The Association between Foveal Morphology and Macular Pigment Spatial Distribution: An Ethnicity Study

Warwick A, Müller PL, Williams KM, Rudnicka AR, Owen CG, Lee AY, Egan C, Tufail A. Foveal Curvature and Its Associations in UK Biobank Participants

Purpose

To examine whether sociodemographic, and ocular factors relate to optical coherence tomography (OCT)-derived foveal curvature (FC) in healthy individuals.

Methods

We developed a deep learning model to quantify OCT-derived FC from 63,939 participants (age range, 39-70 years). Associations of FC with sociodemographic, and ocular factors were obtained using multilevel regression analysis (to allow for right and left eyes) adjusting for age, sex, ethnicity, height (model 1), visual acuity, spherical equivalent, corneal astigmatism, center point retinal thickness (CPRT), intraocular pressure (model 2), deprivation (Townsend index), higher education, annual income, and birth order (model 3). Fovea curvature was modeled as a z-score.

Results

Males had on average steeper FC (0.077; 95% confidence interval [CI] 0.077-0.078) than females (0.068; 95% CI 0.068-0.069). Compared with whites, non-white individuals showed flatter FC, particularly those of black ethnicity. In black males, -0.80 standard deviation (SD) change when compared with whites (95% CI -0.89, -0.71; P 5.2e10-68). In black females, -0.70 SD change when compared with whites (95% CI -0.77, -0.63; p 2.3e10-93). Ocular factors (visual acuity, refractive status, and CPRT) showed a graded inverse association with FC that persisted after adjustment. Macular curvature showed a positive association with FC. Income showed a linear trend increase in males (P for linear trend = 0.005).

Conclusions

We demonstrate marked differences in FC with ethnicity on the largest cohort studied for this purpose to date. Ocular factors showed a graded association with FC. Implementation of FC quantification in research and on the clinical setting can enhance the understanding of clinical macular phenotypes in health and disease.

Macular thickness varies with age-related macular degeneration genetic risk variants in the UK Biobank cohort

To evaluate the influence AMD risk genomic variants have on macular thickness in the normal population. UK Biobank participants with no significant ocular history were included using the UK Biobank Resource (project 2112). Spectral-domain optical coherence tomography (SD-OCT) images were taken and segmented to define retinal layers. The influence of AMD risk single-nucleotide polymorphisms (SNP) on retinal layer thickness was analysed. AMD risk associated SNPs were strongly associated with outer-retinal layer thickness. The inner-segment outer segment (ISOS)-retinal pigment epithelium (RPE) thickness measurement, representing photoreceptor outer segments was most significantly associated with the cumulative polygenic risk score, composed of 33 AMD-associated variants, resulting in a decreased thickness (p = 1.37 × 10-67). Gene-gene interactions involving the NPLOC4-TSPAN10 SNP rs6565597 were associated with significant changes in outer retinal thickness. Thickness of outer retinal layers is highly associated with the presence of risk AMD SNPs. Specifically, the ISOS-RPE measurement. Changes to ISOS-RPE thickness are seen in clinically normal individuals with AMD risk SNPs suggesting structural changes occur at the macula prior to the onset of disease symptoms or overt clinical signs.

A Multitask Deep-Learning System to Classify Diabetic Macular Edema for Different Optical Coherence Tomography Devices: A Multicenter Analysis

OBJECTIVE

Diabetic macular edema (DME) is the primary cause of vision loss among individuals with diabetes mellitus (DM). We developed, validated, and tested a deep learning (DL) system for classifying DME using images from three common commercially available optical coherence tomography (OCT) devices.

RESEARCH DESIGN AND METHODS

We trained and validated two versions of a multitask convolution neural network (CNN) to classify DME (center-involved DME [CI-DME], non-CI-DME, or absence of DME) using three-dimensional (3D) volume scans and 2D B-scans, respectively. For both 3D and 2D CNNs, we used the residual network (ResNet) as the backbone. For the 3D CNN, we used a 3D version of ResNet-34 with the last fully connected layer removed as the feature extraction module. A total of 73,746 OCT images were used for training and primary validation. External testing was performed using 26,981 images across seven independent data sets from Singapore, Hong Kong, the U.S., China, and Australia.

