ESTRO 37 Abstract book

the PTV margins created at planning are neither to small nor to large. In this group we found that one planning CT is enough to adequately predict the bladder expansion. Daily variations in bladder filling show that all three plans should be available as half the patients required them at some stage of treatment. EP-2352 Evaluation of an automatic delineation software for organs at risk and lymph nodes in breast cancer A. Arsène-Henry 1 , H.P. Xu 2 , M. Robilliard 1 , W. El Amine 1 , E. Costa 1 , Y.M. Kirova 1 1 Institut Curie, Radiation oncology, Paris cedex 05, France 2 Ruijin Hospital, Radiation oncology, Shangai, China Purpose or Objective The intensity modulated radiation therapy (IMRT) needs a precise delineation of target volumes and organs at risk. Therefore there is an increase in the time spent on this procedure. The purpose of this study is to evaluate an automatic delineation software, Workflow Box (Mirada Medical, UK), for the everyday use of OAR and lymph nodes (LN) delineation in patients treated for early stage breast cancer. Material and Methods Twenty patients’ CT scans in treatment position for their breast cancer radiotherapy were delineated in respect of the ESTRO delineation guidelines to begin the creation of automatic delineation atlas. Then 30 others CT scans were delineated this time by the automatic delineation system and by the radiation oncologist (reference delineation plan). The precision of the delineation was evaluated using Overlap Volume (OV) Index and evaluation of standard deviation (SD). Results Regarding the OAR, the mean OV were between 0.49 (SD = 0.21) and 0.97 (ET= 0.03). Five OAR of 9 have shown OV ≥ 0.8. The mean OV for all OAR was 0.77 (SD=0.17). The system was less performing for the lymph nodes volumes with a mean OV of 0.43 (SD=0.1) and the OV were between 0.23 (SD= 0.13) and 0.52 (SD= 0.1). The use of this software reduced the delineation time by 40% per patient. Conclusion For patients with breast cancer, the automatic delineation software Workflow Box (Mirada Medical, UK) has allowed to reduce the time devoted to delineation with acceptable OAR contours. Improvement of delineation of the LN volumes is needed. A new evaluation will be realized after the use of the system in the routine practice. EP-2353 Influence of MRI/CT fusion image on target volumes delineation for RTof cervical vertebra tumor P. Jiang 1 , M. Na 1 , W. Junjie 1 , Z. Xile 1 , S. Haitao 1 , J. Weijuan 1 1 Peking University Third Hospital, radiation oncology department, Beijing, China Purpose or Objective To compare the differences of target volumes and dose distribution between outlined based on MRI/CT fusion image and based on CT image. To investigate the accuracy of target delineation on MRI based image Material and Methods 10 cases with primary cervical malignant tumor received IMRT in our center. Patients commit cervical spine column MRI examination(GE Discovery MR 750 3.0T) follow routine procedural of radiology department. The images of MRI of each patient transferred to our TPS work station(Varian) by PACS (Picture Archiving and Communication Systems) to registrate with CT image of treatment plan. 5 physicians outline the target on MRI/CT fusion based image and CT based image

Conclusion Due to organ motion, rectal and bladder constraints were not met for some patients. Applying strict dose constraints at the planning stage ensures dose received by OAR is kept to a minimum on treatment. EP-2351 Reducing organ at risk dose by implementing a bladder plan-of-the-day procedure P.L. Twickler 1 , D. Scandurra-Karssens 1 , H.A.M. Vanhauten 1 , H.E. Van Herpt 1 , J.A. Langendijk 1 , H.P. Van der Laan 1 1 UMCG, Radiotherapy, Groningen, The Netherlands Purpose or Objective In order to reduce the percentage of bladder cancer patients experiencing normal tissue complications from conventional radiotherapy, we have implemented a relatively simple and resource-efficient adaptive radiotherapy (ART) procedure. The target is to reduce the irradiated volume and subsequent side effects. Material and Methods For 20 patients treated with this plan-of-the-day (POTD) technique between 2015-2017, a planning CT was performed with an empty bladder. Three PTVs were delineated (Small [S], Medium [M], Large [L]) using anisotropic margins, derived from work published by Lalondrelle et al 2011, to estimate the target volume under different bladder filling conditions. The L-PTV is slightly smaller than the conventional-PTV used prior to POTD. Image registration match structures, based on the bladder + margin, were delineated for all 3 plans. At each fraction, the most conformal plan which achieves the required target coverage is selected from a library. A cone beam CT (CBCT) was performed and the images were registered initially on bony anatomy and then using the match structures the plan of the day was chosen. Results Overall the S-plan was delivered 18%, the M-plan 62% and the L-plan 20% of the times. The S-plan was delivered three times as often in the final five fractions than at the start of treatment, which might be explained by increased patient compliance to bladder filling instructions. On a per-patient basis, there is some variation in daily filling during the treatment course: all patients required at least two plans and 50% of the patients required all three plans at some stage of treatment. However, the L-plan is only in 5% of the fractions, indicating that extremes in bladder filling are rare. In 95% a plan smaller than the L-plan was chosen. In only 1% of all fractions the chosen plan changed from one extreme to another (e.g. S to L) the next day. In none of the patients repeat CTs were deemed necessary for treatment adaptations. In a random sample of 7 bladder patients the mean doses, in the nominal plan and therefore not the accumulated dose for each fraction, with the L, M and S-plans were for the rectum 37,3 Gy, 32,7 Gy and 27,8 Gy, for the sigmoid 58 Gy, 49,5 Gy and 33,7 Gy and for the small bowel 40 Gy, 25,5 Gy and 12,5 Gy, respectively. Conclusion There is a clear benefit in using POTD with bladder carcinoma, without which in 95% of all fractions these patients would have received a higher OAR dose. In most fractions the M-plan was chosen, giving confidence that