Monte Carlo-driven predictions of neurocognitive and hearing impairments following proton and photon radiotherapy for pediatric brain-tumor patients.

As proton radiotherapy (RT) remains a limited resource, predictive estimates of the potential reduction in adverse treatment-related outcomes compared to photon RT could potentially help improve treatment selection. The aim of this study was to predict the magnitude of the neurocognitive and hearing deficits associated with proton and photon RT for children with brain tumors. The existing RT plans for 50 children treated with photon intensity modulated RT were compared with generated intensity modulated proton RT plans. The proton and photon RT dose distribution was used to estimate the Full Scale Intelligence Quotient (IQ) via a Monte Carlo model and the probability of hearing loss per ear, based on previously published dose-risk relationships. Compared to photon plans, the mean brain dose was found to be reduced in all proton plans, translating into a gain of [Formula: see text] IQ points for the whole cohort at 5 years post-RT for dose regimens of 54 Gy, or [Formula: see text] IQ points for dose regimens of 59.4 Gy, where the errors shown represent statistical and systematic uncertainties. The probability of hearing loss ≥20 dB per ear was less for proton versus photon RT: overall (9 ± 4) versus (17 ± 6)%, respectively, based on dose regimens of 54 Gy, and (13 ± 5) versus (23 ± 9)% for dose regimens of 59.4 Gy. Proton RT is thus expected to reduce the detrimental effect of RT upon IQ and hearing as compared to photon RT for pediatric brain tumors.