A Monte Carlo intercomparison of peak-to-valley dose ratios and output factors for microbeam radiation therapy

A comparison between EGSnrc, Penelope and Geant4 has been made for dosimetry applied to Microbeam Radiation Therapy (MRT). A simple geometry is defined to limit the number of influential parameters and to focus primarily on the dose associated with scattered photons. Use was made of a precalculated photon spectrum for the ESRF ID17 Medical beamline ranging from 40 to 300 keV with a mean energy of 107 keV. In MRT, Compton scattering is the main photon interaction in soft tissue, with the photoelectric effect contributing more substantially in bone. The study investigates differences in the way Compton scattering is handled by the different codes which lead to differences of up to 4% for the simulation of relevant dosimetric quantities in MRT, despite the fact that the cross-section data comes from the same source. There is no significant pattern in the way the codes behave and depending on the dosimetric quantity involved, the agreement between the codes varies. The agreement for each dosimetric quantities is enhanced at large depths where beam-hardening increases the mean energy of the beam and lowers the influence of Doppler broadening and electron binding effects, allowing the codes to use less corrections to the Klein-Nishina model which the three codes implement in the same way.