Is Iterative Reconstruction an Improvement Over Filtered Back Projection in Processing Gated Myocardial Perfusion SPECT?

Abstract

The impact of iterative reconstruction (IR) on myocardial perfusion imaging (MPI) interpretation and func- tional results is under investigated. We evaluated the effect of IR on the interpretation and functional results of MPI com- pared with filtered back-projection (FBP). Material and Methods: Sixty patients with previously acquired MPI were ran- domly selected. Studies were processed using FBP and 2D ordered-subsets expectation-maximization (OSEM). Two phy- sicians interpreted the images after IR and FBP processing. Results were confirmed by coronary angiography and/or clini- cal course. Left ventricular ejection fraction (LVEF), end-diastolic volume (EDV) and end-systolic volume (ESV) were calculated with IR and FBP. Results: The sensitivity, specificity, positive and negative predictive values and accuracy of physician 1 were 95%, 87%, 80%, 97%, and 90% with FBP and 100%, 80%, 72%, 100%, and 87% with IR. For physician 2, they were 95%, 95%, 91%, 97%, and 95% with FBP and 81%, 95%, 90%, 90%, and 90% with IR. There were no statis- tically significant differences between physicians’ interpretations (P=0.71 for FBP and 0.09 for IR). There was good cor- relation between IR and FBP calculated LVEF (r=0.98), EDV (r=0.99), and ESV (r=0.99) but significant difference per patient (P=0.02, 0.03 and 0.02, respectively). Conclusion: There is no statistically significant difference in interpretation of MPI with IR versus FBP but there were statistically significant differences in functional results. Keywords: Myocardial perfusion scintigraphy, SPECT, Iterative reconstruction, Filtered back-projection, SPECT reconstruc- tion algorithms.