The relative impact of scatter on absolute myocardial perfusion quantitation: an EG84 Monte Carlo study

Increased inferior activity in attenuation-corrected myocardial images has been attributed to scatter from the liver. To study this effect, the authors used EGS4 to simulate the SPECT acquisition of a 1.5-cm thick spherical shell placed inside a 24-cm diameter tank. The distance between the shell and a liver-simulating hemisphere was varied from 0.1 to 2 cm and the shell:liver activity concentration ratio was varied from 4:1 to 1:1 to estimate the impact of liver scatter. Profiles through the projection data demonstrated that about 20% of detected counts were scatter, with almost no bias on the side of the shell closest to the liver, except when the shell was very close to the liver. Resulting MLEM-reconstructed images showed only a very small bias proximal to the liver, with quantitative errors dominated by resolution effects. The authors' modeled quantitation procedure underestimated the overall activity concentration by 15% but showed only small regional bias due to liver scatter.