may result in diagnostic and therapeutic delays. Serological tumor markers are useful, although it has been shown that the serum CEA level has only 60-70% sensitivity for the detection of colorectal cancer recurrence [19]. The mor-

phology-based information obtained with CT does not permit a distinction between post-surgical changes and tumor recurrence, nor can it detect tumor involvement of normal-sized lymph nodes [20]. Colonoscopy is only useful in the detection of local recurrence. The suitability of FDG-PET in identifying recurrence and metastases has been confirmed in several studies. Despite the obvious advantage of PET/CT over PET alone, a dedicated ceCT is often requested by clinicians. Soyka et al. found that cePET/CT, as a single-step examination, has the same diagnostic confidence and impact as a sequential approach, with ceCT first and non-cePET/CT afterward [21]. Although the lesion detection rate on ceCT images is high, evaluation by ceCT alone can be challenging because of the possibility of inconclusive results that require further diagnostic evaluation (56% of our patient population). The reason for this is predominantly related to specificity issues regarding the structural abnormalities depicted by this modality. Consequently, patients with inconclusive ceCT findings are now frequently referred for further evaluation with 18F-FDG-PET/CT. More importantly, the same study showed that in 21% of the patients with

223

PET in Tumors of the Digestive Tract

a

b

Fig. 3 a, b. A 71-year-old male patient with a history of colon cancer in whom an increase in serum CEA was detected. In the maximumintensity-projection image (a), focal uptake is seen in projection to the liver. The axial contrast-enhanced CT and fused PET/CT image demonstrate a singular liver metastasis in the right liver lobe, segment VII (b)

apparently conclusive findings on ceCT, the addition of non-cePET/CT information led to appropriate changes in therapy. In clinical routine, in those cases in which ceCT was judged to be conclusive, the patient would not routinely be referred for further evaluation with 18F-FDGPET/CT. However, if cePET/CT had been used as the initial imaging modality, 65% of the patients would have had a clear benefit, including changes in management and in diagnostic confidence. Therefore, one could argue that cePET/CT should be the first-line diagnostic tool in the restaging of colorectal cancer. Nonetheless, one could also argue that in 35% of the patients both the radiation exposure and the costs of the procedure would have been futile. However, the former argument holds true only if ceCT and non-cePET/CT are performed within 2-4 weeks. In general, surgeons insist on ceCT studies not older than 4 weeks before taking a patient into the operating room. Thus, another, additional scan with contrast enhancement (ceCT or cePET/CT) would be needed in the majority of patients. Post-surgical and radiotherapy-induced changes in the small pelvis are the most challenging for morphological imaging studies in recurrent rectal cancer, since tumor recurrence cannot be differentiated from benign scar tissue. In a study by Even-Sapir et al., PET/CT was used to distinguish benign from malignant pre-sacral abnormalities. The sensitivity, specificity, positive predictive value, and negative predictive value were 100, 96, 88, and 100%, respectively, and PET/CT findings were clinically relevant in 47% of 62 patients [22]. A

comparison with other “conventional” imaging studies was not performed. In our own study, by Seltzner et al., the diagnostic value of ceCT and non-enhanced PET/CT was prospectively evaluated and compared in 76 patients referred for pre-operative evaluation for liver resection for metastatic colorectal cancer [23]. Extrahepatic disease was missed by ceCT in one-third of the patients (sensitivity 64%), while PET/CT failed to detect extrahepatic lesions in only 11% (sensitivity 89%; p=0.02). New findings derived from PET/CT resulted in a change in the therapeutic strategy in 21% of the patients. This study also demonstrated the well known limitation in spatial resolution of around 4-6 mm of PET imaging, since small tumours (e.g., <5 mm) were often not detected. Also, patients who underwent chemotherapy within the month prior to PET/CT had a high incidence of false-negative results. Alternatively, this effect might be used as a predictor of success in neo-adjuvant chemotherapy before resection. The above-mentioned studies clearly demonstrate the advantages of PET/CT imaging in colorectal cancer.

Therapy Response Assessment In general, a decrease or reduction to normal of FDG uptake levels in tumor tissue is correlated with response to treatment. Systematic reviews have only been performed in patients with rectal cancer before and after neo-adjuvant radio-chemotherapy. In a study by Kalff et al., the prognostic information obtained from the degree

224

of change in tumor FDG uptake induced by chemoradiation before radical curative surgery was evaluated in patients with T3/T4 rectal cancer. In 34 consecutive patients with T3/T4 Nx M0 rectal cancer, FDG-PET was performed at baseline and after radiochemotherapy before planned curative surgery. The change in FDG uptake was measured by SUV as well as by visual grading as complete (CMR), partial (PMR), or no metabolic response. Histopathological findings were available in 30 patients. After an estimated median 3.1 years of followup, all 17 CMR patients were free of disease. The PET response was highly significantly associated with overall survival duration (p<0.0001) and time to progression (p<0.0001). Pathological complete response was the only other statistically significant prognostic factor (p<0.03). The percentage of maximum SUV change after chemoradiation was not predictive of survival in PMR patients. Based on a simple qualitative assessment, post-chemoradiation 18F-FDG-PET provides good medium-term prognostic information in patients with advanced rectal cancer undergoing radical surgery with curative intent [24].

Conclusions FDG-PET/CT imaging is becoming more established in the work-up of several abdominal malignancies of the gastrointestinal tract. The main advantage lies in its comprehensive evaluation of the patient, including all body compartments, and therefore the detection of pivotal, therapy-deciding lesions. The performance of FDG-PET/CT in the evaluation of primary tumors of the gastrointestinal tract is characterized by a high sensitivity in the detection of distant metastases. Secondary liver tumors such as gastrointestinal metastases are detected by FDG-PET/CT at a high rate, making this imaging technology a primary tool in the evaluation of patients with suspicion of recurrent colon cancer. Furthermore, full integration of ceCT protocols improves diagnostic confidence and reduces the sometimes cumbersome diagnostic pathway for patients. FDG-PET/CT can be reliably used for therapy response assessment. New tracers, such as Ga-68-DOTA-TATE or 18F-DOPA, will bring significantly improved diagnostic confidence in the notoriously difficult evaluation of patients with NET of the small bowel.

References 1. von Schulthess GK, Steinert HC, Hany TF (2006) Integrated PET/CT: current applications and future directions. Radiology 238:405-422 2. Wei JT, Shaheen N (2003) The changing epidemiology of esophageal adenocarcinoma. Semin Gastrointest Dis 14: 112-127 3. Lerut T, Coosemans W, Decker G et al (2001) Cancer of the esophagus and gastro-esophageal junction: potentially curative therapies. Surg Oncol 10:113-122

Thomas F. Hany

4. Vazquez-Sequeiros E, Wiersema MJ, Clain JE et al (2003) Impact of lymph node staging on therapy of esophageal carcinoma. Gastroenterology 125:1626-1635 5. van Vliet EP, Heijenbrok-Kal MH, Hunink MG et al (2008) Staging investigations for oesophageal cancer: a meta-analysis. Br J Cancer 98:547-557 6. Lordick F, Ott K, Krause BJ et al (2007) PET to assess early metabolic response and to guide treatment of adenocarcinoma of the oesophagogastric junction: the MUNICON phase II trial. Lancet Oncol 8:797-805 7. Dassen AE, Lips DJ, Hoekstra CJ et al (2009) FDG-PET has no definite role in preoperative imaging in gastric cancer. Eur J Surg Oncol 35:449-455 8. Sim SH, Kim YJ, Oh DY et al (2009) The role of PET/CT in detection of gastric cancer recurrence. BMC Cancer 9:73 9. Radan L, Fischer D, Bar-Shalom R et al (2008) FDG avidity and PET/CT patterns in primary gastric lymphoma. Eur J Nucl Med Mol Imaging 35:1424-1430 10. Ambrosini V, Tomassetti P, Castellucci P et al (2008) Comparison between 68Ga-DOTA-NOC and 18F-DOPA PET for the detection of gastro-entero-pancreatic and lung neuroendocrine tumours. Eur J Nucl Med Mol Imaging 35:14311438 11. Joensuu H, Roberts PJ, Sarlomo-Rikala M et al (2001) Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N Engl J Med 344:1052-1056 12. Goerres GW, Stupp R, Barghouth G et al (2004) The value of PET, CT and in-line PET/CT in patients with gastrointestinal stromal tumours: long-term outcome of treatment with imatinib mesylate. Comparison of PET, CT, and dual-modality PET/CT imaging for monitoring of imatinib (STI571) therapy in patients with gastrointestinal stromal tumors. Eur J Nucl Med Mol Imaging 4:4 13. Veit-Haibach P, Kuehle CA, Beyer T et al (2006) Diagnostic accuracy of colorectal cancer staging with whole-body PET/CT colonography. JAMA 296:2590-2600 14. Choi H, Charnsangavej C, Faria SC et al (2007) Correlation of computed tomography and positron emission tomography in patients with metastatic gastrointestinal stromal tumor treated at a single institution with imatinib mesylate: proposal of new computed tomography response criteria. J Clin Oncol 25: 1753-1759 15. Bade MA, Ohki T, Cynamon J, Veith FJ (2001) Hypogastric artery aneurysm rupture after endovascular graft exclusion with shrinkage of the aneurysm: significance of endotension from a “virtual,” or thrombosed type II endoleak. J Vasc Surg 33:1271-1274 16. Clarke MP, Kane RA, Steele G Jr et al (1989) Prospective comparison of preoperative imaging and intraoperative ultrasonography in the detection of liver tumors. Surgery 106:849855 17. Abdel-Nabi H, Doerr RJ, Lamonica DM et al (1998) Staging of primary colorectal carcinomas with fluorine-18 fluorodeoxyglucose whole-body PET: correlation with histopathologic and CT findings. Radiology 206:755-760 18. Kantorova I, Lipska L, Belohlavek O et al (2003) Routine (18)F-FDG PET preoperative staging of colorectal cancer: comparison with conventional staging and its impact on treatment decision making. J Nucl Med 44:1784-1788 19. Zervos EE, Badgwell BD, Burak WE Jr et al (2001) Fluorodeoxyglucose positron emission tomography as an adjunct to carcinoembryonic antigen in the management of patients with presumed recurrent colorectal cancer and nondiagnostic radiologic workup. Surgery 130:636-643; discussion 643634 20. Goldberg RM, Fleming TR, Tangen CM et al (1998) Surgery for recurrent colon cancer: strategies for identifying resectable recurrence and success rates after resection. Eastern Cooperative Oncology Group, the North Central Cancer Treatment

PET in Tumors of the Digestive Tract

Group, and the Southwest Oncology Group. Ann Intern Med 129:27-35 21. Soyka JD, Veit-Haibach P, Strobel K et al (2008) Staging pathways in recurrent colorectal carcinoma: is contrast-enhanced 18F-FDG PET/CT the diagnostic tool of choice? J Nucl Med 49:354-361 22. Even-Sapir E, Parag Y, Lerman H et al (2004) Detection of recurrence in patients with rectal cancer: PET/CT after abdominoperineal or anterior resection. Radiology 232:815-822

225

23. Selzner M, Hany TF, Wildbrett P et al (2004) Does the novel PET/CT imaging modality impact on the treatment of patients with metastatic colorectal cancer of the liver? Ann Surg 240:1027-1034; discussion 1035-1026 24. Kalff V, Duong C, Drummond EG et al (2006) Findings on 18F-FDG PET scans after neoadjuvant chemoradiation provides prognostic stratification in patients with locally advanced rectal carcinoma subsequently treated by radical surgery. J Nucl Med 47:14-22

IDKD 2010-2013

Tumors of the Adrenergic System: Imaging and Therapy Cornelis A. Hoefnagel Department of Nuclear Medicine, The Netherlands Cancer Institute, Amsterdam, The Netherlands

Targeting of Neuroendocrine Tumors Neuroendocrine tumors (NET) include pheochromocytoma, neuroblastoma, carcinoid, paraganglioma, chemodectoma, medullary thyroid carcinoma, islet cell tumors, gastrinoma, small cell lung cancer, melanoma, and Merkel cell tumor. NET vary considerably in their clinical presentation, location, and histology but share a common embryonic tissue origin, i.e., the neural crest. In addition, all of these tumors express several unique characteristics that may be exploited in the targeting of radiopharmaceuticals for diagnostic as well as therapeutic purposes [1]. The specific targeting of NET may be achieved via the metabolic route (meta-iodo-benzylguanidine, MIBG), through receptor binding (peptides), or immunologically (antibodies). In metabolic targeting, 123I- or 131I-MIBG and 111Inpentetreotide are sensitive and highly specific tracers, and therefore the most widely used. In addition, comparative studies have demonstrated a complementary role for these procedures [1]. An active uptake mechanism at the cell membrane and by neurosecretory storage granules in the cytoplasm of neural crest tumors is responsible for the uptake and retention of 123I- or 131I-MIBG, respectively. Although the radiopharmaceutical may be released from the granules, its specific re-uptake maintains a prolonged intracellular concentration; this is in contrast to nonadrenergic tissues, in which uptake relies on passive diffusion only. The high selective accumulation results in high tumor/non-tumor ratios. However, it should be kept in mind that a number of drugs may interfere with the uptake and/or the retention of MIBG [2]. Somatostatin analogs target peptide receptors on the cell surface and therefore can be used in the diagnosis and therapy not only of several NET but also in other tumors in which peptide receptors have been demonstrated autoradiographically [3]. Unlike MIBG- and antibodybased targeting, the peptides are not specific for neural crest tumors. 111In-pentetreotide is mostly used for diagnostic scintigraphy, while 90Y-labeled octreotide or lanreotide and 177Lu-labeled octreotate are administered for targeted therapy. Other peptides used for diagnostic scintigraphy include 111In-lanreotide, 99mTc-depreotide, and 123I-vasoactive intestinal peptide.

Radiolabeled monoclonal antibodies target antigens on the cell surface of some NET. Examples are the murine monoclonal antibodies 131I-UJ13A and 131I-3F8, used in the 1980s, and, more recently, chimeric antibodies, such as 131I-chCE7, in the treatment of neuroblastoma. Radioiodinated anti-CEA antibodies/fragments may target CEA-producing medullary thyroid carcinoma (MTC). More recently, bispecific anti-DTPA/anti-CEA immunoconjugates have been shown to improve tumor/non-tumor ratios and to prolong retention in MTC [4]. Non-specific tumor-seeking agents, such as 67Ga-citrate, 201Tl-chloride, 99mTc-pentevalent DMSA, and 99mTc-sestamibi, are also used for diagnostic scintigraphy. Although these tracers may be sensitive, they lack specificity and therefore cannot be used therapeutically. New specific tracers for positron emission tomography (PET) imaging have recently been developed, for example, 124I-MIBG, 11C-hydroxyephedrin, 6-[18F]-fluorodopamine, and 68Ga-octreotide (the latter from a generator). These agents combine great sensitivity and specificity with the high-quality hybrid imaging provided by PET/computed tomography (CT). These techniques will significantly improve the quality and reliability of diagnostic imaging of NET. While 18F-fluorodeoxyglucose (18F-FDG) does not have this high degree of specificity, it may be used to detect dedifferentiating, rapidly growing tumors [5]. As the role of radiolabeled peptides for imaging is discussed elsewhere in this volume, the remainder of this chapter focuses on the diagnosis and therapy of NET using 123I-MIBG or 131I-MIBG.

Techniques Currently, there are several nuclear medicine procedures involving the use of MIBG in the diagnosis and therapy of NET. Diagnostic scintigraphy of the whole body uses either 123I-MIBG (γ-emitter, t 1/2fys = 13 h, photon energy 159 Kev) or 131I-MIBG (β/γ-emitter, t1/2fys = 8 days, photon energy 364 KeV). 123I-MIBG scintigrams are of better quality and the results are more readily available, whereas 131I-MIBG enables delayed imaging over several days.

227

Tumors of the Adrenergic System: Imaging and Therapy

Single photon emission tomography (SPECT) or SPECT/ CT using 123I-MIBG provides improved detection in addition to accurate localization of NET sites by hybrid (fusion) imaging. Positron emission tomography (PET) or PET/CT using novel, specific PET tracers (see above) is currently the most accurate diagnostic modality for NET. Following re-injection of 123I-MIBG, an intraoperative gamma-probe can be used to guide the surgical resection of NET. Finally, radionuclide therapy with high doses of 131I-MIBG is an option for tumors retaining a high concentration of the radiopharmaceutical for a prolonged period of time (as demonstrated on the diagnostic scintigram). The procedure can be monitored by post-therapy total-body scintigraphy with or without 131I-MIBG SPECT/CT fusion imaging.

Diagnostic Imaging

rule out Wilms tumor, Ewing sarcoma, rhabdomyosarcoma, osteosarcoma, and malignant lymphoma [9]. A recent prospective trial of 123I-MIBG scintigraphy in 100 neuroblastoma patients showed an overall sensitivity of 88%, which increased slightly, to 91%, by the addition of SPECT. In patients with a recent diagnosis of neuroblastoma, the sensitivity of the procedure was 93% and its specificity 92% [10]. At present, 123I/131I-MIBG scintigraphy has an established role in the staging of disease and as a parameter in the response criteria. Discrepant findings of MIBG and bone scintigrams have been described, in favor of either the former or the latter. Since a positive finding on an MIBG scintigram is the more specific one, the initial use of 123I/131I-MIBG is preferred, but complementary bone scintigraphy may be indicated. Radioimmunoscintigraphy with 131I-3F8 [11] and, more recently, radioiodinated chimeric antibodies directed against neuroblastoma have yielded results complementary to those obtained with MIBG imaging [12].

