Abstract 6382: Risk prediction of dyslipidemia in long-term survivors of childhood cancer: A report from the St. Jude Lifetime Cohort

Abstract

Abstract Background: Dyslipidemia is a significant contributor to the increased risk of cardiovascular morbidity and premature mortality in long-term survivors of childhood cancer. Identifying individuals at risk at an early stage provides an opportunity for tailored prevention approaches. Methods: Long-term childhood cancer survivors from the St. Jude Lifetime Cohort, (SJLIFE, n=2,559; median age 41 years, range 20-77 years) were directly assessed for dyslipidemia. Risk factors under consideration included demographic characteristics, cancer treatment exposures (cumulative doses of anthracyclines, platinum agents, glucocorticoids, and radiation therapy), comorbid conditions at the time of assessment (hypertension, diabetes mellitus, growth hormone deficiency, hypogonadism, hypothyroidism, cardiomyopathy, and chronic kidney disease), and externally validated multi-ancestry general-population polygenic risk scores (PRSs) for lipid values (total cholesterol, LDL-C, HDL-C, and triglycerides). Multivariable logistic regression predicted a 15-year risk of dyslipidemia (CTCAE grade ≥2 hypercholesterolemia and/or hypertriglyceridemia) including variables selected by the stepwise approach. Performance of the model was assessed by the area under the receiver operating characteristic curve (AUC). Results: Dyslipidemia was clinically identified in 543 (25.1%) SJLIFE survivors. AUC of a clinical model with sex, race, age at prediction, and comorbid conditions (diabetes, growth hormone deficiency, hypogonadism, hypothyroidism, and hypertension) was 0.80 (95% CI, 0.78-0.82). Inclusion of cancer treatment exposures (cumulative doses of anthracyclines, cisplatin, carboplatin, glucocorticoids, abdominal and cranial radiation therapy) significantly increased AUC to 0.82 (95% CI, 0.80-0.84; P=2.1×10-4). Addition of two PRSs for LDL-C (one included only genome-wide significant variants adjusted for statins and another consisted of all genome-wide variants) and the PRS for triglycerides provided further significant improvement in AUC to 0.86 (95% CI, 0.84-0.88; P=4.2×10-12). Conclusions: We developed a prediction model with good performance for identifying the risk of dyslipidemia in childhood cancer survivors. In addition to demographics and comorbidities, cancer treatments and inherited genetic factors contributed significantly to improving the risk prediction of dyslipidemia. Results may refine risk stratification for dyslipidemia in survivors and present opportunities for more precise and personalized preventive strategies to address cardiovascular health. Independent validation in survivors from the Childhood Cancer Survivor Study and the Dutch Childhood Cancer Survivor Study (DCCSS)-LATER is currently underway. Citation Format: Kateryna Petrykey, Kendrick Li, Lu Xie, Achal Neupane, Angela Delaney, Christine Yu, Stephanie B. Dixon, Bonnie Ky, Cindy Im, Isaac B. Rhea, Jeffrey A. Towbin, Jason N. Johnson, Melissa Bolier, Vincent G. Pluimakers, Linda Broer, Sebastian J. Neggers, Marry M. van den Heuvel-Eibrink, Rebecca M. Howell, Kirsten K. Ness, Melissa M. Hudson, Gregory T. Armstrong, Yutaka Yasui, Yadav Sapkota. Risk prediction of dyslipidemia in long-term survivors of childhood cancer: A report from the St. Jude Lifetime Cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6382.