Deep Learning Classifiers for Automated Detection of Gonioscopic Angle Closure Based on Anterior Segment OCT Images

Generative Artificial Intelligence Enhancements for Reducing Image-based Training Data Requirements

Objective

Training data fuel and shape the development of artificial intelligence (AI) models. Intensive data requirements are a major bottleneck limiting the success of AI tools in sectors with inherently scarce data. In health care, training data are difficult to curate, triggering growing concerns that the current lack of access to health care by under-privileged social groups will translate into future bias in health care AIs. In this report, we developed an autoencoder to grow and enhance inherently scarce datasets to alleviate our dependence on big data.

Design

Computational study with open-source data.

Subjects

The data were obtained from 6 open-source datasets comprising patients aged 40-80 years in Singapore, China, India, and Spain.

Methods

The reported framework generates synthetic images based on real-world patient imaging data. As a test case, we used autoencoder to expand publicly available training sets of optic disc photos, and evaluated the ability of the resultant datasets to train AI models in the detection of glaucomatous optic neuropathy.

Main Outcome Measures

Area under the receiver operating characteristic curve (AUC) were used to evaluate the performance of the glaucoma detector. A higher AUC indicates better detection performance.

Results

Results show that enhancing datasets with synthetic images generated by autoencoder led to superior training sets that improved the performance of AI models.

Conclusions

Our findings here help address the increasingly untenable data volume and quality requirements for AI model development and have implications beyond health care, toward empowering AI adoption for all similarly data-challenged fields.

Financial Disclosures

The authors have no proprietary or commercial interest in any materials discussed in this article.

Novel hyperparallel optical coherence tomography for angle closure assessment: comparison with swept-source OCT and gonioscopy

AIM

To evaluate reproducibility and agreement of angle closure assessment by a novel hyperparallel optical coherence tomography (OCT) system (HP-OCT, Cylite Optics, Melbourne, Australia), in comparison with swept-source OCT (SS-OCT, CASIA SS-1000, Tomey Corporation, Nagoya, Japan) and gonioscopy.

METHODS

Cross-sectional study. Phakic subjects >40 years, with no relevant ophthalmic history were consecutively recruited from the glaucoma clinic. Subjects underwent same-day evaluation with HP-OCT, SS-OCT and gonioscopy. The primary outcome was the presence of angle closure, defined as iridotrabecular contact in HP-OCT and SS-OCT images at 0°-180° meridional and as non-visibility of the posterior trabecular meshwork (TM) by gonioscopy. Visibility of TM was also assessed (secondary outcome). Intra and interdevice agreement analysis (Gwet AC1) and logistic regression analysis were performed for primary and secondary outcomes, respectively.

RESULTS

154 sectors from horizontal scans of 77 subjects were analysed. The reproducibility of angle closure assessment by HP-OCT was excellent (AC1 of 0.95 for temporal angle and 1.00 for nasal). Agreement for angle closure detection was very good between HP-OCT and SS-OCT (AC1 of 0.88 for temporal and 0.81 for nasal angle) and good between HP-OCT and gonioscopy (AC1 of 0.71 for temporal and 0.78 for nasal angle). TM was identifiable in 64.4% (94/146) of unprocessed HP-OCT images (both open and closed angles), however not visible in any of the SS-OCT unprocessed images.

CONCLUSIONS

HP-OCT showed excellent reproducibility for angle closure assessment and good agreement with SS-OCT and gonioscopy. HP-OCT technology also provides a unique capability to visualise regions around TM and Schlemm's canal, opening new avenues for clinical research of distal outflow pathways.

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.

Diagnostic criteria of anterior segment swept-source optical coherence tomography to detect gonioscopic angle closure

AIMS

To compare the diagnostic performance of 360° anterior segment optical coherence tomography assessment by applying normative percentile cut-offs versus iris trabecular contact (ITC) for detecting gonioscopic angle closure.

METHODS

In this multicentre study, 394 healthy individuals were included in the normative dataset to derive the age-specific and angle location-specific normative percentiles of angle open distance (AOD500) and trabecular iris space area (TISA500) which were measured every 10° for 360°. 119 healthy participants and 170 patients with angle closure by gonioscopy were included in the test dataset to investigate the diagnostic performance of three sets of criteria for detection of gonioscopic angle closure: (1) the 10th and (2) the 5th percentiles of AOD500/TISA500, and (3) ITC (ie, AOD500/TISA500=0 mm/mm2). The number of angle locations with angle closure defined by each set of the criteria for each eye was used to generate the receiver operating characteristic (ROC) curve for the discrimination between gonioscopic angle closure and open angle.

RESULTS

Of the three sets of diagnostic criteria examined, the area under the ROC curve was greatest for the 10th percentile of AOD500 (0.933), whereas the ITC criterion AOD500=0 mm showed the smallest area under the ROC (0.852) and the difference was statistically significant with or without adjusting for age and axial length (p<0.001). The criterion ≥90° of AOD500 below the 10th percentile attained the best sensitivity 87.6% and specificity 84.9% combination for detecting gonioscopic angle closure.

CONCLUSIONS

Applying the normative percentiles of angle measurements yielded a higher diagnostic performance than ITC for detecting angle closure on gonioscopy.

Development of the AI Pipeline for Corneal Opacity Detection

Ophthalmological services face global inadequacies, especially in low- and middle-income countries, which are marked by a shortage of practitioners and equipment. This study employed a portable slit lamp microscope with video capabilities and cloud storage for more equitable global diagnostic resource distribution. To enhance accessibility and quality of care, this study targets corneal opacity, which is a global cause of blindness. This study has two purposes. The first is to detect corneal opacity from videos in which the anterior segment of the eye is captured. The other is to develop an AI pipeline to detect corneal opacities. First, we extracted image frames from videos and processed them using a convolutional neural network (CNN) model. Second, we manually annotated the images to extract only the corneal margins, adjusted the contrast with CLAHE, and processed them using the CNN model. Finally, we performed semantic segmentation of the cornea using annotated data. The results showed an accuracy of 0.8 for image frames and 0.96 for corneal margins. Dice and IoU achieved a score of 0.94 for semantic segmentation of the corneal margins. Although corneal opacity detection from video frames seemed challenging in the early stages of this study, manual annotation, corneal extraction, and CLAHE contrast adjustment significantly improved accuracy. The incorporation of manual annotation into the AI pipeline, through semantic segmentation, facilitated high accuracy in detecting corneal opacity.

In vivo identification of angle dysgenesis and its relation to genetic markers associated with glaucoma using artificial intelligence

PURPOSE

To predict the presence of angle dysgenesis on anterior-segment optical coherence tomography (ADoA) by using deep learning (DL) and to correlate ADoA with mutations in known glaucoma genes.

PARTICIPANTS

In total, 800 high-definition anterior-segment optical coherence tomography (AS-OCT) images were included, of which 340 images were used to build the machine learning (ML) model. Images used to build the ML model included 170 scans of primary congenital glaucoma (16 patients), juvenile-onset open-angle glaucoma (62 patients), and adult-onset primary open-angle glaucoma eyes (37 patients); the rest were controls (n = 85). The genetic validation dataset consisted of another 393 images of patients with known mutations that were compared with 320 images of healthy controls.

METHODS

ADoA was defined as the absence of Schlemm's canal, the presence of hyperreflectivity over the region of the trabecular meshwork, or a hyperreflective membrane. DL was used to classify a given AS-OCT image as either having angle dysgenesis or not. ADoA was then specifically looked for on AS-OCT images of patients with mutations in the known genes for glaucoma.

RESULTS

The final prediction, which was a consensus-based outcome from the three optimized DL models, had an accuracy of >95%, a specificity of >97%, and a sensitivity of >96% in detecting ADoA in the internal test dataset. Among the patients with known gene mutations, ( MYOC, CYP1B1, FOXC1, and LTBP2 ) ADoA was observed among all the patients in the majority of the images, compared to only 5% of the healthy controls.

CONCLUSION

ADoA can be objectively identified using models built with DL.

Anterior Segment OCT for Detection of Narrow Angles: A Community-Based Diagnostic Accuracy Study

PURPOSE

To assess the diagnostic accuracy of anterior segment OCT (AS-OCT) screening for detecting gonioscopically narrow angles.

DESIGN

Population-based cross-sectional study.

