Background: This study investigates the clinical performance of routine 201Tl gated single photon emission computed tomographic (201Tl GSPECT) myocardial perfusion imaging. Equilibrium radionuclide angiography (ERNA) was used as the standard for comparison.
Methods and results: One hundred and seventy-two consecutive patients were submitted to both myocardial 201Tl GSPECT imaging, at stress and in redistribution, and ERNA. Left ventricular ejection fractions (LVEF) and regional wall motion were assessed from both stress and redistribution 201Tl GSPECT datasets, and from ERNA. Linear regression analysis showed a good correlation between LVEF calculated by ERNA and 201Tl GSPECT (r=0.73 at stress, r=0.75 in redistribution, P<0.0001). However, the 95% prediction intervals of 201Tl GSPECT LVEF from ERNA LVEF were wide (minimum 35.4% at stress and 33.2% in redistribution). Moreover, a difference in LVEF > or =10% between ERNA and 201Tl GSPECT was found in 26.4% of cases at stress and 28.6% of cases in redistribution. A fair agreement between ERNA and 201Tl GSPECT was found in regional wall motion assessment in segments with normal or mildly reduced tracer uptake (kappa=0.32 at stress and kappa=0.33 in redistribution). In segments with moderately to severely reduced tracer uptake, a moderate agreement was found in regional wall motion assessment between ERNA and 201Tl GSPECT (kappa=0.44 at stress and kappa=0.42 in redistribution).
Conclusions: Left ventricular function may be misinterpreted in a significant proportion of patients if the calculation of LVEF is based on 201Tl GSPECT. Moreover, the evaluation of regional wall motion by 201Tl GSPECT appears unsatisfactory.