Performance of Artificial Intelligence in Detecting Diabetic Macular Edema From Fundus Photography and Optical Coherence Tomography Images: A Systematic Review and Meta-analysis.

BACKGROUND: Diabetic macular edema (DME) is the leading cause of vision loss in people with diabetes. Application of artificial intelligence (AI) in interpreting fundus photography (FP) and optical coherence tomography (OCT) images allows prompt detection and intervention. PURPOSE: To evaluate the performance of AI in detecting DME from FP or OCT images and identify potential factors affecting model performances. DATA SOURCES: We searched seven electronic libraries up to 12 February 2023. STUDY SELECTION: We included studies using AI to detect DME from FP or OCT images. DATA EXTRACTION: We extracted study characteristics and performance parameters. DATA SYNTHESIS: Fifty-three studies were included in the meta-analysis. FP-based algorithms of 25 studies yielded pooled area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of 0.964, 92.6%, and 91.1%, respectively. OCT-based algorithms of 28 studies yielded pooled AUROC, sensitivity, and specificity of 0.985, 95.9%, and 97.9%, respectively. Potential factors improving model performance included deep learning techniques, larger size, and more diversity in training data sets. Models demonstrated better performance when validated internally than externally, and those trained with multiple data sets showed better results upon external validation. LIMITATIONS: Analyses were limited by unstandardized algorithm outcomes and insufficient data in patient demographics, OCT volumetric scans, and external validation. CONCLUSIONS: This meta-analysis demonstrates satisfactory performance of AI in detecting DME from FP or OCT images. External validation is warranted for future studies to evaluate model generalizability. Further investigations may estimate optimal sample size, effect of class balance, patient demographics, and additional benefits of OCT volumetric scans.

Developing a privacy-preserving deep learning model for glaucoma detection: a multicentre study with federated learning.

BACKGROUND: Deep learning (DL) is promising to detect glaucoma. However, patients' privacy and data security are major concerns when pooling all data for model development. We developed a privacy-preserving DL model using the federated learning (FL) paradigm to detect glaucoma from optical coherence tomography (OCT) images. METHODS: This is a multicentre study. The FL paradigm consisted of a 'central server' and seven eye centres in Hong Kong, the USA and Singapore. Each centre first trained a model locally with its own OCT optic disc volumetric dataset and then uploaded its model parameters to the central server. The central server used FedProx algorithm to aggregate all centres' model parameters. Subsequently, the aggregated parameters are redistributed to each centre for its local model optimisation. We experimented with three three-dimensional (3D) networks to evaluate the stabilities of the FL paradigm. Lastly, we tested the FL model on two prospectively collected unseen datasets. RESULTS: We used 9326 volumetric OCT scans from 2785 subjects. The FL model performed consistently well with different networks in 7 centres (accuracies 78.3%-98.5%, 75.9%-97.0%, and 78.3%-97.5%, respectively) and stably in the 2 unseen datasets (accuracies 84.8%-87.7%, 81.3%-84.8%, and 86.0%-87.8%, respectively). The FL model achieved non-inferior performance in classifying glaucoma compared with the traditional model and significantly outperformed the individual models. CONCLUSION: The 3D FL model could leverage all the datasets and achieve generalisable performance, without data exchange across centres. This study demonstrated an OCT-based FL paradigm for glaucoma identification with ensured patient privacy and data security, charting another course toward the real-world transition of artificial intelligence in ophthalmology.

What can GPT-4 do for Diagnosing Rare Eye Diseases? A Pilot Study.

INTRODUCTION: Generative pretrained transformer-4 (GPT-4) has gained widespread attention from society, and its potential has been extensively evaluated in many areas. However, investigation of GPT-4's use in medicine, especially in the ophthalmology field, is still limited. This study aims to evaluate GPT-4's capability to identify rare ophthalmic diseases in three simulated scenarios for different end-users, including patients, family physicians, and junior ophthalmologists. METHODS: We selected ten treatable rare ophthalmic disease cases from the publicly available EyeRounds service. We gradually increased the amount of information fed into GPT-4 to simulate the scenarios of patient, family physician, and junior ophthalmologist using GPT-4. GPT-4's responses were evaluated from two aspects: suitability (appropriate or inappropriate) and accuracy (right or wrong) by senior ophthalmologists (> 10 years' experiences). RESULTS: Among the 30 responses, 83.3% were considered "appropriate" by senior ophthalmologists. In the scenarios of simulated patient, family physician, and junior ophthalmologist, seven (70%), ten (100%), and eight (80%) responses were graded as "appropriate" by senior ophthalmologists. However, compared to the ground truth, GPT-4 could only output several possible diseases generally without "right" responses in the simulated patient scenarios. In contrast, in the simulated family physician scenario, 50% of GPT-4's responses were "right," and in the simulated junior ophthalmologist scenario, the model achieved a higher "right" rate of 90%. CONCLUSION: To our knowledge, this is the first proof-of-concept study that evaluates GPT-4's capacity to identify rare eye diseases in simulated scenarios involving patients, family physicians, and junior ophthalmologists. The results indicate that GPT-4 has the potential to serve as a consultation assisting tool for patients and family physicians to receive referral suggestions and an assisting tool for junior ophthalmologists to diagnose rare eye diseases. However, it is important to approach GPT-4 with caution and acknowledge the need for verification and careful referrals in clinical settings.

