Respiratory motion-corrected quiescent period gating magnetic resonance attenuation correction (Q.MRAC) for enhancing quantitative accuracy in hepatic [18F]FDG PET/MRI: a dual-domain motion compensation study.

BACKGROUND: Respiratory motion-induced misalignment between positron emission tomography (PET) and magnetic resonance attenuation maps remains a critical challenge in abdominal PET/magnetic resonance imaging (PET/MRI) quantification. This study evaluated the performance of a novel dual-domain motion correction strategy combining emission-domain [quiescent phase gating PET acquisition (Q.Static)] and attenuation-domain [quiescent period gating magnetic resonance attenuation correction (Q.MRAC)] approaches in enhancing standardized uptake value (SUV) accuracy in hepatic fluorine-18 fluorodeoxyglucose ([18F]FDG) imaging. METHODS: Fifty patients underwent [18F]FDG PET/MRI with two respiratory motion correction protocols: dual-domain correction (Q.Static + Q.MRAC) and conventional MRAC [conventional static PET acquisition (Static) + conventional magnetic resonance attenuation correction (C.MRAC)]. Three 3-cm-diameter spherical volumes of interest (VOIs) were positioned in the right hepatic lobe, including the upper lobe (VOI1), middle lobe (VOI2), and lower lobe (VOI3) to evaluate SUV quantification, spatial alignment of attenuation maps against end-expiratory T2-weighted imaging (T2WI), and motion correction efficacy. Analysis of variance was applied for comparisons of the mean SUV (SUVmean) across VOIs and protocols, the Kruskal-Wallis test to compare the differences in SUVmean, paired t-tests to analyze alignment metrics, and the intraclass correlation coefficient (ICC) to measure interobserver consistency. RESULTS: VOI2 demonstrated minimal SUV variability [coefficient of variation (CV) =2.29-2.55%], validating its use as a reference region as per the relevant guidelines. Respiratory motion caused consistent SUV underestimation: VOI1 underestimation relative to VOI2 was 16.4% for Q.Static + Q.MRAC and 26.3% for Static + C.MRAC, while the maximum underestimation rate for VOI3 was 11.4% for Static + C.MRAC. Q.Static + Q.MRAC reduced VOI1 underestimation by 38.9% as compared to Static + C.MRAC (P=0.006). Q.MRAC significantly improved anatomical alignment as compared to C.MRAC (P<0.001), reducing mismatch ratio by 48.7% and the mismatch area by 52.4% and 33.6% in the coronal and axial views, respectively. Eight percent of Q.MRAC acquisitions failed due to irregular breathing. CONCLUSIONS: Dual-domain motion correction (Q.Static + Q.MRAC) significantly improves SUV quantification accuracy in the upper lobe of right liver. These findings establish a foundation for achieving greater accuracy for tumor assessment and metabolic imaging in clinical practice through the use of [18F]FDG PET/MRI.