Can we predict late risk of cancer-therapy related cardiac dysfunction including heart failure in survivors of adolescent and young adult cancer? A systematic review

Survivors of adolescent and young adult (AYA, age 15-39y) cancer are at risk for late-onset cancer therapy-related cardiac dysfunction (CTRCD) years after therapy completion which can progress to heart failure. It is unclear if any existing CTRCD risk prediction model is applicable to survivors of AYA cancer. Our objective was to identify and evaluate risk prediction models for late-onset CTRCD and examine applicability to survivors of AYA cancer. We searched multiple databases to identify CTRCD risk prediction models developed and/or validated in cancer survivors of any age, ≥ 1 year from diagnosis. We included full-text English studies using real-life patient data and routine clinical variables to predict late-onset CTRCD. We used the Prediction model Risk of Bias Assessment Tool to evaluate models for bias and applicability. We described studies, models, included risk factors, and participants overall and by age group. We screened 12,740 records and identified 22 studies which developed and/or validated 64 models for the prediction of late-onset CTRCD (54.7% symptomatic) among 139,765 cancer survivors (9.4% AYA). Age, sex, anthracycline exposure, and chest radiation were the most common predictors. Most models (84.4%) were at high risk of bias, often due to insufficient reporting of statistical methods. Only two studies developed models ( n = 8) both at low risk of bias and high applicability to survivors of AYA cancer, although neither were validated across the full AYA age- or disease-spectrum. We did not identify any model for late-onset CTRCD risk prediction both at low risk of bias and validated for use in all survivors of AYA cancer. Development and/or validation of high-quality risk prediction models for CTRCD across all ages and diseases within the AYA population is warranted. CRD42023458516. Systematic review of risk prediction models for late onset cancer treatment related cardiac dysfunction in cancer survivors. Risk prediction models for late-onset chemotherapy-related cardiac dysfunction in cancer survivors are limited by bias and have not been effectively validated in most AYA or adult populations. Future research should focus on validation of existing models in diverse populations by age, diagnosis, and geography.