3-D Monte Carlo-based scatter compensation in quantitative I-131 SPECT reconstruction

We have implemented highly accurate Monte Carlo based scatter modeling (MCS) with 3-D ordered subsets expectation maximization (OSEM) reconstruction. The scatter is included in the statistical model as an additive term and attenuation and detector response are included in the forward/backprojector. In the present implementation of MCS, a simple multiple window-based estimate is used for the initial iterations and in the later iterations the Monte Carlo estimate is used for several iterations before it is updated. For I-131, MCS was evaluated and compared with triple energy window (TEW) scatter compensation using simulation studies of a mathematical phantom and a clinically realistic voxel-phantom. Even after just two Monte Carlo runs, excellent agreement was found between the MCS estimate and the true scatter distribution. Accuracy and noise of the reconstructed images were superior with MCS compared to TEW. However, the improvement was not large, and in some cases may not justify the large computational requirements of MCS. Finally clinical application of MCS was demonstrated by applying the method to a radioimmunotherapy (RIT) patient study.