P218] Effect of flattening filter mounting surface position in Monte Carlo model of linear accelerators

Abstract

Listen

Translate article

Purpose Monte Carlo (MC) is considered as the gold standard for dose calculations of complex radiotherapy and the accuracy is mainly restricted by the statistical uncertainty. However, accurate MC simulations of radiation transport rely on a detailed description of the linear accelerator. The linac head is modelled in the specific MC code by using a geometrical description provided by the vendor. Interpreting the vendor provided drawings is, however, sometimes subject to uncertainties and vendors are not always willing to answer detail oriented questions. The position of the flattening filter (FF) mounting surface could be one such example. This study aimed at investigating the potential impact of an error in modelling the linac FF mounting surface. Methods Vendor-provided drawings of a modern linac head were interpreted and the 6 MV FF mounting surface was determined to have three possible positions. The positional uncertainty was in the distance from the linac target at which the mounting surface was placed. Discussions with the vendor and local engineers revealed one main candidate to be the correct choice, but nevertheless with doubt. The linac was therefore modelled in the BEAMnrc/EGSnrc user code with the three different geometries. Simulations of the radiation transport through the entire upper static part of the linac head were conducted and beam specific parameters were extracted and compared at a distance of 80 cm from the target. Results Interpreting the vendor provided drawings in three different ways revealed a maximum difference in the 6 MV FF mounting surface position of 1.6 mm from the most likely option. Analysis of the beam energy fluence (function of energy and off-axis position) at 80 cm distance from target revealed small differences (⩽1%). The beam was therefore considered to be independent of the mounting surface position within the distance variations investigated. Conclusions This study demonstrates that MC simulations are only marginally effected by various interpretations of the linac FF mounting surface position if positional differences are as low as 1.6 mm. Nonetheless, thorough and critical study of vendor provided drawings is necessary in order to conduct accurate MC simulations for radiotherapy purposes.