This study investigates the value of ECG-gated 99mTc-tetrofosmin SPECT in the assessment of resting left ventricular (LV) function by comparison with cine MRI.
Methods: Twenty-eight patients were recruited prospectively from those referred for routine myocardial perfusion scintigraphy. Eight had three-vessel coronary artery disease, two had two-vessel disease, five had single-vessel disease and thirteen had not previously undergone coronary angiography. Twelve patients had previous myocardial infarction. After i.v. injection at rest of 750 MBq 99mTc-tetrofosmin, ECG-gated tomograms (16 frames per cardiac cycle) were acquired after 30 min. A nine-segment model of the LV was used and images were interpreted by two observers independently. Wall motion was assessed using a six-point scale (including unclassified where no judgment was possible), and systolic wall thickening was assessed from count changes through the cycle using a five-point scale. Tracer uptake was scored using a four-point scale. Diastolic wall thickness was assessed using a four-point scale. Cine magnetic resonance images were acquired in the same planes and analyzed in an identical fashion.
Results: There was good overall agreement between the techniques for wall motion, thickness and thickening (kappa = 0.55-0.66), although 15 of the 252 (6%) segments were unclassified on radionuclide imaging. While there was absolute agreement in the assessment of all parameters in 10 patients with normal wall motion by MRI, agreement was less good in the 8 patients with three-vessel disease and poor left ventricular function (mean LVEF = 26%, mean LVEDV = 241 ml) (kappa = 0.37-0.48). Where tracer uptake was normal, there was good agreement between imaging, techniques (kappa = 0.64-0.75), but where uptake was absent or nearly absent, agreement was poor (kappa = 0-0.61), and 15 of 22 segments were unclassified on SPECT.
Conclusion: Gated 99mTc-tetrofosmin imaging provides an accurate assessment of myocardial wall motion, thickening and thickness in normal left ventricles but is less valuable in poorly functioning ventricles. Six percent of segments could not be assessed because of inadequate tracer uptake.