Stroke: Researchers develop AI model to predict a person’s 10-year risk
Recently, researchers from Massachusetts General Hospital developed a deep learning model using artificial intelligence (AI) and a single chest X-ray to help predict a person’s 10-year risk of dying from a stroke or heart attack.
Each year, about 15 million people worldwide have a #stroke, causing about 5 million deaths and leaving another 5 million permanently disabled.
A person’s outcome following a stroke depends on how quickly they receive treatment once symptoms begin.
The results of the study were presented in late November at the annual meeting of the Radiological Society of North America (RSNA) in Chicago.
Developing a predictive model for stroke
The team used a CXR-CVD system that was “trained” to search more than 147,000 chest X-ray images from almost 41,000 participants in a cancer screening trial and spot patterns associated with cardiovascular disease.
Once developed, the system could predict a person’s 10-year risk of having a stroke or heart attack from a single chest X-ray.
Lead study author Jakob Weiss, Ph.D., a radiologist with the Cardiovascular Imaging Research Center at Massachusetts General Hospital and the AI in Medicine program at the Brigham and Women’s Hospital in Boston, explained to MNT:
“Current guidelines of the American College of Cardiology and American Heart Association on the primary prevention of cardiovascular disease recommend the use of a risk calculator to estimate the risk of future CVD. This risk calculator is based on the ASCVD risk score a multivariable regression model requiring nine variables as input, such as age, sex, smoking, lipids, blood pressure, and diabetes. However, these variables are often not available, which makes novel and more practical screening approaches desirable.”
Dr. Weiss added that the new AI model takes a routine chest X-ray as the only input, which “allows for opportunistic screening of CVD risk if the variables to calculate the guideline-recommended risk score are missing.”
Next steps in AI research for stroke detection
Although compelling, the new research is preliminary, and more long-term studies are still needed.
Dr. Jennifer Wong, a cardiologist and medical director of noninvasive cardiology at MemorialCare Heart and Vascular Institute at Orange Coast Medical Center in Fountain Valley, CA, who was not involved in the study, said that future longitudinal studies could examine outcomes for people with different levels of risk as identified on a chest X-ray.
“Identifying people with higher risk for cardiovascular disease early is important, and more tools to help identify those at risk is always useful,” Dr. Wong told MNT.
“This study did a great job of correlating with the current tool that cardiologists use, the ASCVD risk score, but to see based on where they risk-stratify somebody what happens to that person’s risk of heart attack and stroke 10 (or) 20 years down the road with just chest X-ray as the tool might be helpful,” she added.
Source : Medical News Today
