A comparative study of magnetic resonance cholangiography and direct cholangiography

OBJECTIVE: magnetic resonance cholangiopancreatography (MRCP) is a rapidly developing method for the noninvasive assessment of the biliary tree and pa ncreatic duct that obviates the need for contrast medium. We describe our e xperience with this new diagnostic imaging method in patients with obstruct ion of the biliary tree. We assessed both the location and cause of obstruc tion: and compared the results with direct cholangiography. METHODS: between 1997 and 1998, 81 patients under went MRCP at our facility . Two different image acquisition protocols (half-Fourier acquisition singl e-shot turbo spin-echo -HASTE- and rapid acquisition with relaxation enhanc ement -RARE-) for T2-weighted turbo spin echo sequences as well as cross-se ctional fast multiplanar gradient-echo pulse (TI-weighted FL2D) and T2-weig hted fast spin echo (T2 TSE) sequences were used. All patients underwent di rect (either percutaneous or endoscopic retrograde) cholangiography or surg ery for confirmation and/or treatment. The images obtained with MRCP were e valuated by two radiologists with expertise in biliary tree imaging who wer e unaware of the patient's clinical characteristics, and their diagnostic i nterpretations were compared with the findings obtained upon surgery or dir ect cholangiography. RESULTS: the sensitivity and specificity of MRCP in ruling out pathologies and detecting the presence of dilatation of the biliary tree were 100%. In assessing the level of the obstruction, sensitivity and specificity varied with location (intrahepatic/hilar, suprapancreatic, intrapancreatic or ampu llary). Sensitivity in these locations was 100, 92, 69 and 86%, respectivel y, whereas specificity was 100, 94, 92 and 91%, respectively. In determinin g the cause of the obstruction, the results were variable depending on the cause: choledocholithiasis (sensitivity, 89%; specificity, 90%) malignant o bstruction (sensitivity, 92%; specificity, 88%), benign structure (sensitiv ity, 63%; specificity, 90%), and chronic pancreatitis (sensitivity, 50%; sp ecificity, 99%). CONCLUSIONS: MRCP offered high diagnostic accuracy in the assessment of the occurrence and location of biliary obstruction. Sensitivity and specificit y in establishing the cause varied, and were highest for choledocholithiasi s and malignant obstruction. MRCP may be used instead of ERCP, which may th en be reserved for patients who are likely to require surgery.