ESTRO 2021 Abstract Book

S67

ESTRO 2021

Conclusion Absolute displacement was restricted when HCHS was used as shown by the quantitative motion analysis and the lower percentage of patients that required a repeated scan. In practice, the HCHS turned out to be easy to use. OC-0092 Evaluation of MRI sequences for image guided radiotherapy in a MR-only radiotherapy planning pathway R. Brooks 1 , R. Pearson 1,2 , K. Pilling 1 , S. West 1 , B. Ormston 1 , A. Beardmore 1 , D. Redding 1 , J. Wyatt 1,2 1 Newcastle upon Tyne Hospitals NHS Foundation Trust, Northern Centre for Cancer Care, Newcastle upon Tyne, United Kingdom; 2 Newcastle University, Translational and Clinical Research Institute, Newcastle upon Tyne, United Kingdom Purpose or Objective Magnetic Resonance (MR)-only radiotherapy with MR-Cone Beam Computed Tomography (CBCT) soft-tissue matching enables the benefits of MR-only for delineation without requiring invasive fiducial markers. However, unlike CT, different MR sequences can produce images with very different appearances. In an MR- only pathway, several sequences may be acquired for different tasks such as delineation and synthetic CT generation. In the clinical MR-only prostate pathway at our centre a SPACE sequence (3D Turbo Spin Echo T2- weighted sequence) and a DIXON sequence (T1-weighted 3D gradient Echo sequence) were acquired for prostate radiotherapy planning. The aim of this study was to compare the inter-observer variability in MR- CBCT soft-tissue matching for these two sequences. Materials and Methods Four therapeutic radiographers with MR-CBCT matching experience participated. The first fraction CBCT for 10 patients were independently matched to the SPACE image and the in-phase DIXON image. The limits of agreement between radiographers were calculated using modified Bland-Altman methodology which assessed the disagreement of each observer with the mean of all observers. Results Limits of agreement on the SPACE sequence were 1.24 mm, 2.07 mm, 0.64 mm and on the DIXON sequence 2.04 mm, 2.43 mm and 0.67 mm (vertical, longitudinal, lateral). The SPACE sequence had better agreement in all translations than the DIXON sequence, likely due to the better soft-tissue contrast in the image. However, the difference in inter-observer variability was <1 mm in all directions and not likely to be clinically OC-0093 Automated organ at risk delineation in T2w head and pelvis MR images for MR-only radiation therapy L. Ruskó 1 , V. Czipczer 1 , B. Kolozsvári 1 , B. Deák-Karancsi 1 , R. Czabány 1 , B. Gyalai 1 , D. Hajnal 1 , Z. Karancsi 1 , M.E. Capala 2 , G.M. Verduijn 2 , R. Pearson 3 , J.J. Wyatt 3 , E. Borzasi 4 , G. Kelemen 4 , R. Kószó 4 , V. Paczona 4 , Z. Végváry 4 , C. Cozzini 5 , T. Tan 6 , R. Maxwell 3 , J.A. Hernandez Tamames 7 , S.F. Petit 2 , H. Mccallum 3 , K. Hideghéty 4 , F. Wiesinger 5 1 GE Healthcare, Digital AI Data Science, BUDAPEST, Hungary; 2 Erasmus MC Cancer Institute, Department of Radiation Oncology, Rotterdam, The Netherlands; 3 Newcastle University, Translational and Clinical Research Institute, Newcastle upon Tyne, United Kingdom; 4 University of Szeged, Department of Oncotherapy, Szeged, Hungary; 5 GE Healthcare, Imaging MR Applications and Workflow, Munich, Germany; 6 GE Healthcare, Digital AI Data Science, Hoevelaken, The Netherlands; 7 Erasmus MC, Department of Radiology and Nuclear Medicine, Rotterdam, The Netherlands Purpose or Objective MR images are frequently used to improve the accuracy of target and OAR delineation for RT planning that is currently based on CT images. The latest achievements on synthetic CT solutions enable MR-only RT planning by performing dose calculation using MR images only. This emerging scenario can be further augmented by developing automated OAR delineation for MR images. Since most commercially available auto-contouring tools are CT-based, the goal of this work was to demonstrate the feasibility of deep learning (DL) based head and pelvis OAR delineation in T2w MR images. Materials and Methods The input of the auto-contouring was a standard T2w MR image. For both anatomical sites the DL models were trained and tested using one dataset with 90-10% separation. The head models were developed on a public dataset [DOI:10.7937/tcia.2019.bcfjqfqb] including 55 T2w cases. The number of manually contoured cases per organ varied between 22 and 41 depending on coverage. The pelvis models were developed on a set of 48 T2w prostate cases, where all organs were manually contoured. Both datasets were contoured by medical students trained by radiation oncologists. The segmentation method was an ensemble of 2D and 3D convolutional neural networks. The organs were segmented in their bounding box using a 3D model, and the center of the bounding box was localized using 2D (axial, coronal, and sagittal) models. If the full coverage of a structure was not guaranteed (body, brain, bowel-bag) only 2D axial model was used. All 2D and 3D models were trained for each organ, separately. The auto-contours were compared with the manual contours using DICE and Surface DICE (SDICE) metrics. The first measures volumetric overlap, while the second measures how the surface of two objects overlaps within a predefined tolerance. The tolerance was set to 1 mm for head and 2 mm for pelvis organs. For each organ the mean metrics were computed on all test cases. The auto-contours were rated by 2 radiation oncologists using 1-5 score based on their clinical usability and the mean score was computed for each organ. significant. Conclusion The difference in appearance in images produced by the SPACE and DIXON MR sequences translated into small differences in MR-CBCT inter-observer variability. The impact on MR-CBCT should be evaluated prior to introducing new MR sequences in a MR-only pathway.