Reliability of left ventricular ejection fraction calculated with gated myocardial perfusion single photon emission computed tomography in patients with extensive perfusion defect

Abstract

INTRODUCTION: Gated myocardial perfusion scintigraphy (g-MPS) provides functional information on the left ventricle (LV) apart from the perfusion status. Computer algorithm-based LV ejection fraction (EF) calculated from resting g-MPS has been found to be reliable in patients with normal perfusion. However, its role in patients with extensive perfusion defects is not clear. AIM: To find the reliability of LVEF calculated from resting g-MPS in patients with extensive perfusion defects (>25% of LV myocardium) and to correlate this with echocardiography and multigated radionuclide ventriculography (RNV). MATERIALS AND METHODS: Thirty patients with fixed perfusion defect of size greater than or equal to 25% of LV myocardium on rest g-MPS study were included. EF was calculated using three software packages: Emory Cardiac Toolbox, Myometrix, and quantitative gated single photon emission computed tomography (SPECT)/quantitative perfusion SPECT. The patients underwent RNV (gold standard) and echocardiography within a week of the g-MPS. Agreement among the EF values obtained by different methods was determined using Bland–Altman analysis. Correlation among the EF values was measured using Spearman's rank correlation. RESULTS: Thirty patients (23 male; seven female; mean age 51 years, range 32–70 years) were included prospectively. The average size of perfusion defect was 38% of the LV myocardium (range 25–56% of LV). The average ejection fraction values were 33% for Emory Cardiac Toolbox (range 11–50%), 31% for Myometrix (range 18–46%), and 33% for quantitative gated SPECT/quantitative perfusion SPECT (range 17–49%). The mean EF on echocardiography was 37% (range 22–60%), whereas that of RNV was 33% (range 10–50%). The data were normally distributed. There was statistically significant positive agreement between algorithm-based EF measurements to the gold standard RNV. CONCLUSION: This study suggests that EF calculations, from g-MPS SPECT data using different software, have high agreement with the gold standard RNV even in patients with extensive perfusion defects.