ESTRO 2020 Abstract book

S31 ESTRO 2020

Furthermore, it ultimately unburdens the RTT or radiation oncologist from the task of manual delineation and reduces the duration of daily treatment for patients, thereby lessening healthcare costs. To this end, the purpose of this work was to investigate the accuracy of automated contour propagation during daily online plan adaption using the Monaco TPS 1 within a cohort of rectal cancer patients. Material and Methods T2-weighted transverse MR images (in-plane resolution: 0.49x0.49 mm; slice spacing: 3.3 mm) of the abdomen were acquired for 14 patients with rectal cancer once per week over a 5 week period. The rectum and tumor in all images were manually delineated by a RTT (M) and checked by a radiation oncologist. Delineations in the MR images of the first week were automatically deformed and propagated to the MR images of each week (P) using the Monaco TPS (Version 5.40). The similarity between the propagated and manual contours was evaluated using the mean surface distance, Hausdorff distance, Jaccard index, Dice coefficient and volume similarity metrics. Given M and P, the set of voxels bounded by the manually delineated and propagated contours respectively, the volume similarity was calculated as VS = 2 * (|M| - |P|) / (|M| + |P|) , where |M| and |P| represents the cardinality of sets M and P. Results The group mean performance of the Monaco contour propagation algorithm is summarized in Table 1. Figure 1 shows boxplots of the Dice coefficient and Hausdorff distance between the manual and propagated contours over the evaluated period for each patient. Generally, the performance of the contour propagation algorithm was poorer for the tumor than for the rectum, likely due to the limited contrast available in the MR images between the tumor and the rectum. The negative values found for the volume similarity metric reflects an enlargement of the propagated contour relative to the true anatomy. This enlargement could result in a higher radiation delivered to the surrounding normal tissue if left uncorrected, increasing toxicity risk. The relatively wide spread of values observable in the intrasubject boxplots, and the large group standard deviation for the Hausdorff, mean surface distance and volume similarity metrics further suggests inconsistent performance of the deformable registration algorithm for automated contour propagation.

Conclusion The accuracy of the automatic Monaco TPS deformable contour propagation was found to be limited for rectal cancer patients, particularly for the tumor delineations. Further work is necessary to investigate the potential impact of the limited quality of contour propagation on the dosimetric requirements of the treatment plan. References [1]www.elekta.com/software-solutions/treatment- management/external-beam-planning/monaco PD-0071 The MOMENTUM study, an international registry of patients treated on an MR-linac: 9-month experience S. De Mol Van Otterloo 1 , E.L.A. Blezer 2 , C.D. Fuller 3 , C. Faivre-Finn 4 , A. Sahgal 5 , U.A. Van der Heide 6 , J.P. Christodouleas 7 , H. Akhiat 7 , S. Mook 1 , M.P.W. Intven 1 , B. Van Triest 6 , A.C. Tree 8 , A.M. Kirby 8 , S. Lalondrelle 8 , C. Schultz 9 , B. Erickson 10 , H. Emma 11 , M.E. Nowee 6 , S. Hafeez 8 , U. Oelfke 8 , R.H.A. Tersteeg 12 , D. Eggert 7 , W.A. Hall 10 , H.M. Verkooijen 1 1 University Medical Centre Utrecht, Department of Radiation Oncology, Utrecht, The Netherlands ; 2 Universitair Medisch Centrum Utrecht, Radiation oncology, Utrecht, The Netherlands ; 3 University of Texas MD Anderson Cancer Center Houston, Department of Radiation Oncology, Houston, USA ; 4 The Christie National Health Service Foundation Trust, Department of Radiation Oncology, Manchester, United Kingdom ; 5 Sunnybrook Health Sciences Centre, Radiation Oncology, Toronto, Canada ; 6 Netherlands Cancer Institute- Antoni van Leeuwenhoek Hospital, Department of Radiation Oncology, Amsterdam, The Netherlands ; 7 Elekta, Research Software Engineer, Stockholm, Sweden ; 8 The Royal Marsden NHS Foundation Trust and The Institute of Cancer Research, Department of Radiotherapy and Imaging Oncology, London, United Kingdom ; 9 Medical College of Wisconsin, Department of Radiation-Oncology, Wisconsin, USA ; 10 Medical College of Wisconsin, Department of Radiation-Oncology, Milwaukee, USA ; 11 The Institute of Cancer Research, Clinical Trials and Statistics Unit, London, United Kingdom ; 12 University Medical Centre Utrecht, Radiation oncology, Utrecht, The Netherlands