Comparison of GATE and MCNP Monte Carlo codes for internal dosimetery

Introduction: GATE is practically the only nuclear medicine dedicated code with the option to determine the dose distribution inside the body. This code has been designed as upper layer of the GEANT4 toolkit and has been used for internal dosimetery. However, its results have not been compared to other well-developed codes. Moreover anthropomorphic voxel phantoms have never been used with GATE for internal dosimetry. The aim of present study was to compare the dose calculated with GATE to MCNP4B published data. Methods: The Zubal voxel phantom was used to model a typical adult male. Activity was assumed uniformly distributed in liver, kidneys, lungs, spleen, pancreas and adrenals. GATE Monte Carlo package was used for estimation of doses to the organs of phantom. Simulations were performed for photon energy of 0.01-1 MeV and mono-energetic electrons of 935 keV. Specific absorbed fractions for photons and s-values for electrons were calculated. Results: On average, GATE produces higher SAF values (+2.7%) for self-absorption and lower values (-2.9%) for cross-absorption. The SAF values calculated with MCNP4B for lungs as source organ at the energy of 200 and 500 keV was considerably higher than GATE data. Conclusion: Despite of differences between the GATE4 and MCNP4B, the results can be considered ensuring. This difference was almost at the same level as the reported difference between MCNPX and MCNP4b. As a result, this may be considered as validation of GATE as a proprietary code in nuclear medicine for radionuclide dosimetry applications.