High Myopia Normative Database of Peripapillary Retinal Nerve Fiber Layer Thickness to Detect Myopic Glaucoma in a Chinese Population.

PURPOSE: To develop and validate the performance of a high myopia (HM)-specific normative database of peripapillary retinal nerve fiber layer (pRNFL) thickness in differentiating HM from highly myopic glaucoma (HMG). DESIGN: Cross-sectional multicenter study. PARTICIPANTS: A total of 1367 Chinese participants (2325 eyes) with nonpathologic HM or HMG were included from 4 centers. After quality control, 1108 eyes from 694 participants with HM were included in the normative database; 459 eyes from 408 participants (323 eyes with HM and 136 eyes with HMG) and 322 eyes from 197 participants (131 eyes with HM and 191 eyes with HMG) were included in the internal and external validation sets, respectively. Only HMG eyes with an intraocular pressure > 21 mmHg were included. METHODS: The pRNFL thickness was measured with swept-source (SS) OCT. Four strategies of pRNFL-specified values were examined, including global and quadrantic pRNFL thickness below the lowest fifth or the lowest first percentile of the normative database. MAIN OUTCOMES MEASURES: The accuracy, sensitivity, and specificity of the HM-specific normative database for detecting HMG. RESULTS: Setting the fifth percentile of the global pRNFL thickness as the threshold, using the HM-specific normative database, we achieved an accuracy of 0.93 (95% confidence interval [CI], 0.90-0.95) and 0.85 (95% CI, 0.81-0.89), and, using the first percentile as the threshold, we acheived an accuracy of 0.85 (95% CI, 0.81-0.88) and 0.70 (95% CI, 0.65-0.75) in detecting HMG in the internal and external validation sets, respectively. The fifth percentile of the global pRNFL thickness achieved high sensitivities of 0.75 (95% CI, 0.67-0.82) and 0.75 (95% CI, 0.68-0.81) and specificities of 1.00 (95% CI, 0.99-1.00) and 1.00 (95% CI, 0.97-1.00) in the internal and external validation datasets, respectively. Compared with the built-in database of the OCT device, the HM-specific normative database showed a higher sensitivity and specificity than the corresponding pRNFL thickness below the fifth or first percentile (P < 0.001 for all). CONCLUSIONS: The HM-specific normative database is more capable of detecting HMG eyes than the SS OCT built-in database, which may be an effective tool for differential diagnosis between HMG and HM. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Deep learning in optical coherence tomography: Where are the gaps?

Optical coherence tomography (OCT) is a non-invasive optical imaging modality, which provides rapid, high-resolution and cross-sectional morphology of macular area and optic nerve head for diagnosis and managing of different eye diseases. However, interpreting OCT images requires experts in both OCT images and eye diseases since many factors such as artefacts and concomitant diseases can affect the accuracy of quantitative measurements made by post-processing algorithms. Currently, there is a growing interest in applying deep learning (DL) methods to analyse OCT images automatically. This review summarises the trends in DL-based OCT image analysis in ophthalmology, discusses the current gaps, and provides potential research directions. DL in OCT analysis shows promising performance in several tasks: (1) layers and features segmentation and quantification; (2) disease classification; (3) disease progression and prognosis; and (4) referral triage level prediction. Different studies and trends in the development of DL-based OCT image analysis are described and the following challenges are identified and described: (1) public OCT data are scarce and scattered; (2) models show performance discrepancies in real-world settings; (3) models lack of transparency; (4) there is a lack of societal acceptance and regulatory standards; and (5) OCT is still not widely available in underprivileged areas. More work is needed to tackle the challenges and gaps, before DL is further applied in OCT image analysis for clinical use.

Deep Reinforcement Learning-Based Retinal Imaging in Alzheimer's Disease: Potential and Perspectives.

