ESTRO 37 Abstract book

S989

ESTRO 37

criteria G(ΔD, DTA) < 1. Threshold criteria used were: Dose difference ΔD = 3%, 2%, 1%, and distance to agreement DTA = 3 mm, 2 mm, 1 mm. Furthermore, the mean dose deviation and SD were calculated. Results

After the shielding disk great experimental errors are due to a low sensitivity of radiochromic films. MC results are so the most reliable showing a good shielding properties for all solutions. Conclusion MC simulations play a fundamental role in IORT dosimetric studies. Shielding disk for breast cancer therapy have to take into account backscattering effects: wrong positioning of asymmetric disks leads to 10% increase in dose to the target. EP-1832 Validation of Acuros XB dose calculation algorithm with Monte Carlo for clinical treatment plans U.V. Elstrøm 1 , M. Alber 1,2,3 , M. Söhn 3 , L. Hoffmann 1 1 Aarhus University Hospital, Department of Medical Physics, Aarhus C, Denmark 2 Heidelberg University, Clinic for Radiation Oncology, Heidelberg, Germany 3 Scientific RT, Scientific RT GmbH, Munich, Germany Purpose or Objective Modeling of patient outcome data after RT is dependent on the precision of the dosimetric data. Inaccuracies originate from CT calibration and dose calculation, whereby especially modeling of the lateral electron dose deposition formerly has been inaccurate. The Acuros XB dose calculation algorithm integrated in the Eclipse treatment planning system (TPS) (VMS, Palo Alto) is based on solution of the linear Boltzmann transport equation. The purpose of this study was to benchmark Acuros against Monte Carlo (MC) calculations, as implemented in SciMoCa (Scientific RT, Munich), for a variety of clinical treatment plans. Material and Methods Twenty five treatment plans were selected from the clinical data base. The plans included four anatomical sites; head and neck, lung (hypo and normo fractionation), cervix and rectum, and the delivery techniques used were 3D conformal, IMRT and VMAT. The patients were treated at a Varian Clinac (VMS) equipped with a 120 leaf collimator (MLC). All plans were created in the Eclipse TPS v.13.7 and dose to medium was used for the calculation with Acuros. SciMoCa was setup based on measured cross profiles, depth dose curves, absolute calibration, dynamic leaf gap and inter-leaf-leakage of MLC to accurately reproduce the output of a Clinac. All treatment plans were exported to SciMoCa and the dose was recalculated using the MLC movement, jaw position and monitor units of the original plan, to an uncertainty of 0.5%. The 3D dose distributions from Acuros and MC were compared using Gamma evaluation (with Acuros as reference). The Gamma agreement index (GAI) was calculated as the percentage of points fulfilling the

In Fig 1, radar plots are shown for the full body and PTV for each of the anatomical regions. Very similar values of G(3,3) and G(2,2) are seen for all patients. The lowest value of G(2,2) for the body evaluation was seen in a Cervix patient. For this group of patients the range of G(2,2) was [97.0;99.4%]. The highest values of G(2,2) was seen for hypoLung patients where the range of G(2,2) was [99.9;100%]. In 80% of the patients, G(2,2) ≥ 99.0% for the body. Values for G(1,1) varies among the patients. The lowest values were observed for H&N and Cervix.