Remote Perimetry in a Virtual Reality Metaverse Environment for Out-of-Hospital Functional Eye Screening Compared Against the Gold Standard Humphrey Visual Fields Perimeter: Proof-of-Concept Pilot Study.

BACKGROUND: The growing global burden of visual impairment necessitates better population eye screening for early detection of eye diseases. However, accessibility to testing is often limited and centralized at in-hospital settings. Furthermore, many eye screening programs were disrupted by the COVID-19 pandemic, presenting an urgent need for out-of-hospital solutions. OBJECTIVE: This study investigates the performance of a novel remote perimetry application designed in a virtual reality metaverse environment to enable functional testing in community-based and primary care settings. METHODS: This was a prospective observational study investigating the performance of a novel remote perimetry solution in comparison with the gold standard Humphrey visual field (HVF) perimeter. Subjects received a comprehensive ophthalmologic assessment, HVF perimetry, and remote perimetry testing. The primary outcome measure was the agreement in the classification of overall perimetry result normality by the HVF (Swedish interactive threshold algorithm-fast) and testing with the novel algorithm. Secondary outcome measures included concordance of individual testing points and perimetry topographic maps. RESULTS: We recruited 10 subjects with an average age of 59.6 (range 28-81) years. Of these, 7 (70%) were male and 3 (30%) were female. The agreement in the classification of overall perimetry results was high (9/10, 90%). The pointwise concordance in the automated classification of individual test points was 83.3% (8.2%; range 75%-100%). In addition, there was good perimetry topographic concordance with the HVF in all subjects. CONCLUSIONS: Remote perimetry in a metaverse environment had good concordance with gold standard perimetry using the HVF and could potentially avail functional eye screening in out-of-hospital settings.

Retinal neural dysfunction in diabetes revealed with handheld chromatic pupillometry.

BACKGROUND: To evaluate the ability of handheld chromatic pupillometry to reveal and localise retinal neural dysfunction in diabetic patients with and without diabetic retinopathy (DR). METHODS: This cross-sectional study included 82 diabetics (DM) and 93 controls (60.4 ± 8.4 years, 44.1% males). DM patients included those without (n = 25, 64.7 ± 6.3 years, 44.0% males) and with DR (n = 57, 60.3 ± 8.5 years, 64.9% males). Changes in horizontal pupil radius in response to blue (469 nm) and red (640 nm) light stimuli were assessed monocularly, in clinics, using a custom-built handheld pupillometer. Pupillometric parameters (phasic constriction amplitudes [predominantly from the outer retina], maximal constriction amplitudes [from the inner and outer retina] and post-illumination pupillary responses [PIPRs; predominantly from the inner retina]) were extracted from baseline-adjusted pupillary light response traces and compared between controls, DM without DR, and DR. Net PIPR was defined as the difference between blue and red PIPRs. RESULTS: Phasic constriction amplitudes to blue and red lights were decreased in DR compared to controls (p < 0.001; p < 0.001). Maximal constriction amplitudes to blue and red lights were decreased in DR compared to DM without DR (p < 0.001; p = 0.02), and in DM without DR compared to controls (p < 0.001; p = 0.005). Net PIPR was decreased in both DR and DM without DR compared to controls (p = 0.02; p = 0.03), suggesting a wavelength-dependent (and hence retinal) pupillometric dysfunction in diabetic patients with or without DR. CONCLUSIONS: Handheld chromatic pupillometry can reveal retinal neural dysfunction in diabetes, even without DR. Patients with DM but no DR displayed primarily inner retinal dysfunction, while patients with DR showed both inner and outer retinal dysfunction.

Global Prevalence of Diabetic Retinopathy and Projection of Burden through 2045: Systematic Review and Meta-analysis.

