Purpose: Monte Carlo (MC) simulation is often based on phase-space files (PSF), as they are able to accurately represent particle characteristics. However, the large-sized PSF creates a bottleneck for speeding up calculations, as the reading of the PSF takes appreciable time. In addition, the number of independent particles is limited by the PSF, preventing further reduction of statistical uncertainty. The purpose of this study is to develop and validate a virtual source model (VSM) to address these limitations.
Methods: Particles from existing phase space files for Varian 6X, 6FFF, 10X, and 10FFF beams were tallied and analyzed. The particle density distribution, energy spectrum, particle direction, and the correlations between characteristics were computed. Virtual phase space models were created for these energies. The VSM approach was validated with water tank simulations as well as clinical cases against the original PSF. Computation time and gamma index passing rate were recorded for each case.
Results: The new VSM only takes 2MB of disk space for each beam energy, compared to 50GB for PSF. The VSM approach reduces the calculation time by 10 mins on our server. At 5% MC uncertainty, the VSM approach was 45% faster in water tank simulations and reduces the calculation time for most clinical cases from ~15 mins to ~1.5 mins. Gamma index passing rates remain high at 5% statistical uncertainty for both approaches, with 98.5% +- 1.0% for VSM and 99.1% +- 0.9% for PSF.
Conclusion: We have created VSM model for Varian Linacs. The VSM approach significantly saves computation time without sacrificing accuracy for second dose calculation.
Funding Support, Disclosures, and Conflict of Interest: Dr. Castle and Dr. Feng are employees of Carina Medical LLC. Dr. Chen is a co-founder of Carina Medical LLC.