Post-Reconstruction Partial Volume Correction of PET Images Using Iterative Deconvolution Algorithm and Anatomical Priors

PET image quality is affected by partial volume effect (PVE). Moreover, PVE can lead to inaccurate estimate of radiotracer concentration in each voxel of the reconstructed PET volume. Several methods could be used to correct PET images for PVE and can be mainly grouped into two categories: pre- and post-reconstruction approaches. The paper presents a novel post-reconstruction partial volume correction (PVC) method for PET images based on an iterative deconvolution algorithm (Van Cittert method, VC) including anatomical priors (VC-MAP). The proposed method was tested on simulated and real PET acquisition. Preliminary results from simulated images show that the VC-MAP provide better recovery of the true image with respect to the classical VC deconvolution scheme. Moreover, the VC-MAP method provides images with a better edge definition that appear also less noisier than the VC ones. The proposed method was validated using a real PET image of the NEMA IEC Body Phantom. Also in this case the VC-MAP method provides PVC images of good quality. Thus, the proposed method allows at the same time to improve the spatial resolution and to reduce the noise in the image.