Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 31.
Published in final edited form as: Asia Pac J Ophthalmol (Phila). 2022 Mar-Apr;11(2):140–148. doi: 10.1097/APO.0000000000000505

Retinal Biomarkers for Alzheimer’s Disease: the Facts and the Future

Yuan Amy *, Cecilia S Lee *,†
PMCID: PMC9889204  NIHMSID: NIHMS1859938  PMID: 35533333

Abstract

Alzheimer’s disease (AD) is a significant cause of morbidity and mortality worldwide, with limited treatment options and considerable diagnostic challenges. Identification and validation of retinal changes that correlate with clinicopathologic features of AD could provide a non-invasive method of screening and monitoring progression of disease, with notable implications for developing new therapies, particularly in its preclinical stages. Retinal biomarkers that have been studied to date include structural changes in neurosensory retinal layers, alterations in vascular architecture and function, and pathologic deposition of proteins within the retina, which have all demonstrated variable correlation with the presence of preclinical or clinical AD. Evolution of specialized retinal imaging modalities and advances in artificial intelligence hold great promise for future study in this burgeoning field. The current status of research in retinal biomarkers, and some of the challenges that will need to be addressed in future work, are reviewed herein.

Introduction

Dementia is a significant cause of morbidity worldwide, afflicting 55 million people globally, and expected to increase in population to 78 million in the next decade. Alzheimer’s disease (AD) is the most common cause of dementia1, and manifests as a slowly progressive, irreversible neurodegenerative disease that is defined by specific histopathologic changes in brain tissue2. AD is now the sixth leading cause of death in the United States3. Current benchmarks for diagnosis of dementia and AD are inadequate, and the majority of cases of dementia remain undiagnosed4. The gold standard of diagnosis for AD remains neuropathologic examination of brain tissue. Post-mortem analyses have shown up to a 30% rate of misdiagnosis. Over the last decade, noninvasive imaging of cerebral amyloid using positron emission tomography (PET) and analysis of cerebrospinal fluid biomarkers have been established as valid biomarkers for pathologic states of cerebral amyloid deposition5. Nonetheless, current diagnostic metrics are limited by poor sensitivity and specificity6, lack of standardization and scalability, insufficient access to specialized imaging equipment and subspecialty expertise, and cost, to name just a few barriers1.

In spite of the profound burden of disease both to individuals and to society, treatment options for AD have been sparse, with historically approved pharmacological agents directed towards symptomatic management only. Disappointing clinical trial results over the last two decades have hindered attempts at bringing therapy that can alter disease progression to market. In June 2021, the first disease modifying pharmacological therapy, aducanumab, was granted accelerated approval by the US Food and Drug Administration in spite of equivocal clinical benefit in phase three clinical trials7. Aducanumab is an antibody against amyloid-β peptide, a hallmark pathologic feature of AD, and was shown to decrease amyloid-β burden as a surrogate for clinical efficacy. A post approval randomized trial to directly demonstrate effects on clinical endpoints is pending, but results will likely not be available for many years.

The controversy of adacanumab’s approval underscores multiple facets of the need for validated biomarkers in AD at different stages of disease. The pathological changes of AD are thought to begin decades before the manifestation of cognitive changes, accompanying a stage of disease that has been termed the “preclinical phase”8. With clinical manifestation of cognitive impairment, the spectrum of AD progresses from mild cognitive impairment (MCI) to frank AD dementia. MCI carries nearly a 50% chance of converting to AD within 30 months9, and the preclinical phase represents a window of opportunity to intervene before disease is clinically manifested10. Interest in identifying new biomarkers for preclinical and clinical disease is motivated not just by the need for improved diagnostic access and accuracy, but also by the dire need for metrics by which to identify and recruit patients for clinical trials designed to modify disease progression at early stages, as well as endpoints to assess the efficacy of potential therapeutic regimens.

Hypothesis for identifying retinal biomarkers

Alzheimer’s disease is characterized by abnormal extracellular deposition of misfolded amyloid-β protein as amyloid plaques and intracellular accumulation of hyperphosphorylated tau proteins as neurofibrillary tangles in the central nervous system, affecting the neurovascular unit and resulting in progressive loss of cognitive function. The downstream effect of these abnormal proteins, and how it results in neurodegeneration, is multifactorial and complex, with microvascular changes, inflammation, oxidative stress, and loss of neural tissue all implicated in the continuum of disease11–12. Given that the neurosensory retina is an extension of the central nervous system, the possibility of seeing manifestations of AD directly in the eye is a reasonable hypothesis that has prompted the study of retinal biomarkers over the last decade.

Early histopathologic evidence of changes in retinal tissue in AD was demonstrated in postmortem studies in 1990, which showed axonal degeneration in 80% of optic nerves, degeneration of retinal ganglion cells (RGC), and attenuation of retinal neural fiber layer (RNFL) thickness in AD patients compared to age matched controls, with predominant loss of M-type retinal ganglion cells (Figure 1)13–15. The authors posited that studying the retina and optic nerve might be valuable for better diagnostic and pathophysiologic understanding of AD. A proof of principle for using retinal imaging to identify AD came shortly thereafter, in which red-free fundus photography demonstrated RNFL defects in five of 22 patients with AD compared to only one of 24 age matched controls, and furthermore found a trend between increased optic nerve pallor and functional deficits as measured by the Alzheimer’s Disease Assessment Scale.16

Figure 1:

Figure 1:

Open in a new tab

Early histopathologic evidence of ganglion cell degeneration in Alzheimer’s disease (AD). Retinal cross sections from normal (A) and AD (B) patients demonstrates marked thinning of ganglion cell layer (GCL) and nerve fiber layer (NFL) in the AD patient. Reproduced with permission from Ref. 15.

Since these initial studies, evolution of retinal imaging technology and computational advances have driven the study of different retinal biomarkers to correlate with presence of preclinical or clinical AD, in a search for fast, noninvasive and accurate ways to identify disease. (Figure 2a) In parallel, the biochemical mechanisms underlying retinal pathology in AD continue to be elucidated, and are described well in recent publications17. This review will summarize the major recent advances by categories of biomarkers.

