50. Retrospective study of late radiation damages after focal radiotherapy for childhood brain tumors

Abstract

Purpose Investigate late radiation damages after focal radiotherapy (RT) for childhood brain cancer in terms of neurocognitive outcome. Methods 45 retrospective children (age at RT:median = 6.2 years, range 1.1–22.5, age at valuations:median = 11.1 years, range 4.4–25.9) who received focal RT at our Institution with at least 3-years follow-up without relapse, were recruited to undergo cognitive, neuropsychological and psychological testing. Planning CT images were retrieved and patients’ brains were parceled in 116 regions of interest (ROIs) using the Tzourio-Mazoyer atlas [1] and a dedicated home-made, multimodal, highly deformable registration framework specifically developed for the study. Mean RT dose value was collected for each ROI. The pattern of association between the categorized cognitive tests scores and RT dose was assessed in each ROI through the Kruskal Wallis test. Subsequently, the preliminary threshold of RT dose (optimal cut-off value) discriminating patients with and without neurocognitive changes (impaired/non-impaired patients), was selected by using a Receiver Operating Characteristic curve (ROC). Results The statistical analysis allowed us to identify the ROIs where mean dose values and cognitive tests results were significantly related. For the 29 ROIs showing the most significant associations, preliminary threshold dose values related to a specific cognitive or neuropsychological damage were obtained, ranging from 3.5 to 29.8 Gy. As an example, Figure1 shows the distributions of dose values in the ROI “Frontal_ Sup_Orb_ R” for patients with impaired/non-impaired performances in a specific cognitive test (Processing Speed Index, WISC III). Conclusions The preliminary analysis of the retrospective children suggested that the main aim (identify brain structures highly susceptible to radiation damages and the respective radiation dose thresholds) is achievable, justifying a prospective study currently open to children enrollment. Partially funded by AIRC.