Comparison of linear energy transfer scoring techniques in Monte Carlo simulations of proton beams

Monte Carlo (MC) simulations are commonly used to study linear energy transfer (LET) distributions in therapeutic proton beams. Various techniques have been used to score LET in MC simulations. The goal of this work was to compare LET distributions obtained using different LET scoring techniques and examine the sensitivity of these distributions to changes in commonly adjusted simulation parameters. We used three different techniques to score average proton LET in TOPAS, which is a MC platform based on the Geant4 simulation toolkit. We determined the sensitivity of each scoring technique to variations in the range production thresholds for secondary electrons and protons. We also compared the depth-LET distributions that we acquired using each technique in a simple monoenergetic proton beam and in a more clinically relevant modulated proton therapy beam. Distributions of both fluence-averaged LET (LET_Φ) and dose-averaged LET (LET_D) were studied. We found that LET_D values varied more between different scoring techniques than the LET_Φ values did, and different LET scoring techniques showed different sensitivities to changes in simulation parameters.