Around 15 million people across the globe suffer from a heart stroke every year, which accounts for about 5 million permanently disabled and nearly 5 million deaths worldwide. The severity of a heart attack mainly depends on the kind of prompt treatment the patient receives once the symptoms start developing.
Heart attack and stroke risk factors include:
- Hypertension
- Genetics or family history
- Diabetes
- Age
- Gender
- Pre-existing heart disease
Common signs of heart attack and stroke include:
- Dizziness
- Vision issues
- Consistent and severe headache
- Numbness
- Inability to respond, speak, and/or understand
With an aim to detect a stroke early and initiate prompt cardiovascular treatment, researchers from Massachusetts General Hospital, Women’s Hospital, and the Brigham used Artificial Intelligence to develop a deep-learning model that utilizes a single chest X-ray to predict a patient’s 10-year risk of death from a stroke or heart attack due to atherosclerotic cardiovascular disease (the buildup of cholesterol, fats and other substances inside the artery walls).
Researchers from Massachusetts General Hospital, Women’s Hospital, and Brigham recently presented the latest research findings at the Radiological Society of North America (RSNA) in Chicago. The study is still in its final draft stages, and the final results have yet to be published in any journal.
Cardiovascular or heart disease risk is a significant population chronic and health disease concern. The press release further states the current guidelines, which recommend evaluating the 10-year risk of principal adverse cardiovascular disease events to ascertain who should get statin drug therapy for primary prevention.
Researchers used approximately 150,000 chest X-rays to train a deep learning and artificial intelligence model to identify patterns in the images linked to risk factors stemming from major cardiovascular disease events. The model was tested on a different group of about 11,000 people and discovered a “significant connection” between the risk levels predicted by the deep learning AI model and the actual event of a significant cardiovascular disease occurrence.
Researchers used the atherosclerotic cardiovascular disease (ASCVD) risk score. This statistical model evaluates multiple patient variables, such as age, sex, systolic blood pressure, smoking, hypertension treatment, blood tests, and type 2 diabetes, to calculate the risk factors. For patients with a 10-year risk percentage of 7.5% or higher, statin drug therapy is recommended.
Dr. Jakob Weiss, the lead researcher, said, “The variables necessary to calculate ASCVD risk are often not available, which makes approaches for population-based screening desirable. As chest X-rays are commonly available, our approach may help identify high-risk individuals.” He added, “We’ve long acknowledged that X-rays capture information well beyond traditional diagnostic findings, but we have yet to use this data because we haven’t had robust, reliable methods. However, “Advances in AI are making it possible now.”
The beauty of this technology is that you only need one X-ray, which is generally acquired millions of times a day worldwide, said Dr. Jakob. This research is the latest in the efforts to leverage deep learning and AI to predict cardiovascular disease risk.
The researchers finally concluded that further research is necessary to validate the deep learning model. However, these findings indicate AI’s potential as an optimum clinical decision-support tool.
