Artificial intelligence and deep learning in glaucoma: Current state and future prospects

Automatic retinal image analysis methods using colour fundus images for screening glaucomatous optic neuropathy

OBJECTIVES

Train an automatic retinal image analysis (ARIA) method to screen glaucomatous optic neuropathy (GON) on non-mydriatic retinal images labelled with the additional results of optical coherence tomography (OCT) and assess different models for the GON classification.

METHODS

All the images were obtained from the hospital for training and 10-fold cross-validation. Two methods were used to improve the classification performance: (1) using images labelled with the additional results of OCT as the reference standard and (2) generating models using retinal features from the entire images, the region of interest (ROI) of the optic disc, and the ROI of the macula, and the combination of all the features.

RESULTS

Overall, we collected 1338 images with paired OCT scans. In 10-fold validation, ARIA achieved sensitivities of 92.2 %, 92.7% and 85.7%, specificities of 88.8%, 86.7% and 80.2% and accuracies of 90.6%, 89.9% and 83.1% using the retinal features from the entire images, the ROI of the optic disc and the ROI of the macula, respectively. We found the model combining all the features has the best classification performance and obtained a sensitivity of 92.5%, a specificity of 92.1% and an accuracy of 92.4%, which is significantly different from other models (p<0.001).

CONCLUSION

We used two methods to improve the classification performance and found the best model to detect glaucoma on colour fundus retinal images. It can become a cost-effective and relatively more accurate glaucoma screening tool than conventional methods.

The AI revolution in glaucoma: Bridging challenges with opportunities

Recent advancements in artificial intelligence (AI) herald transformative potentials for reshaping glaucoma clinical management, improving screening efficacy, sharpening diagnosis precision, and refining the detection of disease progression. However, incorporating AI into healthcare usages faces significant hurdles in terms of developing algorithms and putting them into practice. When creating algorithms, issues arise due to the intensive effort required to label data, inconsistent diagnostic standards, and a lack of thorough testing, which often limits the algorithms' widespread applicability. Additionally, the "black box" nature of AI algorithms may cause doctors to be wary or skeptical. When it comes to using these tools, challenges include dealing with lower-quality images in real situations and the systems' limited ability to work well with diverse ethnic groups and different diagnostic equipment. Looking ahead, new developments aim to protect data privacy through federated learning paradigms, improving algorithm generalizability by diversifying input data modalities, and augmenting datasets with synthetic imagery. The integration of smartphones appears promising for using AI algorithms in both clinical and non-clinical settings. Furthermore, bringing in large language models (LLMs) to act as interactive tool in medicine may signify a significant change in how healthcare will be delivered in the future. By navigating through these challenges and leveraging on these as opportunities, the field of glaucoma AI will not only have improved algorithmic accuracy and optimized data integration but also a paradigmatic shift towards enhanced clinical acceptance and a transformative improvement in glaucoma care.

Deep-learning-based prediction of glaucoma conversion in normotensive glaucoma suspects

BACKGROUND/AIMS

To assess the performance of deep-learning (DL) models for prediction of conversion to normal-tension glaucoma (NTG) in normotensive glaucoma suspect (GS) patients.

METHODS

Datasets of 12 458 GS eyes were reviewed. Two hundred and ten eyes (105 eyes showing NTG conversion and 105 without conversion), followed up for a minimum of 7 years during which intraocular pressure (IOP) was lower than 21 mm Hg, were included. The features of two fundus images (optic disc photography and red-free retinal nerve fibre layer (RNFL) photography) were extracted by convolutional auto encoder. The extracted features as well as 15 clinical features including age, sex, IOP, spherical equivalent, central corneal thickness, axial length, average circumpapillary RNFL thickness, systolic/diastolic blood pressure and body mass index were used to predict NTG conversion. Prediction was performed using three machine-learning classifiers (ie, XGBoost, Random Forest, Gradient Boosting) with different feature combinations.

RESULTS

All three algorithms achieved high diagnostic accuracy for NTG conversion prediction. The AUCs ranged from 0.987 (95% CI 0.978 to 1.000; Random Forest trained with both fundus images and clinical features) and 0.994 (95% CI 0.984 to 1.000; XGBoost trained with both fundus images and clinical features). XGBoost showed the best prediction performance for time to NTG conversion (mean squared error, 2.24). The top three important clinical features for time-to-conversion prediction were baseline IOP, diastolic blood pressure and average circumpapillary RNFL thickness.