Alzheimer's disease (AD) remains a global health challenge in the 21st century due to its increasing prevalence as the major cause of dementia. State-of-the-art artificial intelligence (AI)-based tests could potentially improve population-based strategies to detect and manage AD. Current retinal imaging demonstrates immense potential as a non-invasive screening measure for AD, by studying qualitative and quantitative changes in the neuronal and vascular structures of the retina that are often associated with degenerative changes in the brain. On the other hand, the tremendous success of AI, especially deep learning, in recent years has encouraged its incorporation with retinal imaging for predicting systemic diseases. Further development in deep reinforcement learning (DRL), defined as a subfield of machine learning that combines deep learning and reinforcement learning, also prompts the question of how it can work hand in hand with retinal imaging as a viable tool for automated prediction of AD. This review aims to discuss potential applications of DRL in using retinal imaging to study AD, and their synergistic application to unlock other possibilities, such as AD detection and prediction of AD progression. Challenges and future directions, such as the use of inverse DRL in defining reward function, lack of standardization in retinal imaging, and data availability, will also be addressed to bridge gaps for its transition into clinical use.

Deep Learning in Optical Coherence Tomography Angiography: Current Progress, Challenges, and Future Directions.

Optical coherence tomography angiography (OCT-A) provides depth-resolved visualization of the retinal microvasculature without intravenous dye injection. It facilitates investigations of various retinal vascular diseases and glaucoma by assessment of qualitative and quantitative microvascular changes in the different retinal layers and radial peripapillary layer non-invasively, individually, and efficiently. Deep learning (DL), a subset of artificial intelligence (AI) based on deep neural networks, has been applied in OCT-A image analysis in recent years and achieved good performance for different tasks, such as image quality control, segmentation, and classification. DL technologies have further facilitated the potential implementation of OCT-A in eye clinics in an automated and efficient manner and enhanced its clinical values for detecting and evaluating various vascular retinopathies. Nevertheless, the deployment of this combination in real-world clinics is still in the "proof-of-concept" stage due to several limitations, such as small training sample size, lack of standardized data preprocessing, insufficient testing in external datasets, and absence of standardized results interpretation. In this review, we introduce the existing applications of DL in OCT-A, summarize the potential challenges of the clinical deployment, and discuss future research directions.

Clinically relevant factors associated with a binary outcome of diabetic macular ischaemia: an OCTA study.

AIMS: We investigated the demographic, ocular, diabetes-related and systemic factors associated with a binary outcome of diabetic macular ischaemia (DMI) as assessed by optical coherence tomography angiography (OCTA) evaluation of non-perfusion at the level of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) in a cohort of patients with diabetes mellitus (DM). MATERIALS AND METHODS: 617 patients with DM were recruited from July 2015 to December 2020 at the Chinese University of Hong Kong Eye Centre. Image quality assessment (gradable or ungradable for assessing DMI) and DMI evaluation (presence or absence of DMI) were assessed at the level of the SCP and DCP by OCTA. RESULTS: 1107 eyes from 593 subjects were included in the final analysis. 560 (50.59%) eyes had DMI at the level of SCP, and 647 (58.45%) eyes had DMI at the level of DCP. Among eyes without diabetic retinopathy (DR), DMI was observed in 19.40% and 24.13% of eyes at SCP and DCP, respectively. In the multivariable logistic regression models, older age, poorer visual acuity, thinner ganglion cell-inner plexiform layer thickness, worsened DR severity, higher haemoglobin A1c level, lower estimated glomerular filtration rate and higher low-density lipoprotein cholesterol level were associated with SCP-DMI. In addition to the aforementioned factors, presence of diabetic macular oedema and shorter axial length were associated with DCP-DMI. CONCLUSION: We reported a series of associated factors of SCP-DMI and DCP-DMI. The binary outcome of DMI might promote a simplified OCTA-based DMI evaluation before subsequent quantitative analysis for assessing DMI extent and fulfil the urge for an updating diabetic retinal disease staging to be implemented with OCTA.

Federated Learning in Ocular Imaging: Current Progress and Future Direction.

Advances in artificial intelligence deep learning (DL) have made tremendous impacts on the field of ocular imaging over the last few years. Specifically, DL has been utilised to detect and classify various ocular diseases on retinal photographs, optical coherence tomography (OCT) images, and OCT-angiography images. In order to achieve good robustness and generalisability of model performance, DL training strategies traditionally require extensive and diverse training datasets from various sites to be transferred and pooled into a "centralised location". However, such a data transferring process could raise practical concerns related to data security and patient privacy. Federated learning (FL) is a distributed collaborative learning paradigm which enables the coordination of multiple collaborators without the need for sharing confidential data. This distributed training approach has great potential to ensure data privacy among different institutions and reduce the potential risk of data leakage from data pooling or centralisation. This review article aims to introduce the concept of FL, provide current evidence of FL in ocular imaging, and discuss potential challenges as well as future applications.

