Multimodal Foundation Models For Echocardiogram Interpretation

TL;DR

This paper introduces EchoCLIP, a multimodal foundation model trained on extensive cardiac ultrasound data, demonstrating strong zero-shot performance in cardiac assessment and report generation, advancing automated echocardiogram interpretation.

Contribution

Developed EchoCLIP, a novel multimodal foundation model for echocardiography, leveraging large-scale cardiac ultrasound videos and expert interpretations to enable zero-shot diagnostic capabilities.

Findings

Achieved a mean absolute error of 7.1% in cardiac function assessment.

AUC of 0.84-0.98 for identifying intracardiac devices.

High accuracy in patient identification and clinical change detection.

Abstract

Multimodal deep learning foundation models can learn the relationship between images and text. In the context of medical imaging, mapping images to language concepts reflects the clinical task of diagnostic image interpretation, however current general-purpose foundation models do not perform well in this context because their training corpus have limited medical text and images. To address this challenge and account for the range of cardiac physiology, we leverage 1,032,975 cardiac ultrasound videos and corresponding expert interpretations to develop EchoCLIP, a multimodal foundation model for echocardiography. EchoCLIP displays strong zero-shot (not explicitly trained) performance in cardiac function assessment (external validation left ventricular ejection fraction mean absolute error (MAE) of 7.1%) and identification of implanted intracardiac devices (areas under the curve (AUC) between 0.84 and 0.98 for pacemakers and artificial heart valves). We also developed a long-context variant (EchoCLIP-R) with a custom echocardiography report text tokenizer which can accurately identify unique patients across multiple videos (AUC of 0.86), identify clinical changes such as orthotopic heart transplants (AUC of 0.79) or cardiac surgery (AUC 0.77), and enable robust image-to-text search (mean cross-modal retrieval rank in the top 1% of candidate text reports). These emergent capabilities can be used for preliminary assessment and summarization of echocardiographic findings.