Search

Radiologic findings of six cases of malignancy associated with chronic empyema 5-39 years in duration were reviewed. Pathologic examination confirmed three B-cell non-Hodgkin lymphomas, one round-cell sarcoma, one mesothelioma, and one adenocarcinoma. Retrospective findings on plain chest radiographs suggested the occurrence of malignancy: increased radiopacity in the thoracic cavity, soft-tissue bulgings and/or unsharpness of fat planes in the chest walls, destruction of bone near the empyema, and extensive medial deviation of the calcified pleurae. Computed tomography delineated masses with soft-tissue attenuation more clearly than radiography in all cases. Magnetic resonance images of three cases were informative because empyema cavities were surrounded by low-intensity rims, and two of them showed a signal intensity different from that of necrotic tumors. Scintigraphy revealed increased uptake of gallium in all cases. Ultrasonography was useful for biopsy guidance. Every radiologist should know this entity in observation of chest radiographs obtained in patients with chronic empyema, and further radiologic assessment and aggressive biopsy are recommended if malignancy is suspected.

Five cases of localized cystic dilatation of a side branch of the main pancreatic duct due to a new entity ("ductectatic" mucinous cystadenoma and cystadenocarcinoma) are reported. The dilated duct was widely covered by epithelium indistinguishable from that of mucinous cystadenoma (n = 4) or cystadenocarcinoma (n = 1) of the pancreas. All lesions were located in the uncinate process and were about 3 cm in size. On computed tomographic scans and sonograms, lesions were difficult to distinguish from simple cyst of the pancreas unless lobulated or irregular margins were demonstrated. Endoscopic retrograde pancreatography (ERP) or operative pancreatography clearly demonstrated characteristic findings: localized, prominent cystic dilatation of a side branch of the main pancreatic duct with grape-like clusters or pear-shaped pools of contrast material associated with filling defects of various sizes. When a cystic lesion is noted in the uncinate process of the pancreas, ERP is mandatory to confirm or rule out this potentially or overtly malignant disease.

Forty patients with hepatocellular carcinoma (HCC) (n = 22) or cavernous hemangioma (diameter, 3 cm or less) (n = 18) were examined with serial magnetic resonance (MR) imaging at 1.5 T with the fast low-angle shot (FLASH) technique and an intravenously administered bolus injection of 0.05 mmol/kg Gd-DTPA. Two images per minute were obtained for 5 minutes, and one per minute thereafter until about 10 minutes. Dynamic MR studies revealed five criteria for differentiating these tumors. With HCC there was a hyperintense mass before contrast material enhancement (32%), peak contrast enhancement at about 10 seconds after injection (55%), slight to moderate peak contrast enhancement (73%), absent or minimal delayed enhancement (100%), and, at morphologic study, a capsule or nodules-in-nodule appearance (59%). With hemangioma there was a hypointense mass before contrast enhancement (72%), peak contrast enhancement more than 2 minutes after injection (72%), marked peak contrast enhancement (83%), moderate to marked delayed enhancement (100%), and, at morphologic study, spreading phenomenon (39%). Seventy-seven percent of HCCs and 83% of hemangiomas satisfied three or more criteria and thus could be differentiated from each other with certainty.

Seventy-two patients with hepatocellular carcinoma (hepatoma) and 56 with hemangioma were studied with magnetic resonance (MR) imaging at 0.35 or 1.5 T to evaluate the efficacy of T2 values in differential diagnosis. T2 values were calculated with the two-point method. The mean T2 values of hepatoma and hemangioma were 58.9 msec +/- 8.9 and 101.6 msec +/- 25.8 at 0.35 T and 49.1 msec +/- 9.8 and 85.3 msec +/- 21.2 at 1.5 T. The difference in the T2 values for hepatoma and hemangioma was statistically significant (P less than .001) at both 0.35 and 1.5 T. Fifty-three of 56 lesions (94.6%) at 0.35 T and 86 of 102 lesions (84.3%) at 1.5 T were correctly classified when the T2 borderline between hepatoma and hemangioma was set at 80 msec. All misdiagnosed lesions were hemangioma, and all but one were smaller than 2 cm. However, over 90% of lesions smaller than 2 cm were correctly diagnosed when 70 msec at 0.35 T and 60 msec at 1.5 T were used as borderline T2 values. MR imaging with T2 measurement was very useful for differentiating between hepatoma and hemangioma (including small lesions) at 1.5 T as well as at 0.35 T.