CONCLUSION

DL models, trained with both fundus images and clinical data, showed the potential to predict whether and when normotensive GS patients will show conversion to NTG.

Learning the cellular activity representation based on gene regulatory networks for prediction of tumor response to drugs

Predicting the response of tumor cells to anti-tumor drugs is critical to realizing cancer precision medicine. Currently, most existing methods ignore the regulatory relationships between genes and thus have unsatisfactory predictive performance. In this paper, we propose to predict anti-tumor drug efficacy via learning the activity representation of tumor cells based on a priori knowledge of gene regulation networks (GRNs). Specifically, the method simulates the cellular biosystem by synthesizing a cell-gene activity network and then infers a new low-dimensional activity representation for tumor cells from the raw high-dimensional expression profile. The simulated cell-gene network mainly comprises known gene regulatory networks collected from multiple resources and fuses tumor cells by linking them to hotspot genes that are over- or under-expressed in them. The resulting activity representation could not only reflect the shallow expression profile (hotspot genes) but also mines in-depth information of gene regulation activity in tumor cells before treatment. Finally, we build deep learning models on the activity representation for predicting drug efficacy in tumor cells. Experimental results on the benchmark GDSC dataset demonstrate the superior performance of the proposed method over SOTA methods with the highest AUC of 0.954 in the efficacy label prediction and the best R2 of 0.834 in the regression of half maximal inhibitory concentration (IC50) values, suggesting the potential value of the proposed method in practice.

A deep learning approach to investigate the filtration bleb functionality after glaucoma surgery: a preliminary study

PURPOSE

To distinguish functioning from failed filtration blebs (FBs) implementing a deep learning (DL) model on slit-lamp images.

METHODS

Retrospective, cross-sectional, multicenter study for development and validation of an artificial intelligence classification algorithm. The dataset consisted of 119 post-trabeculectomy FB images of whom we were aware of the surgical outcome. The ground truth labels were annotated and images splitted into three outcome classes: complete (C) or qualified success (Q), and failure (F). Images were prepared implementing various data cleaning and data transformations techniques. A set of DL models were trained using different ResNet architectures as the backbone. Transfer and ensemble learning were then applied to obtain a final combined model. Accuracy, sensitivity, specificity, area under the ROC curve, and area under the precision-recall curve were calculated to evaluate the final model. Kappa coefficient and P value on the accuracy measure were used to prove the statistical significance level.

RESULTS

The DL approach reached good results in unraveling FB functionality. Overall, the model accuracy reached a score of 74%, with a sensitivity of 74% and a specificity of 87%. The area under the ROC curve was 0.8, whereas the area under the precision-recall curve was 0.74. The P value was equal to 0.00307, and the Kappa coefficient was 0.58.

CONCLUSIONS

All considered metrics supported that the final DL model was able to discriminate functioning from failed FBs, with good accuracy. This approach could support clinicians in the patients' management after glaucoma surgery in absence of adjunctive clinical data.

Endpoints for clinical trials in ophthalmology

With the identification of novel targets, the number of interventional clinical trials in ophthalmology has increased. Visual acuity has for a long time been considered the gold standard endpoint for clinical trials, but in the recent years it became evident that other endpoints are required for many indications including geographic atrophy and inherited retinal disease. In glaucoma the currently available drugs were approved based on their IOP lowering capacity. Some recent findings do, however, indicate that at the same level of IOP reduction, not all drugs have the same effect on visual field progression. For neuroprotection trials in glaucoma, novel surrogate endpoints are required, which may either include functional or structural parameters or a combination of both. A number of potential surrogate endpoints for ophthalmology clinical trials have been identified, but their validation is complicated and requires solid scientific evidence. In this article we summarize candidates for clinical endpoints in ophthalmology with a focus on retinal disease and glaucoma. Functional and structural biomarkers, as well as quality of life measures are discussed, and their potential to serve as endpoints in pivotal trials is critically evaluated.

Multimodal Deep Learning Classifier for Primary Open Angle Glaucoma Diagnosis Using Wide-Field Optic Nerve Head Cube Scans in Eyes With and Without High Myopia

PRCIS

An optical coherence tomography (OCT)-based multimodal deep learning (DL) classification model, including texture information, is introduced that outperforms single-modal models and multimodal models without texture information for glaucoma diagnosis in eyes with and without high myopia.

BACKGROUND/AIMS

To evaluate the diagnostic accuracy of a multimodal DL classifier using wide OCT optic nerve head cube scans in eyes with and without axial high myopia.

