ESTRO 36 Abstract Book

S484 ESTRO 36 2017 _______________________________________________________________________________________________

decreasing pre- or mid-therapy K 1 spatial heterogeneity, higher but decreasing pre- or mid-therapy overall V b parameter value, and lower pre-therapy V b spatial heterogeneity.

Conclusion The post-treatment and response SUV max of the LNs were found to be significant prognostic factors for regional failure and OS in patients with locally advanced NSCLC treated with hypofractionated CCRT. These parameters might be useful in the selection of patients for additional therapy. PO-0889 FLT PET kinetic analysis biomarkers of resistance to radiotherapy for nasal tumours in canines U. Simoncic 1 , T.J. Bradshaw 2 , L. Kubicek 3 , L.J. Forrest 4 , R. Jeraj 5 1 Jozef Stefan Institute, F-8, Ljubljana, Slovenia 2 University of Wisconsin, Department of Radiology, Madison, USA 3 Angell Animal Medical Center, Angell Animal Medical Center, Boston, USA 4 University of Wisconsin, Department of Surgical Sciences- School of Veterinary Medicine, Madison, USA 5 University of Wisconsin, Department of Medical Physics, Madison, USA Purpose or Objective Imaging biomarkers of resistance to radiotherapy are prerequisite for precise treatment. Multiple imaging biomarkers are typically provided by a separate multi- tracer or multimodal imaging. This study assessed kinetic analysis as a means to create multiple imaging biomarkers of resistance to radiotherapy from a dynamic 3’- ( 18 F)fluoro-3’-deoxy-L-thymidine (FLT) positron emission Sixteen canine cancer patients with spontaneous nasal tumours were imaged dynamically with FLT PET before and during the radiotherapy. Images were analysed for kinetics on a voxel basis using a two tissue, four rate- constant compartmental model. Overall parameter values (mean and median over the region of intrests (ROI)) and heterogeneity measures (coefficient of variation (COV), ratio of interquartile range to median (IQR/median)) were evaluated over the tumour gross target volume for the transport (K i =K 1 k 3 /(k 2 +k 3 )), perfusion/permeability (K 1 ) and vascular fraction (V b ) parametric images. Response biomarkers were evaluated as a ratio of mid-therapy to pre-therapy regional values, (i.e. mean, median, COV, IQR/median). Alternative, spatial responses were evaluated as a mean, median, COV or IQR/median taken on a ratio of mid-therapy to pre-therapy prametric images. The time to progression after radiotherapy (TTP) was estimated by assessing the therapy response according to the RECIST. Kaplan-Meier analysis and univariate Cox proportional hazards (PH) regression were used to assess the impact of each imaging biomarker on the TTP. Results Pre- or mid-therapy overall K i parameters were significant predictors of TTP after the radiotherapy. However, many imaging biomarkers based on K 1 and V b parameters had higher predictive power for the radiation therapy response. Table shows results of univariate Cox proportional hazard regression for imaging biomarkers derived from FLT PET parametric images. Hazard is significantly increased for higher pre- or mid-therapy overall K i parameter values, higher or increasing pre- or mid-therapy overall K 1 parameter value, lower or tomography (PET) scan. Material and Methods

Figure shows selected results of Kaplan-Meier analyses that illustrates prognostic power of some imaging biomarkers based on FLT PET parametric images.

Conclusion Worse outcome after radiotherapy was significantly associated with higher pre- or mid-therapy overall K i . Additionally, we found that various imaging biomarkers derived from vascular parameters or their change through the therapy, contains even stronger prognostic information than the FLT transport parameter, which justify use of kinetic analysis. PO-0890 PET-based radiobiological modeling of changes in tumor hypoxia during chemoradiotherapy M. Crispin Ortuzar 1 , M. Grkovski 1 , B.J. Beattie 1 , N.Y. Lee 2 , N. Riaz 2 , J.L. Humm 1 , J. Jeong 1 , A. Fontanella 1 , J.O. Deasy 1 1 Memorial Sloan Kettering Cancer Center, Medical Physics, New York, USA 2 Memorial Sloan Kettering Cancer Center, Radiation Oncology, New York, USA Purpose or Objective To develop a mechanistic radiobiological model of tumor control probability (TCP) for predicting changes in tumor hypoxia during chemoradiotherapy, based on pre- treatment imaging of perfusion and hypoxia with 18 F- Fluoromisonidazole (FMISO) dynamic PET and of glucose metabolism with 18 F-Fluorodeoxyglucose (FDG) PET.