Reference Ranges for LVEF and LV Volumes from Electrocardiographically Gated 82Rb Cardiac PET/CT Using Commercially Available Software

Electrocardiographic gating is increasingly used for Rb-82 cardiac PET/CT, but reference ranges for global functional parameters are not well defined. We sought to establish reference values for left ventricular ejection fraction (LVEF), end systolic volume (ESV), and end diastolic volume (EDV) using 4 different commercial software packages. Additionally, we compared 2 different approaches for the definition of a healthy individual. Methods: Sixty-two subjects (mean age +/- SD, 49 +/- 9 y; 85% women; mean body mass index +/- SD, 34 +/- 10 kg/m(2)) who underwent Rb-82-gated myocardial perfusion PET/CT were evaluated. All subjects had normal myocardial perfusion and no history of coronary artery disease (CAD) or cardiomyopathy. Subgroup 1 consisted of 34 individuals with low pretest probability of CAD (<10%), and subgroup 2 comprised 28 subjects who had no atherosclerosis on a coronary CT angiogram obtained concurrently during the PET/CT session. LVEF, ESV, and EDV were calculated at rest and during dipyridamole-induced stress, using CardIQ Physio (a dedicated PET software) and the 3 major SPECT software packages (Emory Cardiac Toolbox, Quantitative Gated SPECT, and 4DM-SPECT). Results: Mean LVEF was significantly different among all 4 software packages. LVEF was most comparable between CardIQ Physio (62% +/- 6% and 54% +/- 7% at stress and rest, respectively) and 4DM-SPECT (64% +/- 7% and 56% +/- 8%, respectively), whereas Emory Cardiac Toolbox yielded higher values (71% +/- 6% and 65% +/- 6%, respectively, P < 0.001) and Quantitated Gated SPECT lower values (56% +/- 8% and 50% +/- 8%, respectively, P, 0.001). Subgroup 1 (low likelihood) demonstrated higher LVEF values than did subgroup 2 (normal CT angiography findings), using all software packages (P < 0.05). However, mean ESV and EDV at stress and rest were comparable between both subgroups (p = NS). Intra-and interobserver agreement were excellent for all methods. Conclusion: The reference range of LVEF and LV volumes from gated Rb-82 PET/CT varies significantly among available software programs and therefore cannot be used interchangeably. LVEF results were higher when healthy subjects were defined by a low pretest probability of CAD than by normal CT angiography results.