Figure 2:

Figure 2:

Open in a new tab

(A) Cross-sectional view of retina captured by optical coherence tomography (OCT). Assessment of the macular ganglion cell-inner plexiform layer (GC-IPL) and peripapillary retinal nerve fiber layer (RNFL) and in a patient with mild cognitive impairment and positive amyloid PET imaging is demonstrated in (B) and (C) respectively. Choroidal thinning in a cognitively normal patient compared to a patient with Alzheimer’s disease is demonstrated in (D) and (E), respectively (red arrows delineate the thickness of the choroid). Reproduced with permission from Ref. 75.

Evidence for retinal biomarkers

1. RNFL thickness

Peripapillary and macular RNFL thickness in AD has been the most extensively studied retinal biomarker in AD, with the preponderance of data supporting a correlation between decreased thickness in the presence of clinical AD. The first report of in in vivo application was in 2006, in which clinical exam and scanning laser ophthalmoscopy (SLO) evaluation of optic nerve fibers found a greater odds ratio for large cup-to-disc ratio in patients with AD, albeit with low sensitivity and specificity (45% and 84% respectively)18.

Spectral domain optic coherence tomography (SD-OCT) analysis of peripapillary and macular RNFL thickness emerged at the same time with a similarly sized study of 28 AD and 30 age-matched control eyes, and found that RNFL thickness in all quadrants of AD patients were thinner than control. (Figure 2c) This study further correlated decreased total macular volume with lower mini-mental state examination scores, suggesting a relationship between RNFL thickness and severity of cognitive impairment19. Since this initial OCT study, the majority of studies have found RNFL thickness in AD to be decreased on OCT, albeit with some discrepancies surrounding the topographic distribution of thinning. A meta-analysis of 24 studies that measured RNFL thickness in AD using SD-OCT found a significant reduction in mean peripapillary RNFL thickness20 in AD compared to controls.

The relationship between RNFL thickness in preclinical AD or MCI is more nuanced and less clear. A meta-analysis of 6 studies showed an overall trend of RNFL thinning in MCI patients, but the correlation was not statistically significant21, with some reports showing no change22–23 or paradoxical thickening24 in MCI vs control. One hypothesis for the observation of paradoxical thickening is that gliosis precedes neuronal loss and atrophy, as seen with AD histopathology findings in the brain25, thus the temporal changes in RNFL thickness in the years preceding manifest AD dementia is likely a more complex story than a simple linear relationship. Adaptive optics SLO might have a role in delineating the presence of gliosis to further elucidate the relationship between RNFL thickness in preclinical AD26.

Functional implications of RNFL thinning have been studied as a clinical endpoint of interest. Prospective population studies have shown a relationship between RNFL thinning and future risk of cognitive decline. The UK Biobank study found that thinner baseline RNFL measurements were associated with worse performance on baseline cognitive testing, and that the two thinnest quintiles were twice as likely to perform more poorly on cognitive testing at 3 year follow up compared to patients in the thickest RNFL quintile27. The Rotterdam study, evaluating the relationship between RNFL thickness and incident versus prevalent dementia in a Dutch population found that thinner RNFL at baseline was associated with a significant increased risk of developing AD, but interestingly did not find any association with RNFL thinning and prevalent dementia28. A third prospective study in Shanghai of preclinical and MCI patients found that greater reduction in RNFL thickness over a 25 month period was associated with deterioration of cognitive status29. A recent study of retinal structural changes in preclinical patients with in vivo AD biomarkers demonstrated a reduced RNFL thickness in cognitively normal patients with positive amyloid-PET/CT (amyloid-labeled C-PiB positron emission tomography/computed tomography imaging) compared to those with normal PET/CT30. Taken together, these studies suggest that RNFL thinning might potentially be useful as an early screening tool for incident development of dementia, and cautions against interpretation of associations seen in small cross sectional studies in determining the utility of this biomarker as a diagnostic tool.

However, if the intuition behind RNFL thinning serving as a biomarker for AD is that the neurosensory retina serves as a window for central nervous system degeneration in dementia, then it is important to note that RNFL thinning has been found in other neurodegenerative diseases as well and is likely not specific to AD31,32. Larger studies encompassing different causes of dementia will need to be performed to understand the role of RNFL thickness as a specific biomarker of AD vs any neurodegenerative conditions and the timeline in which the biomarker may be sensitive and/or specific for each condition.

2. Inner retinal layer and total macular thickness

In addition to RNFL thickness, other retinal layers have been studied as potential structural biomarkers in AD. The thickness of 10 retinal layers in 150 patient with AD were compared to age-matched controls, and correlations were found between the presence of AD and reduced thickness of RNFL, ganglion cell layer (GCL), inner plexiform layer (IPL), and the outer nuclear layer (ONL). Longer AD duration was correlated with greater thinning of RNFL, GCL and IPL33.

Some subsequent studies have shown decreased macular ganglion cell-inner plexiform layer (GC-IPL) thickness on OCT be more sensitive for identifying AD compared to RNFL thickness (Figure 2b). The Rotterdam study, for example, evaluated GC-IPL thickness in addition to RNFL thickness, and found that thinner GC-IPL was associated with the presence of dementia while this trend was not observed in the same patients for RNFL thickness18. For patients with either MCI or AD, macular GC-IPL thickness was found to be decreased compared to control, with greater area under the receiver operator characteristic (AUROC) for GC-IPL than RNFL34 thickness. Furthermore, there is some suggestion that analyzing the spatial distribution of GC-IPL thickness might be different in patients with MCI compared to AD, which might provide insight into the progression of disease.35

As a preclinical marker, however, no association was found between GC-IPL thickness and increased risk of incident dementia (HR 1.13, CI [0.9–1.43]) in the Rotterdam study (in contrast to RNFL thickness as previously discussed)18. Nonetheless, GC-IPL thinning has been associated with the prevalence of MCI in a meta-analysis21, has been shown to be more sensitive than RNFL thickness in detecting MCI in patient with positive in vivo biomarkers for AD36, and has recently been associated with cognitively normal patients with AD-related neurodegeneration30. One possible explanation for the discrepancy between the positive structural and negative functional findings is the temporal lag between onset of dementia and the manifestation of GC-IPL thinning. The role for GC-IPL thickness as a biomarker in prodromal disease needs to be better defined.

Total macular thickness measurements have similarly been reported to be decreased in patients with MCI, AD20, and preclinical disease with positive in vivo biomarkers30. However, some studies have demonstrated only weak correlations between macular thickness in AD versus control populations, and predominantly in specific regions of the macula. In such studies, supervised machine learning methods were able to discriminate between AD and controls using OCT data, suggesting that OCT conveys a wealth of data that can be interpreted for classification of dementia even in the absence of apparent univariate changes37.