Eight cases of choledochal cyst associated with biliary malignancy (gallbladder carcinoma in three and bile duct carcinoma in five) were reviewed to evaluate the roles and limitations of computed tomography (CT) (n = 8), ultrasound (US) (n = 6), cholangiography (n = 8), and angiography (n = 6). In cases of gallbladder carcinoma, both CT and US revealed mass lesions in the gallbladder consistent with cancer. CT also depicted either a mass lesion or an irregular thickened wall in all cases of bile duct carcinoma; however, US failed to demonstrate a thickened wall in one of three cases in which CT findings were positive. Cholangiography universally revealed malunion of the pancreatico-biliary duct, and the findings were suggestive of malignancy in seven cases in which CT depicted abnormalities. Angiography showed tumor stain in three of six cases. Lymph node metastases were present in four cases but were detected preoperatively in only one. One case showing a thickened bile duct wall was erroneously diagnosed as malignancy among 27 cases of choledochal cyst unassociated with biliary malignancy.

The authors retrospectively evaluated magnetic resonance (MR) images obtained at 1.5 T in 233 patients with portal hypertension and 91 subjects without it and pathologic findings in four resected spleens (one normal). Multiple, tiny (3-8 mm in diameter), low-intensity spots in the spleen were observed in 21 of 233 patients. Among the imaging studies performed in these 21 patients, the spots were seen on five of 14 T1-weighted images, 11 of 20 proton density images, and 12 of 20 T2-weighted images obtained with spin-echo techniques and on 14 of 14 fast-scan images obtained with gradient-echo rephasing. MR images in the 91 subjects did not show such lesions. MR images of the three spleens resected from patients with portal hypertension showed the low-intensity spots, which corresponded to siderotic nodules found at pathologic analysis. Despite limited pathologic confirmation, siderotic nodules (so-called Gamna-Gandy nodules) are considered the most likely cause of multiple low-intensity spots in the spleen.

The radiologic findings are described in six patients with "mucin-hypersecreting" intrahepatic biliary neoplasms (biliary cystadenocarcinoma or papillary cholangiocarcinoma). Massive amounts of mucin were confirmed in the dilated bile ducts at surgery or autopsy. Computed tomography (CT) and ultrasound (US) demonstrated severe dilatation of the intrahepatic bile ducts distal to the tumor in all patients and of extrahepatic ducts in five. On cholangiograms, large, amorphous filling defects caused by retention of mucin were found in the dilated ducts in three patients. The tumor was multilocular and cystic in five patients; in the other patient it was too small to be detected at CT or US. The presence of liver tumor (mostly multilocular and cystic), marked biliary dilatation distal to the tumor, and filling defects in the dilated bile ducts are important findings that indicate a mucin-hypersecreting intrahepatic biliary neoplasm.

We describe an application of simple scintigraphic localization of therapeutic gelatin emboli during and after transcatheter arterial embolization using gelatin particles labeled with technetium 99m. This scintigraphy required an additional 15-30 minutes and can be useful for prevention of serious complications and for early detection of adverse embolisms.

The computed tomographic (CT) and ultrasonographic (US) appearances of retained surgical sponges are described. In each case, the presence of a sponge was confirmed at repeat operation. CT examinations were performed in nine patients. In six cases, a low-density mass was demonstrated; in two cases, a complex mass with areas of both low and medium density was seen. The remaining case was a high-density mass. After intravenous administration of contrast material, dense and prolonged enhancement of the rim of the mass was noted in five of eight cases. Calcification was found in four cases, and spongiform gas in one. US performed in six patients revealed an echogenic area with strong acoustic shadowing in three cases and a complex mass, a hypoechoic mass, and a cystic mass with irregular internal echoes in one case each. Granuloma caused by a retained surgical sponge should be considered as a cause of an abdominal mass in patients with a history of prior abdominal surgery.

