PEACE: Cross-modal Enhanced Pediatric-Adult ECG Alignment for Robust Pediatric Diagnosis

2026-05-01

AI Summary

This work addresses the challenges of domain shift and scarce annotations in pediatric ECG (pECG) automated diagnosis, which arise from the dominance of adult data in training. The authors propose PEACE, a novel framework that introduces structured clinical semantic supervision and cross-modal alignment for the first time. PEACE decomposes clinical semantics along three axes, extracts label-query features, and employs curriculum-gated optimization, leveraging large language model–generated (Gemini) semantic descriptions as auxiliary supervision during training while requiring only raw ECG signals at inference. Experiments demonstrate that PEACE achieves AUCs of 59.39%, 79.03%, and 90.89% under zero-shot, 50-sample fine-tuning, and full fine-tuning settings on the ZZU-pECG dataset, respectively, and attains a 96.65% AUC on PTB-XL within a shared label space, substantially mitigating cross-population transfer difficulties in low-resource scenarios.

Abstract

Automated pediatric electrocardiogram (ECG) diagnosis remains challenging because models trained predominantly on adult data suffer from substantial cross-population mismatch, while pediatric labels are often scarce. We present PEACE (Pediatric-Adult ECG Alignment via Cross-modal Enhancement), a structured cross-modal alignment framework for adult-to-pediatric ECG transfer. PEACE integrates tri-axial clinical semantic decomposition, label-query feature extraction, and curriculum-gated optimization to align transferable adult ECG representations with pediatric diagnostic targets. Since ZZU-pECG provides no paired clinical reports, we generate label-conditioned semantic descriptors using Gemini with concise clinical prompts and use them only as auxiliary training supervision; inference remains ECG-only. On ZZU-pECG, PEACE achieves 59.39%, 79.03%, and 90.89% AUC under zero-shot, 50-shot, and full fine-tuning settings, respectively, and reaches 96.65% AUC on the shared PTB-XL label space. These results suggest that structured clinical semantic supervision can improve low-resource adult-to-pediatric ECG transfer, while prospective clinical validation and more explicit age-aware modeling remain necessary before real-world deployment.

Problem

• pediatric ECG diagnosis
• cross-population mismatch
• data scarcity
• adult-to-pediatric transfer
• automated diagnosis

Innovation

• cross-modal alignment
• clinical semantic decomposition
• label-query feature extraction
• curriculum-gated optimization
• adult-to-pediatric transfer