Small animal PET with a clinical PET/CT: Optimizing image quality with MAP reconstruction and super-resolution

Previous work has shown that super-resolution (SR) can significantly improve image quality of small animal images acquired on a clinical PET/CT. Another approach is to utilize an optimized image reconstruction algorithm that is better suited for small animal imaging. We investigate improvements in the quality of phantom and animal images using maximum a posteriori (MAP) iterative reconstruction and SR compared to the scanner workstation´s native reconstruction. Significant gains in image quality were obtained with MAP reconstruction and were attributed to a finer voxel size and a more accurate projector model. SR offered similar improvement when applied to native reconstruction. Additional gain in image quality was obtained when combining the two techniques (applying SR to MAP-reconstructed images). In phantom scans, combined MAP+SR enabled 2.4mm hot rods and cold rods to be visualized.