Structural and Functional Evaluations for the Early Detection of Glaucoma

Persistence with medical glaucoma therapy in newly diagnosed patients

Background

Monotherapy, age, and side effects are significant risk factors for the discontinuation of antiglaucoma therapy. Long-term therapy persistence is crucial for slowing disease progression and preventing irreversible blindness. Therefore, it is essential to identify patients at higher risk of discontinuation. In this study, we aimed to evaluate the real-world persistence of antiglaucoma therapy in patients diagnosed with glaucoma in the primary healthcare units of the Lisbon and Tagus Valley regions.

Methods

We conducted a retrospective longitudinal study by collecting data from the prescription records of new antiglaucoma drug users diagnosed with glaucoma between 2012 and 2013 in the Primary Health Care Units of the Lisbon and Tagus Valley Region. These patients were followed over 3 years. Therapy persistence was measured as the proportion of patients remaining on any antiglaucoma drug, regardless of any modifications or switching of drugs over time. Persistence was assessed at three time points: the end of the first, second, and third years of the observation period.

Results

A total of 2138 patients treated using new antiglaucoma drugs (867 [40.6%] male patients; 1271 [59.4%] female patients) were included in the study. Over the observation period, the overall persistence rate decreased from 91.9% (n = 1965) in the first year to 67.3% (n = 1439) in the third year. Older patients (≥ 65 years) showed higher persistence rates, although there was a decrease over the 3-year follow-up period (from 1481 [92.7%] to 1124 [70.4%]). Additionally, participants initially treated with monotherapy showed higher persistence rates, ranging from 92.4% (n = 1186) in the first year to 70.2% (n = 901) in the third year.

Conclusions

The findings highlight the importance of patient follow-up over time, as almost one in three new antiglaucoma therapy users completely discontinued treatment, potentially risking disease progression. This could be mitigated with proper use of these drugs. Further studies should utilize recent health information systems to explore the impact of medication adherence and persistence on the functional and structural outcomes in patients with glaucoma.

Association of dietary intake of B vitamins with glaucoma

This cross-sectional study investigated the association between glaucoma and B vitamin dietary intake. A total of 5025 enrolled individuals participated in self-reported glaucoma questionnaire and 3264 participated in International Society Geographical and Epidemiological Ophthalmology (ISGEO) criteria. In self-reported glaucoma, the risk of having self-reported glaucoma was lower in the third quartile of vitamin B1 intake (odds ratio [odds ratio [OR] 0.63, 95% confidence interval [CI] 0.40-0.97), and P trend (P trend = 0.004) for vitamin B12 was significant; in males, the third quartile of vitamin B1 intake (OR 0.44, 95% CI 0.24-0.83) and the fourth quartile of vitamin B2 intake (OR 0.39, 95% CI 0.17-0.89) were associated with a lower risk. In glaucoma based on ISGEO criteria, the increase of niacin intake (OR 0.94, 95% CI 0.89-0.99) was negatively associated with the odds of self-reported glaucoma. After sex-stratified analysis, the third quartile of vitamin B6 intake (OR 0.21, 95% CI 0.08-0.60) in males were associated with reduced odds of glaucoma. The restricted cubic spline analysis revealed a nonlinear association of vitamin B2 (p for nonlinearity = 0.04) and B9 (p for nonlinearity = 0.024) intake with glaucoma diagnosed by ISGEO criteria in females.

Artificial intelligence in glaucoma: opportunities, challenges, and future directions

Artificial intelligence (AI) has shown excellent diagnostic performance in detecting various complex problems related to many areas of healthcare including ophthalmology. AI diagnostic systems developed from fundus images have become state-of-the-art tools in diagnosing retinal conditions and glaucoma as well as other ocular diseases. However, designing and implementing AI models using large imaging data is challenging. In this study, we review different machine learning (ML) and deep learning (DL) techniques applied to multiple modalities of retinal data, such as fundus images and visual fields for glaucoma detection, progression assessment, staging and so on. We summarize findings and provide several taxonomies to help the reader understand the evolution of conventional and emerging AI models in glaucoma. We discuss opportunities and challenges facing AI application in glaucoma and highlight some key themes from the existing literature that may help to explore future studies. Our goal in this systematic review is to help readers and researchers to understand critical aspects of AI related to glaucoma as well as determine the necessary steps and requirements for the successful development of AI models in glaucoma.

