Cancer treatment-related cardiotoxicity among survivors of childhood cancer: A comparative and integrated view of multiple measures of biological age acceleration

Introduction: Mounting evidence suggests that childhood cancer survivors experience biological age acceleration (BioAgeAccel) and are at increased risk for cardiovascular disease. We performed an integrated analysis of multiple BioAgeAccel measures to examine how aging relates to treatment-related toxicity and cardiovascular risk, and how biomarkers developed under different methodological schemes capture this pathway. Methods: Cardiovascular disease (CVD) was defined as either cardiomyopathy or myocardial infarction (MI). DNA methylation (DNAm) profiles of peripheral-blood-mononuclear-cells from 2,941 St. Jude Lifetime Cohort participants were generated using Illumina EPIC BeadChips. Biological age was assessed using 43 DNAm-based biomarkers, including epigenetic clocks such as GrimAge, PhenoAge, and DunedinPACE, as well as molecular biomarkers such as C-reactive protein (DNAmCRP). BioAgeAccel was defined as residuals from regressing biological age on chronological age at DNAm blood draw (except DunedinPACE). Cardiotoxic treatment exposures included radiation potentially exposing the heart (heart-RT; mean dose >15 Gy considered high-dose) and anthracycline chemotherapy (cumulative dose ≥250 mg/m 2 doxorubicin equivalents considered high-dose). Multivariable logistic regression models evaluated the associations of each biomarker with CVD. BioAgeAccel was also evaluated as a mediator between heart-RT or anthracycline exposure and CVD to examine its role in the causal pathway. Mediation analyses were conducted for each biomarker individually and for a composite measure derived from principal component analysis. Results: Median age at cancer diagnosis was 6.1 years (range 0-23.6; Q1-Q3: 2.8-12.6), and at DNAm blood draw was 30.1 years (range 0.2-66.6; Q1-Q3: 21.7-38.0). Overall, 9.4% developed cardiomyopathy (median onset 34.7 years), and 2.3% developed MI (median onset 40.7 years). Most BioAgeAccel measures were elevated after high-dose heart-RT, with fewer associations for high-dose anthracyclines. PC version of GrimAge showed the strongest association with MI (OR=1.85, 95%CI: 1.45-2.36; Bonferroni P =4.06E-05), and DNAmCRP showed the strongest association with cardiomyopathy (OR=1.31, 95%CI: 1.14-1.50; Bonferroni P =5.17E-03). Mediation estimates varied across biomarkers. A composite BioAgeAccel measure accounted for 31.6% of the heart-RT-MI association, 35.9% of the heart-RT-cardiomyopathy association, and 7.7% of the anthracycline-cardiomyopathy association. Conclusions: BioAgeAccel measures vary in their associations with cardiotoxicities and treatment-related aging pathways. These findings support further refinement and validation of DNAm-based aging biomarkers to advance future risk stratification and targeted prevention of long-term CVD in survivors.