PARTICIPANTS

A stratified random sample of individuals aged ≥ 60 years, selected from a door-to-door census performed in low-lying Nepal.

TESTING

Participants underwent AS-OCT, posterior segment OCT, and intraocular pressure (IOP) testing in the community. Those meeting referral criteria in either eye were invited to have a comprehensive eye examination including gonioscopy. Referral criteria included (i) the lowest 2.5% of AS-OCT measurements, (ii) retinal OCT results suggestive of glaucomatous optic neuropathy, diabetic retinopathy, or age-related macular degeneration, and (iii) elevated IOP.

MAIN OUTCOME MEASURES

Sensitivity and specificity of 5 semiautomated AS-OCT parameters relative to gonioscopically narrow angles, defined as the absence of visible trabecular meshwork for ≥ 180° on nonindentation gonioscopy.

RESULTS

Of 17 656 people aged ≥ 60 years enumerated from 102 communities, 12 633 (71.6%) presented for AS-OCT testing. Referral was recommended for 697 participants based on AS-OCT criteria and 2419 participants based on other criteria, of which 858 had gonioscopy performed by a glaucoma specialist. Each of the 5 AS-OCT parameters offered good diagnostic information for predicting eyes with gonioscopically narrow angles, with areas under the receiver operating characteristic curve ranging from 0.85 to 0.89. The angle opening distance at 750 μm from the scleral spur (AOD750) provided the most diagnostic information, providing an optimal sensitivity of 87% (95% confidence interval [CI], 75%-96%) and specificity of 77% (71%-83%) at a cutpoint of 367 μm, and a sensitivity of 65% (95% CI, 54%-74%) when specificity was constrained to 90% (cutpoint, 283 μm).

CONCLUSIONS

On AS-OCT, the AOD750 parameter detected approximately two-thirds of cases of gonioscopically narrow angles when test specificity was set to 90%. Although such a sensitivity may not be sufficient when screening solely for narrow angles, AS-OCT requires little additional effort if posterior segment OCT is already being performed and thus could provide incremental benefit when performing OCT-based screening.

FINANCIAL DISCLOSURE(S)

The authors have no proprietary or commercial interest in any materials discussed in this article.

Prevalence and Risk Factors of Blindness Among Primary Angle Closure Glaucoma Patients in the United States: An IRIS Registry Analysis

PURPOSE

To assess the prevalence and risk factors of blindness among patients newly diagnosed with primary angle closure glaucoma (PACG) in the United States.

DESIGN

Retrospective cross-sectional study.

METHODS

Eligible patients from the American Academy of Ophthalmology (AAO) Intelligent Research in Sight (IRIS) Registry had newly diagnosed PACG, defined as: 1) observable during a 24-month lookback period from index date of PACG diagnosis; 2) no history of eye drops, laser, or cataract surgery unless preceded by a diagnosis of anatomical narrow angle (ANA); and 3) no history of glaucoma surgery. Logistic regression models were developed to identify risk factors for any (one or both eyes) or bilateral (both eyes) blindness (visual acuity ≤20/200) at first diagnosis of PACG.

RESULTS

Among 43,901 eligible patients, overall prevalence of any and bilateral blindness were 11.5% and 1.8%, respectively. Black and Hispanic patients were at higher risk of any (odds ratios [ORs] 1.42 and 1.21, respectively; P < .001) and bilateral (ORs 2.04 and 1.53, respectively; P < .001) blindness compared with non-Hispanic White patients adjusted for ocular comorbidities. Age <50 or >80 years, male sex, Medicaid or Medicare insurance product, and Southern or Western practice region also conferred a higher risk of blindness (OR > 1.28; P ≤ .01).

CONCLUSIONS

Blindness affects 1 of 9 patients with newly diagnosed PACG in the IRIS Registry. Black and Hispanic patients and Medicaid and Medicare recipients are at significantly higher risk. These findings highlight the severe ocular morbidity among patients with PACG and the need for improved disease awareness and detection methods.

Deep Learning Classification of Angle Closure based on Anterior Segment OCT

PURPOSE

To assess the performance and generalizability of a convolutional neural network (CNN) model for objective and high-throughput identification of primary angle-closure disease (PACD) as well as PACD stage differentiation on anterior segment swept-source OCT (AS-OCT).

DESIGN

Cross-sectional.

PARTICIPANTS

Patients from 3 different eye centers across China and Singapore were recruited for this study. Eight hundred forty-one eyes from the 2 Chinese centers were divided into 170 control eyes, 488 PACS, and 183 PAC + PACG eyes. An additional 300 eyes were recruited from Singapore National Eye Center as a testing data set, divided into 100 control eyes, 100 PACS, and 100 PAC + PACG eyes.

METHODS

Each participant underwent standardized ophthalmic examination and was classified by the presiding physician as either control, primary angle-closure suspect (PACS), primary angle closure (PAC), or primary angle-closure glaucoma (PACG). Deep Learning model was used to train 3 different CNN classifiers: classifier 1 aimed to separate control versus PACS versus PAC + PACG; classifier 2 aimed to separate control versus PACD; and classifier 3 aimed to separate PACS versus PAC + PACG. All classifiers were evaluated on independent validation sets from the same region, China and further tested using data from a different country, Singapore.

MAIN OUTCOME MEASURES

Area under receiver operator characteristic curve (AUC), precision, and recall.

RESULTS

Classifier 1 achieved an AUC of 0.96 on validation set from the same region, but dropped to an AUC of 0.84 on test set from a different country. Classifier 2 achieved the most generalizable performance with an AUC of 0.96 on validation set and AUC of 0.95 on test set. Classifier 3 showed the poorest performance, with an AUC of 0.83 and 0.64 on test and validation data sets, respectively.

CONCLUSIONS

Convolutional neural network classifiers can effectively distinguish PACD from controls on AS-OCT with good generalizability across different patient cohorts. However, their performance is moderate when trying to distinguish PACS versus PAC + PACG.

FINANCIAL DISCLOSURES

The authors have no proprietary or commercial interest in any materials discussed in this article.

Assessment of angle closure disease in the age of artificial intelligence: A review

Primary angle closure glaucoma is a visually debilitating disease that is under-detected worldwide. Many of the challenges in managing primary angle closure disease (PACD) are related to the lack of convenient and precise tools for clinic-based disease assessment and monitoring. Artificial intelligence (AI)- assisted tools to detect and assess PACD have proliferated in recent years with encouraging results. Machine learning (ML) algorithms that utilize clinical data have been developed to categorize angle closure eyes by disease mechanism. Other ML algorithms that utilize image data have demonstrated good performance in detecting angle closure. Nonetheless, deep learning (DL) algorithms trained directly on image data generally outperformed traditional ML algorithms in detecting PACD, were able to accurately differentiate between angle status (open, narrow, closed), and automated the measurement of quantitative parameters. However, more work is required to expand the capabilities of these AI algorithms and for deployment into real-world practice settings. This includes the need for real-world evaluation, establishing the use case for different algorithms, and evaluating the feasibility of deployment while considering other clinical, economic, social, and policy-related factors.

Deep Learning-based Quantification of Anterior Segment OCT Parameters

Objective

To develop and validate a deep learning algorithm that could automate the annotation of scleral spur (SS) and segmentation of anterior chamber (AC) structures for measurements of AC, iris, and angle width parameters in anterior segment OCT (ASOCT) scans.

Design

Cross-sectional study.

Subjects

Data from 2 population-based studies (i.e., the Singapore Chinese Eye Study and Singapore Malay Eye Study) and 1 clinical study on angle-closure disease were included in algorithm development. A separate clinical study on angle-closure disease was used for external validation.

Method

Image contrast of ASOCT scans were first enhanced with CycleGAN. We utilized a heat map regression approach with coarse-to-fine framework for SS annotation. Then, an ensemble network of U-Net, full resolution residual network, and full resolution U-Net was used for structure segmentation. Measurements obtained from predicted SSs and structure segmentation were measured and compared with measurements obtained from manual SS annotation and structure segmentation (i.e., ground truth).

Main Outcome Measures

We measured Euclidean distance and intraclass correlation coefficients (ICC) to evaluate SS annotation and Dice similarity coefficient for structure segmentation. The ICC, Bland-Altman plot, and repeatability coefficient were used to evaluate agreement and precision of measurements.

