Clinical evaluation of a new concept: resting myocardial perfusion heterogeneity quantified by markovian analysis of PET identifies coronary microvascular dysfunction and early atherosclerosis in 1,034 subjects

Abstract

Coronary endothelial dysfunction is an early marker of coronary artery disease (CAD) but its noninvasive assessment is limited. We tested the hypothesis that diffuse patchy heterogeneous resting myocardial perfusion by noninvasive cardiac PET, quantified objectively by Markovian homogeneity analysis, or its improvement during dipyridamole stress, is a predictor of even mild stress perfusion abnormalities, consistent with coronary microvascular dysfunction as an early marker of CAD.

Methods: Rest-dipyridamole PET with (82)Rb was performed on 1,034 consecutive subjects for possible CAD or follow-up, for second opinion on revascularization procedures, or for screening because of risk factors and on 50 healthy control subjects. Objective, automated software quantified myocardial PET perfusion images for (i) patchy diffuse perfusion heterogeneity by Markovian homogeneity analysis separately from, independently of, and around significant localized regional perfusion defects; (ii) size and severity of localized regional perfusion defects caused by flow-limiting stenosis; and (iii) the graded base-to-apex longitudinal perfusion gradient due to early diffuse CAD without flow-limiting stenosis. History of vascular risk factors was obtained for all subjects.

Results: Resting myocardial perfusion heterogeneity with a homogeneity index outside 1 SD of healthy reference subjects and its improvement with dipyridamole correlated closely with CAD documented by stress-induced regional myocardial perfusion abnormalities outside 1 SD independently of other risk factors by multivariate logistic regression analysis (P < 0.001), by multivariate linear regression analysis (P < 0.001), and by chi(2) analysis (P < 0.001). The relative odds ratios of having stress-induced myocardial perfusion abnormalities for a resting homogeneity index outside 1 SD of healthy reference subjects was highly predictive and substantially greater than for standard risk factors.

Conclusion: Patchy heterogeneous resting myocardial perfusion by noninvasive cardiac PET quantified objectively using Markovian homogeneity analysis, and its improvement after dipyridamole, are powerful independent predictors of even mild stress-induced perfusion abnormalities, more than standard risk factors, consistent with coronary microvascular dysfunction as an early marker of preclinical CAD for potential preventive treatment.