Diagnostic accuracy of magnetic resonance cholangiopancreatography and ultrasound compared with direct cholangiography in the detection of choledocholithiasis

AIM: To determine the diagnostic accuracy of magnetic resonance cholangiopancreatography (MRCP) and trans-abdominal ultrasound in the detection of choledocholithiasis, and to compare bile duct stone characteristics using endoscopic retrograde cholangiopancreatography (ERCP), MRCP and ultrasound. MATERIALS AND METHODS: Of 191 consecutive patients referred for diagnostic ERCP, choledocholithiasis was diagnosed in 34 patients using direct cholangiography. The latter took the form of ERCP (n = 29), intraoperative cholangiography (n = 3) or percutaneous transhepatic cholangiography (n = 2). All patients underwent MRCP and ultrasound examinations and their findings for choledocholitiasis were compared with those at direct cholangiography. Finally, in the 29 patients with choledocholithiasis diagnosed under ERCP, stone characteristics were compared across the three investigations of ERCP, MRCP and ultrasound. RESULTS: Compared with direct cholangiography, MRCP showed a sensitivity, specificity and diagnostic accuracy of 91%, 98% and 97%, respectively, in the diagnosis of choledocholithiasis. MRCP resulted in three false-negative and three false-positive findings, four of which occurred due to confusion with lesions at the ampulla. Ultrasound showed a sensitivity, specificity and diagnostic accuracy of 38%, 100% and 89%, respectively, in the diagnosis of choledocholithiasis. ERCP revealed a greater number of stones and these were more proximally distributed within the bile ducts when compared to MRCP. CONCLUSIONS: MRCP is highly accurate in the diagnosis of choloedocholithiasis and has the potential to replace diagnostic ERCP. MRCP underestimates the number of bile duct stones present. (C) 2000 The Royal College of Radiologists.