Fine-tuning an ECG Foundation Model to Predict Coronary CT Angiography Outcomes

A multicenter study developed an AI-enabled electrocardiography (AI-ECG) model that predicts vessel-specific hemodynamically significant stenosis using coronary computed tomographic angiography (CCTA) as the reference. The model demonstrated strong discrimination in internal and external cohorts, including normal ECGs, and produced low-, intermediate-, and high-risk strata that correlated with stenosis severity and major adverse cardiovascular events. Calibration, decision curve analyses, and integration with guideline-based pre-test probability showed clinical utility, while waveform and attribution analyses revealed physiologically meaningful ECG features linked to high-risk predictions.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• fine-tuning