Abstract WP487: Retinal Vessel Tortuosity Correlates With White Matter Hyperintensities and Cerebral Microbleeds in Subjects With Cognitive Decline
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
Objective:
Retinal microvasculature analysis is suitable to study cognitive disorders given retinal and brain vessels share similar properties. Using novel automated technologies, we aimed to investigate whether automatically-quantified retinal vascular changes from retinal fundus photographs predict brain imaging abnormalities in patients with cognitive impairment; and what specific retinal vascular parameters are more predictive of dementia.
Methods:
In this cross-sectional study, 34 subjects with cognitive complaints underwent neuropsychometric cognitive testing, retinal fundus photography and brain MRI. Retinal vessel tortuosity index (VTI), inflection index (VII) and length (VL) were automatically computed. The number and volume of cerebral microbleeds (CMB) and white matter hyperintensities (WMHI) were measured using SPIN Software (SpinTech, Bingham Farms, MI). The diagnostic groups were compared for demographic, cognitive, retinal and brain imaging parameters using analysis of variance. The association between brain (intracranial volume (ICV), hippocampal volume (HV), lateral ventricle volume (LVV), WMHI and CMB (volume, number), Fazekas scores) and retinal vascular parameters were assessed by using multiple linear regression models.
Results:
Nine subjects were cognitively normal (CN) and 25 cognitively impaired (CI) (22 with mild cognitive impairment and 3 with dementia). VTI was the only retinal parameter and HV and LVV were the brain parameters that were significantly different between CN and CI subjects (p=0.02). VTI correlated with VL (p=0.02, r=0.33) and branching angle (p=0.004, r=-0.45). The cognitive domains predicted by VTI were abstraction and orientation. VTI correlated with total intracranial volume (p=0.04, r=-0.12). A significant association was demonstrated for VII and WMHI volume (p=0.03, r=0.007), but not Fazekas score. VII and VTI significantly correlated with CMB volume (p=0.04 and 0.03, respectively).
Conclusion:
Automatically-detected retinal vessel tortuosity indexes predict neuroimaging markers of microvascular disease and total intracranial volume in cognitively impaired individuals. Retinal imaging may be used to screen patients with cognitive disorders.