Results

For SS annotation, our algorithm achieved a Euclidean distance of 124.7 μm, ICC ≥ 0.95, and a 3.3% error rate. For structure segmentation, we obtained Dice similarity coefficient ≥ 0.91 for cornea, iris, and AC segmentation. For angle width measurements, ≥ 95% of data points were within the 95% limits-of-agreement in Bland-Altman plot with insignificant systematic bias (all P > 0.12). The ICC ranged from 0.71-0.87 for angle width measurements, 0.54 for IT750, 0.83-0.85 for other iris measurements, and 0.89-0.99 for AC measurements. Using the same SS coordinates from a human expert, measurements obtained from our algorithm were generally less variable than measurements obtained from a semiautomated angle assessment program.

Conclusion

We developed a deep learning algorithm that could automate SS annotation and structure segmentation in ASOCT scans like human experts, in both open-angle and angle-closure eyes. This algorithm reduces the time needed and subjectivity in obtaining ASOCT measurements.

Financial Disclosures

The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Deep Learning-Based Model for Automatic Assessment of Anterior Angle Chamber in Ultrasound Biomicroscopy

OBJECTIVE

The goal of the work described here was to develop and assess a deep learning-based model that could automatically segment anterior chamber angle (ACA) tissues; classify iris curvature (I-Curv), iris root insertion (IRI), and angle closure (AC); automatically locate scleral spur; and measure ACA parameters in ultrasound biomicroscopy (UBM) images.

METHODS

A total of 11,006 UBM images were obtained from 1538 patients with primary angle-closure glaucoma who were admitted to the Eye Center of Renmin Hospital of Wuhan University (Wuhan, China) to develop an imaging database. The UNet++ network was used to segment ACA tissues automatically. In addition, two support vector machine (SVM) algorithms were developed to classify I-Curv and AC, and a logistic regression (LR) algorithm was developed to classify IRI. Meanwhile, an algorithm was developed to automatically locate the scleral spur and measure ACA parameters. An external data set of 1,658 images from Huangshi Aier Eye Hospital was used to evaluate the performance of the model under different conditions. An additional 439 images were collected to compare the performance of the model with experts.

RESULTS

The model achieved accuracies of 95.2%, 88.9% and 85.6% in classification of AC, I-Curv and IRI, respectively. Compared with ophthalmologists, the model achieved an accuracy of 0.765 in classifying AC, I-Curv and IRI, indicating that its high accuracy was as high as that of the ophthalmologists (p > 0.05). The average relative errors (AREs) of ACA parameters were smaller than 15% in the internal data sets. Intraclass correlation coefficients (ICCs) of all the angle-related parameters were greater than 0.911. ICC values of all iris thickness parameters were greater than 0.884. The accurate measurement of ACA parameters partly depended on accurate localization of the scleral spur (p < 0.001).

CONCLUSION

The model could effectively and accurately evaluate the ACA automatically based on fully automated analysis of UBM images, and it can potentially be a promising tool to assist ophthalmologists. The present study suggested that the deep learning model can be extensively applied to the evaluation of ACA and AC-related biometric risk factors, and it may broaden the application of UBM imaging in the clinical research of primary angle-closure glaucoma.

Automatic interpretation and clinical evaluation for fundus fluorescein angiography images of diabetic retinopathy patients by deep learning

BACKGROUND/AIMS

Fundus fluorescein angiography (FFA) is an important technique to evaluate diabetic retinopathy (DR) and other retinal diseases. The interpretation of FFA images is complex and time-consuming, and the ability of diagnosis is uneven among different ophthalmologists. The aim of the study is to develop a clinically usable multilevel classification deep learning model for FFA images, including prediagnosis assessment and lesion classification.

METHODS

A total of 15 599 FFA images of 1558 eyes from 845 patients diagnosed with DR were collected and annotated. Three convolutional neural network (CNN) models were trained to generate the label of image quality, location, laterality of eye, phase and five lesions. Performance of the models was evaluated by accuracy, F-1 score, the area under the curve and human-machine comparison. The images with false positive and false negative results were analysed in detail.

RESULTS

Compared with LeNet-5 and VGG16, ResNet18 got the best result, achieving an accuracy of 80.79%-93.34% for prediagnosis assessment and an accuracy of 63.67%-88.88% for lesion detection. The human-machine comparison showed that the CNN had similar accuracy with junior ophthalmologists. The false positive and false negative analysis indicated a direction of improvement.

CONCLUSION

This is the first study to do automated standardised labelling on FFA images. Our model is able to be applied in clinical practice, and will make great contributions to the development of intelligent diagnosis of FFA images.

Advancements in high-resolution imaging of the iridocorneal angle

High-resolution imaging methods of the iridocorneal angle (ICA) will lead to enhanced understanding of aqueous humor outflow mechanisms and a characterization of the trabecular meshwork (TM) morphology at the cellular level will help to better understand glaucoma mechanics (e.g., cellular level biomechanics of the particulate glaucomas). This information will translate into immense clinical value, leading to more informed and customized treatment selection, and improved monitoring of procedural interventions that lower intraocular pressure (IOP). Given ICA anatomy, imaging modalities that yield intrinsic optical sectioning or 3D imaging capability will be useful to aid in the visualization of TM layers. This minireview examines advancements in imaging the ICA in high-resolution.

Deep Learning-Based Classification of Subtypes of Primary Angle-Closure Disease With Anterior Segment Optical Coherence Tomography

PRCIS

We developed a deep learning-based classifier that can discriminate primary angle closure suspects (PACS), primary angle closure (PAC)/primary angle closure glaucoma (PACG), and also control eyes with open angle with acceptable accuracy.

PURPOSE

To develop a deep learning-based classifier for differentiating subtypes of primary angle closure disease, including PACS and PAC/PACG, and also normal control eyes.

MATERIALS AND METHODS

Anterior segment optical coherence tomography images were used for analysis with 5 different networks including MnasNet, MobileNet, ResNet18, ResNet50, and EfficientNet. The data set was split with randomization performed at the patient level into a training plus validation set (85%), and a test data set (15%). Then 4-fold cross-validation was used to train the model. In each mentioned architecture, the networks were trained with original and cropped images. Also, the analyses were carried out for single images and images grouped on the patient level (case-based). Then majority voting was applied to the determination of the final prediction.

RESULTS

A total of 1616 images of normal eyes (87 eyes), 1055 images of PACS (66 eyes), and 1076 images of PAC/PACG (66 eyes) eyes were included in the analysis. The mean ± SD age was 51.76 ± 15.15 years and 48.3% were males. MobileNet had the best performance in the model, in which both original and cropped images were used. The accuracy of MobileNet for detecting normal, PACS, and PAC/PACG eyes was 0.99 ± 0.00, 0.77 ± 0.02, and 0.77 ± 0.03, respectively. By running MobileNet in a case-based classification approach, the accuracy improved and reached 0.95 ± 0.03, 0.83 ± 0.06, and 0.81 ± 0.05, respectively. For detecting the open angle, PACS, and PAC/PACG, the MobileNet classifier achieved an area under the curve of 1, 0.906, and 0.872, respectively, on the test data set.

CONCLUSION

The MobileNet-based classifier can detect normal, PACS, and PAC/PACG eyes with acceptable accuracy based on anterior segment optical coherence tomography images.

The application of artificial intelligence in glaucoma diagnosis and prediction

Artificial intelligence is a multidisciplinary and collaborative science, the ability of deep learning for image feature extraction and processing gives it a unique advantage in dealing with problems in ophthalmology. The deep learning system can assist ophthalmologists in diagnosing characteristic fundus lesions in glaucoma, such as retinal nerve fiber layer defects, optic nerve head damage, optic disc hemorrhage, etc. Early detection of these lesions can help delay structural damage, protect visual function, and reduce visual field damage. The development of deep learning led to the emergence of deep convolutional neural networks, which are pushing the integration of artificial intelligence with testing devices such as visual field meters, fundus imaging and optical coherence tomography to drive more rapid advances in clinical glaucoma diagnosis and prediction techniques. This article details advances in artificial intelligence combined with visual field, fundus photography, and optical coherence tomography in the field of glaucoma diagnosis and prediction, some of which are familiar and some not widely known. Then it further explores the challenges at this stage and the prospects for future clinical applications. In the future, the deep cooperation between artificial intelligence and medical technology will make the datasets and clinical application rules more standardized, and glaucoma diagnosis and prediction tools will be simplified in a single direction, which will benefit multiple ethnic groups.