Pheochromocytoma Carcinoid Tumors The role of 123I/131I-MIBG-scintigraphy in the diagnosis of pheochromocytoma is not as a screening test. Instead it is the best initial procedure in patients who, on the basis of a clinical or familial history, are suspected of having pheochromocytoma and with high plasma levels or urinary excretion rates of catecholamines and catecholamine metabolites. The cumulative sensitivity of 123I/ 131I-MIBG scintigraphy in potential pheochromocytoma patients is 88% [1]. Although CT and magnetic resonance imaging of adrenal masses provide the surgeon with better anatomical detail, a positive 123I/131I-MIBG scan is a highly specific finding. The scintigraphic technique is superior for localizing extraadrenal, recurrent, multifocal, and malignant disease [6]. Although cumulative results of 111In-pentetreotide scintigraphy also show a sensitivity for this technique of 88% in pheochromocytoma, a disadvantage in the detection of an adrenal tumor is the renal, hepatic, and splenic accumulation of the tracer. A recent prospective multicenter evaluation of 123I-MIBG in 150 patients with confirmed or suspected pheochromocytoma or paraganglioma demonstrated a sensitivity of 82-88% and a specificity of 82-84%. In this series, the addition of SPECT hardly affected these values [7].

Neuroblastoma The cumulative findings of 131I-MIBG scintigraphy reported in the literature [1] indicate that 92% of neuroblastomas concentrate MIBG. 123I/131I-MIBG imaging allows the detection of primary tumors, residual or recurrent disease, and metastases, regardless of localization, in a single procedure. When used together with urinalysis for catecholamine metabolites, MIBG imaging is the most sensitive and highly specific indicator of neuroblastoma [8]. The uptake of MIBG is so tissuespecific that in a child presenting with a tumor of unknown origin 123I/131I-MIBG-scintigraphy can noninvasively establish the diagnosis of neuroblastoma and

The cumulative sensitivity of 111In-pentetreotide scintigraphy in patients with carcinoid (86%) is higher than that of 131I-MIBG scintigraphy (70%) [1], thus favoring use of the former in the initial diagnosis. 131I-MIBG scintigraphy should not be used as a screening test for the initial diagnosis of carcinoid, nor can it be relied upon to exclude disease. However, the combined use of the two techniques are possibly of therapeutic interest. A positive 111In-pentetreotide scintigram may predict a response to palliative octreotide therapy and indicate the feasibility of 90Y-octreotide or 177Lu-octreotate therapy, whereas in the work-up of patients with proven carcinoid 131I-MIBG scintigraphy allows the selection of those patients who may benefit from 131I-MIBG therapy.

Other Neuroendocrine Tumors In MTC, radioimmunoassays of serum calcitonin and CEA levels are currently the most sensitive parameters in diagnosis and follow up, but in recent years many nuclear medicine procedures have emerged that allow the disease to be localized. This is especially the case for the detection of liver metastases and adrenal pheochromocytomas. Total-body scintigraphy with SPECT using 201Tl-chloride, 99mTc-pentavalent DMSA, and/or 99mTc-sestamibi and PET using 18F-FDG are best used initially. These are all relatively non-specific procedures, but their sensitivities are in the range of 80-90%. 111In-pentetreotide and radiolabeled anti-CEA antibodies, with sensitivities of 60-70%, may have a complementary role. 131I-MIBG has the lowest sensitivity (35%) and thus should only be used once MTC metastases have been confirmed, to evaluate its potential therapeutic role [1,13]. A comparison of the results of 131I-MIBG and 111Inpentetreotide in a variety of other neural crest tumors [1] showed highest sensitivities for 111In-pentetreotide in

228

Cornelis A. Hoefnagel

availability and feasibility of other treatment modalities as well as the patient’s condition determine the indication. The principle indications for 131I-MIBG therapy are malignant pheochromocytoma and paraganglioma, neuroblastoma stage III and IV, MTC, and symptomatic, metastatic carcinoid tumors [14]. Contraindications for radionuclide therapy in general are: pregnancy, continued breast feeding, myelosuppression, and renal failure. In addition, relative contraindications apply to those patients whose condition is unstable or who fail to understand or cooperate with the radiation protection guidelines, or if isolation facilities are lacking.

paraganglioma (97%), small cell lung cancer and, to a lesser degree, endocrine gastroenteropancreatic (GEP) tumors, Merkel cell tumor, melanoma, and functioning pituitary tumors. 123I/131I-MIBG scintigraphy is useful for the detection of ganglioneuroma, paraganglioma, and chemodectoma; it is of limited use in pancreatic islet cell tumors, retinoblastoma, schwannoma, and Merkel cell tumors. It has no place in the diagnosis of small cell lung cancer and melanoma [1]. Although high sensitivities for 111In-pentetreotide scintigraphy have been reported in non-neural crest tumors, e.g., non-small cell lung cancer, brain tumors, and lymphomas, as well as in granulomatous and autoimmune diseases, 131I-MIBG scintigraphy, as a highly specific procedure for neural crest tumors, is virtually always negative in non-neural crest tumors [1, 9].

Malignant Pheochromocytoma and Paraganglioma The objective of 131I-MIBG therapy includes objective tumor volume reduction (complete or partial response), tumor arrest (stabilization of previously progressive disease), a reduction of the tumor’s metabolic function (as the prognosis in pheochromocytoma may depend on the long-term consequences of catecholamine hypersecretion, this may actually prolong survival), and palliation of symptoms (e.g., hypertension, bone pain, sweating, constipation) [14]. In 1991, the results of 131I-MIBG therapy in 117 patients with pheochromocytoma treated in 14 centers worldwide were pooled [15]. An objective response, defined as a >50% decrease in catecholamine excretion, a >50% reduction of tumor volume, or, if lesions could not be measured, significant scintigraphic improvement, was determined in 56% of the patients. The response of softtissue metastases was better than that of skeletal metastases. In addition, a subjective improvement of symptoms, decrease in blood pressure, and pain relief were achieved in >60% of the patients. Long-lasting objective responses have been reported also in malignant paraganglioma, in secreting and in non-secreting types [16]. These tumors may be treated either with 131I-MIBG or 90Y-/177Lu-labeled octreotide/octreotate. At a European Association of Nuclear Medicine (EANM) Radionuclide Therapy Committee workshop on 131I-MIBG therapy, held in 1999, treatment results were gathered for 534 patients with neural crest tumors, including 77 patients with malignant pheochromocytoma and 34 with paraganglioma (Table 1). The cumulative objective response rates with respect to tumor volume

Rationale for Using MIBG and Somatostatin-Receptor Imaging Procedures Scintigraphy using 111In-pentetreotide is the best initial procedure in patients with carcinoid, endocrine gastroenteropancreatic tumors, and (benign) paraganglioma. 131I-MIBG can be reserved to evaluate the feasibility of therapy and for radionuclide treatment of these tumors. 123I/131I-MIBG scintigraphy remains the best initial procedure for pheochromocytoma, neuroblastoma, and malignant paraganglioma, because of its high sensitivity/ specificity as well as its effective therapeutic application in these conditions. Both tracers play a modest role in MTC, i.e., complementary to the more sensitive but non-specific tracers and in the evaluation of the various therapeutic options.

Radionuclide Therapy Indications/Contraindications for Therapy Any malignant neural crest tumor showing sufficient uptake and prolonged retention of 131I-MIBG on a diagnostic tracer study (ideally >1% of the administered dose, depending on tumor volume) is a candidate for radionuclide therapy. Apart from tracer concentration, the Table 1. Pooled results of October 1999)

131I-MIBG

Disease Pheochromocytoma Paraganglioma Neuroblastoma Medullary thyroid carcinoma Carcinoid Other Total

therapy in neural crest tumors (EANM Radionuclide Therapy Committee Workshop, Barcelona,

Patients (N)

Objective response: tumor volume (%)

Objective response: biochemical (%)

Subjective response: palliation (%)

77 34 229 29 159 6 534

51 48 51 23 8 2/6

68 51 NA 60 24 NA

68 70 Most patients 60 60 NA

229

Tumors of the Adrenergic System: Imaging and Therapy

were 51 and 48%, respectively; a >50% decrease in catecholamine excretion was observed in 68 and 51%, respectively, while symptomatic palliation occurred in 68% of the patients. These results compare favorably with the best reported results of combination chemotherapy and were attained with a treatment that is non-invasive and associated with minimal side effects. Recently published results in a group of 20 patients with malignant pheochromocytoma or paraganglioma who were treated with moderate administered doses (7.4 GBq) at the Netherlands Cancer Institute (objective response 47%, metabolic response 67%, subjective response 89%) [17] compare well with those reported by the group at Duke University (Durham, NC, USA), who treated 18 patients with moderate doses (7.4 GBq) and 15 with high doses (18.5 GBq). The objective response was 38%, metabolic response 60%, and subjective response 86% [18]. Moreover, both a metabolic response and a subjective response were suggested to have an important influence on survival and quality of life, even in the absence of an objective volume response.

toxicity and the early induction of drug resistance. Chemotherapy is reserved for the post-operative treatment of minimal residual disease. Initial results have demonstrated the feasibility and effectiveness of this approach, i.e., a higher objective response rate (>70%) and considerably less toxicity than obtained with 131I-MIBG therapy after conventional treatment [22]. By 2001, results in 56 patients showed that 131I-MIBG is as effective as chemotherapy in attaining operable neuroblastoma: 43 of 56 evaluable patients (77%) had complete or >95% resection of the primary tumor or did not require surgery at all. At follow-up (13-144 months), 5-year survival was 37%. Based upon these results, two new multicenter studies have been initiated in which 131I-MIBG therapy is integrated up-front in the treatment protocol of neuroblastoma. Patients with favorable parameters receive a less aggressive therapy than before, consisting of two cycles of 131I-MIBG followed by surgery, whereas in patients with unfavorable parameters (high-risk group) 131I-MIBG therapy is intensified and combined with the topoisomerase I inhibitor Topotecan to enhance radiationinduced cytotoxicity.

Neuroblastoma Medullary Thyroid Carcinoma Since 1984, therapeutic doses of 131I-MIBG have been administered to children with metastatic or recurrent neuroblastoma that failed to respond to conventional treatment. In 1991, the pooled results of the major centers (273 patients) indicated an objective response rate of 35% [15]; more recently, the response rate increased to 51% (Table 1). Most of these patients had stage IV, progressive, and intensely pre-treated disease, and were administered 131I-MIBG only after other treatment modalities had failed. Both 131I-MIBG therapy and isolation are generally well tolerated by children; however, hematological side effects may occur. Apart from the objective response, the palliative effect was often impressive. Thus, for patients with recurrent and progressive disease after conventional treatment 131I-MIBG therapy is probably the best palliative treatment, as its invasiveness and toxicity compare favorably with that of chemotherapy and external beam radiotherapy [19]. Some groups have combined 131I-MIBG therapy with chemotherapy and/or total-body irradiation, accepting more toxicity, as well as with myeloablative chemotherapy requiring autologous bone marrow or stem-cell rescue [20]. Voûte et al. [21] combined 131I-MIBG therapy with oxygen treatment under hyperbaric conditions. Their aim was to improve survival in patients with recurrent stage IV neuroblastoma by adding the toxic effect of hydroxyl radicals to the radiation effect. Subsequently, high-dose vitamin C therapy was added to this regimen. More recently, 131I-MIBG therapy has been integrated in the treatment protocol as the initial therapy instead of its use in pre-operative combination chemotherapy in children presenting with advanced/inoperable neuroblastoma. The objective is to reduce the tumor volume, thereby enabling adequate surgical resection, and to avoid

In the abdomen, medullary thyroid carcinoma (MTC) may present with liver metastases. Results of combination chemotherapy are disappointing whereas radionuclide therapy using 131I-MIBG or 131I-anti CEA antibodies may provide both tumor regression and palliation. Pooled results in 29 patients with MTC treated with 131I-MIBG (Table 1) showed that an objective response rate occurred in only 23% and tumor marker response in 60%; nevertheless, palliative effects, which may be quite meaningful, were achieved in 60% of the patients. However, only a minority of patients demonstrated sufficient uptake of 131I-MIBG. More patients may be amenable to radioimmunotherapy. In a phase I/II study of treatment using bispecific anti-DTPA/anti-CEA immunoconjugates followed by 131I-hapten in a two-step procedure, 26 MTC patients showed mixed responses. Stabilization of disease and palliation were attained with limited hematological toxicity, but a HAMA (human anti-mouse antibody) response was reported in more than half of the patients [23]. As patients may require several such treatments, the use of chimeric or humanized immunoconjugates would be more appropriate. Carcinoid Tumors Palliative treatments for metastatic carcinoid tumors include long-acting somatostatin analogs (Sandostatin), α-interferon, hepatic artery embolization, 131I-labeled and unlabeled MIBG, and 90Y- or 177Lu-labeled octreotide therapy. The cumulative results of 131I-MIBG therapy in 159 patients with symptomatic, metastatic disease showed an objective response rate of only 8% and a >50% decrease

230

in 5-hydroxyindoleacetic acid (5-HIAA) excretion in 24% (Table 1). Despite the absence of an objective response, palliation was achieved in 60% of patients and without significant side effects [24]. In view of the often indolent character of this disease, the value of a prolonged symptomatic response should not be underestimated. In a study at Duke University Medical Center, 98 patients with metastatic carcinoid were treated with 131I-MIBG. In this group, a subjective response was found to correlate with prolonged survival [25]. In patients with carcinoid tumors not qualifying for 131I-MIBG therapy because of no or insufficient uptake by the tumor, palliative treatment with high doses of unlabeled MIBG also proved beneficial in 60% of the cases, albeit with a shorter mean duration [26]. Improved biochemical and palliative effects of 131I-MIBG treatment due to enhanced tumor/non-tumor ratios by pre-dosing with non-labeled MIBG have also been reported [27]. A combination of higher doses of 131I-MIBG and unlabeled MIBG is used for therapy whenever comparative scintigraphy demonstrates a >20% increase of the tumor/nontumor-ratio following the addition of unlabeled MIBG.

References 1. Hoefnagel CA (1994) Metaiodobenzylguanidine and somatostatin in oncology: role in the management of neural crest tumours. Eur J Nucl Med 21:561-581 2. Khafagi FA, Shapiro B, Fig LM et al (1989) Labetalol reduces Iodine-131 MIBG uptake by pheochromocytoma and normal tissues. J Nucl Med 30:481-489 3. Reubi JC (1995) Neuropeptide receptors in health and disease: the molecular basis for in vivo imaging. J Nucl Med 36:1825-1835 4. Bardiès M, Bardet S, Faivre-Chauvet A et al (1996) Bispecific antibody and Iodine-131-labeled bivalent hapten dosimetry in patients with medullary thyroid or small-cell lung cancer. J Nucl Med 37:1853-1859 5. Goldsmith SJ (2009) Update on nuclear medicine imaging of neuroendocrine tumors. Future Oncol 5:75-84 6. Troncone L, Rufini V, Montemaggi P et al (1990) The diagnostic and therapeutic utility of radioiodinated metaiodobenzylguanidine (MIBG). 5 years experience. Eur J Nucl Med 16:325-335 7. Wiseman GA, Pacak K, O’Dorisio MS et al (2009) Usefulness of 123I-MIBG scintigraphy in the evaluation of patients with known or suspected primary or metastatic pheochromocytoma or paraganglioma: results from a prospective multicenter trial. J Nucl Med 50:1448-1454 8. Hoefnagel CA, De Kraker J (2004) Pediatric tumors. In: Ell PJ and Gambhir SS (eds) Nuclear medicine in clinical diagnosis and treatment, 3rd edition. Churchill Livingstone, Edinburgh, pp 195-206 9. Leung A, Shapiro B, Hattner R et al (1997) The specificity of radioiodinated MIBG for neural crest tumors in childhood. J Nucl Med 38:1352-1357 10. Vik TA, Pfluger T, Kadota R et al (2009) (123)I-mIBG scintigraphy in patients with known or suspected neuroblastoma: results from a prospective multicenter trial. Pediatr Blood Cancer 52:784-790

