LOR-interleaving image reconstruction for PET with collimation

We propose using collimation in a PET scanner to achieve high-resolution images in a central region of interest. The collimator is designed to partially mask detector crystals to detect lines of response (LORs) within fractional crystals. A sequence of collimator-encoded LORs is measured with different collimation configurations. This novel collimated scanner geometry makes the reconstruction problem challenging, as collimator effects need to be modeled to restore high-resolution LORs from collimator-encoded LORs. In this study, we presented an LOR-interleaving technique to reconstruct high-resolution images from a set of collimated sinograms. We developed a 3-D ray-tracing model for the collimator incorporating both the collimator and crystal penetration. Each mean measured LOR can be modeled as a mixture of high-resolution LORs with mixing coefficients determined by collimator attenuation. By registering the measured LORs with the collimator configurations, high-resolution LORs are restored based on the modeled transfer coefficients using the nonnegative least-squares method. For validation, we applied the approach to A-PET, a small-animal PET scanner, with a specially designed collimator. Details regarding the implementation of the algorithm are provided. We show that our approach can obtain reconstructions with significantly improved spatial resolution and quantification compared with uncollimated PET.