Correction of soft tissue attenuation using differential attenuation between detectors with a CZT camera: A SPECT phantom study

Objectives Soft tissue attenuation remains problematic for single photon emission computed tomography (SPECT)-myocardial perfusion imaging (MPI). The differential attenuation between detectors of Cadmium Zinc Telluride (CZT) camera might result in discrepancy between the estimated and acquired projection images. In this phantom study, we present a novel method, without any transmission scan, to correct soft tissue attenuation using differential attenuation between detectors of a CZT scanner.

Methods Image acquisition of an anthropomorphic torso phantom with a cardiac insert filled with Tc-99m using a CZT camera (GE Discovery NM 530c) was performed. SPECT images were reconstructed using a maximum a posteriori expectation maximization algorithm. Attenuation map was generated from the ratio of each pixel on acquired and estimated projection images. The MPI was applied on the 20-segment polar map provided by the commercially available QPS software (Cedars-Sinai Medical Center, CA, USA). Heterogeneity was expressed by means of coefficient of variation (CV, %).

Results In phantom experiments with torso, decreased radioactivity at mid to basal inferior and mid to basal inferolateral segments (27±2% reduction, in comparison with those without torso) was noted in images without attenuation correction (NAC). Those artifacts could be completely corrected with attenuation correction (AC, mean improvement=27±6%). In addition, homogeneity could be restored after AC (CV, 19 to 9%). Meanwhile, in the phantom with a true filling defect, the reduction of radioactivity of MPI remained the same even after AC (NAC, 49% vs. AC, 52%).

Conclusions In the torso phantom study, the new method for attenuation map without transmission scan significantly improves attenuation-corrected perfusion images by using a CZT camera.