Quantification and reduction of attenuation related artifacts in SPET by applying attenuation model during iterative image reconstruction: a Monte Carlo study.

Photon attenuation is one of the main causes of the quantitative errors and artifacts in SPET. A transmission or CT based attenuation map is necessary to correct for the effects of attenuation accurately. In this research, some important attenuation related artifacts are described. A fast and memory efficient iterative algorithm is proposed for attenuation correction. Ordered subset expectation maximization (OSEM) algorithm with attenuation model was applied for image reconstruction. Monte Carlo simulation was used to create the projections in this study. Different voxel based phantoms with uniform and non-uniform activity distributions and attenuation maps were employed to evaluate the accuracy of this algorithm. The NCAT digital phantom was also used to investigate the attenuation effects on myocardial perfusion SPET in men and women. Projections free from the effect of attenuation were also simulated. The reconstructed image from these attenuation-free projections was considered as reference image. Our attenuation correction algorithm was evaluated by its ability to recover activity and to remove attenuation related artifacts. The mean-square-error (MSE) between reference and corrected image and image contrast were calculated for quantitative evaluation of this algorithm. A variety of attenuation related artifacts were observed. Moreover anterior wall of myocardial perfusion images of female phantom and inferior wall in male phantom were affected by attenuation. All of the attenuation related artifacts were removed after attenuation correction. Quantitatively, the MSE values between reference images and corrected images were reduced by about 900% for all phantoms. In conclusion, by applying our new method for incorporating attenuation model during OSEM, we were able to eliminate a variety of artifacts and errors, which is a necessary step for quantitative SPET.