| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gastrointestinal Imaging |
1 From the Departments of Radiologic Pathology (A.D.L., R.M.A.) and Hepatic and Gastrointestinal Pathology (L.H.S.), Armed Forces Institute of Pathology, 6825 16th St NW, Washington, DC 20306-6000; Department of Radiology and Nuclear Medicine (A.D.L., R.M.A.) and School of Medicine (L.D.T.), Uniformed Services University of the Health Sciences, Bethesda, Md; and Department of Radiology, University of Maryland School of Medicine, Baltimore (R.M.A.). Received November 2, 2004; revision requested January 4, 2005; revision received January 11; accepted February 2. Address correspondence to A.D.L. (e-mail: levya{at}afip.osd.mil).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: The institutional review board approved this study; informed consent was not required. The study was HIPAA compliant. The authors retrospectively reviewed the barium studies (n = 20), computed tomographic (CT) scans (n = 16), magnetic resonance (MR) images (n = 2), pathology reports (n = 33), gross pathology photographs (n = 15), and clinical data (n = 33) from 33 patients (16 men and 17 women; age range, 19–90 years; mean age, 52.6 years) with a confirmed diagnosis of duodenal carcinoid admitted into our institution during a 52-year period. The imaging studies were evaluated by consensus of two abdominal radiologists for the number of masses and their location and morphologic characteristics (polypoid or mural). The CT and MR images were also assessed for contrast enhancement characteristics.
RESULTS: Most carcinoids were located in the proximal duodenum (10 in the bulb, 19 in the second portion, two in the third portion, and two in the fourth portion). Seventeen patients (52%) had focal intraluminal polypoid masses and 13 (39%) had mural masses; in three patients (9%), the tumor was not visualized at CT. Five of the 33 patients (15%) had multiple carcinoids. CT showed heterogeneous contrast enhancement in all patients who received intravenous contrast material in the arterial or portal venous phases of enhancement. Nonenhancing masses were present in patients who underwent CT during the equilibrium phase. Two patients had Zollinger-Ellison syndrome. Five patients (15%) had neurofibromatosis type 1 (NF-1); four of the five patients (80%) were women, and four patients were African American. In all five patients with NF-1, the carcinoids were located in the periampullary region.
CONCLUSION: Duodenal carcinoids are uncommon tumors with a wide clinical-pathologic spectrum. They occur most commonly in the proximal duodenum and manifest as an intraluminal polyp or a mural mass.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Most duodenal carcinoids (62%) are gastrin cell (G cell) tumors, and one-third of these are functioning tumors that manifest as Zollinger-Ellison syndrome at clinical examination (3). G cell carcinoids are also found in 90% of patients with MEN-1 (2). Somatostatin-producing (D cell) tumors represent 21% of duodenal carcinoids and occur exclusively in and around the ampulla of Vater (3). Although D cells produce somatostatin, systemic manifestations of excess somatostatin production are extremely rare. There is a well-established relationship between D cell carcinoids and NF-1. The remaining types of duodenal carcinoids are uncommon and include tumors that contain serotonin, calcitonin, cholecystokinin, vasoactive intestinal polypeptide, bombesin, pancreatic polypeptide, or multiple hormones.
There is a case report and small series in the radiology literature from the early and mid-1980s in which duodenal carcinoids are described as multiple polypoid defects in the duodenal bulb, intramural masses, and large ulcerative lesions at barium examination (4,5). More recent case reports have described duodenal carcinoid tumors in patients with NF-1 as having marked enhancement on intravenous contrast material–enhanced computed tomographic (CT) scans (6,7). Therefore, we undertook our study to retrospectively evaluate the imaging features of duodenal carcinoids with clinical-pathologic comparison.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The study population consisted of 17 women (age range, 19–72 years; mean age, 49.2 years) and 16 men (age range, 33–90 years; mean age, 56.2 years). The entire study population ranged in age from 19 to 90 years (mean age, 52.6 years; median age, 54 years). Twenty-four patients were white, seven were African American, and one was Latin American; the ethnicity of one patient was unknown.
