Inter-institution preference-based variability of ejection fraction and volumes using quantitative gated SPECT with 99mTc-tetrofosmin: a multicentre study involving 106 hospitals

Abstract

Inter-institution reproducibility of gated SPECT quantification based on institutional preferences was evaluated. This sort of variability is crucial for a multicentre study when many hospitals are involved. A total of 106 institutes participated in this study and were grouped according to their use of five workstation types. Fifteen sets of 99mTc-tetrofosmin gated projection images with normal ejection fraction (EF) (approximately 70%, group A, n = 5), borderline low EF (approximately 50%, group B, n = 5) and low EF with large perfusion defects (approximately 30%, group C, n = 5) were prepared. The projection images were processed by QGS software in each institute based on its own routine settings. Based on 318 QGS results, the reproducibility of EF and volumes was analysed for each group and workstation. The reproducibility of EF was good in 14 of 15 cases, showing a standard deviation (SD) of <3.6%, and the coefficient of variance of the end-diastolic volume (EDV) was <9.3% in all cases. When the deviation from the average value was analysed, the difference between EF at each institute and the average EF of the workstation (dEF) showed an SD of 2.2-3.7% for each group. The ratio of the EDV divided by the average EDV (rEDV) showed an SD of 0.061-0.069 for each group. One case in group C that had a large anterior defect with low EF showed bimodal EF distribution in one of the five workstations. The SD of EF was workstation dependent, owing to the SPECT reconstruction conditions. The reproducibility in EF and volumes within a workstation was good, even though the gated SPECT preferences varied. This reproducibility study supports the use of gated SPECT as a standard of ventricular function in multicentre studies.