Quantification of the normal range of myocardial blood flow and flow reserve with 82rubidium versus 13N-ammonia PET

Coronary artery disease (CAD) can be diagnosed by comparing myocardial perfusion scans with a database defining the lower limit of normal myocardial blood flow and flow reserve (MFR). Both ¹³N-ammonia and ⁸²rubidium tracers can be used to generate flow images, however only ¹³N-ammonia has been fully validated for quantifying blood flow and MFR using compartmental models. Normal databases have thus only been reported using ¹³N-ammonia PET and compartmental modeling. This study aimed to establish a lower limit of normal MFR for an ⁸²Rb database using a compartmental model, and to determine if a simplified model would reduce the measured range of normal MFR for both tracers, improving identification of regional flow defects. 14 subjects with <5% risk of CAD underwent rest and dipyridamole stress ⁸²Rb and 13N-ammonia dynamic PET imaging in a randomized order within a 2-week period. MBF was quantified using a one-compartment model for ⁸²Rb, and a two- compartment model for ¹³N-ammonia. A simplified model was used to estimate the net retention rate for both tracers. Model- specific extraction functions were determined to obtain flow estimates. It was found that the retention reserve variability the was lowest and was equivalent for both tracers (plusmn 15% globally, plusmn 16% regionally) indicating that the retention model may be preferable for detection and localization of flow reductions. The two-compartment model for ¹³N-ammonia had the smallest normal MFR range (mean-2sd = 2.27 globally, 1.48 regionally) confirming its precision for absolute flow quantification.