Researchers have developed a deep learning model that can predict a person’s 10-year risk of death from a stroke or heart attack. It can factor that information from a single chest X-ray (CXR). The model is an advanced type of artificial intelligence (AI) that is trained to detect patterns associated with the disease through X-ray images.
How the Model Works
Jakob Weiss, M.D., the study’s lead author and a radiologist affiliated 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, stated, “Our deep learning model offers a potential solution” for those who need screening. This “opportunistic screening of cardiovascular disease risk using existing [CXR] images.”
“This type of screening could be used to identify individuals who would benefit from statin medication but are currently untreated,” Dr. Weiss added.
Current Ways to Predict
Currently, guidelines recommend doctors estimate a 10-year risk of major adverse cardiovascular disease events to establish who should get medication for primary prevention. They use the atherosclerotic cardiovascular disease (ASCVD) risk score to calculate a person’s risk.
The ASCVD is a statistical model that considers a host of variables, including systolic blood pressure, sex, hypertension treatment, age, Type 2 diabetes, smoking, and blood tests. Statin medication is recommended for patients with a 10-year risk of at least 7.5% or higher.
The information to appropriately “calculate ASCVD risk [that] are often not available, which makes approaches for population-based screening desirable,” Dr. Weiss revealed. Although CXRs are more commonly obtainable. “Our approach may help identify individuals at high risk,” said Dr. Weiss.
CXR-CVD Risk Study
The model known as CXR-CVD risk was developed by Dr. Weiss and a team of researchers. Together they trained a deep learning model using a single CXR as input. Researchers created the model to foresee the risk of death from cardiovascular disease. They used 147,497 chest X-rays from 40,643 participants in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. The trial is a multi-center, randomized controlled trial designed. The whole thing was sponsored by the National Cancer Institute.
They tested the model on a second set of test subjects’ CXRs. Researchers ran 11,430 outpatients which consisted of 42.9% males and an average of 60.1 years old. All of whom were potentially eligible for statin therapy at Mass General Brigham.
One-thousand-ninety-six (9.6%) of the 11,430 patients ended up suffering a major adverse cardiac event over the average follow-up of 10.3 years. There was a compelling association between the model’s predictions and observed major cardiac events.
The model’s usage of an X-ray is a unique approach. X-rays are “acquired millions of times a day across the world,” Dr. Weiss said. “Based on a single existing [CXR] image, our [CXR-CVD risk] model predicts future major adverse cardiovascular events with similar performance and incremental value to the established clinical standard.”
What This Means
A CXR “is more than a chest X-ray,” added Dr. Weiss. Co-authors of the study are Vineet Raghu, Ph.D., Kaavya Paruchuri, M.D., Pradeep Natarajan, M.D., M.M.S.C., Hugo Aerts, Ph.D., and Michael T. Lu, M.D., M.P.H. All the investigators were supported in part by funding from the National Academy of Medicine and the American Heart Association.
By Sheena Robertson
News Medical Life Sciences: Deep learning model uses a single chest X-ray to predict heart disease risk
Mayo Clinic: Statin side effects: Weigh the benefits and risks