Evaluation of the uncertainty in calculating nanodosimetric quantities due to the use of different interaction cross sections in Monte Carlo track structure codes

TL;DR

This paper assesses how variations in interaction cross sections in Monte Carlo simulations affect nanodosimetric calculations, revealing significant discrepancies especially at low electron energies and emphasizing the importance of standardized datasets for accurate biological damage prediction.

Contribution

It systematically evaluates the impact of different interaction cross sections on nanodosimetric quantities across multiple Monte Carlo codes, highlighting the primary source of variability.

Findings

Significant discrepancies in ionization cluster size distributions among codes.

Using common cross sections reduces variability, indicating dataset differences as a key factor.

Cross section variations can cause up to 15% differences in simulated DNA double-strand breaks.

Abstract

This study evaluates the uncertainty in nanodosimetric calculations caused by variations in interaction cross sections within Monte Carlo Track Structure (MCTS) simulation codes. Nanodosimetry relies on accurately simulating particle interactions at the molecular scale. Different MCTS codes employ distinct physical models and datasets for electron interactions in liquid water, a surrogate for biological tissues. The paper focuses on the Ionization Cluster Size Distribution (ICSD) generated by electrons of varying energies in nanometric volumes. Seven MCTS codes were tested using their native cross sections and a common dataset derived from averaging data used in the participating codes. The results reveal significant discrepancies among the codes in ICSDs and derived biologically relevant nanodosimetric quantities such as mean ionization numbers (M1) and probabilities of obtaining two or more ionizations (F2). The largest variations were observed for low-energy electrons, where the contribution from interaction cross sections dominates the overall uncertainties. For instance, M1 values for ICSDs of electron of 20 eV can differ by around 45 % (RSD) and 34 % (RSD) was found for F2 values of ICSDs of electrons of 50 eV. Using common cross sections substantially reduced the discrepancies, suggesting that cross section datasets are the primary source of variability. Finally, estimates of deoxyribonucleic acid (DNA) damage using the PARTRAC code highlight tht cross section variations have a non-negligible impact simulated biological outcomes, particularly for double-strand breaks (DSBs) Indeed, despite the fact that many other parameters in the simulation that can greatly differ from one code to another, the different interaction cross-sections studied in this work can lead to differences in the number of DSBs calculated with the PARTRAC code of up to 15%.