Optical Coherence Tomography Angiography (OCTA) Retinal Imaging Associations with Burden of Small Vessel Disease and Amyloid Positivity in the Brain

Abstract

BACKGROUND

Alzheimer’s Disease (AD) and other dementias are associated with vascular changes and amyloid deposition, which may be reflected as density changes in the retinal capillaries. These changes may can be directly visualized and quantified with optical coherence tomography angiography (OCTA), making OCTA a potential noninvasive preclinical biomarker of small vessel disease and amyloid positivity. Our objective was to investigate the feasibility of retinal imaging metrics as noninvasive biomarkers of small vessel disease and amyloid positivity in the brain.

METHODS

We investigated associations between OCTA and neuroimaging and cognitive metrics in 41 participants without dementia from the Mayo Clinic Study of Aging and Alzheimer’s Disease Research Center. OCTA metrics included superficial, deep, and full retina capillary density of the fovea, parafovea, and macula as well as the area of the foveal avascular zone (FAZ). Neuroimaging metrics included a high burden of white matter hyperintensity (WMH), presence of cerebral microbleeds, lacunar infarcts, and amyloid positivity as evidenced on positron emission tomography (PET) while cognitive metrics included mini-mental status exam (MMSE) score. We performed generalized estimating equations to account for measurements in each eye while controlling for age and sex to estimate associations between OCTA metrics and neuroimaging and cognitive scores.

RESULTS

Associations between OCTA and neuroimaging metrics were restricted to the fovea. OCTA showed decreased capillary density with high burden of WMH in both the superficial (p = 0.003), deep (p = 0.004) and full retina (p = 0.01) in the fovea, but not the parafovea or whole macula. Similarly, participants with amyloid PET positivity had significantly decreased capillary density in the superficial fovea (p=0.027) and deep fovea (p = 0.03), but higher density in the superficial parafovea (p = 0.038). Participants with amyloid PET positivity also had a significantly larger FAZ (p = 0.031) whereas in those with high WMH burden the difference did not reach statistical significance (p = 0.075). There was also a positive association between MMSE and capillary density of the full retina within the fovea (p = 0.037) as well as the superficial parafovea (p = 0.046). No associations were found between OCTA metrics and presence of cerebral microbleeds or presence of lacunar infarcts.

CONCLUSION

The associations of lower foveal capillary density with cerebral WMH and amyloid positivity suggest further research is warranted to evaluate for shared mechanisms of disease between small vessel disease and Alzheimer’s disease pathologies.

1. Introduction

Optical coherence tomography angiography (OCTA) is a novel noninvasive tool that allows for visualization and density measurement of the capillaries at various levels of the retina. Retinal imaging provides a unique perspective to study the microvascular environment in the central nervous system due to shared embryologic origins and physiology [1]. Cerebral microvascular disease and amyloid deposition in the brain may possibly correlate with changes in the density of retinal capillaries, making OCTA a potential noninvasive preclinical biomarker of common risk factors for dementia. There have been several studies recently showing OCTA changes in subjects with dementia, which were summarized by Katsimpis et al in a 2021 meta-analysis. This work showed significantly decreased vessel density in the superficial and deep capillary plexus as well as a larger foveal avascular zone in Alzheimer’s disease (AD) compared to cognitively unimpaired controls in 10 retrospective studies[2]. However, there is currently limited evidence about the applicability of using OCTA to detect changes associated with small vessel disease in the brains of non-AD participants. A study by Lahme et al, showed a negative correlation between superficial macula density and white matter hyperintensity (WMH) burden [3]. Similarly, one study found a larger foveal avascular zone (FAZ) in patients who were amyloid positive [4] whereas others found no difference [5]. These studies have been limited by small sample sizes, usually 10–20 participants, and incomplete data acquisition of key imaging biomarkers of small vessel disease as well as OCTA metrics. Also, a larger study by van de Kreeke et al in monozygotic twins showed discrepant findings: capillary density was positively correlated with amyloid status and there was no difference in FAZ area between amyloid subgroups [6].