3. Choroidal thickness and vasculature

The pathophysiologic basis of studying choroidal and retinal vasculature is the observation that decreased cerebral blood flow precedes clinical dementia38. Histopathology of post mortem eyes with advanced AD demonstrated decreased thickness in the nasal choroid, but significant increase in thickness within the macula compared to age matched controls. The increased macular choroidal thickness correlated with increased stromal vascularity, which might represent a compensatory vasoproliferative response to retinal atrophy in severe AD30.

Choroidal thickness was first studied in vivo using enhanced depth imaging (EDI) SD-OCT as a potential biomarker for AD in 2014 (Figure 2d,e). Contrary to the histopathology findings in severe AD eyes, decreased subfoveal choroidal thickness was demonstrated in patients with mild-to-moderate AD compared to control in 21 eyes, in a population where no difference in RNFL thickness was seen between the groups39. This finding has been redemonstrated in MCI patients with positive amyloid PET in all quadrants around the fovea40, and this loss of thickness has been correlated with poor cognitive testing41–42. However, the relationship between choroidal thickness and AD has been challenged by other studies which show no difference in subfoveal choroidal thickness and presence of AD or MCI when adjusted for age, sex, and visual acuity43.

Optical coherence tomography angiography (OCTA) is an emerging imaging modality to directly evaluate choroidal vascularity. For example, OCTA findings recapitulated the aforementioned post mortem findings by performing in vivo analysis of luminal area of choroidal vessels and quantifying the choroidal vascularity index (CVI, ratio of luminal area to total choroidal area), which demonstrated increased luminal area in AD and MCI compared to controls. Interestingly, while decreased CVI was seen in MCI compared to control, no difference in CVI was found between AD and controls.31 Another OCTA study evaluating choroidal flow rate found a trend towards lower choroidal flow rate in patients with AD compared to control, but the difference did not reach statistical significance44.

Interpretation of the conflicting data surrounding choroidal biomarkers is confounded by limitations of individual studies, which are generally limited in sample size, the cross-sectional natures of their design, and analyses seeking to compare large numbers of variables without accounting for the problem of multiple comparisons without employing appropriate statistical methods (such as the Bonferroni correction). The inconsistent data might also reflect the complex relationship between vascular biology and AD. The interplay between choroidal vascular loss, choroidal thinning, and retinal thinning might provide insights into the mechanism and evolution of AD. For example, if retinal vascular changes reflect a diffuse loss of systemic circulation, then choroidal vascular attenuation would be expected to precede retinal vascular attenuation. Microvascular disease and subclinical ischemia might then transiently lead to compensatory increase in choroidal vascularity in preclinical disease or MCI, followed by later attenuation of neurosensory retina. On the other hand, if all the microvascular changes in AD are sequelae of neurodegenerative pathology, we might expect to see structural changes in the inner retina preceding changes in the retinal and choroidal vasculature. Temporal changes in choroidal biomarkers over the spectrum of AD warrant further study.

4. Retinal vasculature

Alzheimer’s disease is associated with deposition of amyloid and collagen within cerebral capillaries. Vascular density and architecture can be quantitatively analyzed from fundus photos using automated software with equivocal results45. Decreased arteriolar fractal dimension and venular fractal dimension have been shown to be associated with AD46 in some studies, while others have found no difference between AD and control47. A systematic review from 2013 noted decreased vascular density and presence of microvascular changes to be weakly associated with the presence of dementia, but its utility as a screening tool was difficult to conclude48.

As with choroidal vascularity, OCTA has been a natural imaging modality for studying retinal vasculature, and has more consistently shown changes associated with AD. In particular, decreased vessel density and increased foveal avascular zone (FAZ) has been demonstrated in AD49. A twin discordance study described monozygotic twins discordant for presentation of AD, and found significantly reduced vessel density and larger FAZ in the superficial capillary plexus in the twin with AD50. Decreased retinal vascular density (most notably in the fovea) and significantly enlarged FAZ correlated also with decreased cognitive function on MMSE44. Vascularity of different vascular layers in the retina has also been correlated with structural OCT biomarkers. Loss of vessel density in the deep venous plexus and superficial capillary plexus were associated with GC-IPL thinning in AD patients, but was not observed in control or MCI groups in separate studies51–52.

Preclinical changes in retinal vasculature have been studied in cognitively normal patients with positive in vivo biomarkers. FAZ area increased in patients with positive amyloid PET/CT or cerebrospinal fluid compared to age matched controls without AD biomarkers53. Another study found no difference in FAZ size in patients with positive amyloid PET/CT, but noted a higher vessel density in preclinical biomarker positive patients compared to controls – a finding that resonates with the choroidal vascularity index trends in MCI noted above54.

In addition to structural changes, functional differences in retinal vasculature of AD patients have also been studied. The dynamic response of vessel dilation to a flicker light stimulation, for example, characteristically results in a primary vasodilatory response followed by vasoconstriction. A study of dynamic vessel analysis found that arterial dilation was decreased in AD compared to age-matched controls, and the reaction amplitude was decreased in both AD and MCI patients (controlling for confounding retinal diseases that are known to be associated with reduced flicker-induced vasodilation), which suggests a decoupling of neurovascular response in patients with cognitive impairment55. Oxygen saturation in retinal arterioles and venules has also been studied using noninvasive spectrophotometric oximetry. Two studies have shown increased oxygen saturation in AD and MCI patients compared to age matched controls, suggesting possible decrease in metabolic activity in the retina56–57.

5. Amyloid and tau imaging

While amyloid plaques and neurofibrillary tangles are well-established findings in the central nervous system, data for corresponding retinal protein deposition in Alzheimer’s disease have been less consistent. Several post mortem studies that looked at immunostaining of amyloid-β and phosphorylated-tau ocular tissue from autopsy-positive AD patients found no abnormal amyloid or tau deposits in any ocular structures to correlate with CNS histopathology58,59,60. Other studies have identified amyloid plaques in retinal tissue of post mortem eyes of both AD and suspected AD patients61. Yet others have reported equivocal findings in retinas of patients with autopsy-confirmed clinicopathologic diagnoses of AD62. A meta-analysis evaluating the evidence of retinal amyloid plaques for diagnosis in AD found existing studies to be too varied in study design and tissue preparation and staining techniques to draw definite conclusions63.

