Establishment And Validation Of Prediction Models For Different Operative Timing Of Laparoscopic Cholecystectomy For Mild Acute Biliary Pancreatitis

Objective:Objective To establish and validate the prediction model of different operative timing of laparoscopic cholecystectomy for mild acute biliary pancreatitis.Methods:The clinical cases of mild acute biliary pancreatitis in the First Affiliated Hospital of Nanchang University were retrospectively analyzed,and the clinical cases of related patients were collected.A total of 480 patients were enrolled in the study and were randomized into a training set(320)and a validation set(160)in a 2:1 ratio.Taking the timing of surgery as the dependent variable,univariate analysis was firstly used to screen out the significant influencing factors(i.e.,risk factors)on the timing of surgery,and then multivariate analysis was performed with Logistic regression model.To further determine which independent risk factors have a definite effect on surgical timing and to develop a predictive model based on these independent risk factors.Then the ROC curve of the training set was drawn and Kappa consistency test was carried out.At the same time,the ROC curve of the validation set was used to verify the prediction model more accurately.Results:In the training set of 320 patients,five independent risk factors were identified,including: blood amylase(OR=1.087,95%CI:1.056-1.117),albumin(OR=0.841,95%CI: 0.812-0.870),blood calcium(OR=0.895,95%CI:0.870～0.921),total cholesterol(OR=1.231,95%CI:1.065～1.350)and gallbladder wall thickness(OR=1.261,95%CI:1.130～1.404).Through collinearity diagnosis,it is proved that there is no linear relationship between these 5 risk factors(VIF is all less than 10),and on this basis,a prediction model is established.The ROC curve of the prediction model was drawn and Kappa consistency test was carried out.The results showed that both the training set and the validation set had good differentiation(AUC=0.852 and 0.827)and good consistency(Kappa values were 0.467 and 0.479,P <0.05).These results indicate that the predictive model can be used to determine the operative timing of laparoscopic cholecystectomy in patients with mild acute biliary pancreatitis.Conclusions:The timing of laparoscopic cholecystectomy for patients with mild acute biliary pancreatitis is affected by several independent risk factors,and the Logit model composed of these independent risk factors can help to determine the timing of the operation.