Predictive model identifies factors for ERCP post-cholecystectomy

A machine learning-based predictive model has been developed to estimate the incidence of endoscopic retrograde cholangiopancreatography (ERCP) after emergency laparoscopic cholecystectomy. Analyzing data from 8,854 patients, the model revealed a postoperative ERCP incidence of 5.7% in training and 6.4% in testing datasets. The gradient-boosting decision tree model excelled, with common bile duct dilatation, serum albumin, and lactate dehydrogenase as significant predictors. This approach addresses critical challenges in managing post-operative biliary complications.