A deep learning model for detection of Alzheimer's disease based on retinal photographs: a retrospective, multicentre case-control study.

BACKGROUND: There is no simple model to screen for Alzheimer's disease, partly because the diagnosis of Alzheimer's disease itself is complex-typically involving expensive and sometimes invasive tests not commonly available outside highly specialised clinical settings. We aimed to develop a deep learning algorithm that could use retinal photographs alone, which is the most common method of non-invasive imaging the retina to detect Alzheimer's disease-dementia. METHODS: In this retrospective, multicentre case-control study, we trained, validated, and tested a deep learning algorithm to detect Alzheimer's disease-dementia from retinal photographs using retrospectively collected data from 11 studies that recruited patients with Alzheimer's disease-dementia and people without disease from different countries. Our main aim was to develop a bilateral model to detect Alzheimer's disease-dementia from retinal photographs alone. We designed and internally validated the bilateral deep learning model using retinal photographs from six studies. We used the EfficientNet-b2 network as the backbone of the model to extract features from the images. Integrated features from four retinal photographs (optic nerve head-centred and macula-centred fields from both eyes) for each individual were used to develop supervised deep learning models and equip the network with unsupervised domain adaptation technique, to address dataset discrepancy between the different studies. We tested the trained model using five other studies, three of which used PET as a biomarker of significant amyloid ß burden (testing the deep learning model between amyloid ß positive vs amyloid ß negative). FINDINGS: 12 949 retinal photographs from 648 patients with Alzheimer's disease and 3240 people without the disease were used to train, validate, and test the deep learning model. In the internal validation dataset, the deep learning model had 83·6% (SD 2·5) accuracy, 93·2% (SD 2·2) sensitivity, 82·0% (SD 3·1) specificity, and an area under the receiver operating characteristic curve (AUROC) of 0·93 (0·01) for detecting Alzheimer's disease-dementia. In the testing datasets, the bilateral deep learning model had accuracies ranging from 79·6% (SD 15·5) to 92·1% (11·4) and AUROCs ranging from 0·73 (SD 0·24) to 0·91 (0·10). In the datasets with data on PET, the model was able to differentiate between participants who were amyloid ß positive and those who were amyloid ß negative: accuracies ranged from 80·6 (SD 13·4%) to 89·3 (13·7%) and AUROC ranged from 0·68 (SD 0·24) to 0·86 (0·16). In subgroup analyses, the discriminative performance of the model was improved in patients with eye disease (accuracy 89·6% [SD 12·5%]) versus those without eye disease (71·7% [11·6%]) and patients with diabetes (81·9% [SD 20·3%]) versus those without the disease (72·4% [11·7%]). INTERPRETATION: A retinal photograph-based deep learning algorithm can detect Alzheimer's disease with good accuracy, showing its potential for screening Alzheimer's disease in a community setting. FUNDING: BrightFocus Foundation.

Effect of Text Messaging Parents of School-Aged Children on Outdoor Time to Control Myopia: A Randomized Clinical Trial.

