A sophisticated eye scanner might detect early signs of Alzheimer’s disease, at least according to a paper in the August 23 JAMA Ophthalmology. Researchers led by Gregory van Stavern and Rajendra Apte, both at Washington University, St. Louis, studied the architecture of the retina in normal adults. They detected thinning of the tissue and a dearth of teeny blood vessels in its center in people who had tested positive for brain amyloid. The finding hints that a simple eye test could help diagnose prodromal AD. “Though this was a small study and very preliminary, it is certainly telling us that we can detect retinal changes in the preclinical stage of Alzheimer’s disease,” van Stavern told Alzforum.

Others have previously reported changes to the retinal architecture in people already diagnosed with AD or with mild cognitive impairment (Cheung et al., 2014; Golzan et al., 2017). Van Stavern and colleagues wondered how early in disease such changes occur. First author Bliss Elizabeth O’Bryhim and colleagues recruited 30 cognitively normal older adults from the Knight Alzheimer’s Disease Research Center at WashU and examined their eyes by optical coherence tomographic angiography. OCT-A, now commonly available, combines the structural information on the eye obtained from tomography with measures of blood flow and microvasculature structure.

The volunteers, 16 of them women, were between 62 and 92 years old. The average age was 75. Seven had had positive florbetapir PET scans, while 10 had low cerebrospinal fluid Aβ42 levels. O’Bryhim found that the inner fovea was thinner in the Aβ-positive than -negative volunteers. A small depression in the middle of the retina, the fovea is where a person’s visual acuity is highest. It is normally devoid of blood vessels, and this foveal avascular zone was wider in amyloid-positive people, as well. The findings suggest that vascular and structural alterations in the retina occur in AD patients before the onset of symptoms, said van Stavern.

Maya Koronyo-Hamaoui, Cedars Sinai Medical Center, Los Angeles, said the study was novel and could have important implications. “It’s so exciting to see that changes in OCTA could be described so early on in disease,” she said. Koronyo- Hamaoui was not involved in this study. While others in the field agreed the results were enticing, they also noted some limitations, as did the authors themselves. Mojtaba Golzan, University of Technology Sydney, wrote to Alzforum that a major challenge for retinal scanning in preclinical AD is the occurrence of ocular comorbidities. “Glaucoma, diabetic retinopathy, and macular degeneration are all well-known ocular conditions that alter the retinal structure and microcirculation,” she wrote. People with these conditions were excluded from the study.

Nevertheless, van Stavern sees potential for an early noninvasive test that would be simpler than PET scans or lumbar taps. He thinks the foveal thinning might be a sign of neurodegeneration and the widening of the avascular zone might be more specific to AD, since Aβ is known to aggregate in blood vessels.

He said that larger cohorts and longitudinal analyses are needed to prove any connections. Cecilia Lee, Washington University, Seattle, agreed. “Longitudinal studies of larger cohorts are needed to confirm who progresses from preclinical AD to clinical manifestation of AD (versus who remains resilient), and to evaluate longitudinal changes in retinal imaging markers using OCTA,” she wrote (see full comment below). Just recently, other researchers used OCT to document thinning of the retina in Parkinson’s disease, for example (Aug 2018 news). 

Koronyo-Hamaoui thinks a combination of OCT and scans for retinal protein aggregates is the way to go. “The real opportunity I see here is to combine measurements that show vascular changes, which appear to happen very early on, with more specific amyloid or tau analysis,” she said. She has reported finding retinal plaques in AD patients and has founded a company, NeuroVision, to develop a retinal scanning test (Sep 2017 news). In a JAMA Ophthalmology editorial accompanying the paper, Christine Curcio, University of Alabama, Birmingham, noted that O’Bryhim and colleagues scanned only the very center of the retina, i.e., the macula. “Thus, obtaining imaging and histopathology data from the same affected retinal regions and layers in AD remains a priority for this line of research,” she wrote.

Others are not so convinced that amyloid plaques occur in the retina. “It’s unfortunate that people cite very limited studies of the retina compared to rigorous studies we have done by histology,” Koronyo-Hamaoui told Alzforum. She claimed to have data showing amyloid, tau, vascular amyloid, and inflammation in the retina. “All the signs you see in the brain you find in the retina,” she said.

Jurre den Haan from VU Medical Center, Amsterdam, noted that a major problem in this field is lack of agreed-upon tests. “I advocate for a multidisciplinary international workgroup to standardize retinal studies in terms of phenotyping, imaging, and postmortem protocols,” he wrote (see full comment below). “This could finally make studies from different labs comparable, and thus advance this emerging field of research.”—Tom Fagan


  1. O’Bryhim et al. evaluated retinal microvasculature of 14 patients with preclinical stages of AD and 16 controls using optical coherence tomography angiography (OCTA). Preclinical AD was defined by positive biomarkers (PET imaging for PiB or F-AV-45 compound or CSF analysis for Aβ42) with normal cognition. All participants had a Clinical Dementia Rating score of zero (cognitively normal) and were free from confounding eye conditions such as glaucoma, age-related macular degeneration, retinal laser therapy.

    The following OCTA parameters were analyzed in a total of 58 eyes: the thickness of nerve fiber and ganglion cell layer, inner and outer foveal thickness, vascular density, macular volume, and the area of the foveal avascular zone (FA). The authors found that the FAZ was larger and the inner fovea was thinner in participants with preclinical AD than controls.

    The study results are exciting and point to the potential of ophthalmic imaging in AD screening in the future. Longitudinal studies of larger cohorts are needed to confirm who progresses from preclinical AD to clinical manifestation of AD (versus who remains resilient) and to evaluate longitudinal changes in retinal imaging markers using OCTA.