Predicting glaucoma progression using deep learning framework guided by generative algorithm

Glaucoma is a slowly progressing optic neuropathy that may eventually lead to blindness. To help patients receive customized treatment, predicting how quickly the disease will progress is important. Structural assessment using optical coherence tomography (OCT) can be used to visualize glaucomatous optic nerve and retinal damage, while functional visual field (VF) tests can be used to measure the extent of vision loss. However, VF testing is patient-dependent and highly inconsistent, making it difficult to track glaucoma progression. In this work, we developed a multimodal deep learning model comprising a convolutional neural network (CNN) and a long short-term memory (LSTM) network, for glaucoma progression prediction. We used OCT images, VF values, demographic and clinical data of 86 glaucoma patients with five visits over 12 months. The proposed method was used to predict VF changes 12 months after the first visit by combining past multimodal inputs with synthesized future images generated using generative adversarial network (GAN). The patients were classified into two classes based on their VF mean deviation (MD) decline: slow progressors (< 3 dB) and fast progressors (> 3 dB). We showed that our generative model-based novel approach can achieve the best AUC of 0.83 for predicting the progression 6 months earlier. Further, the use of synthetic future images enabled the model to accurately predict the vision loss even earlier (9 months earlier) with an AUC of 0.81, compared to using only structural (AUC = 0.68) or only functional measures (AUC = 0.72). This study provides valuable insights into the potential of using synthetic follow-up OCT images for early detection of glaucoma progression.

Normative Percentiles of Retinal Nerve Fiber Layer Thickness and Glaucomatous Visual Field Loss

Purpose

Circumpapillary retinal nerve fiber layer thickness (RNFLT) measurement aids in the clinical diagnosis of glaucoma. Spectral domain optical coherence tomography (SD-OCT) machines measure RNFLT and provide normative color-coded plots. In this retrospective study, we investigate whether normative percentiles of RNFLT (pRNFLT) from Spectralis SD-OCT improve prediction of glaucomatous visual field loss over raw RNFLT.

Methods

A longitudinal database containing OCT scans and visual fields from Massachusetts Eye & Ear glaucoma clinic patients was generated. Reliable OCT-visual field pairs were selected. Spectralis OCT normative distributions were extracted from machine printouts. Supervised machine learning models compared predictive performance between pRNFLT and raw RNFLT inputs. Regional structure-function associations were assessed with univariate regression to predict mean deviation (MD). Multivariable classification predicted MD, pattern standard deviation, MD change per year, and glaucoma hemifield test.

Results

There were 3016 OCT-visual field pairs that met the reliability criteria. Spectralis norms were found to be independent of age, sex, and ocular magnification. Regional analysis showed significant decrease in R2 from pRNFLT models compared to raw RNFLT models in inferotemporal sectors, across multiple regressors. In multivariable classification, there were no significant improvements in area under the curve of receiver operating characteristic curve (ROC-AUC) score with pRNFLT models compared to raw RNFLT models.

Conclusions

Our results challenge the assumption that normative percentiles from OCT machines improve prediction of glaucomatous visual field loss. Raw RNFLT alone shows strong prediction, with no models presenting improvement by the manufacturer norms. This may result from insufficient patient stratification in tested norms.

Translational Relevance

Understanding correlation of normative databases to visual function may improve clinical interpretation of OCT data.

Association of Glaucoma with Poor Quality of Sleep in an Ethiopian Glaucoma Population – A Comparative Cross-Sectional Study

Background

Glaucoma is a group of ocular disorders characterized by progressive optic nerve damage resulting in irreversible visual field defects. Poor quality of sleep in glaucoma patients could be explained by the reduction of the light input to the circadian system as a result of damage to photosensitive retinal ganglion cells in the retina. Information is limited on the association of poor quality of sleep with glaucoma in general and the Ethiopian glaucoma population in particular.

Objective

The study aimed to explore the association between poor quality of sleep and glaucoma at a Tertiary Eye Care Center in Ethiopia.

Methods

An institutional-based comparative cross-sectional study was conducted among 200 glaucoma and 201 non-glaucoma participants recruited by systematic random sampling. Each group was administered with a Pittsburgh Sleep Quality Index (PSQI) questionnaire. Stata-14 was employed for data analysis; an independent t-test was used to show the statistical difference in the global mean PSQI score for the two groups. A binary logistic regression model was applied to identify factors associated with poor quality of sleep. Statistical significance was declared at a 95% confidence interval and a p-value of <0.05.

Results

The prevalence of poor quality of sleep was 82.5% among the glaucoma population, which statistically differed (p<0.001) from the non-glaucomatous population (55.7%). Poor quality of sleep in glaucoma was associated with older age (adjusted odds ratio (AOR)=4.4, 95% confidence interval (CI): 1.5–5.4), depression (AOR=2.9, 95% CI: 1.1–7.3), visual impairment (AOR=3.9, 95% CI: 1.3–12.3) and severe glaucoma (AOR=2.5, 95% CI: 1.1–5.9).

