Purpose: One of the components of radiotherapy (RT) in medulloblastoma/primitive neuroectodermal tumors is the prophylactic irradiation of the whole brain (WBI). With the aim of reducing late neuropsychologic morbidity a CT-scan-based dosimetric study was undertaken in which treatment was confined mainly or exclusively to supratentorial sites considered at high risk for disease recurrence.
Methods and materials: A comparative dosimetric study is presented in which a three field (two laterals and one posterior) proton plan (spot scanning method) is compared with a two-field conventional WBI 6 MV x-ray plan, to a 6-field "hand-made" 6 MV x-ray plan, and to a computer-optimized 9-field "inverse" 15 MV x-ray plan. For favorable patients, 30 Gy were delivered to the ventricles and main cisterns, the subfrontal and subtemporal regions, and the posterior fossa. For the unfavorable patients, 10 Gy WBI preceeded a boost to 30 Gy to the same treatment volume chosen for favorable patients. The dose distribution was evaluated with dose-volume histograms to examine the coverage of the targets as well as the dose to the nontarget brain and optical structures. In addition, the risks of radiation-related late neuropsychologic effects after WBI were collected from the literature and used to predict normal tissue complication probabilities (NTCPs) for an intelligence quotient deficit after treatment with photon or proton beams.
Results: Proton beams succeeded better in reducing the dose to the brain hemispheres and eye than any of the photon plans. A 25.1% risk of an IQ score <90 was predicted after 30 Gy WBI. Almost a 10% drop in the predicted risk was observed when using proton beams in both favorable and unfavorable patients. However, predicted NTCPs for both optimized photon plans ("hand made" and "inverse") were only slightly higher (0.3-2.5%) than those of proton beams. An age-modifying factor was introduced in the predictive NTCP model to assess for IQ differences in relation with age at irradiation. Children with ages between age 4 to 8 benefitted most from the dose reduction in this exercise (similar NTCP predictions for both proton and "inverse" plans).
Conclusion: Modulated proton beams may help to significantly reduce the irradiation of normal brain while optimally treating the supratentorial subsites at higher risk for relapse. A decrease in morbidity can be expected from protons and both optimized proton plans compared to WBI.