Left ventricular ejection fraction and volumes from gated blood-pool SPECT: Comparison with planar gated blood-pool imaging and assessment of repeatability in patients with heart failure

Gated blood-pool SPECT (GBPS) has several potential advantages over planar radionuclide ventriculography (PRNV), including the possibility of greater repeatability of left ventricular ejection fraction (LVEF) and the noninvasive calculation of left ventricular end-systolic volume and left ventricular end-diastolic volume (LVEDV). The aim of this study was to assess the repeatability of LVEF and LVEDV from GBPS and to compare LVEF with those from PRNV. Methods: Fifty patients underwent PRNV and GBPS, 23 of whom also had repeated studies in the same session. GPBS studies were processed using the Cedars Sinai Quantitative Blood-Pool SPECT (QBS) software that automatically calculates LVEF and LVEDV. Automatic processing with QBS was successful in 70% of the GBPS studies, with the remaining studies processed using the manual option in QBS. All PRNV studies were processed using a manual processing technique. Results: Comparison of LVEF from PRNV and GBPS yielded a correlation coefficient of 0.80. Bland-Altman analysis demonstrated a mean difference of 0.74% +/- 7.62% (mean SD) between LVEF from the 2 techniques. The 95% limits of agreement are therefore -14.50% to +15.98%. The correlation between repeated measurements was 0.87 for GBPS and 0.95 for PRNV. Bland-Altman analysis revealed poorer repeatability for GBPS (95% limits of agreement, -9.63% to +14.97% vs. -4.66% to +5.92%; P = 0.003). The mean LVEDV was 198 94 mL, with a mean difference of 9 +/- 47 mL between repeated measurements. The 95% limits of agreement are therefore -85 to +103 mL. Conclusion: GBPS provides a less repeatable measurement of LVEF than PRNV. Repeatability of LVEDV measurements from GBPS is poor.