A correction algorithm for partial volume effects in 3D PET imaging: principle and validation

Presents a method for correcting 3D PET data for partial volume effects. The authors' method assumes that the observed signal within a region of interest is a weighted sum of the true activities in the FOV of the scanner. The weighting coefficients represent the recovery and the cross-contamination factors between structures, and are estimated by 3D analytical simulation of a high resolution emitting model, such as a MRI segmented volume registered with the corresponding PET volume. The validation of the algorithm using well-controlled data sets, proved its efficiency to recover true time-activity curves in small structures, with errors smaller than 5%.