adult-to-pediatric fine-tuned 12-lead ECG foundation model
Activity counts measure attention, not model quality. Missing counts are unknown. Publication status follows the linked source; a listed venue alone does not confirm peer review.
Fine-tunes an adult ECG foundation model on 339k pediatric 12-lead ECGs. Fine-tuning (mean AUC 0.9695) beats direct transfer (0.8256) and slightly beats training from scratch (0.9661), and remains strong with 1% of the pediatric data (0.9079). External mean AUC 0.9385 over 30 labels; age MAE 1.12 y; CHD AUC 0.9562.
1-D CNN components in net1d.py; adult ECG FM fine-tuned on pediatric 12-lead @ 500 Hz (notch 50 Hz, band-pass 0.5–75 Hz, per-lead Z-score)
Pediatric electrocardiogram (ECG) interpretation is complicated by age-dependent electrophysiological changes that limit direct transfer of adult AI models. We asked whether an adult ECG foundation model could be adapted to pediatric ECGs. Using 339,563 12-lead ECGs from 238,189 children, we compared direct application, pediatric fine-tuning, and training from scratch under an identical architecture. Direct transfer was poor (mean AUC 0.8256). Fine-tuning reached 0.9695, above de novo training (0.9661), and attained 0.9079 using only 1% of training data. On 10,999 external ECGs, fine-tuning achieved a mean AUC of 0.9385 across 30 labels. It generalized to age regression (MAE 1.12 years) and congenital heart disease classification (AUC 0.9562). Across 17 clinicians and 100 cases, diagnostic accuracy rose from 56.8% to 65.4% (p < 0.001) and did not decline when AI erred. Adult ECG foundation models require pediatric adaptation; fine-tuning offers a data-efficient route to scalable pediatric ECG AI.
Verbatim fromarXiv· arXiv metadata (abstracts are released under CC0)
| Task | Dataset | Metric | Value | Rank |
|---|---|---|---|---|
| Ecg diagnosis | zju-pediatric-ecg-external-10999 | AUROC ↑ | 0.9385 | #1 / 1 |
| Ecg diagnosis | zju-pediatric-ecg-internal | AUROC ↑ | 0.9695 | #1 / 1 |
Pediatric ECG foundation model via adult-to-pediatric fine-tuning (Cao, Xu, Li et al., arXiv 2611.07717). (2026). arXiv 2611.07717v1. https://arxiv.org/abs/2611.07717
@misc{pediatricecgfm2026pediatric,
title = {{Pediatric ECG foundation model via adult-to-pediatric fine-tuning (Cao, Xu, Li et al., arXiv 2611.07717)}},
year = {2026},
howpublished = {arXiv 2611.07717v1},
eprint = {2611.07717},
archivePrefix = {arXiv}
}Added to the ChatPPG research hub
Pediatric ECG foundation model via adult-to-pediatric fine-tuning (Cao, Xu, Li et al., arXiv 2611.07717)Paper on this modelarXiv 2611.07717v1 · 2026
Holtercare-Bench: A Multimodal Benchmark for Evaluating Long-Term Dynamic ECG Analysis (Xie et al., ACM MM 2026)Shared topic: no licenceACM Multimedia 2026 · 2026