Cornelis A. Hoefnagel

11. Yeh SDJ, Larson SM, Burch L et al (1991) Radioimmunodetection of neuroblastoma with Iodine-131-3F8: correlation with biopsy, Iodine-131-Metaiodobenzylguanidine and standard diagnostic modalities. J Nucl Med 32:769-776 12. Hoefnagel CA, Rutgers M, Buitenhuis CKM et al (2001) A comparison of targetting neuroblastoma with mIBG and anti L1-CAM antibody mAB chCE7: therapeutic efficacy in a neuroblastoma xenograft model and imaging of neuroblastoma patients. Eur J Nucl Med 28:359-368 13. Hoefnagel CA, Delprat CC, Zanin D, van der Schoot JB (1988) New radionuclide tracers for the diagnosis and therapy of medullary thyroid carcinoma. Clin Nucl Med 13:159-165 14. Hoefnagel CA and Lewington VJ (2004) MIBG therapy. In: Ell PJ and Gambhir SS (eds) Nuclear medicine in clinical diagnosis and treatment, 3rd edn. Churchill Livingstone, Edinburgh, pp 445-457 15. Troncone L, Galli G (1991) Proceedings International Workshop on The Role of [131I]metaiodobenzylguanidine in the Treatment of Neural Crest Tumors. J Nucl Biol Med 35:177362 16. Baulieu J-L, Guilloteau D, Baulieu F et al (1988) Therapeutic effectiveness of Iodine-131 MIBG metastases of a nonsecreting paraganglioma. J Nucl Med 29:2008-2013 17. Gedik GK, Hoefnagel CA, Bais E, Olmos RA (2008) 131IMIBG therapy in metastatic phaeochromocytoma and paraganglioma. Eur J Nucl Med Mol Imaging 35:725-733 18. Safford SD, Coleman RE, Gockerman JP et al (2003) Iodine131 metaiodobenzylguanidine as an effective treatment for malignant pheochromocytoma and paraganglioma. Surgery 134:956-962 19. Hoefnagel CA (1999) Nuclear medicine therapy of neuroblastoma. Q J Nucl Med 43:336-343 20. Yanik GA, Levine JE, Matthay KK et al (2002) Pilot study of iodine-131 metaiodobenzylguanidine in combination with myeloablative chemotherapy and autologous stem-cell support for the treatment of neuroblastoma. J Clin Oncol 20:21422149 21. Voûte PA, van der Kleij AJ, de Kraker J et al (1995) Clinical experience with radiation enhancement by hyperbaric oxygen in children with recurrent neuroblastoma stage IV. Eur J Cancer 31A:596-600 22. Hoefnagel CA, de Kraker J, Valdés Olmos RA, Voûte PA (1994) 131I-MIBG as a first-line treatment in high-risk neuroblastoma patients. Nucl Med Commun 15:712-717 23. Kraeber-Bodéré F, Bardet S, Hoefnagel CA et al (1999) Radioimmunotherapy in medullary thyroid cancer using bispecific antibody and iodine-131-labeled bivalent hapten: Preliminary results of a phase I/II clinical trial. Clin Cancer Res 5:3190s-3198s 24. Zuetenhorst H, Taal BG, Boot H et al (1999) Longterm palliation in metastatic carcinoid tumours with various applications of meta-idobenzylguanidine: pharmacological MIBG, 131I-labeled MIBG and the combination. Eur J Gastroenterol Hepatol 11:1157-1164 25. Safford SD, Coleman RE, Gockerman JP et al (2004) Iodine131 metaiodobenzylguanidine treatment for metastatic carcinoid. Results in 98 patients. Cancer 101:1987-1993 26. Taal BG, Hoefnagel CA, Valdés Olmos RA et al (1996) Palliative effect of Metaiodobenzylguanidine in metastatic carcinoid tumors. J Clin Oncol 14:1829-1838 27. Taal BG, Hoefnagel CA, Boot H et al (2000) Improved effect of 131I-MIBG treatment by predosing with non-radiolabeled MIBG in carcinoid patients, and studies in xenografted mice. Ann Oncol 11:1437-1443

IDKD 2010-2013

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy Dik J. Kwekkeboom Department of Nuclear Medicine, Erasmus Medical Center, Rotterdam, The Netherlands

Introduction The development of peptide receptor scintigraphy in combination with radioiodinated somatostatin analogues allowed the in vivo demonstration of somatostatin-receptorpositive tumors in patients [1]. Later, other radiolabeled somatostatin analogues were developed, two of which subsequently became commercially available. With the advent, over the past decade, of positron emission tomography (PET) tracers for somatostatin receptor imaging, superior image quality and increased sensitivity in tumor site detection have become possible, as confirmed by several research groups. In the 1990’s, attempts at treatment with radiolabeled somatostatin analogues were undertaken in patients with inoperable and/or metastasized neuroendocrine tumors. Improvements in the peptides (higher receptor affinity) and the available radionuclides (β instead of γ emission), together with precautions to limit the radiation dose to the kidneys and bone marrow, led to better results and a virtually negligible percentage of serious adverse events.

between a successful localizing study and a disappointing one. For details of the scanning protocol, the reader is referred to the procedural guidelines for somatostatin receptor scintigraphy with [111In-DTPA0]octreotide, published by the Society of Nuclear Medicine [2].

[111In-DTPA0]Octreotide Scintigraphy: Normal Scintigraphic Findings and Artifacts Normal scintigraphic features include visualization of the thyroid, spleen, liver, and kidneys, and in some patients the pituitary gland (Fig. 1). In addition, the urinary bladder and

Somatostatin-Receptor-Based Radionuclide Imaging The many drawbacks of [123I,Tyr3]octreotide, the radioiodinated somatostatin analogue first used for imaging in patients, resulted in its replacement by the chelated and 111In-labeled somatostatin analogue [111In-DTPA0]octreotide (OctreoScan, Covidien, Petten, The Netherlands), which is commercially available and is now the most commonly used agent for somatostatin receptor imaging (SRI). The preferred dose of [111In-DTPA0]octreotide (containing at least 10 mg of the peptide) is about 200 MBq. This dose is appropriate for single photon emission computed tomography (SPECT), which shows increased sensitivity in the detection of somatostatinreceptor-positive tissues and provides better anatomical delineation than planar views. The acquisition of sufficient counts per view and the generation of spot images with a sufficient counting time – as opposed to the low count density of whole-body scans – are other important advantages of this approach and may make the difference

Fig. 1. Normal distribution in somatostatin receptor imaging (SRI). Variable visualization of the pituitary and thyroid (arrows, upper panels). Faint breast uptake can sometimes be seen in women (right middle panel, arrow). Normal uptake in the liver, spleen, and kidneys, and also some bowel activity is seen in the lower panel. Gallbladder visualization in the lower right panel (arrow). Anterior views

232

Table 1. Pitfalls and causes of potential misinterpretation of positive results with [111In-DTPA0]octreotide scintigraphy Radiation pneumonitis Accessory spleen Focal collection of stools Surgical scar tissue Gallbladder uptake Nodular goiter Ventral hernia Bacterial pneumonia Respiratory infections Common old (nasal uptake) Cerebrovascular accident Concomitant granulomatous disease Diffuse breast uptake Adrenal uptake Urine contamination Concomitant second primary tumor

Table 2. Causes of potential misinterpretation of negative results with [111In-DTPA0]octreotide scintigraphy The presence of unlabeled somatostatin, because of octreotide therapy or due to the production of somatostatin by the tumor itself, may lower tumor detectability Different somatostatin receptor subtypes have different affinities for the radioligand; variable tumor differentiation and receptor expression also influences tumor detectability. This may be important especially in patients with insulinomas and medullary thyroid carcinomas Liver metastases of neuroendocrine tumors may appear isointense because of a similar degree of tracer accumulation by the normal liver. Correlation with anatomical imaging and/or SPECT imaging may be helpful

the bowel are usually visualized to variable degrees. Visualization of the pituitary, thyroid, and spleen is due to receptor binding whereas uptake in the kidneys is for the most part due to re-absorption of the radiolabeled peptide by the renal tubular cells after glomerular filtration. While there is predominant renal clearance of the somatostatin analogue, hepatobiliary clearance via the bowel also occurs, thus necessitating the administration of laxatives in order to facilitate the interpretation of abdominal images. False-positive results of SRI with [111In-DTPA0]octreotide have been reported in virtually all cases; however, the term “false-positive” is a misnomer because it includes somatostatin-receptor-positive lesions unrelated to the pathology for which the investigation was performed (see the review by Gibril et al. [3]). The most common of these are listed in Table 1. The potential causes of a falsenegative study interpretation are given in Table 2.

Imaging Results of [111In-DTPA0]Octreotide Scintigraphy in Neuroendocrine and Other Tumors Somatostatin receptors have been identified in vitro in a large number of human neoplasias, in particular, neuroendocrine tumors (NETs) have a high incidence and

Dik J. Kwekkeboom

density of somatostatin receptors. NETs comprise a group of tumors that includes pituitary adenoma, pancreatic islet cell tumor, carcinoid, pheochromocytoma, paraganglioma, medullary thyroid cancer, and small cell lung carcinoma [4]. Tumors of the nervous system, including meningioma, neuroblastoma, and medulloblastoma, also very often express a high density of somatostatin receptors, as do tumors not classically originating from endocrine or neural cells, such as lymphoma, breast cancer, renal cell cancer, hepatocellular cancer, prostate cancer, sarcoma, and gastric cancer. In the majority of these tumors, somatostatin receptor (SR) subtype-2 is predominantly expressed, although low amounts of other SR subtypes may be concomitantly present [5]. It should also be emphasized that selected non-tumoral lesions may express SRs. For instance, SRs are expressed on the epithelioid cells of active granulomas in sarcoidosis and by inflamed joints in active rheumatoid arthritis, especially the proliferating synovial vessels [6]. Therefore, SR expression is not specific for tumoral pathologies. The most common indication for [111In-DTPA0]octreotide scintigraphy is the detection and localization of gastroenteropancreatic neuroendocrine tumors (GEPNETs) and their metastases, the staging of these patients, the follow-up of patients with known disease, and, lastly, the selection of patients with inoperable and/or metastatic tumors for peptide receptor radionuclide therapy (PRRT) [7-12].

Newer Ligands for Somatostatin Receptor Imaging 99mTc-depreotide

(Neotect) is a commercially available somatostatin analogue that has been approved specifically for use in the detection of lung cancer in patients with pulmonary nodules [13]. Due to the relatively high abdominal signal background and the impossibility to perform delayed imaging because of the tracer’s short half-life, it is less suited for the detection of abdominal NETs [14]. Analogues that are used for PET or hybrid PET/CT imaging are of particular interest because of two advantages that they have over γ-emitting analogues. First, many of them have a better affinity for SR subtype-2, the subtype most commonly expressed by NETs; likewise, there are some analogues that better target other SR subtypes and are therefore more appropriate for visualizing the respective tumors. Second, PET and the combined anatomical and functional information obtained with PET/CT provide images with high spatial resolution, which results in a higher sensitivity of this type of scanning. However, based on a review of the results obtained with these newer analogues, there are also causes for concern. Importantly, in many studies these newer analogues were compared to [111InDTPA0]octreotide scintigraphy using inadequate scanning protocols or comparisons were made between two or more new analogues such that a validated reference method was lacking. Also, the multitude of newly available PET analogues has led to a situation in which each center has to accumulate its own results on the normal findings and

233

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy

artifacts of their scanning methods. This hampers the exchange of data and their shared interpretation. It is, however, likely that one of the new PET analogues, [68Ga-DOTA0, Tyr3]octreotide or [68Ga-DOTA0, Tyr3]octreotate, will become the new standard for SRI using PET. This is due to the fact that these somatostatin analogues have a high affinity for SR subtype-2, and 68Ga is a generator-produced rather than a cyclotron-produced product, such that labeling of the compound is simplified [15]. An additional reason favoring the use of [68Ga-DOTA0,Tyr3]octreotide or [68Ga-DOTA0, Tyr3]octreotate as the standard analogue for PET imaging is that the 90Y or 177Lu-labeled counterparts of these compounds are administered for PRRT; thus, the peptide used in diagnostic imaging closely mimics the one that is used for therapy. Other radionuclide-coupled ligands that do not rely on the presence of SRs for tumor visualization have also been tested in patients with GEPNETs. The oldest of these, 123I-MIBG, performs poorly compared to [111In-DTPA0]octreotide scintigraphy [16]. PET scanning with (18F)2-fluoro-2-deoxy-D-glucose 18 ( F-FDG) has gained importance for tumor staging and the evaluation of treatment response for a number of tumor types. The method is based on glucose consumption by the tumors, such that fast-growing tumors usually show high tracer uptake. However, 18F-FDG PET is less suited for GEPNETs, because of the slow growing nature of these tumors. Therefore, the technique is recommended only in patients with negative SRI findings [17], a situation that usually correlates with more aggressive tumor behavior and faster tumor growth. Newer PET radioligands that have been clinically tested in patients with GEPNETs include 18F-DOPA and 11C-5-hydroxy-tryptophan [18]. PET with these ligands has been reported to be more sensitive than SRI with [111In-DTPA0]octreotide. These PET ligands, however, have a short half-life and therefore have to be synthesized in the close vicinity of or in the hospital where they are to be administered. Also, both 18F-DOPA and 11C-5hydroxy-tryptophan, unlike the radiolabeled somatostatin analogues used in PET, lack a sequel in PRRT.

Somatostatin-Receptor-Based Radionuclide Therapy As noted above, the functioning and non-functioning endocrine pancreatic tumors and carcinoids that make up

the GEPNETs are usually slow-growing. The treatment of tumor metastases with somatostatin analogues results in reduced hormonal overproduction and symptomatic relief in most cases. Treatment with somatostatin analogues is, however, seldom successful in terms of tumor size reduction [19]. A new treatment modality for patients with inoperable or metastasized endocrine GEPNETs is the use of radiolabeled somatostatin analogues. The majority of endocrine GEP tumors possess SRs and can therefore be visualized with SRI. A logical sequence to tumor visualization in vivo is to then treat these patients with radiolabeled somatostatin analogues. In the early phases of this approach, virtually all patients considered as candidates for PRRT had well-differentiated GEPNETs. At the time of these early studies, in the mid- to late 1990s, no other chelated somatostatin analogues labeled with β-emitting radionuclides were available, such that [111In-DTPA0]octreotide was used for PRRT. The results of these studies, in which high doses of the radionucleotide were administered to patients with metastasized NETs, were encouraging with regard to symptom relief but partial remissions (PRs) were exceptional [20, 21] (Table 3). The next generation of SR-mediated radionuclide therapy was based on the use of the modified somatostatin analogue [Tyr3]octreotide, which has a higher affinity for SR subtype-2, and a different chelator, DOTA instead of DTPA, in order to ensure a more stable binding of the intended β-emitting radionuclide, 90Yttrium (90Y). The resulting compound (90Y-DOTATOC; OctreoTher, Novartis, Switzerland), was used in several phase-1 and phase-2 PRRT trials [22-25] (Table 3) but renal insufficiency and myelodysplastic syndrome were reported as serious adverse events (SAEs). The incidence of these SAEs could, however, be dramatically reduced through adequate renal protection, achieved by the co-infusion of amino acids. Consequently, SAEs have become relatively rare, occurring in <10% of patients [26]. Despite differences in protocols, the rate of complete remission (CR) and PR reported by most of the different studies with [90Y-DOTA0,Tyr3]octreotide are in the same range, between 10 and 30%, which is better than the rates obtained with [111In-DTPA0]octreotide. Reubi et al. [27] reported a nine-fold increase in somatostatin receptor subtype-2 affinity for [DOTA0, Tyr3]octreotate vs. [DOTA0,Tyr3]octreotide, and a six to seven-fold increase in the affinity for the Yttrium-loaded

Table 3. Tumor responses in patients with GEPNETs, treated with different radiolabeled somatostatin analogues Center (reference)

Tumor response ligand

Rotterdam [20] New Orleans [21] Milan [22] Basel [23, 24] Rotterdam [25] Rotterdam [29]

[111In-DTPA0]octreotide [111In-DTPA0]octreotide [90Y-DOTA0,Tyr3]octreotide [90Y-DOTA0,Tyr3]octreotide [90Y-DOTA0,Tyr3]octreotide [177Lu-DOTA0,Tyr3]octreotate

Patient Complete Partial number remission remission 26 26 21 74 58 310

0 0 0 3 (4%) 0 5 (2%)

0 2 (8%) 6 (29%) 15 (20%) 5 (9%) 86 (28%)

Minimal response

Stable disease

Progressive disease

Complete + partial remissions

5 (19%) NA NA NA 7 (12%) 51 (16%)

11 (42%) 21 (81%) 11 (52%) 48 (65%) 33 (61%) 107 (35%)

10 (38%) 3 (12%) 4 (19%) 8 (11%) 10 (19%) 61 (20%)

0% 8% 29% 24% 9% 29%

234

Dik J. Kwekkeboom

counterparts of this compound. In addition, a comparison carried out in patients showed that the uptake of radioactivity, expressed as percentage of the injected dose of [177Lu-DOTA0,Tyr3]octreotate, was comparable to that after [111In-DTPA0]octreotide for kidneys, spleen, and liver, but was 3- to 4-fold higher for four of five tumor types [28]. Therefore, [177Lu-DOTA0,Tyr3]octreotate potentially represents an important improvement because of the higher absorbed doses that can be achieved in most tumors and the essentially equal doses to potentially dose-limiting organs. Moreover, the lower tissue penetration range of 177Lu vs. 90Y may be especially important for small tumors. These findings support the use of 177Lu-octreotate as the radiolabeled somatostatin analogue of choice in PRRT. The side effects and treatment outcome of [177LuDOTA0,Tyr3]octreotate therapy have been analyzed by our group in 504 and 310 patients with GEPNETs, respectively [29]. In the 504 patients, acute side effects occurring within 24 h after administration of the radiopharmaceutical included nausea (following 25% of administrations), vomiting (10%), and abdominal discomfort or pain (10%). Six patients were hospitalized within 2 days of administration of the radiopharmaceutical because of

3-2007

hormone-related crises [30]. All patients recovered after adequate care. Subacute, hematological toxicity of WHO toxicity grade 3 or 4 occurred 4-8 weeks after 3.6% of administrations, or, expressed in a patient-based manner, after at least one of several treatments in 9.5% of patients. Serious delayed toxicities were observed in 9 out of 504 patients. There were two cases of renal insufficiency, both of which were probably unrelated to 177Lu-octreotate treatment. Three patients showed serious liver toxicity, in two of these cases probably treatment-related. Lastly, myelodysplastic syndrome occurred in four patients and was probably treatment-related in three. Treatment responses according to tumor type at 3 months after the last therapy cycle were analyzed in 310 patients. The overall objective tumor response rate, comprising CR, PR, and minimal response (MR), was 46% (for an example, see Fig. 2). Prognostic factors predicting tumor remission, i.e., CR, PR, or MR, as treatment outcome were uptake on the OctreoScan (p <0.01) and Karnofsky performance score (KPS) >70 (p <0.05). A small percentage of patients who had either stable disease (SD) or MR at their first two evaluations after therapy had a further improvement in categorized tumor response at 6 and 12 months follow-up, occurring in 4% and 5%

5-2007

11-2007

Post Tx 1

Post Tx 4

2-2008

11-2008

CT

SRI

Fig. 2. Serial CT scans, SRI, and post peptide receptor radionuclide therapy (PPRT) scans in a patient with metastatic neuroendocrine tumor with unknown primary. Month and year are indicated in the top row. Notice the ongoing tumor regression on CT, and also the improvement on SRI. The last post-therapy scan shows less tumor uptake than in the first; the difference is even more impressive if the same scaling is used (but then, due to the impressive tumor uptake, in May 2007 the tumors would appear as one large hot spot occupying most of the abdomen). This diminishing tumor uptake on subsequent post-therapy scans usually implies tumor shrinkage

235

Neuroendocrine Tumors of the Abdomen: Imaging and Therapy

of patients, respectively. Three of four patients with clinically non-functioning neuroendocrine pancreatic tumors, that were judged inoperable before treatment with 177Luoctreotate and who had PR, successfully underwent surgery 6-12 months after their last treatment; the fourth patient died of postoperative complications. The median time to progression was 40 months from the start of treatment. Median overall survival in our 310 GEP tumor patients was 46 months (median follow-up 19 months; 101 deaths). Median disease-related survival was >48 months (median follow-up 18 months; 81 deaths). Median progression-free survival was 33 months. The most important factor predicting survival was treatment outcome. Low KPS and liver involvement were also very significant predictors.

