Background: This is the first longitudinal study to assess regional cerebral blood flow (rCBF) changes during the progression from normal control (NC) through mild cognitive impairment (MCI) and Alzheimer's disease (AD).
Objective: We aim to determine if perfusion MRI biomarkers, derived from our prior cross-sectional study, can predict the onset and cognitive decline of AD.
Methods: Perfusion MRIs using arterial spin labeling (ASL) were acquired in 15 stable-NC, 14 NC-to-MCI, 16 stable-MCI, and 18 MCI/AD-to-AD participants from the Cardiovascular Health Study (CHS) cognition study. Group comparisons, predictions of AD conversion and time to conversion, and Modified Mini-Mental State Examination (3MSE) from rCBF were performed.
Results: Compared to the stable-NC group: 1) the stable-MCI group exhibited rCBF decreases in the right temporoparietal (p = 0.00010) and right inferior frontal and insula (p = 0.0094) regions; and 2) the MCI/AD-to-AD group exhibited rCBF decreases in the bilateral temporoparietal regions (p = 0.00062 and 0.0035). Compared to the NC-to-MCI group, the stable-MCI group exhibited a rCBF decrease in the right hippocampus region (p = 0.0053). The baseline rCBF values in the posterior cingulate cortex (PCC) (p = 0.0043), bilateral superior medial frontal regions (BSMF) (p = 0.012), and left inferior frontal (p = 0.010) regions predicted the 3MSE scores for all the participants at follow-up. The baseline rCBF in the PCC and BSMF regions predicted the conversion and time to conversion from MCI to AD (p < 0.05; not significant after multiple corrections).
Conclusion: We demonstrated the feasibility of ASL in detecting rCBF changes in the typical AD-affected regions and the predictive value of baseline rCBF on AD conversion and cognitive decline.
Keywords: Alzheimer’s disease; arterial spin labeling; cerebral blood flow; longitudinal study; mild cognitive impairment; prediction.