Echocardiogram Foundation Model -- Application 1: Estimating Ejection Fraction
This addresses the problem of automating cardiac function assessment for healthcare, reducing interoperator variability and labor, but is incremental as it applies existing self-supervised learning to a new medical domain.
The paper tackled the laborious and variable task of quantifying cardiac function from echocardiograms by introducing EchoAI, a foundation model trained on 1.5 million echocardiograms, which achieved a mean absolute percentage error of 9.40% for estimating ejection fraction, matching expert sonographer performance.
Cardiovascular diseases stand as the primary global cause of mortality. Among the various imaging techniques available for visualising the heart and evaluating its function, echocardiograms emerge as the preferred choice due to their safety and low cost. Quantifying cardiac function based on echocardiograms is very laborious, time-consuming and subject to high interoperator variability. In this work, we introduce EchoAI, an echocardiogram foundation model, that is trained using self-supervised learning (SSL) on 1.5 million echocardiograms. We evaluate our approach by fine-tuning EchoAI to estimate the ejection fraction achieving a mean absolute percentage error of 9.40%. This level of accuracy aligns with the performance of expert sonographers.