MATERIALS AND METHODS

Three hundred seventy-one primary open angle glaucoma (POAG) eyes and 86 healthy eyes, all without axial high myopia [axial length (AL) ≤ 26 mm] and 92 POAG eyes and 44 healthy eyes, all with axial high myopia (AL > 26 mm) were included. The multimodal DL classifier combined features of 3 individual VGG-16 models: (1) texture-based en face image, (2) retinal nerve fiber layer (RNFL) thickness map image, and (3) confocal scanning laser ophthalmoscope (cSLO) image. Age, AL, and disc area adjusted area under the receiver operating curves were used to compare model accuracy.

RESULTS

Adjusted area under the receiver operating curve for the multimodal DL model was 0.91 (95% CI = 0.87, 0.95). This value was significantly higher than the values of individual models [0.83 (0.79, 0.86) for texture-based en face image; 0.84 (0.81, 0.87) for RNFL thickness map; and 0.68 (0.61, 0.74) for cSLO image; all P ≤ 0.05]. Using only highly myopic eyes, the multimodal DL model showed significantly higher diagnostic accuracy [0.89 (0.86, 0.92)] compared with texture en face image [0.83 (0.78, 0.85)], RNFL [0.85 (0.81, 0.86)] and cSLO image models [0.69 (0.63, 0.76)] (all P ≤ 0.05).

CONCLUSIONS

Combining OCT-based RNFL thickness maps with texture-based en face images showed a better ability to discriminate between healthy and POAG than thickness maps alone, particularly in high axial myopic eyes.

The role of artificial intelligence in analysis of biofluid markers for diagnosis and management of glaucoma: A systematic review

PURPOSE

This review focuses on utility of artificial intelligence (AI) in analysis of biofluid markers in glaucoma. We detail the accuracy and validity of AI in the exploration of biomarkers to provide insight into glaucoma pathogenesis.

METHODS

A comprehensive search was conducted across five electronic databases including Embase, Medline, Cochrane Central Register of Controlled Trials, Cochrane Database of Systematic Reviews, and Web of Science. Studies pertaining to biofluid marker analysis using AI or bioinformatics in glaucoma were included. Identified studies were critically appraised and assessed for risk of bias using the Joanna Briggs Institute Critical Appraisal tools.

RESULTS

A total of 10,258 studies were screened and 39 studies met the inclusion criteria, including 23 cross-sectional studies (59%), nine prospective cohort studies (23%), six retrospective cohort studies (15%), and one case-control study (3%). Primary open angle glaucoma (POAG) was the most commonly studied subtype (55% of included studies). Twenty-four studies examined disease characteristics, 10 explored treatment decisions, and 5 provided diagnostic clarification. While studies examined at entire metabolomic or proteomic profiles to determine changes in POAG, there was heterogeneity in the data with over 175 unique, differentially expressed biomarkers reported. Discriminant analysis and artificial neural network predictive models displayed strong differentiating ability between glaucoma patients and controls, although these tools were untested in a clinical context.

CONCLUSION

The use of AI models could inform glaucoma diagnosis with high sensitivity and specificity. While insight into differentially expressed biomarkers is valuable in pathogenic exploration, no clear pathogenic mechanism in glaucoma has emerged.

Artificial intelligence in ophthalmology: The path to the real-world clinic

Artificial intelligence (AI) has great potential to transform healthcare by enhancing the workflow and productivity of clinicians, enabling existing staff to serve more patients, improving patient outcomes, and reducing health disparities. In the field of ophthalmology, AI systems have shown performance comparable with or even better than experienced ophthalmologists in tasks such as diabetic retinopathy detection and grading. However, despite these quite good results, very few AI systems have been deployed in real-world clinical settings, challenging the true value of these systems. This review provides an overview of the current main AI applications in ophthalmology, describes the challenges that need to be overcome prior to clinical implementation of the AI systems, and discusses the strategies that may pave the way to the clinical translation of these systems.

Usability and Clinician Acceptance of a Deep Learning-Based Clinical Decision Support Tool for Predicting Glaucomatous Visual Field Progression

PRCIS

We updated a clinical decision support tool integrating predicted visual field (VF) metrics from an artificial intelligence model and assessed clinician perceptions of the predicted VF metric in this usability study.

PURPOSE

To evaluate clinician perceptions of a prototyped clinical decision support (CDS) tool that integrates visual field (VF) metric predictions from artificial intelligence (AI) models.

