Prospective comparison of radial and linear endoscopic ultrasound for diagnosis of chronic pancreatitis

Abstract

Linear endoscopic ultrasonography (EUS) is currently favored by many endosonographers for the evaluation of pancreatic pathology. However, radial EUS was used in early studies validating EUS for chronic pancreatitis. Radial and linear EUS have never been compared for the diagnosis of chronic pancreatitis. The aim of this study was to compare radial and linear EUS for the diagnosis of chronic pancreatitis using the secretin-stimulated endoscopic pancreatic function test (ePFT) as the reference standard. One hundred consecutive patients evaluated for pain of possible pancreatic origin underwent combined radial EUS, linear EUS, and secretin ePFT during a single endoscopic session. EUS images were acquired on videotape and blindly scored by three reviewers. The main outcome measure was diagnostic accuracy. The accuracy of radial EUS and linear EUS (cutpoint > or = 4 criteria) was 84 % and 74 %, respectively. The statistical test for noninferiority was significant ( P < 0.001) suggesting that the accuracy of radial EUS is as good as or superior to linear EUS. The ratio of accuracy (pi (radial)/pi (linear)) was 1.14 (95 % confidence interval [CI] 0.99 to 1.28). No statistically significant differences were found between radial and linear EUS in terms of sensitivity, specificity, or overall discriminative ability (area under receiver operating characteristic curve 0.84 vs. 0.76, P = 0.10). Interobserver variability was similar for radial (Fleiss' kappa = 0.61, 95 %CI 0.43 to 0.79) and linear EUS (kappa = 0.50, 95 %CI 0.28 to 0.72). The accuracy of radial EUS is as good as linear EUS for the diagnosis of chronic pancreatitis.