Study of the influence of photon energy cuts on the PET simulation results

The objective of this work is to study the influence of photon energy cuts on the results and speed of positron emission tomography (PET) simulations. The investigation is performed with two different MC codes: GATE and egs_pet. These codes were used to simulate the response of GE Discovery LS PET scanner to a point, 511 keV back-to-back, photon source placed at the center of a cubic 30 × 30 × 30 cm water phantom. The effect of the photon cuts on the simulation results is studied by making simulations with 32 energy cut values in the interval 0.3-350 keV. The dependence of the number of singles, primary and scattered coincidences on the applied energy cut is evaluated. The effect of the cuts on the simulation speed is also estimated. The results from the two codes are found to agree well within the statistical uncertainties for all the cuts applied. It is found that the simulation of atomic relaxation processes and the propagation of characteristic X rays play important role if an accurate modeling of a scanner system is to be achieved. For this scanner which is made of BGO crystals the usage of cuts below the Ka X-rays of Bi does not lead to statistically significant differences in the number of recorded events. Photon cuts in the interval 80-100 keV lead to less than 1.5% change in the number of singles, and less than 3% difference in the number of true and scattered coincidences. The differences are mainly due to the simulation/no simulation of Bi X-rays in the BGO. Energy cuts higher than 170 keV result in a strong increase in the number of detected events due to the increased absorption of Compton scattered photons and are not appropriate for PET simulations. The simulation time does not change dramatically for photon cuts between 0.3 and 100 keV. In the current studies an acceleration of the order of 35% is achieved by changing the photon cut from 1 to 100 keV.