Correction of partial volume effects for PET imaging: a comparison study

The low spatial resolution of PET scanners results in partial volume (PV) effects limiting the quantification in small structures. In this study, we compare the correction algorithms implemented at the Research Center Juelich (PVC-J) and at the Brain Imaging Centre of Montreal (PVC-M). PVC-J algorithm: The corrected grey matter (GM) activity image is obtained by dividing voxel-wise the uncorrected GM image by the GM probability map, derived from the convolution of the corresponding MR segmented image by a 3D spatially variant gaussian function, which reproduces the actual PET image resolution. PVC-M algorithm: It accounts for the mutual PV effects between any tissue structure. The cross-contamination factors are computed for each of the structures yielding the geometric transfer matrix which is solved to get the true mean activity values. A PET dynamic acquisition of an adenosin receptor study was simulated using the Zubal´s computerized phantom and the SORTEO Monte Carlo PET simulator. A global spatial resolution of 9.5mm was used with both methods. Mean deviations over the dynamic data from the reference values are ranging from +6% for the GM region and PVC-J to -10% for globus pallidus and PVC-M. The data show a very high consistency of the results obtained from the two different methods concerning the adenosine receptor study taken as basis for the simulation.