Quantitative perfusion assessment with arterial spin labeling magnetic resonance imaging for predicting cognitive decline in Alzheimer's disease: A systematic review and meta-analysis.

BACKGROUND

Arterial spin labeling (ASL) magnetic resonance imaging (MRI) is a potential tool for diagnosis and prediction of Alzheimer's disease (AD). This study evaluates effectiveness of ASL in differentiating AD patients or individuals with mild cognitive impairment (MCI) from healthy individuals with normal cognition (HC).

METHODS

After conducting literature search in electronic databases (Google Scholar, PubMed, and Science Direct), studies were selected by following precise eligibility criteria. Meta-analyses of standardized mean difference (SMDs) were performed to examine differences in cerebral blood flow (CBF) between individuals with AD, MCI, and HC. Meta-analyses of correlation coefficients were performed to achieve pooled correlation between CBF and cognitive performance measures.

RESULTS

Fifteen studies (671 HC, 539 MCI, and 463 CE individuals) were included. The ages of these individuals were: 70.6 [95 %CI: 66.6, 74.7], 71.9 [95 %CI: 67.8, 76.1], and 73.2 [95 %CI: 69.2, 77.3] years for HC, MCI, and AD individuals respectively. The percentages of women in HC, MCI and AD groups in these studies were 58 % [95 %CI: 52, 63], 54 % [95 %CI: 49, 60], and 54 % [95 %CI: 48, 61]. As a composite measure for multiple brain regions, CBF was significantly lower in AD patients compared to HC (SMD - 0.70 [95 %CI: -0.92, -0.49]; I2 = 0 %; p < 0.0001) or individuals with MCI (SMD - 0.62 [95 %CI: -0.86, -0.37]; I2 = 0 %; p < 0.0001). However, composite CBF was not significantly lower in individuals with MCI in comparison with HC (SMD - 0.17 [95 %CI: -0.42, 0.08]; I2 = 11.7 %; p = 0.190). In various parts of the brain including the cerebellum, cingulate, deep gray matter, limbic system, and cortices, CBF was significantly lower in AD patients in comparison with HC or MCI individuals. However, CBF was not significantly different in individuals with MCI in comparison with HC in some parts including deep gray matter, limbic system, and frontal lobe. Pooled correlation coefficients between CBF and any cognitive assessment test score were - 0.005 [95 %CI: -0.066, 0.057] (p = 0.878) for HC, -0.114 [95 %CI: -0.236, 0.009] (p = 0.070) for MCI, and 0.297 [95 %CI: 0.213, 0.381] (p < 0.0001) for AD groups.

CONCLUSIONS

ASL is a promising MRI method that can provide an alternative to nuclear methods for the diagnosis of AD. However, more studies are required to evaluate its detecting and predicting role in individuals with MCI.