Do cell culturing influence the radiosensitizing effect of gold nanoparticles part 1: scrutinizing recent evidence for data consistency

TL;DR

This paper critically examines recent claims about the influence of cell culturing methods on the radiosensitizing effects of gold nanoparticles, revealing methodological issues and inconsistencies in simulation and experimental data.

Contribution

It provides a detailed analysis of experimental and simulation approaches, highlighting potential biases and errors affecting the interpretation of gold nanoparticle radiosensitization studies.

Findings

Simulation setup corresponds to cells near water phantom surface

Validation against experimental data is misleading due to quadratic term neglect

Reported discrepancies may be due to incorrect AuNP size assumptions

Abstract

In radiobiological experiments, the cells can either float in suspension or adhere to the walls of the sample holder. When irradiation is performed in the presence of dose modifying agents such as gold nanoparticles (AuNPs), the different shapes of the floating or adherent cells may imply a different dose to the nucleus, with biological consequences such as cell survival. Recently, it has been reported that the survival rate varies by up to a factor of 1.5 for the two cell geometries and by up to a factor of 2 for different orientation of the cells with respect to the incident beam. These results are examined in this paper and possible methodological issues are analyzed. This analysis shows that the simulation setup corresponds to the case of cells in the dose build-up region near the surface of a water phantom, where different depths result in different dose and survival probabilities. The validation of the simulations by comparison with experimental data is misleading, as the apparent agreement is due to a neglect of the quadratic term of the linear-quadratic survival model. The analysis further shows that the reported step-like changes between the survival predicted from the mean dose and the LEM could be explained by the fact that in the entire simulation only in one event an ionizing interaction of a photon took place in an AuNP. It is shown that the probability of this is in the permille range and that the total number of electrons leaving an AuNP, estimated from the reported electron spectrum, is three orders of magnitude higher than the value estimated from the expectation of photon interactions in the AuNPs. This contradiction would be resolved if the AuNP diameter in the simulations were a factor of 10 larger than intended. Another possible explanation for the discrepancies is a hidden bias in the simulation geometry, for example, if the distribution of AuNPs was non-uniform.