Monte Carlo simulation of TrueBeam flattening-filter-free beams using Varian phase-space files: Comparison with experimental data

Purpose: Phase-space files for Monte Carlo simulation of the Varian TrueBeam beams have been made available by Varian. The aim of this study is to evaluate the accuracy of the distributed phase-space files for flattening filter free (FFF) beams, against experimental measurements from ten TrueBeam Linacs. Methods: The phase-space files have been used as input in PRIMO, a recently released Monte Carlo program based on the PENELOPE code. Simulations of 6 and 10 MV FFF were computed in a virtual water phantom for field sizes 3 Chi 3, 6 Chi 6, and 10 Chi 10 cm(2) using 1 Chi 1 Chi Chi 1 mm(3) voxels and for 20 Chi 20 and 40 Chi 40 cm(2) with 2 Chi 2 Chi 2 mm(3) voxels. The particles contained in the initial phase-space files were transported downstream to a plane just above the phantom surface, where a subsequent phase-space file was tallied. Particles were transported downstream this second phase-space file to the water phantom. Experimental data consisted of depth doses and profiles at five different depths acquired at SSD = 100 cm (seven datasets) and SSD = 90 cm (three datasets). Simulations and experimental data were compared in terms of dose difference. Gamma analysis was also performed using 1%, 1 mm and 2%, 2 mm criteria of dose-difference and distance-to-agreement, respectively. Additionally, the parameters characterizing the dose profiles of unflattened beams were evaluated for both measurements and simulations. Results: Analysis of depth dose curves showed that dose differences increased with increasing field size and depth; this effect might be partly motivated due to an underestimation of the primary beam energy used to compute the phase-space files. Average dose differences reached 1% for the largest field size. Lateral profiles presented dose differences well within 1% for fields up to 20 Chi 20 cm(2), while the discrepancy increased toward 2% in the 40 Chi 40 cm(2) cases. Gamma analysis resulted in an agreement of 100% when a 2%, 2 mm criterion was used, with the only exception of the 40 Chi 40 cm(2) field (similar to 95% agreement). With the more stringent criteria of 1%, 1 mm, the agreement reduced to almost 95% for field sizes up to 10 Chi 10 cm(2), worse for larger fields. Unflatness and slope FFF-specific parameters are in line with the possible energy underestimation of the simulated results relative to experimental data. Conclusions: The agreement between Monte Carlo simulations and experimental data proved that the evaluated Varian phase-space files for FFF beams from TrueBeam can be used as radiation sources for accurate Monte Carlo dose estimation, especially for field sizes up to 10 Chi 10 cm(2), that is the range of field sizes mostly used in combination to the FFF, high dose rate beams. (C) 2014 American Association of Physicists in Medicine.