Left ventricular ejection fraction from gated SPET myocardial perfusion studies: a method based on the radial distribution of count rate density across the myocardial wall

Abstract

Left ventricular ejection fraction (LVEF) can be derived from gated single-photon emission tomographic (SPET) myocardial perfusion studies using either manual or edge detection techniques. In the presence of severe perfusion defects, however, difficulties may be encountered. In this article a method based on the assumption that the average position of the myocardial wall can be localized by means of statistical analysis of the distribution count density, and not on edge detection, is used to measure LVEF. SPET myocardial perfusion images, gated in eight time bins, were recorded in 50 patients 60 min after the injection of 925 MBq technetium-99m tetrofosmin. Masking of non-myocardial structures and thresholding resulted in images in which only myocardial walls had significant non-zero values. The distance of the wall relative to the centre of the cavity was calculated in the three-dimentional space as the first moment of the count rate distribution along radii originating in the centre of the cavity. LVEF was calculated using, for each time bin, the sum of the cube of all distances as an estimate of the cavity volume. The method required minimal operator interventions and was successful in all patients, including those with severe perfusion defects. Intraobserver and interobserver variability was excellent, with regression coefficients of 0.97 and standard deviations of 4.5% and 4.7%, respectively. For 30 patients, the measurements were validated against planar equilibrium radionuclide angiography (ERNA) that was obtained within an interval of 1 week. LVEF ranged from 12% to 88%. Agreement between the two methods was excellent (LVEFERNA=1.05+0.92 LVEFGSPET, r=0.93, P=0.023, SEE=7.06). The Bland-Altman analysis did not show any apparent trend in the differences between ERNA and gated SPET over a wide range of ejection fractions. The standard deviation of the differences was 3. 1%. In addition no relationship was found between the two methods and the severity of perfusion defects. In conclusion, accurate measurements of LVEF are obtained from gated SPET perfusion images using a method based on statistical analysis of the count rate density. This method did not deteriorate even in the presence of severe perfusion defects and could therefore be used in following patients after myocardial infarction.