Multimodal Ensemble Learning for Coronary Artery Disease Risk Stratification Using ECG and Clinical Biomarkers

Coronary artery disease (CAD) remains one of the leading causes of mortality worldwide, necessitating accurate and early risk prediction to support timely clinical decision-making. Recent advances in artificial intelligence (AI) have demonstrated promising results in CAD detection using either electrocardiogram (ECG) signals or clinical biomarkers; however, single-modality approaches often fail to capture the complex and multifactorial nature of cardiovascular disease. In this paper, we propose a hybrid ensemble-based AI framework for CAD risk prediction that integrates ECG-derived features with lipid profile parameters to improve predictive performance and interpretability. Separate machine learning models are trained for each modality, and their outputs are combined using an ensemble learning strategy to generate a unified risk score. To enhance clinical transparency, explainable AI techniques are incorporated to identify the contribution of individual features toward model predictions. The proposed framework is evaluated using publicly available datasets, and experimental results demonstrate improved accuracy, robustness, and generalizability compared to standalone modality-specific models. The developed system highlights the potential of multimodal data fusion and explainable ensemble learning for reliable CAD risk stratification and supports its applicability in real-world clinical decision support systems.