The influence of Geant4-DNA toolkit parameters on electron microdosimetric track structure

Multiple Mesh-type Real Human Cell Models for Dosimetric Application Coupled with Monte Carlo Simulations

The mesh-type models are superior to voxel models in cellular dose assessment coupled with Monte Carlo codes. The aim of this study was to expand the micron-scale mesh-type models based on the fluorescence tomography of real human cells, and to investigate the feasibility of these models in the application of various irradiation scenarios and Monte Carlo codes. Six different human cell lines, including pulmonary epithelial BEAS-2B, embryonic kidney 293T, hepatocyte L-02, B-lymphoblastoid HMy2.CIR, Gastric mucosal GES-1, and intestine epithelial FHs74Int, were adopted for single mesh-type models reconstruction and optimization based on laser confocal tomography images. Mesh-type models were transformed into the format of polygon mesh and tetrahedral mesh for the GATE and PHITS Monte Carlo codes, respectively. The effect of model reduction was analyzed by dose assessment and geometry consideration. The cytoplasm and nucleus doses were obtained by designating monoenergetic electrons and protons as external irradiation, and S values with different "target-source" combinations were calculated by assigning radioisotopes as internal exposure. Four kinds of Monte Carlo codes were employed, i.e., GATE with "Livermore," "Standard" and "Standard and Geant4-DNA mixed" models for electrons and protons, as well as PHITS with "EGS" mode for electrons and radioisotopes. Multiple mesh-type real human cellular models can be applied to Monte Carlo codes directly without voxelization when combined with certain necessary surface reduction. Relative deviations between different cell types were observed among various irradiation scenarios. The relative deviation of nucleus S value reaches up to 85.65% between L-02 and GES-1 cells by 3H for the "nucleus-nucleus" combination, while that of 293T and FHs74Int nucleus dose for external beams at a 5.12 cm depth of water is 106.99%. Nucleus with smaller volume is far more affected by physical codes. There is a considerable deviation for dose within BEAS-2B at the nanoscale. The multiple mesh-type real cell models were more versatile than voxel models and mathematical models. The present study provided several models which can easily be extended to other cell types and irradiation scenarios for RBE estimations and biological effect predictions, including radiation biological experiments, radiotherapy and radiation protection.

Research on the proximity functions of microdosimetry of low energy electrons in liquid water based on different Monte Carlo codes

PURPOSE

The proximity function is an important index in microdosimetry for describing the spatial distribution of energy, which is closely related to the biological effects of organs or tissues in the target area. In this work, the impact of parameters, such as physic models, cut-off energy, and initial energy, on the proximity function are quantitated and compared.

METHODS

According to the track structure (TS) and condensed history (CH) low-energy electromagnetic models, this paper chooses a variety of Monte Carlo (Monte Carlo, MC) codes (Geant4-DNA, PHITS, and Penelope) to simulate the track structure of low-energy electrons in liquid water and evaluates the influence of the electron initial energy, cut-off energy, energy spectrum, and physical model factors on the differential proximity function.

RESULTS

The results show that the initial energy of electrons in the low-energy part (especially less than 1 keV) has a greater impact on the differential proximity function, and the choice of cut-off energy has a greater impact on the differential proximity function corresponding to small radius sites (generally less than 10 nm). The difference in the electronic energy spectrum has little effect on the result, and the proximity functions of different physics models show better consistency under large radius sites.

CONCLUSIONS

This work comprehensively compares the differential proximity functions under different codes by setting a variety of simulation conditions and has basic guiding significance for helping users simulate and analyze the deposition characteristics of microscale electrons according to the selection of an appropriate methodology and cut-off energy.

New damage model for simulating radiation-induced direct damage to biomolecular systems and experimental validation using pBR322 plasmid

In this work, we proposed a new damage model for estimating radiation-induced direct damage to biomolecular systems and validated its the effectiveness for pBR322 plasmids. The proposed model estimates radiation-induced damage to biomolecular systems by: (1) simulation geometry modeling using the coarse-grained (CG) technique to replace the minimum repeating units of a molecule with a single bead, (2) approximation of the threshold energy for radiation damage through CG potential calculation, (3) calculation of cumulative absorption energy for each radiation event in microscopic regions of CG models using the Monte Carlo track structure (MCTS) code, and (4) estimation of direct radiation damage to biomolecular systems by comparing CG potentials and absorption energy. The proposed model replicated measured data with an average error of approximately 14.2% in the estimation of radiation damage to pBR322 plasmids using the common MCTS code Geant4-DNA. This is similar to the results of previous simulation studies. However, in existing damage models, parameters are adjusted based on experimental data to increase the reliability of simulation results, whereas in the proposed model, they can be determined without using empirical data. Because the proposed model proposed is applicable to DNA and various biomolecular systems with minimal experimental data, it provides a new method that is convenient and effective for predicting damage in living organisms caused by radiation exposure.