A large number of tiny (0.5-1.5 cm), low-intensity nodules of the liver were retrospectively observed on T2-weighted magnetic resonance (MR) images of eight patients with cirrhosis of the liver. The lesions were not detected with other imaging modalities and were considered to be regenerating nodules. Twenty-five consecutive cases of liver cirrhosis referred for MR examination were prospectively investigated to study the detectability of regenerating nodules and the optimal pulse sequence for detection. A large number of small, low-intensity nodules were demonstrated in nine of 25 cases on T2-weighted spin-echo (SE) images, in nine of 22 cases on T1-T2 complex fast low-angle shot (FLASH) images, and in one of eight cases on T1-weighted SE images. FLASH images obtained with respiratory suspension most clearly revealed regenerating nodules. However, T2-weighted SE images clearly permit differentiation of high-intensity hepatocellular carcinoma, including small daughter lesions, from low-intensity regenerating nodules.

Segmental intensity differences (SIDs) in hepatic parenchyma free of tumor were noted in six patients with hepatic masses (hepatocellular carcinoma in five and metastatic liver cancer in one). Areas of SID were homogeneous in intensity. The intensity of the affected region was high in all six patients on T2-weighted magnetic resonance (MR) images and low in two on T1-weighted images. Three of five patients examined with plain computed tomography (CT) had corresponding segmental areas of low attenuation. Angiograms obtained in five patients showed occlusion of the intrahepatic portal vein, segmental staining corresponding to the region of the SID, or both. Twelve of 82 patients examined with MR imaging and angiography had similar findings on angiograms, and ten of them had abnormal intensity of anatomic distribution around or beside the liver tumors on MR images. MR imaging may be more sensitive than plain CT in the detection of secondary changes caused by intrahepatic portal flow stoppage.

Five cases of prominent, diffuse dilatation of the main pancreatic duct caused by mucin-hypersecreting carcinoma are reported. Adenocarcinoma was noted mainly within one markedly dilated main pancreatic duct. On endoscopic retrograde pancreatograms, dilated main pancreatic ducts with amorphous or well-defined filling defects resulting from mucin (five patients) or tumor (one patient) were noted. The proximal portion of the main pancreatic duct alone was delineated in three cases because of reflux of contrast material as a result of the presence of excessive mucin and a patulous orifice of the papilla of Vater (major duodenal papilla). Diffuse dilatation of the main pancreatic duct was demonstrated or suggested on computed tomographic (CT) and ultrasound scans. Four patients who underwent operations are alive and free from carcinoma (mean, 50 months after diagnosis). Whenever prominent, diffuse dilatation of the main pancreatic duct is encountered on CT or ultrasound scans, endoscopic retrograde pancreatography is mandatory to permit confirmation or ruling out of mucin-hypersecreting carcinoma of the pancreas.

Thirty-six patients with hepatic tumors (28 hepatocellular carcinomas, seven cavernous hemangiomas, one metastatic tumor) were examined with serial magnetic resonance (MR) imaging, after a bolus intravenous injection of 0.05 mmol/kg gadolinium-diethylenetriaminepentaacetic acid. A typical MR imaging pattern for hemangiomas (present in five of seven cases [71.4%]) was a lesion of diminished signal intensity on precontrast images, peripheral contrast enhancement during the bolus dynamic phase, and complete fill-in of high signal intensity on delayed scan images. Twenty-eight hepatocellular carcinomas showed a variety of contrast enhancement patterns during the dynamic phase. In 21 patients (75%), there was no area of high signal intensity within the tumor on the delayed phase. A peripheral halo with delayed enhancement was noticed in 12 patients (42.8%) Histologic correlation in hepatocellular carcinomas showed that the degree of contrast enhancement corresponded to tumor vascularity and that the peripheral halo corresponded to fibrous capsular structure.