Abstract Body (Do not enter title and authors here): Introduction/Background
The AI-CVD initiative aims to extract opportunistic screening information from coronary artery calcium scans to improve cardiovascular disease prediction. Thoracic aortic calcification (TAC) is a known marker of atherosclerotic burden but remains underutilized in routine coronary artery calcium scan interpretation. Automated quantification of TAC using artificial intelligence may enhance cardiovascular risk prediction, particularly when integrated with conventional risk scores.
Research Questions
We evaluated whether AI-derived TAC from coronary artery calcium scans independently predicts incident cardiovascular disease in the Framingham Heart Study Offspring cohort.
Goals
To assess whether automated TAC measured by the AutoTAC component of AI-CVD predicts future cardiovascular events independently of coronary artery calcium and traditional cardiovascular risk factors.
Methods
Baseline coronary artery calcium scans from 1,002 asymptomatic participants in the Framingham Heart Study Offspring cohort were analyzed using AI-enabled thoracic ascending aortic calcification quantification. TAC scores were categorized as 0, 1–99, 100–299, 300–999, and ≥1000. Cox proportional hazards models estimated hazard ratios for cardiovascular disease across TAC categories using unadjusted, age-adjusted, and fully adjusted models accounting for coronary artery calcium and established risk factors.
Results
296 CVD events accrued over 17 years follow-up. In fully adjusted models, compared to participants with zero TAC scores, participants with TAC scores of 100–299 had a hazard ratio of 2.05 (95% CI: 1.19–3.54), those with scores 300–999 had a hazard ratio of 2.29 (95% CI: 1.32–3.97), and those with scores ≥1000 had a hazard ratio of 2.85 (95% CI: 1.66–4.89). Lower categories (1–99) were not statistically significant after adjustment (HR 1.21, 95% CI: 0.73–2.03). The risk of cardiovascular disease increased progressively with higher TAC burden.
Conclusion(s)
In the FHS Offspring cohort, AI-measured TAC from coronary artery calcium scans was independently associated with future cardiovascular disease events over 17 years of follow-up. These findings support the utility of opportunistic AI-enabled aortic calcification assessment as an adjunct to traditional coronary artery calcium scoring in enhancing long-term risk stratification.