To evaluate the diagnostic efficiency of the magnetic resonance cholangiography (MRC) in the detection of main bile duct stones in a set of 102 patients. METHODS Criteria of inclusion were: Clinic and biological suspicion of biliary stones obstruction with exams of first intention no contributive. We used the "turbo spin echo" sequences with thick slices in single shot mode and fine slides with reconstruction in 3D by a computer. Exams of reference were the endoscopic retrograde cholangiography (76.47%), an intraoperative cholangiography (20.59%) and a per-cutaneous cholangiography (2.94%). RESULTS Stones of the main bile duct have been diagnosed at thirty-five patient (35.7%); we had 3 positive forgeries and 6 negative forgeries of the MRC. The sensitivity was 82,9%, the specificity of 95,5%, the positive predictive value and the negative predictive value were, respectively, of 90,6% and 91,4%. The observant variance test was excellent (kappa = 0.83). Mistakes of diagnosis of the MRC were bound to: stones less than 3 mms with a bile duct no dilated, malignant stenosis, structural details as the presence of a duodenal diverticula's or severe duodenitis and a certain difficulty to see the sphincter complex. CONCLUSIONS Performances of the CIRM was good, and only in very particular cases, it was the origin of confusions.