Monte Carlo dose distribution calculation at nuclear level for Auger-emitting radionuclide energies

The distribution of radiopharmaceuticals in tumor cells represents a fundamental aspect for a successful molecular targeted radiotherapy. It was largely demonstrated at microscopic level that only a fraction of cells in tumoral tissues incorporate the radiolabel. In addition, the distribution of the radionuclides at sub-cellular level, namely inside each nucleus, should also be investigated for accurate dosimetry estimation. The most used method to perform cellular dosimetry is the MIRD one, where S-values are able to estimate cellular absorbed doses for several electron energies, nucleus diameters, and considering homogeneous source distributions. However the radionuclide distribution inside nuclei can be also highly non-homogeneous. The aim of this study is to show in what extent a non-accurate cellular dosimetry could lead to misinterpretations of surviving cell fraction vs dose relationship; in this context, a dosimetric case study with Tc-99m is also presented. Methods: The state-of-art MCNP6 Monte Carlo simulation was used in order to model cell structures both in MIRD geometry (MG) and MIRD modified geometries (MMG), where also entire mitotic chromosome volumes were considered (each structure was modeled as liquid water material). In order to simulate a wide energy range of Auger emitting radionuclides, four mono energetic electron emissions were considered, namely 213 eV, 6 keV, 11 keV and 20 keV. A dosimetric calculation for Tc-99m undergoing inhomogeneous nuclear internalization was also performed. Results: After a successful validation step between MIRD and our computed S-values for three Auger-emitting radionuclides (Tc-99m, I-125 and Cu-64), absorbed dose results showed that the standard MG could differ from the MMG from one to three orders of magnitude. These results were also confirmed by considering the Tc-99m spectrum emission (Auger and internal conversion electrons). Moreover, considering an inhomogeneous radionuclide distribution, the average electron energy that maximizes the absorbed dose was found to be different for MG and MMG. Conclusions: The modeling of realistic radionuclide localization inside cells, including a inhomogeneous nuclear distribution, revealed that i) a strong bias in surviving cell fraction vs dose relationships (taking to different radiobiological models) can arise; ii) the alternative models might contribute to a more accurate prediction of the radiobiological effects inherent to more specific molecular targeted radiotherapy strategies.