Refractive Error and Anterior Chamber Depth as Risk Factors in Primary Angle Closure Disease: The Chinese American Eye Study

PRCIS

The risk of primary angle closure disease (PACD) rises rapidly with greater hyperopia while remaining relatively low for all degrees of myopia. Refractive error (RE) is useful for angle closure risk stratification in the absence of biometric data.

PURPOSE

To assess the role of RE and anterior chamber depth (ACD) as risk factors in PACD.

METHODS

Chinese American Eye Study participants received complete eye examinations including refraction, gonioscopy, amplitude-scan biometry, and anterior segment ocular coherence tomography imaging. PACD included primary angle closure suspect (≥3 quadrants of angle closure on gonioscopy) and primary angle closure/primary angle closure glaucoma (peripheral anterior synechiae or intraocular pressure >21 mm Hg). Logistic regression models were developed to assess associations between PACD and RE and/or ACD adjusted for sex and age. Locally weighted scatterplot smoothing curves were plotted to assess continuous relationships between variables.

RESULTS

Three thousand nine hundred seventy eyes (3403 open angle and 567 PACD) were included. The risk of PACD increased with greater hyperopia [odds ratio (OR) = 1.41 per diopter (D); P < 0.001] and shallower ACD (OR = 1.75 per 0.1 mm; P < 0.001). Hyperopia (≥ + 0.5 D; OR = 5.03) and emmetropia (-0.5 D to +0.5 D; OR = 2.78) conferred a significantly higher risk of PACD compared with myopia (≤0.5 D). ACD (standardized regression coefficient = -0.54) was a 2.5-fold stronger predictor of PACD risk compared with RE (standardized regression coefficient = 0.22) when both variables were included in one multivariable model. The sensitivity and specificity of a 2.6 mm ACD cutoff for PACD were 77.5% and 83.2% and of a +2.0 D RE cutoff were 22.3% and 89.1%.

CONCLUSION

The risk of PACD rises rapidly with greater hyperopia while remaining relatively low for all degrees of myopia. Although RE is a weaker predictor of PACD than ACD, it remains a useful metric to identify patients who would benefit from gonioscopy in the absence of biometric data.

Generalisability and performance of an OCT-based deep learning classifier for community-based and hospital-based detection of gonioscopic angle closure

PURPOSE

To assess the generalisability and performance of a deep learning classifier for automated detection of gonioscopic angle closure in anterior segment optical coherence tomography (AS-OCT) images.

METHODS

A convolutional neural network (CNN) model developed using data from the Chinese American Eye Study (CHES) was used to detect gonioscopic angle closure in AS-OCT images with reference gonioscopy grades provided by trained ophthalmologists. Independent test data were derived from the population-based CHES, a community-based clinic in Singapore, and a hospital-based clinic at the University of Southern California (USC). Classifier performance was evaluated with receiver operating characteristic curve and area under the receiver operating characteristic curve (AUC) metrics. Interexaminer agreement between the classifier and two human examiners at USC was calculated using Cohen's kappa coefficients.

RESULTS

The classifier was tested using 640 images (311 open and 329 closed) from 127 Chinese Americans, 10 165 images (9595 open and 570 closed) from 1318 predominantly Chinese Singaporeans and 300 images (234 open and 66 closed) from 40 multiethnic USC patients. The classifier achieved similar performance in the CHES (AUC=0.917), Singapore (AUC=0.894) and USC (AUC=0.922) cohorts. Standardising the distribution of gonioscopy grades across cohorts produced similar AUC metrics (range 0.890-0.932). The agreement between the CNN classifier and two human examiners (Ҡ=0.700 and 0.704) approximated interexaminer agreement (Ҡ=0.693) in the USC cohort.

CONCLUSION

An OCT-based deep learning classifier demonstrated consistent performance detecting gonioscopic angle closure across three independent patient populations. This automated method could aid ophthalmologists in the assessment of angle status in diverse patient populations.

Artificial Intelligence for Anterior Segment Diseases: A Review of Potential Developments and Clinical Applications

Artificial intelligence (AI) technology is promising in the field of healthcare. With the developments of big data and image-based analysis, AI shows potential value in ophthalmology applications. Recently, machine learning and deep learning algorithms have made significant progress. Emerging evidence has demonstrated the capability of AI in the diagnosis and management of anterior segment diseases. In this review, we provide an overview of AI applications and potential future applications in anterior segment diseases, focusing on cornea, refractive surgery, cataract, anterior chamber angle detection, and refractive error prediction.

Anterior segment biometric measurements explain misclassifications by a deep learning classifier for detecting gonioscopic angle closure

BACKGROUND/AIMS

To identify biometric parameters that explain misclassifications by a deep learning classifier for detecting gonioscopic angle closure in anterior segment optical coherence tomography (AS-OCT) images.

METHODS

Chinese American Eye Study (CHES) participants underwent gonioscopy and AS-OCT of each angle quadrant. A subset of CHES AS-OCT images were analysed using a deep learning classifier to detect positive angle closure based on manual gonioscopy by a reference human examiner. Parameter measurements were compared between four prediction classes: true positives (TPs), true negatives (TNs), false positives (FPs) and false negatives (FN). Logistic regression models were developed to differentiate between true and false predictions. Performance was assessed using area under the receiver operating curve (AUC) and classifier accuracy metrics.

RESULTS

584 images from 127 participants were analysed, yielding 271 TPs, 224 TNs, 77 FPs and 12 FNs. Parameter measurements differed (p<0.001) between prediction classes among anterior segment parameters, including iris curvature (IC) and lens vault (LV), and angle parameters, including angle opening distance (AOD). FP resembled TP more than FN and TN in terms of anterior segment parameters (steeper IC and higher LV), but resembled TN more than TP and FN in terms of angle parameters (wider AOD). Models for detecting FP (AUC=0.752) and FN (AUC=0.838) improved classifier accuracy from 84.8% to 89.0%.

CONCLUSIONS

Misclassifications by an OCT-based deep learning classifier for detecting gonioscopic angle closure are explained by disagreement between anterior segment and angle parameters. This finding could be used to improve classifier performance and highlights differences between gonioscopic and AS-OCT definitions of angle closure.

From 2 dimensions to 3rd dimension: Quantitative prediction of anterior chamber depth from anterior segment photographs via deep-learning

Anterior chamber depth (ACD) is a major risk factor of angle closure disease, and has been used in angle closure screening in various populations. However, ACD is measured from ocular biometer or anterior segment optical coherence tomography (AS-OCT), which are costly and may not be readily available in primary care and community settings. Thus, this proof-of-concept study aims to predict ACD from low-cost anterior segment photographs (ASPs) using deep-learning (DL). We included 2,311 pairs of ASPs and ACD measurements for algorithm development and validation, and 380 pairs for algorithm testing. We captured ASPs with a digital camera mounted on a slit-lamp biomicroscope. Anterior chamber depth was measured with ocular biometer (IOLMaster700 or Lenstar LS9000) in data used for algorithm development and validation, and with AS-OCT (Visante) in data used for testing. The DL algorithm was modified from the ResNet-50 architecture, and assessed using mean absolute error (MAE), coefficient-of-determination (R2), Bland-Altman plot and intraclass correlation coefficients (ICC). In validation, our algorithm predicted ACD with a MAE (standard deviation) of 0.18 (0.14) mm; R2 = 0.63. The MAE of predicted ACD was 0.18 (0.14) mm in eyes with open angles and 0.19 (0.14) mm in eyes with angle closure. The ICC between actual and predicted ACD measurements was 0.81 (95% CI 0.77, 0.84). In testing, our algorithm predicted ACD with a MAE of 0.23 (0.18) mm; R2 = 0.37. Saliency maps highlighted the pupil and its margin as the main structures used in ACD prediction. This study demonstrates the possibility of predicting ACD from ASPs via DL. This algorithm mimics an ocular biometer in making its prediction, and provides a foundation to predict other quantitative measurements that are relevant to angle closure screening.

