Monte Carlo tomographic reconstruction in SPECT impact of bootstrapping and number of generated events

In Single Photon Emission Computed Tomography (SPECT), 3D images usually reconstructed by performing a set of bidimensional (2D) analytical or iterative reconstructions can also be reconstructed using an iterative reconstruction algorithm involving a 3D projector. Accurate Monte Carlo (MC) simulations modeling all the physical effects that affect the imaging process can be used to estimate this projector. However, the accuracy of the projector is affected by the stochastic nature of MC simulations. In this paper, we study the accuracy of the reconstructed images with respect to the number of simulated histories used to estimate the MC projector. Furthermore, we study the impact of applying the bootstrapping technique when estimating the projector