Explainable Artificial Intelligence-Driven Risk Assessment for Malignant Ventricular Arrhythmia and Mortality in Acute Myocardial Infarction

Malignant ventricular arrhythmia is a severe complication after acute myocardial infarction, often leading to sudden cardiac death. A total of 4,471 patients from two medical centers were included in this study. The primary endpoint was a composite of MVA and in-hospital death. Seven state-of-the-art AI models were developed and optimized by nested five-fold cross-validation. Predictive performance was evaluated using receiver operating characteristic curve (AUROC), calibration curve and decision analysis curve. Among the enrolled patients, 3,456 were used for model development and validation, and 1,015 cases from another medical center were used for external validation. In the validation group, the eXtreme Gradient Boosting (XGBoost) model achieved the highest AUROC of 0.792 (95% CI: 0.740-0.845) for the composite endpoint. Light Gradient Boosting Machine (LightGBM) demonstrated superior performance in MVA prediction (AUROC: 0.827, 95% CI: 0.768-0.885), while random forest (RF) outperformed others for mortality prediction (AUROC: 0.784, 95% CI: 0.720-0.848). In the external validation group, the AUROC of the XGBoost model with 15 variables for predicting the primary endpoint event was 0.726. The AUROCs were 0.704 for the LightGBM model with 15 variables for predicting MVA and 0.823 for the RF model with 20 variables for predicting in-hospital death. The web-based prediction system showed real-time risk assessment capabilities. This study presents an interpretable AI framework integrating multi-model analysis for AMI risk management. The developed system offers clinicians validated tool for personalized risk assessment, potentially improving patient outcomes through early intervention strategies.