ESTRO 2024 - Abstract Book

S928

Clinical - CNS

ESTRO 2024

2582

Poster Discussion

Alopecia in brain tumour patients treated with proton beam therapy: can we predict and prevent it?

Simona Gaito 1,2 , Anna France 1 , Laura Cella 3 , Serena Monti 3 , Giuseppe Palma 3 , Gillian Whitfield 1,2 , Ed Smith 1,2 , Marianne Aznar 2 1 The Christie NHS FT, Proton Beam Therapy Centre, Manchester, United Kingdom. 2 The University of Manchester, Division of cancer science, Manchester, United Kingdom. 3 National Research Council, Institute of Biostructures and Bioimaging, Naples, Italy

Purpose/Objective:

The psychological experience of having cancer and its treatments is significantly influenced by the cosmetic results of those treatments. According to reports (1), alopecia is the most frequent and upsetting cosmetic side effect of cancer treatments, negatively affecting both mental and physical health. Radiation-induced alopecia (RIA) can in some patients be permanent, whereas chemotherapy-related alopecia is more often only temporary. Clinically, RIA is dose-dependent and typically manifests as a clearly defined area of hair loss that corresponds to the treatment area. Research teams from various institutions have created Proton Beam Therapy (PBT)-specific Normal Complication Probability Models (NTCP) in order to estimate the risk of this side effect. Aim of this work is to externally validate the NTCP models of acute, late and permanent RIA published by Palma et al. (2) in an independent cohort of patients treated with Pencil Beam Scanning (PBS)- PBT. If validated, these models could be easily exploited for individualized treatment planning and for scalp sparing PBT in brain tumour patients. The project received Institutional board approval. The Hospital electronic patient record system was interrogated to extract clinical data for patients of all age groups treated for brain tumours at our PBT Centre between December 2018 and February 2022. The dosimetric data that were relevant for the model validation were extracted from the Treatment Planning System (TPS) Eclipse. The scalp was defined according to Palma et al. (2) as a 5 mm rind of skin covering the cranial vault and corresponds to the area of the body surface that covers the brain parenchyma. RIA was defined as per CTCAE 5.0 grading system (Grade 0, 1 or 2). For the purpose of model validation, the RIA endpoint was dichotomized into non-event (grade 0 – 1) and event (grade 2). Descriptive statistics were used to report information about the proportions of acute, late and permanent rates of RIA in this cohort. As for the original work, two different modelling strategies were used for the purpose of model validation, the Lyman-Kutcher-Burman (LKB) analysis, based on pure dosimetric parameters, and the multivariable logistic regression (MLR) methodology, based on both clinical and dosimetric parameters. For the evaluation of the model discrimination the area under the ROC (Receiving operator characteristics) curve (AUC) was used. Material/Methods:

Results: