Preoperative prediction model for 30 days biliary events after laparoscopic cholecystectomy

Figure 1 Coefficient paths across log(λ). The X-axis shows log(λ); the Y-axis shows standardized coefficients. The numbers at the top indicate the count of non-zero coefficients decreasing from 16 to 0. The two dashed lines indicate λ_min and λ_1se.

Figure 2 Ten-fold cross-validation curve [binomial deviance vs log(λ)]. Red dots represent mean deviance, and vertical bars represent ± 1 standard error. The left dashed line indicates λ_min; the right dashed line indicates λ_1se selected by the one-standard-error rule.

Figure 3  Forest plot of multivariable logistic regression showing odds ratios and 95%CIs for candidate predictors retained after least absolute shrinkage and selection operator selection.

Figure 4 Nomogram based on total bilirubin and the common bile duct/height index. The top axis represents the point scale for each predictor, the middle axis represents total points, and the bottom axis shows the corresponding predicted probability of 30-day clinically relevant biliary events. Higher total points indicate higher predicted risk.

Figure 5 Receiver operating characteristic curve of the final model. A: Training cohort; B: Validation cohort. AUC: Area under the receiver operating characteristic curve.

Figure 6 Calibration plot in the training cohort. A: Training cohort; B: Validation cohort. The gray line indicates perfect calibration; the black line is the logistic calibration curve; and the dashed line is the nonparametric smooth curve.

Figure 7 Decision curve analysis of the final model. A: Training cohort; B: Validation cohort.