Background: A multicenter intercomparison assessment was made of the variation in left ventricular (LV) volumes and ejection fractions (EFs) obtained from gated myocardial perfusion single photon emission computed tomography (SPECT) of the 3-dimensional AGATE (Amsterdam gated) cardiac phantom.
Methods and results: The phantom was configured to produce 3 different standard end-systolic volume and end-diastolic volume combinations (50 mL and 120 mL, 90 mL and 160 mL, and 120 mL and 190 mL) with corresponding EF (58%, 44%, and 37%). Quantitative gated myocardial perfusion SPECT was performed with 39 SPECT systems in 35 departments. In the multicenter study, for all 3 filling conditions, a wide range of results was obtained. The EF was overestimated (by 1% to 15%), and both the end-systolic volume and end-diastolic volume were underestimated (by 1 to 65 mL). The extent of overestimation of EF was related to the extent of underestimation of the volumes and was independent of filling condition. The trend in error per center was comparable for all 3 filling conditions. Acquisition time per projection was the only independent predictor of the difference between measured and expected EF (P = .0001).
Conclusions: Care should be taken before extrapolation of published and accepted cutoff values for LV EF and volumes in clinical decision making. Results should be validated in each center and monitored for accuracy and consistency over time.