The Evolving Role of the Relationship between Optic Nerve Structure and Function in Glaucoma

Deep Learning Estimation of 10-2 Visual Field Map Based on Macular Optical Coherence Tomography Angiography Measurements

PURPOSE

To develop deep learning (DL) models estimating the central visual field (VF) from optical coherence tomography angiography (OCTA) vessel density (VD) measurements.

DESIGN

Development and validation of a deep learning model.

METHODS

A total of 1051 10-2 VF OCTA pairs from healthy, glaucoma suspects, and glaucoma eyes were included. DL models were trained on en face macula VD images from OCTA to estimate 10-2 mean deviation (MD), pattern standard deviation (PSD), 68 total deviation (TD) and pattern deviation (PD) values and compared with a linear regression (LR) model with the same input. Accuracy of the models was evaluated by calculating the average mean absolute error (MAE) and the R2 (squared Pearson correlation coefficients) of the estimated and actual VF values.

RESULTS

DL models predicting 10-2 MD achieved R2 of 0.85 (95% confidence interval [CI], 74-0.92) for 10-2 MD and MAEs of 1.76 dB (95% CI, 1.39-2.17 dB) for MD. This was significantly better than mean linear estimates for 10-2 MD. The DL model outperformed the LR model for the estimation of pointwise TD values with an average MAE of 2.48 dB (95% CI, 1.99-3.02) and R2 of 0.69 (95% CI, 0.57-0.76) over all test points. The DL model outperformed the LR model for the estimation of all sectors.

CONCLUSIONS

DL models enable the estimation of VF loss from OCTA images with high accuracy. Applying DL to the OCTA images may enhance clinical decision making. It also may improve individualized patient care and risk stratification of patients who are at risk for central VF damage.

Nanocarriers for the Delivery of Neuroprotective Agents in the Treatment of Ocular Neurodegenerative Diseases

Retinal neurodegeneration is considered an early event in the pathogenesis of several ocular diseases, such as diabetic retinopathy, age-related macular degeneration, and glaucoma. At present, there is no definitive treatment to prevent the progression or reversal of vision loss caused by photoreceptor degeneration and the death of retinal ganglion cells. Neuroprotective approaches are being developed to increase the life expectancy of neurons by maintaining their shape/function and thus prevent the loss of vision and blindness. A successful neuroprotective approach could prolong patients' vision functioning and quality of life. Conventional pharmaceutical technologies have been investigated for delivering ocular medications; however, the distinctive structural characteristics of the eye and the physiological ocular barriers restrict the efficient delivery of drugs. Recent developments in bio-adhesive in situ gelling systems and nanotechnology-based targeted/sustained drug delivery systems are receiving a lot of attention. This review summarizes the putative mechanism, pharmacokinetics, and mode of administration of neuroprotective drugs used to treat ocular disorders. Additionally, this review focuses on cutting-edge nanocarriers that demonstrated promising results in treating ocular neurodegenerative diseases.

PyVisualFields: A Python Package for Visual Field Analysis

Purpose

Artificial intelligence (AI) methods are changing all areas of research and have a variety of capabilities of analysis in ophthalmology, specifically in visual fields (VFs) to detect or predict vision loss progression. Whereas most of the AI algorithms are implemented in Python language, which offers numerous open-source functions and algorithms, the majority of algorithms in VF analysis are offered in the R language. This paper introduces PyVisualFields, a developed package to address this gap and make available VF analysis in the Python language.

Methods

For the first version, the R libraries for VF analysis provided by vfprogression and visualFields packages are analyzed to define the overlaps and distinct functions. Then, we defined and translated this functionality into Python with the help of the wrapper library rpy2. Besides maintaining, the subsequent versions' milestones are established, and the third version will be R-independent.

Results

The developed Python package is available as open-source software via the GitHub repository and is ready to be installed from PyPI. Several Jupyter notebooks are prepared to demonstrate and describe the capabilities of the PyVisualFields package in the categories of data presentation, normalization and deviation analysis, plotting, scoring, and progression analysis.

Conclusions

We developed a Python package and demonstrated its functionality for VF analysis and facilitating ophthalmic research in VF statistical analysis, illustration, and progression prediction.

Translational Relevance

Using this software package, researchers working on VF analysis can more quickly create algorithms for clinical applications using cutting-edge AI techniques.

Deep Learning Estimation of 10-2 Visual Field Map Based on Circumpapillary Retinal Nerve Fiber Layer Thickness Measurements

PURPOSE

To estimate central 10-degree visual field (VF) map from spectral-domain optical coherence tomography (SD-OCT) retinal nerve fiber layer thickness (RNFL) measurements in glaucoma with artificial intelligence.

DESIGN

Artificial intelligence (convolutional neural networks) study.

METHODS

This study included 5352 SD-OCT scans and 10-2 VF pairs from 1365 eyes of 724 healthy patients, patients with suspected glaucoma, and patients with glaucoma. Convolutional neural networks (CNNs) were developed to estimate the 68 individual sensitivity thresholds of 10-2 VF map using all-sectors (CNN_A) and temporal-sectors (CNN_T) RNFL thickness information of the SD-OCT circle scan (768 thickness points). 10-2 indices including pointwise total deviation (TD) values, mean deviation (MD), and pattern standard deviation (PSD) were generated using the CNN-estimated sensitivity thresholds at individual test locations. Linear regression (LR) models with the same input were used for comparison.

