ESTRO 38 Abstract book

S1103 ESTRO 38

on-line reconstruction software Symmetry (Elekta). For each patient, baseline shifts were obtained by the difference between bone and ROI registrations. Mean tumor motion amplitudes were estimated from the 4D-CT by subtracting the 3D tumor centroid coordinates on extreme breathing phases, and automatically from 4D- CBCT images. A physician delineated the ITV on average images (4D-CT and 4D-CBCT) using a MonacoSim workstation (Elekta) to analyze the variations of the ITV A total of 280 4D-CBCT images from 58 tumors (52 patients) were retrospectively analyzed. The repartition of the tumors in the lungs was: 53% of tumors located in the upper lobe (UL), 9% in the middle lobe (ML) and 38% of tumors located in the inferior lobe (IL). The grand mean (GM) baseline shifts (SD) in the cranio caudal (CC) direction were 0.16 cm (0.12) for UL, 0.53 cm (0.26) for ML and 0.31 cm (0.28) for IL. In the other directions, GM baseline shifts were under 0.21 cm. 13% of UL tumors, 100% of ML tumors and 67% of IL tumors had a baseline shift above 0.30 cm in the CC direction. The mean [range] of amplitude differences between 4D- CT and 4D-CBCT was less than 0.10 cm [-0.93 cm; 1.70 cm] in all directions. 27% of UL tumors, 40% of ML tumors and 52% of IL tumors had amplitude differences larger than 0.30 cm in all directions. The median [range] of ITV volume deviation between 4D-CT and 4D-CBCT was -1.0% [-196.0%;+82.7%]. Conclusion The largest baseline shifts were observed for tumors located in the middle and inferior lobes. After a review of these cases, those tumors were mainly located on the fissure (ML), close to the chestwall, or in areas influenced by gastric filling (left IL). The same observation was found for amplitude variations. The variability of ITV volume between 4D-CT and 4D-CBCT was, in some cases, large and mainly due to delineation difficulties on average CBCT (image quality or atelectasis) or breathing artefacts (duplication) on the 4D-CT. EP-2014 Decision Support System for Checking Online Adaptive Treatments on the Elekta Unity D. McQuaid 1 , R. Niliwar 1 , J. Mohajer 1 , E. Goodwin 1 , S. Nill 1 , U. Oelfke 1 1 The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, Joint Department of Physics, Sutton, United Kingdom Purpose or Objective Online adaptive treatment plans for the Elekta Unity (Elekta AB, Stockholm Sweden) are produced by an integrated workflow involving the Monaco treatment planning system (TPS). The workflow limits on the online planning options and the automated tools to assess the plan against dose constraints reduce the possibility of errors. However, extra quality assurance of the online generated plans including an independent MU check adds confidence that the treatment plan is correct and optimal. The time taken to perform these tests is critical for the efficient execution of the treatment. A method of performing these tests quickly and effectively is presented. Material and Methods The checking system consists of two redundant components: i) a standalone tool developed in C++, and ii) a python script called from within the RayStation TPS (RaySearch, Stockholm, Sweden). Both parts of the system were developed separately by different teams to a joint specification and tests are run simultaneously by both tools so reducing the risk of a false negative test. The results are combined into a single user interface where they are presented as a traffic light system. Settings and test tolerances are defined for each treatment site and dose prescription. The system reads volume. Results

original plan. Single fraction treatment time duration, including also the fraction with re-optimised dose distribution, was 8.8±2.8min : 4.7/17.3min. When the plan underwent re-optimisation, a reduction of the beam-on time in comparison with the original plan has been detected in 44% of the cases (15 plans). Online QA 3%/3mm gamma passing rate was 97.5±1.5% (94.8/100.0%), comparable with the offline QA performed on a dedicated phantom (98.7±1.1% (96.7/100.0%)). DVH metrics results are shown in Table1 for a subgroup of 6 (pts) with same dose per fraction prescription (8 Gy) . Maximum single fraction CTV to duodenum/stomach centre of mass distance was 2.4cm and 1.8cm respectively; SD of CTV to duodenum/stomach centre of mass distance for all the evaluated 50 fractions is 0.9cm and 0.8cm respectively.

Conclusion The results of this evaluation emphasize that different parameters can affect the entire workflow. Most of them can be improved to optimise the current workflow in order to better perform the OA strategy. EP-2013 Lung tumor motion based on 4D-CBCT: baseline shift, interfraction amplitude and volume variation F. Oger 1 , P. Dupuis 1 , E. Mesny 1 , T. Baudier 2 , S. Rit 2 , R. Tanguy 1 , M. Ayadi 1 1 Léon Bérard Cancer Center, Radiotherapy, Lyon, France ; 2 INSA-Lyon- Université Lyon 1, CREATIS- CNRS UMR5220- Inserm U1044, Lyon, France Purpose or Objective 4D-Cone Beam Computed Tomography (CBCT) allows the accurate positioning of moving targets and the study of tumor motion. In this work based on clinical data, we first compared the Internal Target Volume (ITV) volumes and motion amplitude between 4D-Computed Tomography (4D-CT) and 4D-CBCT. Secondly, we estimated the interfraction baseline shift and motion amplitude variations. Material and Methods Patients with early non-small cell lung cancer treated with SBRT were included in the study. Patients were installed in a Bodyfix (Elekta) and underwent a 4D-CT scan (Big Bore, Philips). Treatment plans based on an ITV strategy were delivered on a Versa HD linac (Elekta). Before each fraction, a 4D-CBCT was acquired. A bone registration followed by an automatic region-of-interest (ROI) registration of the tumor were then computed with the