Simple risk score to predict the likelihood of a positive EUS in idiopathic acute pancreatitis

Abstract

We sought to derive a risk score, DORM65, of known variables to predict the likelihood of a positive EUS in patients with idiopathic acute pancreatitis (IAP). A retrospective cohort study of 180 patients with IAP was performed across 3 tertiary care centers between January 2018 and December 2021. Multivariate logistic regression modeling was performed to predict a positive EUS. Accuracy of the models was assessed by the area under the receiver-operating characteristic curve (AUROCC). The diagnostic yield of EUS was 58.9% (95% confidence interval [CI], 51.7-66.1). The DORM65 scores of 5 predictors present before EUS with the best discrimination were a delayed EUS (defined as≥82 days from the last episode of AP), obesity, not having had a repeated transabdominal US, male sex, and age≥65 years at the time of EUS. For those at the lowest risk score group, the positive EUS rate was 13.0% compared with 100% in those at the highest risk group (relative risk, 7.67; P< .001). A score of 3 or more had a positive predictive value of 86.0% with a sensitivity of 34.9% and specificity of 91.9%. The model had a high predictive accuracy (AUROCC, .774; 95% CI, .707-.841). Adding 3 additional predictors (no cholecystectomy, no MRCP, and a single episode of AP) did not increase the accuracy significantly (AUROCC, .805; 95% CI, .742-.867). DORM65 is easily calculated and accurately predicts a positive EUS in patients with IAP. Further validation is needed.