Scientists have developed an artificial intelligence system that may be able to predict a person’s risk of stroke up to 10 years before it happens using a simple heart test that takes only about 10 seconds.
The new tool, called ECG2Stroke, was created by researchers from Mass General Brigham and the Broad Institute of MIT and Harvard. Their findings were published in the Journal of the American College of Cardiology (JACC).
Stroke is one of the leading causes of death and long-term disability worldwide. It happens when blood flow to part of the brain is blocked or when a blood vessel bursts. Finding people at high risk before a stroke occurs gives doctors an opportunity to prescribe treatments and encourage lifestyle changes that may prevent devastating brain damage.
Doctors already have ways to estimate stroke risk, but many of these methods require several medical measurements and complex calculations. Because they can be time-consuming, they are not always used during routine healthcare visits. The research team wanted to create a faster and more practical approach.
The scientists turned to the electrocardiogram, or ECG, one of the most common medical tests. During an ECG, small sensors placed on the skin record the electrical activity of the heart. The test is inexpensive, painless, non-invasive, and is already performed millions of times every year around the world.
Using information from more than 200,000 patients treated at Massachusetts General Hospital, Brigham and Women’s Hospital, and Beth Israel Deaconess Medical Center, the researchers trained a deep learning computer model to recognize tiny patterns in ECG recordings that humans cannot easily detect. The model also included each patient’s age and sex.
The results were impressive. ECG2Stroke predicted future stroke risk with an accuracy similar to established clinical risk scores, even though it relied mainly on a single 10-second ECG recording. The system performed consistently across different hospitals and patient groups, suggesting it may be widely applicable.
The model was especially good at identifying cardioembolic strokes. These strokes occur when a blood clot forms inside the heart, travels through the bloodstream, and blocks an artery in the brain. Because blood-thinning medicines can often prevent these strokes, identifying high-risk patients early could have major health benefits.
The researchers found that changes related to the heart’s upper chambers, known as the atria, played an important role in the AI model’s predictions. This finding may help scientists better understand how hidden abnormalities in the heart contribute to stroke risk years before symptoms appear.
Although the results are encouraging, the researchers emphasize that more studies are needed before the tool is used routinely in hospitals and clinics. Future real-world research will determine whether using ECG2Stroke actually leads to fewer strokes by helping doctors begin preventive treatment earlier.
In review, this study highlights the growing role of artificial intelligence in preventive medicine. A quick, inexpensive ECG combined with advanced AI could one day help doctors identify people at high risk long before a stroke occurs. If future studies confirm these findings, ECG2Stroke could become an important tool for saving lives and reducing disability.
If you care about stroke, please read studies about how to eat to prevent stroke, and diets high in flavonoids could help reduce stroke risk.
For more health information, please see recent studies about how Mediterranean diet could protect your brain health, and wild blueberries can benefit your heart and brain.
Source: Mass General Brigham.