METHODS

Ten ophthalmologists and optometrists from the University of California San Diego participated in 6 cases from 6 patients, consisting of 11 eyes, uploaded to a CDS tool ("GLANCE", designed to help clinicians "at a glance"). For each case, clinicians answered questions about management recommendations and attitudes towards GLANCE, particularly regarding the utility and trustworthiness of the AI-predicted VF metrics and willingness to decrease VF testing frequency.

MAIN OUTCOMES AND MEASURES

Mean counts of management recommendations and mean Likert scale scores were calculated to assess overall management trends and attitudes towards the CDS tool for each case. In addition, system usability scale scores were calculated.

RESULTS

The mean Likert scores for trust in and utility of the predicted VF metric and clinician willingness to decrease VF testing frequency were 3.27, 3.42, and 2.64, respectively (1=strongly disagree, 5=strongly agree). When stratified by glaucoma severity, all mean Likert scores decreased as severity increased. The system usability scale score across all responders was 66.1±16.0 (43rd percentile).

CONCLUSIONS

A CDS tool can be designed to present AI model outputs in a useful, trustworthy manner that clinicians are generally willing to integrate into their clinical decision-making. Future work is needed to understand how to best develop explainable and trustworthy CDS tools integrating AI before clinical deployment.

Characteristics of a large, labeled dataset for the training of artificial intelligence for glaucoma screening with fundus photographs

Purpose

Significant visual impairment due to glaucoma is largely caused by the disease being detected too late.

Objective

To build a labeled data set for training artificial intelligence (AI) algorithms for glaucoma screening by fundus photography, to assess the accuracy of the graders, and to characterize the features of all eyes with referable glaucoma (RG).

Design

Cross-sectional study.

Subjects

Color fundus photographs (CFPs) of 113 893 eyes of 60 357 individuals were obtained from EyePACS, California, United States, from a population screening program for diabetic retinopathy.

Methods

Carefully selected graders (ophthalmologists and optometrists) graded the images. To qualify, they had to pass the European Optic Disc Assessment Trial optic disc assessment with ≥ 85% accuracy and 92% specificity. Of 90 candidates, 30 passed. Each image of the EyePACS set was then scored by varying random pairs of graders as "RG," "no referable glaucoma (NRG)," or "ungradable (UG)." In case of disagreement, a glaucoma specialist made the final grading. Referable glaucoma was scored if visual field damage was expected. In case of RG, graders were instructed to mark up to 10 relevant glaucomatous features.

Main Outcome Measures

Qualitative features in eyes with RG.

Results

The performance of each grader was monitored; if the sensitivity and specificity dropped below 80% and 95%, respectively (the final grade served as reference), they exited the study and their gradings were redone by other graders. In all, 20 graders qualified; their mean sensitivity and specificity (standard deviation [SD]) were 85.6% (5.7) and 96.1% (2.8), respectively. The 2 graders agreed in 92.45% of the images (Gwet's AC2, expressing the inter-rater reliability, was 0.917). Of all gradings, the sensitivity and specificity (95% confidence interval) were 86.0 (85.2-86.7)% and 96.4 (96.3-96.5)%, respectively. Of all gradable eyes (n = 111 183; 97.62%) the prevalence of RG was 4.38%. The most common features of RG were the appearance of the neuroretinal rim (NRR) inferiorly and superiorly.

Conclusions

A large data set of CFPs was put together of sufficient quality to develop AI screening solutions for glaucoma. The most common features of RG were the appearance of the NRR inferiorly and superiorly. Disc hemorrhages were a rare feature of RG.

Financial Disclosures

Proprietary or commercial disclosure may be found after the references.

Application of diffusion tensor imaging technology in glaucoma diagnosis

Glaucoma is the first major category of irreversible blinding eye illnesses worldwide. Its leading cause is the death of retinal ganglion cells and their axons, which results in the loss of vision. Research indicates that glaucoma affects the optic nerve and the whole visual pathway. It also reveals that degenerative lesions caused by glaucoma can be found outside the visual pathway. Diffusion tensor imaging (DTI) is a magnetic resonance imaging (MRI) technique that can investigate the complete visual system, including alterations in the optic nerve, optic chiasm, optic tract, lateral geniculate nuclear, and optic radiation. In order to provide a more solid foundation for the degenerative characteristics of glaucoma, this paper will discuss the standard diagnostic techniques for glaucoma through a review of the literature, describe the use of DTI technology in glaucoma in humans and animal models, and introduce these techniques. With the advancement of DTI technology and its coupling with artificial intelligence, DTI represents a potential future for MRI technology in glaucoma research.

Orbital and eyelid diseases: The next breakthrough in artificial intelligence?