Comparison of Survival Data Since, in our study, treatment with [177Lu-DOTA0,Tyr3] octreotate is still open for new patients and the median follow-up in relation to survival is still relatively short, we analyzed our local, Dutch patients separately, and specifically those patients with longer follow-up. The results from these analyses suggest that both overall and disease-specific survival times are consistently at or above 48 months. These numbers compare favorably to those reported in the literature. A comparison of survival data for our patients, either from time of diagnosis or from time of referral, with data from different epidemiological studies or studies pertaining to a specific intervention, and limiting our data to similar subgroups of patients, showed a benefit in overall survival for patients treated with 177Lu-octreotate, which ranged from 40 to 72 months from the time of diagnosis [29]. We are aware that comparisons with historical controls should be interpreted with caution, but we also think that this consistent difference with many other reports in similar patient groups cannot be ignored, and is most probably caused by a real difference in survival.

Timing of Treatment In patients treated with [177Lu-DOTA0,Tyr3]octreotate, median overall survival was shorter in patients with a poor performance score and in those with extensive liver involvement by the tumor. This implies that treatment with [177Lu-DOTA0,Tyr3]octreotate should preferably be started early in the disease’s evolution. Since GEPNETs can be clinically stable for years, however, it is, in our opinion, good clinical practice to wait for signs of disease progression if the tumor load is moderate. Such signs should not be restricted to CT-assessed tumor growth but should also include rises in serum tumor markers, increase in symptoms, or involuntary weight loss. In patients with very limited tumor load and in whom cure is potentially possible, treatment should be initiated without further delay. The same holds true for patients with extensive tumor load, hepatomegaly, or those with signifi-

cant weight loss, when waiting for formally assessed tumor progression would place these patients in an unfavorable starting position for treatment or would even make them ineligible for treatment.

Conclusions The use of [111In-DTPA0]octreotide in SRI has a proven role in the diagnosis and staging of GEPNETs. Newer radiolabeled somatostatin analogues that can be used in PET imaging, and which have a higher SR affinity, especially for subtype-2, have been developed. It would be desirable, however, if one radiolabeled analogue became the new standard for PET imaging, as the current application of a multitude of analogues implies a fragmented knowledge regarding image interpretation. Treatment with radiolabeled somatostatin analogues is a promising new tool in the management of patients with inoperable or metastasized NETs. The results thus far obtained with [ 90Y-DOTA 0,Tyr 3]octreotide and [177Lu-DOTA0,Tyr3]octreotate have been very encouraging in terms of tumor regression. Also, if kidneyprotective agents are used, the side effects of this form of therapy are few and mild, and the duration of the therapy response for both radiopharmaceuticals may be longer than 30 months. Lastly, compared to historical controls, there appears to be a benefit in overall survival of several years from the time of diagnosis in patients treated with [ 177Lu-DOTA 0,Tyr 3]octreotate. These data compare favorably with the limited number of alternative treatment approaches. If more widespread use of PRRT can be guaranteed, it may well become the therapy of choice in patients with metastasized or inoperable GEPNETs.

References 1. Krenning EP, Bakker WH, Breeman WA et al (1989) Localization of endocrine related tumors with radioiodinated analogue of somatostatin. Lancet 1:242-245 2. Balon HR, Goldsmith SJ, Siegel BA et al (2001) Procedure guideline for somatostatin receptor scintigraphy with 111Inpentetreotide. J Nucl Med 42:1134-1138 3. Gibril F, Reynolds JC, Chen CC et al (1999) Specificity of somatostatin receptor scintigraphy: a prospective study and effects of false-positive localizations on management in patients with gastrinomas. J Nucl Med 40:539-553 4. Reubi JC (1997) Regulatory peptide receptors as molecular targets for cancer diagnosis and therapy. Q J Nucl Med 41:63-70 5. Reubi JC, Waser B, Schaer JC, Laissue JA (2001) Somatostatin receptor sst1-sst5 expression in normal and neoplastic human tissues using receptor autoradiography with subtypeselective ligands. Eur J Nucl Med 28:836-846 6. Reubi JC, Waser B, Krenning EP et al (1994) Vascular somatostatin receptors in synovium from patients with rheumatoid arthritis. Eur J Pharmacol 271:371-378 7. Krenning EP, Kwekkeboom DJ, Bakker WH et al (1993) Somatostatin receptor scintigraphy with [111In-DTPA-D-Phe1]and [123I-Tyr3]-octreotide: the Rotterdam experience with more than 1000 patients. Eur J Nucl Med 20:716-731

236

8. Kwekkeboom DJ, Krenning EP, Bakker WH et al (1993) Somatostatin analogue scintigraphy in carcinoid tumors. Eur J Nucl Med 20:283-292 9. Westlin JE, Janson ET, Arnberg H et al (1993) Somatostatin receptor scintigraphy of carcinoid tumours using the [111In DTPA D Phe1] octreotide. Acta Oncol 32:783-786 10. De Kerviler E, Cadiot G, Lebtahi R et al (1994) Somatostatin receptor scintigraphy in forty eight patients with the Zollinger Ellison syndrome. Eur J Nucl Med 21:1191-1197 11. Gibril F, Reynolds JC, Doppman JL et al (1996) Somatostatin receptor scintigraphy: its sensitivity compared with that of other imaging methods in detecting primary and metastatic gastrinomas. A prospective study. Ann Intern Med 125:26-34 12. Lebtahi R, Cadiot G, Sarda L et al (1997) Clinical impact of somatostatin receptor scintigraphy in the management of patients with neuroendocrine gastroenteropancreatic tumors. J Nucl Med 38:853-858 13. Menda Y, Kahn D (2002) Somatostatin receptor imaging of nonsmall lung cancer with 99mTc depreotide. Semin Nucl Med 32:92-96 14. Lebtahi R, Le Cloirec J, Houzard C et al (2002) Detection of Neuroendocrine Tumors: (99m)Tc-P829 Scintigraphy Compared with (111)In-Pentetreotide Scintigraphy. J Nucl Med 43:889-895 15. Hofmann M, Maecke H, Börner R et al (2001) Biokinetics and imaging with the somatostatin receptor PET radioligand (68)GaDOTATOC: preliminary data. Eur J Nucl Med 28:1751-1757 16. Quigley AM, Buscombe JR, Shah T et al (2005) Intertumoural variability in functional imaging within patients suffering from neuroendocrine tumours. An observational, cross-sectional study. Neuroendocrinology 82:215-220 17. Belhocine T, Foidart J, Rigo P et al (2002) Fluorodeoxyglucose positron emission tomography and somatostatin receptor scintigraphy for diagnosing and staging carcinoid tumours: correlations with the pathological indexes p53 and Ki-67. Nucl Med Commun 23:727-734 18. Koopmans KP, de Vries EG, Kema IP et al (2006) Staging of carcinoid tumours with 18F-DOPA PET: a prospective, diagnostic accuracy study. Lancet Oncol 7:728-734 19. Arnold R, Benning R, Neuhaus C et al (1993) Gastroenteropancreatic endocrine tumours: effect of Sandostatin on tumour growth. The German Sandostatin Study Group. Digestion 54(Suppl 1):72-75

Dik J. Kwekkeboom

20. Valkema R, de Jong M, Bakker WH et al (2002) Phase I study of peptide receptor radionuclide therap1y with [111In-DTPA0]Octreotide: the Rotterdam experience. Semin Nucl Med 32:110-122 21. Anthony LB, Woltering EA, Espanan GD et al (2002) Indium111-pentetreotide prolongs survival in gastroenteropancreatic malignancies. Semin Nucl Med 32:123-132 22. Bodei L, Cremonesi M, Zoboli S et al (2003) Receptor-mediated radionuclide therapy with 90Y-DOTATOC in association with amino acid infusion: a phase I study. Eur J Nucl Med Mol Imag 30:207-216 23. Waldherr C, Pless M, Maecke HR et al (2001) The clinical value of [90Y-DOTA]-D-Phe1-Tyr3-octreotide (90Y-DOTATOC) in the treatment of neuroendocrine tumours: a clinical phase II study. Ann Oncol 12:941-945 24. Waldherr C, Pless M, Maecke HR et al (2002) Tumor response and clinical benefit in neuroendocrine tumors after 7.4 GBq (90)Y-DOTATOC. J Nucl Med 43:610-616 25. Valkema R, Pauwels S, Kvols LK et al (2006) Survival and response after peptide receptor radionuclide therapy with [90YDOTA0,Tyr3]octreotide in patients with advanced gastroenteropancreatic neuroendocrine tumors. Semin Nucl Med 36:147-156 26. Valkema R, Pauwels SA, Kvols LK et al (2005) Long-term follow-up of renal function after peptide receptor radiation therapy with 90Y-DOTA0,Tyr3-octreotide and 177Lu-DOTA0, Tyr3-octreotate. J Nucl Med 46(Suppl 1):83S-91S 27. Reubi JC, Schaer JC, Waser B et al (2000) Affinity profiles for human somatostatin receptor sst1-sst5 of somatostatin radiotracers selected for scintigraphic and radiotherapeutic use. Eur J Nucl Med 27:273-282 28. Kwekkeboom DJ, Bakker WH, Kooij PP et al (2001) [177LuDOTA0Tyr3]octreotate: comparison with [111In-DTPA0]octreotide in patients. Eur J Nucl Med 28:1319-1325 29. Kwekkeboom DJ, de Herder WW, Kam BL et al (2008) Treatment with the radiolabeled somatostatin analog [177LuDOTA0,Tyr3]octreotate: toxicity, efficacy, and survival. J Clin Oncol 26:2124-2130 30. De Keizer B, van Aken MO, Feelders RA et al (2008) Hormonal crises following receptor radionuclide therapy with the radiolabeled somatostatin analogue [177Lu-DOTA0,Tyr3] octreotate. Eur J Nucl Med Mol Imag 35:749-755

PEDIATRIC SATELLITE COURSE “KANGAROO”

IDKD 2010-2013

Imaging Cystic Kidneys in Children Fred E. Avni Department of Radiology, University Clinics of Brussels, Erasme Hospital, Brussels, Belgium

Introduction Renal cystic diseases may be discovered or suspected at any gestational week during fetal life or at any age in childhood. They encompass a large number of conditions that can be separated according to whether or not they are hereditarily transmitted (Table 1). Imaging, mainly ultrasound (US), but also computed tomography (CT) and magnetic resonance imaging (MRI) in selected indications, plays an important role in differentiating between the various types of cystic diseases, by showing the characteristics of renal involvement as well as associated anomalies [1-5].

Cystic Renal Diseases in the Fetus and in the Perinatal Period

associated anomalies. The timing of detection is important as is the amount of amniotic fluid, since both are important features for the final diagnosis and prognosis. A pre-existing familial history of any renal cystic disease or of any syndrome with renal involvement in a fetus in which abnormal kidneys have been detected should raise strong suspicion. Finally, the evaluation should include a detailed clinical inquiry, looking for toxic causes or maternal diseases that may influence fetal health.

Hyperechoic Kidneys in the Perinatal Period This group includes many diseases that are genetically transmitted or acquired during fetal life. Their evaluation requires a step by step approach. Most will be discovered during the second and third trimester. Sonographic Criteria

In the fetus (and perinatal period), cystic renal disease should be suspected whenever bilateral hyperechoic kidneys or cysts (uni- or bilateral) are discovered during an obstetrical ultrasound examination. The imaging approach for the diagnosis should be based on a detailed sonographic analysis that includes measurement of renal length, the presence or absence of corticomedullary differentiation (CMD), and the presence, number, size and location of the cysts. This evaluation should be completed through an analysis of the entire fetus, looking for

Renal cortical hyperechogenicity is determined by comparing the renal cortex to the adjacent liver or to the spleen. While hyperechogenicity is in most cases visually obvious, there are four objective criteria indicative of cystic disease: 1. the renal cortex should not be hyperechoic compared to the liver or spleen during the third trimester; 2. the appearance of the CMD, which can be normal, increased, absent or reversed (due to a hyperechoic medulla);

Table 1. Classification of cystic diseases of the kidney in the fetus and in children Genetic diseases

Non-genetic diseases (congenital or acquired)

Autosomal recessive polycystic kidney disease (ARPKD) Autosomal dominant polycystic kidney disease (ADPKD) Glomerulocystic kidney diseases (including TCF2 anomalies, nephronophitisis/medullary sponge kidney complex) Cystic dysplasia Medullary cystic dysplasia associated with syndromes Renal obstructive dysplasia (associated with urinary tract malformations) Multicystic/dysplastic kidney (some cases with genetic transmission) Localized cystic dysplasia Simple cyst Multilocular cyst Cystic tumor Cysts associated with chronic dialysis

240

3. renal size: markedly increased (>+4 SD), moderately increased (>+2 SD), or normal or small (<–2 SD) [6]; 4. the presence of renal cysts. Finally, any familial history or associated findings constitute important factors for the differential diagnosis [2, 6-9]. Differential Diagnosis In case of markedly enlarged (>+4SD) hyperechoic kidneys diagnosed during the late first and early second trimesters, Meckel-Gruber syndrome should be considered first, especially if the medulla appears enlarged and hypoechoic and if polydactyly and cerebral anomalies are associated. If the condition is detected during the second and third trimesters, the main diagnosis to be added to the differential would be autosomal recessive polycystic kidney disease (ARPKD) and Bardet-Biedl syndrome (BBS). In ARPKD, CMD may be partially absent, completely absent, or even reversed. A few visible cysts are rare but may be seen in utero. Oligohydramnios is a frequent finding and is associated with pulmonary hypoplasia, which confers a very poor prognosis [7-10]. In BBS, the kidneys are enlarged and hyperechoic and there is post-axial polydactyly. The other symptoms of the disease will develop after birth. Cysts can be observed already in utero or appear after birth [7-12]. In case of moderately enlarged hyperechoic kidneys (+2 SD), three diagnoses have to be considered first: 1. TCF2 mutation associated nephropathy; 2. ARPKD; 3. autosomal dominant polycystic kidney disease (ADPKD). An anomaly of TCF2 (leading to HNF-1β-related morphological anomalies) was recently shown to represent the main cause of fetal hyperechoic kidneys [11]. This mutation is associated with a wide spectrum of renal morphological and structural anomalies that histologically include glomerulocystic-type changes, cystic dysplasia, and renal agenesis. Hepatic ductular plate anomalies are commonly associated findings. In such kidneys, besides renal hyperechogenicity, CMD may or not be visible. Cysts may be detected already in utero or, more often, after birth; they are located in the subcortical area. A familial history of diabetes is a frequent finding. The involvement and extent of the kidney lesions related to ARPKD can vary from 10 to 90%, with the US appearances varying accordingly. In cases with mild involvement, the kidneys may be moderately enlarged, with a hyperechoic cortex and a few small cysts mainly within the pyramids. After birth, cysts may also develop, throughout the medulla first and within the cortex thereafter. Fetuses with mildly enlarged kidneys have a better prognosis for survival than those with massive enlargement [9]. Already in utero, ADPKD may be suspected based on a marked hyperechoic renal cortex that increases CMD. The kidneys are usually normal in size or slightly enlarged. Such finding should prompt familial inquiry. Cysts may be observed in utero but usually develop after birth [13].

Fred E. Avni

Table 2. Causes of moderately enlarged hyperechoic kidneys in the fetus TCF2 mutation Autosomal recessive polycystic kidney disease Autosomal dominant polycystic kidney disease Maternally related diseases Infection Ischemia Metabolic diseases Dysplasia Nephrotic syndromes “Transient”

Once these three diagnoses are excluded, there is a wide spectrum of other diseases that can lead to hyperechoic kidneys; clinical inquiry may suggest the diagnosis [14, 15]. Complementary examinations, such as chromosomal analysis, are directed at searching for infectious, toxic, maternally related, or ischemic causes and will help to reach a diagnosis (Table 2).

