An explainable predictive machine learning model of gangrenous cholecystitis based on clinical data: a retrospective single center study

Abstract

Gangrenous cholecystitis (GC) is a serious clinical condition associated with high morbidity and mortality rates. Machine learning (ML) has significant potential in addressing the diverse characteristics of real data. We aim to develop an explainable and cost-effective predictive model for GC utilizing ML and Shapley Additive explanation (SHAP) algorithm. This study included a total of 1006 patients with 26 clinical features. Through 5-fold CV, the best performing integrated learning model, XGBoost, was identified. The model was interpreted using SHAP to derive the feature subsets WBC, NLR, D-dimer, Gallbladder width, Fibrinogen, Gallbladder wallness, Hypokalemia or hyponatremia, these subsets comprised the final diagnostic prediction model. The study developed a explainable predictive tool for GC at an early stage. This could assist doctors to make quick surgical intervention decisions and perform surgery on patients with GC as soon as possible. Using clinical data from 1006 cholecystitis patients, we developed a machine learning-based diagnostic prediction model to help identify patients at high risk for acute gangrenous cholecystitis. During the study, the deficiency and imbalance of actual clinical data were directly addressed, leading to the ultimate selection of the integrated learning model XGBoost as the predictive model exhibiting superior performance and stability on a novel, unidentified validation set and compared to preoperative clinical diagnosis. The model employs variables that are non-specific, readily available, reasonably priced, and appropriate for clinical generalization.