ESTRO 37 Abstract book

S204

ESTRO 37

RFS, MFS, SS and OS rates according to GAC are reported in Table 1. In UVA, vascular embolisms (p = 0.008) and GAC (p = 0.001) were considered as significant prognostic factors for 3 rd ILBTE, while, in MVA, GAC (p = 0.008) was the only prognostic factor.

depths). This was examined with Monte Carlo to understand the consistency of this difference. Results The AAPM group concluded that dose should be reported to muscle because it is the more accurate description of soft tissue, it is what substantial clinical trial data and patient outcomes is based upon, and it is the necessary direction in which to move to improve dose calculation accuracy in the future. To implement a dose-to-muscle framework, the ideal solution is calibration of the linac in water, and having the treatment planning system inherently account for tissue differences to calculate the dose accurately in the patient. When this solution is not possible (i.e., when the TPS does not calculate the dose to muscle in the patient), the dose is overestimated by 1%. This 1% overestimation was found to be broadly accurate (ranging from 0.6% - 1.4%) for different nominal photon and electron beam energies, different depths, and different field sizes. Conclusion Dose to muscle should be reported by the treatment planning system under reference conditions in tissue. If it does not, a multiplicative 0.99 correction factor can be applied to the reference specification in order to approximately achieve the same result as it is generally accurate across all treatment energies, photons and electrons, and depths and field sizes. Understanding which algorithms inherently calculate dose to medium versus dose to water is a complex question that must be understood before final recommendations and implementation strategies can be defined. These issues are explored in the companion abstract. OC-0405 Monte Carlo based Quality Assurance of Base Data for Beam Modeling in Treatment Planing Systems M. Kowatsch 1 , M. Söhn 2 , M. Alber 2,3 1 LKH Feldkirch, Institut of Medical Physics, Feldkirch, Austria 2 Scientific-RT, Scientific-RT, Munich, Germany 3 Universitätsklinikum Heidelberg, Klinik für RadioOnkologie und Strahlentherapie, Heidelberg, Germany Purpose or Objective Errors in beam base data (BBD) can lead to flawed beam models for the treatment planning system and thus to systematic dose computation errors. Despite a number of sophisticated guidelines, systematic BBD errors are difficult to spot and eliminate. An appropriate quality assurance (QA) of BBD would thus be an essential cornerstone to an overall high dosimetric accuracy. Here, we investigate the use of Monte Carlo (MC) simulation to verify self-consistency and physical plausibility of BBD. Material and Methods MC simulations of various field configurations offer the advantage of being inherently self-consistent if they utilize an invariant description of the phase space (PS) above the variable collimator system. To generate such a PS, relatively few measurements are required, that can be chosen to offer optimum measurement conditions regarding field size and depth of measurement, and thereby consistent beam quality, i.e. particle spectra. Using SciMoCa (Scientific RT, Munich), a newly developed virtual MC source model with a unique commissioning method relying on pre-generated BEAMnrc template beam models (BM), 62 BBD sets of distinct beam qualities from 15 different institutions, a mix of Varian, Elekta, Siemens and Cyberknife, were analyzed. For each BBD, a SciMoCa BM was built. For analyzing the BBD, simulated depth dose curves (DDC), output factors (OF) and cross profiles were compared against the original BBD. Results The BM achieve a high accuracy of 0.2% std. dev. in abs. dose for fields ranging from 10x10 mm² up to 400x400 mm² if derived from high quality BBD and even +-1.0% if

Conclusion In case of 2 nd ILBTE, GAC could be used as a helping- decision tool to discuss conservative or radical treatment options.

Proffered Papers: PH 7: Dose measurement and dose calculations

OC-0404 Rationale for AAPM recommendations on medium for TPS reference dose specification S. Kry 1 , V. Feygelman 2 , P. Balter 1 , T. Knoos 3 , C. Ma 4 , M. Snyder 5 , B. Tonner 6 , O. Vassiliev 1 1 UT MD Anderson Cancer Center, Radiation Physcis, Houston- TX, USA 2 Moffitt Cancer Center, Radiation Oncology, Tampa, USA 3 Skane University Hospital, Radiation Oncology, Lund, Sweden 4 Fox Chase Cancer Center, Radiation Oncology, Philadelphia, USA 5 Wayne State University, Radiation Oncology, Detroit, USA 6 Ackerman Cancer Center, Radiation Oncology, Fernandia Beach, USA Purpose or Objective Reference dose calibration has be done as “dose-to- water” or “dose-to-muscle”. At present there is no consistency in the community about which is ideal or how the solution should be clinically implemented. Consequently, there is an additional +/- 1% spread in different reference calibrations. While this is a small uncertainty, it impacts calibration and systematically affects the dose delivered to all patients. As such, a consistent and reasonable approach should be undertaken to manage this question. The American Association of Physicists in Medicine convened a group to review the medium for reference dose calibration. This abstract highlights the thought-process and supporting data on which is the ideal medium for reference dose calibration. How this should be implemented in clinical practice is the subject of a companion abstract. Material and Methods The AAPM group evaluated three issues: First, whether dose to soft tissue in a patient should be conceptually described as dose-to-muscle or dose-to-water. Second, what the most logical workflow was for converting between calibration in water and dose calculation in patient tissue. Third, the magnitude of this difference between dose-to-muscle and dose-to-water (with the same electronic density as muscle) for different photon and electron energies (specifically: different nominal beam energies, different field sizes, and different