AI can estimate heart attack and stroke risk from a single X-ray

With just one X-ray, artificial intelligence can calculate your 10-year risk of cardiovascular disease.

AI can tell from a single X-ray how likely you are to have a heart attack or stroke in the next 10 years.

With just one chest X-ray, an AI model can figure out how likely it is that you will die in 10 years from a heart attack or stroke.

Researchers taught the deep learning AI to look for patterns in X-rays that are linked to atherosclerosis, which is the most common cause of heart disease.

Current health guidelines in the US say that you should estimate your 10-year risk of major heart disease events so that you can take preventive steps, like taking statins, if you need to.

The score is based on things like age, sex, race, blood pressure, treatment for high blood pressure, smoking, Type 2 diabetes, and blood tests.

Patients with a 10-year risk of 7.5% or more are told to take statins.

The study's lead author, Dr. Jakob Weiss, a radiologist at the Cardiovascular Imaging Research Center at Massachusetts General Hospital and the AI in Medicine programme at the Brigham and Women's Hospital in Boston, said that the variables needed to calculate ASCVD risk are often not available. This makes approaches for population-based screening desirable.

"Our deep learning model could be used to find a way to screen a whole population for the risk of cardiovascular disease using chest X-rays."

"Because chest X-rays are easy to get, our method could help find people who are at high risk." This kind of screening could be used to find people who would benefit from taking statins but aren't getting them yet.

AI progress makes it possible.The researchers analysed 147,497 chest X-rays from 40,643 participants in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, a multi-center, randomised controlled trial designed and funded by the National Cancer Institute in the United States.

They used a second group of 11,430 outpatients who had chest X-rays and might be eligible for statin therapy to test the CXR-CVD risk model.

Over the average follow-up time of 10.3 years, 9.6% (1,096) of the 11,430 patients had a major cardiac event that went wrong.

They found that there was a "significant association" between the risk that the CXR-CVD risk model predicted and the major cardiac events that were actually seen.

Weiss said, "The great thing about this method is that all you need is an X-ray, which is taken millions of times a day all over the world."

"Based on a single existing chest X-ray image, our deep learning model predicts future major adverse cardiovascular events with the same accuracy and greater value than the established clinical standard."

He said that people have known for a long time that X-rays can show more than just the usual diagnostic information, but that the information hasn't been used because "we haven't had strong, reliable methods."

"AI progress has now made it possible," he said.

"We have shown that a chest X-ray is more than just that." With this method, we get a quantitative measure, which lets us give both the doctor and the patient useful information about diagnosis and prognosis.

Weiss said that more research, such as a controlled, randomised trial, is needed to confirm the deep learning model, which could eventually help doctors make decisions.

On Tuesday, the study's results were presented at the Radiological Society of North America's (RSNA) annual meeting.