RESULTS

The CNN_A model achieved an average pointwise mean absolute error of 4.04 dB (95% confidence interval [CI] 3.76-4.35) and correlation coefficient (r) of 0.59 (95% CI 0.52-0.64) over 10-2 map and the mean absolute error and r of 2.88 dB (95% CI 2.63-3.15) and 0.74 (95% CI 0.67-0.80) for MD, and 2.31 dB (95% CI 2.03-2.61) and 0.59 (95% CI 0.51-0.65) for PSD estimations, respectively, significantly outperforming the LR_A model.

CONCLUSIONS

The proposed CNN_A model improved the estimation of 10-2 VF map based on circumpapillary SD-OCT RNFL thickness measurements. These artificial intelligence methods using SD-OCT structural data show promise to individualize the frequency of central VF assessment in patients with glaucoma and would enable the reallocation of resources from patients at lowest risk to those at highest risk of central VF damage.

Detecting glaucoma from multi-modal data using probabilistic deep learning

Objective

To assess the accuracy of probabilistic deep learning models to discriminate normal eyes and eyes with glaucoma from fundus photographs and visual fields.

Design

Algorithm development for discriminating normal and glaucoma eyes using data from multicenter, cross-sectional, case-control study.

Subjects and participants

Fundus photograph and visual field data from 1,655 eyes of 929 normal and glaucoma subjects to develop and test deep learning models and an independent group of 196 eyes of 98 normal and glaucoma patients to validate deep learning models.

Main outcome measures

Accuracy and area under the receiver-operating characteristic curve (AUC).

Methods

Fundus photographs and OCT images were carefully examined by clinicians to identify glaucomatous optic neuropathy (GON). When GON was detected by the reader, the finding was further evaluated by another clinician. Three probabilistic deep convolutional neural network (CNN) models were developed using 1,655 fundus photographs, 1,655 visual fields, and 1,655 pairs of fundus photographs and visual fields collected from Compass instruments. Deep learning models were trained and tested using 80% of fundus photographs and visual fields for training set and 20% of the data for testing set. Models were further validated using an independent validation dataset. The performance of the probabilistic deep learning model was compared with that of the corresponding deterministic CNN model.

Results

The AUC of the deep learning model in detecting glaucoma from fundus photographs, visual fields, and combined modalities using development dataset were 0.90 (95% confidence interval: 0.89-0.92), 0.89 (0.88-0.91), and 0.94 (0.92-0.96), respectively. The AUC of the deep learning model in detecting glaucoma from fundus photographs, visual fields, and both modalities using the independent validation dataset were 0.94 (0.92-0.95), 0.98 (0.98-0.99), and 0.98 (0.98-0.99), respectively. The AUC of the deep learning model in detecting glaucoma from fundus photographs, visual fields, and both modalities using an early glaucoma subset were 0.90 (0.88,0.91), 0.74 (0.73,0.75), 0.91 (0.89,0.93), respectively. Eyes that were misclassified had significantly higher uncertainty in likelihood of diagnosis compared to eyes that were classified correctly. The uncertainty level of the correctly classified eyes is much lower in the combined model compared to the model based on visual fields only. The AUCs of the deterministic CNN model using fundus images, visual field, and combined modalities based on the development dataset were 0.87 (0.85,0.90), 0.88 (0.84,0.91), and 0.91 (0.89,0.94), and the AUCs based on the independent validation dataset were 0.91 (0.89,0.93), 0.97 (0.95,0.99), and 0.97 (0.96,0.99), respectively, while the AUCs based on an early glaucoma subset were 0.88 (0.86,0.91), 0.75 (0.73,0.77), and 0.92 (0.89,0.95), respectively.

Conclusion and relevance

Probabilistic deep learning models can detect glaucoma from multi-modal data with high accuracy. Our findings suggest that models based on combined visual field and fundus photograph modalities detects glaucoma with higher accuracy. While probabilistic and deterministic CNN models provided similar performance, probabilistic models generate certainty level of the outcome thus providing another level of confidence in decision making.

Epigenetic Regulation of Optic Nerve Development, Protection, and Repair

Epigenetic factors are known to influence tissue development, functionality, and their response to pathophysiology. This review will focus on different types of epigenetic regulators and their associated molecular apparatus that affect the optic nerve. A comprehensive understanding of epigenetic regulation in optic nerve development and homeostasis will help us unravel novel molecular pathways and pave the way to design blueprints for effective therapeutics to address optic nerve protection, repair, and regeneration.

Structure-Function Relationship between Cluster Mean Defect and Sector Peripapillary Retinal Nerve Fiber Layer Thickness in Primary Open Angle Glaucoma

Purpose

To determine the structure–function relationship between cluster mean defect (MD) offered by standard automated perimetry and corresponding sector peripapillary retinal nerve fiber layer thickness (pRNFLT) measured with optical coherence tomography (OCT) in primary open angle glaucoma (POAG).

Method

39 healthy eyes (control group), 43 early POAG eyes (global MD ≤ 6 dB, early group), 30 moderate POAG eyes (global MD between 6 and 12 dB, moderate group), and 53 advanced POAG eyes (global MD > 12 dB, advanced group) underwent visual field (VF) examination with Octopus perimeter (dynamic strategy/G2 pattern) and peripapillary retinal nerve fiber layer thickness measurements with RTVue-100 FD-OCT. Spearman analysis was used to investigate the correlation between cluster MDs provided by Octopus perimeter and corresponding sector pRNFLT for the total sample and each subgroup, respectively. Then, linear (y = a+ bx) and curvilinear (quadratic, y = a+bx + cx²) regression analyses were employed to investigate the model for the cluster MD-sector pRNFLT pair with significant correlation. The strength of the relationship was characterized with correlation coefficient (ρ) and coefficient of determination (R²). For the cluster–sector pair that could be fitted by both models, Wilcoxon signed rank test of absolute residuals was used to compare the goodness of fit.

