Comparison of PHITS, GEANT4, and HIBRAC simulations of depth-dependent yields of β(+)-emitting nuclei during therapeutic particle irradiation to measured data.

For quality assurance in particle therapy, a non-invasive, in vivo range verification is highly desired. Particle therapy positron-emission-tomography (PT-PET) is the only clinically proven method up to now for this purpose. It makes use of the β(+)-activity produced during the irradiation by the nuclear fragmentation processes between the therapeutic beam and the irradiated tissue. Since a direct comparison of β(+)-activity and dose is not feasible, a simulation of the expected β(+)-activity distribution is required. For this reason it is essential to have a quantitatively reliable code for the simulation of the yields of the β(+)-emitting nuclei at every position of the beam path. In this paper results of the three-dimensional Monte-Carlo simulation codes PHITS, GEANT4, and the one-dimensional deterministic simulation code HIBRAC are compared to measurements of the yields of the most abundant β(+)-emitting nuclei for carbon, lithium, helium, and proton beams. In general, PHITS underestimates the yields of positron-emitters. With GEANT4 the overall most accurate results are obtained. HIBRAC and GEANT4 provide comparable results for carbon and proton beams. HIBRAC is considered as a good candidate for the implementation to clinical routine PT-PET.