RESULTS

In classifying the presence or absence of DME, the DL system achieved area under the receiver operating characteristic curves (AUROCs) of 0.937 (95% CI 0.920-0.954), 0.958 (0.930-0.977), and 0.965 (0.948-0.977) for the primary data set obtained from CIRRUS, SPECTRALIS, and Triton OCTs, respectively, in addition to AUROCs >0.906 for the external data sets. For further classification of the CI-DME and non-CI-DME subgroups, the AUROCs were 0.968 (0.940-0.995), 0.951 (0.898-0.982), and 0.975 (0.947-0.991) for the primary data set and >0.894 for the external data sets.

CONCLUSIONS

We demonstrated excellent performance with a DL system for the automated classification of DME, highlighting its potential as a promising second-line screening tool for patients with DM, which may potentially create a more effective triaging mechanism to eye clinics.

Bardet-Biedl syndrome-7 (BBS7) shows treatment potential and a cone-rod dystrophy phenotype that recapitulates the non-human primate model

Purpose: To provide a detailed ophthalmic phenotype of two male patients with Bardet-Biedl Syndrome (BBS) due to mutations in the BBS7 geneMethods: Two brothers ages 26 (Patient 1, P1) and 23 (P2) underwent comprehensive ophthalmic evaluations over three years. Visual function was assessed with full-field electroretinograms (ffERGs), kinetic and chromatic perimetry, multimodal imaging with spectral domain optical coherence tomography (SD-OCT), fundus autofluorescence (FAF) with short- (SW) and near-infrared (NIR) excitation lights and adaptive optics scanning light ophthalmoscopy (AOSLO).Results: Both siblings had a history of obesity and postaxial polydactyly; P2 had diagnoses of type 1 Diabetes Mellitus, Addison's disease, high-functioning autism-spectrum disorder and -12D myopia. Visual acuities were better than 20/30. Kinetic fields were moderately constricted. Cone-mediated ffERGs were undetectable, rod ERGs were ~80% of normal mean. Static perimetry showed severe central cone and rod dysfunction. Foveal to parafoveal hypoautofluorescence, most obvious on NIR-FAF, co-localized with outer segment shortening/loss and outer nuclear layer thinning by SD-OCT, and with reduced photoreceptors densities by AOSLO. A structural-functional dissociation was confirmed for cone- and rod-mediated parameters. Worsening of the above abnormalities was documented by SD-OCT and FAF in P2 at 3 years. Gene screening identified compound heterozygous mutations in BBS7 (p.Val266Glu: c.797 T > A of maternal origin; c.1781_1783delCAT, paternal) in both patients.Conclusions: BBS7-associated retinal degeneration may present as a progressive cone-rod dystrophy pattern, reminiscent of both the murine and non-human primate models of the disease. Predominantly central retinal abnormalities in both cone and rod photoreceptors showed a structural-functional dissociation, an ideal scenario for gene augmentation treatments.

Detection of Diabetic Retinopathy from Ultra-Widefield Scanning Laser Ophthalmoscope Images: A Multicenter Deep Learning Analysis

PURPOSE

To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO).

DESIGN

Observational, cross-sectional study.

PARTICIPANTS

A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina.

METHODS

All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions.

MAIN OUTCOME MEASURES

Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR.

RESULTS

For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection.

CONCLUSIONS

The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.

Foveal macular pigment dip in offspring of age-related macular degeneration patients is inversely associated with omega-3 index

Background

Offspring of parent(s) with age-related macular degeneration (AMD) have a 45% lifetime risk of developing the disease. High foveal macular pigment optical density (MPOD) is protective, whereas individuals with a “foveal macular pigment dip” (FMPD) are at increased risk. Shortage of the dietary carotenoids lutein, zeaxanthin as well as fish consumption are reported AMD risk factors. This Early Biomarkers of AMD (EBAMD) study evaluates serum factors that protect foveal MPOD architecture in Caucasian offspring of parent(s) with AMD.

