ESTRO 2020 Abstract book

S1094 ESTRO 2020

than 1Gy discrepancy. DMax figures were provided as a supplementary comparison Conclusion There are concerning factors when evaluating DVH data using larger voxels. This study demonstrated over 20% of the studied population have recorded higher dose received by an OAR by up 2Gy using a finer voxel resolution. This translates to a 1 in 5 chances of potential misinterpretation. A possible explanation to this observation could be related to the proximity of the OAR/s to the target volume where smaller voxels will be required to confidently evaluate dose within or near the high gradient zone. Clinicians should exercise higher degree of caution when evaluating DVH data in situation where the target is near any OAR. PO‐1867 Dosimetric study of LINAC‐based SRS/SRT of multiple brain lesions:multiple/single isocenter approach S.W. Tsui 1 , K.S. Yu 1 , L.M. Cheung 1 , S.C. Cheung 2 1 Prince of Wales Hospital, Clinical Oncology, Hong Kong, Hong Kong SAR China ; 2 Queen Mary Hospital, Clinical Oncology, Hong Kong, Hong Kong SAR China Purpose or Objective Linac-based SRS/SRT for multiple brain lesions is typically performed by multiple isocenters. Previous studies have demonstrated comparable dosimetric outcomes and improved efficiency by single-isocenter for simultaneous treatment. However, dosimetric effects after intra- fractional errors induced to both approaches remains uncertain. First, the study compared the dosimetric results, in terms of target coverage, conformity index (CI) and OARs doses, between multiple and single-isocenter approach in SRS/SRT treatment for multiple brain metastases Second, it investigated the effects of intra-fractional errors, whether same magnitude of translational and rotational effects could lead to significant difference For the first part, from 2011-2018, 22 patients with multiple brain metastases were retrospectively recruited. For each site, plans were generated using multiple and single-isocenter by Eclipse. Regarding second part, to simulate the effects of intra- fractional errors, isocenter shifts and rotations were applied to 35 lesions among 17 patients. The plans were recalculated using same beam parameters. Additional plans with 8 types of translational and/or rotational combination errors, including superoinferior, anteroposterior, lateral translations, and roll, yaw and pitch rotations, were generated for each lesion to determine similarities and differences between 2 approaches after the shift. The extent of translational errors within 1 mm and rotational errors within 2 degree were tested. Results With respect to dosimetric comparison as Part 1, no statistically significant differences were found in PTV coverage between multiple- and single-isocenter technique (p-value=0.70). A few OARs doses were resulted in higher doses in single-isocenter plans. For the effect of translational isocenter shift, there were no statistically significant differences between multiple and single-isocenter plans in PTV coverage (p=0.88/0.39 for +1/-1 mm shift) and CI (p=0.24/0.37 for +1/-1 mm shift). Contrary to these, after application of rotational and combined translational and rotational shifts, plans using single-isocenter showed statistically significant weaker performance to target coverage and CI (all p- values <0.01) . While medians of PTV coverage for multiple-isocenter plans could maintain at >99% in 1/2- degree rotation, PTV coverage using single-isocenter decreased from ~98% at 1 degree, to ~95% at 2-degree. The This abstr ct ha been withdrawn between 2 approaches Material and Methods

RW3 and a pinpoint camera; a 3D measurement in 3D with the Octavius 4D phantom (PTW), which allows to perform gamma analysis comparing the calculated and measured doses; and finally portal dosimetry with the tool provided by Varian, which offers a gamma analysis of both measured and calculated fluences.

Results Machine QA: The measurements made are well below the tolerances recommended by the TG-142 for SBRT treatments. Patient QA: The table shows the results of the QC of 100 treatment plans for SBRT with VMAT.

Conclusion Based on the results obtained, the tolerances of the QA for the treatment plans are established by applying a level of confidence of 95% and 99% to our data. Only in one of the 100 analyzed patients the results of the QA were not within the previous level of confidence and therefore a new treatment plan was made. In view of the results we are considering simplify ourQA procedure for future treatments of SBRT. PO‐1866 Comparison of dose evaluation using difference voxel sizes in Stereotactic Radiotherapy Planning T.L. LAI 1 , E. Tsang 2 1 GensisCare Australia, WROC, Melbourne, Australia ; 2 GensisCare Australia, Epping ROC, Melbourne, Australia Purpose or Objective Genesis Care (GC) Australia has developed a national stereotactic team responsible for planning stereotactic brain radiotherapy using Monaco TPS for all Elekta based centres across Australia. All reporting for through Monaco is done using a 0.1cc volumetric matrix. GC Victoria provides stereotactic radiotherapy (SRT) across all sites using Pinnacle TPS and Varian Truebeam linear accelerators. Pinnacle TPS allow dose volume to be reported to as small as 0.01cc. Many international clinical trial report doses to 0.03cc whilst the AAPM Task Group 101 (2010) defined “point” as 0.035cc or less. This study aims to evaluate whether there are major differences in OARs dose reporting between voxel resolution sizes of 0.1cc Vs 0.03cc Material and Methods 22 SRT brain patients planned using Pinnacle TPS were reviewed retrospectively. A script was created to report OAR using dose to 0.1cc (D0.1cc), dose to 0.03cc (D0.03cc) and maximum point dose (DMax). DMax in this case is defined as the maximum dose reported on the dose volume histogram in Pinnacle. The OARs reported include the left and right optic nerves, optic chiasm, left and right lens and brainstem. Data are entered onto spreadsheet for comparison. Results Out of 22 samples, 5 (22.7%) showed 1 to 2Gy discrepancy for at least one OAR and 3 (13.6%) showed larger than 2Gy difference between values obtained using D0.1cc and D0.03cc voxels. The reminder of the cohort showed less