Abstract 19971: Myocardial Blood Flow and Coronary Flow Reserve by 82Rubidium Cardiac Positron Tomography: Validation versus Coronary Angiography
Objective: Absolute myocardial blood flow (MBF) and coronary flow reserve (CFR = stress MBF/rest MBF) can be quantified by analysis of dynamic PET/CT 82Rb (PET) flow data. This study compares a method of deriving global and regional MBF and CFR from PET, using factor analysis, with invasive angiographic findings and conventional perfusion analysis.
Methods: Data were evaluated for 26 consecutive pts with known/suspected coronary disease (17M, 9F, age 70+11 yrs) who had both coronary angiography and vasodilator stress CT attenuation-corrected PET. Data were collected for 20 3-sec frames, 5 12-sec frames and 7 30-sec frames. Flow data were analyzed by semi-automated algorithms which separated RV and LV blood pools and LV myocardial volumes using curve-fitting factor analysis. MBF and CFR values were obtained for each of the standard AHA/ACC 17 LV segments, which were then also grouped into coronary territories. Lumen diameter at angiography was evaluated qualitatively. Summed stress and rest scores (SSS, SRS), from Emory Cardiac Toolbox, quantified the size/intensity of perfusion defects. Unpaired t-tests compared MBF and CFR values versus angiography. Logistic regression assessed strength of association between CFR and angiography.
Results: Angiograms of 20/26 (77%) patients were abnormal (> at least one 50% stenosis): 18/26 (69%) involved the LCX, 15/26 (58%) the LAD, and 9/26 (35%) the RCA. ROC analysis demonstrated significant association between abnormal angiogram and CFR (ROC area = 79%±12%, p<0.0001). Global CFR (for all segments) was significantly reduced in pts with normal vs. abnormal angiograms (1.8±0.4 vs. 1.4±0.3, p=0.02). Regional CFR was reduced for normal vs. abnormal LCX coronary territory (1.6±0.3 vs. 1.2±0.2, p=0.003), but not for the LAD or RCA (p>0.05). Logistic regression indicated that CFR correctly characterized patients as having normal vs. abnormal angiograms in 77% of cases (p=0.02), including patients with normal vs. abnormal LCX (p=0.003). SSS and SRS did not predict abnormal angiography.
Conclusion: Preliminary results indicate that MBF and CFR by PET using factor analysis can accurately identified patients with abnormal coronary angiograms, especially for LCX. Further study is needed to validate the methodology
- © 2010 by American Heart Association, Inc.