The effect of filtrating and reconstruction method on the left ventricular ejection fraction derived from GSPET: a statistical comparison of angiography and echocardiography

Abstract

There are different protocols of reconstruction in myocardial gated imaging that produce different values of left ventricular ejection fraction (EF). We attempted to determine how the parameters of reconstruction affect the calculated EF. The results were statistically compared with the values obtained from angiography and echocardiography. In this retrospective study, the data from 23 patients were used. All the patients had the angiographic and the echocardiographic data within 2 weeks before the test. Imaging was performed using a single-head gamma camera using technetium-99 methoxyisobutylisonitrile. The image data were reconstructed using 50 different combinations of the ramp, Hanning, Butterworth, Wiener, and Metz filters. The ordered subset expectation maximization (OSEM) technique was also examined using 12 combinations of iteration and subset. The calculated EF values were analyzed and compared with the echocardiographic and angiographic results. The backprojection technique produced higher values of EF than those derived from echocardiography and angiography. The OSEM on the other hand produced lower values when compared with echocardiography and angiography. On using the backprojection technique, the maximum correlation between the values derived from gated single-photon emission tomography and echocardiography (r=0.88, P<0.01) and angiography (r=0.81, P<0.01) was observed when using the Metz filter (full width at half maximum=5 mm and order=9) and the Gaussian filter (alpha=3), respectively. In the case of the OSEM technique, the maximum correlation with both angiography and echocardiography was observed when using the iteration=2 and the subset=12. On the average, the backprojection technique produces higher values, and iteration technique produces lower estimation of the EF when compared with angiography and echocardiography.