ESTRO 37 Abstract book

S1215

ESTRO 37

EP-2194 Analysis of treatment planning feasibility of the multinational end-to-end IMRT audit methodology P. Kazantsev 1 , V. Hernandez 2 , L. Luketin 1 , J. Izewska 1 1 International Atomic Energy Agency, Dosimetry and Medical Radiation Physics Section, Vienna, Austria 2 Hospital Sant Joan, Servei de Física, Reus, Spain Purpose or Objective The purpose of this study was to analyse the feasibility of the treatment planning part of the newly developed multinational end-to-end IMRT audit methodology. The methodology utilizes a “Shoulders, Head And Neck, End- to-end” anthropomorphic phantom (SHANE, CIRS Inc.) which passes all the stages of a routine IMRT treatment from imaging to irradiation. During the pilot testing of the methodology, its treatment planning stage was noted to be considerably challenging and time-consuming and it therefore required a deeper check within a broader group of participants. Material and Methods Participants of this study were supplied with the CT dataset and the structures of the SHANE phantom as well as the set of dose constraints and asked to prepare a clinically acceptable IMRT plan, recording the time needed to create it. The plans had to pass the local pre- treatment QA and then be submitted together with the questionnaire for review. Several participants were also asked to submit the results of ion chamber measurements in the SHANE phantom. The submitted plans were analysed in terms of quality (using PlanIQ, Sun Nuclear Corp.) as well as complexity (using the in-house software PlanAnalyser to compute several reported complexity indices). There were 16 treatment planning constraints for PTVs and OARs which were checked using linear function metrics so the maximum plan quality score was 100. Complexity indices depended on the IMRT technique and were evaluated as predictors of potential discrepancies between dose calculations and measurements. Results Participants from 12 institutions submitted 42 plans; in terms of IMRT delivery modes, 21 plans were Sliding window, 12 – IMAT, 6 – Step and Shoot, 3 – Tomotherapy. 11 plans were submitted together with ion chamber measurement results. The pre-treatment QA was performed by different institutions using both measurement and computational approaches. Only in two cases the QA results with portal dosimetry were unsatisfactory with gamma passing rates (3%/3mm, global gamma, 20% threshold) just under 95%. The planning time varied from 1 to 12 hours with the average value of 3.5 hours. About 80% of the plans achieved a quality score >90 with 9 plans fully satisfying the planning constraints. The unsatisfactory plans were related to the experience of the planner or the quality of the beam model. There was no correlation between planning time, plan quality score, plan complexity and pre-treatment QA results. Nevertheless, for those plans accompanied with ion chamber measurement results it was clearly observed that increased number of MUs leads to larger discrepancies in point doses. Conclusion The treatment planning stage of the multinational end- to-end IMRT audit methodology is feasible in terms of achieving a clinically acceptable plan within a reasonable period of time of less than 4 hours. More complex, overmodulated plans with greater number of MUs are prone to errors during treatment delivery. EP-2195 Feasibility analysis of a novel off-line approach for MR-guided radiotherapy T. Bostel 1 , A. Pfaffenberger 2 , M. Stoll 2 , P. Haering 2 , M. Splinter 2 , C. Lang 2 , G. Echner 2 , S. Delorme 3 , F. Sterzing 1 , P.E. Huber 4 , J. Debus 1 , N.H. Nicolay 1

Ten patients, prescribed 78 Gy in 39 fractions to the prostate, underwent an MR-only workflow. All preparation, usually performed in connection to the CT, i.e. patient information, fixation, tattoo etc., was performed at the MR. MRI acquisition protocol included sequences for target delineation, synthetic CT (sCT) generation (MriPlanner TM ) and fiducial marker identification. Total time for dedicated MR-only acquisition sequences was 18 minutes. Automatic check of MRI acquisition parameters was done to assure compliance with the defined MR-protocol. With a dedicated fiducial marker sequence, markers were identified directly on the images intended for sCT generation. A treatment plan was created using the sCT according to clinical practice. To evaluate the performance of our MR-only workflow a CT was acquired directly after the MR scan. This enabled general quality checks of the generated sCT, including dose recalculations on the CT and verification of the identified fiducial markers. This also ensured no delay of treatment in case of exclusion from the study. Patient positioning was done based on sCT-DRRs with synthetic markers. The patient positioning method was verified by simulation of the corresponding positioning based on CT references. A cone-beam CT (CBCT) dedicated for QA development was acquired prior to treatment. Results Nine patients successfully received an MR-only treatment. One patient was too obese for the sCT sequence FOV and was therefore excluded. Marker identification and target delineation was performed successfully using the sCT images. Mean dose deviation between sCT and CT treatment plans were <0.8% for target volumes and risk organs. Markers were identified with a maximum deviation of 4 mm total distance between two markers. Identified marker positions were used to generate synthetic markers visible in sCT-DRR. Patient positioning was successfully performed with a difference of maximum 1.6 mm compared to CT-based patient positioning, mean difference 0.0 mm (s.d 0.6 mm). Conclusion A successful MR-only implementation relies on a fine detailed work plan with tasks to be carried out in a well- defined order. We have presented a method for implementation of an MR-only workflow and shown that the CT could successfully be entirely excluded. Future development will include QA procedures for marker identification and HU-control with CBCT. This study was conducted within the national consortium "Gentle Radiotherapy”.