Background: Rubidium-82 ((82)Rb) PET imaging has been proposed for routine myocardial blood flow (MBF) quantification. However, few studies have investigated the test-retest repeatability of this method. The aim of this study was to optimize same-day repeatability of rest MBF imaging with a highly automated analysis program (FlowQuant) using image-derived input functions and dual spillover corrections (SOC).
Methods: Test-retest repeatability of resting left-ventricle (LV) MBF was measured in patients (n = 27) with suspected coronary artery disease (CAD) and healthy volunteers (n = 9). The effects of scan-time, reconstruction, and quantification methods were assessed with correlation and Bland-Altman repeatability coefficients.
Results: Factors affecting rest MBF included gender, suspected CAD, and SOC (P < .001). Significant test-retest correlations were found using all analysis methods tested (r > 0.79). The best repeatability coefficient for same-day MBF was 0.20 mL/minute/g using a 6-minute scan-time, iterative reconstruction, SOC, resting rate-pressure-product (RPP) adjustment, and left atrium input function. This protocol was significantly less variable than standard protocols using filtered back-projection reconstruction, longer scan-time, no SOC, or LV input function.
Conclusion: Absolute MBF can be measured with good repeatability using FlowQuant analysis of (82)Rb PET scans with a 6-minute scan time, iterative reconstruction, dual SOC, RPP-adjustment, and an image-derived input function in the left atrium cavity.