Precision Medicine in Glaucoma: Artificial Intelligence, Biomarkers, Genetics and Redox State

Glaucoma is a multifactorial neurodegenerative illness requiring early diagnosis and strict monitoring of the disease progression. Current exams for diagnosis and prognosis are based on clinical examination, intraocular pressure (IOP) measurements, visual field tests, and optical coherence tomography (OCT). In this scenario, there is a critical unmet demand for glaucoma-related biomarkers to enhance clinical testing for early diagnosis and tracking of the disease's development. The introduction of validated biomarkers would allow for prompt intervention in the clinic to help with prognosis prediction and treatment response monitoring. This review aims to report the latest acquisitions on biomarkers in glaucoma, from imaging analysis to genetics and metabolic markers.

Applications of Artificial Intelligence and Deep Learning in Glaucoma

Diagnosis and detection of progression of glaucoma remains challenging. Artificial intelligence-based tools have the potential to improve and standardize the assessment of glaucoma but development of these algorithms is difficult given the multimodal and variable nature of the diagnosis. Currently, most algorithms are focused on a single imaging modality, specifically screening and diagnosis based on fundus photos or optical coherence tomography images. Use of anterior segment optical coherence tomography and goniophotographs is limited. The majority of algorithms designed for disease progression prediction are based on visual fields. No studies in our literature search assessed the use of artificial intelligence for treatment response prediction and no studies conducted prospective testing of their algorithms. Additional challenges to the development of artificial intelligence-based tools include scarcity of data and a lack of consensus in diagnostic criteria. Although research in the use of artificial intelligence for glaucoma is promising, additional work is needed to develop clinically usable tools.

Patch-based CNN for corneal segmentation of AS-OCT images: Effect of the number of classes and image quality upon performance

Anterior segment optical coherence tomography (AS-OCT) is a fundamental ophthalmic imaging technique. AS-OCT images can be examined by experts and segmented to provide quantitative metrics that inform clinical decision making. Manual segmentation of these images is time-consuming and subjective, encouraging software developers in the field to automate segmentation procedures. Traditional programing segmentation approaches are being replaced by deep learning methods, which have shown state-of-the-art performance in AS-OCT image analysis. In this study, a method based on patch-based convolutional neural networks (CNN) was used to segment the three main boundaries of the cornea: the epithelium, Bowman's layer, and the endothelium. To assess the effect of the number of classes on performance, the model was designed as a patch-based boundary classifier using 4 and 8 classes. The effect of image quality was also assessed using different data distributions during the training process. While the Dice coefficient and probability revealed greater precision for the 8 class models, the boundary error metric indicated comparable performance. Additionally, for 8 class models, the image quality test had only a small negative effect on performance, which may be an indication of the robustness of the model and could also suggest that the data augmentation methods did not show significant improvement. These findings contribute to the development of automatic segmentation techniques for AS-OCT images, since patch-based methods have been largely unexplored in favor of other deep learning techniques. The overall performance of the proposed method is comparable to other well-established segmentation methods.

Anterior Segment Parameter Changes after Cataract Surgery in Open-Angle and Angle-Closure Eyes: A Prospective Study

Background: To investigate the anterior segment parameters before and after cataract surgery in open-angle eyes and different subtypes of primary angle-closure glaucoma (PACG) eyes and to further explore the potential relationship between the anterior rotation of the ciliary process and crystalline lens. Methods: An observational, prospective study was performed on 66 patients who had cataract surgery including 22 chronic PACG patients, 22 acute PACG patients, and 22 open-angle cataract patients. Anterior segment parameters including the trabecular-ciliary process distance, ciliary process area, trabecular-ciliary angle (TCA), maximum ciliary body thickness (CBTmax), and so on, were measured using ultrasound biomicroscopy preoperatively and 3 months postoperatively. Results: After the surgery, there were significant increases in TCA (p < 0.001) and CBTmax (p < 0.05) in all three groups, while there was no significant change in the trabecular-ciliary process distance (p > 0.05) in all three groups. No significant difference in the changes of ciliary process area, TCA, and CBTmax (p > 0.05) pre- and postoperatively among the three groups were identified. Conclusions: Extractions of crystalline lenses played similar roles in terms of decreasing the anterior rotation of ciliary processes in open-angle eyes and angle-closure eyes. A natural anatomical abnormality may be a more important factor in the anterior rotation of ciliary processes in PACG patients.

Performance of the Deep Neural Network Ciloctunet, Integrated with Open-Source Software for Ciliary Muscle Segmentation in Anterior Segment OCT Images, Is on Par with Experienced Examiners

Anterior segment optical coherence tomography (AS-OCT), being non-invasive and well-tolerated, is the method of choice for an in vivo investigation of ciliary muscle morphology and function. The analysis requires the segmentation of the ciliary muscle, which is, when performed manually, both time-consuming and prone to examiner bias. Here, we present a convolutional neural network trained for the automatic segmentation of the ciliary muscle in AS-OCT images. Ciloctunet is based on the Freiburg U-net and was trained and validated using 1244 manually segmented OCT images from two previous studies. An accuracy of 97.5% for the validation dataset was achieved. Ciloctunet's performance was evaluated by replicating the findings of a third study with 180 images as the test data. The replication demonstrated that Ciloctunet performed on par with two experienced examiners. The intersection-over-union index (0.84) of the ciliary muscle thickness profiles between Ciloctunet and an experienced examiner was the same as between the two examiners. The mean absolute error between the ciliary muscle thickness profiles of Ciloctunet and the two examiners (35.16 µm and 45.86 µm) was comparable to the one between the examiners (34.99 µm). A statistically significant effect of the segmentation type on the derived biometric parameters was found for the ciliary muscle area but not for the selective thickness reading ("perpendicular axis"). Both the inter-rater and the intra-rater reliability of Ciloctunet were good to excellent. Ciloctunet avoids time-consuming manual segmentation, thus enabling the analysis of large numbers of images of ample study cohorts while avoiding possible examiner biases. Ciloctunet is available as open-source.

Artificial intelligence and deep learning in ophthalmology: Current status and future perspectives

Background

The ophthalmology field was among the first to adopt artificial intelligence (AI) in medicine. The availability of digitized ocular images and substantial data have made deep learning (DL) a popular topic.

Main text

At the moment, AI in ophthalmology is mostly used to improve disease diagnosis and assist decision-making aiming at ophthalmic diseases like diabetic retinopathy (DR), glaucoma, age-related macular degeneration (AMD), cataract and other anterior segment diseases. However, most of the AI systems developed to date are still in the experimental stages, with only a few having achieved clinical applications. There are a number of reasons for this phenomenon, including security, privacy, poor pervasiveness, trust and explainability concerns.

Conclusions

This review summarizes AI applications in ophthalmology, highlighting significant clinical considerations for adopting AI techniques and discussing the potential challenges and future directions.

Deep learning and optical coherence tomography in glaucoma: Bridging the diagnostic gap on structural imaging

Glaucoma is a leading cause of progressive blindness and visual impairment worldwide. Microstructural evidence of glaucomatous damage to the optic nerve head and associated tissues can be visualized using optical coherence tomography (OCT). In recent years, development of novel deep learning (DL) algorithms has led to innovative advances and improvements in automated detection of glaucomatous damage and progression on OCT imaging. DL algorithms have also been trained utilizing OCT data to improve detection of glaucomatous damage on fundus photography, thus improving the potential utility of color photos which can be more easily collected in a wider range of clinical and screening settings. This review highlights ten years of contributions to glaucoma detection through advances in deep learning models trained utilizing OCT structural data and posits future directions for translation of these discoveries into the field of aging and the basic sciences.

Artificial Intelligence for Glaucoma: Creating and Implementing Artificial Intelligence for Disease Detection and Progression

On September 3, 2020, the Collaborative Community on Ophthalmic Imaging conducted its first 2-day virtual workshop on the role of artificial intelligence (AI) and related machine learning techniques in the diagnosis and treatment of various ophthalmic conditions. In a session entitled "Artificial Intelligence for Glaucoma," a panel of glaucoma specialists, researchers, industry experts, and patients convened to share current research on the application of AI to commonly used diagnostic modalities, including fundus photography, OCT imaging, standard automated perimetry, and gonioscopy. The conference participants focused on the use of AI as a tool for disease prediction, highlighted its ability to address inequalities, and presented the limitations of and challenges to its clinical application. The panelists' discussion addressed AI and health equities from clinical, societal, and regulatory perspectives.