In spite of the as-yet unclear value of retinal amyloid or tau deposition in the diagnosis of AD, several novel methods to noninvasively detect amyloid on retinal imaging have been described. Koronyo et al described using curcumin as a contrast agent that binds amyloid-β with high affinity, in combination with modified SLO, to directly image retinal amyloid-β protein. The authors developed a quantitative measure of increased curcumin fluorescence – the “retinal amyloid index” – as a potential biomarker to delineate AD compared to control. In their proof of principle study with 10 mild-moderate AD patients and 6 healthy controls, and found a 2 fold curcumin fluorescence increase in the AD population64. Of interest, curcurmin-labeled plaques were identified in the retina earlier than it was detected in the brains of a murine model of AD, which suggests its potential utility as a retinal biomarker in early stages of disease61. Hyperspectral imaging is another potential modality that has been studied for identification of early pathological changes in the retina based on the differential scattering of light from abnormal amyloid aggregates65.

6. Artificial intelligence

The advent of computer vision and deep learning algorithms has proven to be a powerful tool in medical imaging analyses. Artificial intelligence research in ophthalmology to date has largely focused on common ophthalmic diseases such as glaucoma, age-related macular degeneration, and diabetic retinopathy, and has demonstrated comparable performance between deep learning algorithms and human experts in classifying disease66.

In the context of Alzheimer’s disease, machine learning has been used to classify fundus photos of patients with AD compared to control from the UK Biobank open-access database with 82% accuracy67. Multimodal imaging has also been used to train AI models, and found improved area under the curve compared to a single imaging modality input68. However, both studies were limited by low samples sizes of AD patients, exclusion of poor image quality, and exclusion of common comorbid retinal diseases – all of which confounds the applicability of the findings to real life screening applications.

Nonetheless, AI holds great promise for furthering the study of biomarkers in AD. Potential applications of AI include providing new insights based on known features, or discerning novel features and biomarkers that are not currently known. As an example of the former, deep learning algorithms applied to histologic sections from brain biopsies of AD patients were able to successfully quantify levels of phosphorylated tau and neurofibrillary tangles, and found that neurofibrillary tangle burden was strongly associated with cognitive impairment in AD patients69. As an example of the latter, parallel work on identifying early biomarkers for age related macular degeneration described a framework agnostic to prespecified parameters that successfully discovered new structural biomarkers on OCT that correlated with function70.

Challenges and future directions

Research on the topic of retinal biomarkers for diagnosis of Alzheimer’s disease has shed light on interesting associations between retinal imaging findings and AD, but validation of these findings in the clinical setting has yet to be demonstrated. Furthermore, conflicting data for many of the potential biomarkers discussed in this review undermines the strength of evidence supporting their relevance and potential utility.

Multiple challenges impede interpretation of present data, and highlight areas that need to be addressed in future studies of retinal biomarkers. For example, diagnostic criteria for the spectrum of Alzheimer’s disease has evolved over time, and there might be a discordance between population of patients with AD and MCI diagnosed clinically as opposed to those diagnosed according to more rigid classifications of disease on the basis of in vivo biomarkers such as CSF or PET/CT amyloid. On the contrary, data from autopsy studies of patients with significant AD neuropathology has identified sizeable populations of people without manifestations of clinical disease, termed “resilient” populations71. Both of these discrepancies might account for inadvertent misclassifications of patients and obscures the ability to generalize conclusions. A list of studies to date that have evaluated prodromal disease (as defined by positive in vivo biomarker in patients with normal cognition and mild cognitive impairment) is summarized in Table 1, and will be an important area of development in future in order to more rigorously define the relationship between retinal biomarker and the specific stage of AD.

Table 1:

Summary of Studies Evaluating Retinal Biomarkers in Patients with Prodromal Disease as Defined by Normal Cognition or Mild Cognitive Impairment and Positive In Vivo Biomarkers for Alzheimer’s Disease.

Retinal biomarker Imaging modality Cognitive impairment Biomarker Authors Year Summary of findings (compared to control)
Retinal structural parameters OCT MCI ApoE carrier Shin et al20 2021 No difference in thickness of any retinal layers in biomarker+ patients
Aβ PET Lopez-de-Eguileta et al33 2019 RNFL and GC-IPL thickness decreased in biomarker+ patients
Normal Aβ PET Byun et al27 2021 RNFL and macular thickness decreased in biomarker+ patients
Aβ PET Van de Kreeke et al70 2021 No differences in retinal thickness change over 22 months in biomarker+ patients
Aβ PET Snyder et al71 2016 Increased GC-IPL thickness and inclusion body surface area in biomarker+ patients
OCTA Normal CSF Aβ/Tau Asanad et al72 2020 RNFL thickness decreased in biomarker+ patients
Aβ PET or CSF Aβ/Tau O’Bryhim et al73 2018 Inner foveal thickness decreased in biomarker+ patients
Aβ PET Santos et al74 2018 Larger reduction in RNFL and GC-IPL volume in biomarker+ patients over 27 month follow up
AO-SLO MCI CSF Aβ/Tau or ApoE carrier Zhang et al23 2019 Increased number and area of granular membranes in biomarker+ patients
Retinal vascular parameters OCTA MCI ApoE carrier Shin et al20 2021 Lower vessel density of deep capillary plexus in biomarker+ patients
CSF Aβ/Tau Querques et al51 2019 Positive biomarkers were correlated with arterial dilation and reaction amplitude on dynamic vessel analysis
Normal ApoE carrier or Aβ PET Elahi et al75 2021 Lower capillary density in biomarker+ patients
Aβ PET Van de Kreeke et al76 2019 Higher vessel density in biomarker+ patients
Aβ PET or CSF Aβ/Tau O’Bryhim et al73 2018 Enlarged FAZ in biomarker+ patients
Fundus photography Normal Aβ PET Frost et al77 2013 Differences in retinal vascular parameters in biomarker+ patients
Choroidal parameters OCT MCI Aβ PET Lopez-de-Eguileta et al37 2020 Focal choroidal thinning in biomarker+ patients
Open in a new tab

Aβ PET indicates amyloid beta positron emission tomography; AOSLO, adaptive optics scanning laser ophthalmoscope; APOE ε4, Apolipoprotein E ε4; CSF, cerebrospinal fluid; FAZ, foveal avascular zone; GC-IPL, ganglion cell–inner plexiform layer; MCI, mild cognitive impairment; OCT, optical coherence tomography; OCTA, optical coherence tomography angiography; RNFL, retinal neural fiber layer.

