The relationship between ischemia-induced left ventricular dysfunction, coronary flow reserve, and coronary steal on regadenoson stress-gated 82Rb PET myocardial perfusion imaging

Abstract

BackgroundGated rubidium-82 (82Rb) positron emission tomography (PET) imaging studies are acquired both at rest and during pharmacologic stress. Stress-induced ischemic left ventricular dysfunction (LVD) can produce a significant decrease in left ventricular ejection fraction (LVEF) from rest to stress. We determined the prevalence on PET of stress LVD with reduced ejection fraction (EF) and its association with absolute global and regional coronary flow reserve (CFR), and with relative perfusion defect summed difference score (SDS). Methods and ResultsWe studied 205 patients with known or suspected coronary disease (120 M, 75 F, age 69 ± 13 years) who had clinically indicated rest/regadenoson stress 82Rb PET/CT studies. Data were acquired in dynamic gated list mode. Global and 17-segment regional CFR values were computed from first-pass flow data using a 2-compartment model and factor analysis applied to auto-generated time-activity curves. Rest and stress LVEF and SDS were quantified from gated equilibrium myocardial perfusion tomograms using Emory Cardiac Toolbox software. LVD was defined as a change in LVEF of ≤−5% from rest to stress. A subgroup of 109 patients also had coronary angiography. Stress LVD developed in 32 patients (16%), with mean EF change of −10 ± 5%, vs +6 ± 7% for patients without LVD (P < .0001). EF was similar at rest in patients with and without stress LVD (57 ± 18% vs 56 ± 16%, P = .63), but lower during stress for patients with LVD (47 ± 20% vs 61 ± 16%, P = .0001). CFR was significantly lower in patients with LVD (1.61 ± 0.67 vs 2.21 ± 1.03, Wilcoxon P = .002), and correlated significantly with change in EF (r = 0.35, P < .0001), but not with SDS (r = −0.13, P = .07). The single variable most strongly associated with high risk of CAD (i.e., left main stenosis ≥50%, LAD % stenosis ≥70%, and/or 3-vessel disease) was stress EF (χ2 = 17.3, P < .0001). There was a higher prevalence of patients with territorial CFR values ≤1.0, consistent with coronary steal, in the LVD group than in the non-LVD group (39% vs 12%, P = .001). ConclusionsLVD developed in 16% of patients undergoing 82Rb PET myocardial perfusion imaging, and was associated with multivessel coronary artery disease. There was a significant relationship between LVD and coronary blood flow during stress, with LVD corresponding to a low CFR. Territorial CFR ≤1.0 was more common in patients with LVD than those without, suggesting that coronary steal is an important pathophysiologic mechanism contributing to pharmacologic stress-induced LVD.