TOPIC: To provide updated estimates on the global prevalence and number of people with diabetic retinopathy (DR) through 2045. CLINICAL RELEVANCE: The International Diabetes Federation (IDF) estimated the global population with diabetes mellitus (DM) to be 463 million in 2019 and 700 million in 2045. Diabetic retinopathy remains a common complication of DM and a leading cause of preventable blindness in the adult working population. METHODS: We conducted a systematic review using PubMed, Medline, Web of Science, and Scopus for population-based studies published up to March 2020. Random effect meta-analysis with logit transformation was performed to estimate global and regional prevalence of DR, vision-threatening DR (VTDR), and clinically significant macular edema (CSME). Projections of DR, VTDR, and CSME burden were based on population data from the IDF Atlas 2019. RESULTS: We included 59 population-based studies. Among individuals with diabetes, global prevalence was 22.27% (95% confidence interval [CI], 19.73%-25.03%) for DR, 6.17% (95% CI, 5.43%-6.98%) for VTDR, and 4.07% (95% CI, 3.42%-4.82%) for CSME. In 2020, the number of adults worldwide with DR, VTDR, and CSME was estimated to be 103.12 million, 28.54 million, and 18.83 million, respectively; by 2045, the numbers are projected to increase to 160.50 million, 44.82 million, and 28.61 million, respectively. Diabetic retinopathy prevalence was highest in Africa (35.90%) and North American and the Caribbean (33.30%) and was lowest in South and Central America (13.37%). In meta-regression models adjusting for habitation type, response rate, study year, and DR diagnostic method, Hispanics (odds ratio [OR], 2.92; 95% CI, 1.22-6.98) and Middle Easterners (OR, 2.44; 95% CI, 1.51-3.94) with diabetes were more likely to have DR compared with Asians. DISCUSSION: The global DR burden is expected to remain high through 2045, disproportionately affecting countries in the Middle East and North Africa and the Western Pacific. These updated estimates may guide DR screening, treatment, and public health care strategies.

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.

Generative adversarial networks in ophthalmology: what are these and how can they be used?

PURPOSE OF REVIEW: The development of deep learning (DL) systems requires a large amount of data, which may be limited by costs, protection of patient information and low prevalence of some conditions. Recent developments in artificial intelligence techniques have provided an innovative alternative to this challenge via the synthesis of biomedical images within a DL framework known as generative adversarial networks (GANs). This paper aims to introduce how GANs can be deployed for image synthesis in ophthalmology and to discuss the potential applications of GANs-produced images. RECENT FINDINGS: Image synthesis is the most relevant function of GANs to the medical field, and it has been widely used for generating 'new' medical images of various modalities. In ophthalmology, GANs have mainly been utilized for augmenting classification and predictive tasks, by synthesizing fundus images and optical coherence tomography images with and without pathologies such as age-related macular degeneration and diabetic retinopathy. Despite their ability to generate high-resolution images, the development of GANs remains data intensive, and there is a lack of consensus on how best to evaluate the outputs produced by GANs. SUMMARY: Although the problem of artificial biomedical data generation is of great interest, image synthesis by GANs represents an innovation with yet unclear relevance for ophthalmology.

Vision Impairment in CKD Patients: Epidemiology, Mechanisms, Differential Diagnoses, and Prevention.

Eyes and kidneys have numerous structural, developmental, physiologic, and pathogenic pathways in common, suggesting that many kidney and eye diseases may be interlinked. Studies suggest that the prevalence of eye diseases and vision impairment are higher among persons with end-stage kidney disease and earlier stages of chronic kidney disease (CKD) than in those without. Ocular morbidity in persons with CKD and end-stage kidney disease may be due to the following risk factors: (1) underlying conditions and risk factors for CKD such as diabetes or hypertension, (2) metabolic disorders associated with CKD, (3) uremia and anemia, and (4) CKD treatment. Among the chief eye diseases, diabetic retinopathy and age-related macular degeneration are most consistently associated with CKD. Further research for eye diseases such as glaucoma and cataract is needed to determine their relationships with CKD. Despite the high prevalence and burden of vision impairment among persons with CKD, eye screening in patients with CKD is not currently recommended as standard practice. This review suggests that patients with CKD should be encouraged to undergo a complete eye examination. Furthermore, physicians should be aware that patients undergoing dialysis may develop acute eye problems such as acute glaucoma, and appropriate referral to ophthalmologists should be considered in those with a history of glaucoma or recent ocular surgery. Interdisciplinary collaboration between nephrologists and ophthalmologists will ensure enhanced and appropriate management of patients with CKD.

