Abstract
Animal studies have demonstrated a lower myocardial extraction fraction and earlier "roll-off' in the myocardial uptake of Tc-99m tetrofosmin compared to Tc-99m sestamibi, giving rise to concerns that the consequent reduction in defect contrast could result in underestimation of perfusion defects by Tc-99m tetrofosmin. There are no large studies comparing the two agents in the same group of patients. It has been suggested that differences between these two agents are most likely to manifest in the setting of mild to moderate coronary disease, where defect contrast is already relatively low, and during the use of vasodilator stress where the higher rates of myocardial blood flow induced make tracer roll off more likely. Accordingly, in this prospective, two-centre (UK and Canada) study, 81 carefully chosen patients with mild to moderate coronary stenosis (50-90% stenosis in up to 2 major epicardial vessels, no previous myocardial infarction) and 7 patients with less than 5% probability of coronary disease underwent dipyridamole SPECT imaging with Tc-99m sestamibi and Tc-99m tetrofosmin in random order. Images were interpreted blindly by a panel of independent, experienced readers, and were analysed visually to determine the presence of perfusion defects and reversibility, and quantitatively using the CEqual programme to detemine defect extent and severity. Two additional readers undertook a second blinded reading of the scans of patients recruited by the UK centre. These were also analysed using the Liege quantitative programme. This subgroup analysis was intended to serve as a control to corroborate the results of the larger study. Tc-99m sestamibi detected reversible perfusion defects in 365 segments compared with 285 by Tc-99m tetrofosmin (p<0.0001), and demonstrated a larger extent of perfusion defect as determined by the percentage of left ventricular myocardium involved (mean +/- SD, 15.8%+/- 12.3 and 12% +/-11.9 for Tc-99m sestamibi and Tc-99m tetrofosmin, respectively, p<0.03). The defect /normal wall ratio was significantly lower with Tc-99m sestamibi (0.60 +/- 0.15) compared to Tc-99m tetrofosmin (0.73 +/- 0.14, p = 0.01) indicating better defect contrast with Tc-99m sestamibi. In addition, Tc-99m sestamibi more often correctly predicted the presence of disease in more than one coronary artery. Diagnostic sensitivity was 63% and 58% for Tc-99m sestamibi and Tc-99m tetrofosmin, respectively (p=0.09), and the accuracy for detection of disease in individual coronary tenitories was similar. There was overall concordance of 78% (Kappa =0.67) and 86% (kappa = 0.67) for categorisation of scans and individual segments, respectively, as normal, reversible and fixed. The quality of images produced by the two agents was comparable. Such differences in ability of two commonly used perfusion tracers to determine the presence, extent and severity of reversible perfusion defects could have significant diagnostic and prognostic implications.