Use of methylcellulose in small bowel follow-through examination: comparison with conventional series in normal subjects

Background: To evaluate and optimize a modified small bowel follow-through examination (SBFT) by using an oral administration of a large amount (600 mL) of methylcellulose after taking a small amount of barium. Methods: Ninety-nine normal subjects underwent modified SBFT with an oral administration of 600 mL of 0.5% methylcellulose after taking 100 mL of 120% or 100, 150, or 200 mL of 70% w/v barium. The transradiancy and distensibility of the bowel and the transit time were compared with those of 39 other normal subjects who underwent conventional SBFT with 500 mL of 70% w/v barium. Results: Except for two subjects who developed diarrhea immediately after examination, no patients complained of acute symptoms, such as abdominal pain or vomiting. Modified SBFT was much superior to conventional series for obtaining good bowel transradiancy and rapid transit time (range = 37–49 min), but bowel distention was not significantly improved. The use of 150 mL of 70% w/v barium was better than the other three modified techniques for achieving good bowel transradiancy, rapid transit time, and less occurrence of flocculation. Conclusion: Modified SBFT is a simple method for easily improving the image quality in terms of bowel transradiancy and transit time. Received: 11/6/96/Accepted after revision: 2/19/97     

Methodological assessment of combined spiral CT angiography and CT arterial portography

Background: To assess an optimal methodology of combined spiral computed tomographic (CT) angiography (CTA) and CT arterial portography (CTAP) for detection and characterization of liver tumors. Methods: We performed spiral CTAP only in five patients with 30–32% contrast (subset A), CTAP combined with preceding spiral CTA using 30–32% contrast in 19 (subset B), and CTAP combined with preceding spiral CTA with 60–64% contrast in seven (subset C). The CT numbers of the aorta immediately before preceding CTA and subsequent CTAP and the CT numbers of malignant tumor and liver parenchyma with CTAP were measured. Results: The differences of the CT number between the malignant tumor and liver parenchyma on CTAP were 61.1–161.8 (mean ± SD, 114.5 ± 39.3) HU, 50.7–164.8 (104.2 ± 31.2) HU, and 101.2–368.3 (219.5 ± 90.5) HU in subsets A, B, and C, respectively. Two cavernous hemangiomas showed pathognomonic findings with preceding CTA. Conclusion: Combination of preceding spiral CTA and subsequent spiral CTAP using 30% contrast with a 5-min interval is an optimal method for detection and characterization of liver tumors. Received: 14 December 1995/Accepted after revision: 13 February 1996     

Helical CT of the small bowel with an alternative oral contrast material in patients with Crohn disease

Background: We assessed the usefulness of helical computed tomography (CT) with a negative oral contrast material for detecting Crohn disease. Methods: Thirty-eight patients with proven Crohn disease were examined. We administered a large volume of a new negative oral contrast material (Mucofalk suspended in water) and then proceeded with helical CT scanning. This technique is an alternative to CT and conventional enteroclyses that use a nasojejunal tube. Two radiologists interpreted the scans, and patients were interviewed about their tolerance of the procedure. We created multiplanar reformatted images in all cases. Potential of small bowel distention by Mucofalk was evaluated by two observers on a three-point scale, and interobserver agreement was calculated with κ statistics. Results: All patients who underwent enteroclysis stated that CT was the more comfortable method, the taste of the peroral contrast medium was considered good by 52.6% and acceptable by 47.4%. Small bowel distention was excellent in 55% of cases, moderate in 26%, and poor in 19%, with an interoberserver agreement of 78%. CT findings correlated with enteroclysis in 27 patients who underwent both methods. Analysis of CT versus enteroclysis showed a sensitivity of 89% for CT versus 78% for small bowel enteroclysis. Conclusion: Mucofalk CT is a simple, rapid, noninvasive, and accurate method of evaluating extramucosal manifestations of Crohn disease. The tubeless procedure improved patients' comfort and decreased time, cost, and radiation exposure.     

