Identifying genetic susceptibility for severe cardiac events in testicular cancer survivors.

Abstract

12048 Background: Testicular cancer (TC) survivors face an increased risk of cardiovascular diseases (CVD) as a late adverse effect of oncological treatment, especially after cisplatin-based chemotherapy. To explore genetic susceptibility, we performed a genome-wide association study using hierarchical clustering and assessed impact of polygenic risk scores (PRS). Methods: A subcohort of TC patients treated at expert centers in the Netherlands between 1976 and 2007 was established by hospital stratified random sampling, and enriched with all identified cases of severe CVD (coronary artery disease, myocardial infarction, heart failure) of the entire cohort (Lubberts et al J Clin Oncol 41:3512-3522, 2023). Genotyping was performed using the Global Screening Array. The study population consisted of 396 TC patients: 92 survivors with cardiac disease and 304 survivors without cardiac disease. Seven general population derived PRSs from the CARDIoGRAMplusC4D consortium were applied to assess cardiac risk, followed by hierarchical SNP aggregation based on linkage disequilibrium to identify cardiac risk-associated SNPs. These SNPs were used for co-functionality analysis and decision tree analyses to develop novel PRSs. Results: Five of the seven external PRSs were significantly associated with cardiac risk, explaining 8%-12% of the variance, with the CARDIoGRAM GWAS PRS showing the strongest association. Patient’s age at diagnosis explained the largest variance in cardiac risk (21%). Hierarchical SNP aggregation identified 67 cardiac risk-associated SNPs, individually explaining 3% - 12% of the variance. Co-functionality analysis revealed biological processes, including cardiovascular development, immune signaling, and DNA damage repair, as potential mediators of these SNPs' effects. Using decision tree analysis, a novel PRS was developed incorporating CVD risk-associated SNPs and clinical risk factors at start of treatment. This PRS+ achieved the highest concordance statistic of 0.783 (standard error 0.022) and explained maximum variance (50%) among all tested models. This PRS+ identified that patients diagnosed at age > 41 years had higher cardiac risk, independent of genotype data. Whereas a subset of patients diagnosed at age ≤41 years could be identified as having increased cardiac risk based on the minor allele frequency of five SNPs (rs6830970, rs17666409, rs10782601, rs2446862, rs767707). Conclusions: This study introduces a novel PRS to identify TC patients with a higher risk of CVD, enabling the development of personalized cardiovascular risk management strategies at the start of their cancer treatment. It also advances understanding of the biological mechanisms underlying this increased risk. Clinical trial information: NCT02276430 .