Objective: To measure left ventricular (LV) function parameters in heart of healthy rats by three different positron emission tomography (PET) imaging techniques and by magnetic resonance imaging (MRI).
Methods: ECG-gated microPET examinations were obtained in seven healthy rats with 2-deoxy-2-[(18)F]fluoro-D-glucose (FDG) for calculation of LV-function from the blood-pool phase of the dynamic recording (FDGBP), and also from the later myocardial uptake (FDGMyo). On subsequent days, we re-measured LV-function using the novel blood-pool tracer (68)Ga-albumin (AlbBP) and again by FDG (FDGMyo2) in one setting. Cine-MRI examination provided the reference standard measurement.
Results: The mean LV ejection fractions (LVEF) were 56 ± 3 (FDGBP), 55 ± 3 (FDGMyo), 56 ± 3 (FDGMyo2), 57 ± 3 (AlbBP), and 57 ± 2 (MRI). There were good to excellent correlations found between the LVEF-values as compared to MRI reference standard for FDGBP (r = 0.71), FDGMyo (r = 0.86) and AlbBP (r = 0.88). Both of the blood-pool methods significantly overestimated the magnitudes of end-diastolic-volume and end-systolic-volume, whereas FDGMyo matched closely to the MRI reference standard. There was no significant bias for both blood-pool methods and a minor negative bias for FDGMyo regarding the LV ejection fraction (LVEF) when compared to cine-MRI results. There was no significant difference between the means of FDGMyo and FDGMyo2 (P = .50).
Conclusions: Relative to reference standard MRI measurements of LVEF, there was excellent agreement between PET-based measurements, notably for the novel blood-pool tracer (68)Ga-albumin.