Signal and noise delineation for prompt-gamma detection during hadrontherapy

摘要

Proton and carbon therapies of a human brain were simulated using FLUKA, with particular emphasis on treatment monitoring using the prompt gamma method where activities of the source proton/carbon in the patient's body are to be extracted from the positional profile of exit photons. Tissue heterogeneity ranging from hydrogen to zinc was represented in a VIP-Man anthropomorphic voxel phantom. Each photon exiting the head was scored with ancestry attributes such as the particle type of its parent, the originating collision type and site. These data, not accessible via physical detectors, were analysed to characterise the escaping photons, delineating signal from noise according to the gamma origin and scatter history. Combinations of energy, geometry, time and angle filters, aided by raytracing, were studied for the optimal compromise between a sufficiently featured exit profile and a statistically adequate count. These will be put in context with ongoing research in the field; our perspective for the prompt gamma method will be discussed in detail.