ESTRO 37 Abstract book

S992

ESTRO 37

of variation of the RED values, from 0.4 to 0.9, obtained for carbon fiber from the CT due to its thin thickness around the foam. The attenuation of the couch top extension is comparable to the rest of the couch, the average discrepancy between the attenuation from both parts of the couch was 0.242% (0.002% SD). The comparison of the attenuation between measured and TPS dose calculation showed mean and standard deviation of 0.039%±0.002% (within one sigma). The measurements on the buildup region, with the beam passing through the couch, showed a shift in the dose maximum (d max ) of about 4 mm towards the surface, when compared with the d max without the couch. The PDD measurements agreed with TPS calculation within 1%, after applying Gerbi´s method to correct the PDD in the buildup region thus, taking into account the over- response dose of the parallel-plate ionization chamber in this region. Conclusion After the introduction of the new values of RED for carbon fiber of 0.6 and foam 0.05 to the couch template on the TPS, it was verified good agreement between TPS dose calculation and the dose measurements, which shows the couch top is well modeled on the system. EP-1837 Characterization of VMAT plan complexity P.I. Lønne 1 , E. Waldeland 1 , T.P. Hellebust 1 , N.I. Hoven 1 1 Oslo University Hospital, Department of Medical Physics, Oslo, Norway Purpose or Objective In order to ensure precise delivery of dose using volumetric arc therapy (VMAT), our clinic enforces patient specific quality assurance (QA) for the majority of plans. If metrics able to predict the dosimetric accuracy of VMAT plans are identified ahead of plan delivery, treatment quality and safety may be increased while reducing the time used for patient specific QA. In our clinic we are investigating plan complexity through various parameters that may impact dosimetric accuracy, in order to increase the efficiency and quality of VMAT. Material and Methods 105 clinical prostate and prostate bed VMAT plans, with and without pelvic lymph nodes, were evaluated, including 76 dual- and 29 single arc plans. The plans were created using Eclipse V13.6, optimized using PRO11.0 and calculated with AAA11.0. The parameters were extracted with the Eclipse Scripting API V13.6. The metrics presented in this study are the Modulation Complexity Score (MCS) [1] adapted for VMAT plans [2] , Leaf Travel (LT) [2] and the product of these two parameters (LTMCS) [2] . In addition, the average dose rate (DR) vs. average change in dose rate per control point (ΔDR CP ) [3] was calculated for the 76 dual arc plans. Patient specific QA was performed by a Γ analysis, comparing calculated dose from the treatment plan and dose measured by an ArcCHECK (Sun Nuclear, Melbourne, FL) 3D diode array. The plans in the study were delivered by four Varian linacs (three Clinac 2300IX and one Truebeam). Results In this study, no correlation of the gamma pass rate and MCS, LT or LTMCS was observed, indicating a high homogeneity of the plans, as seen in figure 1. The same was observed comparing DR vs ΔDR CP , seen in figure 2, as the majority of plans were found to be clustered, with no clear discrimination between plans with respect to Γ pass rates. Calculated MCS and LTMCS values of the plans indicate a generally high plan complexity, but all plans passed the Γ analysis where 95% of the points were required to have a maximum dose deviation of 3% at a distance below 3mm (Γ(3%, 3mm)). Implementing more stringent requirements, Γ(2%, 2mm) provided a larger range of passing rates, differentiating the plans further. Using this requirement, 81% of the plans had passing rates above 90%. Filtering the plans based on LTMCS, DR and ΔDR CP thresholds, either separately or combined, did

Figure 2. Average ratios of photon Specific Absorbed Fractions between NCINM and OLINDA, over five models, in four different source/target combinations. Conclusion The significant dose differences between our method and OLINDA’s highlighted the inadequacy of the simplistic models for a use in epidemiological investigations. This new dosimetry tool is critical for ongoing and future epidemiological studies: it will allow dosimetrists to conveniently calculate organ doses for patients with lower uncertainty and better assess cancer risks. As the most advanced method of its kind, we anticipate that NCINM will become the go-to nuclear medicine dosimetry tool for research and clinical applications worldwide. EP-1836 Validation of Elekta couch modeling for dose calculation in the Monaco treatment planning system E. Poli 1 , P. Reis 1 , L. Prudencio 1 , J. Galhardas 1 , T. Ribeiro 1 , R. Malveiro 1 1 CHLN - Santa Maria Hospital, Medical Physics Unit, Lisbon, Portugal Purpose or Objective This work intends to validate Elekta iBEAM evo couch top template and determine relative electron density to water (RED) for the different structures of the couch top, namely carbon fiber and foam, to adjust couch modeling in the Monaco v5.0 treatment planning system (TPS) for 6 MV photon beam from a Synergy S, improving dose calculation for oblique beams. The motivation to reanalyze the RED values introduced in the TPS was the implementation of VMAT in the radiotherapy routine. Material and Methods The first step was to obtain a 3D CT image reconstruction of the couch top extension and compare its geometry and the RED values for carbon and foam to the TPS. The methodology used to evaluate the attenuation dose and buildup region was based on AAPM TG-176. The beam attenuation was measured with a Farmer ion chamber 0.6 cm 3 in the center of a cylindrical phantom CIRS 002H9K, for gantry angles from 90 to 180 degrees with intervals of 10 degrees, and normalized to 90 degree gantry angle, where the beam do not pass through the couch. To verify that couch top extension has the same attenuation of the fixed part of the couch top, the attenuation measurements were taken on both parts of the couch. For the measurements in the buildup region, it was used a Markus parallel-plate ionization chamber inside a solid water phantom with and without the couch placed between the 6 MV beam and the ion chamber. The parameters used to calculate dose in the TPS with Monte Carlo Photon algorithm were: grid 0.2 cm, statistical uncertainty 2% per control point. The RED for carbon fiber was 0.6 and for the foam 0.05. The RED values chosen were an approximation of the values obtained from the 3D CT image of the couch top. Results The geometry of the couch top is consistent with the template provided by the vendor. It was noticed a range