Results

Correlation between cluster MDs and corresponding sector pRNFLT was significant for all clusters in the total sample (ρ values: −0.572 to 0.832, P < 0.001) and in the POAG group (ρ values: −0.551 to −0.777, P < 0.001). The highest ρ values were found for cluster-sector pair 9 and pair 3, respectively. The curvilinear (quadratic) model provided better fit for all 10 cluster-sector pairs in the total sample (R² values: 0.431–0.687, P < 0.001) and in the POAG group (R² values: 0.364–0.594, P < 0.01). The highest R² values were found also for cluster–sector pair 9 and pair 3, respectively. In the control group, no significant correlation was found for any cluster–sector pair (P > 0.01). In the early group, correlation was significant for cluster–sector pairs 3, 8, and 9 (ρ values: −0.449, −0.627, and −0.815, resp., P < 0.01). In the moderate group, correlation was significant for pairs 2, 3, 8, and 9 (ρ values: −0.703, −0.556, −0.680, and −0.637, resp., P < 0.01). In the advanced group, correlation was significant (P < 0.01) for all 10 pairs (ρ values: −0.395 to −0.699, P < 0.001) except for pairs 2, 3, and 8, and the highest ρ value was found for pair 1. For all cluster–sector pairs with significant correlation in the early, moderate, and advanced groups, only linear model could be fitted (P < 0.01), except for pair 9 in the early group and pair 5 in the advanced group.

Conclusions

Cluster MD of the Octopus visual field showed significant moderate-to-strong negative correlation and curvilinear (quadratic) relationship with the corresponding sector pRNFLT for POAG. This type of regional structure–function relationship varied according to the severity of POAG, and at each stage, the significantly correlated cluster–sector pairs mainly showed linear relationship. The results could provide guidance for better utilization of this regional structure–function method in the management of different stages of POAG.

Degeneration of retina-brain components and connections in glaucoma: Disease causation and treatment options for eyesight preservation

Eyesight is the most important of our sensory systems for optimal daily activities and overall survival. Patients who experience visual impairment due to elevated intraocular pressure (IOP) are often those afflicted with primary open-angle glaucoma (POAG) which slowly robs them of their vision unless treatment is administered soon after diagnosis. The hallmark features of POAG and other forms of glaucoma are damaged optic nerve, retinal ganglion cell (RGC) loss and atrophied RGC axons connecting to various brain regions associated with receipt of visual input from the eyes and eventual decoding and perception of images in the visual cortex. Even though increased IOP is the major risk factor for POAG, the disease is caused by many injurious chemicals and events that progress slowly within all components of the eye-brain visual axis. Lowering of IOP mitigates the damage to some extent with existing drugs, surgical and device implantation therapeutic interventions. However, since multifactorial degenerative processes occur during aging and with glaucomatous optic neuropathy, different forms of neuroprotective, nutraceutical and electroceutical regenerative and revitalizing agents and processes are being considered to combat these eye-brain disorders. These aspects form the basis of this short review article.

Neuroprotection Against NMDA-Induced Retinal Damage by Philanthotoxin-343 Involves Reduced Nitrosative Stress

N-methyl-D-aspartate receptor (NMDAR) overstimulation is known to mediate neurodegeneration, and hence represents a relevant therapeutic target for neurodegenerative disorders including glaucoma. This study examined the neuroprotective effects of philanthotoxin (PhTX)-343 against NMDA-induced retinal injury in rats. Male Sprague Dawley rats were divided into three groups; group 1 received phosphate buffer saline as the negative control, group 2 was injected with NMDA (160 nM) to induce retinal excitotoxic injury, and group 3 was pre-treated with PhTX-343 (160 nM) 24 h before NMDA exposure. All treatments were given intravitreally and bilaterally. Seven days post-treatment, rats were subjected to visual behaviour assessments using open field and colour recognition tests. Rats were then euthanized, and the retinas were harvested and subjected to haematoxylin and eosin (H&E) staining for morphometric analysis and 3-nitrotyrosine (3-NT) ELISA protocol as the nitrosative stress biomarker. PhTX-343 treatment prior to NMDA exposure improved the ability of rats to recognize visual cues and preserved visual functions (i.e., recognition of objects with different colours). Morphological examination of retinal tissues showed that the fractional ganglion cell layer thickness within the inner retina (IR) in the PhTX-343 treated group was greater by 1.28-fold as compared to NMDA-treated rats (p < 0.05) and was comparable to control rats (p > 0.05). Additionally, the number of retinal cell nuclei/100 μm² in IR for the PhTX-343-treated group was greater by 1.82-fold compared to NMDA-treated rats (p < 0.05) and was comparable to control group (p > 0.05). PhTX-343 also reduced the retinal 3-NT levels by 1.74-fold compared to NMDA-treated rats (p < 0.05). In conclusion, PhTX-343 pretreatment protects against NMDA-induced retinal morphological changes and visual impairment by suppressing nitrosative stress as reflected by the reduced retinal 3-NT level.

