Groundbreaking AI-Method Finds a Way to People’s Hearts

Osaka Metropolitan University researchers have unveiled a groundbreaking AI application that accurately identifies heart functions and valvular heart disease from chest X-rays, as detailed in The Lancet Digital Health. By training the AI on over 22,000 chest X-rays and echocardiograms from multiple institutions, the team achieved high precision in classifying six types of valvular heart disease, with an Area Under the Curve (AUC) ranging from 0.83 to 0.92—a significant advancement. Dr. Daiju Ueda envisions this AI as a complement to echocardiography, potentially aiding diagnostics in various healthcare settings and emergency situations, benefiting patients without access to specialized care.