Optical coherence tomography angiography as a potential screening tool for cerebral small vessel diseases

Abstract

Background The retina and the brain share anatomic, embryologic, and physiologic characteristics. Therefore, retinal imaging in patients with brain disorders has been of significant interest. Using optical coherence tomography angiography (OCTA), a novel quantitative method of measuring retinal vasculature, we aimed to evaluate radial peripapillary capillary (RPC) network density and retinal nerve fiber layer (RNFL) thickness in cognitively impaired patients and determine their association with brain imaging markers. Methods In this prospective cross-sectional study, a total of 69 patients (138 eyes) including 29 patients with amyloid-positive Alzheimer's disease-related cognitive impairment (ADCI), 25 patients with subcortical vascular cognitive impairment (SVCI), and 15 amyloid-negative cognitively normal (CN) subjects were enrolled. After excluding eyes with an ophthalmologic disease or poor image quality, 117 eyes of 60 subjects were included in the final analyses. Retinal vascular [capillary density (CD) of the radial peripapillary capillary (RPC) network] and neurodegeneration markers [retinal nerve fiber layer (RNFL) thickness at four quadrants] were measured using OCTA and OCT imaging. Brain vascular (CSVD score) and neurodegeneration markers (cortical thickness) were assessed using 3D brain magnetic resonance imaging. The CD and RNFL thickness and their correlation with brain imaging markers were investigated. Results The SVCI group showed lower CD in the temporal quadrant of the RPC network compared to the CN group (mean (SD), 42.34 (6.29) vs 48.45 (7.08);p = 0.001). When compared to the ADCI group, the SVCI showed lower CD in the superior quadrant (mean (SD), 60.14 (6.42) vs 64.15 (6.39);p = 0. 033) as well as in the temporal quadrant (ADCI 45.76, SVCI 42.34;p = 0.048) of the RPC network. The CD was negatively correlated with CSVD score in the superior (B (95%CI), - 0.059 (- 0.097 to - 0.021);p = 0.003) and temporal (B (95%CI), - 0.048 (- 0.080 to - 0.017);p = 0.003) quadrants of the RPC network. RNFL thickness did not differ among the groups nor did it correlate with cortical thickness. Conclusions and relevance The microvasculature of the RPC network was related to the CSVD burden. However, the RNFL thickness did not reflect cerebral neurodegeneration. Noninvasive and rapid acquisition of the OCTA image might have the potential to be used as a screening tool to detect CSVD.