Background: Diagnosis of coronary disease and microvascular dysfunction may be improved by comparing myocardial perfusion scans with a database defining the lower limit of normal myocardial blood flow and flow reserve (MFR). To maximize disease detection sensitivity, a small normal range is desirable. Both (13)N-ammonia and (82)Rb tracers are used to quantify blood flow and MFR using positron emission tomography (PET). The goal of this study was to investigate the trade-off between noise and accuracy in both (82)Rb and (13)N-ammonia normal databases formed using a net retention model.
Methods: Fourteen subjects with <5% risk of CAD underwent rest and stress (82)Rb and (13)N-ammonia dynamic PET imaging in a randomized order within 2 weeks. Myocardial blood flow was quantified using a one-compartment model for (82)Rb, and a two-compartment model for (13)N-ammonia. A simplified model was used to estimate tracer retention, with tracer-specific net extraction functions derived to obtain flow estimates.
Results: Normal variability of retention reserve was equivalent for both tracers (±15% globally, ±16% regionally) and was lower in comparison to compartment model results (P < .05). The two-compartment model for (13)N-ammonia had the smallest normal range of global blood flow resulting in a lower limit of normal MFR = 2.2 (mean - 2 SD).
Conclusion: These results suggest that the retention model may have higher sensitivity for detection and localization of abnormal flow and MFR using (82)Rb and (13)N-ammonia, whereas the (13)N-ammonia two-compartment model has higher precision for absolute flow quantification.