A novel model of persistent retinal neovascularization for the development of sustained anti-VEGF therapies.

Anti-vascular endothelial growth factor (VEGF) therapies lead to a major breakthrough in treatment of neovascular retinal diseases such as age-related macular degeneration or diabetic retinopathy. Current management of these conditions require regular and frequent intravitreal injections to prevent disease recurrence once the effect of the injected drug wears off. This has led to a pressing clinical need of developing sustained release formulations or therapies with longer duration. A major drawback in developing such therapies is that the currently available animal models show spontaneous regression of vascular leakage. They therefore not only fail to recapitulate retinal vascular disease in humans, but also prevent to discern if regression is due to prolonged therapeutic effect or simply reflects spontaneous healing. Here, we described the development of a novel rabbit model of persistent retinal neovascularization (PRNV). Retinal Müller glial are essential for maintaining the integrity of the blood-retinal barrier. Intravitreal injection of DL-alpha-aminoadipic acid (DL-AAA), a selective retinal glial (Müller) cell toxin, results in persistent vascular leakage for up to 48 weeks. We demonstrated that VEGF concentrations were significantly increased in vitreous suggesting VEGF plays a significant role in mediating the leakage observed. Intravitreal administration of anti-VEGF drugs (e.g. bevacizumab, ranibizumab and aflibercept) suppresses vascular leakage for 8-10 weeks, before recurrence of leakage to pre-treatment levels. All three anti-VEGF drugs are very effective in re-ducing angiographic leakage in PRNV model, and aflibercept demonstrated a longer duration of action compared with the others, reminiscent of what is observed with these drugs in human in the clinical setting. Therefore, this model provides a unique tool to evaluate novel anti-VEGF formulations and therapies with respect to their duration of action in comparison to the currently used drugs.

COMPARISON OF OPTICAL COHERENCE TOMOGRAPHY ANGIOGRAPHIC CHANGES AFTER ANTI-VASCULAR ENDOTHELIAL GROWTH FACTOR THERAPY ALONE OR IN COMBINATION WITH PHOTODYNAMIC THERAPY IN POLYPOIDAL CHOROIDAL VASCULOPATHY.

PURPOSE: To compare changes in optical coherence tomography angiography in eyes with polypoidal choroidal vasculopathy after treatment with anti-vascular endothelial growth factor monotherapy or combined with photodynamic therapy. METHODS: This is a longitudinal case-controlled study. The authors performed optical coherence tomography angiography at baseline and Month 3 in patients with treatment-naive polypoidal choroidal vasculopathy undergoing monotherapy (n = 10) or combination therapy (n = 13). We analyzed flow signal within the outer retina and choriocapillaris using automated segmentation. The authors analyzed the presence of pachyvessels using a 10.4-µm segment through Haller layer. The changes in each layer were compared between treatments. RESULTS: At Month 3, both groups showed similar improvement in best-corrected visual acuity and central retinal thickness. However, flow signal within the polypoidal choroidal vasculopathy complex was decreased in more eyes after combination therapy than after monotherapy (84.6% vs. 40.0%, P = 0.04). Patchy reduction in flow signal within the choriocapillaris layer was noted in 15.4% and 10.0% after combination therapy and monotherapy, respectively (P = 0.61). Significant reduction in pachyvessel caliber was seen only after combination therapy but not after monotherapy (75.0% vs. 0.0%, P = 0.01). CONCLUSION: Longitudinal optical coherence tomography angiography demonstrates more significant reduction in lesion flow and pachyvessels in the short term after combination therapy than after monotherapy, although visual and structural OCT showed similar improvement.

