A 10-second heart test may reveal long-term risk
Researchers in the United States have developed and validated an artificial intelligence model that can estimate a person’s risk of stroke up to 10 years in advance using a single 10-second electrocardiogram.
The research team was co-led by investigators at Mass General Brigham in Boston and the Broad Institute of Massachusetts Institute of Technology and Harvard University in Cambridge, Massachusetts.
How ECG2Stroke works
The model, called ECG2Stroke, analyzes information from an electrocardiogram, or ECG. This low-cost and non-invasive test records the heart’s electrical activity through electrode sensors placed on the skin.
Rather than relying on lengthy clinical scoring systems, the model uses ECG data along with a patient’s age and sex to identify patterns linked to future stroke risk.
Researchers sought a more scalable tool
Dr Rahul Mahajan, co-lead author, neurologist with the Mass General Brigham Neuroscience Institute and member of the Broad Cardiovascular Disease Initiative, said current stroke-risk tools can be difficult to use routinely.
“Existing tools to identify which patients are at the highest risk of stroke often require cumbersome clinical score calculations, are not easily scalable, and are therefore not used widely in routine practice,” said Dr Rahul Mahajan.
Deep learning detects hidden ECG signals
Mahajan and his colleagues built the model using patient data from Massachusetts General Hospital. The system relies on deep learning, a type of artificial intelligence designed to detect complex patterns across large datasets.
In this case, the model was trained to find subtle ECG waveform features that may not stand out during standard clinical review but could still contain information about long-term cardiovascular and cerebrovascular risk.
Model tested across Boston hospitals
The investigators validated ECG2Stroke using patients from Brigham and Women’s Hospital and Beth Israel Deaconess Medical Center, both in Boston.
Overall, the team used information from more than 200,000 patients to train and validate the model. Its performance was assessed across different hospitals and patient subgroups.
Performance matched a clinical risk score
The results showed that ECG2Stroke could consistently predict stroke risk up to 10 years into the future. Its performance was reported to be similar to that of a validated clinical risk score.
The advantage is that ECG2Stroke relies on a test that is already common in routine medical care, making it potentially easier to use at scale.
Atrial signals played an important role
The researchers found that signals related to dysfunction in the atria had a strong influence on the model’s predictions. The atria are the upper chambers of the heart that receive blood returning from the body.
This suggests that small electrical changes in the heart may reveal hidden abnormalities connected to stroke risk before symptoms appear.
Strong results for cardioembolic stroke
ECG2Stroke was especially accurate at predicting cardioembolic stroke. This type of stroke happens when a clot forms in the heart, breaks away, travels to the brain and blocks blood flow.
That finding is clinically important because some cardioembolic strokes may be preventable with anticoagulant medicines in selected patients.
Future studies will be needed
Dr Shaan Khurshid, co-senior author, cardiologist with the Mass General Brigham Heart and Vascular Institute and member of the Broad Cardiovascular Disease Initiative, said the model could help guide prevention if future studies confirm its value.
“If confirmed after prospective, real-world studies, tools like this could identify which patients should be prioritized for intensive prevention efforts,” said Dr Shaan Khurshid.
“The tool could also be helpful in driving future mechanistic research into abnormalities in the upper chambers of the heart and links to stroke,” Khurshid concluded.
