J Nucl Med. 2007; 48 (Supplement 2):50P
Cardiovascular: Basic Science
Instrumentation and Data Analysis |
Reproducibility of absolute myocardial blood flow (MBF) quantitation using RB-82 PET
Georges El Fakhri1,
Nathalie Abi-Hatem2,
Bastien Guérin1,
Arkadiusz Sitek1 and
Marcelo Di Carli1
1 Radiology/Nuclear Medicine, Harvard Medical School, Brigham&Women's Hospital, Boston, Massachusetts; ;
2 Kaslik University, Kaslik, Lebanon
167
Objectives: Absolute MBF quantitation with Rb-82 PET has great potential in the assessment of the extent and severity of ischemia in patients without known coronary artery disease (CAD) but little data exists regarding its reproducibility. The aim of this work was to assess the reproducibility of our quantitative approach and that of Rb-82 PET itself in repeated patient studies. Methods: 22 subjects (3 CAD patients and 19 with no evidence of CAD) were injected with 60±5mCi of Rb-82 and imaged twice, within 10 days, dynamically for 6 minutes at rest and during dipyridamole stress (IV infusion of 0.14 mg/kg/min for 4 min). Each rest or stress study was processed 4 times by each of 4 independent operators in order to estimate the left and right ventricular input functions and regional MBF (ml/g/min) in the standard 17-segment polar map using generalized factor analysis of dynamic sequences (GFADS) and 2-compartment kinetic modeling while modeling extraction fraction and spillover. We assessed the reproducibility of the GFADS+compartment analysis quantitative approach, by computing the mean standard deviation of MBF values over the 4 estimates obtained for each rest or stress study by each operator. The reproducibility of Rb-82 imaging itself was assessed at rest and during stress by comparing MBF (averaged over the 4 estimates) associated with the two consecutive scans. Results: The variability in estimation of coronary flow reserve (CFR=K1stress/K1rest) associated with our quantitative approach was 4.1±3.0% in LAD, 3.3±3.0% in RCA and 3.6±3.7% in LCX. The variability associated with the CFR estimate for Rb-82 PET itself was 12.1±6.6% in LAD, 11.8±7.6% in RCA and 9.9±9.4% in LCX. Likewise, the reproducibility of GFADS+compartment analysis quantitation for a given operator (Y= 1.011 X – 0.013, r=0.99) was significantly higher than the reproducibility of Rb-82 imaging itself in repeated rest/stress studies in all 3 territories (Y = 0.751 X + 0.461, r=0.854), p< 0.05. Conclusions: Absolute quantitation of MBF in dynamic Rb-82 PET using generalized factor and compartment analyses is reproducible within 12%.
Research Support (if any): Supported by grants from NIH (RO1-EB005876), AHA (GIA-0655909T) and Bracco Diagnostics.