Early detection of microvascular disease burden and amyloid deposition in the brain can help identify patients at risk of developing dementia for candidacy in clinical trials or emerging anti-amyloid therapies[7]. Current strategies for the detection of β-amyloid (Aβ) in the central nervous system rely on analysis of cerebrospinal fluid or plasma biomarkers as well as positron emission tomography [8]. Both strategies have shown high concordance with underlying neuropathology, but are limited by cost, invasiveness, or availability, which has prompted the field to search for less invasive biomarkers including plasma[9, 10]. Similarly, there is currently no in vivo technique to visualize the microvasculature of the brain. Instead, several proxy imaging biomarkers extracted from structural brain magnetic resonance imaging (MRI) are used to estimate the burden of small vessel disease in the brain. Such imaging biomarkers include WMH [11], the presence of lacunar infarcts[12], or presence of cerebral microbleeds (CMB)[13]. Neuropathologic studies in Alzheimer’s disease have shown changes in the microvascular architecture of the brain, with disruption of the blood-brain barrier and decreased capillary density that may facilitate amyloid deposition [14].

In this work, we aimed to investigate the utility of retinal imaging with OCTA metrics as noninvasive biomarkers of small vessel disease and amyloid positivity in the brain. We hypothesized that OCTA density measurements are associated with the presence of small vessel disease in the brain as indicated by imaging biomarkers as well as the presence of amyloid as indicated by PET imaging. We leverage a subgroup population from the Mayo Clinic Study of Aging (MCSA), a population-based cohort, and the Alzheimer’s Disease Research Center (ADRC), a clinic-based observational cohort, that includes regular imaging and cognitive assessments as well as complete retinal imaging to test these associations.

2. Methods

2.1. Study Participants

All imaging and OCTA metrics were collected from participants in the MCSA or the ADRC. Clinical diagnoses of cognition were determined by the consensus of a committee composed of a physician, a study coordinator, and a neuropsychologist employing predetermined criteria [15, 16]. Mini-mental status exam (MMSE) score was computed as part of the evaluation. The Institutional Review Boards of both Mayo Clinic and Olmsted Medical Center approved this study and participants provided written informed consent. A subset of 41 MCSA and ADRC randomly selected participants underwent retinal OCTA imaging. Roughly half were PET amyloid positive, but no participants had AD. The small vessel disease biomarkers and OCTA metrics were obtained within 3 years of each other.

2.2. OCTA Acquisition

Optovue OCT angiography was used to measure the retinal capillary network. The macula was imaged with a 6.4×6.4mm scan centered on the fovea with 512 × 512 A-scans. Split-spectrum amplitude decorrelation angiography (SSADA) images were calculated after coregistration of A-scans followed by decorrelation calculations on consecutive B-scans. Capillary density in the retina was calculated using Optovue’s automated software, AngioAnalytics (version 2015.1.1.98), to create a flow density map. OCTA metrics evaluated included superficial and deep capillary density of the fovea, parafovea, and macula as well as the FAZ area for each eye. The fovea zone is made of a region that is a 300µm ring around the FAZ.

2.3. Brain MRI Acquisition, Small Vessel Disease Biomarkers, and Amyloid Status

All images were obtained using 3-tesla MRI scanners. The complete details of the acquisitions were previously published by Jack et al [17]. On the brain MRI, three imaging biomarkers of small vessel disease were defined: high white matter hyperintensity (WMH) burden, presence of lacunar infarcts and cerebral microbleeds (CMB). MRI assessment and determination of high WMH burden as well as presence of CMBs and examination of lacunar infarcts in this cohort has been previously published by Graff-Radford et al [18]. Acquisition and processing of amyloid PET and determination of positivity has also been previously published by Graff-Radford et al [13].

2.4. Statistical Analysis

Characteristics of the participants were summarized using means and standard deviations for continuous variables and counts and percentages for categorical variables. We used generalized estimating equation (GEE) modeling as the primary analysis of this study to account for left and right eye measurements of OCTA when looking for associations between OCTA outcomes and various neuroimaging and cognitive variables. The neuroimaging and cognitive variables of interest were: high WMH burden, presence of CMBs, presence of lacunar infarcts, amyloid PET positivity, and MMSE scores. GEE models were run while adjusting for age and sex. The OCTA predictors of interest consisted of superficial, deep, and full retina capillary density measurement in overlapping concentric discs extending from smallest to largest to include the fovea, parafovea, and macula, while excluding measurements from the FAZ. The area of the FAZ was also independently measured (Figure 1). Interaction models examining the combined effect of WMH and amyloid were not pursued due to be lack of power from small sample size.

