Abstract 4366745: Artificial Intelligence-Derived Myosteatosis on Coronary Artery Calcium CT Scans Predicts Incident COPD: An AI-CVD Study within the Multi-Ethnic Study of Atherosclerosis (MESA)

Introduction/Background: Myosteatosis, defined as pathological fat infiltration into skeletal muscle, is an emerging marker of metabolic dysfunction and cardiovascular risk, particularly when measured in abdominal CT. However, its association with lung health and risk of chronic obstructive pulmonary disease (COPD) is not well established. The AI-CVD initiative aims to extract all useful opportunistic screening information from coronary artery calcium scans and combines them with traditional risk factors to create a stronger predictor of cardiovascular diseases. We hypothesized that myosteatosis measured from cardiac CT scans using AI-CVD is associated with increased risk of incident COPD in a population free of clinical cardiovascular disease at baseline. Methods/Approach: A retrospective cohort analysis was conducted using baseline data from Exam 1 of the Multi-Ethnic Study of Atherosclerosis including men and women aged 45 to 84 free of cardiovascular disease at baseline. Myosteatosis was quantified using AI-CVD to segment muscle and compute thoracic skeletal muscle density as a proxy for fat infiltration. Chronic obstructive pulmonary disease (COPD) was defined using clinical diagnosis with ICD codes. Proportional hazards models were used to assess the association between myosteatosis and incident COPD disease over 15 years. Models were adjusted for confounders including age, sex, pack years of smoking, emphysema, body mass index, inflammation, diabetes, and socioeconomic status. Results/Data: A total of 283 cases of incident COPD were identified. Individuals in the lowest quartile of muscle attenuation had significantly higher cumulative incidence compared to other quartiles. In minimally adjusted models, the hazard ratio comparing the lowest to highest quartile was 1.87. In fully adjusted models, the association remained significant with a hazard ratio of 1.32. Conclusion(s): AI-based quantification of myosteatosis on routine cardiac CT scans independently predicts future risk of COPD. Adverse muscle composition in the pectoralis, intercostal, and paraspinal muscles may precede lung function decline. Opportunistic assessment of myosteatosis could enable early identification of individuals at elevated risk and support preventive interventions at elevated risk for COPD and guide preventive strategies before clinical disease onset.