Background: Because myocardial wall thickness is smaller than the spatial resolution of single photon emission computed tomography (SPECT) imaging, changes in myocardial wall thickness are related to changes in maximum pixel counts via the partial volume effect, allowing for quantification of regional systolic wall thickening. We have developed a new gated SPECT method for computing the global left ventricular ejection fraction (LVEF) based entirely on changes in maximum regional myocardial counts during systolic contraction. This new method is independent of endocardial edge detection or other geometric measurements.
Methods and results: In 23 patients the gated SPECT method was validated by comparison with radionuclide angiography. The correlation between computed LVEFs was excellent (slope = 0.97, r = 0.91). The measurement of LVEF by gated SPECT was highly reproducible, with minimal intraoperator (slope = 0.97, r = 0.97) or interoperator (slope = 1.00, r = 0.97) variability. Measurements of regional thickening indexes were also reproducible, with a mean intraoperator correlation coefficient of 0.89 +/- 0.05 (range 0.79 to 0.95) for the 14 myocardial regions. Finally, the measurement of LVEF was not significantly influenced by changes in reconstruction filter parameters over a range of cutoff frequencies from 0.16 to 0.28.
Conclusions: This new counts-based gated SPECT method for measuring global left ventricular systolic function correlates well with radionuclide angiography, is highly reproducible, and has theoretic advantages over geometric methods.