Dynamical Predictive Modelling of Cardiovascular Disease Progression Post-Myocardial Infarction via ECG-Trained Artificial Intelligence Model

2026-05-13

AI Summary

This study addresses the challenge of predicting adverse cardiovascular events following myocardial infarction, which is hindered by scarce labeled data and suboptimal performance of existing electrocardiogram (ECG)-based models. To overcome these limitations, the authors propose a deep temporal modeling approach that integrates self-supervised contrastive learning with multi-task fine-tuning. The method first pretrains on large-scale unlabeled ECG data using contrastive learning to effectively capture patient-specific temporal patterns, then employs a multi-task supervised head for few-shot fine-tuning to jointly predict multiple clinical outcomes. Evaluated on post-myocardial infarction prognosis prediction, the proposed model achieves an AUC of 0.794, substantially outperforming a baseline trained from scratch (AUC = 0.608), thereby demonstrating the efficacy and generalizability of the pretraining strategy in low-resource medical settings.

Abstract

Myocardial infarction (MI) is a leading cause of death, and its adverse outcomes are urgent to predict. Yet ECG-based prognostic models underperform because deep learning requires large, labelled datasets, which are scarce in medicine. Foundation models can learn from unlabelled ECGs via selfsupervision, but medically relevant training strategies remain underexplored. We propose a pretrained artificial intelligence model that combines patient-specific temporal information using contrastive learning with supervised multitask heads, then fine-tunes on post-MI outcome prediction. The proposed model outperformed a model trained from scratch (0.794 vs 0.608 AUC) showing that clinically structured ECG modelling improves classification in limited data regimes.

Problem

• myocardial infarction
• ECG
• predictive modeling
• cardiovascular disease progression
• limited labelled data

Innovation

• contrastive learning
• foundation model
• ECG-based prognosis
• multitask learning
• self-supervised pretraining