ESTRO 2021 Abstract Book

S401

ESTRO 2021

Conclusion The internationally adopted DAHANCA guidelines with geometric target definition do not lead to more local failures and have been safely implemented. The resulting smaller high-dose volumes may result in a reduction of morbidities. OC-0517 Importance of dose to the parotid gland stem cell rich region in preventing xerostomia in RT for HNC M.I. van Rijn - Dekker 1 , P. van Luijk 1 , L. Van den Bosch 1 , J.G. van den Hoek 1 , E. Oldehinkel 1 , T.W. Meijer 1 , J.A. Langendijk 1 , R.J. Steenbakkers 1 1 University Medical Center Groningen (UMCG), Radiation Oncology, Groningen, The Netherlands Purpose or Objective Radiotherapy (RT) for head and neck cancer (HNC) is associated with salivary gland damage and subsequent xerostomia. The radiation response of parotid glands of rats, mice, and patients critically depends on dose to its stem cells, mainly located in the gland’s main ducts. In a double-blind randomized controlled trial (RCT), dose to this stem cell rich (SCR) region in the contralateral parotid gland was the most important dosimetric predictor for daytime xerostomia (ClinicalTrials.gov number NCT01955239). However, besides dose to the SCR region, other risk factors are important as well. In the current analysis we aimed to investigate the role of the SCR region in addition to dose to other organs at risk (OAR) and clinical risk factors, in development of xerostomia. Materials and Methods The study population consisted of 1,103 patients treated with definitive RT with or without systemic treatment included in a prospective data registration program. NTCP-models were developed using multivariable analysis considering e.g. nonlinearity of and collinearity between risk factors and overfitting. Patient-rated xerostomia was assessed with the H&N35 and GRIX questionnaires, evaluating moderate-to- severe general, daytime and nighttime xerostomia (resp. genXER12M, dayXER12M and nightXER12M). Physician-rated grade ≥2 xerostomia (physXER12M) was assessed 12 months after RT. The NTCP model was developed in two steps. First, the 102 patients included in the RCT were used to determine whether dose to the SCR region was more predictive for development of xerostomia, based on the BIC, than dose to the entire parotid gland. Secondly, the role of other risk factors and doses to OARs was tested on 561 additional patients. The resulting model was externally validated in the 350 remaining patients. Results In the RCT cohort dose to the contralateral SCR region was more predictive for dayXERM12 than dose to the entire parotid gland. Expanding the NTCP model in step two added baseline xerostomia, mean dose to the oral cavity and required a square root transformation of the mean dose to the contralateral SCR region ( figure 1, table 1 ). The resulting model showed moderate discrimination and calibration performation in external validation ( table 1 ). The SCR region dose was not predictive for nightXER12M and genXER12M. PhysXER12M, was associated with