Amyloidosis of the alimentary canal: radiologic–pathologic correlation of CT findings

Background: The purpose of the study was to describe the computed tomographic (CT) findings of the alimentary canal and mesentery in amyloid infiltration of the gastrointestinal (GI) tract and to correlate the CT findings with histologic extent and distribution and with amyloid subtype. Methods: Abdominal CT scans performed between 1988 and 1997 on patients with pathologically proven amyloidosis of the alimentary canal were reviewed for abnormalities of the alimentary canal and mesentery. Histology was graded for extent of mucosal, submucosal, and muscularis propria involvement and for degree of interstitial and vascular distribution. CT findings were correlated with histologic extent, histologic distribution, and amyloid histochemical type. Results: Twenty-three patients were included. Four (17%) had bowel wall thickening, which was associated with a higher submucosal extent and interstitial distribution than in patients with normal bowel by CT. Four (17%) patients had bowel wall dilatation without thickening, which was not associated with statistically significantly different histology than in patients with normal bowel by CT. There was no statistically significant correlation between CT findings and histochemical subtype. Mesenteric soft tissue infiltration was seen in two patients, and mesenteric adenopathy was seen in one patient. Conclusions: Normal bowel is a common abdominal CT finding in amyloidosis of the alimentary canal. When findings are present, GI wall thickening and/or bowel wall dilatation without wall thickening may be seen. Bowel wall thickening on CT correlates with submucosal extent and interstitial distribution of disease. Soft tissue infiltration and adenopathy are also occasionally seen. Received: 15 January 1999/Accepted: 10 March 1999     

Cisapride or metoclopramide to accelerate small bowel transit during barium follow-through examination?

Background: Metoclopramide is commonly used to accelerate small bowel transit during barium follow-through (BaFT) examinations, but its action is unpredictable. Cisapride, commonly used to treat gastroesophageal reflux disease, also accelerates small bowel transit and may be a viable alternative. The two were compared in a prospective, randomized, blind study. Methods: Patients attending for BaFT were randomized to receive either 10 mg cisapride or 20 mg metoclopramide orally 1 h before the barium suspension. BaFT was performed by using a standard technique, and small bowel transit and study quality were compared. Patients also noted any side effects experienced. Results: Of 45 patients recruited, 27 received cisapride and 18 metoclopramide. Median transit time for the cisapride group was 30 min (range = 10–130 min) versus 67.5 min (range = 30–290 min) for the metoclopramide group (p= 0.019). Study quality was comparable. However, nine patients (33%) receiving cisapride experienced nausea versus only one subject (6%) receiving metoclopramide (p= 0.034). Conclusions: This study suggests that cisapride is a more effective prokinetic agent than metoclopramide, but this benefit is offset by a higher incidence of side effects. Received: 17 August 1999/Revision accepted: 3 November 1999     

Pirenzepine versus scopolamine methyl bromide in double-contrast barium enema study of large bowel

To evaluate the usefulness of pirenzepine for diagnostic double-contrast barium enema study of the large bowel, pirenzepine and scopolamine methyl bromide (SMB) were compared in a single, blind, randomized trial. Sixty consecutive patients were enrolled in the study. Quantitative analysis of bowel distention was done by measuring the maximum diameter of the transverse colon before and after drug administration. Four independent observers blindly evaluated distention and mucosal coating of the large bowel and global quality of the images. No differences were found in the diagnostic performance between the two drugs. However, pirenzepine induced a slight but significantly larger distention of the large bowel (68 ± 12 vs. 65 ± 8 mm, p= 0.02). Heart rate and rhythm during the study were recorded by ECG. SMB induced tachycardia in all patients (from 72 ± 15 to 98 ± 24 beats/min, p < 0.01), whereas pirenzepine did not (from 76 ± 13 to 78 ± 20, p= NS). After SMB, one-patient exhibited faintness, and some patients complained of visual accommodation defects, dryness of the mouth, and dizziness. Pirenzepine had a diagnostic performance similar to SMB in avoiding adverse effects elicited by SMB. Received: 14 December 1994/Accepted: 2 February 1995     

Iliopsoas abscess: a report of 24 patients diagnosed by CT

Background: We wanted to define the role of computed tomography (CT) in the diagnosis, etiology, and treatment of iliopsoas abscess. Methods: Twenty-four patients (18 men, six women; age range = 17–86 years) with iliopsoas abscesses diagnosed over 8 years were retrospectively reviewed. All presented with fever and elevated white blood cell counts. Twenty-one had abdominal, flank or pelvic pain and nine had specific psoas signs suggesting the diagnosis. Results: Seventeen of the abscesses were right-sided. Twenty were regarded as secondary to various underlying causes that were clearly demonstrated on CT and related to gastrointestinal (n= 12), skeletal (n= 5), or urinary tract (n= 3) diseases. All patients received appropriate antibiotic treatment. Thirteen also had their abscesses drained and eight had definitive surgical procedures. Conclusion: CT is an effective imaging technique for diagnosing iliopsoas abscess, even when classic clinical signs are absent. Treatment by percutaneous drainage under CT guidance is another advantage. When a psoas abscess is a complication of Crohn's disease, resection of the affected bowel segment is recommended in addition to drainage because drainage alone even in conjunction with appropriate medical therapy is usually not effective. Received: 30 August 2000/Accepted: 1 November 2000     

