Calculation of cellular S-values using Geant4-DNA: The effect of cell geometry

Abstract

PurposeGeant4-DNA is used to calculate S-values for different subcellular distributions of low-energy electron sources in various cell geometries. MethodCalculations of cellular S-values for monoenergetic electron sources with energy from 1 to 100keV and the Auger-electron emitting radionuclides Tc-99m, In-111, and I-125 have been made using the Geant4 Monte Carlo toolkit. The Geant4-DNA low-energy extension is employed for simulating collision-by-collision the complete slowing-down of electron tracks (down to 8eV) in liquid water, used as a surrogate of human cells. The effect of cell geometry on S-values is examined by simulating electron tracks within different cell geometries, namely, a spherical, two ellipsoidal, and an irregular shape, all having equal cellular and nuclear volumes. Algorithms for randomly sampling the volume of the nucleus, cytoplasm, surface, and whole cell for each cell phantom are presented. ResultsDifferences between Geant4-DNA and MIRD database up to 50% were found, although, for the present radionuclides, they mostly remain below 10%. For most source–target combinations the S-values for the spherical cell geometry were found to be within 20% of those for the ellipsoidal cell geometries, with a maximum deviation of 32%. Differences between the spherical and irregular geometries are generally larger reaching 100–300%. Most sensitive to the cell geometry is the absorbed dose to the nucleus when the source is localized on the cell surface. Interestingly, two published AAPM spectra for I-125 yield noticeable differences (up to 19%) in cellular S-values. ConclusionMonte Carlo simulations of cellular S-values with Geant4-DNA reveal that, for the examined radionuclides, the widely used approximation of spherical cells is reasonably accurate (within 20–30%) even for ellipsoidal geometries. For irregular cell geometries the spherical approximation should be used with caution because, as in the present example, it may lead to erroneous results for the nuclear dose for the commonly encountered situation where the source is localized to the cell surface.