Estimation accuracy of ejection fraction in gated cardiac SPECT/CT imaging using iterative reconstruction with 3D resolution recovery in rapid acquisition Protocols

Iterative reconstruction methods with 3D resolution recovery, and attenuation and scatter compensations are now common in clinical practice. Still, the reconstruction of ECG gated cardiac SPECT data is often done with filtered backprojection (FBP) to save reconstruction time. The reconstruction times of the 2007 release of Flash3D (Siemens' OSEM reconstruction with 3D distance dependent resolution recovery and optional scatter and attenuation corrections) have been significantly improved and it is now possible to process an entire clinical gated cardiac data set in clinically acceptable times. It has been shown that Flash3D enables the use of rapid acquisition protocols, yet maintains clinical diagnostic ability. In this work we evaluated the estimation accuracy of the cardiac ejection fraction of gated myocardial SPECT/CT perfusion images, acquired with currently used clinical protocols, as well as rapid acquisition protocols. For this, we acquired ECG gated image data of a dynamic cardiac torso phantom manufactured by Data Spectrum. Images were reconstructed using FBP and the new Flash3D with and without scatter and attenuation corrections. We analyzed the reconstructed volumes of the dynamic phantom using our own quantitative analysis tool as well as 4D-MSPECT and compared the image-based estimated ejection fraction (EF) to the true ejection fraction delivered by the dynamic phantom. Results show high correlation (r
²
>0.97) between conventional clinical protocols with FBP reconstruction and half time protocols using Flash3D except for 16 gates an 128times128 matrix (r
²
=0.95). Best ejection fraction accuracy was found for 16 gates and 64times64 matrix with a mean deviation of 0.5plusmn1.2% (full time FBP) 0.9plusmn1.6% (half time Flash3D) from the true value.