Development of a Hybrid Image Reconstruction Algorithm Combining PET and Compton Events for Whole Gamma Imaging

The goal of whole gamma imaging (WGI) is to utilize all detectable gamma rays from radioactive decays for imaging. By inserting a scatterer ring into a PET ring, the WGI detects not only 511 keV annihilation photons of positron emitters but also single gamma rays with various energies as Compton events. The Compton events include one photon of a pair of the annihilation photons in that the other photon is not detected, as well as a single gamma-ray photon emitted from a non-pure positron emitter. To realize the WGI concept, it is necessary to develop an image reconstruction method combining the two different types of events. In this study, therefore, we developed a hybrid image reconstruction algorithm which utilized PET and Compton events efficiently in each iteration based on the list-mode ordered subset expectation maximization framework. For each sub-iteration of the proposed algorithm, a weighted sum of update images individually calculated for PET and Compton events was used as an update image. To demonstrate the effectiveness, we used our WGI prototype to measure a contrast phantom filled with a ⁸⁹Zr solution of 9.4 MBq having a hot to background ratio of 4:1. We extracted 511 keV PET events and 909 keV Compton events by setting different energy windows in software-based coincidence detection. Reconstructed images obtained using both PET and Compton events were compared with those using PET events only and those using Compton events only. The contrast recovery against background variability for each iteration was used as a figure of merit. The results showed the image reconstructed using both PET and Compton events had better image contrast compared with other images.