Methods

N = 130 subjects [mean (SD) age 62.8 (8.6) years; 36/94 male/female] were recruited from Scripps Health/ Scripps Memorial Hospital/ Scripps Mericos Eye Institute between 2012 and 2017. Macula pigment 3D topography was evaluated using specular reflectance. Buccal genetic cheek swab, circulating serum dietary carotenoids and long-term RBC omega-3 fatty acid status, as well as common secondary clinical structural and vision function parameters were obtained.

Results

41 % of offspring of AMD parent(s) presented with FMPD. These offspring were about 4 years younger than those without FMPD (controls; P = 0.012) and had thinner foveas (P = 0.010). There were no differences in gender, BMI, % body fat, visual acuity or contrast sensitivity between those with and without FMPD. % RBC membrane docosahexaenoic acid (DHA) was reduced in FMPD offspring vs. control offspring (P = 0.04). The Omega-3 Index was significantly decreased in the FMPD group (P = 0.03).

Conclusions

The percentage of FMPD in AMD offspring is nearly twice that reported for the general population in the scientific literature. Offspring presenting FMPD had similar AMD genetic risk, but significantly reduced % RBC membrane omega-3 fatty acids and thinner foveas compared with those without FMPD. Our data supports the importance of ‘essential fatty’ acids as an independent AMD risk factor.

Local Abundance of Macular Xanthophyll Pigment Is Associated with Rod- and Cone-Mediated Vision in Aging and Age-Related Macular Degeneration

Purpose

We assessed the association between the abundance of macular xanthophyll carotenoid pigment using dual-wavelength autofluorescence and multimodal vision testing including rod-mediated dark adaptation (RMDA), a measure of retinoid re-supply, in adults ≥60 years old with and without age-related macular degeneration (AMD).

Methods

AMD severity was determined using the nine-step Age-Related Eye Disease Study grading. Tests probed cones (best-corrected visual acuity, contrast sensitivity), cones and rods (low-luminance visual acuity, low-luminance deficit, mesopic light sensitivity), or rods only (scotopic light sensitivity, RMDA). Signal attenuation by macular pigment optical density (MPOD) was estimated using a ratio of blue and green autofluorescence signal to yield mean MPOD in a 1°-diameter fovea-centered disk, mean MPOD in a 2°-diameter disk centered on a perifoveal RMDA test location, and macular pigment optical volume (MPOV, or integrated MPOD) in a 4°-diameter fovea-centered disk. Age-adjusted associations between vision and imaging measures were determined.

Results

In 88 eyes of 88 subjects (age, 74.9 ± 5.8 years) with normal eyes (n = 32), early AMD (n = 23), or intermediate AMD (n = 33), foveal and perifoveal MPOD and MPOV were higher in the AMD eyes than in the normal eyes. At the RMDA test location, higher MPOD was unrelated to AMD severity but was associated with faster RMDA.

Conclusions

In older adults with and without AMD, higher macular xanthophyll concentrations are associated with better best-corrected visual acuity and RMDA. Data are consistent with a model of cone resilience and rod vulnerability in aging and AMD and can be further explored in a larger sample study.

Spatial distribution of macular pigment estimated by autofluorescence imaging in elderly Japanese individuals

PURPOSE

To determine the spatial distribution types of macular pigment (MP) in elderly Japanese individuals and to consider their origin.

STUDY DESIGN

Observational case series.

METHODS

Local MP optical density (MPOD) at some eccentricities and MP volume were measured using the MPOD module of a MultiColor Spectralis in 96 pseudophakic eyes of 96 participants (age range, 52-86 years; mean age, 72.8 ± 8.3 years). The MP distribution types were determined from the MP spatial profiles. The retinal thickness (RT) at the foveal center, at both 0.5° and 0.9° eccentricities, and the foveal width were measured using optical coherence tomography.

