ESTRO 38 Abstract book

S972 ESTRO 38

ranged between 0.2 and 2.7 cm. The Acuros dose calculation algorithm was used after optimization with a dose grid resolution of 1.25 mm. Patient-specific measurements were carried out in the CIRS film QA phantom with GafChromic EBT3 film (Ashland Specialty Ingredients) located centrally in the frontal slice of the GTV and used for evaluation. The film calibration was cross-validated with ionization chamber (measurements) according to the IAEA TRS-398 protocol prior to measurements. Fine-tuning of the calculation model was performed by varying the dosimetric leaf separation (dls) in the algorithm to best fit the measured dose distribution by visually inspection of the dose gradient together with gamma evaluation. A gamma criterion of 5%/1mm and 2%/2mm local normalization was used. Results The mean fraction of γ 5%,1 mm ≤1 and of γ 2%,2 mm ≤1 was 94.5 % and 93.8 %, respectively, for all measurements after modifying the dls in the beam model. An adjustment of +0.5 mm was carried out on the dls parameter compared to the measured value at the machine, which corresponds to an increase in the effective MLC field size. Before dls adjustment, the measured maximum dose deviated around 8% as compared to the planned dose. For the case with smallest target size (0.2 cm), an overestimation of about 4% in the maximum calculated dose as compared to measurements was observed. A tolerance of 5 mm in minimum equivalent sphere diameter of the PTV has therefor been introduced for treatment of SRS with the millennium MLC. Conclusion For accurate modeling of the beam penumbra, which may have influence on the critical organ dosimetry, adjustments are warranted on the MLC modeling in the calculation model. High spatial resolution measurements therefore play an essential part in a safe delivery of HyperArc™. EP-1796 Comparison of Treatment Planning Systems’ shallow depth dose prediction for IMRT E. Bogaert 1 , G. Pittomvils 1 , C. De Wagter 1 1 Ghent University Hospital, Radiotherapy and Oncology, Gent, Belgium Purpose or Objective Validation of a Treatment Planning System (TPS) for IMRT is well described e.g. NCS 22 report [1], but challenging. Dose contribution from outside IMRT segments’ edges needs to be predicted correctly. QA patterns like Inverse Pyramids [2] are recommended at reference depths of 10 and 5 cm. However, at shallow depths primary scattered electron contamination causes head scatter conditions to differ. During photon beam modelling in RayStation v6.1 a rather high electron contribution was found to be necessary for modelling large open fields’ depth dose curves. Our goal was to validate the need of this higher electron contribution by assessing the effects at shallow depths and by comparing RayStation planned doses to those retrieved from the two other TPS in use at our department. Material and Methods For a Varian Clinac-IX (Millennium MLC; 6 and 18 MV), beam models were available in RayStation v6.1 (RaySearch Laboratories) and Eclipse v8.1 (Varian), whereas for Elekta Synergy (MLCi2; 6 and 15MV) models were available in RayStation v6.1 and Pinnacle v9.8 (Elekta). Predicted line dose profiles were evaluated against those measured in a polystyrene slab phantom with a PTW linear array LA48. Calculation grids were 2x2x2mm 3 for Synergy and 2.5x2.5x2.5mm 3 for Clinac-IX. The effective point of measurement (0.75mm below surface) was taken into account in all profiles. A local criterion of 3%/3mm (max dose) was used. For high energies and 1 cm depths, the effect of 0.5 mm shift in depth on the position for

created are not biased by measurement uncertainty and have the same resolution as real measurements. Artificial measurement files were compared to the clinical (original) verification plan using gamma analysis (Verisoft v.6.1) with criteria: 1%/1mm, 2%/2mm, 3%/3mm both for local and max gamma. Dose threshold was set to 5%. DVHs for plans with introduced errors were also analyzed in order to find: 2% change in PTV mean dose, PTV D98%<95% and PTV D2%>107% (considered as clinically relevant). Results Wilcoxon signed rank test was used to perform statistical comparison between array, p-value less than 0.005 was considered as significant. Significant difference between 729 and 1500 array was observed for most of the error types for 3%/3mm and 2%/2mm gamma analysis. It might be explained by impact of interpolated points included in evaluation. There is no significant differences between 1000SRS and 729 array for most of plans with shift error. However significant difference is noticed for gap width error showing grater capability for 1000SRS to detect plan with error (Fig. 1). The MLC errors without clinically relevant effect are marked in the Figure.

Conclusion Analysis results for array 1500 and 1000SRS with Verisoft 6.1 may lead to false negative result (plan rejection with correct linac performance) due to including into gamma analysis not only measurement points but interpolated also. Higher detector resolution matters for gap width error but has no effect in case of systematic shift in MLC. The 1000SRS array, although dedicated to verification SRS plans, can also be used for larger PTV plans with high dose gradient even if it does no cover all irradiation area. EP-1795 HyperArc™ commissioning necessitates high- resolution measurements M. Sjölin 1 , J.M. Edmund 1 1 University Hospital Herlev, Oncology, Herlev, Denmark Purpose or Objective An integrated solution for linac-based stereotactic radiosurgery (SRS) has recently been introduced by Varian Medical Systems with an automated multiple non-coplanar arc beam arrangement. This technique calculates a high dosimetric target conformity with a rapid dose fall-off to critical organs and therefore requires carefully dosimetric evaluation. The purpose of this study was to experimentally establish the accuracy of the dose modeling performed in the Eclipse treatment planning system (TPS) of this small target dosimetry technique manifested by high resolution measurements. Material and Methods A set of 10 HyperArc™ treatment plans with single and multiple metastasis were optimized in Eclipse (v. 15.6) with the millennium MLC and 6 MV flattening filter free (FFF) beamlets. The planning target volume (PTV) measured as the effective equivalent sphere diameter,