ESTRO 36 Abstract Book

S477 ESTRO 36 _______________________________________________________________________________________________

GTV position. Purpose of this study is to investigate the impact of a varying rectum filling on tumor position and quantify potential tumor shifts. Material and Methods For the analysis, nine patients were included who were scanned twice on MRI in supine position. First on a 1.5 T MRI for diagnostic purposes and next on a 3T MRI for treatment planning. For the diagnostic MRI, the rectum was filled using an ultra sound transducer gel (MRI full ), and for the planning MRI no rectal preparation was performed (MRI standard ). On both MRIs the tumor was delineated. To evaluate tumor displacement, for both MRI standard and MRI full , three distances in cranial-caudal (CC) direction were determined between the bony anatomy; i.e. the sacrum promontory and the tumor cranial border, the tumor caudal border and the center of mass (COM), (figure 1, C & D). For each distance measure, displacements were then determined by taking the difference in distance between both MRI scans. Results In all patients a shift in tumor COM in CC direction was observed, ranging between 6.9 and 28.3 mm. Mean tumor displacements between MRI standard and MRI full were found to be 16.7 mm, 16.5 mm and 17.7 mm for the cranial and caudal tumor border and the COM, respectively (figure 1 C & D). Displacements were all found to be significantly different from zero (p<0.002 for all distance measures). Displacement was larger for tumors situated higher up in the rectum (figure 2). Conclusion In all patients, tumor position changes considerably under influence of rectal filling. The found mean displacements are larger than the typical PTV-margins for rectal GTV (Brierley et. al 2011). The higher situated rectal tumors show the largest displacements under influence of rectal filling. To avoid geometrical miss of the tumor, rectal volume preparation prior to boost radiotherapy or adaptive RT with online tumor visualization using MRI (Lagendijk et al. 2008) seems beneficial. Especially for tumors located high in the rectum.

Poster: Physics track: Adaptive radiotherapy for inter- fraction motion management

PO-0875 Dosimetric effects of anatomical changes in proton therapy of head and neck (H&N) cancer G. Miori 1,2 , L. WIdesott 1 , F. Fracchiolla 1 , S. Lorentini 1 , P. Farace 1 , R. Righetto 1 , C. Algranati 1 , M. Schwarz 1,3 1 Trento Hospital, Protontherapy, Trento, Italy 2 University of Rome Tor Vergata, Postgraduate School of Purpose or Objective Anatomical changes in H&N patients can affect dose distributions especially in proton therapy. A retrospective analysis of H&N patients undergoing repeat CTs and treated at our Proton Therapy Center was done to evaluate dose changes and to identify a dosimetric index for the need of replanning. Furthermore, TCP analysis was performed to evaluate the magnitude of changes with radiobiological parameters. Finally, non-adapted and adapted plans were compared. Material and Methods All H&N patients treated in our center between October 2014 and September 2016 with at least one repeat CT (eCT) were considered. 21 patients were identified: 18 patients had at least one eCT (1 to 6 eCTs), but did not need replanning, and 3 patients needed replanning at some stage of the treatment. The original plan was recalculated on each eCT. Differences were calculated for each treatment fraction, considering a stepwedge interpolation on fractions where the eCT was missing. D1 variations (ΔD1) for cord, brainstem, optic chiasm and optic nerves, and Dmax differences (ΔDmax) for lenses were considered. Target coverage analysis was based on differences in CTV V95 (ΔV95). ΔV95 values were included in Non-replanned ( controls ) if they came from non- replanned patients or from replanned patient calculations on CT preceding the replanning CT (rCT). On the contrary, ΔV95 were included in Replanned ( cases ) if they came from replanned patients on the rCT and the following CTs. The choice was made to consider the trend in target coverage after the point identified for replanning. A cut- off ΔV95 for the need of replanning was identified by the maximum Youden’s index on the ROC analysis between control and cases. Next, TCP differences with respect to the planning TCP (ΔTCP) were calculated. ΔTCP values were divided in Non-replanned and Replanned as for DV95 analysis. Finally, a comparison between adapted and non- adapted plans for the 3 replanned patients was done. All statistics were made by t-Student tests. Medical Physics, Rome, Italy 3 INFN, TIFPA, Trento, Italy