A Decrease in Bruch's Membrane Opening-Minimum Rim Area Precedes Decreased Retinal Nerve Fiber Layer Thickness and Visual Field Loss in Glaucoma

PRCIS

A decrease in Bruch's membrane opening-minimum rim area, which represents the optic nerve head (ONH), preceded a decrease in the peripapillary retinal nerve fiber layer thickness (RNFLT) and the visual field index (VFI).

PURPOSE

This study aimed to investigate the relative comparison between a decrease in BMO-MRA, the peripapillary RNFLT, and the VFI, according to the severity of glaucoma.

MATERIALS AND METHODS

This retrospective cross-sectional study included 121 eyes (73 with open-angle glaucoma and 48 normal eyes). The ONH and retinal nerve fiber layer were analyzed using spectral domain optical coherence tomography, and VFI was obtained using the Humphrey Field Analyzer. The tipping points of RNFLT for VFI and BMO-MRA were estimated using broken-stick regression models. Polynomial regression analysis was performed, and the changes in the 3 parameters were expressed as a graph.

RESULTS

The tipping point of the RNFLT for the VFI was 88.62 μm [95% confidence interval (CI): 79.59-97.65; P=0.001]. The tipping point of the RNFLT for BMO-MRA was 60.00 μm (95% CI: 48.28-71.72; P=0.220). Above the tipping point, BMO-MRA decreased with a decrease in the RNFLT (slope=0.0135; 95% CI: 0.0115-0.0155; P<0.001); below the tipping point, BMO-MRA did not decrease significantly (slope=0.0002; 95% CI: -0.0177 to 0.0181; P=0.983). Polynomial regression analysis showed that with the progression of glaucoma, BMO-MRA decreased more rapidly, and this preceded a decrease in the RNFLT followed by a decrease in the VFI.

CONCLUSION

The ONH parameter, BMO-MRA, showed a faster decrease than RNFLT and VFI in early glaucoma. BMO-MRA may help detect early glaucomatous damage and its progression.

Deep level set method for optic disc and cup segmentation on fundus images

Glaucoma is a leading cause of blindness. The measurement of vertical cup-to-disc ratio combined with other clinical features is one of the methods used to screen glaucoma. In this paper, we propose a deep level set method to implement the segmentation of optic cup (OC) and optic disc (OD). We present a multi-scale convolutional neural network as the prediction network to generate level set initial contour and evolution parameters. The initial contour will be further refined based on the evolution parameters. The network is integrated with augmented prior knowledge and supervised by active contour loss, which makes the level set evolution yield more accurate shape and boundary details. The experimental results on the REFUGE dataset show that the IoU of the OC and OD are 93.61% and 96.69%, respectively. To evaluate the robustness of the proposed method, we further test the model on the Drishthi-GS1 dataset. The segmentation results show that the proposed method outperforms the state-of-the-art methods.

Deep Learning Estimation of 10-2 and 24-2 Visual Field Metrics Based on Thickness Maps from Macula OCT

PURPOSE

To develop deep learning (DL) systems estimating visual function from macula-centered spectral-domain (SD) OCT images.

DESIGN

Evaluation of a diagnostic technology.

PARTICIPANTS

A total of 2408 10-2 visual field (VF) SD OCT pairs and 2999 24-2 VF SD OCT pairs collected from 645 healthy and glaucoma subjects (1222 eyes).

METHODS

Deep learning models were trained on thickness maps from Spectralis macula SD OCT to estimate 10-2 and 24-2 VF mean deviation (MD) and pattern standard deviation (PSD). Individual and combined DL models were trained using thickness data from 6 layers (retinal nerve fiber layer [RNFL], ganglion cell layer [GCL], inner plexiform layer [IPL], ganglion cell-IPL [GCIPL], ganglion cell complex [GCC] and retina). Linear regression of mean layer thicknesses were used for comparison.

MAIN OUTCOME MEASURES

Deep learning models were evaluated using R² and mean absolute error (MAE) compared with 10-2 and 24-2 VF measurements.

RESULTS

Combined DL models estimating 10-2 achieved R² of 0.82 (95% confidence interval [CI], 0.68-0.89) for MD and 0.69 (95% CI, 0.55-0.81) for PSD and MAEs of 1.9 dB (95% CI, 1.6-2.4 dB) for MD and 1.5 dB (95% CI, 1.2-1.9 dB) for PSD. This was significantly better than mean thickness estimates for 10-2 MD (0.61 [95% CI, 0.47-0.71] and 3.0 dB [95% CI, 2.5-3.5 dB]) and 10-2 PSD (0.46 [95% CI, 0.31-0.60] and 2.3 dB [95% CI, 1.8-2.7 dB]). Combined DL models estimating 24-2 achieved R² of 0.79 (95% CI, 0.72-0.84) for MD and 0.68 (95% CI, 0.53-0.79) for PSD and MAEs of 2.1 dB (95% CI, 1.8-2.5 dB) for MD and 1.5 dB (95% CI, 1.3-1.9 dB) for PSD. This was significantly better than mean thickness estimates for 24-2 MD (0.41 [95% CI, 0.26-0.57] and 3.4 dB [95% CI, 2.7-4.5 dB]) and 24-2 PSD (0.38 [95% CI, 0.20-0.57] and 2.4 dB [95% CI, 2.0-2.8 dB]). The GCIPL (R² = 0.79) and GCC (R² = 0.75) had the highest performance estimating 10-2 and 24-2 MD, respectively.

