A novel artificial intelligence (AI) model may be able to accurately predict the risk of death from lung cancer, cardiovascular disease, and other causes by using data from low-dose computed tomography (CT) scans of the lungs, according to a recent study published by Xu et al in Radiology.
Background
The U.S Preventive Services Task Force currently recommends annual lung screenings with low-dose CT scans of the chest for patients aged 50 to 80 at high risk of lung cancer, such as those with a long history of smoking. Along with images of the lungs, the scans also provide information about other structures within the chest.
Abnormal body composition such as obesity and loss of muscle mass has been linked to chronic health conditions like metabolic disorders. Previous studies have shown that body composition is a useful measurement in risk stratification and prognosis for cardiovascular disease and chronic obstructive pulmonary disease. In lung cancer therapy, body composition has been shown to affect survival and quality of life.
“When we’re looking at the CT images, the primary focus is on identifying nodules suspicious for lung cancer, but there is much more anatomical information coded in the space—including information on body composition,” explained lead study author Kaiwen Xu, MS, a PhD candidate in the Department of Computer Science at Vanderbilt University.
Mr. Xu and his colleagues previously developed, tested, and publicly released an AI algorithm designed to automatically derive body composition measurements from low-dose CT scans of the lungs.
Study Methods and Results
In the recent study, the researchers examined the CT scans of more than 200,000 patients who participated in the National Lung Screening Trial to determine the potential benefits of the AI-derived body composition measurements. The researchers found that including these measurements in the patients’ assessments improved the risk prediction for lung cancer mortality, cardiovascular disease mortality, and all-cause mortality.
Additionally, the measurements associated with myosteatosis—now thought to be more predictive of health outcomes than reduced muscle bulk—were particularly strong predictors of mortality, a finding consistent with previous studies, according to the investigators.
Conclusions
“The images in a CT [scan] ordered for quite a different purpose—in our case, early detection of lung cancer—contain much more information,” Mr. Xu suggested. “In the space of the chest CT [scan] used for lung cancer screenings, you can also check other information like body composition or coronary artery calcification that is directly associated with cardiovascular disease risk,” he noted.
The body composition measurements from low-dose CT scans of the lungs may provide greater opportunistic screenings, where the images obtained for one purpose can offer information about other conditions. The new AI model may have the potential to aid physicians in routine clinical settings.
“Automatic AI body composition potentially extends the value of lung screenings with low-dose CT beyond the early detection of lung cancer,” Mr. Xu highlighted. “It can help us identify high-risk [patients] for interventions like physical conditioning or lifestyle modifications, even at a very early stage before the onset of disease,” he emphasized.
The researchers concluded that they hope to further evaluate the potential benefits of their AI model in future studies with longer follow-up periods to better understand how changes in body composition may relate to health outcomes.
Disclosure: For full disclosures of the study authors, visit pubs.rsna.org.
