Smartphone-enabled AI tool can detect widely underdiagnosed heart condition

A new AI platform developed by Yale University can diagnose a rarely detected heart condition using a simple heart test accessed via a smartphone. Called amyloid cardiomyopathy, this disease arises from misshapen proteins collecting on the heart muscle and can be fatal if left untreated. Symptoms often mimic those of other heart issues, leading to delayed diagnoses.

To detect the condition, researchers utilized electrocardiogram (ECG) images captured through the noninvasive test. The AI platform, described in JAMA, screens for the disease by analyzing patterns in ECGs, helping doctors identify at-risk patients sooner to prevent extensive heart damage. Lead investigator Dr. Rohan Khera emphasized the importance of early detection, stating that the condition progresses aggressively, akin to some cancers, and has a five-year average life expectancy without intervention.

The AI tool, which can be accessed through a smartphone, was trained on data from thousands of patients and refined using a few hundred diagnosed cases. In tests across eight U.S. and European patient groups, the platform accurately identified individuals with transthyretin amyloid cardiomyopathy, a subtype caused by misfolded proteins in the liver.

This breakthrough, backed by the FDA's Breakthrough Devices Program, aims to expedite the deployment of lifesaving therapies by identifying more at-risk individuals.