Figure 1.

Representative optical coherence tomography angiography (OCTA) capillary density measurements of the retina. Measurements are taken within the area of three concentric circles composing the fovea (innermost, red area), parafovea (middle, blue area), and macula (outermost, yellow area) while excluding the fovea avascular zone (FAZ). Each disk has measurements in the superficial and deep retina. OCTA output metrics include capillary density in each of these six partitions of the retina as well as the area of the FAZ.

3. Results

Participant demographics are summarized in Table 1. The average age in our cohort was 75.2 years, ranging from 62 to 93 years old. Twenty-three of the 41 participants (56.1%) were women and 23 of the 41 participants (56.1%) were amyloid PET positive. There was no statistical association between sex and amyloid PET positivity (chi-squared test, p = 0.88). With regards to cognitive testing, the average MMSE score was 28.1 points ranging from 23 to 30 points. Seven out of 41 participants (17.1%) had a diagnosis of mild cognitive impairment (MCI), six of which were amyloid PET positive. With regards to cognition and OCTA metrics, there was a positive association between MMSE and capillary density of the full fovea (β = 1.05 [0.08, 2.03], p = 0.037) as well as the superficial parafovea (β = 0.64 [0.02, 1.26], p = 0.046), but not the FAZ. The MCI subgroup was significantly older than the CU subgroup (80.9 vs 74.0, p = 0.03) (Supplementary Table 1). There was a significant association with small capillary density in the superficial fovea of the MCI subgroup in the age- and sex-controlled models (Supplementary Table2). Otherwise, there was no difference between groups in the neuroimaging metrics.

Table 1.

Participant demographics listed as mean (SD) for continuous variables and percent (count) for categorical variables. OCTA: Optical coherence tomography angiography.

Demographics	Participants (N = 41)
Age (years)	75.1 +/− 8.01
Sex (% Female)	56.1% (23/41)
Amyloid (% positive)	56.1% (23/41)
MCI (% positive)	17.1% (7/41)
MMSE (score)	28.1 +/− 1.49
Small Vessel Disease Biomarkers on Brain MRI
Cerebral Microbleeds (% positive)	39.0% (16/41)
Lacunar Infarct (% positive)	7.32% (3/41)
White Matter Hyperintensity (% with high burden)	14.6% (6/41)
OCTA Measurements
Foveal Avascular Zone (area mm2)	27.9 +/− 14.1
Foveal Capillary Density
Superficial Fovea	20.4 +/− 8.08
Deep Fovea	36.2 +/− 7.87
Full Fovea	32.7 +/− 7.23
Parafoveal Capillary Density
Superficial Parafovea	45.7 +/− 4.71
Deep Parafovea	49.2 +/− 4.02
Full Parafovea	50.3 +/− 5.13
Macular Capillary Density
Superficial Macula	43.4 +/− 3.27
Deep Macula	42.9 +/− 4.29
Full Macula	46.8 +/− 3.76

The comparison between neuroimaging metrics and OCTA biomarkers are summarized in Table 2. High burden of WMH was associated with decreased capillary density in the superficial (β = −7.56 [−12.60, −2.52], p = 0.003), deep (β = −9.66 [−16.17, −3.15], p = 0.004), and full fovea (β = −8.56 [−15.07, −2.05], p = 0.01), but not in the parafovea or whole macula (p > 0.05) (Figure 2). Similarly, participants with amyloid PET positivity had significantly decreased capillary density in the superficial (β = −4.93 [−9.30, −0.56], p = 0.027) and deep fovea (β = −5.20 [−9.96, −0.44], p = 0.032), but not the full retina (β = −3.55 [−8.00, 0.90], p = 0.117). There was a positive association between amyloid presence and capillary density in the superficial parafovea (β = 2.20 [0.12, 4.32], p = 0.038) but not other areas of the parafovea or macula (p > 0.05) (Figure 2). No statistical associations were found between OCTA metrics and the presence of cerebral microbleeds or lacunar infarcts. (Supplementary Figure 1).