RESULTS

The mean local MPOD at the foveal center was 0.79. Spatial distribution was classified into four types: central peak (24.0%), ring-like (40.6%), intermediate (22.9%), and central dip (12.5%). The ring-like type was the most frequent in these Japanese participants. The central-peak type showed lower MPOD than did the other types in the area outside 0.9°. The ring-like type occurred frequently in eyes with small RT at 0.5° and wider foveal width. A rough contour of the Müller cell cone was found more frequently in the central-dip type than in the other types.

CONCLUSIONS

The present characteristics of the different distribution patterns could be explained by the hypothesis that MP presents mainly in the Müller cell cone within 0.5° and in Müller cells in the outer and inner plexiform layers in the area outside 0.5°. The anatomic characteristics of Müller cells at the fovea and parafovea likely affect the MP distribution.

Objective method for measuring the macular pigment optical density in the eye

Macular pigment is a yellowish pigment of purely dietary origin, which is thought to have a protective role in the retina. Recently, it was linked to age-related macular degeneration and improved visual function. In this work, we present a method and a corresponding optical instrument for the rapid measurement of its optical density. The method is based on fundus reflectometry and features a photodetector for the measurement of reflectance at different wavelengths and retinal locations. The method has been tested against a commercially available instrument on a group of healthy volunteers and has shown good correlation. The proposed instrument can serve as a rapid, non-midriatic, low-cost tool for the measurement of macular pigment optical density.

Investigating the potential of Zernike polynomials to characterise spatial distribution of macular pigment

It has been postulated that particular patterns of macular pigment (MP) distribution may be associated with the risk for eye diseases such as age-related macular degeneration (AMD). This work investigates the potential of Zernike polynomials (ZP) to characterise the level and distribution of MP, and their suitability as a representation for analysis of the effects of age and AMD on MP patterns. As the case study, MP distribution maps computed using an experimental method based on fundus reflectance (MRIA) were obtained for ninety volunteers representing three groups: under-fifty without AMD, fifty and over without AMD, and fifty and over with AMD. ZP with 105 coefficients were fitted to the maps using least-squares optimisation and found to represent MP maps accurately (RMSE<10⁻¹). One-way MANOVA analysis carried out on ZP representations showed that the three subject groups have significantly different means (Wilk’s Lambda 0.125, p<0.0001). Linear discriminant analysis with leave-one-out scheme resulted in accuracy, sensitivity and specificity of classification according to, respectively, disease status regardless of age (81% all); disease status in the age-matched groups (87%, 88%, 86%); age irrespective of disease status (81%, 83%, 73%); and age for subjects without AMD (83%, 88%, 80%). Mean MP distributions computed from ZP coefficients for the three groups showed more elevated and more peaked MP for the healthy under-fifty group; more irregular and more elevated peripheral levels in over-fifty AMD group than in over-fifty non-AMD group; and moderate radial asymmetry in non-AMD over-50 group. The results suggest that ZP coefficients are capable of accurately representing MP in a way that captures certain spatial patterns of its distribution. Using the ZP representation MP maps could be classified according to both age and disease status with accuracy significantly greater than chance, with peak elevation, pattern irregularity and radial asymmetry identified as important features.

Repeatability of the macular pigment spatial profile: A comparison of objective versus subjective classification

PURPOSE

Classification of macular pigment (MP) spatial profile phenotypes varies and is often based on subjective visualisation. We investigated repeatability of MP optical density (MPOD) comparing an objective versus subjective profiling system.

METHODS

The coefficient of repeatability (CoR) was calculated for point MPOD values (0-3.8°) obtained by dual-wavelength fundus autofluorescence (FAF) from two scans obtained in a single visit of 40 healthy individuals (39 ± 9 years). For each individual's dataset, the MP profile was classified as exponential, ring-like or central dip using an objective method (based on deviations away from an exponential fit), as well as by subjective visual profiling. Existing FAF images of 88 monozygotic (MZ) and 69 dizygotic (DZ) twin pairs were reanalysed using the objective profiling method and concordance and heritability of ring-like profiles determined.

