Background: Increased left ventricular mass (LVM) has been correlated with adverse cardiac events, such as sudden cardiac death. However, LVM quantitation with widely utilized gated SPECT myocardial perfusion imaging (MPI) software, has little validation and clinical application. Thus, we compared LVM from two commonly employed gated SPECT packages [4D-MSPECT (4DM) and Quantitative Perfusion SPECT (QPS)] with the 3-dimensional reference standard, CT angiography (CTA).
Methods: Comparisons were made in 56 patients (mean age 61.4 +/- 14.6; 32% female) referred for dual-isotope or low-dose/high-dose Tc-99m-tetrofosmin rest/stress MPI and cardiac CTA (mean 1.5 +/-4.5 months apart). LVM measurement was performed for both CTA and MPI by two independent observers blinded to clinical information.
Results: Correlation with CTA was best for post-stress MPI than at rest; thus, post-stress values are reported. Values obtained with each of the techniques were very highly reproducible (interobserver correlation r = 0.99 for each technique). The mean LVM values were 142 g by CTA, 145 g by 4DM, and 135 g by QPS (P = NS for CTA vs SPECT, but P < .001 for 4DM vs QPS). There was moderately good correlation between CTA and SPECT LVM data (r = 0.74 and 0.72 for 4DM and QPS, respectively; both P < .001). However, on Bland-Altman analysis there was significant overestimation of lower values and underestimation of higher CT LVM values by both QPS and 4DM (both r = 0.68 and 0.69, P < .001). The limits of agreement relative to CT LVM were wide (-52.1 g to 64.1 g for QPS; and -60.0 g to 53.5 g for 4DM).
Conclusions: SPECT and CTA give reproducible measures of LVM. Using CTA as the reference standard, the mean SPECT LVM values are similar, but lower values are overestimated and higher values are underestimated. Thus, the SPECT values are not substitutable for CTA without mathematical correction.