ESTRO 2020 Abstract book

S793 ESTRO 2020

tolerance result. There were 10 occasions where MPC detected a true out-of-tolerance event on a machine.

rest of the teeth, mandible and maxilla received mean doses <50Gy. Map 2: A 61 year old male with pT2N2b left oropharyngeal squamous cell carcinoma (SCC). He was treated adjuvantly with 66Gy to high-risk regions, 60Gy to left levels II-V, and 56Gy to right levels II-V. No tooth or region of maxilla or mandible received a mean dose of >50Gy. Conclusion We believe this to be the first study generating dental radiation dosimetric maps of estimated doses to each tooth and each third of the mandible and the maxilla for common examples of head and neck cancer faced by radiation oncologists. Adoption of these dental maps may help improve clinical work flow efficiency for dental providers and radiation oncologists by providing estimates of dental doses for common primary head and neck cancers, both for patients pending future radiation treatment and for previously irradiated patients when radiation plans are not available. PO-1402 Conventional Linac QC v Automated Linac QC: 3 years of experience of daily Varian MPC M. Pearson 1 , S. Mission-Yates 1 , D. Eaton 1 , T. Greener 1 1 Guys and St Thomas', Medical Physics, London, United Kingdom Purpose or Objective Machine Performance Check (MPC) is automated QC tool that is integrated into the Varian Truebeam. MPC utilises the Electronic Portal Imaging Device to measure a number of parameters relating to the beam and geometry of the Linac. This work compares the MPC results with conventional testing results over a period of 3 years for 8 Linacs in order to assess MPC’s ability at detecting faults and therefore the suitability of MPC as a QC tool. Material and Methods MPC was run daily for all 8 Varian Truebeam (v2.5) Linacs at the Authors’ centre. The MPC output measurements were compared with the monthly chamber measurements over the 3 year period. The 6X Beam and Geometry results have been compared with the results of the conventional QC testing program. The conventional QC testing program of daily/monthly/annual QC was recorded in a QATrack+ database and followed guidance outlined in AAPM TG142 and IPEM report 81v2. The QC records in MPC and QATrack+ were analysed to investigate the correlation of out-of-tolerance results for MPC and conventional QC. All the conventional monthly tests results in QATrack that are comparable to the checks carried out by MPC were analysed. The QC tests were: Symmetry measured with the PTW 1500 array, Light-Radiation coincidence, VMAT MLC Picket fence, collimator/gantry/couch calibration scales and MV-kV isocentre using Isoverification. All the MPC test failures were then further analysed to determine where out-of-tolerance MPC results had occurred and compared with QATrack+ to investigate the cause of any failures. Results There was a total of approximately 6000 MPC measurements test across all the machines. MPC outputs were within 2.8% of the monthly chamber measurements over the 3 year period, with 2.3% of measurements > 1.5%, (MPC was rebaselined if the output differed by more than 1.5%). The MPC v chamber output is shown in the Figure, and the agreement has improved over the 3 years. For the remaining QC tests, there were a 4 occasions where the conventional QC tests were out-of- tolerance. These QC failures were for radiation field size, beam symmetry, and kV and MV isocentres tests. MPC detected all these failures, with an out-of-tolerance result for the relevant test. The maximum failure rate of all MPC tests was 2.5% (BeamOutputChange), the next highest was 1.1% (BeamCenterShift), and are 0.6% or less for the remaining parameters. There were 6 occasions where the repeat MPC measurement cleared the particular out-of- This abstract has been withdrawn

Conclusion The analysis of 3 years of conventional QATrack+ records against the MPC 6X Beam and Geometry results has demonstrated that MPC is a robust and sensitive method of performing beam and mechanical checks in a clinical setting. PO-1403 FMEA of a Linac QC program based on Varian MPC, DailyQA3, Delta4 and Conventional QC checks. PO-1404 Safe number of transfers between Truebeam & Clinac for different treatment sites without replanning T. Newbold 1 1 Poole Hospital NHS foundation trust, Radiotherapy, Poole, United Kingdom Purpose or Objective Our department currently has one TrueBeam (TB) linac and two iX linacs. The two iX linacs are matched but the iX & TB linacs are not matched; the beam models for iX and TB in the Eclipse planning system are similar, though not identical. Having a single TB linac represents a single point of failure for which a contingency, other than a total re-plan for an iX linac, is desirable. Having a similar beam model suggests that it may be possible to treat a number of fractions on the ‘wrong’ machine without leading to detrimental patient dosimetry. The aim of this project is to ascertain whether treatment planned for one type of linac is deliverable on the other, and if so, for each treatment site, the safe number of fractions for which it is safe to transfer patients without re-planning the treatment. Material and Methods Treatment planning is performed in Eclipse. Plans originally produced for one type of linac were recalculated for the other by changing the beam model and fixing the MUs to the original value. Reported MLC errors were ignored. Appropriate DVH parameters were compared between the original plan and the same plan on the alternative machine. A predicted dose to the ArcCheck was calculated in Eclipse from the original plan. The original treatment plan was delivered on both linacs and an ArcCheck measurement taken. Gamma analysis was used to compare to the predicted and measured dose distributions for both deliveries. Results Analysis of the data obtained enabled recommendations for the maximum number of unplanned machine transfer fractions to be suggested. Outcome-based options for each site were given to clinicians for consideration. Table 1 shows the outcome based options for transfer number. Abstract withdrawn