Effects of temporal sampling on PET myocardial blood flow estimates

Objectives Undersampling of the left ventricle (LV) blood pool time activity curve (TAC) can produce inaccurate estimates of myocardial blood flow (MBF) and coronary flow reserve (CFR) from dynamically acquired PET myocardial perfusion images. The objective of this study was to identify the minimum temporal sampling required for adequate estimation of MBF and CFR.

Methods We studied 10 normal volunteers undergoing dynamic stress/rest PET (Rb-82 or N-13 ammonia) imaging. All reconstructions were performed with attenuation-weighted 3D-OSEM (24 subsets, 3 iterations) without post-filtering. From a dynamic sequence of 32 images in two phases (24x5sec, 8x30sec), the injection duration (W) was estimated from the full-width at half maximum of the LV blood pool TAC. Fourier analysis of the first phase determined the longest sampling interval (Ts), which preserved 95% of the spectral energy, as a function of W. Analysis of dynamic images reconstructed with frame durations varying from 4 to 15 seconds (T1) for phase 1 and 30 to 120 seconds (T2) for phase 2. The LV blood pool and tissue TACs were sampled using ROI and fit to compartment models for MBF and CFR quantification using Corridor4DM software (INVIA). Temporal sampling effects on mean MBF, mean CFR, and their relative changes were evaluated.

Results The mean W was 15.5 +/- 8.0s and the mean Ts was 15.4 +/- 5.2s. Ts increased linearly with W with Ts=0.61*W+5.9 (R=0.94). First phase sampling periods T1 of greater than the smallest Ts of 9.0s produced increased stress MBFs and CFRs, while T1 less than 9.0s had relative changes of less than 5% in MBF and CFR. The MBF and CFR magnitude changes from varying T2 were less than 5%.

Conclusions A simple two-phase framing of dynamic PET images with temporal sampling for the blood pool phase as function of the injection bolus duration was shown to be feasible. The frame durations for the second phase can be as large as 120 seconds. Further investigation with more patients is warranted for clinical validation.