The effect of filtering and zooming on 201TI-gated SPECT was evaluated in patients with major myocardial infarction.
Methods: Rest thallium (TI)-gated SPECT was performed with a 90 degrees dual-head camera, 4 h after injection of 185 MBq 201TI in 32 patients (mean age 61 +/- 11 y) with large myocardial infarction (33% +/- 17% defect on bull's eye). End diastolic volume (EDV), end systolic volume (ESV) and left ventricular ejection fraction (LVEF) were calculated using a commercially available semiautomatic validated software. First, images were reconstructed using a 2.5 zoom, a Butterworth filter (order = 5) and six Nyquist cutoff frequencies: 0.13 (B5.13), 0.15 (B5.15), 0.20 (B5.20), 0.25 (B5.25), 0.30 (B5.30) and 0.35 (B5.35). Second, images were reconstructed using a zoom of 1 and a Butterworth filter (order = 5) (cutoff frequency 0.20 [B5.20Z1]) (total = 32 x 7 = 224 reconstructions). LVEF was calculated in all patients using equilibrium radionuclide angiocardiography (ERNA). EDV, ESV and LVEF were measured with contrast left ventriculography (LVG).
Results: LVEF was 39% +/- 2% (mean +/- SEM) for ERNA and 40% +/- 13% for LVG (P = 0.51). Gated SPECT with B5.20Z2.5 simultaneously offered a mean LVEF value (39% +/- 2%) similar to ERNA (39% +/- 2%) and LVG (40% +/- 3%), optimal correlations with both ERNA (r = 0.83) and LVG (r = 0.70) and minimal differences with both ERNA (-0.9% +/- 7.5% [mean +/- SD]) and LVG (1.1% +/- 10.5%). As a function of filter and zoom choice, correlation coefficients between ERNA or LVG LVEF, and gated SPECT ranged from 0.26 to 0.88; and correlation coefficients between LVG and gated SPECT volumes ranged from 0.87 to 0.94. There was a significant effect of filtering and zooming on EDV, ESV and LVEF (P < 0.0001). Low cutoff frequency (B5.13) overestimated LVEF (P < 0.0001 versus ERNA and LVG). Gated SPECT with 2.5 zoom and high cutoff frequencies (B5.15, B5.20, B5.25, B5.30 and B5.35) overestimated EDV and ESV (P < 0.04) compared with LVG. This volume overestimation with TI-gated SPECT in patients with large myocardial infarction was correlated to the infarct size. A zoom of 1 underestimated EDV, ESV and LVEF compared with a 2.5 zoom (P < 0.02).
Conclusion: Accurate LVEF measurement is possible with TI-gated SPECT in patients with major myocardial infarction. However, filtering and zooming greatly influence EDV, ESV and LVEF measurements, and TI-gated SPECT overestimates left ventricular volumes, particularly when the infarct size increases.