Precision of Myocardial Blood Flow and Flow Reserve Measurement During CZT SPECT Perfusion Imaging Processing: Intra- and Interobserver Variability

The aim of this study was to evaluate the reproducibility of myocardial blood flow (MBF) and myocardial flow reserve (MFR) measurement in patients referred for dynamic SPECT. Methods: We retrospectively analyzed patients referred for myocardial perfusion imaging. SPECT data were acquired on a cadmium zinc telluride–based pinhole cardiac camera in list mode using a stress (251 ± 15 MBq)/rest (512 ± 26 MBq) 1-d (99m)Tc-tetrofosmin protocol. Kinetic analyses were done with software using a 1-tissue-compartment model and converted to MBF using a previously determined extraction fraction correction. MFR was analyzed and compared globally and regionally. Motion detection was applied, but not attenuation correction. Results: In total, 124 patients (64 male, 60 female) were included, and SPECT acquisitions were twice reconstructed by the same nuclear medicine board-certified physician for 50 patients and by 2 different physicians for 74. Both intra- and interobserver measurements of global MFR had no significant bias (−0.01 [P = 0.94] and 0.01 [P = 0.67], respectively). However, rest MBF and stress MBF were significantly different in global left ventricular evaluation (P = 0.001 and P = 0.002, respectively) and in the anterior territory (P < 0.0001) on interuser analysis. The average coefficient of variation was 15%–30% of the mean stress MBF if the analysis was performed by the same physician or 2 different physicians and was around 20% of the mean MFR independently of the processing physician. Using the MFR threshold of 2, we noticed good intrauser agreement, whereas it was moderate when the users were different (κ = 0.75 [95% CI, 0.56–0.94] vs. 0.56 [95% CI, 0.36–0.75], respectively). Conclusion: Repeated measurements of global MFR by the same physician or 2 different physicians were similar, with an average coefficient of variation of 20%. Better reproducibility was achieved for intrauser MBF evaluation. Automation of processing is needed to improve reproducibility.