A deep learning approach for detection of shallow anterior chamber depth based on the hidden features of fundus photographs

BACKGROUND AND OBJECTIVES

Patients with angle-closure glaucoma (ACG) are asymptomatic until they experience a painful attack. Shallow anterior chamber depth (ACD) is considered a significant risk factor for ACG. We propose a deep learning approach to detect shallow ACD using fundus photographs and to identify the hidden features of shallow ACD.

METHODS

This retrospective study assigned healthy subjects to the training (n = 1188 eyes) and test (n = 594) datasets (prospective validation design). We used a deep learning approach to estimate ACD and build a classification model to identify eyes with a shallow ACD. The proposed method, including subtraction of the input and output images of CycleGAN and a thresholding algorithm, was adopted to visualize the characteristic features of fundus photographs with a shallow ACD.

RESULTS

The deep learning model integrating fundus photographs and clinical variables achieved areas under the receiver operating characteristic curve of 0.978 (95% confidence interval [CI], 0.963-0.988) for an ACD ≤ 2.60 mm and 0.895 (95% CI, 0.868-0.919) for an ACD ≤ 2.80 mm, and outperformed the regression model using only clinical variables. However, the difference between shallow and deep ACD classes on fundus photographs was difficult to be detected with the naked eye. We were unable to identify the features of shallow ACD using the Grad-CAM. The CycleGAN-based feature images showed that area around the macula and optic disk significantly contributed to the classification of fundus photographs with a shallow ACD.

CONCLUSIONS

We demonstrated the feasibility of a novel deep learning model to detect a shallow ACD as a screening tool for ACG using fundus photographs. The CycleGAN-based feature map showed the hidden characteristic features of shallow ACD that were previously undetectable by conventional techniques and ophthalmologists. This framework will facilitate the early detection of shallow ACD to prevent overlooking the risks associated with ACG.

Transfer learning for medical image classification: a literature review

BACKGROUND

Transfer learning (TL) with convolutional neural networks aims to improve performances on a new task by leveraging the knowledge of similar tasks learned in advance. It has made a major contribution to medical image analysis as it overcomes the data scarcity problem as well as it saves time and hardware resources. However, transfer learning has been arbitrarily configured in the majority of studies. This review paper attempts to provide guidance for selecting a model and TL approaches for the medical image classification task.

METHODS

425 peer-reviewed articles were retrieved from two databases, PubMed and Web of Science, published in English, up until December 31, 2020. Articles were assessed by two independent reviewers, with the aid of a third reviewer in the case of discrepancies. We followed the PRISMA guidelines for the paper selection and 121 studies were regarded as eligible for the scope of this review. We investigated articles focused on selecting backbone models and TL approaches including feature extractor, feature extractor hybrid, fine-tuning and fine-tuning from scratch.

RESULTS

The majority of studies (n = 57) empirically evaluated multiple models followed by deep models (n = 33) and shallow (n = 24) models. Inception, one of the deep models, was the most employed in literature (n = 26). With respect to the TL, the majority of studies (n = 46) empirically benchmarked multiple approaches to identify the optimal configuration. The rest of the studies applied only a single approach for which feature extractor (n = 38) and fine-tuning from scratch (n = 27) were the two most favored approaches. Only a few studies applied feature extractor hybrid (n = 7) and fine-tuning (n = 3) with pretrained models.

CONCLUSION

The investigated studies demonstrated the efficacy of transfer learning despite the data scarcity. We encourage data scientists and practitioners to use deep models (e.g. ResNet or Inception) as feature extractors, which can save computational costs and time without degrading the predictive power.

Performance evaluation of various deep learning based models for effective glaucoma evaluation using optical coherence tomography images

Glaucoma is the dominant reason for irreversible blindness worldwide, and its best remedy is early and timely detection. Optical coherence tomography has come to be the most commonly used imaging modality in detecting glaucomatous damage in recent years. Deep Learning using Optical Coherence Tomography Modality helps in predicting glaucoma more accurately and less tediously. This experimental study aims to perform glaucoma prediction using eight different ImageNet models from Optical Coherence Tomography of Glaucoma. A thorough investigation is performed to evaluate these models' performances on various efficiency metrics, which will help discover the best performing model. Every net is tested on three different optimizers, namely Adam, Root Mean Squared Propagation, and Stochastic Gradient Descent, to find the best relevant results. An attempt has been made to improvise the performance of models using transfer learning and fine-tuning. The work presented in this study was initially trained and tested on a private database that consists of 4220 images (2110 normal optical coherence tomography and 2110 glaucoma optical coherence tomography). Based on the results, the four best-performing models are shortlisted. Later, these models are tested on the well-recognized standard public Mendeley dataset. Experimental results illustrate that VGG16 using the Root Mean Squared Propagation Optimizer attains auspicious performance with 95.68% accuracy. The proposed work concludes that different ImageNet models are a good alternative as a computer-based automatic glaucoma screening system. This fully automated system has a lot of potential to tell the difference between normal Optical Coherence Tomography and glaucomatous Optical Coherence Tomography automatically. The proposed system helps in efficiently detecting this retinal infection in suspected patients for better diagnosis to avoid vision loss and also decreases senior ophthalmologists' (experts) precious time and involvement.

Diagnostic Performance of Deep Learning Classifiers in Measuring Peripheral Anterior Synechia Based on Swept Source Optical Coherence Tomography Images

Purpose

To develop deep learning classifiers and evaluate their diagnostic performance in detecting the static gonioscopic angle closure and peripheral anterior synechia (PAS) based on swept source optical coherence tomography (SS-OCT) images.

Materials and Methods

Subjects were recruited from the Glaucoma Service at Zhongshan Ophthalmic Center of Sun Yat-sun University, Guangzhou, China. Each subject underwent a complete ocular examination, such as gonioscopy and SS-OCT imaging. Two deep learning classifiers, using convolutional neural networks (CNNs), were developed to diagnose the static gonioscopic angle closure and to differentiate appositional from synechial angle closure based on SS-OCT images. Area under the receiver operating characteristic (ROC) curve (AUC) was used as outcome measure to evaluate the diagnostic performance of two deep learning systems.

Results

A total of 439 eyes of 278 Chinese patients, which contained 175 eyes of positive PAS, were recruited to develop diagnostic models. For the diagnosis of static gonioscopic angle closure, the first deep learning classifier achieved an AUC of 0.963 (95% CI, 0.954–0.972) with a sensitivity of 0.929 and a specificity of 0.877. The AUC of the second deep learning classifier distinguishing appositional from synechial angle closure was 0.873 (95% CI, 0.864–0.882) with a sensitivity of 0.846 and a specificity of 0.764.

Conclusion

Deep learning systems based on SS-OCT images showed good diagnostic performance for gonioscopic angle closure and moderate performance in the detection of PAS.

Hybrid Variation-Aware Network for Angle-Closure Assessment in AS-OCT

Automatic angle-closure assessment in Anterior Segment OCT (AS-OCT) images is an important task for the screening and diagnosis of glaucoma, and the most recent computer-aided models focus on a binary classification of anterior chamber angles (ACA) in AS-OCT, i.e., open-angle and angle-closure. In order to assist clinicians who seek better to understand the development of the spectrum of glaucoma types, a more discriminating three-class classification scheme was suggested, i.e., the classification of ACA was expended to include open-, appositional- and synechial angles. However, appositional and synechial angles display similar appearances in an AS-OCT image, which makes classification models struggle to differentiate angle-closure subtypes based on static AS-OCT images. In order to tackle this issue, we propose a 2D-3D Hybrid Variation-aware Network (HV-Net) for open-appositional-synechial ACA classification from AS-OCT imagery. Specifically, taking into account clinical priors, we first reconstruct the 3D iris surface from an AS-OCT sequence, and obtain the geometrical characteristics necessary to provide global shape information. 2D AS-OCT slices and 3D iris representations are then fed into our HV-Net to extract cross-sectional appearance features and iris morphological features, respectively. To achieve similar results to those of dynamic gonioscopy examination, which is the current gold standard for diagnostic angle assessment, the paired AS-OCT images acquired in dark and light illumination conditions are used to obtain an accurate characterization of configurational changes in ACAs and iris shapes, using a Variation-aware Block. In addition, an annealing loss function was introduced to optimize our model, so as to encourage the sub-networks to map the inputs into the more conducive spaces to extract dark-to-light variation representations, while retaining the discriminative power of the learned features. The proposed model is evaluated across 1584 paired AS-OCT samples, and it has demonstrated its superiority in classifying open-, appositional- and synechial angles.

