Domain-Adapted Fine-Tuning of ECG Foundation Models for Multi-Label Structural Heart Disease Screening
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914507631099904 |
|---|---|
| author | Do, Duc N. Do, Minh N. Nguyen, Dang Le, Khanh T. Q. Pham, Khoa D. Huynh, Hung N. Pham-Van-Hoang, Phi Huynh, Quan K. Odat, Ramez M. Ashar, Perisa Lowder, Ethan Philip Le, Minh H. N. Le, Hoang Nguyen, Phat V. H. Le, Quan Kpodonu, Jacques Huynh, Phat K. |
| author_facet | Do, Duc N. Do, Minh N. Nguyen, Dang Le, Khanh T. Q. Pham, Khoa D. Huynh, Hung N. Pham-Van-Hoang, Phi Huynh, Quan K. Odat, Ramez M. Ashar, Perisa Lowder, Ethan Philip Le, Minh H. N. Le, Hoang Nguyen, Phat V. H. Le, Quan Kpodonu, Jacques Huynh, Phat K. |
| contents | Transthoracic echocardiography is the reference standard for confirming structural heart disease (SHD), but first-line screening is limited by cost, workflow burden, and specialist availability. We evaluated whether open pretrained electrocardiogram (ECG) foundation models can support echo-confirmed multi-label SHD detection using the public EchoNext Mini-Model benchmark. Six echocardiography-derived abnormalities were targeted: reduced left ventricular ejection fraction, increased left ventricular wall thickness, aortic stenosis, mitral regurgitation, tricuspid regurgitation, and right ventricular systolic dysfunction. Under a common pipeline, we compared engineered ECG features with gradient boosting, end-to-end waveform learning from scratch, and transfer from open ECG foundation models. We then applied in-domain self-supervised adaptation of an ECG foundation model (ECG-FM) on EchoNext waveforms followed by selective supervised fine-tuning, and evaluated trade-offs between discrimination and adaptation cost. Adapted ECG-FM models achieved the best overall performance: peak macro-AUROC 0.8509 and macro-AUPRC 0.4297, while a parameter-efficient operating point preserved AUROC (0.8501) and attained the highest fixed-threshold macro-F1 0.3691. Late fusion with covariates did not improve threshold-independent discrimination, and evaluated LoRA, alternative backbones, and mixture-of-foundations strategies did not surpass the best adapted single-backbone models. These results indicate that for ECG-based case finding and echocardiography triage, combining target-domain self-supervised adaptation with selective supervised updating of a pretrained ECG backbone is the most effective transfer strategy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_23385 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Domain-Adapted Fine-Tuning of ECG Foundation Models for Multi-Label Structural Heart Disease Screening Do, Duc N. Do, Minh N. Nguyen, Dang Le, Khanh T. Q. Pham, Khoa D. Huynh, Hung N. Pham-Van-Hoang, Phi Huynh, Quan K. Odat, Ramez M. Ashar, Perisa Lowder, Ethan Philip Le, Minh H. N. Le, Hoang Nguyen, Phat V. H. Le, Quan Kpodonu, Jacques Huynh, Phat K. Machine Learning Transthoracic echocardiography is the reference standard for confirming structural heart disease (SHD), but first-line screening is limited by cost, workflow burden, and specialist availability. We evaluated whether open pretrained electrocardiogram (ECG) foundation models can support echo-confirmed multi-label SHD detection using the public EchoNext Mini-Model benchmark. Six echocardiography-derived abnormalities were targeted: reduced left ventricular ejection fraction, increased left ventricular wall thickness, aortic stenosis, mitral regurgitation, tricuspid regurgitation, and right ventricular systolic dysfunction. Under a common pipeline, we compared engineered ECG features with gradient boosting, end-to-end waveform learning from scratch, and transfer from open ECG foundation models. We then applied in-domain self-supervised adaptation of an ECG foundation model (ECG-FM) on EchoNext waveforms followed by selective supervised fine-tuning, and evaluated trade-offs between discrimination and adaptation cost. Adapted ECG-FM models achieved the best overall performance: peak macro-AUROC 0.8509 and macro-AUPRC 0.4297, while a parameter-efficient operating point preserved AUROC (0.8501) and attained the highest fixed-threshold macro-F1 0.3691. Late fusion with covariates did not improve threshold-independent discrimination, and evaluated LoRA, alternative backbones, and mixture-of-foundations strategies did not surpass the best adapted single-backbone models. These results indicate that for ECG-based case finding and echocardiography triage, combining target-domain self-supervised adaptation with selective supervised updating of a pretrained ECG backbone is the most effective transfer strategy. |
| title | Domain-Adapted Fine-Tuning of ECG Foundation Models for Multi-Label Structural Heart Disease Screening |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2604.23385 |