Selection for proton radiotherapy of grade 1–3 glioma patients

•National workshops ensure consensus when neurooncologists select patients with brain cancer for proton therapy.•A logistic regression-based model was developed to aid in the selection of grade 1–3 glioma patients for proton therapy.•Age at radiotherapy and difference in mean dose to healthy brain are the two most important selection criteria. For adult patients with grade 1–3 gliomas, identifying patients with an indication for proton therapy (PT) can be challenging due to sparse evidence supporting its benefits. In this study, we aimed to ensure national consensus and develop a decision support tool to aid clinicians in identifying patients with grade 1–3 gliomas eligible for PT. Sixty-one historic patients referred for postoperative radiotherapy for glioma grade 1–3 were included in this study and had new photon therapy and PT plans calculated. These plans along with clinical parameters were presented to neurooncologists with experience in treating brain tumours. The patients were presented at three workshops (WSs), where each neurooncologist individually had to choose between photon and proton therapy. Important parameters were selected using cross validation. Multivariable logistic regression was used to predict the neurooncologists’ treatment modality choice. At the three WSs 23, 24 and 19 randomly selected patients were presented. Seventy-five percent of the neurooncologists agreed for 14 patients (61%), 16 patients (67%) and 15 patients (79%) at WS1, WS2 and WS3. Age at radiotherapy and difference in mean dose (ΔDmean) to the residual brain were significant predictors of the choice of treatment modality, p