Table 2.

Regression coefficients for associations between small vessel disease neuroimaging metrics and optical coherence tomography angiography (OCTA) metrics.

	Amyloid PET Positivity	Presence of Cerebral Microbleeds	Presence of Lacunar Infarcts	High Burden of White Matter Hyperintensity
Superficial Fovea Capillary Density	−4.93 [−9.28, −0.58] (p = 0.027)	2.52 [−2.16, 7.20] (p = 0.292)	−3.29 [−8.85, 2.27] (p = 0.247)	−7.56 [−12.6, −2.55] (p = 0.003)
Deep Fovea Capillary Density	−5.2 [−9.94, −0.46] (p = 0.032)	2.30 [−1.83, 6.43] (p = 0.277)	−1.25 [−6.69, 4.19] (p = 0.654)	−9.66 [−16.1, −3.19] (p = 0.004)
Full Retina Foveal Capillary Density	−3.55 [−7.98, 0.88] (p = 0.117)	2.65 [−1.04, 6.33]	−0.82 [−5.58, 3.94] (p = 0.737)	−8.56 [−15.0, −2.09] (p = 0.01)
Superficial Parafovea Capillary Density	2.22 [0.13, 4.31] (p = 0.038)	1.45 [−0.97, 3.87] (p = 0.240)	1.50 [−2.22, 5.22] (p = 0.433)	−2.55 [−6.70, 1.60] (p = 0.231)
Deep Parafovea Capillary Density	2.13 [−0.11, 4.37] (p = 0.064)	1.19 [−1.05, 3.43] (p = 0.301)	2.13 [−1.38, 5.64] (p = 0.236)	−0.68 [−4.42, 3.06] (p = 0.722)
Full Retina Parafoveal Capillary Density	2.28 [−0.51, 5.07] (p = 0.110)	2.06 [−0.61, 4.73] (p = 0.133)	0.66 [−2.85, 4.17] (p = 0.715)	−2.84 [−8.07, 2.39] (p = 0.289)
Superficial Macula Capillary Density	1.13 [−0.45, 2.71] (p = 0.164)	1.27 [−0.43, 2.97] (p = 0.147)	0.98 [−1.50, 3.46] (p = 0.441)	−2.11 [−4.71, 0.48] (p = 0.112)
Deep Macula Capillary Density	0.87 [−1.94, 3.68] (p = 0.546)	1.88 [−0.62, 4.38] (p = 0.141)	2.43 [−1.55, 6.41] (p = 0.235)	−2.20 [−6.06, 1.66] (p = 0.267)
Full Retina Macular Capillary Density	0.93 [−1.39, 3.25] (p = 0.432)	1.68 [−0.31, 3.67] (p = 0.101)	0.60 [−1.74, 2.94] (p = 0.618)	−2.60 [−6.09, 0.89] (p = 0.146)
Foveal Avascular Zone	0.09 [0.01, 0.17] (p = 0.031)	−0.06 [−0.12, 0.0] (p = 0.100)	−0.01 [−0.09, 0.07] (p = 0.865)	0.19 [−0.02, 0.40] (p = 0.075)

Figure 2.

Optical coherence tomography angiography (OCTA) measurements of capillary density measurements in different retinal layers separated by high and low white matter hyperintensity burden and by amyloid PET positivity status. * indicates a significant difference between the two groups (p<0.05).

When looking at associations with the FAZ, participants with amyloid PET positivity had a larger FAZ (β = 0.09 [0.01, 0.17], p = 0.031). There was no association between FAZ area and high burden of WMH (β = 0.19 [−0.02, 0.40], p = 0.075) as shown in Figure 3. There was also no significant association between FAZ area the presence of CMB or lacunar infarcts (p > 0.05) (Supplementary Figure 2).

Figure 3.

Representative optical coherence tomography angiography (OCTA) capillary density images and foveal avascular zone (FAZ) measurements from three participants: amyloid negative and low white matter hyperintensity (WMH) burden (left), amyloid positive and low WMH burden (center), and amyloid positive and high WMH burden (right). Of note, there were no subjects in this cohort who were amyloid negative with high WMH burden.

