Left ventricular ejection fraction calculated from gated technetium-99-sestamibi SPECT.

BACKGROUND The accuracy of the left ventricular ejection fraction (LVEF) calculated from gated single photon emission computed tomography (GSPECT) studies is dependent on the accuracy of the determination of the end-diastolic volume (EDV) and end-systolic volume (ESV) of the left ventricle (LV). In this study we evaluated the feasibility of calculating the EDV, ESV and LVEF from the area under the polar graph (APC) of the edges of the LV image determined by a first derivative edge-detection method. METHODS AND RESULTS Technetium-99m ((99m)Tc) sestamibi GSPECT studies and planar equilibrium radionuclide ventriculography (ERNV) were performed on 15 male subjects in whom the LVEF ranged from 19% to 75%. Images were reconstructed to obtain short axis slices of the LV spanning the cardiac cycle. On each slice the LV edge points were determined at 10 degrees intervals using the APC method. The area of each short axis slice was determined by conversion to polar co-ordinates, interpolation and numerical integration of the graphs and multiplication by a pre-determined conversion factor. RESULTS Edges were successfully determined in all 15 patients using the APC method. The LVEF results correlate well with conventional planar ERNV studies (r = 0.96, LVEF(GSPECT) = 8.80 + 0.66 LVEF( ERNV)). The absolute difference between the LVEF for ERNV and for the APC method was 6.1% with a standard deviation of 7.6%. The reproducibility of SPECT LVEF using the APC method was good (intra-observer r = 0.99, inter-observer r = 0.99). CONCLUSIONS The APC method provides for easy and accurate ejection fraction determination with limited underlying mathematical assumptions. The ability to interpolate the edge points provides for stable edge detection even in hypoperfused myocardium.