AI Uses Lung CT Data to Predict Risk of Death from Cancer and Cardiovascular Disease

S Preventive Services Task Force advises yearly lung screening with low-dose CT (LDCT) for individuals aged 50 to 80 years at high risk of lung cancer, such as long-term smokers. These scans, while focused on the lungs, also offer information about other chest structures. Now, a new study has revealed that artificial intelligence (AI) can harness data from these low-dose CT scans of the lungs to improve risk predictions for death from lung cancer, cardiovascular disease, and other causes.

Researchers at Vanderbilt University (Nashville, TN, USA) had earlier developed, tested, and publicly released an AI algorithm that automatically extracts body composition measurements from LDCT scans used in lung screening. Body composition refers to the percentage of fat, muscle, and bone in the body. Abnormal body composition, like obesity or muscle mass loss, is associated with chronic health conditions including metabolic disorders. Prior research has shown that body composition is valuable for risk stratification and prognosis in cardiovascular disease and chronic obstructive pulmonary disease. In lung cancer therapy, body composition has been shown to influence survival and quality of life.

In the new study, the researchers evaluated the added value of AI-derived body composition measurements by examining CT scans of over 20,000 individuals from the National Lung Screening Trial. Their findings showed that incorporating these measurements improved risk prediction for death from lung cancer, cardiovascular disease, and all-cause mortality. Measurements associated with fat within muscle were particularly strong predictors of mortality, which is in line with existing research. The infiltration of skeletal muscle with fat, a condition known as myosteatosis, is now considered more predictive for health outcomes than reduced muscle bulk.

The use of body composition measurements from lung screening LDCT serves as an example of opportunistic screening, where imaging intended for one purpose provides information about other conditions. This practice is considered highly promising for routine clinical use. This study assessed individuals at a baseline screening only. For future research, the scientists aim to conduct a longitudinal study, tracking individuals over time to observe how changes in body composition relate to health outcomes.

"Automatic AI body composition potentially extends the value of lung screening with low-dose CT beyond the early detection of lung cancer," said study lead author Kaiwen Xu, a Ph.D. candidate in the Department of Computer Science at Vanderbilt University. "It can help us identify high-risk individuals for interventions like physical conditioning or lifestyle modifications, even at a very early stage before the onset of disease."

Related Links:
Vanderbilt University