Purpose  Absolute quantification of myocardial blood flow expands the diagnostic potential of PET for assessment of coronary artery disease. ⁸²Rb has significantly contributed to increasing utilization of PET; however, clinical studies are still mostly analysed qualitatively. The aim of this study was to reevaluate the feasibility of ⁸²Rb for flow quantification, using hybrid PET-CT in an animal model of coronary stenosis.

Methods  Nine dogs were prepared with experimental coronary artery stenosis. Dynamic PET was performed for 8 min after ⁸²Rb(1480–1850 MBq) injection during adenosine-induced vasodilation. Microspheres were injected simultaneously for reference flow measurements. CT angiography was used to determine the myocardial regions related to the stenotic vessel. Two methods for flow calculation were employed: a two-compartment model including a spill-over term, and a simplified retention index.

Results  The two-compartment model data were in good agreement with microsphere flow (y = 0.84x + 0.20; r = 0.92, p<0.0001), although there was variability in the physiological flow range <3 ml/g per minute (y = 0.54x + 0.53; r = 0.53, p = 0.042). Results from the retention index also correlated well with microsphere flow (y = 0.47x + 0.52; r = 0.75, p = 0.0004). Error increased with higher flow, but the correlation was good in the physiological range (y = 0.62x + 0.29; r = 0.84, p = 0.0001).

Conclusion  Using current state-of-the-art PET-CT systems, quantification of myocardial blood flow is feasible with ⁸²Rb. A simplified approach based on tracer retention is practicable in the physiological flow range. These results encourage further testing of the robustness and usefulness in the clinical context of cardiac hybrid imaging.