The AI algorithm, developed by software developer and cardiac monitoring company Eko, analyzes heart sounds data from non-invasive sensors with machine learning, a type of AI in which a computer learns over time, rather than having to be programmed like typical software.
To test the algorithm’s ability to detect heart murmurs, a team of researchers used an independent dataset of pediatric heart sounds. The research team then compared the AI’s findings to independent diagnoses from five pediatric cardiologists, who listened to the same heart sound recordings.
The researchers found Eko’s algorithm outperformed four out of the five cardiologists at detecting heart murmurs.
Eko is pursuing FDA clearance for the algorithm, which it plans to integrate into its existing cardiac monitoring devices. The company said the algorithm will help clinicians who are not specialized in cardiology, such as internal medicine and family practice physicians, screen for heart murmurs.
“Eko imagines a future where every physician is able to augment their clinical judgment and experience, leveraging thousands and thousands of previous medical cases in seconds,” said Eko CEO and co-founder Connor Landgraf.
Eko, which is based in San Francisco, is privately backed by investments from various organizations, including Rochester, Minn.-based Mayo Clinic. In October, Mayo Clinic and Eko entered into a collaboration to create and commercialize a tool that helps physicians detect patients with low ejection fraction, or a weak heart pump.
More articles on artificial intelligence:
Amazon adds 3 ‘HIPAA-eligible’ AI tools to cloud service
New AI model from UCSF can detect Alzheimer’s 6 years before diagnosis
Mayo Clinic partners with AI firm to screen patients for heart conditions
