A Multimodal and Explainable Machine Learning Approach to Diagnosing Multi-Class Ejection Fraction from Electrocardiograms

TL;DR

This study presents a multimodal machine learning framework combining ECG features and EHR data to classify left ventricular ejection fraction into four categories, aiding diagnosis in resource-limited settings.

Contribution

It introduces a novel, explainable multimodal model that outperforms single-modality models in classifying LVEF strata using retrospective clinical data.

Findings

Achieved high AUROCs for all LVEF categories, up to 0.95.

Outperformed ECG-only and EHR-only models in classification accuracy.

Maintained performance under temporal validation, demonstrating robustness.

Abstract

Left ventricular ejection fraction (LVEF) assessment depends on echocardiography, limiting access in primary care and resource-constrained settings. We developed a multimodal machine-learning framework that combines engineered 12-lead ECG timeseries features with structured EHR variables to classify LVEF into four clinically used strata: normal (>50%), mildly reduced (40-50%), moderately reduced (30-40%), and severely reduced (<30%). To support model explainability, we identified the most influential ECG and EHR features via SHAP attributions. Using retrospective data from Hartford HealthCare, we trained XGBoost models on 36,784 ECG-echocardiogram pairs from 30,952 outpatients and evaluated temporal generalizability on 19,966 ECGs from a subsequent period. The multimodal model achieved one-vs-rest AUROCs of 0.95 (severe), 0.92 (moderate), 0.82 (mild), and 0.91 (normal), outperforming ECG-only and EHR-only baselines, and maintained performance under temporal validation. This work supports ECG-based, multimodal LVEF stratification as a practical screening and triage aid to prioritize confirmatory imaging where resources are limited.