Orbital and eyelid disorders affect normal visual functions and facial appearance, and precise oculoplastic and reconstructive surgeries are crucial. Artificial intelligence (AI) network models exhibit a remarkable ability to analyze large sets of medical images to locate lesions. Currently, AI-based technology can automatically diagnose and grade orbital and eyelid diseases, such as thyroid-associated ophthalmopathy (TAO), as well as measure eyelid morphological parameters based on external ocular photographs to assist surgical strategies. The various types of imaging data for orbital and eyelid diseases provide a large amount of training data for network models, which might be the next breakthrough in AI-related research. This paper retrospectively summarizes different imaging data aspects addressed in AI-related research on orbital and eyelid diseases, and discusses the advantages and limitations of this research field.

Use of Artificial Neural Networks to Predict the Progression of Glaucoma in Patients with Sleep Apnea

Aim: To construct neural models to predict the progression of glaucoma in patients with sleep apnea. Materials and Methods: Modeling the use of neural networks was performed using the Neurosolutions commercial simulator. The built databases gather information on a group of patients with primitive open-angle glaucoma and normal-tension glaucoma, who have been associated with sleep apnea syndrome and various stages of disease severity. The data within the database were divided as follows: 65 were used in the neural network training stage and 8 were kept for the validation stage. In total, 21 parameters were selected as input parameters for neural models including: age of patients, BMI (body mass index), systolic and diastolic blood pressure, intraocular pressure, central corneal thickness, corneal biomechanical parameters (IOPcc, HC, CRF), AHI, desaturation index, nocturnal oxygen saturation, remaining AHI, type of apnea, and associated general conditions (diabetes, hypertension, obesity, COPD). The selected output parameters are: c/d ratio, modified visual field parameters (MD, PSD), ganglion cell layer thickness. Forward-propagation neural networks (multilayer perceptron) were constructed with a layer of hidden neurons. The constructed neural models generated the output values for these data. The obtained results were then compared with the experimental values. Results: The best results were obtained during the training stage with the ANN network (21:35:4). If we consider a 25% confidence interval, we find that very good results are obtained during the validation stage, except for the average GCL thickness, for which the errors are slightly higher. Conclusions: Excellent results were obtained during the validation stage, which support the results obtained in other studies in the literature that strengthen the connection between sleep apnea syndrome and glaucoma changes.

Evaluating machine learning classifiers for glaucoma referral decision support in primary care settings

Several artificial intelligence algorithms have been proposed to help diagnose glaucoma by analyzing the functional and/or structural changes in the eye. These algorithms require carefully curated datasets with access to ocular images. In the current study, we have modeled and evaluated classifiers to predict self-reported glaucoma using a single, easily obtained ocular feature (intraocular pressure (IOP)) and non-ocular features (age, gender, race, body mass index, systolic and diastolic blood pressure, and comorbidities). The classifiers were trained on publicly available data of 3015 subjects without a glaucoma diagnosis at the time of enrollment. 337 subjects subsequently self-reported a glaucoma diagnosis in a span of 1–12 years after enrollment. The classifiers were evaluated on the ability to identify these subjects by only using their features recorded at the time of enrollment. Support vector machine, logistic regression, and adaptive boosting performed similarly on the dataset with F1 scores of 0.31, 0.30, and 0.28, respectively. Logistic regression had the highest sensitivity at 60% with a specificity of 69%. Predictive classifiers using primarily non-ocular features have the potential to be used for identifying suspected glaucoma in non-eye care settings, including primary care. Further research into finding additional features that improve the performance of predictive classifiers is warranted.

The Role of Artificial Intelligence in the Diagnosis and Management of Glaucoma

Glaucomatous optic neuropathy is the leading cause of irreversible blindness worldwide. Diagnosis and monitoring of disease involves integrating information from the clinical examination with subjective data from visual field testing and objective biometric data that includes pachymetry, corneal hysteresis, and optic nerve and retinal imaging. This intricate process is further complicated by the lack of clear definitions for the presence and progression of glaucomatous optic neuropathy, which makes it vulnerable to clinician interpretation error. Artificial intelligence (AI) and AI-enabled workflows have been proposed as a plausible solution. Applications derived from this field of computer science can improve the quality and robustness of insights obtained from clinical data that can enhance the clinician's approach to patient care. This review clarifies key terms and concepts used in AI literature, discusses the current advances of AI in glaucoma, elucidates the clinical advantages and challenges to implementing this technology, and highlights potential future applications.