Conclusion and Recommendation

Poor quality of sleep was significantly higher in the glaucoma population compared to their non-glaucoma control. It was associated with older age, depression, visual impairment and advanced glaucoma. Incorporating psychiatric counseling into the existing glaucoma follow-up was recommended.

Diagnostic Accuracy of Artificial Intelligence in Glaucoma Screening and Clinical Practice

PURPOSE

Artificial intelligence (AI) has been shown as a diagnostic tool for glaucoma detection through imaging modalities. However, these tools are yet to be deployed into clinical practice. This meta-analysis determined overall AI performance for glaucoma diagnosis and identified potential factors affecting their implementation.

METHODS

We searched databases (Embase, Medline, Web of Science, and Scopus) for studies that developed or investigated the use of AI for glaucoma detection using fundus and optical coherence tomography (OCT) images. A bivariate random-effects model was used to determine the summary estimates for diagnostic outcomes. The Preferred Reporting Items for Systematic Reviews and Meta-Analysis of Diagnostic Test Accuracy (PRISMA-DTA) extension was followed, and the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool was used for bias and applicability assessment.

RESULTS

Seventy-nine articles met inclusion criteria, with a subset of 66 containing adequate data for quantitative analysis. The pooled area under receiver operating characteristic curve across all studies for glaucoma detection was 96.3%, with a sensitivity of 92.0% (95% confidence interval: 89.0-94.0) and specificity of 94.0% (95% confidence interval: 92.0-95.0). The pooled area under receiver operating characteristic curve on fundus and OCT images was 96.2% and 96.0%, respectively. Mixed data set and external data validation had unsatisfactory diagnostic outcomes.

CONCLUSION

Although AI has the potential to revolutionize glaucoma care, this meta-analysis highlights that before such algorithms can be implemented into clinical care, a number of issues need to be addressed. With substantial heterogeneity across studies, many factors were found to affect the diagnostic performance. We recommend implementing a standard diagnostic protocol for grading, implementing external data validation, and analysis across different ethnicity groups.

Correlation Between Optical Coherence Tomography and Photopic Negative Response of Flash Electroretinography in Ganglion Cell Complex Assessment in Glaucoma Patients

Purpose

To investigate the correlation between the photopic negative response (PhNR) of the light-adapted flash electroretinography (ERG) and measurements of standard automated perimetry (SAP) and optical coherence tomography (OCT) in assessment of retinal ganglion cells’ (RGCs) affection in glaucoma.

Patients and Methods

A cross-sectional study included 40 eyes of glaucoma patients and 40 eyes of age- and gender-matched normal subjects. Participants underwent a complete ophthalmologic assessment, SAP, OCT, and light-adapted flash ERG using the extended PhNR protocol of the International Society for Clinical Electrophysiology of Vision (ISCEV). Glaucomatous eyes were divided into 3 subgroups: mild (n = 15), moderate (n = 11) and severe glaucoma (n = 14) according to the mean deviation (MD) of SAP. Measurements of SAP, OCT and ERG parameters were analyzed, and correlations between PhNR measurements and other study measurements were evaluated.

Results

PhNR amplitudes and PhNR/b-wave ratios were significantly reduced in glaucoma cases compared to healthy controls, and they showed a significant and progressive decline across the three glaucoma subgroups (P < 0.05). An exception to this is PT (b-wave peak to PhNR trough) PhNR amplitude where its reduction was statistically non-significant when comparing between controls and mild glaucoma cases (P = 0.178), and between moderate and severe glaucoma cases (P = 0.714). PhNR amplitudes and PhNR/b-wave ratios correlated significantly with SAP and OCT parameters (P < 0.05).

Conclusion

PhNR correlates well with SAP and OCT parameters in glaucoma assessment. PhNR could be a valuable supplementary tool for objective assessment of the RGCs’ function in glaucoma.

