CMR validation of left ventricular volumes and ejection fraction measured by the IQ-SPECT system in patients with small heart size

BACKGROUND: The IQ-SPECT system is equipped with multifocal collimators and uses ordered-subset conjugate gradient minimization (OSCGM) as its reconstruction algorithm, achieving a shorter acquisition time than conventional SPECT. Left ventricular ejection fraction (LVEF) is overestimated by conventional SPECT in patients with small heart size. In this study, we compared IQ-SPECT with conventional SPECT and cardiovascular magnetic resonance (CMR) for the estimation of LVEF in patients with small hearts (males: EDV ≤ 60 ml, ESV ≤ 25 ml; females: EDV ≤ 45 ml, ESV ≤ 20 ml). METHODS: The study consisted of 49 consecutive patients (20 normal and 29 with small heart size) undergoing gated myocardial perfusion imaging (GMPI) with a 99mTc-labelled agent during stress or rest to assess the risk of coronary artery disease (CAD). The data were reconstructed using filtered back-projection (FBP) for conventional SPECT and OSCGM for IQ-SPECT. ESV, EDV, and LVEF were calculated using quantitative gated SPECT (QGS). To determine the optimal ordered-subset reconstruction parameters, we compared the LVEF from SPECT to the corresponding measurement from CMR. RESULTS: EDV, ESV, and LVEF values obtained from IQ-SPECT and conventional SPECT showed that the results of these two forms of SPECT were significantly correlated, although the EDV and ESV obtained by IQ-SPECT were higher than those obtained by conventional SPECT. IQ-SPECT yielded lower LVEF measurements than conventional SPECT (normal heart size: 50.6 ± 4.3% vs. 73.4 ± 8.4%, P = 0.002; small heart size: 62.1 ± 7.8% vs. 75.0 ± 11.4%, P < 0.001). There were no significant differences in LVEF measurements made by IQ-SPECT and CMR (normal heart size: 50.6 ± 4.3% vs. 53.2 ± 5.8%, P > 0.05; small heart size: 62.1 ± 7.8% vs. 64.6 ± 8.8%, P > 0.05). Five subsets (S) and 12 iterations (I) did not differ significantly in LVEF between CMR and IQ-SPECT for patients with small hearts (64.6 ± 8.8% vs. 62.1 ± 7.8%, P = 0.120), while 3 S and 10 I were the best parameters for patients with normal heart size (50.6 ± 4.3% vs. 53.1 ± 5.8%, P = 0.117). CONCLUSION: With CMR as the standard, IQ-SPECT yields more reliable LVEF values than conventional SPECT for populations with small heart size. The best reconstruction parameters from IQ-SPECT were 5 S and 12 I for patients with small hearts.