[Artificial intelligence and glaucoma: A literature review]

In recent years, research in artificial intelligence (AI) has experienced an unprecedented surge in the field of ophthalmology, in particular glaucoma. The diagnosis and follow-up of glaucoma is complex and relies on a body of clinical evidence and ancillary tests. This large amount of information from structural and functional testing of the optic nerve and macula makes glaucoma a particularly appropriate field for the application of AI. In this paper, we will review work using AI in the field of glaucoma, whether for screening, diagnosis or detection of progression. Many AI strategies have shown promising results for glaucoma detection using fundus photography, optical coherence tomography, or automated perimetry. The combination of these imaging modalities increases the performance of AI algorithms, with results comparable to those of humans. We will discuss potential applications as well as obstacles and limitations to the deployment and validation of such models. While there is no doubt that AI has the potential to revolutionize glaucoma management and screening, research in the coming years will need to address unavoidable questions regarding the clinical significance of such results and the explicability of the predictions.

Updates in deep learning research in ophthalmology

Ophthalmology has been one of the early adopters of artificial intelligence (AI) within the medical field. Deep learning (DL), in particular, has garnered significant attention due to the availability of large amounts of data and digitized ocular images. Currently, AI in Ophthalmology is mainly focused on improving disease classification and supporting decision-making when treating ophthalmic diseases such as diabetic retinopathy, age-related macular degeneration (AMD), glaucoma and retinopathy of prematurity (ROP). However, most of the DL systems (DLSs) developed thus far remain in the research stage and only a handful are able to achieve clinical translation. This phenomenon is due to a combination of factors including concerns over security and privacy, poor generalizability, trust and explainability issues, unfavorable end-user perceptions and uncertain economic value. Overcoming this challenge would require a combination approach. Firstly, emerging techniques such as federated learning (FL), generative adversarial networks (GANs), autonomous AI and blockchain will be playing an increasingly critical role to enhance privacy, collaboration and DLS performance. Next, compliance to reporting and regulatory guidelines, such as CONSORT-AI and STARD-AI, will be required to in order to improve transparency, minimize abuse and ensure reproducibility. Thirdly, frameworks will be required to obtain patient consent, perform ethical assessment and evaluate end-user perception. Lastly, proper health economic assessment (HEA) must be performed to provide financial visibility during the early phases of DLS development. This is necessary to manage resources prudently and guide the development of DLS.

Digital Gonioscopy Based on Three-dimensional Anterior-Segment OCT: An International Multicenter Study

PURPOSE

To develop and evaluate the performance of a 3-dimensional (3D) deep-learning-based automated digital gonioscopy system (DGS) in detecting 2 major characteristics in eyes with suspected primary angle-closure glaucoma (PACG): (1) narrow iridocorneal angles (static gonioscopy, Task I) and (2) peripheral anterior synechiae (PAS) (dynamic gonioscopy, Task II) on OCT scans.

DESIGN

International, cross-sectional, multicenter study.

PARTICIPANTS

A total of 1.112 million images of 8694 volume scans (2294 patients) from 3 centers were included in this study (Task I, training/internal validation/external testing: 4515, 1101, and 2222 volume scans, respectively; Task II, training/internal validation/external testing: 378, 376, and 102 volume scans, respectively).

METHODS

For Task I, a narrow angle was defined as an eye in which the posterior pigmented trabecular meshwork was not visible in more than 180° without indentation in the primary position captured in the dark room from the scans. For Task II, PAS was defined as the adhesion of the iris to the trabecular meshwork. The diagnostic performance of the 3D DGS was evaluated in both tasks with gonioscopic records as reference.

MAIN OUTCOME MEASURES

The area under the curve (AUC), sensitivity, and specificity of the 3D DGS were calculated.

RESULTS

In Task I, 29.4% of patients had a narrow angle. The AUC, sensitivity, and specificity of 3D DGS on the external testing datasets were 0.943 (0.933-0.953), 0.867 (0.838-0.895), and 0.878 (0.859-0.896), respectively. For Task II, 13.8% of patients had PAS. The AUC, sensitivity, and specificity of 3D DGS were 0.902 (0.818-0.985), 0.900 (0.714-1.000), and 0.890 (0.841-0.938), respectively, on the external testing set at quadrant level following normal clinical practice; and 0.885 (0.836-0.933), 0.912 (0.816-1.000), and 0.700 (0.660-0.741), respectively, on the external testing set at clock-hour level.

CONCLUSIONS

The 3D DGS is effective in detecting eyes with suspected PACG. It has the potential to be used widely in the primary eye care community for screening of subjects at high risk of developing PACG.

Optical Coherence Tomography and Glaucoma

Early detection and monitoring are critical to the diagnosis and management of glaucoma, a progressive optic neuropathy that causes irreversible blindness. Optical coherence tomography (OCT) has become a commonly utilized imaging modality that aids in the detection and monitoring of structural glaucomatous damage. Since its inception in 1991, OCT has progressed through multiple iterations, from time-domain OCT, to spectral-domain OCT, to swept-source OCT, all of which have progressively improved the resolution and speed of scans. Even newer technological advancements and OCT applications, such as adaptive optics, visible-light OCT, and OCT-angiography, have enriched the use of OCT in the evaluation of glaucoma. This article reviews current commercial and state-of-the-art OCT technologies and analytic techniques in the context of their utility for glaucoma diagnosis and management, as well as promising future directions.

Deep Learning-Enabled Identification of Autoimmune Encephalitis on 3D Multi-Sequence MRI

BACKGROUND

Autoimmune encephalitis (AE) is a noninfectious emergency with severe clinical attacks. It is difficult for the earlier diagnosis of acute AE due to the lack of antibody detection resources.

PURPOSE

To construct a deep learning (DL) algorithm using multi-sequence magnetic resonance imaging (MRI) for the identification of acute AE.

STUDY TYPE

Retrospective.

POPULATION

One hundred and sixty AE patients (90 women; median age 36), 177 herpes simplex virus encephalitis (HSVE) (89 women; median age 39), and 184 healthy controls (HC) (95 women; median age 39) were included. Fifty-two patients from another site were enrolled for external validation.

FIELD STRENGTH/SEQUENCE

3.0 T; fast spin-echo (T₁ WI, T₂ WI, fluid attenuated inversion recovery imaging) and spin-echo echo-planar diffusion weighted imaging.

ASSESSMENT

Five DL models based on individual or combined four MRI sequences to classify the datasets as AE, HSVE, or HC. Reader experiment was further carried out by radiologists.

STATISTICAL TESTS

The discriminative performance of different models was assessed using the area under the receiver operating characteristic curve (AUC). The optimal threshold cut-off was identified when sensitivity and specificity were maximized (sensitivity + specificity - 1) in the validation set. Classification performance using confusion matrices was reported to evaluate the diagnostic value of the models and the radiologists' assessments before being assessed by the paired t-test (P < 0.05 was considered significant).

RESULTS

In the internal test set, the fusion model achieved the significantly greatest diagnostic performance than single-sequence DL models with AUCs of 0.828, 0.884, and 0.899 for AE, HSVE, and HC, respectively. The model demonstrated a consistently high performance in the external validation set with AUCs of 0.831 (AE), 0.882 (HSVE), and 0.892 (HC). The fusion model also demonstrated significantly higher performance than all radiologists in identifying AE (accuracy between the fuse model vs. average radiologist: 83% vs. 72%).

DATA CONCLUSION

The proposed DL algorithm derived from multi-sequence MRI provided desirable identification and classification of acute AE.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

A Deep Learning System for Automatic Assessment of Anterior Chamber Angle in Ultrasound Biomicroscopy Images

Purpose

To develop and assess a deep learning system that automatically detects angle closure and quantitatively measures angle parameters from ultrasound biomicroscopy (UBM) images using a deep learning algorithm.

