BACKGROUND: Numerous published studies have shown the high diagnostic performance of both EUS and MRCP compared with ERCP for the detection of choledocholithiasis. DESIGN: We undertook a systematic review of all published randomized, prospective trials that compared EUS with MRCP with the primary aim being to compare the overall diagnostic accuracy for the detection of choledocholithiasis in patients with suspected biliary disease. METHODS: A MEDLINE review was performed. We identified 5 randomized, prospective, blinded trials comparing MRCP and EUS for the detection of choledocholithiasis, with subsequent ERCP or intraoperative cholangiography as a criterion standard. The study-specific variables for EUS and MRCP for choledocholithiasis were calculated from the data, and analyses were performed by using aggregated variables (sensitivity, specificity, positive and negative predictive values, and likelihood ratios). RESULTS: The pooled data set consisted of 301 patients. The aggregated sensitivities of EUS and MRCP for the detection of choledocholithiasis were 0.93 and 0.85, respectively, whereas their specificities were 0.96 and 0.93, respectively. The aggregated positive predictive values for EUS and MRCP were 0.93 and 0.87, respectively, with the corresponding negative predictive values of 0.96 and 0.92, respectively. Positive likelihood ratios were >10 for both tests, and corresponding negative likelihood ratios approached 0.10 for both tests. No statistically significant differences between EUS and MRCP were found in our analysis. CONCLUSIONS: EUS and MRCP have high diagnostic performance overall. Our analysis showed no statistically significant difference between the modalities. We recommend taking into consideration other factors, such as resource availability, experience, and cost considerations in deciding between these 2 tests.
BACKGROUND: Numerous published studies have shown the high diagnostic performance of both EUS and MRCP compared with ERCP for the detection of choledocholithiasis. DESIGN: We undertook a systematic review of all published randomized, prospective trials that compared EUS with MRCP with the primary aim being to compare the overall diagnostic accuracy for the detection of choledocholithiasis in patients with suspected biliary disease. METHODS: A MEDLINE review was performed. We identified 5 randomized, prospective, blinded trials comparing MRCP and EUS for the detection of choledocholithiasis, with subsequent ERCP or intraoperative cholangiography as a criterion standard. The study-specific variables for EUS and MRCP for choledocholithiasis were calculated from the data, and analyses were performed by using aggregated variables (sensitivity, specificity, positive and negative predictive values, and likelihood ratios). RESULTS: The pooled data set consisted of 301 patients. The aggregated sensitivities of EUS and MRCP for the detection of choledocholithiasis were 0.93 and 0.85, respectively, whereas their specificities were 0.96 and 0.93, respectively. The aggregated positive predictive values for EUS and MRCP were 0.93 and 0.87, respectively, with the corresponding negative predictive values of 0.96 and 0.92, respectively. Positive likelihood ratios were >10 for both tests, and corresponding negative likelihood ratios approached 0.10 for both tests. No statistically significant differences between EUS and MRCP were found in our analysis. CONCLUSIONS: EUS and MRCP have high diagnostic performance overall. Our analysis showed no statistically significant difference between the modalities. We recommend taking into consideration other factors, such as resource availability, experience, and cost considerations in deciding between these 2 tests.