Renal Cyst(s) Discovered in the Perinatal Period A unilocular, single renal cyst occurring in otherwise normal-appearing kidneys can be detected in utero or after birth. It should be differentiated, especially if the cyst is septated, from a cystic tumor, segmental cystic dysplasia, a dysplastic upper-pole of a duplex kidney, or a urinoma. Associated urinary tract malformation and dilatation may help make the diagnosis. Noteworthy is the fact that ADPKD may start asymmetrically (with a single cyst) [16]. Whenever multiple cysts are detected, the first criterion for the differential diagnosis is uni- or bilateral involvement. Multiple cysts detected in one kidney only most often correspond to a multicystic dysplastic kidney (MCDK), which usually has a straightforward US appearance: multiple cysts of various sizes without interconnection, no recognizable normal renal parenchyma, and no central renal pelvis. MDCK should be differentiated from obstructive dysplasia (associated with urinary tract obstructive malformation), in which the dilated urinary tract is recognizable. The disease can also occur in the upper pole of a duplex kidney. MDCK evolves such that in most cases the kidney will eventually shrink. This can be followed by US [17, 18]. Bilateral multiple renal cysts can be visualized in a large number of isolated renal or syndromic diseases (Table 3) [3, 4] and may or may not be associated with global renal Table 3. Bilateral multiple cysts in the perinatal period Bilateral multicystic dysplastic kidney disease Bilateral obstructive dysplasia (+urinary tract dilatation) Autosomal dominant polycystic kidney disease Autosomal recessive polycystic kidney disease Glomerulocystic disease (subcortical cysts) Syndromes with cystic dysplasia, including (but not limited to) Ivemark syndrome, Zellweger syndrome, Meckel Gruber syndrome, Bardet-Biedl syndrome, tuberous sclerosis complex

241

Imaging Cystic Kidneys in Children

Table 4. Bilateral macrocysts Tuberous sclerosis complex Simpson-Golabi-Behmel TCF2 mutation syndrome Bilateral multicystic dysplastic kidney disease

hyperechogenicity (see above). A step-by-step approach, including detailed US analysis, familial history, and complementary examinations, will lead to the diagnosis. Amniotic fluid volume and associated morphological or chromosomal anomalies are mandatory for the prognosis. An interesting subgroup includes diseases with macro-cysts that are already obvious at birth [19, 20] (Table 4). Many renal cystic diseases present with some type of hepatic ductular anomaly. However, these are rarely detected by perinatal US. Furthermore, in some patients it may be of interest to perform fetal MRI in order to better characterize the renal involvement [21].

Cystic Renal Diseases in Childhood Renal cystic diseases in children are usually discovered during an US examination performed: (1) in the followup of a known, perinatally diagnosed disease, (2) during the workup of syndromes diagnosed after birth, or (3) during screening in an at-risk family. They may also be detected as an incidental finding in an US examination performed for other reasons. The main sonographic features of renal cystic diseases are obviously the presence of visible cysts uni-or bilaterally [1, 3]. The sonographic approach is similar to the one described in the fetus and is based on the characteristics of the cortex, medulla, and CMD as well as the size, location, and number of cysts. Certain features will orient the diagnosis [22, 23]: subcapsular cysts are indicative of glomerulocystic disease; medullary cysts suggest medullary cystic dysplasia (ARPKD); diffuse cortical cysts point to ADPKD, and macro-cysts to ADPKD or tuberous sclerosis complex. The role of US is in the follow-up of these patients once the disease has been diagnosed, including monitoring disease evolution and the potential development of complications (hemorrhage, lithiasis). Another important role for US is to verify the occurrence of hepatic complications of the diseases and the development of portal hypertension [24-26]. For this purpose, MRI may provide additional information important for patient management [27]. Nephronophthisis (NPS) deserves special mention. This genetically transmitted, autosomal recessive disease is being increasingly recognized due to improved genetic mapping. The symptoms of NPS include anemia, polyuria, and polydysia, along with chronic renal failure. The disease can be part of a syndrome (i.e., Joubert syndrome) or an isolated finding. At US examination, the kidneys are relatively small and cysts will develop at the corticomedullary junction [28, 29].

Conclusions The differential diagnosis of renal cystic disease is a difficult challenge. Cystic disease is suspected based on the discovery of hyperechoic kidney or/and cysts. It can be identified by approaching the diagnosis in a step-by-step manner that includes US analysis, familial history, and clinical evaluation.

References 1. De Bruyn R, Gordon R (2000) Imaging in cystic renal disease. Arch Dis Child 83:401-407 2. Avni EF, Garel L, Cassart M et al (2006) Perinatal assessment of hereditary cystic renal diseases: the contribution of sonography. Pediatr Radiol 35:405-414 3. Rizk D, Chapman AB (2003) Cystic and inherited kidney diseases. Am J Kidn Dis 42:1305-1317 4. Deshpande C, Hennekam RCM (2008) Genetic syndromes and prenatally detected renal anomalies. Semin Fetal Neonat Med 13:171-180 5. Winyard P, Chitty LS (2008) Dysplastic kidneys. Semin Fetal Neonat Med 13:142-151 6. Cohen HL, Cooper J, Eisenberg P et al (1991) Normal length of fetal kidneys. AJR Am J Roentgenol 157:545-548 7. De Bruyn R, Marks SD (2008) Post-natal investigation of fetal renal disease. Semin Fetal Neonat Med 12:133-141 8. Tsatsaris V, Gagnadoux MF, Aubry MC et al (2002) Prenatal diagnosis of bilateral isolated fetal hyperechogenic kidneys. It is possible to predict long-term outcome? BJOG 109:1388-1393 9. Chaumoitre K, Brun M, Cassart M et al (2006) Differential diagnosis of fetal hyperechogenic cystic kidneys unrelated to renal tract anomalies. Ultrasound Obstet Gynecol 28:911-917 10. Ickowicz V, Eurin D, Maugey-Laulom B et al (2006) MeckelGruber syndrome: sonography and pathology. Ultrasound Obstet Gynecol 27:296-300 11. Decramer S, Parant O, Beaufils S et al (2007) Anomalies of the TCF2-Gene are the main cause of fetal bilateral hyperechogenic kidneys. J Am Soc Nephrol 18:923-929 12. Cassart M, Eurin D, Didier F et al (2004) Antenatal renal sonographic anomalies and post-natal follow-up of renal involvement in Bardet-Biedl syndrome. Ultrasound Obstet Gynecol 24:51-54 13. Brun M, Maugey-Laulom B, Eurin D et al (2004) Prenatal sonographic patterns in autosomal dominant polycystic kidney disease. Ultrasound Obstet Gynecol 24:55-61 14. Slovis TL, Bernstein J, Gruskin A (1993) Hyperechoic kidneys in the newborn and young infant. Pediatr Nephrol 7:294-302 15. Nortier JL, Debiec H, Tournay Y et al (2006) Neonatal disease in NEP alloimmunization: lessons for immunological monitoring. Pediatr Nephrol 21:1399-1405 16. McHugh K, Stringer DA, Hebert D, Babiak GA (1991) Simple renal cysts in children: diagnosis and follow-up with US. Radiology 178:383-385 17. Kuwertz-Broeking E, Brinkmann OA, VanLengerke HJ et al (2004) Unilateral MDK: experience in children. BJU Internat 93:388-392 18. Aslam M, Watson AR (2006) Unilateral MDK: long-term outcomes. Arch Dis Child 91:820-823 19. Glazier DB, Fleisher MH, Cummings KB, Barone JG (1996) Cystic renal disease and TS in children. Urology 48:613-615 20. Newmann HPH, Schwarzkopf G, Henske EP (1998) Renal angiomyolipomas, cysts and cancer in TSC. Semin Pediatr Neurol 5:269-275 21. Cassart M, Massez A, Metens T et al (2004) Complementary role of MRI after US in assessing bilateral UT anomalies in the fetus. AJR Am J Roentgenol 182:684-695

242

22. Jaim M, Lequesne GW, Bourne AJ, Henning P (1997) High-resolution US in the differential diagnosis of cystic diseases of the kidney in infancy and childhood. J Ultrasound Med 16:235-240 23. Traubici J, Daneman A (2005) High-resolution renal sonography in children with ARPKD. AJR Am J Roentgenol 184:1630-1633 24. Lipschitz B, Berdon WE, Defelice AR, Levy J (1993) Association of congenital hepatic fibrosis with ADPKD. Pediatr Radiol 23:131-133 25. Premkumar A, Berdon WE, Levy J et al (1988) Emergence of hepatic fibrosis and portal HT in ARPKD. Pediatr Radiol 18:123-129

Fred E. Avni

26. Avni EF, Guissard G, Hall M et al (2002) Hereditary polycystic kidney diseases in children: changing sonographic patterns through childhood. Pediatr Radiol 32:169-174 27. Turkbey B, Ocak I, Daryanani K et al (2009) ARPKD and congenital hepatic fibrosi. Pediatr Radiol 39:100-111 28. Salomon R, Saunier S, Niaudet P (2009) Nephronophtisis. Pediatr Nephrol 24:2333-2344 29. Blowey DL, Querfeld U, Geary D et al (1996) US findings in juvenile nephronophtisis. Pediatr Nephrol 10:22-24

IDKD 2010-2013

Understanding Duplication Anomalies of the Kidney Jeanne S. Chow Departments of Urology and Radiology, Children’s Hospital, Boston, MA, USA

Introduction Learning the basic rules governing duplication anomalies of the kidney is particularly rewarding because the fundamental rules can be applied reliably to the great variety of cases and using all different imaging modalities. Although ultrasound is typically the starting point of imaging of the urinary tract in infants and children, renal duplex anomalies are also studied by micturating cystourethrography (MCU), also called voiding cystourethrography (VCUG), intravenous pyelography (IVP), magnetic resonance urography (MRU), computed tomography (CT), MAG-3 lasix renogram, and DMSA scan. While more and more duplex anomalies are being discovered in utero, most people with duplex anomalies are asymptomatic and thus may never be studied at all.

Embryology The ureteric bud must meet the metanephros in order for the kidney to form. Without this interaction, kidney formation is not induced. If two ureteric buds meet the metanephric blastema, then the kidney becomes duplex and the collecting system and ureter become duplicated (Fig. 1). Ureteral duplication may be complete or incomplete.

a

In incomplete ureteral duplication, a single ureteric bud, which is derived from the the mesonephric (Wolffian) duct, bifurcates and meets the metanephros during approximately week 5-6 of gestation. The two branches of the ureter may join at the level of the renal pelvis (bifid pelvis) or the proximal, middle or distal ureter (bifid ureter) and terminate in a single distal ureter that inserts orthotopically into the bladder. Rarely, one of the ureteral buds may be blind-ending and never appear to “reach” the kidney (blind ending-ureteric duplication). Since the upper and lower poles of the duplex kidney with a bifid ureter have a common distal ureter, the upper and lower poles typically appear similar (and normal). If there is reflux or obstruction at the end of the common ureter, both the upper and the lower pole will be affected. In complete ureteral duplication, two separate ureteric buds arise from the mesonephric duct to meet the metanephric blastema. The lower-pole ureter is the analogue of the normal single-system ureter and has a normally located ureteral orifice in the corner of the trigone of the bladder. The upper-pole ureter is the “accessory ureter” and inserts medially and inferiorly to the normal ureteral orifice (Weigert-Meyer rule). Since the lower pole of the duplex kidney is analogous to the normal single-system kidney, abnormalities of the lower pole are

Mesonephric Duct Accessory Ureter

Fig. 1 a, b. Duplex Kidney: embryology of the ectopic pathway. a Embryology of a complete ureteral duplication showing that the accessory ureter joins the upper pole of the metanephros. b As the kidney continues to develop, the bladder insertion of the lower-pole ureter ascends to the normal (orthotopic) position in the bladder trigone. The upper-pole ureter descends so that its orifice in the bladder is medial and inferior to the lowerpole ureter

Metanephros Ureter

b

244

Jeanne S. Chow

a

Fig. 2. Comparison of the calyceal axes

b

similar to those of a single (non-duplex) collecting system, such as vesicoureteral reflux and obstruction at the ureteropelvic or ureterovesical junction. Reflux into the lower-pole ureter and intrarenal collecting system can be easily distinguished from reflux into a single (non-duplex) system kidney by carefully noting the axis of the calyces (Fig. 2). Normally, the axis of the calyces (the line drawn from the lowest to the highest calyx) of a single-system kidney is toward the contralateral shoulder. Since only the lower calyces are opacified in lower-pole reflux, the axis of the visualized calyces is altered and lies toward the ipsilateral shoulder [1]. Ureteral obstruction of the lower pole ureter can occur at the level of either the renal pelvis or the insertion of the ureter into the bladder. Ureteropelvic junction obstruction (UPJO) of the lower pole of the kidney is visualized by ultrasound [2]. Occasionally, when the upper pole is dysplastic, lower-pole UPJO can be mistaken for obstruction of a single collecting system. Ureterovesical junction obstruction (UVJO) can occur due to primary mega-ureter, or it may be secondary to the effect of a dilated obstructed upper-pole ureter. In these cases, when the upper-pole ureter is decompressed, the lower pole obstruction also resolves. The upper pole ureter inserts ectopically, medially, and inferiorly, to the normal ureteral orifice into any mesonephric duct derivative. In addition to forming the ureter, the Wolffian (mesonephric) duct contributes to the formation of the trigone of the bladder, the urethra, and the vagina in females, and the posterior urethra and genital ducts in males. The “ectopic pathway” follows the pathway created by the Wolffian duct (Fig. 3). Girls with an ectopic ureter inserting into the vagina or perineum may present with constant urinary dribbling [3] (Fig. 4). In boys, the ureter can terminate in Wolffian duct derivatives, including the seminal vesicles and vas deferens. However, ectopic ureters in boys never terminate below the urinary sphincter and thus never cause incontinence.

Fig. 3 a, b. The ectopic pathway in boys (a) shows that the ectopic ureter may insert from just below the trigone of the bladder, to the posterior wall of the uretha as low as the veromontanum, and the ejaculatory duct and its branches. The ectopic pathway in the girl (b) shows that the ectopic ureter may insert from just below the trigone of the bladder down to the posterior wall of the urethra, to the vulva and vagina

Ectopic ureters are often obstructed, usually at the level of the ureterovesical junction, but rarely reflux. If the ectopic ureter inserts into the urethra at the level of the urinary sphincter, it is both obstructed and refluxes, depending on whether the sphincter is closed or open, the so-called sphinteric ectopic ureter (Fig. 5) [4]. The more distal the ureteral insertion, the more dysplastic and dysfunctional is the associated renal parenchyma that it drains. A ureterocele is the dilated submucosal terminal segment of the ureter. It is associated with the upper-pole ureter of a double collecting system in girls. In boys, ureteroceles are rare, but when they do occur they are

245

Understanding Duplication Anomalies of the Kidney

Prolapse Fig. 7. Prolapsing ectopic ureterocele

Fig. 4. Vaginal ectopic ureter

Sphincter closed

Sphincter open (Voiding)

b

Fig. 8 a, b. Ureterocele disproportion. a A typical ectopic, obstructed, left upper pole ureter ending in ureterocele. b Ureterocele disproportion

Fig. 5. Sphincteric ectopic ureter

a

a

b

Effacement of ureterocele “Intussusception” of ureterocele Fig. 6 a, b. Effacing (a) and everting (b) ureterocele

most commonly associated with single-system kidneys and their orifices are orthotopic. These are associated with varying degrees of ureteral obstruction and pelvicalyceal dilatation. Although ureteroceles protrude into the lumen of the bladder, when the intravesical pressure equals that of the ureterocele, the ureterocele can flatten and become

imperceptible (efface) (Fig. 6a). When the intravesical pressure exceeds that of the ureterocele, the ureterocele everts or intussuscepts into its ureter (Fig. 6b). Everting ureteroceles are often confused for periureteral diverticula. Ectopic ureteroceles involving the bladder neck can prolapse into the urethra and cause obstruction of the bladder outlet (Fig. 7). Typically, ureteroceles of upper pole ureters are associated with hydroureteronephrosis. Rarely, the upper pole is diminutive and dysplastic, and the ureterocele is relatively large. This combination is described as ureterocele disproportion (Fig. 8). The condition can be mistaken for single ectopic ureters [5]. Ureteroceles can also be associated with multicystic dysplastic kidneys. As the fluid in the multiple cysts resolves, these develop the appearance of ureterocele disproportion. Very rarely, three complete ureters or three incompletely separated ureters form, resulting in complete or incomplete in ureteral triplication [6].

246

Jeanne S. Chow

In paired organs, such as the kidney, when there is a congenital anomaly in one and there is something wrong with the other, it is almost always the same anomaly, but often different in degree. Thus, if one kidney is duplex, the other is more likely to be duplex as well.

8. Ureteroceles associated with duplex kidneys and ectopic upper-pole ureters are more common in girls than in boys. In boys, ureteroceles are usually associated with single (non-duplex) kidneys. 9. Ureteroceles are dynamic.

Conclusions

References

These are the fundamental principles to remember: 1. The ureters are duplicated but the kidney is called “duplex”. 2. Ureteral duplication may be incomplete or complete. 3. The Weigert-Meyer rule applies to complete ureteral duplication and states that the upper-pole ureteral orifice is ectopic. When the ectopia is slight, the upper pole is normal. When the ectopia is moderate or severe, the upper pole is abnormal. 4. The lower pole is the analogue of a single-system kidney. 5. Lower-pole reflux can be distinguished from reflux into a single-system kidney by the axis of the calyces. 6. Ectopic ureters terminate along the ectopic pathway and can cause incontinence in girls, but never in boys. 7. Ureteroceles are caused by obstruction; they do not cause obstruction.