CONCLUSIONS

Deep learning models improved estimates of functional loss from SD OCT imaging. Accurate estimates can help clinicians to individualize VF testing to patients.

Therapeutic Drugs and Devices for Tackling Ocular Hypertension and Glaucoma, and Need for Neuroprotection and Cytoprotective Therapies

Damage to the optic nerve and the death of associated retinal ganglion cells (RGCs) by elevated intraocular pressure (IOP), also known as glaucoma, is responsible for visual impairment and blindness in millions of people worldwide. The ocular hypertension (OHT) and the deleterious mechanical forces it exerts at the back of the eye, at the level of the optic nerve head/optic disc and lamina cribosa, is the only modifiable risk factor associated with glaucoma that can be treated. The elevated IOP occurs due to the inability of accumulated aqueous humor (AQH) to egress from the anterior chamber of the eye due to occlusion of the major outflow pathway, the trabecular meshwork (TM) and Schlemm’s canal (SC). Several different classes of pharmaceutical agents, surgical techniques and implantable devices have been developed to lower and control IOP. First-line drugs to promote AQH outflow via the uveoscleral outflow pathway include FP-receptor prostaglandin (PG) agonists (e.g., latanoprost, travoprost and tafluprost) and a novel non-PG EP2-receptor agonist (omidenepag isopropyl, Eybelis^®). TM/SC outflow enhancing drugs are also effective ocular hypotensive agents (e.g., rho kinase inhibitors like ripasudil and netarsudil; and latanoprostene bunod, a conjugate of a nitric oxide donor and latanoprost). One of the most effective anterior chamber AQH microshunt devices is the Preserflo^® microshunt which can lower IOP down to 10–13 mmHg. Other IOP-lowering drugs and devices on the horizon will be also discussed. Additionally, since elevated IOP is only one of many risk factors for development of glaucomatous optic neuropathy, a treatise of the role of inflammatory neurodegeneration of the optic nerve and retinal ganglion cells and appropriate neuroprotective strategies to mitigate this disease will also be reviewed and discussed.

Open-Source Automatic Biomarker Measurement on Slit-Lamp Photography to Estimate Visual Acuity in Microbial Keratitis

Purpose

To assess clinical applicability of automatic image analysis in microbial keratitis (MK) by evaluating the relationship between biomarker measurements on slit-lamp photography (SLP) and best-corrected visual acuity (BCVA).

Methods

Seventy-six patients with MK with SLP images and same-day logarithm of the minimum angle of resolution (logMAR) BCVA were evaluated. MK biomarkers (stromal infiltrate, white blood cell infiltration, corneal edema, hypopyon, epithelial defect) were segmented manually by ophthalmologists and automatically by a novel, open-source, deep learning–based segmentation algorithm. Five measurements (presence, maximum width, total area, proportion of the corneal limbus area affected, centrality) were calculated. Correlations between the measurements and BCVA were calculated. An automatic regression model estimated BCVA from the measurements. Differences in performance between using manual and automatic measurements were evaluated using William's test (for correlation) and the paired-sample t-test (for absolute error).

Results

Measurements had high correlations of 0.86 (manual) and 0.84 (automatic) with true BCVA. Estimated BCVA had average (mean ± SD) absolute errors of 0.39 ± 0.27 logMAR (manual, median: 0.30) and 0.35 ± 0.28 logMAR (automatic, median: 0.30) and high correlations of 0.76 (manual) and 0.80 (automatic) with true BCVA. Differences between using manual and automatic measurements were not statistically significant for correlations of measurements with true BCVA (P = .66), absolute errors of estimated BCVA (P = .15), or correlations of estimated BCVA with true BCVA (P = .60).

Conclusions

The proposed algorithm measured MK biomarkers as accurately as ophthalmologists. Measurements were highly correlated with and estimative of visual acuity.

Translational Relevance

This study demonstrates the potential of developing fully automatic objective and standardized strategies to aid ophthalmologists in the clinical assessment of MK.

Relationship Between Flicker Modulation Sensitivity and Retinal Ganglion Cell Related Layer Thicknesses

Purpose

Early detection of structural changes in retinal ganglion cells (RGCs) and corresponding changes in visual function is important in early degenerative diseases of the retina, but the sensitivity of both measurements is limited by the inherent variability in healthy subjects. This study investigates the relationships between RGC-related layer thicknesses and foveal and parafoveal flicker modulation sensitivity (FMS) across photopic and mesopic light levels in healthy subjects.

Methods

Photopic and mesopic FMS was measured in 56 young adults, at the point of fixation and at an eccentricity of 5 degrees, in each of the four quadrants. Spectral-domain optical coherence tomography (SD-OCT) was used to measure retinal thicknesses. Relationships between foveal and parafoveal FMS and the retinal thickness in the corresponding region were examined after adjusting for confounding variables.

Results

Total macular and inner retinal layer (IRL) thicknesses in the parafoveal ring were significant predictors of photopic (P = 0.034) and mesopic (P = 0.034) parafoveal FMS, respectively. The superior peripapillary retinal nerve fiber layer (pRNFL) thickness was a contributing factor to the inferior parafoveal FMS (photopic: P = 0.006 and mesopic: P = 0.021) and the inferior pRNFL thickness was also a contributing factor to the superior parafoveal FMS (photopic: P < 0.001 and mesopic: P = 0.015).

Conclusions

The pRNFL thicknesses predict parafoveal FMS for both mesopic and photopic conditions in healthy eyes.

