Skip to content
Conference Open access

AI Based Early Prediction of Cardiac Events using Multi-Modal Clinical Data, ECG Analysis and Explainable AI

Sep 2026 · 2026 10th International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · pp. 349-356 · 0 citations · 18 references

Abstract

However, prediction of cardiac events is still not easy, as clinically meaningful data is spread across structured patient data and the electrocardiogram (ECG) signal data. This paper presents an interpretable multimodal artificial intelligence (AI) system that combines clinical and ECG data to predict the risk of a cardiac event. The approach uses modality-specific feature encoders to capture a complementary clinical, temporal and morphological representation, followed by adaptive clinical–ECG feature fusion, calibrated risk estimation and explainability at the level of the patient. The experiments were performed using 5000 patient records and the accuracy, precision, recall, F1-score and area under the receiver operating characteristic curve (AUC) achieved were 98.62%, 98.47%, 98.31%, 98.39% and 99.08% respectively. Multimodal integration, adaptive fusion and ECG temporal learning are validated via comparative and ablation analyses. Explainability analysis revealed that the ST-segment variation, heart rate, QT interval, age and systolic blood pressure were important predictors. The proposed system will facilitate multimodal cardiac-risk stratification with high accuracy and interpretability, and enable the identification of patients at high risk at an earlier stage.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.