Automatic and visual reproducibility of perfusion and function measures for myocardial perfusion SPECT

Abstract

Background. We define the repeatability coefficients (RC) of key quantitative and visual perfusion and function parameters that can be derived by the QGS/QPS automated software and by expert visual observer from gated myocardial perfusion SPECT (MPS) scans. Methods. Standard QGS/QPS algorithms have been applied to derive quantitative perfusion and function parameters in 200 99mTc-tetrofosmin rest/stress MPS scans, obtained in 100 consecutive patients who underwent 2 separate gated rest/stress scans on the same camera. Variables included stress, rest, and ischemic total perfusion deficit (TPD), ejection fraction, motion, and thickening. Visual perfusion/motion scores were derived by an expert reader using randomized scan order and normalized to % myocardium. Results. Quantitative and visual parameters were highly reproducible with smaller RC for some quantitative measures as compared to visual measures (P < .0001). RC for quantitative measures were 3.3% for stress TPD, 1.8% for rest TPD, and 3.2% for ischemic TPD and for visual scoring 4.8% for stress, 3.8% for rest, and 4.3% for ischemic (P ≤ .002). The results in each vessel territory showed that in the right coronary artery (RCA) territory the quantitative approach had improved reproducibility as compared to visual reading. Visual thickening scoring was more reproducible than motion scoring (P < .0001). Conclusions. This study demonstrates that standard perfusion and function parameters derived from MPS by visual or quantitative analysis are highly reproducible with some advantages to the quantitative approach. Copyright © 2010 The Author(s).