An inverse Monte Carlo algorithm for 3D PET image reconstruction

We have developed a 3D PET reconstruction algorithm based upon inverse Monte Carlo analysis and have tested it on data acquired by high-resolution 3D PET detectors. This algorithm backprojects 3D PET data across a 3D array of voxels, then during each iteration generates Monte Carlo “pseudo-data” which is similarly backprojected across a 3D voxel array. Comparison between the backprojection of the data and the backprojection of the pseudo-data is used to update the source distribution hypothesis. Inverse Monte Carlo methods provide a natural framework for incorporation of detector response functions, and potentially for the incorporation of effects due to scatter and other sources of systematic error. The algorithm is fairly fast, capable of reconstructing a 128×128×128 Hoffman brain phantom image in a few hours on a modern single-processor workstation. We have achieved nearly linear speed-up by implementation of the algorithm on a computer with 8 parallel processors