OCTA-measured retinal microvascular features are associated with cSVD progression and worsening cognition after mild stroke, while baseline OCTA was associated with future brain and cognitive outcomes, suggesting that OCTA holds promise as a noninvasive tool for monitoring retinal microvascular pathology in stroke populations enriched for cSVD.
Key Findings
Results
Longitudinal reductions in retinal vessel density were associated with increases in white matter hyperintensity volume over one year.
Association found in both superficial and deep vascular layers of the retina
Superficial layer standardized β: -0.045 (95% CI: -0.067 to -0.023)
Deep layer standardized β: -0.050 (95% CI: -0.074 to -0.026)
Models adjusted for age, sex, systolic blood pressure, smoking, diabetes, stroke severity, disability, and stroke subtype
Analysis used linear mixed-effects models in a prospective longitudinal cohort
Results
Increases in retinal vessel density and reductions in retinal flow index were associated with increased basal ganglia perivascular space volume over time.
Both vessel density increases and flow index reductions showed associations with perivascular space (PVS) volume
PVS volume was measured as basal ganglia perivascular space volume on brain MRI
This finding was observed using linear mixed-effects models assessing longitudinal relationships
The direction of the vessel density association with PVS was opposite to that seen with WMH volume
Results
Lower MoCA scores over one year tracked with decreases in deep retinal vessel density.
Standardized β: 0.136 (95% CI: 0.036 to 0.236) for deep retinal VD association with MoCA over time
MoCA (Montreal Cognitive Assessment) was used as the cognitive outcome measure
This association was found specifically in the deep vascular layer, not reported for the superficial layer in this context
Assessment conducted at baseline and 1-year follow-up
Results
Lower baseline retinal vessel density in both superficial and deep layers was associated with worse MoCA scores at 1-year follow-up.
Superficial layer standardized β: 0.125 (95% CI: 0.043 to 0.207)
Deep layer standardized β: 0.242 (95% CI: 0.162 to 0.322)
The deep layer association was notably stronger than the superficial layer association
This finding indicates that baseline OCTA measures have predictive value for future cognitive outcomes after mild stroke
Methods
The study enrolled 189 participants with lacunar or mild cortical ischemic stroke who underwent OCTA, brain MRI, and cognitive assessment at baseline and 1 year.
Mean age of participants was 64.8 years; 37% were female
154 participants completed the 1-year follow-up assessment
Participants were recruited from a hospital stroke service with modified Rankin Scale ≤2
OCTA metrics included vessel density, foveal avascular zone area, vessel radius, tortuosity, and flow index from superficial and deep vascular layers
MRI outcomes included white matter hyperintensity volume, basal ganglia perivascular space volume, and mean diffusivity in normal-appearing white matter
Background
Cross-sectional studies have suggested associations between retinal microvascular abnormalities measured by OCTA and cerebral small vessel disease and cognitive impairment, but longitudinal relationships had remained unclear prior to this study.
This study was designed specifically to address the gap in longitudinal evidence
Two analytical questions were addressed: (1) whether retinal measures were associated with brain and cognitive markers over time, and (2) whether baseline retinal measures were associated with 1-year outcomes
The study population was enriched for cSVD by virtue of stroke subtype selection (lacunar or mild cortical ischemic stroke)
What This Means
This research suggests that the tiny blood vessels visible in the back of the eye (the retina) can reflect what is happening to blood vessels in the brain over time. Using a non-invasive eye-imaging technology called optical coherence tomography angiography (OCTA), researchers followed 189 people who had experienced a mild stroke over the course of one year, measuring both retinal blood vessel characteristics and brain changes on MRI scans alongside cognitive (thinking and memory) testing. They found that when the density of retinal blood vessels decreased over the year, there were corresponding increases in white matter hyperintensities — bright spots on brain scans that represent damage to the brain's wiring and are a hallmark of cerebral small vessel disease. Declining retinal vessel density also tracked alongside worsening cognitive scores.
Importantly, the study also found that the state of a person's retinal blood vessels at the time of their stroke predicted how their cognition would look a year later. People with lower retinal vessel density at baseline had worse cognitive scores at the one-year mark, suggesting that OCTA imaging taken shortly after a stroke could help identify who is at greater risk of cognitive decline. This research suggests that the retina may serve as a window into brain vascular health, and that OCTA — already used in eye clinics — could become a useful, non-invasive tool for tracking cerebrovascular disease progression and cognitive risk in stroke patients.
The practical implication of these findings is that eye scanning technology, which is widely available and does not require radiation or invasive procedures, may one day be used alongside or even as an alternative to brain imaging to monitor patients after stroke for signs of ongoing small vessel disease and cognitive decline. The authors note that further investigation in larger groups of patients is needed before OCTA could be adopted for this purpose in clinical practice.
Gibbon S, Giarratano Y, Hamid C, Arteaga-Reyes C, Clancy U, Wiseman S, et al.. (2026). Retinal Optical Coherence Tomography-Angiography and Longitudinal Changes in Cognition and Cerebral Small Vessel Disease.. Neurology. https://doi.org/10.1212/WNL.0000000000218502