Comorbid retinal disease is another obstacle that complicates biomarker research in AD. While most research studies take care to exclude patients with age related macular degeneration (ARMD), for example, the demographics of patients with AD or preclinical AD invariably overlap with those who are at highest risk of having ARMD. When considering biomarkers that might simultaneously have manifestations in both disease processes, such as retinal amyloid deposition in drusen of ARMD patients72, the optimal biomarker must have the ability to discern AD in spite of confounding retinal disease in order to have practical utility in clinical settings. Retinal biomarker profiles might also be confounded by similarities between AD and other causes of cognitive impairment such as vascular dementia and frontotemporal dementia, and similarly will need to be able to differentiate between these etiologies73.

Other challenges include the small numbers of cases described in most studies. Given the heterogeneity in presentation of AD, associations found on the basis of small patient populations might introduce unexpected biases and limit generalizability to communities at large. Sharing of large datasets will be important to overcome these limitations, as well as allow us to harness the potential power of AI in generating new insights. Challenges of obtaining retinal images of sufficient quality for analysis in community cohorts will also need to be addressed prior to practical implementation at a large scale. In the population of interest for AD, confounding obstacles related to patient cooperation and media opacities, for example, might significantly hinder the ability to obtain consistent and useful data. For example, in one recent population based OCT-A study, images of sufficient quality for analysis was obtained in only 55% of patients74.

Finally, a standardized framework to approach future studies, such as that proposed recently by Cheung et al75, would help promote consistency across study design. Such a framework would be relevant not just for evaluation of retinal biomarkers, but also biomarkers involving other ocular tissues such as the lens76, aqueous humor77, and cornea78.

Conclusions

Finding a disease modifying treatment for Alzheimer’s disease with definitive benefit to clinical endpoints has been an elusive goal of research in the field of AD. Discovery and validation of in vivo disease-specific biomarkers has prompted a shift in research interest to early and prodromal stages of disease as an opportune window in which to intervene, prior to the onset of irreversible neurovascular changes. Current clinical standards of diagnosis are limited by cost, accessibility, and inaccuracy. In this context, the possibility of identifying noninvasive, accurate, inexpensive biomarkers to help screen for and follow the progression of disease represents an important goal. Advances in specialized noninvasive retinal imaging and computer processing power positions the search for retinal biomarkers well to contribute to this endeavor.

