ESTRO 37 Abstract book

S177

ESTRO 37

challenging in part due to limited resources, high cost, and difficulties with patient accrual. Therefore, alternative comparative approaches have been proposed to allow appropriate guidance in treatment selection, based on modeling clinical outcome in different patients. A model-based strategy may be a cost-effective strategy to quantify clinical gains with PBT via estimation of potential reduction normal tissue complication probability (NTCP). Such an approach may be optimal in informing patient eligibility for a chosen therapy to enhance clinical outcomes and cost efficiency. The subject of this presentation is to present and discuss a clinical validation of a multivariable NTCP model in patients treated for oropharynx cancer with PBT, and to compare outcomes against similarly-treated patients with IMRT. We also will present and discuss the possible implications of such findings, and how it may inform patient selection and prospective studies as it relates to proton therapy. Methods This study was a joint collaboration between 2 institutions (University Medical Centre of Groningen and the University of Pennsylvania), limited to patients with advanced-stage oropharynx cancer treated with curative intent (PBT, n=30; IMRT, n=175). NTCP models were developed using multivariable logistic regression analysis with backward selection. In the PBT treated patients, an alternative IMRT plan was also made, to serve as a reference to determine the benefit of PBT in terms of NTCP. The models were then applied to the PBT treated patients to compare predicted and observed clinical outcomes. Five, clinically-relevant, binary endpoints were analyzed at 6 months post-treatment: dysphagia ≥ grade 2, dysphagia ≥ grade 3, xerostomia ≥ grade 2, salivary duct inflammation ≥ grade 2, and feeding tube dependence. Corresponding toxicity grading was based on CTCAEv4. Paired t-tests and Wilcoxon rank tests were used to compare mean NTCP results for endpoints between PBT and IMRT. Results The NTCP models developed based on outcomes from all patients were applied to those receiving PBT. For patients receiving PBT, no significant differences were observed between the expected and observed prevalences. In addition, the NTCP-values were calculated for the alternative IMRT plans for all PBT treated patients, revealing significantly higher NTCP- values for the IMRT plans. PBT was associated with statistically significant reductions in the mean NTCP values for each endpoint at 6 months post treatment, with the largest absolute differences in rates of > grade 2 dysphagia and xerostomia. Conclusion At 6-month follow-up, this model-based approach demonstrates significant improvements in treatment- related toxicity with PBT compared to IMRT for oropharyngeal cancer. This study demonstrates the value of NTCP model based approaches in comparing predicted patient outcomes when randomized data are not available. This study also brings up questions on whether randomized data is necessary, and if so, how such a model-based approach could be incorporated into prospective trials. SP-0341 Advanced selection procedures for proton therapy in head and neck cancer patients. D. Scandurra 1 1 UMCG, Radiation Oncology, Groningen, The Netherlands Abstract text In the Netherlands, head and neck cancer is not a standard indication for proton therapy referral; there must first be a demonstrable benefit compared to photon therapy. This benefit can consist of: an expected reduction in treatment-related side-effects as determined by published and validated normal tissue

complication probability (NTCP) models, and/or an increased target dose or coverage due to limitations arising from exceeding adjacent critical organ tolerance doses. A patient selection procedure has been developed with a nation-wide consensus in order to specify the workflow for an inter-clinic photon-proton plan comparison and to define the relevant target coverage criteria, NTCP models and delta-NTCP thresholds which will be used to indicate eligibility for proton therapy. The presentation will consist of: o an overview of the patient selection procedure and plan comparison workflow o a demonstration of the proton planning process at the UMCG, including robust optimisation, evaluation of target coverage under uncertainties, and calculation of NTCP o a summary of planning comparison results since clinical operation o advanced tools and further work in development SP-0342 Carbon ion therapy in adenoid cystic carcinoma A. Jensen 1 1 University of Munich, Radiation Oncology, München, Germany Abstract text Due to their aggressive local growth patterns, adenoid cystic carcinomas require high radiation doses in order to achieve long-term local control. Unfortunately, patients with salivary gland malignancies are often diagnosed with advanced disease; especially in the paranasal sinuses both surgical and radiooncological approaches remain a challenge in view of close proximity to critical structures. In the 1980s, radiotherapy with neutrons and hence high- LET radiation resulted in superior local control rates but observed long-term toxicities were substantial. Charged particle beams with carbon ions (C12) also show increased radiobiological effectiveness as compared to photon RT. In contrast to neutron radiation though, they exhibit sharp dose gradients leading to highly conformal dose distributions with improved normal tissue sparing especially at complex anatomical sites. Therefore, carbon ion therapy (C12) in the treatment of adenoid cystic carcinoma has been intensively investigated by both Japanese and European groups. Recent analyses show significantly improved control and survival rates in patients treated with a combination regimen of C12 plus IMRT vs IMRT only. In patients with T4 tumours treated with C12, control rates did not differ according to resection status indicating that debulking surgery with sometimes substantial morbidity may not be necessary. Results were confirmed by the prospective COSMIC trial and validated in a larger patient cohort. In the head and neck in general and in ACC in particular, management of local recurrence following full course radiotherapy remains an oncological challenge. Chemotherapeutic regimen only achieve limited response rates if there are no surgical options. High-dose re- irradiation with charged particles can be feasible and achieves promising response rates with moderate toxicity. However, further dose escalation needs to be considered carefully as the risk of higher-grade late toxicity increases. In summary, particle therapy is a good treatment option with promising local control especially in advanced ACC. The use of tumour debulking surgery in advanced ACC may have to be reconsidered. In cases of local tumour relapse, C12 may be a good option when surgery is not feasible but has to be used with caution.