Reproducibility study of Monte Carlo simulations for nanoparticle dose enhancement and biological modeling of cell survival curves.

Nanoparticle-derived radiosensitization has been investigated by several groups using Monte Carlo simulations and biological modeling. In this work we replicated the physical simulation and biological modeling of previously published research for 50 nm gold nanoparticles irradiated with monoenergetic photons, various 250 kVp photon spectra, and spread-out Bragg peak (SOBP) protons. Monte Carlo simulations were performed using TOPAS and used condensed history Penelope low energy physics models for macroscopic dose deposition and interaction with the nanoparticle; simulation of the microscopic dose deposition from nanoparticle secondaries was performed using Geant4-DNA track structure physics. Biological modeling of survival fractions was performed using a local effect model-type approach for MDA-MB-231 breast cancer cells. Physical simulation results agreed extraordinarily well at all distances (1 nm to 10 μ m from nanoparticle) for monoenergetic photons and SOBP protons in terms of dose per interaction, dose kernel ratio (often labeled dose enhancement factor), and secondary electron spectra. For 250 kVp photons the influence of the gold K-edge was investigated and found to appreciably affect the results. Calculated survival fractions similarly agreed well within one order of magnitude at macroscopic doses (i.e. without nanoparticle contribution) from 1 Gy to 10 Gy. Several 250 kVp spectra were tested to find one yielding closest agreement with previous results. This highlights the importance of a detailed description of the low energy (< 150 keV) component of photon spectra used for in-silico , as well as in-vitro , and in-vivo studies to ensure reproducibility of the experiments by the scientific community. Both, Monte Carlo simulations of physical interactions of the nanoparticle with photons and protons, as well as the biological modelling of cell survival curves agreed extraordinarily well with previously published data. Further investigation of the stochastic nature of nanoparticle radiosenstiziation is ongoing.