OBJECTIVE The clinical importance of magnetic resonance cholangiopancreatography (MRCP) as a noninvasive diagnostic modality for investigation of the biliary tree and pancreatic duct system is under debate. Using endoscopic retrograde cholangiopancreatography (ERCP) as the gold standard, this study determined in a prospective, blinded fashion the sensitivity and further statistic values of MRCP findings for evaluation of the biliary and pancreatic tract. METHODS Seventy-eight patients referred for ERCP were studied prospectively with MRCP and ERCP during a 12-month period. All images were interpreted on a blinded basis by two radiologists. Any dilations, strictures, and intraductal abnormalities were recorded and correlated with the clinical diagnoses. RESULTS MRCP images of diagnostic quality were obtained in 76 of the 78 patients (97%). Magnetic resonance cholangiography (MRC) showed sensitivities (and positive predictive values) of 71% (62%) for recognition of normal bile ducts, 83% (91%) for recognition of dilation, 85% (100%) for recognition of strictures, 77% (91%) for correct stricture location, and 80% (100%) for diagnosing bile duct calculi. In addition, the sensitivity of MRC in classifying benign and malignant strictures was 50% and 80%, respectively. The statistical values (sensitivity and positive predictive value) for magnetic resonance pancreatography findings were determined for the recognition of normal pancreatic ducts (33% and 50%), recognition of dilation (62% and 100%), recognition of strictures (76% and 87%) and correct location (66% and 100%), diagnosis of benign strictures (87% and 87%) and malignant strictures (60% and 75%), and for diagnosing pancreatic duct stones (60% and 100%). CONCLUSIONS MRCP is capable of providing diagnostic information equivalent to ERCP in many patients, and should be applied whenever established techniques provide no results, or inadequate results.