Influence of mismatched CT anatomy on the accuracy of partial volume compensation in cardiac SPECT perfusion imaging

Necessary conditions for successful partial volume effect (PVE) compensation are: 1) a co-registered high-spatial resolution CT Map, 2) compensation for respiratory motion present in the emission data, and 3) Compton scatter compensation. The aim of this study was to determine what affect a mismatched CT Map will have on the accuracy of PVE compensated cardiac SPECT perfusion images. The MCAT phantom was used to generate two different sets of mismatched segmented binary CT maps. First, normal respiratory motion was simulated during the acquisition of the emission projections, and a 'breath-hold' binary CT map was generated. Secondly, 6 emission data sets with x and y-directional shifts (1/4, 1/2, 3/4, 1, 2, and 3 pixels) of the heart compared to the CT Map were generated. Reconstructed slices were post filtered with a 3-dimensional Gaussian with a sigma of 1 pixel (0.467 cm). The PVE compensated data were again filtered with a 3-dimensional Gaussian with a sigma of 0.5 pixels. Cooling is visible in the inferior and antro-lateral regions of the polar map generated from projections with respiratory motion without motion or PVE compensation. PVE compensation clearly did improve the uniformity of the polar map, but did not undo the effects of respiratory motion. The polar maps generated from the reconstructions with x and y shifts showed significant defects only when the mismatch between the emission and CT maps approached 2 pixels. Therefore, PVE compensation may be possible in cardiac perfusion SPECT imaging especially with respiratory motion compensation and correction of patient body motion.