Comparison of the Efficacy of Random Forest and Decision Tree Model in Predicting the Recurrence Risk of Common Bile Duct Stones After Laparoscopic Combined Choledochoscopy
MA Jianfeng
HEI Tao
ZHAO Zhengguo
YANG Weilong
Abstract：Objective To compare the efficacy of random forest and decision tree model in predicting the recurrence risk of common bile duct stones after laparoscopic combined choledochoscopy.Methods A total of 98 cases of choledocholithiasis were selected and operated by laparoscopy combined with choledochoscope,and were divided into recurrence group and non-recurrence group according to whether the choledocholithiasis recurred within 6 months after operation.The difference of baseline data between the two groups was compared,and the influencing factors were analyzed by logistic regression.The random forest model and decision tree model were constructed respectively,and the predictive efficacy of the two models for the recurrence risk of postoperative common bile duct stones was analyzed by using the receiver operating characteristics(ROC)curve.Results Logistic regression showed that the recurrence factors of postoperative stones were the history of biliary tract operation,infection,bile duct dilatation and juxtapapillary diverticulum,old age,long diameter of stones and common bile duct,and the number of stones ≥2(P＜0.05).ROC curve analysis showed that area under the curve(AUC)of the random forest model and the decision tree model for predicting the recurrence of common bile duct stones after operation were 0.928(95％CI:0.827-0.974)and 0.853(95％CI:0.796-0.916),respectively.The result of Delong test shows that the AUC value of random forest model was higher than that of decision tree model(D=3.452,P=0.012).Conclusion Both random forest and decision tree models have good prediction efficiency,and the prediction efficiency of random forest model is better than decision tree model.
Keywords：common bile duct stonesrecurrencelaparoscopycholedochoscoperandom forest modeldecision tree model
Publication Date：2025-04-28
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：5( 1415-1419 )