Importance: Myopia in school-aged children is a public health issue worldwide; consequently, effective interventions to prevent onset and progression are required. Objective: To investigate whether SMS text messages to parents increase light exposure and time outdoors in school-aged children and provide effective myopia control. Design, Setting, and Participants: This randomized clinical trial was conducted in China from May 2017 to May 2018, with participants observed for 3 years. Of 528 965 primary school-aged children from Anyang, 3113 were randomly selected. Of these, 268 grade 2 schoolchildren were selected and randomly assigned to SMS and control groups. Data were analyzed from June to December 2021. Interventions: Parents of children in the SMS group were sent text messages twice daily for 1 year to take their children outdoors. All children wore portable light meters to record light exposure on 3 randomly selected days (2 weekdays and 1 weekend day) before and after the intervention. Main Outcomes and Measures: The co-primary outcomes were change in axial length (axial elongation) and change in spherical equivalent refraction (myopic shift) from baseline as measured at the end of the intervention and 3 years later. A secondary outcome was myopia prevalence. Results: Of 268 grade 2 schoolchildren, 121 (45.1%) were girls, and the mean (SD) age was 8.4 (0.3) years. Compared with the control group, the SMS intervention group demonstrated greater light exposure and higher time outdoors during weekends, and the intervention had significant effect on axial elongation (coefficient, 0.09; 95% CI, 0.02-0.17; P = .01). Axial elongation was lower in the SMS group than in the control group during the intervention (0.27 mm [95% CI, 0.24-0.30] vs 0.31 mm [95% CI, 0.29-0.34]; P = .03) and at year 2 (0.39 mm [95% CI, 0.35-0.42] vs 0.46 mm [95% CI, 0.42-0.50]; P = .009) and year 3 (0.30 mm [95% CI, 0.27-0.33] vs 0.35 mm [95% CI, 0.33-0.37]; P = .005) after the intervention. Myopic shift was lower in the SMS group than in the control group at year 2 (-0.69 diopters [D] [95% CI, -0.78 to -0.60] vs -0.82 D [95% CI, -0.91 to -0.73]; P = .04) and year 3 (-0.47 D [95% CI, -0.54 to -0.39] vs -0.60 D [95% CI, -0.67 to -0.53]; P = .01) after the intervention, as was myopia prevalence (year 2: 38.3% [51 of 133] vs 51.1% [68 of 133]; year 3: 46.6% [62 of 133] vs 65.4% [87 of 133]). Conclusions and Relevance: In this randomized clinical trial, SMS text messages to parents resulted in lower axial elongation and myopia progression in schoolchildren over 3 years, possibly through increased outdoor time and light exposure, showing promise for reducing myopia prevalence. Trial Registration: Chinese Clinical Trial Registry Identifier: ChiCTR-IOC-17010525.

Three-Dimensional Multi-Task Deep Learning Model to Detect Glaucomatous Optic Neuropathy and Myopic Features From Optical Coherence Tomography Scans: A Retrospective Multi-Centre Study.

Purpose: We aim to develop a multi-task three-dimensional (3D) deep learning (DL) model to detect glaucomatous optic neuropathy (GON) and myopic features (MF) simultaneously from spectral-domain optical coherence tomography (SDOCT) volumetric scans. Methods: Each volumetric scan was labelled as GON according to the criteria of retinal nerve fibre layer (RNFL) thinning, with a structural defect that correlated in position with the visual field defect (i.e., reference standard). MF were graded by the SDOCT en face images, defined as presence of peripapillary atrophy (PPA), optic disc tilting, or fundus tessellation. The multi-task DL model was developed by ResNet with output of Yes/No GON and Yes/No MF. SDOCT scans were collected in a tertiary eye hospital (Hong Kong SAR, China) for training (80%), tuning (10%), and internal validation (10%). External testing was performed on five independent datasets from eye centres in Hong Kong, the United States, and Singapore, respectively. For GON detection, we compared the model to the average RNFL thickness measurement generated from the SDOCT device. To investigate whether MF can affect the model's performance on GON detection, we conducted subgroup analyses in groups stratified by Yes/No MF. The area under the receiver operating characteristic curve (AUROC), sensitivity, specificity, and accuracy were reported. Results: A total of 8,151 SDOCT volumetric scans from 3,609 eyes were collected. For detecting GON, in the internal validation, the proposed 3D model had significantly higher AUROC (0.949 vs. 0.913, p < 0.001) than average RNFL thickness in discriminating GON from normal. In the external testing, the two approaches had comparable performance. In the subgroup analysis, the multi-task DL model performed significantly better in the group of "no MF" (0.883 vs. 0.965, p-value < 0.001) in one external testing dataset, but no significant difference in internal validation and other external testing datasets. The multi-task DL model's performance to detect MF was also generalizable in all datasets, with the AUROC values ranging from 0.855 to 0.896. Conclusion: The proposed multi-task 3D DL model demonstrated high generalizability in all the datasets and the presence of MF did not affect the accuracy of GON detection generally.

Deep Learning for Glaucoma Detection and Identification of Novel Diagnostic Areas in Diverse Real-World Datasets.