4. Discussion

In this study, we analyzed associations between OCTA metrics and imaging biomarkers of small vessel disease on MRI (WMH burden, CMB, lacunar infarcts) as well as amyloid positivity on PET. We found that lower capillary densities in the superficial and deep fovea and superficial parafovea of the retina were associated with high WMH burden, while amyloid PET positivity in the brain was associated with lower capillary density in the superficial and deep fovea but higher capillary density in the superficial parafovea. The tropism for the fovea may indicate that areas of higher concentration of cone cells may be the first to be affected by small vessel disease with possible progression outwards towards the parafovea.

Consistent with existing literature, we observed a significant increase in the size of the FAZ in participants who were amyloid PET positive and a larger FAZ in those with high WMH burden, although the latter did not achieve statistical significance[4, 5]. It is possible that as the vascular capillary density decreases due to small vessel disease, the FAZ area increases; capillary density may be an earlier marker of progression of cerebral microvascular disease than FAZ area. These changes are visually apparent on OCTA capillary density maps and FAZ measurements as shown in Figure 2. Although we observed associations of OCTA metrics with WMH burden, there was no association with other neuroimaging markers of small vessel disease including lacunar infarcts and CMBs. The latter may be heavily influenced by other confounders, which this study was underpowered to address, such as underlying hypertension or APOE genotype, respectively[12]. Moreover, both lacunar infarcts and CMBs result in discrete lesions in the brain tissue, whereas a high WMH burden may represent more widespread neuronal degeneration with effects evidenced in the retinal capillaries.

The relationship between lower capillary density or larger FAZ with amyloid positivity may be driven by indirect relationships such as vascular risk factors, co-existing elevated WMH and neurodegeneration. Ex-vivo brain studies have shown that areas of high WMH burden were associated with colocalized vascular pathologic change and amyloid plaque deposition[19, 20]. Amyloid deposition and cerebrovascular disease may interact to promote breakdown of the blood brain barrier and neurodegeneration [14]. Interestingly, we also observed increased capillary density in the superficial parafovea associated with amyloid positivity, which has been observed before in cognitively unimpaired individuals [6], despite lower density being associated with Alzheimer’s disease [2]. It is possible that if density changes develop radially extending outwards from the fovea, early changes associated with amyloid positivity cause a transient increase in capillary density before decreasing. However, a longitudinal study with serial OCT and cognitive testing would be necessary to answer this question. Overall, OCTA offers a unique opportunity to directly visualize and measure the microvasculature of the central nervous system through the retina. Directly measuring amyloid deposition in the retina may further elucidate the mechanism between neurodegeneration and small vessel disease, particularly in longitudinal studies that include cognitively unimpaired individuals as well as those who will later develop dementia.

Our study has limitations. The small size of our cohort (even if larger than most previous series published on this topic) reduced the statistical power of our analyses and may have prevented some associations from reaching significance. Larger studies are necessary to confirm this association. Similarly, a small subset of our cohort had a diagnosis of MCI. Unfortunately, our work was underpowered to test whether MCI status has an interaction effect between OCTA and amyloid positivity or WMH burden. We were able to show a negative association between MCI status and capillary density of the superficial fovea, consistent with the existing literature. Larger studies would be necessary to appreciate the association of MCI with OCTA and neuroimaging metrics. There were no participants who were amyloid negative but had high WMH burden in our cohort, which limited our ability to discriminate some of the associations identified in our analyses. Unmeasured confounders (such as risk factors of cerebrovascular disease like hypertension, diabetes, and smoking history) may have influenced our results, but this study was not powered to include these covariates. Lack of correlation between retinal capillary density and MMSE scores may have been due to the fact that we excluded patients with scores consistent with diagnosis of dementia. Lastly, we acknowledge that a small sample taken from a population-based cohort may result in a lack of racial and ethnic diversity in our sample, which can influence the generalizability of these results.

In summary, our study supports the value of OCTA as a noninvasive surrogate tool to evaluate the cerebral microvasculature and adds to the growing evidence of a possible relationship between microvascular disease and amyloid deposition that may lead to cognitive decline and dementia.