Introduction

Cerebral small vessel disease (SVD) affects arterioles, capillaries, and venules and can lead to cognitive impairments and clinical symptomatology of vascular cognitive impairment and dementia (VCID)[1,2]. VCID symptoms are similar to Alzheimer’s disease (AD) but the neurophysiologic alterations are less well studied. In this study we used 3D pseudo-continuous ASL (pCASL) perfusion MRI to assess cerebral blood flow (CBF) alterations in association with cognitive impairment in a cohort of aged Latino subjects with varying risks of vascular diseases. Furthermore, we estimated the test-retest repeatability of regional CBF for an interval of ~6weeks.

Methods

50 participants (mean+/-std age: 69.02+/-6.83years; 12m/38f, all Latinos) completed 1^st visit MRI, 37 completed 2^nd visit MRI for retest assessment (interval 45.8 days +/- 32.6 days). 27 subjects completed full neurocognitive assessment including Cognitive dementia rating (CDR), Montreal Cognitive Assessment (MoCA) and NIH Toolbox Flanker Inhibitory Control and Attention Test (Flanker) for executive control and attention, as well as the Dimensional Change Card Sort Test (DCCS) for cognitive flexibility and attention. Furthermore, we assessed whether or not vascular risk factors are present (=1) or not (=0): diabetes, hypertension and hyperlipidemia. MRI was performed on a 3T Siemens Prisma scanner with a 20channel head-coil including an MPRAGE with 1mm³ isotropic resolution, and a 3D GRASE pCASL with 2.5mm³ isotropic resolution, 48 slices, TR/TE 4300ms/36.76ms, Label Time 1500ms, Post Label Delay 2000ms. One M0 image and 8 label/control image pairs were acquired for ASL. ASL data were preprocessed and quantified using SPM12 and in-house Matlab scripts. CBF maps from 1st and 2nd visit were coregisterd to individual T1 and normalized to the MNI template space. CBF values for the main vascular perfusion territories were extracted [3,4](Fig. 1). We also calculated the GM and WM CBF based on tissues probability maps thresholded at 30% and 99% respectively. Correlation between average regional CBF from both test and retest scans and clinical/behavioral assessments were evaluated using mixed effects linear regression model implemented in STATA 13.1 (College Station, Texas), incorporating age, gender and globalCBF (average of GM and WM CBF) as covariates and time (test/retest) as the random variable. Finally, test-Retest repeatability of GM, WM and regional CBF was assessed by Interclass correlation coefficients (A-k criterion).

Results

A high test-retest repeatability was observed for regional CBF measurements ~6weeks apart (Figure2). The ICC for GM and WM was 0.82 and 0.76, respectively. For regional CBF in vascular territories, ICC was on average 0.76 (stdev=0.048). We found significant positive correlations between MoCA scores and regional CBF in, left ACAPerf, leptoMCA and MCAperf as well as bilateral ACHA and POCA territories. (Table1/Figure1). Significant negative correlations were observed between regional CBF and CDR sum-of-boxes scores in left MCA perf (beta=-1.96 p=0.014) and bilateral ACHA (beta=-0.89/-0.95 p=0.025/0.035 for right/left) For CDR global scores no significant correlations were found. Vascular risk factors only showed a significant negative correlation with CBF for diabetes in right MCAperf territory (beta=-2.00 p= 0.038). Cognitive tests only showed a weak association between Flanker and lepto PCA (beta=-0.09 p=0.016).

Discussion

The most consistent finding was regional CBF in subcortical perfusion territories including MCA perforator and ACHA territories related to CDR sum-of-boxes, MoCA as well as diabetes. Other cortical perfusion territories only showed correlations with MoCA but not CDR or vascular risk factors. In our cohort cognitive impairments were mild with CDR being 0 or 0.5 hence MoCA might provide a more sensitive assessment of overall cognitive state than CDR or tests for specific cognitive function such as Flanker and DCCS. The MCA perforator territory are fed by the lenticulostriate arteries, which are end arteries with almost no collaterals that could compensate for impaired perfusion due to SVD. Therefore, the CBF measurements in MCAperf territory might be especially susceptible to vascular impairment caused by SVD. Furthermore, the lenticulostriate arteries are known to be involved in silent strokes, which contribute to progressive cognitive impairment in elderly persons [5]. Finally, in this study we showed that both global and regional CBF are highly reproducible in this cohort of elderly individuals, supporting that 3D pCASL can serve as a viable imaging marker for monitoring the progression and treatment effects of SVD.

Conclusion

MCA perforator perfusion measured by 3D pCASL may serve as a potential imaging marker to identify early small vessel changes related to vascular cognitive impairment and dementia.

Acknowledgements

MarkVCID consortium and grant NINDS UH2NS100614