1. Claudon M, Ben-Sira L, Lebowitz RL (1999) Lower pole reflux in children: uroradiologic appearance and pitfalls. AJR 172:795-801 2. Fernbach SK, Zawin JK, Lebowitz RL (1995) Complete duplication of the ureter with ureteropelvic junction obstruction of the lower pole of the kidney: imaging findings. AJR 164: 701-704 3. Carrico C, Lebowitz RL (1998) Incontinence due to an infrasphincteric ectopic ureter: why the delay in diagnosis and what the radiologist can do about it. Pediatric Radiology 28:942-949 4. Wyly JB, Lebowitz RL (1984) Refluxing urethral ectopic ureters: recognition by the cyclic voidng cystourethrogram. AJR 142:1263-1267 5. Share JC, Lebowitz RL (1989) Ectopic ureterocele without ureteral and calyceal dilatation (ureterocele disproportion): findings on urography and sonography. AJR 152:567-571 6. Gill RD (1952) Triplication of the ureter and renal pelvis. J Urol 68:140-147

IDKD 2010-2013

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management Alan Daneman Department of Radiology, University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada

Introduction The midgut is that part of the bowel supplied by the superior mesenteric artery (SMA) and it extends from the mid-portion of the second part of the duodenum to the distal third of the transverse colon. The term “malrotation” is broadly used to describe the spectrum of developmental abnormalities of the midgut that are associated with abnormal rotation and/or fixation [1-3]. Malrotation may occur as an isolated entity. However, it is also associated with congenital defects in the development of the abdominal wall (e.g., omphalocele and gastroschisis) or diaphragm (e.g., congenital diaphragmatic hernia). Furthermore, it may also be associated with abnormalities of visceral situs and with certain syndromes.

Embryology During embryological development, the midgut undergoes a complicated process of growth and lengthening, herniation, rotation, reduction, and fixation [1]. Growth and lengthening of the midgut result in its herniation into the umbilical cord along the axis of the SMA, such that the apex of the herniated midgut is located at the level of the omphalomesenteric duct. This herniated bowel undergoes a 270° counterclockwise rotation, with subsequent reduction of the midgut back into the abdomen before the end of the first trimester. The proximal loop of herniated bowel (including the small bowel up to the level of the omphalomesenteric duct) is the first to reduce, followed by reduction of the distal loop (including the distal small bowel and colon to the level of the distal third of the transverse colon). The final phase involves fixation of the bowel into its final anatomical position. The cecum, however, is usually initially reduced into the upper abdomen on the right. This precedes elongation of the proximal colon until the cecum finally reaches the right lower quadrant. The latter process may only be finally achieved during early infancy. This complicated sequence is essential for the midgut to assume its normal position in the abdomen. One important aspect is the development of the normal duodenal

loop. Initially, the duodenum rotates to the right, then posteriorly, and finally to the left of the SMA, thereby completing the normal duodenal loop. The third part of the duodenum crosses the midline from right to left in the angle between the SMA (anterior to the duodenum) and the aorta (posterior). The fourth part of the duodenum ascends to the left of the spine to reach the duodeno-jejunal flexure (D-J flexure) at the ligament of Treitz. The position of the D-J flexure has been traditionally used as a landmark for documenting normal rotation of the midgut on contrast examinations of the upper gastrointestinal (GI) tract. The other landmark used is the position of the cecum in the right iliac fossa. When the D-J flexure and cecum are in their normal positions, the small bowel mesentery has a long base extending from the left upper quadrant (D-J flexure) obliquely down to the right lower quadrant (cecum). This long mesenteric base protects against the development of volvulus. Abnormalities due to the arrest of rotation and/or fixation can occur at any phase of the above-described process and may involve only a part or all of the midgut. This leads to many variations of malrotation and/or malfixation and accounts for the spectrum of clinical presentations and radiological appearances.

Causes of Symptoms The majority of these variations of malrotation and /or malfixation are associated with clinical symptoms that usually present within the first few months of life and can be life-threatening [1, 2]. Others may be associated with few or no symptoms, with the latter type often found only incidentally. Abnormalities of rotation and/or fixation usually lead to a situation in which the cecum and duodenum lie closer to each other than normal, such that the base of the small bowel mesentery is much shorter than normal. The obstruction occurs primarily in the duodenum (Figs. 1, 2) and is most often due to peritoneal (Ladd) bands that anchor the cecum to the retroperitoneum across the duodenum in the right upper quadrant. However, even more importantly, obstruction may also be due to volvulus as a re-

248

Alan Daneman

Fig. 1. Series of anteroposterior fluoroscopic images from a contrast examination of the upper GI tract in a young infant with malrotation. In the initial image (top left), before contrast was injected into the nasogastric tube, there is a non-specific bowel gas pattern with diminished gas in the mid and right abdomen. Contrast administration is followed by intermittent dilatation of the second and proximal third parts of the duodenum; initially, no dilatation is visible but subsequent images show varying degrees of dilatation. The dilated third part of the duodenum reaches only to the level of the right pedicles and does not cross the midline. The distal duodenum is on the right. The normal D-J flexure is absent. The findings are diagnostic of midgut malrotation with obstruction of the duodenum by Ladds bands and suggest a volvulus, which was confirmed at surgery. This case illustrates the potential non-specificity of the abdominal radiograph in the presence of duodenal obstruction and the intermittent nature of the duodenal dilatation

sult of the narrow base of the small bowel mesentery (Fig. 2). Much less commonly, there may be an associated internal hernia (Fig. 3). The clinical picture and imaging appearance depend on the nature and degree of the obstruction (which may be intermittent) as well as on the presence or absence of vascular compromise.

Imaging Modalities

Fig. 2. Lateral fluoroscopic view of contrast examination of the upper GI tract in a neonate with malrotation and volvulus. The proximal duodenum is dilated and ends in a beak that leads into a corkscrew pattern of non-dilated small bowel. The beak and corkscrew pattern are typical of obstruction due to volvulus, which was confirmed at surgery. The lateral view is often better than the anteroposterior view shown in Fig. 1 in demonstrating the volvulus but the latter view remains essential to depict the position of the of D-J flexure

As the clinical findings are often non-specific, pediatricians and surgeons rely on the radiologist to confirm or exclude the diagnosis [1-3]. The radiologist thus plays an exceptionally important role in the diagnosis of a malrotation and must be able to recognize the spectrum of appearances of its many variations, as depicted by any imaging modality (Figs. 1-5). Failure to do so may lead to a delay in treatment and thus potentially to bowel necrosis (which may require extensive resection) and even death. On a plain abdominal radiograph, malrotations may show a wide spectrum of appearances [3]. In contrast to what might be expected, the finding of duodenal distention with gas as a typical component of duodenal obstruction is often absent. If the duodenum is fluid-filled or

249

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management

a

b

Fig. 3. Anteroposterior abdominal radiograph from an upper GI series and follow-through contrast examination in a young teenager with midgut malrotation and an internal hernia. Contrast outlines the stomach and duodenal cap and then leads into a very well defined, rounded collection of small bowel loops on the right, representing a paraduodenal internal hernia. This appearance of contained, well defined small bowel is typical of an internal hernia

Fig. 5 a, b. Transverse sonograms of the upper abdomen in a neonate with malrotation and volvulus proven at surgery. The volvulus appears as a characteristic whirlpool in gray-scale (a) imaging and on color Doppler evaluation (b). The small bowel and mesenteric veins are coiled around the superior mesenteric artery

Fig. 4. Contrast enema in a young infant with malrotation proven at surgery. Note that the proximal colon turns back towards the left in the right upper quadrant, and the cecum and appendix lie in the upper abdomen close to the position of the duodenum. As a result of the proximity of the duodenum and cecum, the small bowel mesentery is much shorter than normal, thus predisposing to development of a volvulus

collapsed, it may not be visible and gaseous distension of the stomach alone may suggest gastric obstruction. A volvulus may lead to the presence of a soft-tissue mass due to the fluid-filled bowel, or the abdomen may be almost completely gasless. In children with severe vascular compromise due to volvulus, there is often gaseous

distention (with air fluid levels) of the entire small bowel, resembling a low bowel obstruction or ileus. The appearances of malrotation are indeed often non-specific (Fig. 1) and may be normal even in the presence of volvulus. Due to this wide variation of appearances on plain radiographs, the clinician should never rely on plain film findings to exclude malrotation. Any child in whom there is a clinical suspicion of malrotation should be studied with modalities that will directly depict the position of the bowel and/or the nature of the obstruction. These include following modalities: 1. Contrast examinations of the gastrointestinal (GI) tract: An upper GI series (alone or in combination with a follow-through series) and/or a contrast enema (Figs. 1-4) [4-8]. The most important features of the GI tract to define are the position of the D-J flexure (which requires a straight A-P view on the upper GI series) and/or the position of the cecum and proximal colon.

250

2. Cross-sectional imaging of the abdomen, particularly with sonography (but also with computed tomography and magnetic resonance) [9-15]. These modalities may depict a dilated duodenum, malposition of the D-J flexure, the whirlpool sign indicative of volvulus (Fig. 5), an internal hernia, or abnormalities of the relationship of the SMA and superior mestenteric vein. Meticulous attention to technique is critical with each of these modalities in order to delineate the relevant structures accurately. However, even when a technically perfect examination is performed, none of the above features are 100% accurate in allowing the confirmation or exclusion of malrotation. Accordingly, there has been much debate over the years as to which modality should be used first and what protocol of subsequent modalities may be required in order to enable the radiologist to rapidly make an accurate diagnosis. Independent of which modality is chosen first, the radiologist should never hesitate to ask for or perform other types of examinations aimed at obtaining more information that may increase confidence in the diagnosis – whether malrotation is present or not. As information from the various examinations is accumulated, the balance of evidence from all the examinations should be weighed together so that the correct diagnosis can be determined. Most institutions still use the upper GI series as the modality of choice in children in whom malrotation is suspected clinically [4-8]. This examination is relatively non-invasive, easy to perform, and the position of the D-J flexure is highly accurate in predicting malrotation (Fig. 1). Indeed, the presence or absence of malrotation can be readily made in most children using this modality alone. However, there will remain a group of children in whom the diagnosis of malrotation will be difficult based on the upper GI series alone, either because of technical difficulties in some patients or because of difficulties in differentiating normal variations in duodenal anatomy from true abnormalities of rotation. It is in these children that the radiologist must not be reluctant to extend the GI contrast examination or to perform another type of examination that will generate further diagnostic information. For this purpose, the next most commonly evaluated factor to determine on contrast examinations of the GI tract is the position of the cecum [4, 5]. In the acute clinical situation, particularly in neonates and young infants, the quickest way to achieve this is by performing an immediate contrast enema (Fig. 4), i.e., before too much contrast from the prior upper GI series fills the small bowel. The advantage of the contrast enema is that it relatively quickly provides information on the position of the entire large bowel. However, interpretation of the cecal position may be difficult. It must be remembered that the position of the cecum and proximal small bowel has a wide range of normality, particularly in neonates and young infants, and may well be normal even in patients with malrotation. Furthermore, the cecum and proximal

Alan Daneman

colon may be easily displaced into positions that simulate malrotation by adjacent markedly dilated loops of small bowel. This is seen particularly in neonates with congenital obstruction involving the distal small bowel. In such patients, the presence of a microcolon usually excludes malrotation, as this combination is rare. In older children, especially if the clinical setting is not acute, contrast from the original upper GI series can be followed with serial plain radiographs to determine the cecal position. This approach may take much longer than performing a contrast enema and is usually less fruitful in the neonate and young infant, since in this age group it is not always possible to clearly depict the position of the cecum. In recent years, signs of malrotation have been visualized using sonography and other cross-sectional imaging modalities, but they are still not used as the modality of choice by most radiologists [9-15]. It is true, however, that sonography is being used much more frequently than suggested in the literature. Direct visualization of a volvulus (whirlpool sign) (Fig. 5) may obviate the necessity for contrast examination of the GI tract; but this sign is not present in those children without volvulus but who are symptomatic because of obstruction due to bands. Furthermore, the sign may be difficult to appreciate in children with volvulus and a large amount of dilated, gas-filled bowel. Inversion of the superior mesenteric artery and vein relationship may be present in normal rotation, and a normal relationship may be present in children with malrotation. Therefore, a normal sonogram does not exclude malrotation and, to date, sonography has not been used as a screening procedure for this condition. Nevertheless, it is essential that radiologists be able to recognize these abnormal signs when they are detected as an incidental or unexpected finding on cross-sectional imaging. Yousefzadeh [15] has drawn attention to the sonographic depiction of the third part of the duodenum (D3) in the angle between the SMA and the aorta, and has suggested that documentation of the presence of the D3 in this normal position excludes the presence of malrotation. This is an extremely interesting approach because, if shown to be accurate, then the sonographic depiction of D3 in this position would obviate the necessity for doing a fluoroscopic, contrast examination of the upper GI tract in those children suspected of having this condition. However, to date, there are no data to substantiate this suggestion. Filling the duodenum with fluid administered orally or through a feeding tube may facilitate delineation of the position of the duodenum and D-J flexure on sonography. This approach has been advocated by some as the technique of choice, and some groups have used it quite extensively but have yet to publish their data. Gent and LeQuesne presented their experience with this technique, based on over 100 cases, at the World Federation of Ultrasound in Medicine and Biology Meeting in Sydney in 2009. These authors illustrated their technique, which

251

Malrotation: Techniques, Spectrum of Appearances, Pitfalls, and Management

delineates the anatomy of the duodenum exquisitely. They have been able to easily confirm the diagnosis of malrotation in the vast majority of patients, with no falsenegatives. In a few patients in whom the diagnosis was unsure, an upper GI series was required for confirmation. To the best of this author’s knowledge, there has, to date, been no large single series comparing the accuracy of this technique to the routine contrast upper GI series. Color Doppler sonographic assessment of bowel perfusion may be useful in patients in whom bowel perfusion is in doubt. However, its role is limited as poor perfusion is usually seen in sicker patients and surgeons are keen to get these children to the operating room as soon as possible without necessarily relying on the need for diagnostic imaging.

Conclusions Accurate diagnosis of malrotation requires (i) a high index of suspicion, (ii) an understanding of normal variations of the GI tract, (iii) an appreciation of the spectrum of appearances of malrotation on imaging modalities, and (iv) attention to meticulous technique when performing GI contrast examinations or cross-sectional imaging. Although the diagnosis of malrotation can be accurately made in many instances based on a single examination, the radiologist should never be reticent about using more than one examination (or repeating examinations) when attempting to confirm or exclude malrotation in difficult cases. As information is accumulated from the various diagnostic examinations, the balance of evidence derived from all of them should be weighed together to determine the correct diagnosis. If uncertainty remains after all imaging avenues have been exhausted, as does happen (though uncommonly), then the decision whether to perform laparoscopy or to operate should be left to the surgeon.

References 1. Strouse PJ (2000) Disorders of intestinal rotation and fixation (“malrotation”). Pediatr Radiol 34:837-851 2. Lampl B, Levin TL, Berdon WE, Cowles RA (2009) Malrotation and midgut volvulus: a historical review and current controversies in diagnosis and management. Pediatr Radiol 39:359-366 3. Daneman A (2009) Malrotation: the balance of evidence. Pediatr Radiol 39(2):S164-S166 4. Long FR, Kramer SS, Markowitz RI et al (1996) Intestinal malrotation in children: Tutorial on radiographic diagnosis in difficult cases. Radiology 198:775-780 5. Long FR, Kramer SS, Markowitz RI, Taylor GE (1996) Radiographic Patterns of Intestinal Malrotation in Children. RadioGraphics 16:547-556 6. Katz ME, Siegel MJ, Shackelford GD, McAlister WH (1987) The Position and Mobility of the Duodenum in Children. AJR 148:947-951 7. Donnolly LF, Rencken IO, de Lorimier AA, Gooding CA (1996) Left paraduodenal hernia leading to ileal obstruction. Pediatr Radiol 26:534-536 8. Manji R, Warnock GL (2000) Left paraduodenal hernia: an unusual cause of small-bowel obstruction. Can J Surg 44:455-457 9. Dufour D, Delaet MH, Dassonville M et al (1992) Midgut malrotation, the reliability of sonographic diagnosis. Pediatr Radiol 22:21-23 10. Chao HC, Kong MS, Chen JY et al (2000) Sonographic features related to volvulus in neonatal intestinal malrotation. J Ultrasound Med 19:371-376 11. Shimanuki Y, Aihara T, Takano H et al (1996) Clockwise whirlpool sign at color Doppler US: an objective and definite sign of midgut volvulus. Radiology 199:261-264 12. Yoo SJ, Park KW, Cho SY et al (1999) Definitive diagnosis of intestinal volvulus in utero. Ultrasound Obstet Gynecol 13:200-203 13. Loyer E, Eggli KD (1989) Sonographic evaluation of superior mesenteric vascular relationship in malrotation. Pediatr Radiol 19:173-175 14. Weinberger E, Winters WD, Liddell RM et al (1992) Sonographic diagnosis of intestinal malrotation in infants: importance of the relative positions of the superior mesenteric vein and artery. AJR Am J Roentgenol 159:825-828 15. Yousefzadeh DK (2009) The position of the duodenojejunal junction: the wrong horse to back on in diagnosing or excluding malrotation. Pediatr Radiol 39:S172-S177

IDKD 2010-2013

Pediatric Intestinal Ultrasonography Simon G. Robben Department of Radiology, Maastricht University Medical Centre, Maastricht, The Netherlands

Introduction Ultrasonography (US) is the imaging modality of choice for the initial evaluation of diseases in children for many reasons. First, it is relatively inexpensive and patient friendly. Second, it lacks radiation and motion artifacts. Third, the small size of the child compensates for the limited penetration of sound waves and facilitates the use of high-frequency transducers. Fourth, flow studies are possible using the Doppler mode, while realtime imaging allows the visualization of movements such as peristalsis. Moreover, US involves direct contact with the patient, thus offering a unique opportunity to ask specific questions and to perform additional physical examinations, emphasizing the role of the radiologist as a clinician. Accordingly, ultrasonography has become the most important imaging technique in children and can be considered as the workhorse of pediatric radiology. Initially, US of the stomach and intestines was not popular for obvious reasons: bowel gas has the annoying characteristic of reflecting all sound waves or creating artifacts because of its abnormally low acoustic impedance.

a

However, increased knowledge, improved technique (e.g., graded compression), improved hardware (highfrequency transducers), and improved software (adaptive imaging, compound imaging) have resolved the limitations to US use. Nowadays it is impossible to imagine US without intestinal US, especially in pediatrics! The hallmark of intestinal US is the “gut signature”, i.e., the characteristic appearance of the layers of the gut (Table 1 and Fig. 1).