Purpose: To develop a three-dimensional (3D) deep learning algorithm to detect glaucoma using spectral-domain optical coherence tomography (SD-OCT) optic nerve head (ONH) cube scans and validate its performance on ethnically diverse real-world datasets and on cropped ONH scans. Methods: In total, 2461 Cirrus SD-OCT ONH scans of 1012 eyes were obtained from the Glaucoma Clinic Imaging Database at the Byers Eye Institute, Stanford University, from March 2010 to December 2017. A 3D deep neural network was trained and tested on this unique raw OCT cube dataset to identify a multimodal definition of glaucoma excluding other concomitant retinal disease and optic neuropathies. A total of 1022 scans of 363 glaucomatous eyes (207 patients) and 542 scans of 291 normal eyes (167 patients) from Stanford were included in training, and 142 scans of 48 glaucomatous eyes (27 patients) and 61 scans of 39 normal eyes (23 patients) were included in the validation set. A total of 3371 scans (Cirrus SD-OCT) from four different countries were used for evaluation of the model: the non overlapping test dataset from Stanford (USA) consisted of 694 scans: 241 scans from 113 normal eyes of 66 patients and 453 scans of 157 glaucomatous eyes of 89 patients. The datasets from Hong Kong (total of 1625 scans; 666 OCT scans from 196 normal eyes of 99 patients and 959 scans of 277 glaucomatous eyes of 155 patients), India (total of 672 scans; 211 scans from 147 normal eyes of 98 patients and 461 scans from 171 glaucomatous eyes of 101 patients), and Nepal (total of 380 scans; 158 scans from 143 normal eyes of 89 patients and 222 scans from 174 glaucomatous eyes of 109 patients) were used for external evaluation. The performance of the model was then evaluated on manually cropped scans from Stanford using a new algorithm called DiagFind. The ONH region was cropped by identifying the appropriate zone of the image in the expected location relative to Bruch's Membrane Opening (BMO) using a commercially available imaging software. Subgroup analyses were performed in groups stratified by eyes, myopia severity of glaucoma, and on a set of glaucoma cases without field defects. Saliency maps were generated to highlight the areas the model used to make a prediction. The model's performance was compared to that of a glaucoma specialist using all available information on a subset of cases. Results: The 3D deep learning system achieved area under the curve (AUC) values of 0.91 (95% CI, 0.90-0.92), 0.80 (95% CI, 0.78-0.82), 0.94 (95% CI, 0.93-0.96), and 0.87 (95% CI, 0.85-0.90) on Stanford, Hong Kong, India, and Nepal datasets, respectively, to detect perimetric glaucoma and AUC values of 0.99 (95% CI, 0.97-1.00), 0.96 (95% CI, 0.93-1.00), and 0.92 (95% CI, 0.89-0.95) on severe, moderate, and mild myopia cases, respectively, and an AUC of 0.77 on cropped scans. The model achieved an AUC value of 0.92 (95% CI, 0.90-0.93) versus that of the human grader with an AUC value of 0.91 on the same subset of scans (\(P=0.99\)). The performance of the model in terms of recall on glaucoma cases without field defects was found to be 0.76 (0.68-0.85). Saliency maps highlighted the lamina cribrosa in glaucomatous eyes versus superficial retina in normal eyes as the regions associated with classification. Conclusions: A 3D convolutional neural network (CNN) trained on SD-OCT ONH cubes can distinguish glaucoma from normal cases in diverse datasets obtained from four different countries. The model trained on additional random cropping data augmentation performed reasonably on manually cropped scans, indicating the importance of lamina cribrosa in glaucoma detection. Translational Relevance: A 3D CNN trained on SD-OCT ONH cubes was developed to detect glaucoma in diverse datasets obtained from four different countries and on cropped scans. The model identified lamina cribrosa as the region associated with glaucoma detection.

Brain Activation Induced by Myopic and Hyperopic Defocus From Spectacles.

Purpose: To assess neural changes in perceptual effects induced by myopic defocus and hyperopic defocus stimuli in ametropic and emmetropic subjects using functional magnetic resonance imaging (fMRI). Methods: This study included 41 subjects with a mean age of 26.0 ± 2.9 years. The mean spherical equivalence refraction was -0.54 ± 0.51D in the emmetropic group and -3.57 ± 2.27D in the ametropic group. The subjects were instructed to view through full refractive correction, with values of +2.00D to induce myopic defocus state and -2.00D to induce hyperopic defocus state. This was carried over in three random sessions. Arterial spin labeling (ASL) perfusion was measured using fMRI to obtain quantified regional cerebral blood flow (rCBF). Behavioral tests including distant visual acuity (VA) and contrast sensitivity (CS), were measured every 5 min for 30 min. Results: Myopic defocus induced significantly greater rCBF increase in four cerebral regions compared with full correction: right precentral gyrus, right superior temporal gyrus, left inferior parietal lobule, and left middle temporal gyrus (P < 0.001). The differences were less significant in low myopes than emmetropes. In the hyperopic defocus session, the increased responses of rCBF were only observed in the right and left precentral gyrus. Myopic defocused VA and CS improved significantly within 5 min and reached a plateau shortly after. Conclusion: This study revealed that myopic defocus stimuli can significantly increase blood perfusion in visual attention-related cerebral regions, which suggests a potential direction for future investigation on the relationship between retinal defocus and its neural consequences.

