Artificial Intelligence Can Improve Predictions for Lung Cancer and Cardiovascular Disease Risk

Artificial intelligence (AI) has the potential to revolutionize the medical field once again by effectively utilizing data from low-dose lung CT scans to improve predictions about a person’s risk for death from lung cancer, cardiovascular disease, and other conditions. Currently, annual lung screenings with low-dose chest CT scans are recommended for individuals between the ages of 50 and 80 who are at high risk for lung cancer, such as long-time smokers. These scans not only provide images of the lungs but also offer valuable information about other structures within the chest.

A study led by Kaiwen Xu, a Ph.D. candidate in the Department of Computer Science at Vanderbilt University, demonstrates that AI algorithms can automatically derive body composition measurements from lung screening CT scans. These measurements capture the percentage of fat, bone, and muscle in the body. Abnormal body composition, such as obesity or loss of muscle mass, is associated with chronic health conditions like heart disease. Body composition is known to be a useful indicator for identifying risks and prognosis for cardiovascular disease, chronic obstructive pulmonary disease (COPD), and even survival and quality of life in lung cancer therapy.

During the study, the research team analyzed CT scans of over 20,000 individuals from the National Lung Screening Trial and found that incorporating AI body composition measurements improved predictions for lung cancer, heart disease, and all-cause mortality risk. The measurements related to fat found in the muscle were specifically strong predictors of mortality.

By utilizing AI to assess body composition from lung screening CT scans, healthcare professionals can identify high-risk individuals for interventions like physical conditioning or lifestyle modifications, even at an early stage before the onset of disease. This reveals that CT scans, initially ordered for lung cancer detection, contain a wealth of additional information that can be used to evaluate body composition or assess coronary artery calcification associated with cardiovascular disease risk.

The findings of this study highlight the potential for AI to enhance the value of lung screening CT scans and provide a more comprehensive understanding of a patient’s health by gathering information about various conditions beyond lung cancer. Integrating AI analysis of body composition into routine clinical care could be an essential step in improving risk assessment and guiding personalized interventions.