Table 1. Gut signature, characterized by alternating hyper- and hypoechoic layers Layer

Echogenicity

Mucosal surface Mucosa Submucosa Muscularis propria Serosal surface

Hyperechoic Hypoechoic Hyperechoic Hypoechoic Hyperechoic

b

Fig. 1 a-c. Gut signature in a stomach, b ileum, and c appendix (between arrows)

c

253

Pediatric Intestinal Ultrasonography

Gastroesophageal Junction The gastroesophageal junction (Fig. 2) can be visualized with US in 87-95% of children with suspected gastroesophageal reflux disease (GERD), and reflux of gastric content into the esophagus can be demonstrated ultrasonographically [1, 2]. A threshold of three reflux periods within 10 min corresponds to a sensitivity of 82% and a specificity of 85% for GERD [1]. Farina et al. increased the sensitivity to 98% using color Doppler US [3]. However, in premature infants the results differ: the sensitivity is 38% and the specificity is 100% compared to 24-hour pH-metry [4]. US studies have also shown that there is a positive correlation between GERD and the length of the abdominal esophagus, protrusion of the gastric mucosa, and an increased gastroesophageal angle (angle of His). Therefore, US can be the initial imaging technique in children suspected of having GERD. Moreover, it can evaluate the post-operative situation after a fundoplication and gastric emptying, another important contributing factor to the pathogenesis of GERD.

palpation is accurate but not always successful as it depends on factors such as the experience of the examiner, the presence of gastric distension, and the cooperation of the infant. In virtually all patients, US is very accurate in facilitating the diagnosis and therefore plays a key role in the initial care of these infants. It is important that the radiologist understands the anatomical changes of the pyloric channel in affected infants, as demonstrated by US. Pyloric muscle hypertrophy is shown to a variable degree during the US examination. In addition, a certain amount of thickening of the mucosa is present (Fig. 3). A muscle thickness that is consistently v3 mm is considered to be diagnostic of IHPS, although some clinicians have stated that the overall morphological and dynamic impression, including length of the pyloric canal as well as relaxation and peristalsis, are just as important. The US examination can be performed in a very short time with an accuracy approaching 100%.

Stomach Gastric emptying in vomiting patients can be evaluated with US, which depicts dynamic and anatomical abnormalities. Infantile hypertrophic pyloric stenosis (IHPS) [5] is a condition of unknown etiology that affects young infants ages 2-8 weeks and with a male-to-female ratio of approximately 4:1. In IHPS, the antropyloric portion of the stomach becomes abnormally thickened and manifests as an obstruction to gastric emptying. Typically, infants with IHPS are clinically normal at birth but during the first few weeks of postnatal life, they develop non-bilious “projectile” vomiting that leads to weight loss, dehydration and hypochloremic alkalosis, and eventually death. Surgical treatment is curative. The clinical diagnosis relies on palpation of the thickened pylorus. Abdominal

a

Fig. 2 a, b. Sagittal slice of a normal gastroesophageal junction (arrows) in a 3month-old boy. b Same slice during transit of air. During real-time ultrasonography, the direction of transit can easily be appreciated. S Stomach, L liver

Fig. 3. Hypertrophic pyloric stenosis (between large arrows). Muscle thickening is indicated by small arrows. A Antrum of stomach with retained fluid

b

254

Simon G. Robben

Foveolar hyperplasia consists of a polypoid thickening of the mucosal layer. It may be seen after long-standing prostaglandin therapy, hypertrophic gastropathy, or cow’s milk allergy, but it may also be idiopathic. While the condition may simulate pyloric hypertrophy, on closer US examination the obstruction will appear to be caused by thickened mucosa instead of thickened muscle. Rare causes of gastric outlet obstruction include pylorospasm, (eosinophilic) gastritis, food allergy, chronic granulomatous disease, hyperlipidemia, duplication cysts, ectopic pancreas, benign and malignant tumors, and bezoars.

Small Bowel Conventional radiography and US are the initial imaging modalities in children with abdominal pain or obstruction. The most important additional value of US over conventional abdominal radiographs in these children is its capability to visualize peristalsis, vascularity, bowel wall characteristics, dilatation of fluid-filled loops, and extra-intestinal abnormalities, e.g., ascites and other fluids. The jejunum and ileum can be distinguished from the colon based on anatomical location, caliber, contents, folds, and peristalsis. Anatomical location: The colon has a peripheral location in which the ascending and descending colon lie dorsally in both flanks and the transverse colon is located ventrally in the upper abdomen. The sigmoid colon traverses the left psoas muscle and courses into the pelvis whereas the small bowel has a more central position. Caliber: The diameter of the small bowel is small while, as its name indicates, the diameter of the large bowel is relatively large.

a

Contents: The small bowel is either empty or filled with liquid contents but little air, whereas the colon is generally filled with gas-filled bulky stools. Folds: The folds in the jejunum are more numerous, longer, thinner, and closer together than the ileal folds. In the terminal ileum, the mucosa may be thickened due to hyperplasia of lymphoid tissue. The colon is recognized by its haustrations. Peristalsis: The small bowel moves continuously due to persistaltic waves whereas the colon shows sparse movements. Baud proposed a systematic US approach for identifying small bowel disease, based on wall thickening [6]. 1. Determine wall thickening: normal (f3 mm), mild (36 mm), moderate (6-9 mm), or severe (>9 mm). 2. Determine location (proximal or distal) and extent (focal 5 cm, segmental 6-40 cm, or diffuse >40 cm). 3. Determine stratification. The bowel wall is stratified when the hyperechogenicity of the submucosa is preserved and the mucosa, submucosa, and muscularis propria are visible as separate layers. Non-stratification implies the absence of distinction between mucosa and submucosa or between all three layers (Fig. 4). 4. Determine the valvular fold pattern: normal, thickened, thumb-printing, and hyperplastic valvular folds. In general, thickened small bowel loops show decreased peristalsis and contain little air. They are therefore easily visualized and measured. At least three patterns can be distinguished. Stratified thickening of the small bowel is found in infectious ileitis, advanced appendicitis, early Crohn’s disease, and graft versus host disease. Non-stratified thickening occurs in Henoch-Schönlein purpura, advanced Crohn’s disease, tuberculous ileitis,

b

Fig. 4 a, b. Difference between a stratified wall thickening in early Crohn’s disease and b non-stratified wall thickening in advanced Crohn’s disease

255

Pediatric Intestinal Ultrasonography

3. Determine stratification, which is best seen on transverse images. The bowel wall is stratified when the hyperechogenicity of the submucosa is preserved and the mucosa, submucosa, and muscularis propria are visible as separate layers. Non-stratification implies an absence of distinction between the mucosa and submucosa or between all three layers. 4. Determine the haustral pattern, which is best seen on longitudinal images; also, confirm the absence or presence of the haustral folds and their length (normal or shortened) and aspect. These criteria distinguish three patterns. Fig. 5. Two benign small bowel intussusceptions in a patient with a malabsorption syndrome

protein-losing enteropathy, hereditary angioedema, ischemia, celiac disease, Burkitt’s lymphoma, Kawasaki’s disease, and viral enteritis. Non-stratified thickening with hyperplastic valvular folds can be seen in viral (and sometimes bacterial) lymphoid hyperplasia and Yersinia ileitis. Malrotation and midgut volvulus are discussed in the chapter by Daneman. Benign small bowel intussusception (BSBI) is a recently described entity. It differs from the classical symptomatic ileocolic intussusception in that it occurs predominantly in the right lower quadrant or periumbilical region, has a smaller diameter (mean diameter 1.4 vs. 2.5 cm), a thinner outer rim, and does not contain mesenteric lymph nodes (Fig. 5) [7]. Moreover, peristalsis in the intussuscepted loop persists, in contrast with ileocolic intussusception. Often, BSBI is an incidental finding but it occurs with increased frequency and number in celiac disease. In general, BSBI does not need immediate reduction because of its spontaneously resolving nature. Appendicitis, Meckels diverticulum, necrotizing enterocolitis, and ileocolic intussusception are discussed elsewhere in this volume.

Large Bowel The differences between the normal small and large bowel were described above. In addition to the systematic approach to the small bowel, Baud proposed an analogous approach to the colon [8]. 1. Determine wall thickening: normal (f3 mm), mild (36 mm), moderate (6-9 mm), or severe (>9 mm). 2. Determine the extent and location of disease: diffuse, cecum, ascending colon, proximal and distal transverse colon, descending colon, or rectosigmoid.

Stratified thickening is found in infectious colitis, advanced appendicitis, and inflammatory bowel disease (ulcerative colitis and Crohn’s disease). Non-stratified thickening with loss of haustral folds is found in early hemolytic uremic syndrome (HUS) and advanced Crohn’s disease. Non-stratified thickening with preservation of normal haustral fold length is found in pseudomembranous colitis and neutropenic colitis (typhlitis).

Rectum In several studies, US was used to measure the transverse diameter of the rectum. This seems to be a reliable approach to identifying rectal impaction and may replace digital rectal examination. All children with rectal impaction on digital examination were ultrasonographically shown to have had a rectal diameter >30 mm [9]. Moreover, several studies reported that in children with constipation the mean diameter of the rectum is significantly larger than in normal children. In a series of 225 children, the mean rectal diameter in normal children was 32 mm (SD 8.2) and in children with constipation 43 mm (SD 9.7) [10]. To overcome the problem of agedependency of the rectal diameter in normal children, Bijos et al. proposed the rectopelvic ratio, defined as the ratio of the rectal diameter (as determined with US) to the distance between the anterior superior iliac spines [10]. A rectopelvic ratio >0.189 corresponds to a sensitivity for rectal impaction of 88% compared to proctoscopy. The rectal diameter can also be used to monitor therapy.

Anus Anal atresia is a relatively frequent congenital abnormality in which the anus is absent and a rectoperineal, rectovestibular, rectovaginal, rectourethral, or rectovesical fistula may be identified in almost all cases. Since the fistula demonstrates an internal sphincter, some clinicians instead prefer the term “ectopic anus” or “anorectal

256

malformation”. During pre-operative evaluation, it is important to assess the type of anal atresia, which may be high (distal rectal pouch above the puborectal sling), intermediate (at the sling), or low (through the sling). Transperineal US is a good diagnostic modality for defining the type of anal atresia as it can be used to measure the distance between the rectal pouch and the perineum (P-P distance). A P-P threshold value of 15 mm discriminates the low type of atresia from the intermediate and high types with a sensitivity of 100% and a specificity of 86% [11]. Moreover, internal fistula can be correctly identified in 82% of the patients with the high type of anal atresia [12]. Transperineal sonography is also a useful method for differentiating between an anteriorly displaced anus, which is a normal anatomical variant, and a low-type imperforate anus with perineal fistula, which is a pathological developmental abnormality requiring surgical repair [13]. In adults, transperineal ultrasonography is a simple, painless, cost-effective and real-time method to detect and classify perianal fluid collections, abscesses, fistulas, and sinus tracts [14-16]. These data can probably be extrapolated to the pediatric population, e.g., children with Crohn’s disease.

Cystic Intestinal Masses Cystic intra-abdominal masses originating from the alimentary canal are increasingly recognized because of the advent of routine prenatal US. These masses can be divided into cysts originating from solid organs (mesenchymal hamartoma, congenital splenic cyst, pancreatic pseudocyst, pancreatic cystadenoma, hydronephrosis, multicystic dysplastic kidney, multilocular cystic nephroma, adrenal hemorrhage, ovarian cysts and cystic neoplasms, hematocolpos, urachal cysts, abdominal and sacrococcygeal teratoma, and cerebrospinal fluid pseudocyst) and those originating from the alimentary canal and its appendages (hydrops of the gallbladder, choledochal cyst, mesenteric and omental cysts, gastrointestinal duplication cyst, meconium pseudocyst, and appendiceal abscess) [17] (Table 2).

Simon G. Robben

When a cystic mass is found on US examination, it should be evaluated for its size, shape, location, relation to organs, and contents and wall characteristics. In the majority of cases, US can provide a specific diagnosis or offer a narrow differential diagnosis [18]. Hydrops of the gallbladder is a rare cause of a right upper quadrant mass in children. It has been described in the absence of stones, infections, or congenital abnormalities, in which case it is probably caused by transient obstruction of the cystic duct or increased mucus secretion with ineffective emptying. It has also frequently been associated with Kawasaki’s disease. Rarely, it is associated with childhood infections, Henoch-Schönlein purpura, Cryptosporidium infection in immunocompromised children, Epstein-Barr virus infections, and typhoid fever. US examination reveals a dilated anechoic elliptical gallbladder without wall thickening. The bile ducts are not dilated and no stones are seen [17, 19]. Choledochal cysts are actually focal cystic dilatations of the biliary tree. Most patients present in the first decade of life with symptoms of episodic abdominal pain, mass, and jaundice. US is the best initial method of evaluating dilatation of the bile ducts. On US, the choledochal cyst is located in the porta hepatis, separate from the gallbladder, with bile duct(s) leading into or out of it. The entire biliary tree should be evaluated but intrahepatic bile duct dilatation may be absent. Surgical resection is necessary to prevent the development of ascending cholangitis, stones, or malignant degeneration [17, 18, 20]. Mesenteric cysts are cystic lymphangiomas that are most often found in the small bowel mesentery, especially the ileal mesentery but also in the greater and lesser omentum and occasionally in the mesenteric root and retroperitoneum. About one-third occur in children younger than 15 years. Newborns present with abdominal distention and a palpable mass whereas children are much more likely to present with pain, anorexia, vomiting, or fever. US can characterize the mass as a typical thin-walled unior multilocular cyst that displaces adjacent structures to the periphery of the abdomen (Fig. 6). Calcification of

Table 2. Cystic masses of the gut and gut-related structures Biliary system Choledochal cyst Hydrops of gallbladder Gastrointestinal tract Mesenteric cyst/lymphangioma Enteric/duplication cyst Omental cyst Meconium pseudocyst Miscellaneous Abscess Teratoma Sacrococcygeal teratoma

Fig. 6. Multilocular cystic lymphangioma

257

Pediatric Intestinal Ultrasonography

the wall is rare in children. The cyst content is anechoic but if hemorrhage has occurred then debris may be seen. Lymphangiomas may be so large that they are difficult to distinguish from severe ascites. However, ascites separates individual bowel loops and fills the perihepatic and perisplenic spaces [17, 18]. Gastrointestinal duplication cysts are spherical or tubular masses adherent to the gastrointestinal (GI) tract and sometimes communicating with it [17, 18, 21]. These cysts are lined with intestinal epithelium and contain smooth muscle within their walls. They may occur anywhere along the alimentary tract but the most common location is the ileum, followed by the stomach. Most patients present within the first year of life; their symptoms include GI obstruction and, less commonly, a palpable mass, intussusception, and abdominal distension. The cystic nature of duplication cysts can be easily appreciated on US. The content of the cyst is often anechoic but there may be debris after hemorrhage or due to mucoid material. Rarely, the cyst appears completely hyperechoic after hemorrhage. Two signs are virtually diagnostic of duplication cysts: 1. a double-layered wall consisting of echogenic mucosa and hypoechoic muscularis propria (Fig. 7); 2. peristalsis in the cyst. In 15-20% of cases, the cyst contains gastric mucosa, which accounts for the above-mentioned hemorrhages. Meconium pseudocyst is a manifestation of the cystic type of meconium peritonitis that results from in utero bowel perforation. Bowel perforation may be secondary to intestinal obstruction (meconium ileus, atresia, or volvulus) or idiopathic. The spilled meconium is encapsulated and forms a large meconium-filled (hyperechoic) cyst that is lined by a thick inflammatory and fibrotic membrane, often containing calcifications. The perforation may still communicate with the cyst postnatally. Dilated fluid-filled bowel loops may be seen next to the cyst if obstruction is still present. Distant peritoneal or scrotal calcifications are additional evidence for meconium peritonitis and may also be demonstrated with US or conventional radiographs [17, 18, 22]. Intra-abdominal abscesses often are complications of appendicitis, inflammatory bowel disease, or intra-abdominal surgery. However, they can also occur in patients on immunosuppressive therapy or in those with AIDS or chronic granulomatous disease. The typical presentation is intermittent fever, increased white blood cell count or C-reactive protein, and abdominal tenderness. US shows an ill-defined, fluid-filled mass with a thickened wall and irregular inner surface. The fluid (pus) demonstrates septations, debris, or debris-fluid levels. The presence of gas bubbles in the pus often makes differentiation from bowel loops difficult; the presence of peristalsis rules out an abscess. If feasible, ultrasonographically guided aspiration and/or drainage can be performed, if necessary in the ICU.

a

b

Fig. 7 a, b. Duplication cyst. a Large right-sided cyst on prenatal magnetic resonance imaging shows no typical features. b Postnatal ultrasound demonstrates gut signature in the wall of the cyst suggestive of duplication cyst

Abdominal teratomas usually arise from the retroperitoneum, presacral region, or the ovary, whereas mesenteric and gastric teratomas are rare. These tumors occur mainly in children, and 80% are benign. Mature, immature, and malignant types have been described. Most children present with a palpable abdominal mass that on US is seen to be either (multi)cystic or solid. Teratomas typically contain calcifications and fat. The latter is difficult to appreciate with US but highly echogenic tissue is suggestive of fatty tissue.

