Researchers at Rochester, Minn.-based Mayo Clinic have developed an AI model that estimates left ventricular ejection fraction from a single echocardiographic frame.
Unlike models that rely on video, the Mayo tool uses deep learning and computer vision to assess cardiac efficiency from static 2D ultrasound images, according to a Jan. 20 news release from the health system. The approach could streamline cardiac assessments in point-of-care settings, particularly in which time, resources or expertise are limited.
The model was validated across multiple Mayo sites using standard transthoracic echocardiograms, as well as two cohorts using handheld cardiac ultrasound devices, according to a study published in April in The Lancet Digital Health. The model demonstrated strong diagnostic accuracy across all test groups, suggesting the AI tool could support reliable cardiac assessments in a wide range of clinical settings — including emergency departments, rural clinics and bedside care — without requiring high-end equipment or expert users.
Mayo researchers said in the release the technology may reduce dependence on operators, accelerate results and pave the way for similar AI applications that extract dynamic clinical metrics from static images.
