Image quality characteristics of myocardial perfusion SPECT imaging using state-of-the-art commercial software algorithms: evaluation of 10 reconstruction methods.

Myocardial perfusion imaging (MPI) is widely used as standard of care in patients with coronary artery disease. The availability of hybrid SPECT/CT imaging system and associated advanced reconstruction algorithms serve to improve diagnostic accuracy and enhances image quality of MPI. The aim of this work was to evaluate the relative performance of iterative reconstruction algorithms correcting for different combinations of image degrading factors versus filtered back projection on the quality of myocardial perfusion SPECT imaging. A standard cardiac phantom containing myocardial defects of different sizes and compositions was used to simulate myocardial perfusion SPECT/CT clinical studies. A clinically relevant activity was determined to avoid discordance with real data acquisition. Acquisition parameters including time per projection, angular rotation increment, and iterative reconstruction number of iterations and subsets were varied. The reconstruction was carried out applying different algorithms including 10 variants of analytical (e.g FBP) and iterative reconstructions with and without resolution recovery. Typical figures of merit were used to evaluate the image quality of MPI reconstructed with ten different reconstruction methods. OSEM-RR showed remarkable improvement of image quality of MPI in terms of SNR, CNR and defect contrast percentage compared to FBP algorithm. Full correction scheme IR-RR (i.e. IRACSCRR) provides clinically acceptable image quality of MPI compared to FBP.