References

  • 1.World Alzheimer Report 2021: Journey through the diagnosis of dementia. :314. [Google Scholar]
  • 2.McKhann GM, Knopman DS, Chertkow H, et al. The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement J Alzheimers Assoc. 2011;7(3):263–269. doi: 10.1016/j.jalz.2011.03.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.National Center for Health Statistics, Heron M. Deaths: Leading Causes for 2018. National Center for Health Statistics; 2021. doi: 10.15620/cdc:104186 [DOI] [PubMed] [Google Scholar]
  • 4.Amjad H, Roth DL, Sheehan OC, Lyketsos CG, Wolff JL, Samus QM. Underdiagnosis of Dementia: an Observational Study of Patterns in Diagnosis and Awareness in US Older Adults. J Gen Intern Med. 2018;33(7):1131–1138. doi: 10.1007/s11606-018-4377-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Jack CR Jr. Bennett DA, Blennow K, et al. NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement. 2018;14(4):535–562. doi: 10.1016/j.jalz.2018.02.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Beach TG, Monsell SE, Phillips LE, Kukull W. Accuracy of the clinical diagnosis of Alzheimer disease at National Institute on Aging Alzheimer Disease Centers, 2005–2010. J Neuropathol Exp Neurol. 2012;71(4):266–273. doi: 10.1097/NEN.0b013e31824b211b [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Dunn B, Stein P, Cavazzoni P. Approval of Aducanumab for Alzheimer Disease—The FDA’s Perspective. JAMA Intern Med. 2021;181(10):1276–1278. doi: 10.1001/jamainternmed.2021.4607 [DOI] [PubMed] [Google Scholar]
  • 8.Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease - Sperling - 2011 - Alzheimer’s & Dementia - Wiley Online Library. Accessed November 20, 2021. 10.1016/j.jalz.2011.03.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Fischer P, Jungwirth S, Zehetmayer S, et al. Conversion from subtypes of mild cognitive impairment to Alzheimer dementia. Neurology. 2007;68(4):288–291. doi: 10.1212/01.wnl.0000252358.03285.9d [DOI] [PubMed] [Google Scholar]
  • 10.Sperling R, Mormino E, Johnson K. The Evolution of Preclinical Alzheimer’s Disease: Implications for Prevention Trials. Neuron. 2014;84(3):608–622. doi: 10.1016/j.neuron.2014.10.038 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Kalaria RN, Pax AB. Increased collagen content of cerebral microvessels in Alzheimer’s disease. Brain Res. 1995;705(1):349–352. doi: 10.1016/0006-8993(95)01250-8 [DOI] [PubMed] [Google Scholar]
  • 12.Serrano-Pozo A, Muzikansky A, Gómez-Isla T, et al. Differential Relationships of Reactive Astrocytes and Microglia to Fibrillar Amyloid Deposits in Alzheimer Disease. J Neuropathol Exp Neurol. 2013;72(6):462–471. doi: 10.1097/NEN.0b013e3182933788 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hinton DR, Sadun AA, Blanks JC, Miller CA. Optic-Nerve Degeneration in Alzheimer’s Disease. 10.1056/NEJM198608213150804. doi: 10.1056/NEJM198608213150804 [DOI] [PubMed] [Google Scholar]
  • 14.Sadun AA, Bassi CJ. Optic nerve damage in Alzheimer’s disease. Ophthalmology. 1990;97(1):9–17. doi: 10.1016/s0161-6420(90)32621-0 [DOI] [PubMed] [Google Scholar]
  • 15.Blanks JC, Hinton DR, Sadun AA, Miller CA. Retinal ganglion cell degeneration in Alzheimer’s disease. Brain Res. 1989;501(2):364–372. doi: 10.1016/0006-8993(89)90653-7 [DOI] [PubMed] [Google Scholar]
  • 16.Tsai CS, Ritch R, Schwartz B, et al. Optic nerve head and nerve fiber layer in Alzheimer’s disease. Arch Ophthalmol Chic Ill 1960. 1991;109(2):199–204. doi: 10.1001/archopht.1991.01080020045040 [DOI] [PubMed] [Google Scholar]
  • 17.Ashok A, Singh N, Chaudhary S, et al. Retinal Degeneration and Alzheimer’s Disease: An Evolving Link. Int J Mol Sci. 2020;21(19):7290. doi: 10.3390/ijms21197290 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Danesh-Meyer HV, Birch H, Ku JYF, Carroll S, Gamble G. Reduction of optic nerve fibers in patients with Alzheimer disease identified by laser imaging. Neurology. 2006;67(10):1852–1854. doi: 10.1212/01.wnl.0000244490.07925.8b [DOI] [PubMed] [Google Scholar]
  • 19.Iseri PK, Altinaş O, Tokay T, Yüksel N. Relationship between cognitive impairment and retinal morphological and visual functional abnormalities in Alzheimer disease. J Neuro-Ophthalmol Off J North Am Neuro-Ophthalmol Soc. 2006;26(1):18–24. doi: 10.1097/01.wno.0000204645.56873.26 [DOI] [PubMed] [Google Scholar]
  • 20.Chan VTT, Sun Z, Tang S, et al. Spectral-Domain OCT Measurements in Alzheimer’s Disease: A Systematic Review and Meta-analysis. Ophthalmology. 2019;126(4):497–510. doi: 10.1016/j.ophtha.2018.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Spectral-Domain OCT Measurements in Alzheimer’s Disease | Elsevier Enhanced Reader. doi: 10.1016/j.ophtha.2018.08.009 [DOI] [Google Scholar]
  • 22.Lad EM, Mukherjee D, Stinnett SS, et al. Evaluation of inner retinal layers as biomarkers in mild cognitive impairment to moderate Alzheimer’s disease. PLoS ONE. 2018;13(2):e0192646. doi: 10.1371/journal.pone.0192646 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Shin JY, Choi EY, Kim M, Lee HK, Byeon SH. Changes in retinal microvasculature and retinal layer thickness in association with apolipoprotein E genotype in Alzheimer’s disease. Sci Rep. 2021;11(1):1847. doi: 10.1038/s41598-020-80892-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Knoll B, Simonett J, Volpe NJ, et al. Retinal nerve fiber layer thickness in amnestic mild cognitive impairment: Case-control study and meta-analysis. Alzheimers Dement Diagn Assess Dis Monit. 2016;4:85–93. doi: 10.1016/j.dadm.2016.07.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Bates KA, Fonte J, Robertson TA, Martins RN, Harvey AR. Chronic gliosis triggers Alzheimer’s disease-like processing of amyloid precursor protein. Neuroscience. 2002;113(4):785–796. doi: 10.1016/s0306-4522(02)00230–0 [DOI] [PubMed] [Google Scholar]
  • 26.Zhang YS, Onishi AC, Zhou N, et al. Characterization of Inner Retinal Hyperreflective Alterations in Early Cognitive Impairment on Adaptive Optics Scanning Laser Ophthalmoscopy. Invest Ophthalmol Vis Sci. 2019;60(10):3527–3536. doi: 10.1167/iovs.19-27135 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Ko F, Muthy ZA, Gallacher J, et al. Association of Retinal Nerve Fiber Layer Thinning With Current and Future Cognitive Decline: A Study Using Optical Coherence Tomography. JAMA Neurol. 2018;75(10):1198–1205. doi: 10.1001/jamaneurol.2018.1578 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Mutlu U, Colijn JM, Ikram MA, et al. Association of Retinal Neurodegeneration on Optical Coherence Tomography With Dementia: A Population-Based Study. JAMA Neurol. 2018;75(10):1256–1263. doi: 10.1001/jamaneurol.2018.1563 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Shi Z, Wu Y, Wang M, et al. Greater Attenuation of Retinal Nerve Fiber Layer Thickness in Alzheimer’s Disease Patients. J Alzheimers Dis. 2014;40(2):277–283. doi: 10.3233/JAD-131898 [DOI] [PubMed] [Google Scholar]