258

Simon G. Robben

Sacrococcygeal teratomas usually present as a large external cystic tumor located at the coccyx but in some cases consist of a large cystic intra-abdominal mass without any external mass. These pre-sacral tumors obstruct the rectum and bladder. A pre-sacral extension of an intra-abdominal cyst provides a clue to the diagnosis of sacrococcygeal teratoma. Cystic tumors carry a better prognosis than do solid, hypervascular tumors [17, 18].

ileum. The bowel wall thickening is extensive, asymmetrical, and poorly stratified whereas the mesenteric involvement is bulky and lobulated and appears to be in continuity with the bowel wall. Despite the extensive bowel wall thickening, the lumen may remain wide. Burkitt’s lymphoma can lead to intestinal obstruction and intussusception. It can also involve the liver, spleen, kidneys and pancreas. Extensive involvement of the omentum and peritoneum is rare [25, 26].

Cerebrospinal fluid pseudocyst (liquor cyst) is a complication resulting from ventriculoperitoneal shunt, with a frequency of approximately 3%. Risk factors for pseudocyst formation are related to inflammatory processes and CNS tumors. Pseudocysts tend to occur within 6 months of the last abdominal surgical procedure. Children present with abdominal pain, distention, or mass. US will demonstrate a sonolucent, well-defined mass in noninfected cysts, whereas infected cysts show septa, internal debris, and fluid-fluid levels. There is no statistically significant correlation between pseudocyst size and the presence of infection. It is important to identify the tip of the shunt within the cyst, producing the characteristic “railroad sign” [17, 23, 24].

Intra-abdominal lipomatous tumors (lipoma, lipoblastoma, and the rare liposarcoma) predominantly involve the mesentery and omentum [27-29]. They show hypo- or hyperechoic textures and are finely lobulated, homogeneous, or with fibrovascular septa. In most cases, magnetic resonance imaging is necessary for further evaluation.

Solid Intestinal Masses The many types of solid masses (often of neoplastic origin) that can be found during US examination are summarized in Table 3.

Inflammatory myofibroblastic tumor (inflammatory pseudotumor) most commonly occurs in the mesentery of children or young adults [29, 30]. The typical complaints are of fever, malaise, weight loss, or abdominal pain. The US characteristics of inflammatory pseudotumor are nonspecific: solid, well-defined (sometimes lobulated), with mixed echo-texture and frequent calcifications. Infiltration of the adjacent bowel may occur. Prominent vascularity may be shown with Doppler US.

Lymphoma Burkitt’s lymphoma Non-Hodgkin’s lymphoma

Fibromatosis or abdominal desmoid is part of the clinical-pathological spectrum of deep fibromatoses [29, 30]. The latter encompass a group of benign fibroproliferative processes that are locally aggressive and have the capacity to infiltrate or recur but not to metastasize. Mesenteric structures are the most common sites of origin of intraabdominal fibromatosis. Other locations are the abdominal wall, pelvis, and retroperitoneum. Thirteen percent of patients with mesenteric fibromatosis have familial adenomatous polyposis (FAP), specifically, the Gardner syndrome variant. In these patients, prior abdominal surgery is an important risk factor for the development of mesenteric fibromatosis. The US appearance is a solid, wellcircumscribed mass of variable echo-texture and homogeneity. Locally aggressive fibromatosis infiltrates the mesenteric fat.

Peritoneal, mesenteric, and omental Lipoma, lipoblastoma, and liposarcoma Inflammatory myofibroblastic tumor Fibromatosis (= desmoids) Neurofibromas (NF1-associated) Rhabdomyosarcoma Metastasis Mesenteric lymphadenitis

Neurofibromatous tumors are associated with neurofibromatosis type 1 (NF1). Abdominal involvement is found in 10-25% of patients with NF1, regardless of their age. Intra-abdominal neurofibromas present as hypoechoic heterogeneous masses or as multiple rounded, hypoechoic, well-circumscribed, variably sized mesenteric nodules.

Small and large bowel Juvenile colonic polyps Hamartomatous small bowel polyps Polyposis Carcinoid Vascular malformations

Rhabdomyosarcomas are rare intra-abdominal pediatric tumors involving the mesentery, peritoneum, or omentum, often associated with ascites. Leung et al. described an omental embryonal rhabdomyosarcoma consisting of lobulated round masses surrounded by tissue with a

Burkitt’s lymphoma is a fast-growing and aggressive malignant neoplasm predominantly affecting children and that may be associated with immunodeficiency. US will demonstrate bowel wall thickening combined with a mesenteric mass, predominantly of the cecum and distal

Table 3. Solid masses of the gut and gut-related structures

259

Pediatric Intestinal Ultrasonography

cerebriform pattern [31]. Rhabdomyosarcomas may also occur in the biliary tree, and are actually the most common pediatric tumor of the biliary tree. US findings are nonspecific but if a solid tumor in the biliary tree is detected, it is highly suggestive of rhabdomyosarcoma [32].

a

Intestinal polyps in the small bowel are usually hamartomatous polyps in patients with Peutz-Jeghers syndrome. These polyps occur more often in the jejunum than in the ileum. US is only able to detect polyps >15 mm and is therefore not a screening tool for the detection of uncomplicated polyps. However, it is valuable in diagnosing complications associated with polyps of the small bowel, e.g., intussusceptions. b

Juvenile polyps are the most common neoplasms of the large bowel in children. The sigmoid and rectum are preferential locations. The presenting symptom is rectal bleeding in over 90% of patients. Occasionally, a colo-colic intussusception is the first manifestation of a juvenile polyp. US demonstrates a pedunculated spherical nodule 10-25 mm in diameter and containing multiple 2- to 3-mm cysts. Administration of fluid within the bowel lumen will greatly improve the visualization of these polyps; however, the assessment of rectal polyps is unreliable. Mesenteric lymphadenitis. Abdominal lymph nodes varying from 0 to 10 mm in diameter are detected in almost all asymptomatic children (Fig. 8a). Approximately half of these lymph nodes (43-54%) are >5 mm in their short axis, with the number and shape of the nodes being age-independent [33, 34]. There is a gender dependency in that boys are more affected than girls [33, 35]. In children with recurrent or acute abdominal pain without known cause, the lymph nodes are significantly larger than in asymptomatic children, but there is considerable overlap between the two groups. Children with abdominal pain and abdominal lymph nodes >10 mm in their short axis may be considered as having primary mesenteric lymphadenitis, if no other acute inflammatory process is identified [34, 36]. There are few US criteria that are typical for a specific causative agent. The presence of calcifications and necrosis suggests tuberculous lymphadenitis (Fig. 8b).

Neonatal Bowel Obstruction Many neonatal bowel obstructions are caused by diseases discussed in other chapters, e.g., duplication cysts, malrotation or volvulus, meconium peritonitis, necrotizing enterocolitis, anal atresia, and pyloric hypertrophy. Additional causes of neonatal obstruction include various types of atresia, annular pancreas, meconium ileus, Hirschsprung’s disease, and meconium plug syndrome. These diseases are diagnosed by conventional abdominal radiographs and/or conventional contrast studies but US may be of additional value [37].

Fig. 8 a, b. Enlarged mesenteric lymph nodes. a Asymptomatic child shows lymph nodes with normal texture and ellipsoid shape; the largest node has a short axis of 8 mm. b Tuberculous mesenteric lymph node shows indistinct margins, round shape, and heterogeneous texture with hypoechoic areas of necrosis and a diameter of 20 mm

Atresias are congenital interruptions of the lumen of the alimentary canal caused by a failure of canalization during organogenesis. The result is dilatation proximal to the atresia and complete collapse distal to the atretic segment. While the clinical and radiographic presentation is typical, US may be used to demonstrate accompanying pathology: 1. exclusion of associated malrotation in cases of high obstruction; 2. evaluation of the colon and rectum in cases of high obstruction; single atresias are associated with a normalsized colon, multiple atresias, with microcolon; 3. evaluation of associated malformations of the heart, kidneys, and biliary tree; 4. demonstration of extrinsic duodenal compression in cases of high obstruction, e.g., duplication cysts, preduodenal portal vein, annular pancreas; and 5. demonstration of signs of prenatal meconium peritonitis as a cause of small bowel atresia (meconium cysts, subtle peritoneal or scrotal calcifications).

260

Meconium ileus is a small bowel occlusion caused by inspissated, abnormally tenacious, meconium in the distal ileum, invariably associated with cystic fibrosis. It occurs in 10-20% of neonates with cystic fibrosis. US may demonstrate the microcolon (mean diameter of 4 mm), small bowel dilatation with echogenic contents (in contrast to the hypoechoic contents in ileal atresia), pseudothickening of the bowel wall (caused by a circular layer of tenacious meconium adherent to the bowel wall), granular pattern (air bubbles in the tenacious meconium), and detection of meconium pellets in the distal ileum. Hirschsprung’s disease is caused by an absence of ganglion cells, which results in abnormal gut motility and a lack of gut relaxation. The length of the aganglionic segment is variable but always involves the distal end of the intestinal tract. In a small number of patients, the entire colon and even the ileum and jejunum are involved. US is of limited value because of the air-artifacts in dilated bowel loops. However in very early neonatal US examination, air is not yet present and distention of the colon in a neonate with distal occlusion suggests Hirschsprung’s disease. Meconium plug syndrome can be considered as neonatal constipation. US demonstrates moderate dilatation of the entire colon without a transition zone, echogenic colonic content, and dilatation of the proximal small bowel [37].

Conclusions Ultrasonography is a reliable initial imaging technique to evaluate a variety of gastrointestinal pediatric diseases and malformations. A solid knowledge of the pathophysiology of these conditions is necessary to understand their ultrasonographic manifestations.

References 1. Couture A, Baud C, Ferran J et al (2008) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg 2. Westra SJ, Derkx HH, Taminiau JA (1994) Symptomatic gastroesophageal reflux: diagnosis with ultrasound. J Pediatr Gastroenterol Nutr 19:58-64 3. Farina R, Pennisi F, La Rosa M et al (2008) Contrast-enhanced colour-Doppler sonography versus pH-metry in the diagnosis of gastro-oesophageal reflux in children. Radiol Med 113:591-598 4. Pezzati M, Filippi L, Psaraki M et al (2007) Diagnosis of gastro-oesophageal reflux in preterm infants: sonography vs. pHmonitoring. Neonatology 91:162-166 5. Hernanz-Schulman M (2003) Infantile hypertrophic pyloric stenosis. Radiology 227:319-331 6. Baud C (2008) Small bowel thickening. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 253-296 7. Park NH, Park SI, Park CS et al (2007) Ultrasonographic findings of small bowel intussusception, focusing on differentiation from ileocolic intussusception. Br J Radiol 80:798-802

Simon G. Robben

8. Baud C (2008) Infectious and inflammatory colitis. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 297-339 9. Joensson IM, Siggaard C, Rittig S et al (2008) Transabdominal ultrasound of rectum as a diagnostic tool in childhood constipation. J Urol 179:1997-2002 10. Bijos A, Czerwionka-Szaflarska M, Mazur A, Romanczuk W (2007) The usefulness of ultrasound examination of the bowel as a method of assessment of functional chronic constipation in children. Pediatr Radiol 37:1247-1252 11. Haber HP, Seitz G, Warmann SW, Fuchs J (2007) Transperineal sonography for determination of the type of imperforate anus. AJR Am J Roentgenol 189:1525-1529 12. Choi YH, Kim IO, Cheon JE et al (2009) Imperforate anus: determination of type using transperineal ultrasonography. Korean J Radiol 10:355-360 13. Haber HP, Warmann SW, Fuchs J (2008) Transperineal sonography of the anal sphincter complex in neonates and infants: differentiation of anteriorly displaced anus from low-type imperforate anus with perineal fistula. Ultraschall Med 29:383-387 14. Bonatti H, Lugger P, Hechenleitner P et al (2004) Transperineal sonography in anorectal disorders. Ultraschall Med 25:111-115 15. Maconi G, Ardizzone S, Greco S et al (2007) Transperineal ultrasound in the detection of perianal and rectovaginal fistulae in Crohn's disease. Am J Gastroenterol 102:2214-2219 16. Stewart LK, McGee J, Wilson SR (2001) Transperineal and transvaginal sonography of perianal inflammatory disease. AJR Am J Roentgenol 177:627-632 17. Wootton-Gorges SL, Thomas KB, Harned RK et al (2005) Giant cystic abdominal masses in children. Pediatr Radiol 35:1277-1288 18. Khong PL, Cheung SC, Leong LL, Ooi CG (2003) Ultrasonography of intra-abdominal cystic lesions in the newborn. Clin Radiol 58:449-454 19. Maurer K, Unsinn KM, Waltner-Romen M et al (2008) Segmental bowel-wall thickening on abdominal ultrasonography: an additional diagnostic sign in Kawasaki disease. Pediatr Radiol 38:1013-1016 20. Kim OH, Chung HJ, Choi BG (1995) Imaging of the choledochal cyst. Radiographics 15:69-88 21. Segal SR, Sherman NH, Rosenberg HK et al (1994) Ultrasonographic features of gastrointestinal duplications. J Ultrasound Med 13:863-870 22. Yang WT, Ho SS, Metreweli C (1997) Case report: antenatal sonographic diagnosis of meconium peritonitis and subsequent evolving meconium pseudocyst formation without peritoneal calcification. Clin Radiol 52:477-479 23. Pathi R, Sage M, Slavotinek J, Hanieh A (2004) Abdominal cerebrospinal fluid pseudocyst. Australas Radiol 48:61-63 24. Roitberg BZ, Tomita T, McLone DG (1998) Abdominal cerebrospinal fluid pseudocyst: A complication of ventriculoperitoneal shunt in children. Pediatr Neurosurg 29:267-273 25. Biko DM, Anupindi SA, Hernandez A et al (2009) Childhood Burkitt lymphoma: abdominal and pelvic imaging findings. AJR Am J Roentgenol 192:1304-1315 26. Wong S, Sanchez TR, Swischuk LE, Huang FS (2009) Diffuse peritoneal lymphomatosis: atypical presentation of Burkitt lymphoma. Pediatr Radiol 39:274-276 27. Moholkar S, Sebire NJ, Roebuck DJ (2006) Radiologicalpathological correlation in lipoblastoma and lipoblastomatosis. Pediatr Radiol 36:851-856 28. Prando A, Wallace S, Marins JL et al (1990) Sonographic features of benign intraperitoneal lipomatous tumors in childrenreport of 4 cases. Pediatr Radiol 20:571-574 29. Veyrac C (2008) Intraperitoneal masses. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. Springer-Verlag, Berlin-Heidelberg, pp 511-544

Pediatric Intestinal Ultrasonography

30. Levy AD, Rimola J, Mehrotra AK, Sobin LH (2006) From the archives of the AFIP: benign fibrous tumors and tumorlike lesions of the mesentery: radiologic-pathologic correlation. Radiographics 26:245-264 31. Leung RS, Calder A, Roebuck D (2009) Embryonal rhabdomyosarcoma of the omentum: two cases occurring in children. Pediatr Radiol 39:865-868 32. Roebuck DJ, Yang WT, Lam WW, Stanley P (1998) Hepatobiliary rhabdomyosarcoma in children: diagnostic radiology. Pediatr Radiol 28:101-108 33. Karmazyn B, Werner EA, Rejaie B, Applegate KE (2005) Mesenteric lymph nodes in children: what is normal? Pediatr Radiol 35:774-777

261

34. Watanabe M, Ishii E, Hirowatari Y et al (1997) Evaluation of abdominal lymphadenopathy in children by ultrasonography. Pediatr Radiol 27:860-864 35. Vayner N, Coret A, Polliack G et al (2003) Mesenteric lymphadenopathy in children examined by US for chronic and/or recurrent abdominal pain. Pediatr Radiol 33:864-867 36. Simanovsky N, Hiller N (2007) Importance of sonographic detection of enlarged abdominal lymph nodes in children. J Ultrasound Med 26:581-584 37. Couture A (2008) Bowel obstruction in neonates and children. In: Couture A, Baud C, Ferran J et al (eds) Gastrointestinal tract sonography in fetuses and children. 1st edn. SpringerVerlag, Berlin-Heidelberg, pp 131-251