Intestinal tuberculosis with abdominal complications: radiologic and pathologic features

Background: To investigate radiologic and pathological features of intestinal tuberculosis with abdominal complications. Methods: Twenty-two patients with 23 surgically proven complications (nine intestinal obstructions, eight perforations, three fistulae, and three intestinal bleeds) were analyzed. Medical records, radiologic studies, and pathologic examinations were reviewed with special emphasis on searching for the common features in each group of complication. Results: The most important single feature in seven of the nine patients with intestinal obstruction was the presence of stricture. In the remaining two patients, bowel adhesion was a primary cause of obstruction. In eight patients with intestinal perforation, both obstruction and ulcerations in the dilated proximal loop were the important features in six, and multiple deep ulcerations without obstruction was a primary cause in the remaining two. The common features in three patients with fistulae were focal or multiple strictures, severe adhesions, and fibrotic bowel wall. Intestinal bleeding originated from diffuse mucosal ulcerations. The abdominal complications occurred during antituberculous therapy in 10 of the 22 patients. Conclusion: Understanding the radiologic and pathologic features of intestinal tuberculosis with complications help in making an appropriate clinical decision for the treatment strategy. Close observation is necessary, especially in those patients who are acutely ill during antituberculous medical therapy. Received: 30 June 1997/Accepted 20 August 1997     

Intestinal malrotation as an incidental finding on CT in adults

Background:Intestinal malrotation in adults is usually an incidental finding on computed tomography (CT). We present the CT findings of 18 adult patients with malrotation and discuss the clinical implications. Methods: Abdominal scans of 18 patients (12 women, six men; age range = 15–79 years) with intestinal malrotation were reviewed. Special attention was directed to the location of the superior mesenteric vessels, the location of the small and large bowels, the size of the uncinate process, the situs definition, and additional anomalies. Results: The malrotation was an incidental finding in all but one patient. The malrotation was type Ia in 17 patients and IIc in the one symptomatic patient. The superior mesenteric vessels were vertically oriented in 10, inverted in two, normally positioned in four, and mirror imaged in two cases with situs ambiguus. All patients had aplasia of the pancreatic uncinate process, five had a short pancreas, and two had a preduodenal portal vein. Fourteen patients had a normal situs and four had heterotaxia. Seven patients had polysplenia, six of which with associated inferior vena cava anomalies. Conclusions: Intestinal malrotation can be diagnosed on CT by the anatomic location of a right-sided small bowel, left-sided colon, an abnormal relationship of the superior mesenteric vessels, and aplasia of the uncinate process. Awareness of these abnormalities is necessary to diagnose this anomaly. It should be sought in patients with a situs problem, inferior vena cava anomalies, polysplenia, or preduodenal portal vein. Although usually an incidental finding, it is important to diagnose such a malrotation because it may cause abdominal symptoms. Also, knowledge of associated vascular anomalies is important when abdominal surgery is planned. Received: 1 October 1998/Revision accepted: 27 January 1999     

CT of the mesenteric vascular anatomy

Background: To evaluate the appearance of the arrangement of the superior mesenteric artery (SMA) and superior mesenteric vein (SMV) on computed tomography (CT) in normal patients and in patients with abdominal masses. Methods: One hundred seventy-seven consecutive abdominal CT examinations of 143 adults and two children were reviewed. The relationship of the SMV to the SMA was recorded at four locations: the beginning of the mesenteric vessels and levels 3 cm, 6 cm, and 9 cm caudad to the beginning. The relationship of the SMV to the SMA was divided into four quadrants in relation to the SMA: I, ventral right or directly ventral; II, dorsal right or directly right; III, dorsal left or directly dorsal; and IV, ventral left or directly left. Results: In the beginning of the SMV–SMA complex and levels 3 cm, 6 cm, and 9 cm caudal to the beginning, the SMV was located in quadrant I in 146, 84, 69, and 43 examinations, in quadrant II in 31, 93, 71, and 27 examinations, in quadrant III in zero, zero, five, and three examinations, and in quadrant IV in zero, zero, nine, and 15 examinations, respectively. The cases with SMV inversion had neither malrotation nor adjacent tumor compression. All the cases with an adjacent tumor-induced compression of the SMV–SMA complex had a normal SMV–SMA relationship. Conclusion: In the first 3 cm, the SMV is always to the right of the SMA. Caudal to the level of 6 cm, the SMV may be located to the left of the SMA without evidence of malrotation. A midgut nonrotation is more likely to be present when a proximal SMV inversion is coexistent with a rightward direction of the proximal jejunal vessels. A hypothetical depiction of the step-by-step change of the SMV–SMA relationship during embryologic development may explain the arrangement patterns of the mesenteric vessels in normal rotation and midgut nonrotation. Received: 6 May 1996/Accepted: 22 May 1996     