Clinical and Pathology Data Review
Clinical and pathology data were reviewed by one experienced abdominal radiologist (A.D.L.) and one 4th-year medical student (L.D.T.). Clinical data were reviewed for signs and symptoms at presentation; clinical evidence of excess hormone secretion (peptic ulcer disease, diarrhea, steatorrhea, diabetes mellitus); medical history of malignancy; documented evidence of NF-1, Zollinger-Ellison syndrome, or MEN-1; and clinical evidence of carcinoid syndrome.
All patients underwent surgical excision (n = 30) or biopsy (n = 3) of their tumor. Histopathologic and surgical records were reviewed for tumor location and size and evidence of lymph node or liver metastasis.
Imaging Review
Imaging studies were available in all patients. Two abdominal radiologists (A.D.L., R.M.A.), each with 11 years of experience, reviewed all images retrospectively with final interpretation by consensus. The final diagnosis was known at interpretation. Spot and overhead images from barium examination of the upper gastrointestinal tract were available in 20 patients whose cases were accessioned from January 1951 to June 1994. The barium studies were reviewed for the presence or absence of a duodenal mass or ulcer. If a mass or ulcer was present, its location and the number of masses were noted. Masses were assessed for overall morphologic characteristics (polypoid, intramural, or extrinsic), mass margins (smooth or irregular), and whether ulceration was associated with the mass. Duodenal folds were evaluated for thickness (normal folds [<3 mm], thick regular folds, or thick irregular folds), and the caliber of the duodenum was assessed for dilatation, luminal narrowing, or stricture formation.
CT scans were available in 16 patients. Four patients underwent unenhanced CT, eight underwent intravenous contrast-enhanced CT, and four underwent both unenhanced and intravenous contrast-enhanced CT. One patient underwent CT in July 1978; the remaining patients underwent CT from April 1986 to August 2003. All patients received oral contrast material for scanning ("unenhanced" CT indicates that no intravenous contrast material was used). Because our patients were referred from many institutions and their CT examinations had been performed by using a variety of equipment during a long time period, the CT examinations were not standardized. Section thickness, injection technique, and scanner technique were variable. Nine patients were scanned with incremental CT and section thicknesses of 8–10 mm. Seven patients were scanned with helical CT and section thicknesses of 3–8 mm. Of the 12 patients who received intravenous contrast material, five patients were scanned during the equilibrium phase of contrast enhancement, one patient was scanned during the arterial phase of contrast enhancement, five patients were scanned during the portal venous phase of contrast enhancement, and one patient was scanned during both portal venous and equilibrium phases of contrast enhancement.
CT scans were reviewed for the presence or absence of a duodenal mass. If a mass was present, the location, morphologic characteristics (polypoid or intramural), and contrast enhancement pattern (no enhancement, homogeneous enhancement, or heterogeneous enhancement) were noted. The scans were also evaluated for evidence of lymphadenopathy (nodes > 8 mm in the short axis in the peripancreatic, paraduodenal, retroperitoneal, or mesenteric regions), liver metastases, obstructive biliary dilatation, adjacent organ invasion, and intussusception.
Magnetic resonance (MR) images were available for review in two patients. One patient was imaged with T1-weighted gradient-recalled echo, gadolinium-enhanced T1-weighted inversion recovery, and heavily T2-weighted MR cholangiopancreatography sequences. The second patient was imaged with a T1-weighted sequence and gadolinium-enhanced dynamic fast multiplanar spoiled gradient-recalled acquisition in the steady state. MR images were evaluated for the presence of a duodenal mass, mass location, morphologic characteristics (polypoid or intramural), signal intensity pattern, adenopathy, liver metastases, obstructive biliary dilatation, and adjacent organ invasion.
Comparisons
Immediately after imaging review, gross pathologic photographs (n = 15), endoscopic photographs (n = 3), surgical records (n = 33), and pathology records (n = 33) were compared with all imaging studies by two abdominal radiologists (A.D.L., R.M.A.) and one 4th-year medical student (L.D.T.). Gross photographs were evaluated for the appearance of tumor margins (well defined or poorly defined), the presence of a tumor capsule, and the presence of intratumoral hemorrhage or necrosis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Five patients (four women and one man) had NF-1. These patients ranged in age from 26 to 63 years (mean, 41 years). Four of these patients were African American and one was white. Two patients presented with complaints of abdominal pain, two had jaundice, and one was found to have an elevated serum liver transaminase level during an evaluation of right flank pain. Four of the five patients had concurrent neoplasms or a history of neoplasia. Three patients had subcutaneous neurofibromas, and one patient had previously undergone craniotomy for a falx meningioma.
