Validating the Clinical Utility of CineECG 3D Reconstructions through Cross-Modal Feature Attribution

arXiv cs.LG / 5/1/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper addresses a key gap in deep-learning ECG systems: high diagnostic accuracy is often undermined by limited clinical interpretability.
It proposes a cross-modal feature-attribution mapping method that transfers explanations from a 12-lead ECG model into a CineECG 3D anatomical space.
Experiments show that directly training on CineECG signals can reduce accuracy and produce incoherent attributions, while the proposed mapping restores clinically meaningful feature rankings.
On a ground-truth dataset of 20 expert-annotated cases, the mapped explanations achieve a Dice score of 0.56 versus 0.47 for standard 12-lead attributions, indicating better localization of pathological features.
Overall, the approach combines the diagnostic power of conventional ECG analysis with the intuitive clarity of anatomical visualization, potentially improving clinical integration prospects.

Abstract

Deep learning models for 12-lead electrocardiogram (ECG) analysis achieve high diagnostic performance but lack the intuitive interpretability required for clinical integration. Standard feature attribution methods are limited by the inherent difficulty in mapping abstract waveform fluctuations to physical anatomical pathologies. To resolve this, we propose a cross-modal method that projects feature attributions from high-performance 12-lead ECG models onto the CineECG 3D anatomical space. Our study reveals that while models trained directly on CineECG signals suffer from reduced accuracy and incoherent attributions, the proposed mapping mechanism effectively recovers clinically relevant feature rankings. Validated against a ground-truth dataset of 20 cases annotated by domain experts, the mapped explanations yield a Dice score of 0.56, significantly outperforming the 0.47 baseline of standard 12-lead attributions. These findings indicate that cross-modal averaging mapping effectively filters attribution instability and improves the localization of pathological features, combining the diagnostic expressiveness of standard ECG with the intuitive clarity of anatomical visualization.