Microdosimetry calculations in situ for clinically relevant photon sources and their correlation with the early DNA...

Radiobiological data suggests variations in relative biological effectiveness (RBE) between clinically used photon-based sources. A microdosimetric formalism using Monte Carlo (MC) methods can mechanistically describe the photon RBE. Experimentally derived RBE based on DNA double-strand breaks ( ${\mathrm{RBE}}_{\mathrm{DSB}}$ ) has been shown to scale with the microdosimetry quantity dose-mean lineal energy ( $y_D$ ).

To calculate microdosimetric spectra for clinically relevant photon sources, spanning from soft x-rays produced by a 50 kVp x-ray source through various brachytherapy sources up to a 6 MV medical linac. Furthermore, we investigated the correlation between ${\mathrm{RBE}}_{\mathrm{DSB}}$ and $y_D$ of different photon sources.

Photon sources simulated include low-energy x-rays (50 kVp), orthovoltage x-rays (225 kVp), high-dose-rate brachytherapy sources (⁷⁵Se, ¹⁹²Ir and ⁶⁰Co), and a 6 MV medical linac. Secondary electron spectra at the cellular level were calculated for in vitro cell irradiation setups using Geant4 MC-based packages, RapidBrachyMCTPS and RapidExternalBeam. The obtained spectra were used in MicroDose, a microdosimetry simulation software, to obtain microdosimetric quantities, including single-event lineal energy ( $y$ ) and specific energy ( $z$ ) spectra, and dose-mean and frequency-mean quantities ( $y_F$ , $y_D$ , $z_F^s$ , $z_D^s$ ). Uniform spherical targets (1-14  $\mu m$ radius) and realistic HeLa and PC3 cell nucleus models were simulated using cell size data obtained from literature and nuclei size data from confocal microscopy imaging. Radiobiological experiments using $\gamma \text{H2AX}$ foci quantified DNA double-strand breaks for HeLa and PC3 cells after irradiations with 50 and 225 kVp, ¹⁹²Ir, and 6 MV linac, and ${\mathrm{RBE}}_{\mathrm{DSB}}$ was determined using 225 kVp as the reference.

The calculated $y_D$ ( $y_F$ ) is within the 3.5-1.2 keV/ $\mu m$ range (1.8-0.2 keV/ $\mu m$ ) for 1  $\mu m$ simulated target size between the lowest energy 50 kVp x-ray source and the highest energy 6 MV linac source, respectively. For the HeLa and PC3 cell nuclei models based on microscopy data, $y_D$ ( $y_F$ ) spans from 1.6 to 0.6 keV/ $\mu m$ (0.7 to 0.2 keV/ $\mu m$ ). When compared between different target sizes, $y_D$ ( $y_F$ ) ranges from 3.5 to 1.0 (1.8-0.4) keV/ $\mu m$ between 1 and 10  $\mu m$ radius targets for the 50 kVp x-ray source. A smaller change is observed for 6 MV linac, ranging from 1.2 to 0.5 keV/ $\mu m$ and 0.23 to 0.22 keV/ $\mu m$ for $y_D$ and $y_F$ , respectively. For the simulated ⁷⁵Se source currently under investigation, the calculated $y_D$ values are 11%-24% higher relative to those of ¹⁹²Ir in the range of target sizes between 1 and 14  $\mu m$ in radius. ${\mathrm{RBE}}_{\mathrm{DSB}}$ for HeLa cells was 1.4 $\pm$ 0.7 for 50 kVp x-rays, 0.5 $\pm$ 0.2 for ¹⁹²Ir, and 0.7 $\pm$ 0.4 for 6 MV linac irradiations. For PC3 cells, ${\mathrm{RBE}}_{\mathrm{DSB}}$ was 1.3 $\pm$ 0.6, 0.8 $\pm$ 0.4 and 0.5 $\pm$ 0.3 for 50 kVp, ¹⁹²Ir and 6 MV linac, respectively. Measured ${\mathrm{RBE}}_{\mathrm{DSB}}$ values are consistent with $y_D$ ratios of the corresponding photon sources for HeLa and PC3 nucleus models.

Microdosimetric spectra strongly depend on the simulated energy of photon sources and target size, with $y_D$ and $z_D^s$ decreasing by a factor of $\approx$ 2-3 between diagnostic 50 kVp and 6 MV therapeutic x-rays for target sizes from 1-14 $\mu m$ in radius. The early damage ${\mathrm{RBE}}_{\mathrm{DSB}}$ indicates this stochastic change in energy density between various photon sources as the yields of $\gamma \text{H2AX}$ foci per nucleus scale with $y_D$ of the source.