Background/Objectives: Gated cardiac positron emission tomography (PET) synchronizes PET data to the cardiac cycle based on an electrocardiogram (ECG) signal, providing left ventricular (LV) functional and geometrical parameters. Nevertheless, image artifacts, due to cardiac-, breathing-, and/or patient-motion occurring during image acquisition, undermine the reliability and clinical utility of these parameters. This study aims to elucidate the effect of two motion correction (MC) tools, CardioFreeze (CF) and a data-driven motion correction (DDMC) prototype, on LV functional and geometrical parameters. Methods: ECG-gated rest/stress 13NNH3 PET/CT scans from forty patients with myocardial ischemia and thirty-nine patients with normal myocardial perfusion were included. The following four reconstructions were performed for each patient scan: without motion correction (NMC), with CF, DDMC, and DDMC & CF. Images were processed with Cedars-Sinai QPET software. Results: End-diastolic volume (EDV) in rest and stress increased significantly using DDMC. End-systolic volume (ESV) increased significantly, while LV ejection fraction (LVEF) decreased significantly using any MC tool, regardless of the phase. Shape index end-systole (SI ES) and shape-index end-diastole (SI ED) increased significantly when using MC, except for SI ED in rest, where DDMC did not cause any difference. Eccentricity index end-systole (ECC ES) in rest and stress increased significantly in patients with normal myocardial perfusion, while it did not differ in ischemic patients after applying MC. Conclusions: MC tools significantly increase ESV values and decrease EF values. The highest effect is observed with the combined use of DDMC & CF. Image quality is greatly improved when using MC, regardless of the method, particularly in patients with the highest myocardial displacement.
Martinez-Lucio et al. (Fri,) studied this question.