Abstract 3196: Quantification of Myocardial Blood Flow (MBF) with PET-CT and Rb-82 -Validation in a Canine Model of Myocardial Ischemia
Background - Absolute quantification of MBF would be of significant value but has been considered difficult for Rb-82 in the past due to image noise degrading quality of time-activity curves. We re-evaluated this issue using a recent state-of-the-art clinical PET-CT system with sensitive LYSO crystals, which is expected to improve study quality.
Methods - 8 dogs were prepared with LCA stenosis, followed by hybrid PET-CT imaging using a GE Discovery Rx VCT scanner. Listmode acquisition was done for 8 min after infusion of 35–50mCi Rb-82 at rest and during adenosine (140μg/kg/min). A low dose CT was obtained for attenuation correction and contrast-enhanced CTA was performed for vascular coregistration. For MBF quantification, dynamic Rb-82 data were fitted to a 2-compartment model (2C) including spillover term. Alternatively, retention (RET) was calculated and extrapolated to flow using a previously published equation. Radioactive microspheres (MSP) were injected for assessment of true MBF. Results were compared in 25 LV regions.
Results - Out of 400 segments, 2C fitting was reasonable in 359. MBF was 1.6±1.0 (range 0.5– 4.0) ml/g/min for MSP, 2.0±1.5 (range 0.2–5.5) ml/g/min for 2C and 1.2±.6 (range 0.4 –2.6) ml/g/min for RET at rest. At stress, MBF was 1.9±1.5 (range 0.3– 6.4) ml/g/min for MSP, 1.8±1.4 (range 0.2–5.7) ml/g/min for 2C and 1.3±1.0 (range 0.2–7.0) ml/g/min for RET. 2C MBF was significantly correlated with MSP (r= 0.74, P<0.001, 2C=0.40 + 0.85* flow(MSP)), and Bland Altman analysis showed a mean difference of 0.46 ml/g/min. RET MBF was also correlated with MSP (r=0.58, P<0.0001, RET=0.64 + 0.36*flow(MSP)). Passing-Bablok regression showed no significant deviation from linearity for 2C (P>0.1), however, deviation was significant for RET with increasing error at higher flow values (P<0.01).
Conclusion - Quantification of MBF is feasible with Rb-82 using current state-of-the-art, sensitive PET-CT systems. Our data suggest that 2-compartment modeling is preferable over simplified retention analysis.