THE APPLICATION OF NEURAL NETWORK TECHNOLOGY BASED ON MEA-BP ALGORITHM IN THE PREDICTION OF MICRODOSIMETRIC QUALITIES

The most abundant products of the interaction between radiation and matter are low-energy electrons, and the collisions between these electrons and biomolecules are the main initial source of radiation-based biological damage. To facilitate the rapid and accurate quantification of low-energy electrons (0.1-10 keV) in liquid water at different site diameters (1-2000 nm), this study obtained ${\overline{y}}_{\mathrm{F}}$ and ${\overline{y}}_{\mathrm{D}}$data for low-energy electrons under these conditions. This paper proposes a back-propagation (BP) neural network optimized by the mind evolutionary algorithm (MEA) to construct a prediction model and evaluate the corresponding prediction effect. The results show that the ${\overline{y}}_{\mathrm{F}}$ and ${\overline{y}}_{\mathrm{D}}$ values predicted by the MEA-BP neural network algorithm reach a training precision on the order of ${10}^{-8}$. The relative error range between the prediction results of the validated model and the Monte Carlo calculation results is 0.03-5.98% (the error range for single-energy electrons is 0.1-5.98%, and that for spectral distribution electrons is 0.03-4.4%).

Calculation of microdosimetric spectra for protons using Geant4-DNA and a μ-randomness sampling algorithm for the nanometric structures

PURPOSE

Through introducing stochastic quantities that can be connected to the dimensions of the microscopic structures exposed to radiations, microdosimetry is concerned with the substantive specifications of radiation quality that could help gain insight into radiation effects. Utilizing the μ-randomness method and Geant4-DNA code, we calculated microdosimetry quantities for nanometric structures in a spherical body of water irradiated with protons. To gain more insight into the effects of radiation on microscopic structures and validate the code parameters, we made a comparison between our results obtained within Geant4-DNA and results from other simulations.

MATERIALS AND METHODS

We calculated microdosimetric quantities through irradiating a spherical body of water of 6 μm diameter with 0.5-100 MeV protons. Microdosimetric quantities were derived for cylinders with diameter × height values of 23 × 23, 50 × 100, and 300 × 300 Å × Å, which would resemble the typical sizes of sub-cellular organisms such as the DNA, nucleosome, and chromatin fiber. We exploited the concept of μ-randomness to introduce convex bodies of random positions and directions for calculating microdosimetric quantities. We used the Geant4-DNA Monte Carlo simulation toolkit for transporting protons and secondary particles and calculating the frequency- and dose-mean lineal and specific energies in cylindrical volumes. Specifically, for same-sized cylindrical volumes, microdosimetric parameters obtained by Nikjoo et al. using the KURBUC code were used for evaluation.

RESULTS

For the energy range investigated, the frequency-mean lineal energy, dose-mean lineal energy, frequency-mean specific energy, and dose-mean specific energy vary within [2.34,47.06] (keV/μm), [10.40,68.55] (keV/μm), [0.04,39.38] × 10⁶ cGy, and [0.16,90.29] × 10⁶ cGy, respectively. Regardless of the proton energy, our specific-energy results showed higher sensitivity to volume change, for smaller cylinder volumes rather than larger ones. Regardless of both proton energy and volume of the cylinder under study, we observed a generally better agreement between our frequency-mean, than dose-mean, specific energy results and the KURBUC results.

CONCLUSION

Using Geant4-DNA to account for the stochastic nature of energy depositions due to physical interactions between radiation and matter, we calculated microdosimetry parameters concerning proton irradiation. By employing microdosimetry concepts in conjunction with simulation results of our previous work on radiation effects on the DNA, we pinpointed and quantified correlations between microdosimetry parameters and DNA damage. As such, for a volume with comparable mass and mean chord length to the DNA, we could observe the clear correspondence of the mean lineal and specific energy results with the double-strand-break yields of protons in Gy^-1.Gbp^-1.