Background: The differential diagnosis of biliary strictures remains a challenge. This study evaluated magnetic resonance cholangiopancreatography (MRCP) as a new procedure in comparison with the established methods of diagnosis including ERCP or percutaneous transhepatic cholangiography (PTC), CT, and EUS.
Methods: Fifty patients (21 men, 29 women, mean age 65.7 years) with jaundice but no pain suspected to have biliary strictures were enrolled in this prospective study. MRCP, ERCP/PTC, CT, and EUS were performed prospectively; images and videotapes (EUS) of these tests were reviewed blindly under standardized conditions. Reference standards for comparison were surgery, a biopsy confirming malignancy, or the clinical course during follow-up (at least 12 months) in cases without histopathologic proof of malignancy.
Results: Seven patients ultimately proved to have jaundice caused by parenchymal liver disease and 43 had a biliary stricture (17 benign, 26 malignant). Forty patients underwent all 4 imaging tests. There were 10 patients in whom patient-specific problems precluded some procedures but who were included in an intention-to-diagnose analysis. The sensitivity and specificity for diagnosis of malignancy in the 50 patients were as follows: 85% / 75% for ERCP/PTC, 85% / 71% for MRCP, 77% / 63% for CT, and 79% / 62% for EUS, with similar values in the 40 patients who underwent all 4 imaging methods. The combination of MRCP and EUS improved specificity.
Conclusions: Although MRCP provides the same imaging information as direct cholangiography, it has limited specificity for the diagnosis of malignant strictures. In the differential diagnosis of biliary strictures, several tests including ERCP-guided tissue diagnosis are still required, and MRCP has only a limited clinical role.