Can endoscopic ultrasound or magnetic resonance cholangiopancreatography replace FRCP in patients with suspected biliary disease? A prospective trialand cost analysis

OBJECTIVES: FRCP is the gold standard for pancreaticobiliary evaluation but is associated with complications. Less invasive diagnostic alternatives wi th similar capabilities may be cost-effective, particularly in situations i nvolving low prevalence of disease. The aim of this study was to compare th e performance of endoscopic ultrasound (EUS) with magnetic resonance cholan giopancreatography (MRCP) and ERCP in the same patients with suspected extr ahepatic biliary disease. The economic outcomes of EUS-, MRCP-, and FRCP-ba sed diagnostic strategies were evaluated. METHODS: Prospective cohort study of patients referred for FRCP with suspec ted biliary disease. MRCP and EUS were performed within 24 h before ERCP. T he investigators were blinded to the results of the alternative imaging stu dies. A cost-utility analysis was performed for initial ERCP, MRCP, and EUS strategies for these patients. RESULTS: A total of 30 patients were studied. ERCP cholangiogram failed in one patient, and another patient did not complete MRCP because of claustrop hobia. The final diagnoses (N = 28) were CBD stone (mean = 4 mm; range = 3- 6 mm) in five patients; biliary stricture in three patients, and normal bil iary tree in 20. Two patients had pancreatitis after therapeutic ERCP, one after precut sphincterotomy followed by a normal cholangiogram. EUS was mor e sensitive than MRCP in the detection of choledocolithiasis (80% vs 40%), with similar specificity. MRCP had a poor specificity and positive predicti ve value for the diagnosis of biliary stricture (76%/25%) compared to EUS ( 100%/100%), with similar sensitivity. The overall accuracy of MRCP for any abnormality was 61% (95% CI = 0.41-0.78) compared to 89% (CI = 0.72-0.98) f or EUS. Among those patients with a normal biliary tree, the proportion cor rectly identified with each test was 95% for EUS and 65% for MRCP (p < 0.02 ). The cost for each strategy per patient evaluated was $1346 for FRCP, $11 11 for EUS, and $1145 for MRCP. CONCLUSIONS: In this patient population with a low disease prevalence, EUS was superior to MRCP for choledocholithiasis. EUS was most useful for confi rming a normal biliary tree and should be considered a low-risk alternative to FRCP. Although MRCP had the lowest procedural reimbursement, the initia l EUS strategy had the greatest cost-utility by avoiding unnecessary ERCP e xaminations. (Am J Gastroenterol 2001;96:2900-2904. (C) 2001 by Am. Coll. o f Gastroenterology).