Translational Relevance

The measurement of rapid flicker sensitivity in the parafoveal retina together with the pRNFL thickness profiles measured before the onset of disease, may provide a more sensitive biomarker for detecting loss of sensitivity caused by the earliest neurodegenerative changes in the eyes.

Multimodal Machine Learning Using Visual Fields and Peripapillary Circular OCT Scans in Detection of Glaucomatous Optic Neuropathy

PURPOSE

To develop and validate a multimodal artificial intelligence algorithm, FusionNet, using the pattern deviation probability plots from visual field (VF) reports and circular peripapillary OCT scans to detect glaucomatous optic neuropathy (GON).

DESIGN

Cross-sectional study.

SUBJECTS

Two thousand four hundred sixty-three pairs of VF and OCT images from 1083 patients.

METHODS

FusionNet based on bimodal input of VF and OCT paired data was developed to detect GON. Visual field data were collected using the Humphrey Field Analyzer (HFA). OCT images were collected from 3 types of devices (DRI-OCT, Cirrus OCT, and Spectralis). Two thousand four hundred sixty-three pairs of VF and OCT images were divided into 4 datasets: 1567 for training (HFA and DRI-OCT), 441 for primary validation (HFA and DRI-OCT), 255 for the internal test (HFA and Cirrus OCT), and 200 for the external test set (HFA and Spectralis). GON was defined as retinal nerve fiber layer thinning with corresponding VF defects.

MAIN OUTCOME MEASURES

Diagnostic performance of FusionNet compared with that of VFNet (with VF data as input) and OCTNet (with OCT data as input).

RESULTS

FusionNet achieved an area under the receiver operating characteristic curve (AUC) of 0.950 (0.931-0.968) and outperformed VFNet (AUC, 0.868 [95% confidence interval (CI), 0.834-0.902]), OCTNet (AUC, 0.809 [95% CI, 0.768-0.850]), and 2 glaucomatologists (glaucomatologist 1: AUC, 0.882 [95% CI, 0.847-0.917]; glaucomatologist 2: AUC, 0.883 [95% CI, 0.849-0.918]) in the primary validation set. In the internal and external test sets, the performances of FusionNet were also superior to VFNet and OCTNet (FusionNet vs VFNet vs OCTNet: internal test set 0.917 vs 0.854 vs 0.811; external test set 0.873 vs 0.772 vs 0.785). No significant difference was found between the 2 glaucomatologists and FusionNet in the internal and external test sets, except for glaucomatologist 2 (AUC, 0.858 [95% CI, 0.805-0.912]) in the internal test set.

CONCLUSIONS

FusionNet, developed using paired VF and OCT data, demonstrated superior performance to both VFNet and OCTNet in detecting GON, suggesting that multimodal machine learning models are valuable in detecting GON.

Three-dimensional Neuroretinal Rim Thickness and Visual Fields in Glaucoma: A Broken-stick Model

Supplemental Digital Content is available in the text.

Precis:

In open-angle glaucoma, when neuroretinal rim tissue measured by volumetric optical coherence tomography (OCT) scans is below a third of the normal value, visual field (VF) damage becomes detectable.

Purpose:

To determine the amount of neuroretinal rim tissue thickness below which VF damage becomes detectable.

Methods:

In a retrospective cross-sectional study, 1 eye per subject (of 57 healthy and 100 open-angle glaucoma patients) at an academic institution had eye examinations, VF testing, spectral-domain OCT retinal nerve fiber layer (RNFL) thickness measurements, and optic nerve volumetric scans. Using custom algorithms, the minimum distance band (MDB) neuroretinal rim thickness was calculated from optic nerve scans. “Broken-stick” regression was performed for estimating both the MDB and RNFL thickness tipping-point thresholds, below which were associated with initial VF defects in the decibel scale. The slopes for the structure-function relationship above and below the thresholds were computed. Smoothing curves of the MDB and RNFL thickness covariates were evaluated to examine the consistency of the independently identified tipping-point pairs.

Results:

Plots of VF total deviation against MDB thickness revealed plateaus of VF total deviation unrelated to MDB thickness. Below the thresholds, VF total deviation decreased with MDB thickness, with the associated slopes significantly greater than those above the thresholds (P<0.014). Below 31% of global MDB thickness, and 36.8% and 43.6% of superior and inferior MDB thickness, VF damage becomes detectable. The MDB and RNFL tipping points were in good accordance with the correlation of the MDB and RNFL thickness covariates.

Conclusions:

When neuroretinal rim tissue, characterized by MDB thickness in OCT, is below a third of the normal value, VF damage in the decibel scale becomes detectable.

Computational Analysis of Clinical and Molecular Markers and New Theranostic Possibilities in Primary Open-Angle Glaucoma

Primary open-angle glaucoma (POAG) is a paramount cause of irreversible visual disability worldwide. We focus on identifying clinical and molecular facts that may help elucidating the pathogenic mechanisms of the disease. By using ophthalmological approaches (biomicroscopy, ocular fundus, optical coherence tomography, and perimetry) and experimental tests (enzyme-linked immunosorbent assay (ELISA), high performance liquid chromatography (HPLC), and Western blot/immunoblotting) directed to evaluate the oxidative stress, inflammation, apoptosis, and neurodegeneration processes, we gather information to build a network of data to perform a computational bioinformatics analysis. Our results showed strong interaction of the above players and its downstream effectors in POAG pathogenesis. In conclusion, specific risk factors were identified, and molecules involved in multiple pathways were found in relation to anterior and posterior eye segment glaucoma changes, pointing to new theranostic challenges for better managing POAG progression.

