Abstract 4142483: An Integrated Germline and Somatic Genomic Model Improves Risk Prediction for Coronary Artery Disease

Background: Multiple germline and somatic genomic drivers are associated with coronary artery disease (CAD), but their combined effects have not been thoroughly investigated. Hypothesis: An integrated model using all available information from single DNA biopsy including germline and somatic drivers improves CAD risk prediction. Methods: Based on whole genome sequencing data from UK Biobank, we curated familial hypercholesterolemia variants, clonal hematopoiesis of indeterminate potential, and leukocyte telomere length to calculate CAD polygenic risk score (PRS), proteome PRS, and metabolome PRS. An integrated genomic model (IGM) was developed by incorporating the six germline and somatic drivers for CAD. Results: Among 391,536 participants from UK Biobank (mean age 56.5 years; 53.8% women), 28,346 participants (7.2%) underwent incident CAD over a median follow-up of 12.3 years. The IGM showed better performance in predicting CAD risk (C-statistic 0.734, 95% CI 0.731 to 0.736) compared to the pooled cohort equations (PCE)(C-statistic 0.725, 95% CI 0.723 to 0.728). Addition of IGM to PCE further improved prediction accuracy (C-statistic 0.750, 95% CI 0.748 to 0.753) and resulted in a net reclassification improvement of 3.7% at risk threshold of 7.5%. IGM holistically captured CAD genetic risk, identifying individuals at high risk for CAD who do not have a single high-risk genomic factor but whose overall genomic profile places them at significant risk. Conclusions: We developed a state-of-the-art genomic model that integrates all available DNA information from a single biopsy. By accounting for the cumulative effects of multiple genomic drivers, IGM identifies high-risk individuals who might be overlooked by genomic risk models that rely on a single genetic factor.