Is MRCP worthwhile when ERCP is planned? A prospective trial in a referral center

AIMS: 1) Does magnetic resonance cholangiopancreatography (MRCP) add to ERCP? 2) Does MRCP narrow the differential diagnosis (DDx) and/or obviate the need for diagnostic (Dx) and/or therapeutic (Rx) ERCP?. METHODS: All new patients (pts) referred to our pancreato-biliary service, in whom ERCP was planned, also underwent MRCP when available. PART 1 (DOES MRCP ADD TO ERCP?): Attendings performed ERCP while blinded to MRCP. They assigned pre and post-test probabilities for pre-specified differential diagnoses (6 possible per context) before and after ERCP and then after the MRCP result was received (reported) on standardized form by 1 of 2 radiologists + films available). PART 2 (DOES MRCP REDUCE THE DDx AND/OR PREVENT ERCP?): For each pt, conclusions were made by another attending who was aware of the complete history (Hx) + MRCP, but blinded to ERCP results. At each step, physicians also rated the need for Dx and Rx ERCP as: Indicated / Unsure / No. RESULTS: 40 pts studied to date (study ongoing). Mean age 52,±16; 53% female. Contexts: Abnormal liver enzymes 29%; suspected chronic pancreatitis 26%; pre or post cholecystectomy pain 18%; recurrent acute pancreatitis 16%; jaundice 11%. PART 1: Adding MRCP to ERCP did not reduce the # DDx significantly (mean # DDx: without MRCP 1.7±0.9 vs after adding MRCP 1.4±0.6; p=0.09); and would have prevented no Rx ERCPs. PART 2: Using MRCP alone (without ERCP information) reduced the mean # DDx; but less so than using ERCP alone (mean reduction in DDx: Hx+ERCP -1.1±1.2 vs Hx+MRCP -0.6±1.0; p=0.06). MRCP would have prevented no Dx ERCPs and only 1/40(3%) of Rx ERCPs. PART 1 PART 2 ERCP type Dx Rx Dx Rx ERCP needed? U No U No U No U No (U-unsure) After history only 10% 0% 82% 0% 10% 0% 68% 0% Dx ERCP information added - - 23% 10% - - - - MRCP information added - - 20% 10% 7% 0% 45% 3% The results did not change significantly after stratifying by the perceived need for sphincter of Oddi (SO) manometry or by clinical context. CONCLUSIONS: When ERCP is planned at our tertiary referral center, adding MRCP to ERCP does not reduce the # DDx and does not prevent Dx or Rx ERCP. MRCP alone narrows the DDx (but less so than Dx ERCP), but prevents no Dx ERCP and <5% of Rx ERCPs. These results appear unrelated to the need for SO manometry. These preliminary findings need exploration in different patient groups and practice settings.