Radiation-Induced Contrast Enhancement After Proton Therapy for Paediatric Brain Tumours: Association with Variable...

Background/Objectives: Proton therapy is widely used in paediatric neuro-oncology because of its favourable dose distribution and reduced exposure to radiation of healthy tissue. Nevertheless, current knowledge regarding radiation-induced contrast enhancements (RICE) is largely derived from adult proton therapy cohorts, while evidence in paediatric patients remains limited and the underlying biological mechanisms are still incompletely understood. This study investigated the incidence of RICE in paediatric patients with low-grade glioma (LGG) and ependymoma and evaluated potential associations with linear energy transfer (LET) and variable relative biological effectiveness (RBE)-weighted dose distributions. Methods: In this retrospective single-centre study, paediatric patients with LGG or ependymoma were identified from a prospectively maintained institutional registry patients treated with proton therapy at the Heidelberg Ion-Beam Therapy Centre (HIT) between 2010 and 2024. Follow-up magnetic resonance imaging examinations were systematically reviewed for RICE. Lesions were contoured in RayStation and assessed dosimetrically using a fixed RBE of 1.1 and three variable RBE models. LET-weighted dose metrics and Dose_(RBE) × LET products were compared between RICE regions and planning target volumes (PTVs). Overall survival and progression-free survival were estimated using Kaplan-Meier method. Results: Seventy-six patients met the inclusion criteria, including 33 patients with LGG and 43 with ependymoma. A total of 724 follow-up magnetic resonance imaging examinations were reviewed. RICE occurred in six patients (7.9%), exclusively in the ependymoma subgroup (14.0%). Higher prescribed radiation dose was significantly associated with RICE occurrence (p = 0.031). Across all variable RBE models, biologically weighted dose estimates and Dose_(RBE) × LET products were consistently elevated within RICE regions compared with corresponding PTV regions, particularly at low- and intermediate-dose levels. Five-year overall survival rates were 91.7% for LGG and 86.6% for ependymoma, while corresponding progression-free survival rates were 84.4% and 69.3%, respectively. Conclusions: Variable RBE models identified biologically relevant dose heterogeneities within RICE regions that were not adequately captured by the conventional fixed RBE approach. These findings suggest that LET-associated biological dose escalation may contribute to radiation-related brain tissue changes after proton therapy and support further investigation and prospective validation of LET- and RBE-informed normal-tissue risk assessment in paediatric neuro-oncology.