正常小肠的MSCT表现

目的 总结口服及灌肠双重肠道准备后小肠MSCT各种正常表现.方法 对44名健康志愿者,经结肠、小肠双充盈法准备后,对小肠进行测量,包括肠壁厚度、肠管充盈直径、肠壁分层、肠壁CT值、空肠黏膜皱襞数、肠系膜脂肪密度、肠系膜血管影以及肠系膜淋巴结和后腹膜淋巴结MSCT显示率等.结果 小肠总的肠壁厚度平均值为(1.72±0.20)mm;整个小肠平均充盈肠管直径(21.24±3.05)mm;小肠各组段肠壁平扫、增强动脉期和门静脉期平均CT值分别为(28.13±0.38)HU、(55.42±0.88)HU和(67.86±0.94)HU;正常小肠每厘米肠段内可见的空肠黏膜皱襞数为2.29个,肠系膜血管影为2.59个;平均肠系膜脂肪密度-98.57 HU;肠壁改变以单层为主.结论 对正常小肠MSCT的各种表现和测量指标的正确认识有助于小肠病变的诊断.     

Acute diverticulitis of the small bowel: CT findings

We present the computed tomographic (CT) findings in two cases of small bowel diverticulitis, one affecting the jejunum and the other a Meckel's diverticulum. The main CT finding was that of a mass with an air–fluid collection in contiguity with small bowel loops. Received: 18 September 1998/Accepted: 2 December 1998     

Computed tomographic appearance of sigmoid volvulus

The computed tomographic (CT) appearance of two cases of sigmoid colon volvulus is described. Both underwent plain abdominal radiographs, contrast enema, and CT. The findings of sigmoid volvulus at CT were characteristic, having a whirl pattern of the dilated sigmoid loop around mesocolon and vessels and a bird-beak aspect of the afferent and efferent segments. CT may be valuable in a case of unusual clinical or plain film presentation as an alternative to contrast enema. Received: 24 March 1995/Accepted: 3 May 1995     

CT differentiation between necrotic and nonnecrotic small bowel in closed loop and strangulating obstruction

Background: The purpose of this study was to evaluate computed tomographic (CT) findings for predicting the presence of intestinal necrosis in patients with closed loop and strangulating obstruction of the small bowel. Methods: Twenty-five patients with surgically confirmed closed loop and strangulating obstruction were divided into two groups with (n= 16) and without (n= 9) intestinal necrosis. By using univariate and multivariate statistical procedures, we evaluated the differences in CT findings between the two groups on the basis of the following six findings: bowel dilatation of strangulated loops (bowel dilatation), wall thickening of strangulated intestines (wall thickening), ascites, vascular dilatation of affected mesenteries (vascular dilatation), elevation of mesenteric attenuation (mesenteric attenuation), and radial distribution of the mesenteric vessels (radial distribution). Results: Of the six findings, ascites, vascular dilatation, mesenteric attenuation, and radial distribution provided significant discriminating findings between the two groups on univariate analysis. On multivariate analysis, mesenteric attenuation was the most important discriminative factor, followed by radial distribution and ascites. Using these three parameters, the CT was correlated with the surgical findings in 15 of the 16 patients in the necrosis group (sensitivity = 93.8%) and in eight of the nine patients in the nonnecrosis group (specificity = 88.9%). The overall accuracy was 92.0%. Conclusions: Mesenteric attenuation, radial distribution, and ascites, depicted on CT differentiate well between necrosis and nonnecrosis of the small bowelin patients with closed loop and strangulating obstruction. Received 5 December 1997/Accepted: 14 January 1998     

Cavernous transformation of the portal vein secondary to tumor thrombosis of hepatocellular carcinoma: spiral CT visualization of the collateral vessels