Two patients had MEN-1 syndrome (one presented with diarrhea from a gastrin-producing carcinoid and the other presented with striae and hyperpigmentation from an elevated adrenocorticotrophic hormone level). Two patients had Zollinger-Ellison syndrome. None of the patients had clinical evidence of carcinoid syndrome or laboratory evidence of elevated levels of serotonin or its metabolites.
Results of pathology and surgical record review revealed that the tumor was located in the second portion of the duodenum in 19 patients (58%), in the duodenal bulb in 10 (30%), in the third portion of the duodenum in two (6%), and in the fourth portion of the duodenum in two (6%). Of the 19 patients whose tumors were located in the second portion of the duodenum, 11 (58%) had tumors in the periampullary region; five of these patients had NF-1. Seven of the periampullary carcinoids produced biliary obstruction. Five patients (15%) had multiple carcinoids: Two patients had multiple carcinoids in the second portion of the duodenum, two patients had multiple carcinoids in the duodenal bulb, and one patient had one carcinoid in the duodenal bulb and a second at the ileocecal junction. The maximal diameter of the tumors ranged from 0.7 to 7.5 cm (mean, 2.5 cm).
Twelve of the 33 patients (36%) had pathologic evidence of metastatic disease at the initial diagnosis: Five patients had positive paraduodenal or peripancreatic lymph nodes, three patients had positive peripancreatic lymph nodes and pancreatic gland invasion, two patients had liver metastasis and positive peripancreatic lymph nodes, one patient had tumor within a small bowel mesenteric lymph node, and one patient had positive tumor in a lymph node of the transverse mesocolon.
Barium Study Features
All 20 patients who underwent barium examinations had a focal abnormality. Fifteen of the 20 patients (75%) had focal intraluminal polypoid masses (Fig 1). The polyps were smoothly marginated in all cases. One polyp had an irregular ulcer crater on its surface. Two patients with polyps had multiple lesions (one patient had two lesions and the other had three) in the duodenal bulb. In nine patients (45%), the primary polypoid masses were located in the duodenal bulb (Fig 1). The masses in the remaining patients were distributed in the second (n = 3), third (n = 2), and fourth (n = 1) portions of the duodenum.
|
|
|
|
CT Features
A focal duodenal mass was visualized in 13 of the 16 patients who underwent CT. Of the 13 patients with visualized masses, seven (54%) had carcinoids that manifested as focal mural masses (Fig 3) and six (46%) had carcinoids that manifested as intraluminal polypoid masses (Fig 4). The distribution of carcinoids for those patients who underwent CT was as follows: duodenal bulb (n = 1), second portion of the duodenum (n = 10), third portion of the duodenum (n = 1), and fourth portion of the duodenum (n = 1). All but one patient in whom masses were visualized at CT had solitary lesions. The one patient with multiple carcinoids at CT had multiple polypoid nodules in the second portion of the duodenum (Fig 4a).
|
|
|
|
|
|
Other CT findings included enhancing liver metastases in one patient, extension of the carcinoid into the pancreas such that the lesion mimicked a primary pancreatic neoplasm in one patient, intussusception of a polypoid carcinoid of the third portion of the duodenum in one patient, and bilateral adrenal enlargement in one patient with MEN-1. Obstructive biliary dilatation was present in eight patients. Although 10 patients had pathologic evidence of lymph node metastasis in paraduodenal and/or peripancreatic lymph nodes, only one patient had evidence of pathologic adenopathy at CT.
MR Imaging Features
Both patients who underwent MR imaging had low-signal-intensity masses on T1-weighted images. In one patient, the carcinoid was located in the second portion of the duodenum proximal to the ampulla and showed heterogeneously high signal intensity on T2-weighted images. The mass had well-defined margins that extended into the head of the pancreas and compressed the distal common bile duct. The carcinoid in the second patient was also well defined. It was located on the lateral aspect of the second portion of the duodenum and showed homogeneous enhancement on images obtained with gadolinium-enhanced fast multiplanar spoiled gradient-recalled acquisition in the steady state.
