A frequently observed late side effect of cancer treatment in young patients is an increased risk of infertility or impaired fertility due to chemotherapy-induced toxicity in the reproductive organs. These late effects of cancer treatment can severely impact patients’ long-term quality of life and overall outlook. While many adolescent and young adult (AYA) cancer patients are offered fertility-preserving treatments, the severity and outcomes of gonadotoxic effects vary widely between individuals. Consequently, some patients may receive inadequate care, while others experience overtreatment. Being able to predict an individual’s risk of late chemotherapy-induced organ toxicities would greatly enhance our ability to tailor treatment, improve outcomes, and minimize the burden of unnecessary invasive procedures. To address this, we plan to use existing clinical records and whole-genome sequencing (WGS) data from over 1,400 pediatric cancer patients included in Barntumörbanken (BTB) and the GMS Childhood Initiative. We will analyze this cohort to identify individuals who suffer from late effects of chemotherapy, including the type and timing of toxicities, as well as the interactions between disease characteristics and different types of organ toxicities. By linking genetic markers to clinical outcomes, we aim to identify biomarkers associated with the risk of late organ toxicities. Our overarching goal is to develop a clinically validated prediction model to identify male and female patients with increased risk for long-term effects of reproductive organ toxicities based on their constitutional genotypes. To this end, we will use both an exploratory approach targeted to identify genomic biomarkers associated with gonadal toxicity, and look at association of toxicity outcomes with a set of predefined variations in genes relevant to drug metabolism, DNA damage response, and tissue-specific toxicity responses. Machine learning techniques such as Random Forest, LASSO regression and autoencoders will be used to construct polygenic risk models. As discovery cohort, we will focus on children who were prepubertal at diagnosis and reached pubertal age (defined as 13 years for females and 14 years for males) at the time of data collection. Prediciton models will be tested using an internal validation cohort using individuals who will reach the age of 12 during the study period.