RESULTS

The CoR was 0.23 at 0° and 0.06 at 0.8°. Agreement of objective profiling between scans was excellent (κ = 0.85, 95% CI 0.69 to 1.00; p < 0.0005). Subjective profiling showed moderate agreement between scans (κ = 0.48, 95% CI 0.23 to 0.73; p < 0.0005). Agreement between objective and subjective classification was low (κ = 0.23, 95% CI 0.04 to 0.42; p = 0.02). Concordance for the ring-like profile using objective profiling was 0.74 for MZ compared to 0.36 for DZ twins. Heritability was calculated as 81.5% (95% confidence interval 61.1-93.1%).

CONCLUSION

Compared to visual assessment, objective MP profiling is a more reliable method and should be considered in future observational and interventional studies. In addition, MP profile phenotypes showed high heritability.

The primate fovea: Structure, function and development

A fovea is a pitted invagination in the inner retinal tissue (fovea interna) that overlies an area of photoreceptors specialized for high acuity vision (fovea externa). Although the shape of the vertebrate fovea varies considerably among the species, there are two basic types. The retina of many predatory fish, reptilians, and birds possess one (or two) convexiclivate fovea(s), while the retina of higher primates contains a concaviclivate fovea. By refraction of the incoming light, the convexiclivate fovea may function as image enlarger, focus indicator, and movement detector. By centrifugal displacement of the inner retinal layers, which increases the transparency of the central foveal tissue (the foveola), the primate fovea interna improves the quality of the image received by the central photoreceptors. In this review, we summarize ‒ with the focus on Müller cells of the human and macaque fovea ‒ data regarding the structure of the primate fovea, discuss various aspects of the optical function of the fovea, and propose a model of foveal development. The "Müller cell cone" of the foveola comprises specialized Müller cells which do not support neuronal activity but may serve optical and structural functions. In addition to the "Müller cell cone", structural stabilization of the foveal morphology may be provided by the 'z-shaped' Müller cells of the fovea walls, via exerting tractional forces onto Henle fibers. The spatial distribution of glial fibrillary acidic protein may suggest that the foveola and the Henle fiber layer are subjects to mechanical stress. During development, the foveal pit is proposed to be formed by a vertical contraction of the centralmost Müller cells. After widening of the foveal pit likely mediated by retracting astrocytes, Henle fibers are formed by horizontal contraction of Müller cell processes in the outer plexiform layer and the centripetal displacement of photoreceptors. A better understanding of the molecular, cellular, and mechanical factors involved in the developmental morphogenesis and the structural stabilization of the fovea may help to explain the (patho-) genesis of foveal hypoplasia and macular holes.

Interindividual Variations in Foveal Anatomy and Artifacts Seen on Inner Retinal Probability Maps from Spectral Domain OCT Scans of the Macula

Purpose

We tested the hypothesis that variations in foveal morphology can account for artifacts seen on optical coherence tomography (OCT) retinal ganglion cell (RGC) layer probability maps.

Methods

A total of 126 healthy subjects were tested with spectral domain (sd) OCT. Thickness and probability maps of the macular RGC plus inner plexiform layer (RGC+) were obtained with customized software. Macular b-scans were analyzed to derive three foveal anatomic parameters: width, depth, and slope. The distribution of these parameters was compared between eyes with and without circumfoveal artifacts seen in the central 4° of macular RGC+ probability maps.

Results

Of 126 healthy subjects, 12 (9.5%) had an abnormal circumfoveal region (artifact) on RGC+ probability maps. Based upon the normal distribution of the three anatomic parameters, only three of the 12 eyes (25%) fell outside the 95% confidence interval of one or more of the three foveal morphologic parameters. Multivariable logistic regression revealed that the parameter slope was significantly associated with the presence of these artifacts (odds ratio = 0.26; P = 0.019). However, the combination of these parameters and age explained only 11% of the total variance of these artifacts.

Conclusions

Fovea morphology, as measured based upon width, depth, and slope, has a minor role in explaining artifacts seen on macular scans. Variations in the distribution of RGC+ thickness that are not reflected in our measures warrant further investigation as potential sources of artifacts.

Translational Relevance

A small proportion of circumfoveal artifacts seen on RGC+ probability maps can be explained by variations in foveal anatomy.