ESTRO 2020 Abstract book

S960 ESTRO 2020

dosimetric variations in the rectum only for V70 (-2.13 % ± 6.77%; p=0.006) but much larger in the bladder (12.75% ± 11.92%, 18.98% ± 16.84 % and 21.55% ± 18.75% for V75, V70 and V65 respectively; p <10 -3 ). The results of the linear regression analysis showed that for a 10% bladder volume decrease there was an increase of the V75, V70 and V65 of 3.64%, 5.66% and 5.75% respectively. Conclusion The anatomic changes of the rectum during IMRT could lead to a significant displacement of the prostate that could partially remove it from the planned volume and negatively influence the delivered dose to the target volume. Bladder volume variations could cause dosimetric modifications at its level with an increase of the mean dose while bladder shrinkage. PO-1656 Plan library to manage daily motion in locally advanced cervical cancer Z. Paquier 1 , Y. Jourani 1 , S. Benkhaled 2 , A. De Caluwé 2 , N. Reynaert 1 1 Institut Jules Bordet - Université Libre de Bruxelles, Medical Physics, Brussels, Belgium ; 2 Institut Jules Bordet - Université Libre de Bruxelles, Radiation Oncology, Brussels, Belgium Purpose or Objective Daily image guided radiotherapy for cervical cancer patients can reveal large motion of the uterus due to bladder and rectal volume changes during the treatment course. A plan library strategy can take those motions into account while keeping small margins to spare healthy tissues. The aim of this study was to create a plan library based on a cervix-uterus structure (CTV-T) motion model using non-rigid registration. Material and Methods Twelve patients, who had an empty and a full bladder planning CT available and with more than 2.5 cm displacement at the tip of the uterus, were retrospectively included. A non-rigid registration method was implemented in MICE toolkit (v1.0.9, NONPI Medical AB, Umeå, Sweden), based only on the delineation of the structures to register (CTV-T and bladder) of the two CT, to create a mid position structure. The Dice Similarity Coefficient (DSC), the Hausdorff Distance (HD), the Mean Distance to Agreement (MDA) and the Inverse Consistency Error (ICE) were computed to evaluate the algorithm. Two strategies were designed: a non-adaptive strategy that uses a motion robust plan to all the bladder fillings, and a plan library consisting of a robust plan, an empty to half- full bladder plan and a half-full to full bladder plan. At every treatment fraction, the best plan from the library was selected based on daily CBCT. Plan selection frequencies were reported. Doses in the bladder and in the bowel bag were calculated on the planning CT and compared to the non-adapted treatment. The dose delivered in the CTV-T if the wrong plan is selected was also analyzed. Results The registration result for the DSC is around 0.96, the 95th percentile of the HD is around 2mm and the MDA is below 1mm. The maximal HD and its 99th percentile are larger for the CTV-T (4.71 - 19.53mm and 2.17 - 6.79mm) than for the bladder (3.19 - 6.57mm and 2 - 3.22mm). However, the ICE is better for the CTV-T (1.14 - 2.87mm) compared to the bladder (1.42 - 4.61mm). The mid position structures derived from the registration were acceptable visually (see Figure 1), especially for the CTV-T.

immediately before and during treatment, resulting in the need for fiducial markers or beam alignment based on surrogate structures. Online MR guided radiotherapy using a hybrid MR-Linac may facilitate markerless SBRT and improve treatment precision by the superior tumor visualization. Herein, we report our initial experience with Ten consecutive patients with oligometastatic or oligoprogressive liver metastases who were simulated and treated on a 1.5T MR-Linac (“Unity”, Elekta AB, Stockholm, Sweden) with > 5 Gy per fraction were included in this study. In general, target volume dose, fractionation and organ at risk constraints were chosen according to the UK SABR guidelines. Online plan adaptations were performed based on daily T2 weighted, isotropic MRI scans. Results Five patients had colorectal primaries, the other primaries were esophageal, melanoma, gastrointestinal stromal tumor, head and neck and cystic duct. The median tumor volume was 34.6 cc (IQR 14.5 cc - 84.2 cc). Seven patients were treated with five fractions (6 Gy – 10 Gy per fraction), three patients received 45 Gy in three fractions. Nine of the ten metastases were well visible on T2 scans in the online adaptive workflow. The single metastasis with poor visibility was a melanoma metastasis located in segment I and therefore permitting a precise adaptation. Median time from patient-identity check to “beam-off” was 29 minutes (IQR 27 min – 33 min). Treatment was well tolerated in all instances with no medical intervention required beyond antiemetics. No event of radiation induced liver disease was observed within the first twelve weeks after treatment. Conclusion The majority of liver metastases can be visualized online on a 1.5 T MR-Linac and can be treated without fiducial markers. Funding: German Research Council (DFG), Clinician Scientist program (medical faculty Tübingen) PO-1655 Dosimetric impact of the anatomical pelvic organs variations during prostate cancer IMRT F. Dhouib Trabelsi 1 , W. Mnejja 1 , N. Fourati 1 , T. Sahnoun 1 , W. Siala 1 , L. Farhat 1 , J. Daoud 1 1 University Hospital Habib Bourguiba- Sfax, Radiotherapy, Sfax, Tunisia Purpose or Objective To evaluate the dosimetric impact of pelvic organs changes during the intensity modulated radiotherapy (IMRT) for localized prostate cancers. Material and Methods Ten patients were included in this dosimetric study. Five dosimetric control scans (CT c ) were performed weekly and the delineation of the prostate and pelvic organs (bladder and rectum) was redone on each scan. The volumetric variations of these three organs as well as the motion of the prostate isocenter in the three axes of space (X,Y,Z) were noted. The maximum dose (Dmax) and the dose received by, respectively, 100% and 2% of prostate volume (D100% and D2%) were recalculated on each CT c . The percentage of rectal volume receiving respectively 60 Gy, 70 Gy and 75 Gy (V60, V70 and V75) and bladder volume receiving respectively 65 Gy, 70 Gy and 75 Gy (V65, V70 and V75) were also evaluated. Results The dosimetric impact analysis revealed a mean reduction of the D100% of 1.77 Gy ± 3.33 Gy (p = 10 -3 ) and a decreased dose coverage in the prostate on 12 CT c (24.4%). A significant correlation between the changes of prostate dose coverage and the rectum volumetric variations (r = +0.39; p = 0.05) was found, but no relationship with changes in bladder volume (r= +0.02). The dose volume histograms (DVHs) analysis revealed significant mean this new methodology. Material and Methods