Association between Daily Niacin Intake and Glaucoma: National Health and Nutrition Examination Survey

Background and Aims: To determine the relationship between dietary intake of niacin and glaucoma using the data from the 2005 to 2008 National Health and Nutrition Examination Survey (NHANES). Methods: Subjects aged 40 years and older who participated in the dietary intake interview and vision health questionnaire of NHANES were included in the study. Glaucoma diagnosis by self-report was utilized. Additionally, glaucoma diagnosis by fundus imaging and International Society Geographical and Epidemiological Ophthalmology (ISGEO) criteria was used in subjects with available qualified retinal imaging. Survey logistic regression analyses were used to examine the association between daily niacin consumption and glaucoma. Results: A total of 5768 participants were included in the study. There was a significant decrease in the crude odds of self-reported glaucoma in the third (OR 0.57, 95% Cl 0.43–0.76; p < 0.001) and fourth (OR 0.57, 95% Cl 0.37–0.90; p = 0.018) quartiles of daily niacin consumption, which equated to 21.01 to 28.22 mg/day and greater than 28.22 mg/day, respectively. A similar trend was found using fundus imaging of subjects with niacin intake in the third (OR 0.42, 95% Cl 0.25–0.72; p = 0.002) and fourth (OR 0.36, 95% Cl 0.20–0.67; p = 0.002) quartiles. After adjusting for covariates, the odds of glaucoma based on fundus imaging remained significantly lower for niacin intake in the third (OR 0.49, 95% Cl 0.28–0.87; p = 0.016) and fourth (OR 0.48, 95% Cl 0.26–0.89; p = 0.022) quartile levels. Using ISGEO criteria, there was no significant association between glaucoma and daily niacin consumption. Conclusions: Greater niacin intake may be associated with a lower chance of developing glaucoma.

Multifocal visual evoked potential for evaluation of open-angle glaucoma

Background

To correlate multifocal visual evoked potential (mfVEP) findings with static automated perimetry (SAP) and spectral-domain optical coherence tomography (SD-OCT) in eyes with primary open- angle glaucoma (POAG).

Methods

This cross-sectional study included a consecutive sample of 40 eyes of 40 patients with POAG. The participants underwent a complete ophthalmologic assessment, axial length (AL) measurement, and assessments with SAP, SD-OCT, and mfVEP.

Results

POAG cases were aged 49.70 (14.16) years (mean [SD]) and most were females (n = 24, 60%). For eyes of patients with POAG, the mfVEP upper-ring signal-to-noise ratio (SNR) showed a significant negative correlation with best-corrected logMAR visual acuity (r = - 0.33; P = 0.038), and a significant positive correlation with the superior hemifield of the visual field (VF) and the inferior-quadrant retinal nerve fiber layer (RNFL) thickness (r = + 0.34; P = 0.030; r = + 0.51; P < 0.001, respectively). Similarly, the mfVEP lower-ring SNR showed a significant negative correlation with best-corrected logMAR visual acuity (r = - 0.36; P = 0.024) and a significant positive correlation with the inferior hemifield of the VF and superior quadrant RNFL thickness (r = + 0.55; P < 0.001 and r = + 0.70; P < 0.001, respectively).

Conclusions

mfVEP is a promising tool for objective assessment of the VF in patients with POAG, as it is positively correlated with the VF and OCT RNFL thickness. Future longitudinal studies with a larger sample size and a specific glaucoma subtype, along with multiple follow-up evaluations, are warranted to confirm our preliminary results.

Change of Retinal Vessels in Different Sectors of the Parapapillary Area in Primary Open-Angle Glaucoma

Purpose: To investigate the changes in the retinal vessels (RVs) in different sectors in patients with primary open-angle glaucoma (POAG), and their possible correlations with retinal nerve fiber layer thickness (RNFLT) and visual-field defects in the temporal parapapillary region.

Methods: The RV diameters, RNFLTs, and visual-field parameters were measured. The temporal parapapillary region was divided into the temporal (T, 315°-45°), temporal superior (TS, 45°-90°), and temporal inferior sectors (TI, 270°-315°). The changes in the RV diameters in each sector were determined, and their relationships with RNFLT, the mean deviation (MD), and visual field sensitivity (VFS) were examined.

Results: Fifty POAG patients (50 eyes) and 50 healthy subjects (50 eyes) were included. Compared with the healthy subjects, the POAG group had a significantly smaller accumulated parapapillary RV diameter (P < 0.001), which was positively correlated with the MD and RNFLT. When the different temporal sectors were examined, the accumulated RV diameters were significantly smaller in the POAG group than in the healthy controls in the TI and T sectors, but not in the TS sector. The accumulated diameters in the TI and T sectors were correlated with the corresponding RNFLTs (all P < 0.05), but only the accumulated diameter in the TI sector was correlated with the VFS.

Conclusions: In POAG, the changes in the RVs differed between different temporal sectors, with the most prominent changes occurring in the TI and T sectors.

High interpretable machine learning classifier for early glaucoma diagnosis

AIM

To develop a classifier for differentiating between healthy and early stage glaucoma eyes based on peripapillary retinal nerve fiber layer (RNFL) thicknesses measured with optical coherence tomography (OCT), using machine learning algorithms with a high interpretability.