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

OBJECTIVE: Diabetic macular edema (DME) is the primary cause of vision loss among individuals with diabetes mellitus (DM). We developed, validated, and tested a deep learning (DL) system for classifying DME using images from three common commercially available optical coherence tomography (OCT) devices. RESEARCH DESIGN AND METHODS: We trained and validated two versions of a multitask convolution neural network (CNN) to classify DME (center-involved DME [CI-DME], non-CI-DME, or absence of DME) using three-dimensional (3D) volume scans and 2D B-scans, respectively. For both 3D and 2D CNNs, we used the residual network (ResNet) as the backbone. For the 3D CNN, we used a 3D version of ResNet-34 with the last fully connected layer removed as the feature extraction module. A total of 73,746 OCT images were used for training and primary validation. External testing was performed using 26,981 images across seven independent data sets from Singapore, Hong Kong, the U.S., China, and Australia. RESULTS: In classifying the presence or absence of DME, the DL system achieved area under the receiver operating characteristic curves (AUROCs) of 0.937 (95% CI 0.920-0.954), 0.958 (0.930-0.977), and 0.965 (0.948-0.977) for the primary data set obtained from CIRRUS, SPECTRALIS, and Triton OCTs, respectively, in addition to AUROCs >0.906 for the external data sets. For further classification of the CI-DME and non-CI-DME subgroups, the AUROCs were 0.968 (0.940-0.995), 0.951 (0.898-0.982), and 0.975 (0.947-0.991) for the primary data set and >0.894 for the external data sets. CONCLUSIONS: We demonstrated excellent performance with a DL system for the automated classification of DME, highlighting its potential as a promising second-line screening tool for patients with DM, which may potentially create a more effective triaging mechanism to eye clinics.

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

PURPOSE: To develop a deep learning (DL) system that can detect referable diabetic retinopathy (RDR) and vision-threatening diabetic retinopathy (VTDR) from images obtained on ultra-widefield scanning laser ophthalmoscope (UWF-SLO). DESIGN: Observational, cross-sectional study. PARTICIPANTS: A total of 9392 UWF-SLO images of 1903 eyes from 1022 subjects with diabetes from Hong Kong, the United Kingdom, India, and Argentina. METHODS: All images were labeled according to the presence or absence of RDR and the presence or absence of VTDR. Labeling was performed by retina specialists from fundus examination, according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Three convolutional neural networks (ResNet50) were trained with a transfer-learning procedure for assessing gradability and identifying VTDR and RDR. External validation was performed on 4 datasets spanning different geographical regions. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (AUROC); area under the precision-recall curve (AUPRC); sensitivity, specificity, and accuracy of the DL system in gradability assessment; and detection of RDR and VTDR. RESULTS: For gradability assessment, the system achieved an AUROC of 0.923 (95% confidence interval [CI], 0.892-0.947), sensitivity of 86.5% (95% CI, 77.6-92.8), and specificity of 82.1% (95% CI, 77.3-86.2) for the primary validation dataset, and >0.82 AUROCs, >79.6% sensitivity, and >70.4% specificity for the geographical external validation datasets. For detecting RDR and VTDR, the AUROCs were 0.981 (95% CI, 0.977-0.984) and 0.966 (95% CI, 0.961-0.971), with sensitivities of 94.9% (95% CI, 92.3-97.9) and 87.2% (95% CI, 81.5-91.6), specificities of 95.1% (95% CI, 90.6-97.9) and 95.8% (95% CI, 93.3-97.6), and positive predictive values (PPVs) of 98.0% (95% CI, 96.1-99.0) and 91.1% (95% CI, 86.3-94.3) for the primary validation dataset, respectively. The AUROCs and accuracies for detecting both RDR and VTDR were >0.9% and >80%, respectively, for the geographical external validation datasets. The AUPRCs were >0.9, and sensitivities, specificities, and PPVs were >80% for the geographical external validation datasets for RDR and VTDR detection. CONCLUSIONS: The excellent performance achieved with this DL system for image quality assessment and detection of RDR and VTDR in UWF-SLO images highlights its potential as an efficient and effective diabetic retinopathy screening tool.

Deep learning in glaucoma with optical coherence tomography: a review.

