Knowledge-enhanced Multimodal ECG Representation Learning with Arbitrary-Lead Inputs

TL;DR

K-MERL is a novel framework that enhances multimodal ECG representation learning by integrating structured knowledge from reports and accommodating arbitrary lead inputs, improving performance in zero-shot classification tasks.

Contribution

The paper introduces K-MERL, a knowledge-enhanced ECG learning framework that handles arbitrary leads and improves zero-shot classification accuracy.

Findings

Achieves state-of-the-art zero-shot performance on six datasets.

Provides an average 16% AUC improvement in partial-lead classification.

Effectively leverages large language models for structured report knowledge.

Abstract

Recent advances in multimodal ECG representation learning center on aligning ECG signals with paired free-text reports. However, suboptimal alignment persists due to the complexity of medical language and the reliance on a full 12-lead setup, which is often unavailable in under-resourced settings. To tackle these issues, we propose **K-MERL**, a knowledge-enhanced multimodal ECG representation learning framework. **K-MERL** leverages large language models to extract structured knowledge from free-text reports and employs a lead-aware ECG encoder with dynamic lead masking to accommodate arbitrary lead inputs. Evaluations on six external ECG datasets show that **K-MERL** achieves state-of-the-art performance in zero-shot classification and linear probing tasks, while delivering an average **16%** AUC improvement over existing methods in partial-lead zero-shot classification.