AI Tool Detects Structural Heart Disease Using Smartwatch ECG Sensors

An artificial intelligence algorithm paired with single-lead electrocardiogram sensors on smartwatches accurately diagnosed structural heart diseases including weakened pumping ability, damaged valves, and thickened heart muscle, according to preliminary research to be presented at the American Heart Association's Scientific Sessions 2025. This represents the first prospective study demonstrating that AI can detect multiple structural heart diseases using measurements from the single-lead ECG sensor on the back and digital crown of consumer smartwatches.

Millions of people already wear smartwatches that are primarily used to detect heart rhythm problems such as atrial fibrillation. Structural heart diseases, by contrast, are typically identified through echocardiograms, which are advanced ultrasound imaging tests requiring specialized equipment not widely available for routine screening. The study explored whether everyday smartwatches could help detect these hidden structural heart conditions earlier, before they progress to serious complications or cardiac events.

Researchers developed the AI algorithm using more than 266,000 12-lead ECG recordings from over 110,000 adults. Based on this extensive data library, they created an algorithm capable of identifying structural heart disease from single-lead ECGs similar to those obtained from smartwatch sensors. The team isolated only one of the 12 leads from traditional ECGs to simulate the single-lead format available on consumer devices. They also incorporated random interference or noise into the training data to make the AI model more resilient and reliable when dealing with real-world smartwatch signals.

The AI model underwent external validation using data from patients at community hospitals and participants from the population-based ELSA-Brasil study. The Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) gathers important information about chronic disease development and progression, with particular focus on cardiovascular diseases and diabetes. Following validation, researchers prospectively recruited 600 participants who underwent 30-second, single-lead ECGs using smartwatches on the same day they received heart ultrasounds to assess the algorithm's real-world accuracy.

The analysis revealed compelling results. When using single-lead ECGs obtained from hospital equipment, the AI model demonstrated 92% effectiveness at distinguishing individuals with and without structural heart disease. Among the 600 participants who provided single-lead ECGs from smartwatches, the AI model maintained high performance at 88% for detecting structural heart disease. The algorithm accurately identified most people with heart disease, achieving 86% sensitivity, and was highly accurate in ruling out heart disease with 99% negative predictive value.

Study author Arya Aminorroaya, M.D., M.P.H., an internal medicine resident at Yale New Haven Hospital and research affiliate at the Cardiovascular Data Science (CarDS) Lab at Yale School of Medicine, emphasized the potential impact of this technology. The median age of study participants was 62 years, with approximately half being women and diverse racial and ethnic representation including 44% non-Hispanic white, 15% non-Hispanic Black, 7% Hispanic, 1% Asian, and 33% others. About 5% of participants were found to have structural heart disease on heart ultrasound.

Rohan Khera, M.D., M.S., the study's senior author and director of the CarDS Lab, noted that while a single-lead ECG alone has limitations and cannot replace the comprehensive 12-lead ECG tests available in healthcare settings, the addition of AI makes it powerful enough to screen for important heart conditions. This advancement could enable early screening for structural heart disease on a large scale using devices many people already own and use daily.

The researchers plan to evaluate the AI tool in broader settings and explore how it could be integrated into community-based heart disease screening programs to assess its potential impact on improving preventive care. Study limitations include the small number of patients with actual disease in the prospective study and the number of false positive results. The findings are considered preliminary until published as full manuscripts in peer-reviewed scientific journals, as abstracts presented at American Heart Association scientific meetings are not peer-reviewed.