AI Detects Heart Defects with Just a Smartwatch

For the first time, Yale researchers have been able to diagnose structural heart disease using the simple electrocardiogram (ECG) sensors found in a smartwatch. Their findings will be presented this week at the American Heart Association’s Scientific Sessions 2025. 

While wearable devices are increasingly used to diagnose and monitor heart diseases that affect the heartbeat, such as arrhythmia or atrial fibrillation, their single-lead ECG design has so far limited their ability to diagnose physical defects in the heart. However, with the help of artificial intelligence (AI) algorithms, a group of 600 patients in a real-world hospital setting could be diagnosed with high accuracy from just a 30-second smartwatch read. 

“On its own, a single-lead ECG is limited; it can’t replace a 12-lead ECG test available in health care settings. However, with AI, it becomes powerful enough to screen for important heart conditions,” said Rohan Khera, MD, MS, director of the cardiovascular data science lab at Yale School of Medicine and senior author of the study. “This could make early screening for structural heart disease possible on a large scale, using devices many people already own.”

The AI algorithm was developed using more than 266,000 full ECGs taken from 110,000 patients at Yale New Haven Hospital between 2015 and 2023. Khera and colleagues isolated one of the 12 leads which most resembles the single lead found in smartwatch sensors and used that data to predict the presence of three types of structural heart disease: low left ventricular ejection fraction, severe left-sided valvular disease, and severe left ventricular hypertrophy. 

The model was first validated in over 44,000 adults across four community hospitals as well as 3,000 participants in a population-based study in Brazil. Then, in a prospective study, 600 patients undergoing an echocardiogram were recruited to take a 30-second ECG measurement using a smartwatch before the procedure.

In this patient cohort, the algorithm achieved 86% sensitivity and 87% specificity, with an overall performance of 88% using single-lead reads compared to 92% when using 12-lead hospital ECG equipment. This was achieved by introducing noise in the data used to train the model, which helped the AI become more reliable when analyzing ECG signals taken in a real-world context. 

“We explored whether the same smartwatches people wear every day could also help find these hidden structural heart diseases earlier, before they progress to serious complications or cardiac events,” said Arya Aminorroaya, MD, MPH, internal medicine resident at Yale New Haven Hospital and research affiliate at Yale School of Medicine. “Millions of people wear smartwatches, and they are currently mainly used to detect heart rhythm problems such as atrial fibrillation. Structural heart diseases, on the other hand, are usually found with an echocardiogram, an advanced ultrasound imaging test of the heart that requires special equipment and isn’t widely available for routine screening.” 

While promising, these results will have to be validated in a larger population. In the current prospective study, only five percent of patients were confirmed to have structural heart disease during the ultrasound procedure: 15 had low left ventricular ejection fraction, five were diagnosed with severe left-sided valvular disease, and just one patient had severe left ventricular hypertrophy. 

“We 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,” said Aminorroaya.