Fast vs slow rubidium-82 infusion profiles and test-retest precision of myocardial perfusion using contemporary 3D cardiac analog PET-CT imaging.

Abstract

BackgroundOn legacy 2D PET systems utilizing a 50 mL/min Rb-82 profile, test-retest precision of quantitative perfusion is ∼10%. It is unclear whether Rb-82 infusion rate significantly impacts quantitative perfusion and/or image quality on modern analog 3D PET-CT systems. We aimed to determine whether the Rb-82 infusion profile significantly impacts test-retest precision of quantitative perfusion, perfusion metrics, and/or image quality on a modern analog 3D PET-CT scanner. MethodsNinety-eight volunteers from 3 distinct groups: healthy volunteers (Normals), patients with risk factors and/or coronary disease (Clinicals) and patients with prior transmural myocardial infarctions (Infarcts), underwent cardiac stress testing on an analog 3D PET-CT. Participants received 3 consecutive resting scans and 2 consecutive stress scans, minutes apart, with two randomly assigned Rb-82 infusion profiles: 50 mL/min (fast [F]) and 20 mL/min (slow [S]). Perfusion metrics (resting (rMBF) and stress myocardial blood flow (sMBF)) were calculated using HeartSee software. Coefficients of variance (COV), repeatability (RC), MBF and image quality metrics were compared. ResultsrMBF correlated well between F and S profiles, with intraclass correlation coefficients (ICC) ranging 0.91-0.93. sMBF was highly correlated between F and S profiles (ICC=0.97). Fast and slow profiles were associated with similar same-day test-retest precision (COV 11.5% vs. 11.3% (p=0.77); RC 21.5% vs. 22.6%, for F-F vs S-S). There were no clinically significant differences in MBF values between F and S profiles. Image quality metrics were similar between the 2 profiles. ConclusionsThere are no clinically significant differences in precision, perfusion metrics or image quality between Rb-82 fast and slow infusions using a contemporary analog 3D PET-CT.