  2. With interest we read the article of O’Bryhim and colleagues assessing the retina for possible non-invasive biomarkers in preclinical AD, adding to the literature changes in microvasculature as a possible early sign of retinal involvement in AD pathology. Great to see that the urgency of AD research on patient-friendly biomarkers has also reached the ophthalmological scientific community and that currently used AD biomarkers are being integrated in a retinal imaging study. We agree with the authors‘ caution in interpreting these results. With individual layer segmentation and different vascular parameters available in different anatomical regions of the retina, large amounts of parameters are derived from OCT imaging. Caution should therefore be taken in (over)interpreting findings of a selection of these parameters without correction for multiple testing. Future studies are needed to assess the reproducibility of these interesting findings.

    Given the discrepant findings between different studies assessing retinal involvement in different stages of AD (retinal layer thinning measured with OCT, microvasculature changes on OCT/fundus photography/OCT-A, and amyloid deposition in postmortem and in vivo studies), there is a need for  studies assessing cohorts in the light of the ATN criteria and taking important confounders for retinal changes into account (Jack et al., 2016). These confounders include age and age-related diseases such as diabetes mellitus, glaucoma, and age-related macular degeneration.

    We (in our center and in collaboration with UCL Dementia Research Center) are currently working on correlating retinal measurements with established AD biomarkers (atrophy on MRI, amyloid, and tau in CSF/on PET) in cohorts of patients in different stages of AD (preclinical, SCD, and AD/PCA) taking the above confounders in account. Hopefully these data can help separate disease from aging effects and unravel the precise role of retinal biomarkers for diagnosis, prognosis, and therapeutic read-out in AD.

    Lastly, I advocate for a multidisciplinary international workgroup aiming to standardize retinal studies in terms of phenotyping, imaging, and postmortem protocols. This could help improve comparability between different studies and advance this emerging research field further.


    . A/T/N: An unbiased descriptive classification scheme for Alzheimer disease biomarkers. Neurology. 2016 Aug 2;87(5):539-47. Epub 2016 Jul 1 PubMed.

  3. Due to a physiological and embryological connection between the eye and the brain, eye imaging, or more specifically scanning, the retina, the light-sensitive layer of the eye, in preclinical Alzheimer’s disease has gained momentum over the past few years. Many studies have utilized current techniques optimized in ophthalmology to investigate the association between retinal changes and signs of AD pathology in the brain. Optical Coherence Tomography (OCT), a noninvasive eye imaging test, uses light wave to reconstruct the structure of the retina. OCT angiography (OCTA) is a functional extension of OCT that is capable of visualising the microcirculation of the retina. O’Bryhim et al. have employed OCTA to study retinal alterations in preclinical AD and control participants. They use results from brain amyloid imaging and/or analyzing levels of amyloid in the cerebrospinal fluid (CSF) to determine an individual’s disease status.

    Results demonstrate specific retinal vascular and structural changes that occur in the preclinical AD group and not in the control group. While these are very exciting findings, there are a number of limitations that need to be addressed prior to establishing the validity of such biomarkers in preclinical AD screening. First, a consistent method to demonstrate preclinical AD presence needs to be used. This study uses two different brain-imaging tracers and/or analysis of the CSF to establish disease presence. A recent study has shown CSF analysis to be superior to brain imaging to determine amyloid-positivity (Lewczuk et al., 2017). Second, a significant proportion of cognitively normal individuals will have positive amyloid imaging with no other symptoms of cognitive impairment (Chételat et al., 2013). As such, a battery of tests including laboratory, imaging, and neuropsychology should be used to determine disease status. Third, a major challenge in using retinal scanning for preclinical AD screening is eliminating local ocular comorbidities. Glaucoma, diabetic retinopathy, and macular degeneration are all well-known ocular conditions that alter the retinal structure and microcirculation. While the study presented has excluded such patients, further longitudinal studies are required to determine the specificity of retinal screening for preclinical AD assessment. Finally, a larger sample size and group is required to draw a meaningful interpretation.

    Collectively, it appears that retinal scanning, taken with other established markers for AD, has merit in establishing the presence of AD. Whether retinal scans alone can be used to determine preclinical AD status needs further longitudinal studies to determine their potential in successfully sub-groupings in mild cognitive impairment—those who have AD-like changes, and those who don’t.


    . Cerebrospinal Fluid Aβ42/40 Corresponds Better than Aβ42 to Amyloid PET in Alzheimer's Disease. J Alzheimers Dis. 2017;55(2):813-822. PubMed.

    . Amyloid imaging in cognitively normal individuals, at-risk populations and preclinical Alzheimer's disease. Neuroimage Clin. 2013;2:356-65. Epub 2013 Mar 5 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Thinning Retina … Could It Be Parkinson’s?
  2. Retinal Plaques May Enable Noninvasive Screening for AD

Paper Citations

  1. . Microvascular network alterations in the retina of patients with Alzheimer's disease. Alzheimers Dement. 2014 Mar;10(2):135-42. Epub 2014 Jan 15 PubMed.
  2. . Retinal vascular and structural changes are associated with amyloid burden in the elderly: ophthalmic biomarkers of preclinical Alzheimer's disease. Alzheimers Res Ther. 2017 Mar 1;9(1):13. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Association of Preclinical Alzheimer Disease With Optical Coherence Tomographic Angiography Findings. JAMA Ophthalmol. 2018 Nov 1;136(11):1242-1248. PubMed.