Estimating Global Visual Field Indices in Glaucoma by Combining Macula and Optic Disc OCT Scans Using 3-Dimensional Convolutional Neural Networks

PURPOSE

To evaluate the accuracy at which visual field global indices could be estimated from OCT scans of the retina using deep neural networks and to quantify the contributions to the estimates by the macula (MAC) and the optic nerve head (ONH).

DESIGN

Observational cohort study.

PARTICIPANTS

A total of 10 370 eyes from 109 healthy patients, 697 glaucoma suspects, and 872 patients with glaucoma over multiple visits (median = 3).

METHODS

Three-dimensional convolutional neural networks were trained to estimate global visual field indices derived from automated Humphrey perimetry (SITA 24-2) tests (Zeiss, Dublin, CA), using OCT scans centered on MAC, ONH, or both (MAC + ONH) as inputs.

MAIN OUTCOME MEASURES

Spearman's rank correlation coefficients, Pearson's correlation coefficient, and absolute errors calculated for 2 indices: visual field index (VFI) and mean deviation (MD).

RESULTS

The MAC + ONH achieved 0.76 Spearman's correlation coefficient and 0.87 Pearson's correlation for VFI and MD. Median absolute error was 2.7 for VFI and 1.57 decibels (dB) for MD. Separate MAC or ONH estimates were significantly less correlated and less accurate. Accuracy was dependent on the OCT signal strength and the stage of glaucoma severity.

CONCLUSIONS

The accuracy of global visual field indices estimate is improved by integrating information from MAC and ONH in advanced glaucoma, suggesting that structural changes of the 2 regions have different time courses in the disease severity spectrum.

Graph convolutional network based optic disc and cup segmentation on fundus images

Calculating the cup-to-disc ratio is one of the methods for glaucoma screening with other clinical features. In this paper, we propose a graph convolutional network (GCN) based method to implement the optic disc (OD) and optic cup (OC) segmentation task. We first present a multi-scale convolutional neural network (CNN) as the feature map extractor to generate feature map. The GCN takes the feature map concatenated with the graph nodes as the input for segmentation task. The experimental results on the REFUGE dataset show that the Jaccard index (Jacc) of the proposed method on OD and OC are 95.64% and 91.60%, respectively, while the Dice similarity coefficients (DSC) are 97.76% and 95.58%, respectively. The proposed method outperforms the state-of-the-art methods on the REFUGE leaderboard. We also evaluate the proposed method on the Drishthi-GS1 dataset. The results show that the proposed method outperforms the state-of-the-art methods.

BDNF improves axon transportation and rescues visual function in a rodent model of acute elevation of intraocular pressure

Optic neuropathies lead to blindness; the common pathology is the degeneration of axons of the retinal ganglion cells. In this study, we used a rat model of retinal ischemia-reperfusion and a one-time intravitreal brain-derived neurotrophic factor (BDNF) injection; then we examined axon transportation function, continuity, physical presence of axons in different part of the optic nerve, and the expression level of proteins involved in axon transportation. We found that in the disease model, axon transportation was the most severely affected, followed by axon continuity, then the number of axons in the distal and proximal optic nerve. BDNF treatment relieved all reductions and significantly restored function. The molecular changes were more minor, probably due to massive gliosis of the optic nerve, so interpretation of protein expression data should be done with some caution. The process in this acute model resembles a fast-forward of changes in the chronic model of glaucoma. Therefore, impairment in axon transportation appears to be a common early process underlying different optic neuropathies. This research on effective intervention can be used to develop interventions for all optic neuropathies targeting axon transportation.

Deep Learning Approaches Predict Glaucomatous Visual Field Damage from OCT Optic Nerve Head En Face Images and Retinal Nerve Fiber Layer Thickness Maps

PURPOSE

To develop and evaluate a deep learning system for differentiating between eyes with and without glaucomatous visual field damage (GVFD) and predicting the severity of GFVD from spectral domain OCT (SD OCT) optic nerve head images.

DESIGN

Evaluation of a diagnostic technology.

PARTICIPANTS

A total of 9765 visual field (VF) SD OCT pairs collected from 1194 participants with and without GVFD (1909 eyes).

METHODS

Deep learning models were trained to use SD OCT retinal nerve fiber layer (RNFL) thickness maps, RNFL en face images, and confocal scanning laser ophthalmoscopy (CSLO) images to identify eyes with GVFD and predict quantitative VF mean deviation (MD), pattern standard deviation (PSD), and mean VF sectoral pattern deviation (PD) from SD OCT data.

MAIN OUTCOME MEASURES

Deep learning models were compared with mean RNFL thickness for identifying GVFD using area under the curve (AUC), sensitivity, and specificity. For predicting MD, PSD, and mean sectoral PD, models were evaluated using R² and mean absolute error (MAE).

RESULTS

In the independent test dataset, the deep learning models based on RNFL en face images achieved an AUC of 0.88 for identifying eyes with GVFD and 0.82 for detecting mild GVFD significantly (P < 0.001) better than using mean RNFL thickness measurements (AUC = 0.82 and 0.73, respectively). Deep learning models outperformed standard RNFL thickness measurements in predicting all quantitative VF metrics. In predicting MD, deep learning models based on RNFL en face images achieved an R² of 0.70 and MAE of 2.5 decibels (dB) compared with 0.45 and 3.7 dB for RNFL thickness measurements. In predicting mean VF sectoral PD, deep learning models achieved high accuracy in the inferior nasal (R² = 0.60) and superior nasal (R² = 0.67) sectors, moderate accuracy in inferior (R² = 0.26) and superior (R² = 0.35) sectors, and lower accuracy in the central (R² = 0.15) and temporal (R² = 0.12) sectors.