  • 30.Byun MS, Park SW, Lee JH, et al. Association of Retinal Changes With Alzheimer Disease Neuroimaging Biomarkers in Cognitively Normal Individuals. JAMA Ophthalmol. 2021;139(5):548–556. doi: 10.1001/jamaophthalmol.2021.0320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Ferrari L, Huang SC, Magnani G, Ambrosi A, Comi G, Leocani L. Optical Coherence Tomography Reveals Retinal Neuroaxonal Thinning in Frontotemporal Dementia as in Alzheimer’s Disease. J Alzheimers Dis JAD. 2017;56(3):1101–1107. doi: 10.3233/JAD-160886 [DOI] [PubMed] [Google Scholar]
  • 32.Moreno-Ramos T, Benito-León J, Villarejo A, Bermejo-Pareja F. Retinal nerve fiber layer thinning in dementia associated with Parkinson’s disease, dementia with Lewy bodies, and Alzheimer’s disease. J Alzheimers Dis JAD. 2013;34(3):659–664. doi: 10.3233/JAD-121975 [DOI] [PubMed] [Google Scholar]
  • 33.Garcia-Martin E, Bambo MP, Marques ML, et al. Ganglion cell layer measurements correlate with disease severity in patients with Alzheimer’s disease. Acta Ophthalmol (Copenh). 2016;94(6):e454–e459. doi: 10.1111/aos.12977 [DOI] [PubMed] [Google Scholar]
  • 34.CY, Ong YT, Hilal S, et al. Retinal Ganglion Cell Analysis Using High-Definition Optical Coherence Tomography in Patients with Mild Cognitive Impairment and Alzheimer’s Disease. J Alzheimers Dis. 2015;45(1):45–56. doi: 10.3233/JAD-141659 [DOI] [PubMed] [Google Scholar]
  • 35.Shao Y, Jiang H, Wei Y, et al. Visualization of Focal Thinning of the Ganglion Cell-Inner Plexiform Layer in Patients with Mild Cognitive Impairment and Alzheimer’s Disease. J Alzheimers Dis JAD. 2018;64(4):1261–1273. doi: 10.3233/JAD-180070 [DOI] [PubMed] [Google Scholar]
  • 36.López-de-Eguileta A, Lage C, López-García S, et al. Ganglion cell layer thinning in prodromal Alzheimer’s disease defined by amyloid PET. Alzheimers Dement Transl Res Clin Interv. 2019;5:570–578. doi: 10.1016/j.trci.2019.08.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Nunes A, Silva G, Duque C, et al. Retinal texture biomarkers may help to discriminate between Alzheimer’s, Parkinson’s, and healthy controls. PLoS ONE. 2019;14(6):e0218826. doi: 10.1371/journal.pone.0218826 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Zlokovic BV. Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat Rev Neurosci. 2011;12(12):723–738. doi: 10.1038/nrn3114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Gharbiya M, Trebbastoni A, Parisi F, et al. Choroidal thinning as a new finding in Alzheimer’s disease: evidence from enhanced depth imaging spectral domain optical coherence tomography. J Alzheimers Dis JAD. 2014;40(4):907–917. doi: 10.3233/JAD-132039 [DOI] [PubMed] [Google Scholar]
  • 40.López-de-Eguileta A, Lage C, López-García S, et al. Evaluation of choroidal thickness in prodromal Alzheimer’s disease defined by amyloid PET. PLOS ONE. 2020;15(9):e0239484. doi: 10.1371/journal.pone.0239484 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Jonas JB, Wang YX, Wei WB, Zhu LP, Shao L, Xu L. Cognitive Function and Subfoveal Choroidal Thickness: The Beijing Eye Study. Ophthalmology. 2016;123(1):220–222. doi: 10.1016/j.ophtha.2015.06.020 [DOI] [PubMed] [Google Scholar]
  • 42.Trebbastoni A, Marcelli M, Mallone F, et al. Attenuation of Choroidal Thickness in Patients With Alzheimer Disease: Evidence From an Italian Prospective Study. Alzheimer Dis Assoc Disord. 2017;31(2):128–134. doi: 10.1097/WAD.0000000000000176 [DOI] [PubMed] [Google Scholar]
  • 43.Robbins CB, Grewal DS, Thompson AC, et al. Choroidal Structural Analysis in Alzheimer Disease, Mild Cognitive Impairment, and Cognitively Healthy Controls. Am J Ophthalmol. 2021;223:359–367. doi: 10.1016/j.ajo.2020.09.049 [DOI] [PubMed] [Google Scholar]
  • 44.Bulut M, Kurtuluş F, Gözkaya O, et al. Evaluation of optical coherence tomography angiographic findings in Alzheimer’s type dementia. Br J Ophthalmol. 2018;102(2):233–237. doi: 10.1136/bjophthalmol-2017-310476 [DOI] [PubMed] [Google Scholar]
  • 45.Arnould L, Guillemin M, Seydou A, et al. Association between the retinal vascular network and retinal nerve fiber layer in the elderly: The Montrachet study. PLOS ONE. 2020;15(10):e0241055. doi: 10.1371/journal.pone.0241055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.CY, Ong YT, Ikram MK, et al. Microvascular network alterations in the retina of patients with Alzheimer’s disease. Alzheimers Dement. 2014;10(2):135–142. doi: 10.1016/j.jalz.2013.06.009 [DOI] [PubMed] [Google Scholar]
  • 47.den Haan J, van de Kreeke JA, van Berckel BN, et al. Is retinal vasculature a biomarker in amyloid proven Alzheimer’s disease? Alzheimers Dement Diagn Assess Dis Monit. 2019;11:383–391. doi: 10.1016/j.dadm.2019.03.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Heringa SM, Bouvy WH, van den Berg E, Moll AC, Kappelle LJ, Biessels GJ. Associations Between Retinal Microvascular Changes and Dementia, Cognitive Functioning, and Brain Imaging Abnormalities: A Systematic Review. J Cereb Blood Flow Metab. 2013;33(7):983–995. doi: 10.1038/jcbfm.2013.58 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Jin Q, Lei Y, Wang R, Wu H, Ji K, Ling L. A Systematic Review and Meta-Analysis of Retinal Microvascular Features in Alzheimer’s Disease. Front Aging Neurosci. 2021;13. Accessed January 17, 2022. 10.3389/fnagi.2021.683824 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Grewal DS, Polascik BW, Hoffmeyer GC, Fekrat S. Assessment of Differences in Retinal Microvasculature Using OCT Angiography in Alzheimer’s Disease: A Twin Discordance Report. Ophthalmic Surg Lasers Imaging Retina. 2018;49(6):440–444. doi: 10.3928/23258160-20180601-09 [DOI] [PubMed] [Google Scholar]
