Abstract
Electrocardiograms (ECGs) are a cornerstone of cardiac assessment; however, the clinical deployment of deep learning models is constrained by limited interpretability and the hallucination risk of large language models (LLMs). Existing CNN+Grad-CAM+multimodal LLM frameworks can generate ECG reports, but their explanations are often weakly grounded in established diagnostic criteria, reducing trustworthiness and reproducibility.
We propose a guide-grounded multimodal framework that explicitly anchors report generation in curated clinical knowledge. A convolutional neural network (CNN) and Grad-CAM produce class probabilities and class-specific heatmaps from 12-lead ECG images. In parallel, authoritative ECG textbooks and guideline materials are distilled offline into a structured ECG Interpretation Guide, which is injected as a fixed knowledge block for every sample.
Conditioned on the ECG image, Grad-CAM overlay, CNN-derived fact pack, and the injected guide, a multimodal LLM generates structured diagnostic reports with guideline-consistent terminology and criteria usage. Experiments on the full PTB-XL test set demonstrate that guide grounding improves semantic quality and perceived consistency of generated reports while preserving competitive classification performance. In particular, our method increases the average BERTScore of generated impressions from 0.818 to 0.953 relative to a strong CNN+Grad-CAM+MLLM baseline, indicating closer alignment with reference reports. These findings suggest that injecting a distilled interpretation guide into the multimodal prompting pipeline offers a practical pathway to reduce hallucinations and enhance the clinical plausibility of LLM-based ECG explanations, bringing explainable cardiac diagnosis closer to real-world deployment.
Blogger's Review: This paper presents an innovative guide-grounded multimodal framework that effectively enhances the interpretability and consistency of ECG reports by integrating CNN, Grad-CAM, and LLM. It not only improves report quality but also provides new insights for clinical applications, highlighting the potential of deep learning in the medical field.