Development and Validation of a Deep Learning System for Diabetic Retinopathy and Related Eye Diseases Using Retinal Images From Multiethnic Populations With Diabetes.

Importance: A deep learning system (DLS) is a machine learning technology with potential for screening diabetic retinopathy and related eye diseases. Objective: To evaluate the performance of a DLS in detecting referable diabetic retinopathy, vision-threatening diabetic retinopathy, possible glaucoma, and age-related macular degeneration (AMD) in community and clinic-based multiethnic populations with diabetes. Design, Setting, and Participants: Diagnostic performance of a DLS for diabetic retinopathy and related eye diseases was evaluated using 494 661 retinal images. A DLS was trained for detecting diabetic retinopathy (using 76 370 images), possible glaucoma (125 189 images), and AMD (72 610 images), and performance of DLS was evaluated for detecting diabetic retinopathy (using 112 648 images), possible glaucoma (71 896 images), and AMD (35 948 images). Training of the DLS was completed in May 2016, and validation of the DLS was completed in May 2017 for detection of referable diabetic retinopathy (moderate nonproliferative diabetic retinopathy or worse) and vision-threatening diabetic retinopathy (severe nonproliferative diabetic retinopathy or worse) using a primary validation data set in the Singapore National Diabetic Retinopathy Screening Program and 10 multiethnic cohorts with diabetes. Exposures: Use of a deep learning system. Main Outcomes and Measures: Area under the receiver operating characteristic curve (AUC) and sensitivity and specificity of the DLS with professional graders (retinal specialists, general ophthalmologists, trained graders, or optometrists) as the reference standard. Results: In the primary validation dataset (n = 14 880 patients; 71 896 images; mean [SD] age, 60.2 [2.2] years; 54.6% men), the prevalence of referable diabetic retinopathy was 3.0%; vision-threatening diabetic retinopathy, 0.6%; possible glaucoma, 0.1%; and AMD, 2.5%. The AUC of the DLS for referable diabetic retinopathy was 0.936 (95% CI, 0.925-0.943), sensitivity was 90.5% (95% CI, 87.3%-93.0%), and specificity was 91.6% (95% CI, 91.0%-92.2%). For vision-threatening diabetic retinopathy, AUC was 0.958 (95% CI, 0.956-0.961), sensitivity was 100% (95% CI, 94.1%-100.0%), and specificity was 91.1% (95% CI, 90.7%-91.4%). For possible glaucoma, AUC was 0.942 (95% CI, 0.929-0.954), sensitivity was 96.4% (95% CI, 81.7%-99.9%), and specificity was 87.2% (95% CI, 86.8%-87.5%). For AMD, AUC was 0.931 (95% CI, 0.928-0.935), sensitivity was 93.2% (95% CI, 91.1%-99.8%), and specificity was 88.7% (95% CI, 88.3%-89.0%). For referable diabetic retinopathy in the 10 additional datasets, AUC range was 0.889 to 0.983 (n = 40 752 images). Conclusions and Relevance: In this evaluation of retinal images from multiethnic cohorts of patients with diabetes, the DLS had high sensitivity and specificity for identifying diabetic retinopathy and related eye diseases. Further research is necessary to evaluate the applicability of the DLS in health care settings and the utility of the DLS to improve vision outcomes.

Cost-effectiveness of a National Telemedicine Diabetic Retinopathy Screening Program in Singapore.

