Effects of Distorted PET Projection Data on the Reconstructed Image Using Different Reconstruction Algorithms

Summary

In a simulation study, the effects of multiplicative and additive distortions of the projection data on the reconstructed image quality and quantitation capability were investigated. Three different reconstruction algorithms were compared: (i) filtered back-projection (FBP), (ii) an iterative method with multiplicative correction term (ML-EM), and (iii) an iterative method with additive correction term (AR-AC). Distortions were uniformly applied to all projections with projection angles ranging from 72° through 108°. Multiplicative distortion factors ranged from 0.5 to 1.4, additive distortions were −30% to +30%. The reconstructed images and activity concentrations were not affected by the multiplicative distortions if ML-EM was applied, whereas additive distortions were best tolerated when FBP or AR-AC were used for image reconstruction.