Predictive model enhances early diagnosis of gangrenous cholecystitis

A novel machine learning model has been developed to predict gangrenous cholecystitis, utilizing clinical data from 1006 patients. Employing the XGBoost algorithm and SHAP for interpretability, key clinical features were identified, including white blood cell count and d-dimer levels. This explainable model aims to assist clinicians in making timely surgical decisions, thereby potentially improving patient outcomes by facilitating early intervention for this serious condition associated with high morbidity and mortality rates.