ESTRO 2021 Abstract Book

S404

ESTRO 2021

selection have not been identified. The Danish Head and Neck Cancer Group (DAHANCA) has prepared a randomised trial in an enriched population of head-neck cancer patients aimed at reducing the risk of late dysphagia and xerostomia (NCT04607694). The enrichment lies in the plan comparison, where only patients with a normal tissue complication probability (NTCP) reduction in favor of protons can be randomised. With one national proton center, the plan comparisons are performed at local head-neck cancer centers and the ability of these to select and refer patients were tested in a non-randomised feasibility study. This study presents the results of the patients selection- and referral process. Materials and Methods The national proton center offered an educational program for dose planners and phycisians in patient selection and proton planning, and developed and updated a planning template. Comparative doseplans were reviewed at national video conferences prior to selection. From June 2019 to December 2020, selected patients with squamous cell carcinoma of the pharynx or larynx at all six Danish head-neck cancer centers were offered proton-photon plan comparison. In case of an NTCP reduction of at least 7%-point (until Nov 2019) or 5%-point (after Nov 2019) in favor of proton treatment for dysphagia (6 months) ≥ grade II (DAHANCA toxicity score) or xerostomia (6 months) ≥ grade II (EORTC HN35), the patient was offered proton treatment at the national proton therapy center. Results Proton-photon plan comparisons were performed in 141 patients, constituting 113 oropharynx, 12 hypopharynx, 12 larynx and 4 nasopharynx cancers. A total of 71 (50%) patients achieved an NTCP reduction above the clinical goal, constituting 55% (62/113) of oropharynx cancers, 17% (2/12) of hypopharynx, 33% (4/12) of larynx, and 3/4 of nasopharynx. In general, more patients were selected based on a reduction in the risk of dysphagia (Table 1). Of the 71 patients, 55 (77%) accepted referral to proton treatment. In these 55 patients, the median time from radiotherapy decision to the first proton fraction was 19 calendar days (interquartile range 18-23) and from referral to the national proton therapy center to first proton fraction 9 days (interquartile range 8-11). All patients started proton treatment within the time limits set by the Danish national integrated cancer pathways.

Conclusion Proton-photon plan comparisons with NTCP modelling at local head and neck cancer centers can be used to select patients for a national proton trial without exceeding time limits.

Proffered papers: Proffered papers 32: Image acquisition and processing

OC-0520 CT calibration for precise proton therapy planning in children E. Baer 1 , C. Collins-Fekete 1 , V. Rompokos 2 , Y. Zhang 1 , M. Gaze 3 , A. Warry 2 , A. Poynter 2 , G. Royle 1 1 University College London, Medical Physics and Biomedical Engineering, London, United Kingdom; 2 University College London Hospitals NHS Foundation Trust, Department of Radiotherapy Physics, London, United Kingdom; 3 University College London Hospitals NHS Foundation Trust, Department of Oncology, London, United Kingdom Purpose or Objective To assess dose and range errors in paediatric proton therapy treatments arising from using a conventional single-energy CT (SECT) calibration, derived from standardised tissues of the adult male, for RSP estimation in children. Dual-energy CT (DECT) is investigated to increase the RSP prediction accuracy and reduce dose and A data base of 51 human tissues (26 soft tissues, 25 bone tissues) for age groups newborn, 1-, 5-, 10-, 15-year- old and adult, with given compositions and densities (ICRP reports 143&110) is used. CT numbers are simulated using XCOM attenuation coefficients for 100 kV and 140 kV/Sn (tin filtered) spectra. For SECT, RSP values of all tissues are estimated using the stoichiometric CT calibration (Schneider et al. , PMB 41.1) derived with standardised tissues of an adult male (Woodard and White, BJR 59.708). For DECT, the calibration by Bourque et al. is used (Bourque et al ., PMB 59.8). The SECT- and DECT-predicted RSP values of the tissues are compared to theoretical RSPs. Resulting range/dose errors are evaluated on 3 computational phantoms range errors from SECT. Materials and Methods