RESUMO
BACKGROUND: Radiotherapy (RT) in the pediatric brain tumor population causes late neurocognitive effects. In the current study, the authors investigated associations between clinical and dosimetric risk factors and memory outcomes in a cohort of patients treated with proton radiotherapy (PRT). METHODS: A total of 70 patients (median age at PRT, 12.1 years [range, 5.0-22.5 years]) who were treated with PRT were identified with baseline and follow-up evaluations of visual and verbal memory (Children's Memory Scale and the third edition of the Wechsler Memory Scale). Whole-brain as well as bilateral hippocampal and temporal lobe contours were delineated for the calculation of dosimetric indices. Multivariate analyses were performed to assess associations of score changes over time with clinical factors and dosimetric indices. RESULTS: The median neurocognitive follow-up was 3.0 years (range, 1.1-11.4 years). For the entire cohort, delayed and immediate verbal memory scaled scores demonstrated small declines. The mean decline for delayed verbal memory scores was 0.6 (P = .01), and that for immediate verbal memory scores was 0.5 (P = .06). Immediate and delayed visual memory scores were not found to change significantly (+0.1 and -0.3, respectively; P>.30). A higher left hippocampal V20GyE (percentage of the volume of a particular anatomical region receiving at least a 20 gray equivalent) was correlated with a score decline in all 4 measures. Female sex was found to be predictive of lower delayed verbal memory follow-up scores (P = .035). CONCLUSIONS: Only delayed verbal memory scores were found to have declined statistically significantly at follow-up after PRT, reflecting some weakness in verbal memory retrieval. Given a correlation of left hippocampal dosimetry and memory outcomes after PRT, left hippocampal-sparing PRT plans may assist patients with pediatric brain tumors in preserving memory-retrieval abilities. Cancer 2018;124:2238-45. © 2018 American Cancer Society.