The effect of non-uniform attenuation compensation on myocardial SPECT defect analysis

Abstract
In SPECT myocardial imaging, the measurement of defect size helps to evaluate the existence and severity of suspected cardiovascular disease. To determine whether non-uniform attenuation and its compensation after the apparent defect size, and thus potentially impact clinical evaluations, phantom studies were performed with defects in a 'myocardium'. Non-uniform attenuation compensation was performed with weighted backprojection and a cone-beam CT attenuation map. Polar maps with defects were compared to maps without defects but otherwise identically acquired and processed. Attenuation compensation generally altered the defect size by no more than 10%, except that (i) a truncated attenuation map altered the size of some defects, and (ii) the size was increased when the polar map with a defect was more attenuated than the defect-free polar map (simulating the comparison of a large patient to a smaller normal). This increase was removed by attenuation compensation, which also produced much more circularly symmetric polar maps.