AI-driven Models with Effective Feature Selection Accurately Predict ICU Admission after Laparoscopic Cholecystectomy

Laparoscopic Cholecystectomy (LC) is widely recognized for its benefits over open cholecystectomy, such as reduced hospital stay, postoperative pain, and mortality. Despite its safety, the risk of postoperative complications, particularly those necessitating ICU admission, remains a concern. This study aims to develop AI-driven models to predict ICU admission following LC and identify key patient-specific risk factors. The study population included 1,411 patients who underwent LC at the Irvine Medical Center between 2017 and 2022. The primary outcome was ICU admission after LC. The study analyzed 51 variables comprising patient demographics, medical history, medications, lab results, and postoperative complications using ‘recursive feature elimination with cross-validation, and five AI-driven models (Random Forest, Decision Tree, Support Vector Machine, Neural Network, and Logistic Regression). Shapley additive exPlanation (SHAP) was performed to visualize and interpret the model’s performance. Model performance was assessed using the area under the receiver operating characteristics (AUROC), recall, accuracy, F1 score, and Matthews Correlation Coefficient (MCC). Results demonstrated length of stay, anesthesia duration, and discharge disposition as the top risk factors impacting ICU admission after LC. The Random Forest model demonstrated the highest predictive performance with an AUROC of 0.83. This study is the first of its kind to develop and validate AI-driven models for predicting ICU admission after LC. Future research should focus on external validation and development of AI-driven clinical decision support tools to enhance postoperative risk stratification and resource management in clinical settings.