ESTRO 36 Abstract Book

S492 ESTRO 36 2017 _______________________________________________________________________________________________

Figure 1.

by the 3 observers showed a median dice similarity coefficient of 0.71, 0.69 and 0.72 respectively. For all 3 observers the mean hausdorff distance was small with median (range) distances between PET and DW of 2.3 (1.5 – 6.8), 2.5 (1.6 – 6.9) and 2.0 (1.35 – 7.6) mm respectively. Over all patients, the median 95 th percentile distances were 6.0 (3.0 – 13.4), 6.6 (4.0 – 24.0) and 5.3 (3.4 – 26.0) mm.

Conclusion ABAS is a clinically useful tool for segmenting structures in breast cancer loco-regional radiation therapy in a multi- institutional setting. The introduction of ABAS in daily clinical practice will significantly reduce the workload especially in departments where the radiation therapy technologists (RTTs) are responsible for target volume delineation and treatment planning. Manual correction of some structures is important before clinical use. Moreover, applying ABAS may be a reasonable alternative for consistent segmentation and easy quality assurance testing in multi-institutional trials. Careful selection and stratification of atlas subjects seems to be the most influencing factor in the outcome of the ABAS. Further investigation to find out the best stratification factors is encouraged. Based on these results, ABAS is now made available for Danish patients. PO-0899 Tumor volume delineation us ing non-EPI diffusion weighted MRI and FDG-PET in he ad-and-neck patients. B. Peltenburg 1 , T. Schakel 1 , J.W. Dankbaar 2 , M. Aristophanous 3 , C.H.J. Terhaard 1 , J.M. Hoogduin 2 , M.E.P. Philippens 1 1 UMC Utrecht, Radiation Oncology, Utrecht, The Netherlands 2 UMC Utrecht, Radiology, Utrecht, The Netherlands 3 MD Anderson Cancer Center, Radiation Physics, Houston, USA Purpose or Objective Diffusion weighted (DW) MRI shows high contrast between tumor and the surrounding tissue, which makes it a candidate to facilitate target volume delineation in head- and-neck (HN) radiotherapy treatment planning. In this study we assess the performance of geometrically undistorted DW MRI for target delineation in terms of interobserver agreement and spatial concordance with automatic delineation on 18 F-fluorodeoxyglucose (FDG) positron emission tomography (PET). Material and Methods Fifteen head-and-neck cancer patients underwent both standard echo-planar imaging based (EPI) and undistorted fast spin-echo based (SPLICE) DW MRI in addition to FDG- PET for RT treatment planning. Target delineation on DW MRI was performed by 3 observers, while for PET a semi- automatic segmentation was performed using a Gaussian mixture model. Volumes, overlap metrics, defined as dice similarity coefficient and generalized conformity index, and hausdorff distances were calculated from the delineations. Results The median volumes delineated by the 3 observers on DW MRI were 10.8, 10.5 and 9.0 mL respectively. The median conformity index over all patients was 0.73 (range 0.38 – 0.80). On PET, a significantly smaller median volume of 8.0 mL was found. Compared with PET, the delineations

Conclusion Diffusion weighted imaging optimized for geometric accuracy resulted in target volume delineation with good interobserver agreement and a large similarity with PET. PO-0900 Quantifying the Effect of MRI Geometrical Distortions on Radiotherapy Treatment Planning Doses. M. Adjeiwaah 1 , M. Bylund 1 , J. Lundman 1 , J. Jonsson 1 , T. Nyholm 1 1 Umeå University, Radiation Sciences, Umea, Sweden Purpose or Objective The use of MRI for Radiotherapy Treatment Planning (RTP) is increasing and the proposed MR-only workflow could be beneficial. One worry of an MR-only RTP is geometrical distortions. There are at present few studies focusing on the effect of MR geometrical distortions on planned doses in an MR-only treatment planning and to our knowledge, none fully takes into account both gradient non-linearities and Patient-induced Susceptibility effects. This study focused on quantifying the effect of gradient non- linearities and Patient–induced Susceptibility effects on dose distributions for Prostate Cancers. Material and Methods The deformation field was generated by adding measured machine-specific and simulated patient-induced susceptibility effect deformation fields for a 3T scanner as shown in Fig. 1. Different bandwidths and simulated gradient readouts in the anterior/posterior (A/P) and right/left (R/L) directions were used. To isolate the effect of the distortions, the deformation fields were applied to 17 Prostate Patient CT images and their corresponding clinically delineated structures, giving a distorted CT (dCT). VMAT optimized plans were generated for all distorted cases and recalculated on the undistorted CT