The quantification of myocardial blood flow (MBF) and coronary flow reserve (CFR) is crucial in the evaluation of coronary heart disease. Aim of this study was to evaluate the effect of reconstruction parameters on quantitative assessment of MBF and CFR by dynamic PET with 13N-ammonia, acquired with last-generation PET/CT systems that can account for time of flight (TOF) information and for a 3D model of the PET point spread function (PSF).
Thirty-six patients underwent both dynamic stress and rest cardiac PET studies with 13N-ammonia. Patients were imaged on one of two models of 64-slices PET/CT scanners (Discovery 710 and Discovery 690; GE Healthcare), in 3D list-mode, for a scan time of 10 minutes. A low dose CT scan was obtained before each emission scan, for attenuation and scatter correction.
For each study, seventeen dynamic frames were reconstructed: 12 x 10 s, 4 x 30 s, 1 x 6 min. Data were reformatted to 128x128x47 image volumes within a 42 cm field of view (3.27 mm voxel). Three different reconstruction settings were applied: (a) filtered back-projection (FBP) with a 4.8-mm Hanning filter; (b) 3D ordered subsets expectation maximization (OSEM) with 2 iterations, 18 subsets; (c) 3D OSEM with TOF and PSF, 2 iterations, 24 subsets.
To estimate stress/rest MBF and CFR, all images were analyzed with the PMOD software package Version 3.3 (PMOD Technologies), using the 1-compartment model.
Regression analysis was applied to correlate results among the three reconstruction settings. Variations of MBF and CFR estimations among the three different reconstructions were evaluated for statistical significance with the non-parametric Wilcoxon test for paired samples.
Linear regression analysis between paired series (a) vs. (b), (a) vs. (c) and (b) vs. (c) returned correlation values r in the range 0.97 – 0.99 for total stress and rest MBF and for CFR (P < 0.0005 for all regressions).
Non-parametric test returned the following results:
Stress MBF: (b) and (c) were found significantly higher than (a) (P < 0.0005 and P = 0.001, respectively); also (c) was found significantly higher than (b) (P = 0.005).
Rest MBF: (b) was found significantly lower than (a) (P = 0.037) and (c) significantly higher than (a) and (b) (P = 0.004 and P < 0.0005, respectively).
CFR: (b) and (c) were found significantly higher than (a) (P < 0.0005 in both cases), while no significant difference was found between (b) and (c) (P = 0.712).
For total stress/rest MBF and CFR values, excellent correlation was found between data obtained by FBP and by each of OSEM reconstruction settings. MBF estimation was found significantly higher when TOF and PSF corrections were applied, according to the expected signal-to-noise ratio and spatial resolution improvements. Further investigation is needed to evaluate which of the reconstruction settings better approaches the true quantitative value.