Background: We investigated the constituting collateral vessels in cavernous transformation of the portal vein (CTPV) caused by tumor thrombosis of hepatocellular carcinoma (HCC) by using contrast-enhanced spiral computed tomographic (CT) examination. Methods: Fifty-four histopathologically proven HCC patients with tumor thrombosis-induced CTPV were retrospectively included and assigned to cirrhosis negative (n= 31) and positive (n= 23) groups. Another 15 cirrhotic patients with portal hypertension but no HCC and CTPV were used for comparison. Standardized dual-phase contrast-enhanced spiral CT was performed for all patients. CT appearances of the collateral vessels of CTPV were observed, and their visualization rates were analyzed. Results: Biliary (cystic and paracholedochal veins) and gastric (left and right gastric veins) branches of the portal vein were the most frequently visualized collateral vessels of CTPV. There was a marked difference in CT visualization rates for biliary branches between patients with and without CTPV (83–94% vs. 0). No difference existed in visualization rates for gastric branches across the three groups (77–87% for left gastric, 58–61% for right gastric vein). Conclusions: Biliary and gastric branches of the portal vein are the major collateral vessels of CTPV. The intergroup differences in CT visualization rates may provide clues to the roles that they might play in the hemodynamic adaptation process of CTPV. Received: 13 October 1999/Accepted: 12 January 2000     

Perforations of the rectosigmoid colon induced by cleansing enema: CT findings in 14 patients

Background: We report the computed tomographic (CT) features of colorectal perforations caused by cleansing enema. Methods: We reviewed the medical records and CT studies of all patients with colorectal perforations caused by a cleansing enema. Results: Fourteen patients (10 men, four women; average age = 80 years) were included in the study. The most common presenting symptoms were severe abdominal pain and fever. CT was performed within 48 h after the event in most patients. Extraluminal air in the perirectal fat was the most frequent finding on CT and was present in all patients. Additional findings were extraperitoneal (n = 9), intraperitoneal (n = 3) and /or subcutaneous (n = 3) air, free fluid (n = 9), extraluminal feces (n = 8), and focal bowel wall thickening (n = 4). No leak of contrast from the rectum was observed in any patient including the one patient in whom contrast was administered rectally. Ten patients were treated by surgery; five recovered and the other five died. The other four patients were treated conservatively and all four died. Conclusion: The diagnosis of colorectal perforation can be made on CT. Because the diagnosis is not always suspected by the clinician, the radiologist may be the first to suggest it. Therefore, the radiologist should be familiar with the CT features of a potentially lethal, rectally induced perforation. Received: 19 April 2001/Revision accepted: 4 July 2001     

Adult Crohn disease: can ileoscopy replace small bowel radiology?

Background: This study aimed to document the radiological features and distribution of small bowel Crohn disease (CD) in adults by using a barium follow-through (BaFT) technique and to determine whether disease would be missed or its distribution underestimated if only colonoscopy with ileoscopy were performed. Methods: The BaFT examinations of 121 adults with proven CD were reviewed retrospectively with respect to the stage and distribution of disease. Colonoscopy with attempted ileoscopy was performed in 37 of these subjects, and the results were compared with radiological findings. Results: A normal villous pattern was visualized in 89 studies (74%). BaFT showed small bowel CD in 71 (59%) of 121 patients studied. The terminal ileum (TI) was the most common site of disease, affecting 62 (87%) of patients with small bowel CD. Forty-six patients (65%) had more proximal small bowel disease, including nine (13%) with a normal TI. BaFT showed early mucosal changes of CD in 52 subjects (73%), which was the sole manifestation in 15 (21%). Ileoscopy was possible in the majority of patients colonoscoped but was not achieved in 14 (38%), nine of whom had CD on BaFT. Of the 23 patients in whom ileoscopy was performed, findings agreed with BaFT assessment of the TI in 22. Conclusion: BaFT adequately demonstrates the stage and extent of small bowel CD. The majority of patients with small bowel CD have disease proximal to the TI, which cannot be diagnosed by ileoscopy. Received: 27 August 1996/Accepted: 16 October 1996     

