ESTRO 36 Abstract Book

S534 ESTRO 36 _______________________________________________________________________________________________

IX was significantly correlated with overall stage (p=0.03) and N classification (p=0.04). There was a significant inverse correlation between MVD and CA-IX expression (r=- 0.22, p = 0.03). Multivariate analysis showed that low MVD combined with high CA IX-expression was a significant independent prognostic factor for worse loco-regional control (HR=2.6, 95%CI 1.1-5.0, p = 0.02) in the whole population. However, in the p16-positive subgroup, the difference was not significant (85.7% vs. 89.7%, p=0.73). Patients treated with CRT had a better LRC than those with BRT independent of MVD or CA-IX expression. Conclusion The combination of MVD and CA-IX status might give additional prognostic information in HNSCC patients with known HPV status. PO-0967 Analysis of tumour microenvironment using multi-parametric PET/MR imaging in HNSCC xenograft models S. Boeke 1 , R. Winter 2 , A. Menegakis 1 , P. Mena-Romano 2,3 , M. Krueger 4 , E.C. Sezgin 1 , G. Reischl 4 , B. Pichler 4 , D. Zips 1 , D. Thorwarth 2 1 University Hospital Tübingen, University Department of Radiation Oncology, Tübingen, Germany 2 University Hospital Tübingen, University Department of Radiation Oncology - Section for Biomedical Physics, Tübingen, Germany 3 Pontificia Universidad Católica de Chile, Institute of Physics, Santiago, Chile 4 Werner Siemens Imaging Center, Department of Preclinical Imaging and Radiopharmacy, Tübingen, Germany Purpose or Objective Hypoxia is a major determinant of outcome in radiotherapy (RT) especially in head and neck squamous cell carcinoma (HNSCC). Non-invasive imaging of tumour microenvironment with multi-parametric PET/MRI, using e.g. hypoxia specific tracers, is a potentially powerful technology for personalisation of RT. The aim of this study is to investigate simultaneously fMRI and hypoxia PET in HNSCC xenografts during the course of fractionated RT. Material and Methods FaDu tumours (n=7) were xenografted on the right hind leg of immunodeficient nude mice. After a growth period of 4-6 weeks multi-parametric FMISO-PET/MRI (7T, Bruker) was performed before and after RT (10 x 2 Gy in two weeks, small animal image guided RT platform, SAIGRT, Dresden, Germany). Following the second imaging, tumours were excised after injection of Pimonidazole and Hoechst for further histological analysis. The imaging protocol included a 80-90 min dynamic FMISO-PET acquisition, anatomical T2w and diffusion-weighted MRI (DWI, 9 b-values from 0 to 800 s/mm²) as well as DCE MRI. T2w anatomical MRI data was used for precise manual segmentation of the actual tumour region of interest (ROI). Within each tumour ROI, mean and maximum tumour-to-muscle ratios (TMR, TMR max ) as well as mean ADC values were analysed prior and post fractionated RT treatment. Results Two animals presented with very small tumor volume (< 10 mm³) which did not allow for ROI-based analysis before (n=1) or after (n=1) RT, respectively. The mean (SD) volume was 479.3 (651.7) and 808.0 mm³ (1146.3), mean ADC was 760.0 (138.3) and 950.0 10-³mm²/sec (176.9), mean TMR at 80 min post injection (pi) was 1.42 (0.27) and 0.98 (0.17) and mean TMR max at 80 min pi was 2.47 (0.18) and 1.75 (0.75) before and after 2 weeks of RT respectively (cf. figure 1). Mean changes (SD) during the two weeks of irradiation were 25.0% (71.64%), 19.7% (24.0%), -31.7% (10.5%), -21.5 % (86.9%) for tumour volume, ADC, TMR and TMR max , respectively.

Fig. 1: Tumour volume, TMR max and mean ADC before and after 10 fractions of RT. Red dashed-line depicts the mean, black the median, lower and upper hinges the 25 and 75 percentile, respectively. Conclusion Intra- and intertumoural heterogeneity in hypoxia distribution and fMRI parameters were observed. FMISO uptake after irradiation decreased and ADC-value increased during radiotherapy as a surrogate for reoxygenation and lower cell density or increasing necrotic areas, respectively. Our data will be extended with various tumour models and will form the preclinical data base for the development of an integrated multiparametric prediction model for personalised RT in HNSCC. PO-0968 The Role of epithelial to mesenchymal transition (EMT) as Biomarker for Radioresistance in HNSCC I. Kurth 1,2 , D. Digomann 2 , L. Hein 2 , A. Linge 1,2,3,4 , L. Koi 5 , S. Loeck 2 , K. Maebert 2 , H. Stephan 2 , C. Peitzsch 1 , M. Krause 1,2,3,4,5 , M. Baumann 2,3,4,5,6 , A. Dubrovska 2,5 1 NCT Partnerstandort Dresden, Translationale Onkologie, Dresden, Germany 2 OncoRay-National Center for Radiation Research in Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Dresden, Germany 3 Department of Radiation Oncology, Faculty of Medicine and University Hospital Carl Gustav Carus, Dresden, Germany 4 German Cancer Consortium DKTK Dresden, German Cancer Research Center DKFZ, Dresden, Germany 5 Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiation Oncology, Dresden, Germany 6 Deutsches Krebsforschungszentrum DKFZ, Heidelberg, Germany Purpose or Objective It is described that epithelial-to-mesenchymal transition (EMT) plays an important role in head and neck squamous carcinomas (HNSCC) progression and resistance to therapy [1] . Recent studies suggest that for instance the expression of EMT related microRNAs may cause intrinsic radioresistance in HNSCC [2] . During the process of EMT epithelial cancer cells obtain a more mesenchymal-like motile and invasive phenotype, which has been argued to sustain survival and therapy resistance of those tumor cells and facilitate cancer progression. Radiotherapy is one of the main approaches to treat HNSCC. However, tumor radioresistance often impedes the success of radiotherapy and has been found to drive tumor aggressiveness and expansion. In this study we asked the question, if radioresistant HNSCC populations display EMT features on a molecular as well as on a functional level and whether we can correlate those characteristics to

treatment outcome. Material and Methods

We used multiple irradiated HNSCC lines (IR) as an established model to investigate the traits of radioresistance [3] . Global gene expression, protein