INFRASTRUCTURE DEVELOPMENT FOR PERSONALIZED RISK PREDICTION TO REDUCE CARDIOVASCULAR DISEASE IN CHILDHOOD CANCER SURVIVORS

Although childhood cancer survivors have lengthy life expectancies, they run the risk of experiencing long-term health issues as a result of their treatment. The most frequent non-cancerous cause of morbidity and mortality for these survivors is cardiac disease. Radiation therapy (RT) has been linked in numerous cohort studies to a higher chance of developing a late cardiac disease in these survivors, and this risk rises with higher mean heart doses and increased RT exposure to larger cardiac volumes. Since, the heart is a heterogeneous organ made up of several distinct substructures, RT dose received by the entire heart does not accurately represent the doses to the various cardiac substructures with likely different risk profiles. Given the ability of modern RT techniques to limit organ exposure and dose, contemporary RT plans should include dosimetric constraints specific to individual substructures. A limiting factor in developing cardiac substructure level dose-response models is that cardiac substructure doses are not available for long-term survivors for whom late cardiac outcomes are well characterized. Moreover, existing whole-heart dose-response models were established based on dosimetry from reconstructing survivors’ RT on computational phantoms with a simple heart model. Thus, the main objective of this research was to expand the dose reconstruction infrastructure for late effects studies and enhance the heart model to improve the accuracy of whole heart dosimetry and enable cardiac substructure dose reconstructions for large multi-institutional childhood cancer survivor cohorts and establish sub-structure level dose-response relationships