Imaging Features of Carcinoids in NF-1
The maximal diameter of the carcinoids in the five patients with NF-1 ranged from 1.5 to 2.5 cm, and all were located in the periampullary region. Four of the five patients with NF-1 underwent CT and had evidence of biliary obstruction. Morphologically, the carcinoids were intraluminal polyps in two patients and mural masses in two patients. The carcinoids were not visualized with CT in one patient.
Summary of Imaging Features with Pathologic Correlation
The morphologic features of carcinoids identified at barium examination, CT, and MR imaging were as follows: 17 patients (52%) had focal intraluminal polypoid masses and 13 (39%) had mural masses. The tumor was not visualized at CT in three patients (9%). Gross pathologic and endoscopic photographs showed correlation with the findings at barium examination and CT of well-defined masses in all cases. There was no evidence of tumor capsule in cases with gross pathologic correlation (n = 15). The color of the tumors on cut sections ranged from white to yellow. Gross photographs showed hemorrhagic ulcers (Fig 2b) on the mucosal margin of the tumor in eight of the 15 patients (53%) and showed foci of hemorrhage on the cut surface of the tumor in two (13%). No evidence of ulceration was present at barium examination or CT in six of these eight patients. Endoscopic photographs (n = 3) showed well-defined sessile polyps that showed correlation with findings at barium examination and CT of polypoid lesions (Fig 4b).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The clinical presentation of our patients was variable. The most common presenting symptoms (abdominal pain and gastrointestinal bleeding) are frequent complaints of patients with upper gastrointestinal disease. There were no clinical symptoms specific for carcinoid tumor. The results of our study show that duodenal carcinoids occur most commonly in the proximal duodenum; in 89% of our patients, tumors were localized to the first and second portions of the duodenum. The morphologic characteristics of duodenal carcinoids were closely divided between intraluminal polyps (52%) and mural masses (39%).
The striking exception to our findings occurred in patients with NF-1. The occurrence of duodenal carcinoids in patients with NF-1 is well described in the clinical and pathology literature (8,9). Similar to results in published reports, most of our patients with carcinoids and NF-1 were women and African American, and all tumors in these patients were located in the periampullary region. Morphologically, the tumors were evenly divided between intraluminal polyps and mural masses.
Our results show that duodenal carcinoids enhance during the arterial and portal venous phases of intravenous contrast-enhanced CT and lose enhancement during the equilibrium phase. To our knowledge, the CT enhancement features of duodenal carcinoids have only been described in two case reports of periampullary carcinoids in patients with NF-1 (6,7). In each of those cases, the carcinoid was described as a hyperenhancing periampullary mass during the arterial phase of contrast enhancement. Early-phase contrast enhancement is an important feature to consider when imaging patients suspected of having a duodenal carcinoid or ampullary and/or periampullary abnormality because these lesions may be difficult to detect at CT when they are small. Modern multi–detector row CT technology affords the opportunity to improve lesion detection because of the ability to scan in multiple phases of contrast enhancement and view images in multiplanar two-dimensional reconstructions.
The early contrast enhancement pattern at CT may suggest the diagnosis of a carcinoid rather than an adenocarcinoma of the periampullary duodenum, ampulla, or pancreas when one is evaluating a patient with biliary obstruction and an ampullary region mass. The histologic distinction between a carcinoid and an adenocarcinoma may be difficult with hematoxylin-eosin–stained slide material from endoscopic biopsy specimens because of crush artifact and the nature of some adenocarcinomas to have an insular growth pattern that mimics that of a carcinoid at histopathologic examination. The accurate preoperative distinction between these tumors is important for making management decisions and determining the prognosis. Therefore, the knowledge of the imaging pattern characteristic of carcinoids may be helpful to the pathologist interpreting a biopsy specimen.
