Ultra-widefield imaging of the retinal vasculature in mild cognitive impairment and Alzheimer’s disease: A pilot study

Abstract

Purpose : Changes in the retinal vasculature in Alzheimer’s disease (AD) have previously been observed in small pilot studies using ultra-widefield (UWF) imaging. Our aim was to assess whether such vascular changes are evident in mild cognitive impairment (MCI) as well as AD, to highlight the potential of early retinal imaging biomarkers for use in future large-scale studies.

Methods : Using UWF imaging, the retinal vasculature was imaged in 90 healthy controls, 35 individuals with MCI and 48 subjects with AD. Images were analyzed using custom software (VAMPIRE-UWF). Central retinal artery equivalent (CRAE), central retinal vein equivalent (CRVE) and their ratio (AVR) were computed from the vascular network. Global branching complexity of arteriolar (FDa) and venular (FDv) network was computed along with measurements of global tortuosity of the main arteriolar (TortA) and venular (TortV) paths. Width gradients of main arteriolar (WGa) and venular (WGv) paths of superior temporal (ST), inferior temporal (IT), superior nasal (SN) and inferior nasal (IN) quadrants were also measured. One-way ANOVA (normal distribution) and Kruskal-Wallis (non-normal distribution) tests were used to compare groups. Post-hoc analysis was conducted on statistically significant results using Tukey-Kramer (for ANOVA) and Dunn-Sidak (for Kruskal-Wallis) tests to identify where the differences between groups occurred.

Results : A sparser venular network as indicated by a lower fractal dimension (FDv) was observed in AD compared to controls (right eye; p = 0.02). There was an increased narrowing of arterioles as indicated by a lower width gradient (WGa-SN) in MCI compared to controls (left eyes; p = 0.01) and AD compared to MCI (left eye; p = 0.05). There was an increase in CRVE in MCI compared to controls (right eye; p = 0.03) , but no evidence of a statistically significant difference in CRVE between AD and controls or between AD and MCI.

Conclusions : Although our sample size was modest, we observed significant differences and trends in several venular and arteriolar parameters. Measures need further validation and influence of within and between subject variations on vascular measurements, such as inherent retinal characteristics and age need to be evaluated. This pilot study indicates a utility for developing retinal vasculature measurements for future studies of MCI and AD.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.