This research demonstrates enhanced risk assessment and prediction for acute coronary syndrome using ecg foundation models, suggesting benefits of model fusion.
Key Points
Main finding: The fusion of ecg foundation model embeddings significantly improves the early detection of acute coronary syndrome.
Key evidence: The fusion approach achieved the highest performance with AUROC of 0.843 and AUCPR of 0.674, surpassing the baseline model.
Approach: This study utilized individual and fused ecg foundation models based on self-supervised learning to analyze prehospital ecg data.
Significance: These findings underscore the importance of utilizing advanced ecg models for improved clinical outcomes in acute coronary syndrome cases.