Multicenter comparison of the interobserver agreement of standard EUS scoring and Rosemont classification scoring for diagnosis of chronic pancreatitis

Background EUS has less than optimal interobserver agreement for the diagnosis of chronic pancreatitis. The newly developed Rosemont consensus scoring system includes weighted criteria and stricter definitions for individual features. Objective The primary aim was to compare the interobserver agreement of standard and Rosemont scoring. Setting Multiple tertiary-care institutions. Intervention Fifty EUS videos were interpreted by 14 experts. Each expert interpreted the videos on two occasions: First, the videos were read by using standard scoring (9 criteria). Second, after viewing a presentation of the Rosemont classification, the same experts re-read the videos by using Rosemont scoring. Main Outcome Measurements Fleiss' kappa (K) statistics are reported with 95% confidence intervals (CI). Results The interobserver agreement was "substantial" (K = 0.65 [95% CI, 0.52-0.77]) for Rosemont scoring and "moderate" (K = 0.54 [95% CI, 0.44-0.66]) for standard scoring; however, the difference was not statistically significant ( P = 0.12). Limitations The sample size does not allow detection of differences in K of Conclusion Use of the Rosemont classification did not significantly increase interobserver agreement for EUS diagnosis of chronic pancreatitis compared with standard scoring.