Methods

A total of 3788 UBM images (2146 open angle and 1642 angle closure) from 1483 patients were collected. We developed a convolutional neural network (CNN) based on the InceptionV3 network for automatic classification of angle closure and open angle. For nonclosed images, we developed a CNN based on the EfficienttNetB3 network for the automatic localization of the scleral spur and the angle recess; then, the Unet network was used to segment the anterior chamber angle (ACA) tissue automatically. Based on the results of the latter two processes, we developed an algorithm to automatically measure the trabecular-iris angle (TIA500 and TIA750), angle-opening distance (AOD500 and AOD750), and angle recess area (ARA500 and ARA750) for quantitative evaluation of angle width.

Results

Using manual labeling as the reference standard, the ACA classification network's accuracy reached 98.18%, and the sensitivity and specificity for angle closure reached 98.74% and 97.44%, respectively. The deep learning system realized the automatic measurement of the angle parameters, and the mean of differences was generally small between automatic measurement and manual measurement. The coefficients of variation of TIA500, TIA750, AOD500, AOD750, ARA500, and ARA750 measured by the deep learning system were 5.77%, 4.67%, 10.76%, 7.71%, 16.77%, and 12.70%, respectively. The within-subject standard deviations of TIA500, TIA750, AOD500, AOD750, ARA500, and ARA750 were 5.77 degrees, 4.56 degrees, 155.92 µm, 147.51 µm, 0.10 mm², and 0.12 mm², respectively. The intraclass correlation coefficients of all the angle parameters were greater than 0.935.

Conclusions

The deep learning system can effectively and accurately evaluate the ACA automatically based on fully automated analysis of a UBM image.

Translational Relevance

The present work suggests that the deep learning system described here could automatically detect angle closure and quantitatively measure angle parameters from UBM images and enhancing the intelligent diagnosis and management of primary angle-closure glaucoma.

Circumpapillary OCT-focused hybrid learning for glaucoma grading using tailored prototypical neural networks

Glaucoma is one of the leading causes of blindness worldwide and Optical Coherence Tomography (OCT) is the quintessential imaging technique for its detection. Unlike most of the state-of-the-art studies focused on glaucoma detection, in this paper, we propose, for the first time, a novel framework for glaucoma grading using raw circumpapillary B-scans. In particular, we set out a new OCT-based hybrid network which combines hand-driven and deep learning algorithms. An OCT-specific descriptor is proposed to extract hand-crafted features related to the retinal nerve fibre layer (RNFL). In parallel, an innovative CNN is developed using skip-connections to include tailored residual and attention modules to refine the automatic features of the latent space. The proposed architecture is used as a backbone to conduct a novel few-shot learning based on static and dynamic prototypical networks. The k-shot paradigm is redefined giving rise to a supervised end-to-end system which provides substantial improvements discriminating between healthy, early and advanced glaucoma samples. The training and evaluation processes of the dynamic prototypical network are addressed from two fused databases acquired via Heidelberg Spectralis system. Validation and testing results reach a categorical accuracy of 0.9459 and 0.8788 for glaucoma grading, respectively. Besides, the high performance reported by the proposed model for glaucoma detection deserves a special mention. The findings from the class activation maps are directly in line with the clinicians' opinion since the heatmaps pointed out the RNFL as the most relevant structure for glaucoma diagnosis.

Telehealth and Screening Strategies in the Diagnosis and Management of Glaucoma

Telehealth has become a viable option for glaucoma screening and glaucoma monitoring due to advances in technology. The ability to measure intraocular pressure without an anesthetic and to take optic nerve photographs without pharmacologic pupillary dilation using portable equipment have allowed glaucoma screening programs to generate enough data for assessment. At home, patients can perform visual acuity testing, web-based visual field testing, rebound tonometry, and video visits with the physician to monitor for glaucomatous progression. Artificial intelligence will enhance the accuracy of data interpretation and inspire confidence in popularizing telehealth for glaucoma.

Semi-supervised generative adversarial networks for closed-angle detection on anterior segment optical coherence tomography images: an empirical study with a small training dataset

BACKGROUND

Semi-supervised learning algorithms can leverage an unlabeled dataset when labeling is limited or expensive to obtain. In the current study, we developed and evaluated a semi-supervised generative adversarial networks (GANs) model that detects closed-angle on anterior segment optical coherence tomography (AS-OCT) images using a small labeled dataset.

METHODS

In this cross-sectional study, a semi-supervised GANs model was developed for automatic closed-angle detection training on a small labeled and large unsupervised training dataset collected from the Joint Shantou International Eye Center of Shantou University and the Chinese University of Hong Kong (JSIEC). The closed-angle was defined as iris-trabecular contact beyond the scleral spur in AS-OCT images. We further developed two supervised deep learning (DL) models training on the same supervised dataset and the whole dataset separately. The semi-supervised GANs model and supervised DL models' performance were compared on two independent testing datasets from JSIEC (515 images) and the Department of Ophthalmology (84 images), National University Health System, respectively. The diagnostic performance was assessed by evaluation matrices, including the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).

RESULTS

For closed-angle detection using clinician grading of AS-OCT imaging as the reference standard, the semi-supervised GANs model showed comparable performance, with AUCs of 0.97 (95% CI, 0.96-0.99) and 0.98 (95% CI, 0.94-1.00), compared with the supervised DL model (using the whole dataset) [AUCs of 0.97 (95% CI, 0.96-0.99), and 0.97 (95% CI, 0.94-1.00)]. When training on the same small supervised dataset, the semi-supervised GANs achieved performance at least as well as, if not better than, the supervised DL model [AUCs of 0.90 (95% CI: 0.84-0.96), and 0.92 (95% CI, 0.86-0.97)].

CONCLUSIONS

The semi-supervised GANs method achieves diagnostic performance at least as good as a supervised DL model when trained on small labeled datasets. Further development of semi-supervised learning methods could be useful within clinical and research settings.

TRIAL REGISTRATION NUMBER

ChiCTR2000037892.

Automatic Anterior Chamber Angle Classification Using Deep Learning System and Anterior Segment Optical Coherence Tomography Images

Purpose

The purpose of this study was to develop a software package for the automatic classification of anterior chamber angle using anterior segment optical coherence tomography (AS-OCT).

Methods

AS-OCT images were collected from subjects with open, narrow, and closure anterior chamber angles, which were graded based on ultrasound biomicroscopy (UBM) results. The Inception version 3 network and the transfer learning technique were applied in the design of an algorithm for anterior chamber angle classification. The classification performance was evaluated by fivefold cross-validation and on an independent test dataset.

Results

The proposed algorithm reached a sensitivity of 0.999 and specificity of 1.000 in the judgment of closed and nonclosed angles. The overall classification of the proposed method in open angle, narrow angle, and angle-closure classifications reached a sensitivity of 0.989 and specificity of 0.995. Additionally, the sensitivity and specificity reached 1.000 and 1.000 for angle-closure, 0.983 and 0.993 for narrow angle, and 0.985 and 0.991 for open angle.

Conclusions

The experimental results showed that the proposed method can achieve a high accuracy of anterior chamber angle classification using AS-OCT images, and could be of value in future practice.

Translational Relevance

The proposed deep learning-based method that automate the classification of anterior chamber angle can facilitate clinical assessment of glaucoma.

Machine Learning-Guided Prediction of Central Anterior Chamber Depth Using Slit Lamp Images from a Portable Smartphone Device

There is currently no objective portable screening modality for narrow angles in the community. In this prospective, single-centre image validation study, we used machine learning on slit lamp images taken with a portable smartphone device (MIDAS) to predict the central anterior chamber depth (ACD) of phakic patients with undilated pupils. Patients 60 years or older with no history of laser or intraocular surgery were recruited. Slit lamp images were taken with MIDAS, followed by anterior segment optical coherence tomography (ASOCT; Casia SS-1000, Tomey, Nagoya, Japan). After manual annotation of the anatomical landmarks of the slit lamp photos, machine learning was applied after image processing and feature extraction to predict the ACD. These values were then compared with those acquired from the ASOCT. Sixty-six eyes (right = 39, 59.1%) were included for analysis. The predicted ACD values formed a strong positive correlation with the measured ACD values from ASOCT (R² = 0.91 for training data and R² = 0.73 for test data). This study suggests the possibility of estimating central ACD using slit lamp images taken from portable devices.