METHODS

Ninety patients with early glaucoma and 85 healthy eyes were included. Early glaucoma eyes showed a visual field (VF) defect with mean deviation >-6.00 dB and characteristic glaucomatous morphology. RNFL thickness in every quadrant, clock-hour and average thickness were used to feed machine learning algorithms. Cluster analysis was conducted to detect and exclude outliers. Tree gradient boosting algorithms were used to calculate the importance of parameters on the classifier and to check the relation between their values and its impact on the classifier. Parameters with the lowest importance were excluded and a weighted decision tree analysis was applied to obtain an interpretable classifier. Area under the ROC curve (AUC), accuracy and generalization ability of the model were estimated using cross validation techniques.

RESULTS

Average and 7 clock-hour RNFL thicknesses were the parameters with the highest importance. Correlation between parameter values and impact on classification displayed a stepped pattern for average thickness. Decision tree model revealed that average thickness lower than 82 µm was a high predictor for early glaucoma. Model scores had AUC of 0.953 (95%CI: 0.903-0998), with an accuracy of 89%.

CONCLUSION

Gradient boosting methods provide accurate and highly interpretable classifiers to discriminate between early glaucoma and healthy eyes. Average and 7-hour RNFL thicknesses have the best discriminant power.

Attention-Guided 3D-CNN Framework for Glaucoma Detection and Structural-Functional Association Using Volumetric Images

The direct analysis of 3D Optical Coherence Tomography (OCT) volumes enables deep learning models (DL) to learn spatial structural information and discover new bio-markers that are relevant to glaucoma. Downsampling 3D input volumes is the state-of-art solution to accommodate for the limited number of training volumes as well as the available computing resources. However, this limits the network's ability to learn from small retinal structures in OCT volumes. In this paper, our goal is to improve the performance by providing guidance to DL model during training in order to learn from finer ocular structures in 3D OCT volumes. Therefore, we propose an end-to-end attention guided 3D DL model for glaucoma detection and estimating visual function from retinal structures. The model consists of three pathways with the same network architecture but different inputs. One input is the original 3D-OCT cube and the other two are computed during training guided by the 3D gradient class activation heatmaps. Each pathway outputs the class-label and the whole model is trained concurrently to minimize the sum of losses from three pathways. The final output is obtained by fusing the predictions of the three pathways. Also, to explore the robustness and generalizability of the proposed model, we apply the model on a classification task for glaucoma detection as well as a regression task to estimate visual field index (VFI) (a value between 0 and 100). A 5-fold cross-validation with a total of 3782 and 10,370 OCT scans is used to train and evaluate the classification and regression models, respectively. The glaucoma detection model achieved an area under the curve (AUC) of 93.8% compared with 86.8% for a baseline model without the attention-guided component. The model also outperformed six different feature based machine learning approaches that use scanner computed measurements for training. Further, we also assessed the contribution of different retinal layers that are relevant to glaucoma. The VFI estimation model achieved a Pearson correlation and median absolute error of 0.75 and 3.6%, respectively, for a test set of size 3100 cubes.

A feature agnostic approach for glaucoma detection in OCT volumes

Optical coherence tomography (OCT) based measurements of retinal layer thickness, such as the retinal nerve fibre layer (RNFL) and the ganglion cell with inner plexiform layer (GCIPL) are commonly employed for the diagnosis and monitoring of glaucoma. Previously, machine learning techniques have relied on segmentation-based imaging features such as the peripapillary RNFL thickness and the cup-to-disc ratio. Here, we propose a deep learning technique that classifies eyes as healthy or glaucomatous directly from raw, unsegmented OCT volumes of the optic nerve head (ONH) using a 3D Convolutional Neural Network (CNN). We compared the accuracy of this technique with various feature-based machine learning algorithms and demonstrated the superiority of the proposed deep learning based method. Logistic regression was found to be the best performing classical machine learning technique with an AUC of 0.89. In direct comparison, the deep learning approach achieved a substantially higher AUC of 0.94 with the additional advantage of providing insight into which regions of an OCT volume are important for glaucoma detection. Computing Class Activation Maps (CAM), we found that the CNN identified neuroretinal rim and optic disc cupping as well as the lamina cribrosa (LC) and its surrounding areas as the regions significantly associated with the glaucoma classification. These regions anatomically correspond to the well established and commonly used clinical markers for glaucoma diagnosis such as increased cup volume, cup diameter, and neuroretinal rim thinning at the superior and inferior segments.