Risk modeling of imaging changes after proton beam therapy for childhood brain tumors.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND AND PURPOSE: In childhood brain tumors, minimizing long-term side effects of cancer therapy is a critical objective. Radiation-related imaging changes (ICs), indicative of potential radionecrosis, remain an area of active investigation in proton beam therapy (PBT). This study aimed to identify and correlate post-therapeutic ICs and radio-biological and dosimetric factors, including linear energy transfer (LET) and variable relative biological effectiveness (RBE), as well as clinical factors. MATERIALS AND METHODS: A 3:1 matched-pair cohort of 93 pediatric PBT patients from a register study was retrospectively analyzed. The cohort comprised various brain tumor entities, with follow-up MRI data available up to 14 months post-treatment. Potential clinical risk factors for therapy-induced ICs in pediatric brains were analyzed using logistic regression at both patient and voxel levels. Dosimetric parameters were evaluated for the entire brain, periventricular region (PVR), and brainstem. RESULTS: A total of 15 cases with post-therapeutic ICs from various childhood tumor entities were identified and localized in the brainstem, the PVR, and other brain regions. At the voxel level, the key predictor linked to increased IC probability was the product of dose D and proton dose-averaged LETd (D · LETdproton)σ=6 mm, excluding voxels below 5 Gy (RBE). The Gaussian filtering with a standard deviation σ of 6 mm served as a practical approach to account for spatial uncertainties. At the patient level, the median dose (D50%) within the volume of the healthy brain receiving more than 20 Gy (RBE) was most significant. CONCLUSION: The identified univariate voxel- and patient-level risk factors provide a foundation for predicting post-therapeutic ICs in pediatric CNS tumor patients treated with PBT. Our findings contribute to refining risk prediction models and optimizing treatment planning strategies, ultimately aiming to minimize long-term radiation-induced effects in pediatric brain tumor patients.