  • 51.Jiang H, Wei Y, Shi Y, et al. Altered Macular Microvasculature in Mild Cognitive Impairment and Alzheimer Disease. J Neuroophthalmol. 2018;38(3):292–298. doi: 10.1097/WNO.0000000000000580 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Yoon SP, Grewal DS, Thompson AC, et al. Retinal Microvascular and Neurodegenerative Changes in Alzheimer’s Disease and Mild Cognitive Impairment Compared with Control Participants. Ophthalmol Retina. 2019;3(6):489–499. doi: 10.1016/j.oret.2019.02.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.O’Bryhim BE, Apte RS, Kung N, Coble D, Van Stavern GP. Association of Preclinical Alzheimer Disease With Optical Coherence Tomographic Angiography Findings. JAMA Ophthalmol. 2018;136(11):1242–1248. doi: 10.1001/jamaophthalmol.2018.3556 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Kreeke van de JA, Nguyen HT, Konijnenberg E, et al. Optical coherence tomography angiography in preclinical Alzheimer’s disease. Br J Ophthalmol. 2020;104(2):157–161. doi: 10.1136/bjophthalmol-2019-314127 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Querques G, Borrelli E, Sacconi R, et al. Functional and morphological changes of the retinal vessels in Alzheimer’s disease and mild cognitive impairment. Sci Rep. 2019;9(1):63. doi: 10.1038/s41598-018-37271-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Olafsdottir OB, Saevarsdottir HS, Hardarson SH, et al. Retinal oxygen metabolism in patients with mild cognitive impairment. Alzheimers Dement Diagn Assess Dis Monit. 2018;10:340–345. doi: 10.1016/j.dadm.2018.03.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Retinal Oximetry Imaging in Alzheimer’s Disease - IOS Press. Accessed October 10, 2021. https://content.iospress.com/articles/journal-of-alzheimers-disease/jad150457
  • 58.Schön C, Hoffmann NA, Ochs SM, et al. Long-Term In Vivo Imaging of Fibrillar Tau in the Retina of P301S Transgenic Mice. PLOS ONE. 2012;7(12):e53547. doi: 10.1371/journal.pone.0053547 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Ho CY, Troncoso JC, Knox D, Stark W, Eberhart CG. Beta-amyloid, phospho-tau and alpha-synuclein deposits similar to those in the brain are not identified in the eyes of Alzheimer’s and Parkinson’s disease patients. Brain Pathol Zurich Switz. 2014;24(1):25–32. doi: 10.1111/bpa.12070 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Williams EA, McGuone D, Frosch MP, Hyman BT, Laver N, Stemmer-Rachamimov A. Absence of Alzheimer Disease Neuropathologic Changes in Eyes of Subjects With Alzheimer Disease. J Neuropathol Exp Neurol. 2017;76(5):376–383. doi: 10.1093/jnen/nlx020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Koronyo-Hamaoui M, Koronyo Y, Ljubimov AV, et al. Identification of Amyloid Plaques in Retinas from Alzheimer’s Patients and Noninvasive In Vivo Optical Imaging of Retinal Plaques in a Mouse Model. NeuroImage. 2011;54S1:S204–S217. doi: 10.1016/j.neuroimage.2010.06.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 62.Tsai Y, Lu B, Ljubimov AV, et al. Ocular Changes in TgF344-AD Rat Model of Alzheimer’s Disease. Invest Ophthalmol Vis Sci. 2014;55(1):523–534. doi: 10.1167/iovs.13-12888 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Jiang J, Wang H, Li W, Cao X, Li C. Amyloid Plaques in Retina for Diagnosis in Alzheimer’s Patients: a Meta-Analysis. Front Aging Neurosci. 2016;8:267. doi: 10.3389/fnagi.2016.00267 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Koronyo Y, Biggs D, Barron E, et al. Retinal amyloid pathology and proof-of-concept imaging trial in Alzheimer’s disease. JCI Insight. 2(16):e93621. doi: 10.1172/jci.insight.93621 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.More SS, Beach JM, McClelland C, Mokhtarzadeh A, Vince R. In Vivo Assessment of Retinal Biomarkers by Hyperspectral Imaging: Early Detection of Alzheimer’s Disease. ACS Chem Neurosci. 2019;10(11):4492–4501. doi: 10.1021/acschemneuro.9b00331 [DOI] [PubMed] [Google Scholar]
  • 66.Ting DSW, Pasquale LR, Peng L, et al. Artificial intelligence and deep learning in ophthalmology. Br J Ophthalmol. 2019;103(2):167–175. doi: 10.1136/bjophthalmol-2018-313173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67.Tian J, Smith G, Guo H, et al. Modular machine learning for Alzheimer’s disease classification from retinal vasculature. Sci Rep. 2021;11(1):238. doi: 10.1038/s41598-020-80312-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 68.Convolutional neural network to identify symptomatic Alzheimer’s disease using multimodal retinal imaging | British Journal of Ophthalmology. Accessed November 24, 2021. https://bjo.bmj.com/content/early/2020/11/25/bjophthalmol-2020-317659 [DOI] [PubMed] [Google Scholar]
  • 69.Lee CS, Latimer CS, Henriksen JC, et al. Application of deep learning to understand resilience to Alzheimer’s disease pathology. Brain Pathol. 2021;31(6):e12974. doi: 10.1111/bpa.12974 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Lee AY, Lee CS, Blazes MS, et al. Exploring a Structural Basis for Delayed Rod-Mediated Dark Adaptation in Age-Related Macular Degeneration Via Deep Learning. Transl Vis Sci Technol. 2020;9(2):62. doi: 10.1167/tvst.9.2.62 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71.Aiello Bowles EJ, Crane PK, Walker RL, et al. Cognitive Resilience to Alzheimer’s Disease Pathology in the Human Brain. J Alzheimers Dis. 2019;68(3):1071–1083. doi: 10.3233/JAD-180942 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 72.Ohno-Matsui K. Parallel findings in age-related macular degeneration and Alzheimer’s disease. Prog Retin Eye Res. 2011;30(4):217–238. doi: 10.1016/j.preteyeres.2011.02.004 [DOI] [PubMed] [Google Scholar]
  • 73.Chalkias E, Topouzis F, Tegos T, Tsolaki M. The Contribution of Ocular Biomarkers in the Differential Diagnosis of Alzheimer’s Disease versus Other Types of Dementia and Future Prospects. J Alzheimers Dis. 2021;80(2):493–504. doi: 10.3233/JAD-201516 [DOI] [PubMed] [Google Scholar]
  • 74.Abraham AG, Guo X, Arsiwala LT, et al. Cognitive decline in older adults: What can we learn from optical coherence tomography (OCT)-based retinal vascular imaging? J Am Geriatr Soc. 2021;69(9):2524–2535. doi: 10.1111/jgs.17272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Cheung CY, Mok V, Foster PJ, Trucco E, Chen C, Wong TY. Retinal imaging in Alzheimer’s disease. J Neurol Neurosurg Psychiatry. 2021;92(9):983–994. doi: 10.1136/jnnp-2020-325347 [DOI] [PubMed] [Google Scholar]
  • 76.Melov S, Wolf N, Strozyk D, Doctrow SR, Bush AI. Mice transgenic for Alzheimer disease β-amyloid develop lens cataracts that are rescued by antioxidant treatment. Free Radic Biol Med. 2005;38(2):258–261. doi: 10.1016/j.freeradbiomed.2004.10.023 [DOI] [PubMed] [Google Scholar]
  • 77.Kwak DE, Ko T, Koh HS, et al. Alterations of aqueous humor Aβ levels in Aβ-infused and transgenic mouse models of Alzheimer disease. PLOS ONE. 2020;15(1):e0227618. doi: 10.1371/journal.pone.0227618 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 78.Dong Z, Luo A, Gan Y, Li J. Amyloid Beta Deposition Could Cause Corneal Epithelial Cell Degeneration Associated with Increasing Apoptosis in APPswePS1 Transgenic Mice. Curr Eye Res. 2018;43(11):1326–1333. doi: 10.1080/02713683.2018.1501070 [DOI] [PubMed] [Google Scholar]

RESOURCES