Abstract Book

S265

ESTRO 37

validation. Confidence interval was obtained from bootstrap using 2000 replicates. Results Gender, acceleration, current smoker, tumour in the vicinity of the oral cavity, use of Nimorazole, and the two first principal dose components were selected as predictors using LASSO (bold values in table). Acceleration is a well-known risk factor while the tumour position indicates an increased risk beyond the prediction related to the oral cavity dose. The protective value of being male and current smoker is in line with previous findings of toxicity in oesophagus and lung. The risk related to dose is dominated by the PC1. The model calibration plot (predicted vs observed risk) based on 10 bins of equal number of patients ranked according to predicted value, show good agreement with the line of identity (grey line). The bootstrap adjusted area under the curve (AUC) was 0.77 (95% CI 0.73 – 0.81). PC1 and PC2 approximately describe the mean oral cavity dose and an almost linear balance between high and low doses to the oral cavity, respectively. Positive sign of PC1 coefficient increases the mean dose, while a positive value for PC2 increases the high dose component.

Conclusion A robust logistic regression model for prediction of radiation induced mucositis of H&N has been developed, which can be used as risk assessment of mucosal toxicity during treatment plan optimisation. The AUC value of 0.77 is significantly larger than previous published models on mucositis. OC-0510 The validity of photon-based rectum NTCP models together with a constant RBE for proton therapy J. Pedersen 1 , N. Mendenhall 2 , C. Bryant 2 , Z. Li 2 , S. Flampouri 2 , L.P. Muren 1 1 Aarhus University Hospital, Department of Medical Physics, Aarhus C, Denmark 2 University of Florida, Health Proton Therapy Institute, Jacksonville, USA Purpose or Objective Proton therapy (PT) is currently being explored to improve normal tissue (NT) sparing beyond what can be achieved with conventional photon-based therapy. When planning and evaluating PT dose distributions it is currently assumed that the relative biological effectiveness (RBE) of protons is uniformly 10% higher than photons for all tumours and NTs. However, there is emerging evidence – mostly from pre-clinical and simulation studies – that the RBE increases towards the distal end of the beam. This could increase RBE-weighted doses in critical NTs and lead to higher-than-expected rates of morbidity. In addition, PT dose distributions have a reduced NT low-to-intermediate ‘dose bath’ compared to photons, questioning the applicability of photon-based NT complication probability (NTCP) models to PT. The aim of this study was therefore to explore whether photon-based NTCP models combined with a constant RBE (of 1.1) were valid for PT, by comparison to prospectively recorded rectal morbidity data in a large cohort of prostate cancer patients treated with PT. Material and Methods Treatment planning and morbidity data from 1102 prostate cancer patients treated with passive scattering PT between 2006 and 2010 were analysed. Prospectively