CONCLUSIONS

Deep learning models had high accuracy in identifying eyes with GFVD and predicting the severity of functional loss from SD OCT images. Accurately predicting the severity of GFVD from SD OCT imaging can help clinicians more effectively individualize the frequency of VF testing to the individual patient.

Variable structure and function relationship of compressive optic neuropathy at the time of diagnosis

Purpose

To illustrate the structure–function relationship of compressive optic neuropathy (CON) at the time of diagnosis.

Patients and methods

Thirty-two eyes of newly diagnosed suprasellar CON and 60 healthy eyes were included in the study. The peripapillary retinal nerve fiber layer (RNFL) thickness and macular ganglion cell-inner plexiform layer (GCIPL) thickness were obtained using Cirrus spectral domain optical coherence tomography (SD-OCT). CON eyes were stratified based on the similar degree and pattern of both RNFL and GCIPL.

Results

From 32 eyes of newly diagnosed suprasellar CON eyes, 27 eyes had a predominantly nasal hemiretina thinning of macular GCIPL, 4 eyes showed a generalized macular thinning, and 1 eye showed a predominantly superior macular thinning. The corresponding temporal peripapillary RNFL thinning with nasal hemiretina GCIPL thinning were inconsistently manifested. Structure–function analysis of stratified CON eyes with similar thinning profiles showed that a range rather than a fixed value of visual field loss based on mean deviation (MD) index was associated to each thinning profile. The maximal limit of visual field loss range was ubiquitously nonrestricted to any structural thinning profile. While the minimal limit of the associated MD range was gradually reduced from 0 to about −16.0 dB, the nasal hemiretina macular GCIPL thinning was the only manifestation and decreased from 75 to 45 µm. However, the different degrees of temporal hemiretina macular GCIPL and superior–inferior peripapillary RNFL thinning were only seen in 10 of 32 eyes of which their nasal hemiretina GCIPL and temporal RNFL thinning had reached significant thinning. Interestingly when present, the minimal limit of associated MD range continued to decrease from −16.0 to −32.0 dB.

Conclusion

CON eyes can present with variable structure and function relationship at the time of diagnosis. Using structural parameters at the time of diagnosis to predict the prognosis should be used with caution.

Point your SmartPhone at the code above. If you have a QR code reader the video abstract will appear. Or use:

https://youtu.be/pp-J8F7Rw2s

Effect of Ocular Hypertension on D-β-Aspartic Acid-Containing Proteins in the Retinas of Rats

Purpose

To investigate the effect of ocular hypertension-induced isomerization of aspartic acid in retinal proteins.

Methods

Adult Wistar rats with ocular hypertension were used as an experimental model. D-β-aspartic acid-containing proteins were isolated by SDS-PAGE and western blot with an anti-D-β-aspartic acid antibody and identified by liquid chromatography-mass spectrometry analysis. The concentration of ATP was measured by ELISA.

Results

D-β-aspartic acid was expressed in a protein band at around 44.5 kDa at much higher quantities in the retinas of rats with ocular hypertension than in those of normotensive rats. The 44.5 kDa protein band was mainly composed of α-enolase, S-arrestin, and ATP synthase subunits α and β, in both the ocular hypertensive and normotensive retinas. Moreover, increasing intraocular pressure was correlated with increasing ATP concentrations in the retinas of rats.

Conclusion

Ocular hypertension affected the expression of proteins containing D-β-aspartic acid, including ATP synthase subunits, and up-regulation of ATP in the retinas of rats.

Macular Inner Retinal Layer Thickness in Relation to Photopic and Mesopic Contrast Sensitivity in Healthy Young and Older Subjects

Purpose

To examine relationships between the thicknesses of ganglion cell (GC)-related macular layers and central photopic or mesopic contrast sensitivity (CS) in healthy eyes.

Methods

Measurements were made in 38 young and 38 older healthy individuals. Total, inner, and outer retinal layer (IRL) thicknesses were measured in the macula region through spectral-domain optical coherence tomography (SD-OCT) across three subfields, or rings, centered at the fovea: central foveal, pericentral, and peripheral. Ganglion cell complex and circumpapillary retinal nerve fiber layer thicknesses were also measured. Low-spatial-frequency CS for gratings presented at the central 10° visual field were measured through computerized psychophysical tests under photopic and mesopic conditions. Relationships were examined by uni- and multivariate regression analysis.

Results

Peripheral IRL thickness emerged as the only independent predictor of photopic CS (P = 0.001) in the young group and of photopic (P = 0.026) and mesopic CS (P = 0.001) in the older group. The slopes of regression lines used to predict CS from peripheral IRL thickness were significantly different for pair-wise comparisons of both photopic CS and age group (P = 0.0001) and mesopic CS (P = 0.0001) and age group. These models explained 37% of the variability in photopic CS and 36% of the variability in mesopic CS.

Conclusions

Macular IRL thinning likely due to GC loss was related to reduced photopic and mesopic CS in older healthy eyes. In contrast, in the young eyes, a thicker macular IRL, possibly indicating transient gliosis, was associated with reduced CS.