Enhanced arterial spin labeling in mild cognitive impairment phenotypes

Abstract

AbstractBackgroundArterial spin labeling perfusion MRI is increasingly proving to be a promising tool for exploring neurovascular changes in neurodegenerative conditions. Using an advanced 3D pseudo‐continuous arterial spin labeling sequence with Hadamard‐encoded multiple post‐labeling delays (eASL), this study evaluated both cerebral blood flow (CBF) and arterial transit time (ATT) alterations in mild cognitive impairment (MCI) phenotypes.MethodMCI patients (N = 19;Mage = 74.63±5.65) and controls (N = 20;Mage = 70.04±7.06) underwent a 3T MR scan with the eASL sequence. ATT and CBF were simultaneously measured. Voxel‐wise group comparisons of CBF and ATT were performed. Results were thresholded at a cluster‐size corrected p<0.001 (a<0.05). A region of interest (ROI) was defined as the overlap of clusters showing a significant difference in CBF and ATT between MCI and controls. Independent t‐tests between MCI phenotypes (amnestic, non‐amnestic) and controls on CBF and ATT ROIs were conducted. Pearson correlations of ROIs with measures of cognition, including executive functioning, were calculated. Correlations informed regression analysis.ResultsVoxelwise analyses (Figure1) demonstrated decreased CBF in MCI compared to controls in the default mode network, middle and superior frontal gyri, inferior parietal lobule, and temporal lobe. Increased ATT was seen in similar regions, with clusters centered in inferior and middle temporal lobe and visual cortex. T‐tests using created ROIs (as described above) revealed group differences between controls and MCI phenotypes in both CBF and ATT (ps≤.01); there was a non‐significant trend for higher ATT in naMCI versus aMCI. The highest correlational values emerged between CBF and ATT ROIs with a measure of executive functioning (TrailsB). IVs in regression analysis were age (step 1) and ATT and CBF overlapping ROI values (step 2, stepwise); TrailsB = DV . Across the entire cohort, CBF (R2 change = .239, p<.001), not ATT, was included in the model.ConclusionAdvanced 3D pCASL with Hadamard encoded multiple PLDs (eASL) can be used to simultaneously measure CBF and ATT in MCI and controls. MCI patients showed different patterns of reduced CBF and prolonged ATT. While CBF and ATT values were significantly correlated with TrailsB, regression analysis revealed that CBF was a more robust predictor of TrailsB. Future research should examine eASL as a biomarker for MCI phenotypes.