24-frame gated myocardial perfusion SPECT (GSPECT) with ordered subset expectation maximum (OSEM) and resolution recovery

Objectives Previously we reported that for half-time (6 min) reduced count rate GSPECT, OSEM incorporating resolution recovery (RR) and noise modeling provides GSPECT quality equal to standard full-time (12 min) OSEM. As an alternative to a reduced-time acquisition, we postulate that by combining OSEM-RR (Evolution®, GE Healthcare) with a full-time acquisition, 24 frame (fr) GSPECT is feasible, providing temporal resolution adequate to quantify ventricular functional parameters.

Methods 36 hypertensive adults without known CAD underwent 32 mCi Tc-99m sestamibi post-stress full-time GSPECT with (1) 8fr standard OSEM, then (2) 24fr acquisitions processed with standard OSEM and also OSEM-RR. Gated image quality was graded (1=poor to 4=excellent) based upon myocardial count homogeneity and endocardial edge definition. LVEF, EDV, and ESV were calculated using Emory Toolbox®. Using CardioGRAF®, global rates of peak emptying (PER) and peak filling (PFR), and diastolic synchrony (% coefficient of variation of time to peak filling among the 17 myocardial segments) were quantified.

Results 8fr OSEM image quality was good (3.3+/-0.7) but temporal resolution was inadequate to evaluate filling and emptying rates. Temporal resolution improved with 24fr OSEM, but image quality was poorer (2.2+/-0.8) due to low counting statistics. With 24fr OSEM-RR image quality (3.1+/-0.8) was nearly as good as 8fr OSEM. The pre-ejection period, diastasis plateau, and atrial kick were identified in 78%, 69%, and 53% of patients. PFR was abnormal (<2.5 EDV/sec) in 61% of these hypertensive patients.

Results In TABLE *=p<.0001 vs. 8fr OSEM.

Conclusions OSEM-RR 24fr GSPECT provides images of good quality from which functional parameters may be quantified.