Repeatability of regional myocardial blood flow calculation in (82)Rb PET imaging

BACKGROUND: We evaluated the repeatability of the calculation of myocardial blood flow (MBF) at rest and pharmacological stress, and calculated the coronary flow reserve (CFR) utilizing (82)Rb PET imaging. The aim of the research was to prove high repeatability for global MBF and CFR values and good repeatability for regional MBF and CFR values. The results will have significant impact on cardiac PET imaging in terms of making it more affordable and increasing its use. METHODS: 12 normal volunteers were imaged at rest and during pharmacological stress, with 2220 MBq of (82)Rb each. A GE Advance PET system was used to acquire dynamic 50-frame studies. MBF was calculated with a 2-compartmental model using a modified PMOD program (PMOD; University Hospital Zurich, Zurich, Switzerland). Two differential equations, describing a 2-compartmental model, were solved by numerical integration and using Levenberg-Marquardt's method for fitting data. The PMOD program defines 16 standard segments and calculates myocardial flow for each segment, as well as average septal, anterior, lateral, inferior and global flow. Repeatability was evaluated according to the method of Bland and Altman. RESULTS: Global rest and stress MBF, as well as global CFR, showed very good repeatability. No significant differences were found between the paired resting global MBF (0.63 ± 0.13 vs. 0.64 ± 0.13 mL/min/g; mean difference, -1.0% ± 2.6%) and the stress global MBF (1.37 ± 0.23 vs. 1.37 ± 0.24; mean difference, 0.1% ± 2.3%). Global CFR was highly reproducible (2.25 ± 0.56 vs. 2.22 ± 0.54, P = not statistically significant; mean difference, 1.3% ± 14.3%). Repeatability coefficients for global rest MBF were 0.033 (5.2%) and stress MBF 0.062 (4.5%) mL/min/g. Regional rest and stress MBF and CFR have shown good reproducibility. The average per sector repeatability coefficients for rest MBF were 0.056 (8.5%) and stress MBF 0.089 (6.3%) mL/min/g, and average repeatability coefficient for CFR was 0.25 (10.6%). CONCLUSION: The results of the study show that software calculation of MBF and CFR with (82)Rb myocardial PET imaging is highly repeatable for global values and has good repeatability for regional values.