Role of magnetic resonance in characterising extrahepatic cholangiocarcinomas

PURPOSE: The purpose of this study was to evaluate the accuracy of magnetic resonance (MR) in correctly locating and characterising biliary strictures in patients affected by extrahepatic cholangiocarcinoma, identify findings suggestive of the disease, identify lesions with similar MR features and possible criteria for differential diagnosis and establish prospective MR accuracy in diagnosis of malignant obstruction of extrahepatic bile ducts. MATERIALS AND METHODS: We retrospectively reviewed the MR examinations of 39 patients affected by extrahepatic cholangiocarcinoma confirmed by histology or cytology. The studies were evaluated for the following parameters: site of obstruction (hilar, proximal or distal), presence of intra- or extrahepatic dilation of bile ducts, morphology of ductal stenosis (gradual tapering or abrupt ending), morphology of the lesion (mass like or circumferential), dimension, signal intensity before contrast medium administration and lesion enhancement after administration of contrast medium. Finally, we assessed the most useful sequence for the diagnosis. In order to evaluate MR accuracy in the diagnosis of malignant obstruction of extrahepatic bile ducts, we prospectively reviewed MR examinations of 74 patients affected by obstructive jaundice (55 malignant lesions and 19 inflammatory lesions). MR diagnosis was compared with histology or cytology considered as the gold standard. RESULTS: MR allowed identification and localisation of 41/41 extrahepatic cholangiocarcinomas. Fifty-four percent of the lesions showed gradual duct tapering; the remaining lesions showed an abrupt ending. Fifty-six percent of the lesions appeared as a circumferential thickening (infiltrative growth); the remaining lesions had a mass-like appearance (expansile growth). Most lesions were hypo- (49%) or isointense (49%) in T1-weighted sequences and hyper- (49%) or isointense (51%) in T2-weighted sequences. Ninety-five percent of the lesions did not enhance significantly in the arterial phase while 98% showed late enhancement (10 min). The most diagnostic sequence (in 76% of cases) was the late-phase gradient-echo (GRE) T1 fat-saturated sequence. MR had good sensitivity (91%) but poor specificity (47%) in characterising stenosis as malignant, given the large number (10/19) of benign lesions evaluated as neoplastic lesions. CONCLUSIONS: MR almost always identified the cause of stenosis and suggested its neoplastic nature if it exhibited a mass-like appearance (extraductal or growing into the choledochus). On the other hand, lesions with parietal thickening, particularly if smaller than 1 cm, require endoscopic cytology or histology because of the high risk of unnecessary procedures for benign lesions.