Deep learning (DL), a subset of artificial intelligence (AI) based on deep neural networks, has made significant breakthroughs in medical imaging, particularly for image classification and pattern recognition. In ophthalmology, applying DL for glaucoma assessment with optical coherence tomography (OCT), including OCT traditional reports, two-dimensional (2D) B-scans, and three-dimensional (3D) volumetric scans, has increasingly raised research interests. Studies have demonstrated that using DL for interpreting OCT is efficient, accurate, and with good performance for discriminating glaucomatous eyes from normal eyes, suggesting that incorporation of DL technology in OCT for glaucoma assessment could potentially address some gaps in the current practice and clinical workflow. However, further research is crucial in tackling some existing challenges, such as annotation standardization (i.e., setting a standard for ground truth labelling among different studies), development of DL-powered IT infrastructure for real-world implementation, prospective validation in unseen datasets for further evaluation of generalizability, cost-effectiveness analysis after integration of DL, the AI "black box" explanation problem. This review summarizes recent studies on the application of DL on OCT for glaucoma assessment, identifies the potential clinical impact arising from the development and deployment of the DL models, and discusses future research directions.

A 3D Deep Learning System for Detecting Referable Glaucoma Using Full OCT Macular Cube Scans.

Purpose: The purpose of this study was to develop a 3D deep learning system from spectral domain optical coherence tomography (SD-OCT) macular cubes to differentiate between referable and nonreferable cases for glaucoma applied to real-world datasets to understand how this would affect the performance. Methods: There were 2805 Cirrus optical coherence tomography (OCT) macula volumes (Macula protocol 512 × 128) of 1095 eyes from 586 patients at a single site that were used to train a fully 3D convolutional neural network (CNN). Referable glaucoma included true glaucoma, pre-perimetric glaucoma, and high-risk suspects, based on qualitative fundus photographs, visual fields, OCT reports, and clinical examinations, including intraocular pressure (IOP) and treatment history as the binary (two class) ground truth. The curated real-world dataset did not include eyes with retinal disease or nonglaucomatous optic neuropathies. The cubes were first homogenized using layer segmentation with the Orion Software (Voxeleron) to achieve standardization. The algorithm was tested on two separate external validation sets from different glaucoma studies, comprised of Cirrus macular cube scans of 505 and 336 eyes, respectively. Results: The area under the receiver operating characteristic (AUROC) curve for the development dataset for distinguishing referable glaucoma was 0.88 for our CNN using homogenization, 0.82 without homogenization, and 0.81 for a CNN architecture from the existing literature. For the external validation datasets, which had different glaucoma definitions, the AUCs were 0.78 and 0.95, respectively. The performance of the model across myopia severity distribution has been assessed in the dataset from the United States and was found to have an AUC of 0.85, 0.92, and 0.95 in the severe, moderate, and mild myopia, respectively. Conclusions: A 3D deep learning algorithm trained on macular OCT volumes without retinal disease to detect referable glaucoma performs better with retinal segmentation preprocessing and performs reasonably well across all levels of myopia. Translational Relevance: Interpretation of OCT macula volumes based on normative data color distributions is highly influenced by population demographics and characteristics, such as refractive error, as well as the size of the normative database. Referable glaucoma, in this study, was chosen to include cases that should be seen by a specialist. This study is unique because it uses multimodal patient data for the glaucoma definition, and includes all severities of myopia as well as validates the algorithm with international data to understand generalizability potential.

Towards multi-center glaucoma OCT image screening with semi-supervised joint structure and function multi-task learning.

Glaucoma is the leading cause of irreversible blindness in the world. Structure and function assessments play an important role in diagnosing glaucoma. Nowadays, Optical Coherence Tomography (OCT) imaging gains increasing popularity in measuring the structural change of eyes. However, few automated methods have been developed based on OCT images to screen glaucoma. In this paper, we are the first to unify the structure analysis and function regression to distinguish glaucoma patients from normal controls effectively. Specifically, our method works in two steps: a semi-supervised learning strategy with smoothness assumption is first applied for the surrogate assignment of missing function regression labels. Subsequently, the proposed multi-task learning network is capable of exploring the structure and function relationship between the OCT image and visual field measurement simultaneously, which contributes to classification performance improvement. It is also worth noting that the proposed method is assessed by two large-scale multi-center datasets. In other words, we first build the largest glaucoma OCT image dataset (i.e., HK dataset) involving 975,400 B-scans from 4,877 volumes to develop and evaluate the proposed method, then the model without further fine-tuning is directly applied on another independent dataset (i.e., Stanford dataset) containing 246,200 B-scans from 1,231 volumes. Extensive experiments are conducted to assess the contribution of each component within our framework. The proposed method outperforms the baseline methods and two glaucoma experts by a large margin, achieving volume-level Area Under ROC Curve (AUC) of 0.977 on HK dataset and 0.933 on Stanford dataset, respectively. The experimental results indicate the great potential of the proposed approach for the automated diagnosis system.