Gated SPECT evaluation of left ventricular function using a CZT camera and a fast low-dose clinical protocol: comparison to cardiac magnetic resonance imaging

CZT technology allows ultrafast low-dose myocardial scintigraphy but its accuracy in assessing left ventricular function is still to be defined.

The study group comprised 55 patients (23 women, mean age 63 ± 9 years) referred for myocardial perfusion scintigraphy. The patients were studied at rest using a CZT camera (Discovery NM530c; GE Healthcare) and a low-dose ^99mTc-tetrofosmin clinical protocol (mean dose 264 ± 38 MBq). Gated SPECT imaging was performed as a 6-min list-mode acquisition, 15 min after radiotracer injection. Images were reformatted (8-frame to 16-frame) using Lister software on a Xeleris workstation (GE Healthcare) and then reconstructed with a dedicated iterative algorithm. Analysis was performed using Quantitative Gated SPECT (QGS) software. Within 2 weeks patients underwent cardiac magnetic resonance imaging (cMRI, 1.5-T unit CVi; GE Healthcare) using a 30-frame acquisition protocol and dedicated software for analysis (MASS 6.1; Medis).

The ventricular volumes obtained with 8-frame QGS showed excellent correlations with the cMRI volumes (end-diastolic volume (EDV), r = 0.90; end-systolic volume (ESV), r = 0.94; p < 0.001). However, QGS significantly underestimated the ventricular volumes (mean differences: EDV, −39.5 ± 29 mL; ESV, −15.4 ± 22 mL; p < 0.001). Similarly, the ventricular volumes obtained with 16-frame QGS showed an excellent correlations with the cMRI volumes (EDV, r = 0.92; ESV, r = 0.95; p < 0.001) but with significant underestimations (mean differences: EDV, −33.2 ± 26 mL; ESV, −17.9 ± 20 mL; p < 0.001). Despite significantly lower values (47.9 ± 16 % vs. 51.2 ± 15 %, p < 0.008), 8-frame QGS mean ejection fraction (EF) was closely correlated with the cMRI values (r = 0.84, p < 0.001). The mean EF with 16-frame QGS showed the best correlation with the cMRI values (r = 0.91, p < 0.001) and was similar to the mean cMRI value (49.6 ± 16 %, p not significant). Regional analysis showed a good correlation between both 8-frame and 16-frame QGS and cMRI wall motion score indexes (8-frame WMSI, r = 0.85; 16-frame WMSI, r = 0.89; p < 0.01).

Low-dose gated SPECT with a CZT camera provides ventricular volumes that correlate well with cMRI results despite significant underestimation in the measure values. EF estimation appeared to be more accurate with 16-frame reformatted images than with 8-frame images.