Duodenal carcinoids are distinctly different tumors compared with their counterparts in the jejunum and ileum. None of our patients had clinical or laboratory evidence of serotonin production by the tumor. Unlike duodenal carcinoids, jejunal and ileal carcinoids arise from serotonin-producing enterochromaffin cells and commonly manifest imaging evidence of local serotonin production such bowel kinking, mesenteric desmoplasia, elastic vascular sclerosis, and intestinal ischemia (10,11). These features were not present in any of our patients with duodenal carcinoids.
Thirty-six percent of our patients had evidence of metastatic disease at initial presentation. Local lymph node metastases were the most common site for spread. Although lymph node metastases were not identified on our CT studies when present, this may be explained by our limitations in CT technology and intravenous contrast material injection protocols. New multi–detector row CT scanners, intravenous contrast material protocols, somatostatin-receptor scintigraphy, and positron emission tomography have improved the preoperative detection of local lymph node and distant metastases in patients with carcinoid tumors (12–14).
The differential diagnosis of duodenal carcinoids is dependent on the imaging pattern. For those lesions that manifest as solitary polyps, the differential diagnosis includes Brunner gland hamartoma, gastric and pancreatic heterotopias, adenoma, adenocarcinoma, and gastrointestinal stromal tumor, lymphoma, and a prolapsed gastric neoplasm. For multiple duodenal polyps, the differential diagnosis includes Brunner gland hyperplasia, lymphoid hyperplasia, heterotopia, carcinoid, adenoma, polyposis syndrome, and metastatic disease. For a mural mass in the duodenum, the differential diagnosis includes carcinoid, gastrointestinal stromal tumor, lymphoma, metastatic disease, and rare mesenchymal neoplasms such as neurofibromas and schwannomas.
There is an inherent limitation to our study because of the nature of patients in a referral population. Our images and data were obtained at various institutions during a 52-year period with different equipment and protocols. The lack of standardization of technique and contrast material injection protocols cannot be overcome in a study population of this type.
The introduction of flexible endoscopy and CT in the evaluation of upper gastrointestinal tract disease is clearly reflected in our data and affects the spectrum of disease on which we are reporting. The use of barium examinations dominated the initial 30 years of our cases. In those patients evaluated with barium examination, 75% of the carcinoids were polyps. In the latter 20 years with CT, and to a lesser extent MR imaging, as the imaging study of choice, 46% of the duodenal carcinoids were polyps and 54% were mural masses. We hypothesize that the shift in morphologic characteristics over time reflects the changes in the evaluation of upper gastrointestinal tract disease after the introduction of flexible endoscopy. Patients with complex masses (periampullary lesions that cause jaundice and large lesions that are not amenable to endoscopic removal) undergo CT or MR imaging in addition to endoscopy. Patients with small intraluminal polyps are no longer referred to the radiology department for evaluation.
In conclusion, duodenal carcinoids are neuroendocrine tumors that encompass a wide clinical-pathologic spectrum of hormonally functioning and nonfunctioning tumors that differ from jejunal and ileal carcinoids. They occur most commonly in the proximal duodenum and have variable morphologic characteristics on imaging studies. The CT feature of an enhancing duodenal mass during the early phases of contrast material enhancement is suggestive of the possibility of a carcinoid.
|
FOOTNOTES |
|---|
Abbreviations: MEN-1 = multiple endocrine neoplasia type 1 • NF-1 = neurofibromatosis type 1
2 Current address: Department of Radiology, Madigan Army Medical Center, Tacoma, Wash. 
Authors stated no financial relationship to disclose.
The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Departments of the Army or Defense.
Author contributions: Guarantors of integrity of entire study, A.D.L., L.H.S.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, A.D.L., L.D.T.; clinical studies, A.D.L., R.M.A.; and manuscript editing, A.D.L., R.M.A., L.H.S.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. F. Scarsbrook, A. Ganeshan, J. Statham, R. V. Thakker, A. Weaver, D. Talbot, P. Boardman, K. M. Bradley, F. V. Gleeson, and R. R. Phillips Anatomic and Functional Imaging of Metastatic Carcinoid Tumors RadioGraphics, March 1, 2007; 27(2): 455 - 477. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. D. Levy and L. H. Sobin From the Archives of the AFIP: Gastrointestinal Carcinoids: Imaging Features with Clinicopathologic Comparison RadioGraphics, January 1, 2007; 27(1): 237 - 257. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