Nonpathological focal enhancements on spiral CT hepatic angiography

Background: To assess the frequency and characteristics of nonpathological focal enhancements seen on spiral computed tomographic (CT) hepatic angiography (CTA). Methods: Spiral CTA and spiral CT arterial portography (CTAP) were performed in 31 patients with suspected liver malignancy prior to potential liver resection. The CTA images were retrospectively reviewed for focal enhancements by two radiologists. After determining nonpathological focal enhancements on CTA images based on the other radiographic tests, surgical exploration including intraoperative sonography, follow-up imagings, the frequency, size, site, and shape of nonpathological focal enhancements with CTA were assessed. Results: Thirty-six nonpathological focal enhancements with CTA from 4 to 23 (mean = 11.4) mm were seen in 14 (45.2%) of 31 patients. Thirteen (36.1%) of 36 nonpathological focal enhancements with CTA were not depicted with CTAP. Nonpathological focal enhancements with CTA were frequent in Couinaud segments III (27.8%), V (22.2%), and VI (19.4%). Twenty-three (63.9%) of 36 nonpathological focal enhancements were located in the edge of the liver. Shapes of 36 nonpathological focal enhancements with CTA included circular (n = 16), worm (n = 7), irregular (n = 6), dot (n = 6), and wedge (n = 1). Conclusion: In nearly half of patients, spiral CTA shows various shapes of small nonpathological focal enhancements more frequently in the liver edge. Received: 6 June 1996/Accepted: 14 July 1996     

Xanthogranulomatous cholecystitis: a radiological study of 12 cases and a review of the literature

Background: Twelve cases of xanthogranulomatous cholecystitis (XGC) are presented, and their radiologic appearance is described. Methods: Four men and eight women, aged 31–82 years old, with XGC were reviewed. Abdominal ultrasound (US) was performed in all patients. Computed tomography (CT) was performed in five patients, barium enema examination in two, and percutaneously CT-guided fine-needle aspirative biopsy of the gallbladder in one. Results: Barium enema examination showed an indentation of the hepatic flexure. Cholelithiasis was present in all patients, and sludge was present in six. The gallbladder wall was thickened in all patients, irregular in nine, and could not be properly differentiated from surrounding liver parenchyma or from other adjacent structures in most patients. A curvilinear halo, hypoechoic on US and with low attenuation on CT, within the gallbladder wall was found in three patients and pericholecystic fluid in two others. On CT, the pericholecystic fat had streaky soft tissue densities in three cases. Percutaneously CT-guided fine-needle aspirative biopsy of the gallbladder was nondiagnostic. The diagnosis of gallbladder carcinoma was considered preoperatively in three patients. Conclusion: Despite the characteristic histologic appearance of XCG, radiologic findings are nonspecific, varying from signs observed in other forms of cholecystitis to the appearance of a gallbladder neoplasm. Received: 5 April 1995/Accepted: 15 May 1995     

Evaluation of pancreatic carcinoma with FDG PET

Background: To assess the diagnostic usefulness and clinical impact of positron emission tomography with [F-18]fluorodeoxyglucose (FDG PET) on the management of patients with known or suspected pancreatic carcinoma. Methods: Attenuation-corrected FDG PET was performed in 20 patients (12 male, eight female) with pancreatic carcinoma at the time of initial diagnosis (n = 7), for tumor surveillance after Whipple surgery (n = 11), and for reevaluation after chemoradiation therapy (n = 2). Visual analysis of PET images were correlated with the results of abdominal computed tomography (CT) and carbohydrate antigen (CA) 19-9 serum tumor marker level that were obtained within 1 month of the PET study. Diagnostic validation was by histology in nine patients and by clinical or radiologic follow-up (5–48 months) in 11 patients. Changes in therapeutic management that were prompted by PET were tabulated. Results: PET was concordant with the findings of abdominal CT in 14 patients (13 true positive, 1 true negative). PET detected clinically unsuspected lung lesions, confirmed subsequently by a chest CT, in one of these 14 patients. PET was discordant with CT in six patients. PET detected tumor recurrence in three patients in this group (15% of total) with nondiagnostic CT findings and elevated CA 19-9 serology. In two of these three patients, chemotherapy with gemcitabine was initiated based on PET localization of disease. Tumor was confirmed in the remaining one of the three patients at autopsy shortly after the PET study. FDG localization in a displaced loop of bowel resulted in an apparent false-positive hepatic lesion in one of six patients in the discordant group. PET underestimated the extent of metastatic disease in the remaining two of six patients due to hyperglycemia. Conclusion: In patients with suspected pancreatic carcinoma at the time of initial presentation, PET is complementary to abdominal CT and allows detection of unsuspected distant metastases. In patients with suspected recurrent pancreatic carcinoma, based on elevated or rising CA 19-9 serology, PET can localize the disease when abdominal CT is nondiagnostic as a result of posttherapy anatomic alteration. Imaging evaluation with PET may impact the clinical management of patients with pancreatic carcinoma. Received: 1 August 2000/Accepted: 20 September 2000