PURPOSE: To determine the incremental cost-effectiveness of a new telemedicine technician-based assessment relative to an existing model of family physician (FP)-based assessment of diabetic retinopathy (DR) in Singapore from the health system and societal perspectives. DESIGN: Model-based, cost-effectiveness analysis of the Singapore Integrated Diabetic Retinopathy Program (SiDRP). PARTICIPANTS: A hypothetical cohort of patients aged 55 years with type 2 diabetes previously not screened for DR. METHODS: The SiDRP is a new telemedicine-based DR screening program using trained technicians to assess retinal photographs. We compared the cost-effectiveness of SiDRP with the existing model in which FPs assess photographs. We developed a hybrid decision tree/Markov model to simulate the costs, effectiveness, and incremental cost-effectiveness ratio (ICER) of SiDRP relative to FP-based DR screening over a lifetime horizon. We estimated the costs from the health system and societal perspectives. Effectiveness was measured in terms of quality-adjusted life-years (QALYs). Result robustness was calculated using deterministic and probabilistic sensitivity analyses. MAIN OUTCOME MEASURES: The ICER. RESULTS: From the societal perspective that takes into account all costs and effects, the telemedicine-based DR screening model had significantly lower costs (total cost savings of S$173 per person) while generating similar QALYs compared with the physician-based model (i.e., 13.1 QALYs). From the health system perspective that includes only direct medical costs, the cost savings are S$144 per person. By extrapolating these data to approximately 170 000 patients with diabetes currently being screened yearly for DR in Singapore's primary care polyclinics, the present value of future cost savings associated with the telemedicine-based model is estimated to be S$29.4 million over a lifetime horizon. CONCLUSIONS: While generating similar health outcomes, the telemedicine-based DR screening using technicians in the primary care setting saves costs for Singapore compared with the FP model. Our data provide a strong economic rationale to expand the telemedicine-based DR screening program in Singapore and elsewhere.

Biomarkers of Diabetic Retinopathy.

Diabetic retinopathy (DR), a leading cause of acquired vision loss, is a microvascular complication of diabetes. While traditional risk factors for diabetic retinopathy including longer duration of diabetes, poor blood glucose control, and dyslipidemia are helpful in stratifying patient's risk for developing retinopathy, many patients without these traditional risk factors develop DR; furthermore, there are persons with long diabetes duration who do not develop DR. Thus, identifying biomarkers to predict DR or to determine therapeutic response is important. A biomarker can be defined as a characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic intervention. Incorporation of biomarkers into risk stratification of persons with diabetes would likely aid in early diagnosis and guide treatment methods for those with DR or with worsening DR. Systemic biomarkers of DR include serum measures including genomic, proteomic, and metabolomics biomarkers. Ocular biomarkers including tears and vitreous and retinal vascular structural changes have also been studied extensively to prognosticate the risk of DR development. The current studies on biomarkers are limited by the need for larger sample sizes, cross-validation in different populations and ethnic groups, and time-efficient and cost-effective analytical techniques. Future research is important to explore novel DR biomarkers that are non-invasive, rapid, economical, and accurate to help reduce the incidence and progression of DR in people with diabetes.

Transition probabilities of diabetic retinopathy and death in an Asian population with diabetes.

PURPOSE: To evaluate the dynamic transitions in diabetic retinopathy (DR) severity over time and associated risk factors in an Asian population with diabetes. DESIGN: Longitudinal cohort study METHODS: We analyzed data from 9481 adults in the Singapore Integrated Diabetic Retinopathy Screening Program (2010-2015) with linkage to death registry. A multistate Markov model adjusted for age, sex, systolic blood pressure (SBP), diabetes duration, HbA1c, and body mass index (BMI) was applied to estimate annual transition probabilities between four DR states (no, mild, moderate, and severe/proliferative) and death, and the mean sojourn time in each state. RESULTS: The median assessment interval was 12 months, with most patients having 3 assessments. Annual probabilities for DR progression (no-to-mild, mild-to-moderate and moderate-to-severe/proliferative) were 6.1 %, 7.0 % and 19.3 %, respectively; and for regression (mild-to-no, moderate-to-mild and severe-to-moderate) were 55.4 %, 17.3 % and 4.4 %, respectively. Annual mortality rates from each DR state were 1.2 %, 2.0 %, 18.7 %, and 30.0 %. The sojourn time in each state were 8.2, 0.8, 0.8 and 2.2 years. Higher HbA1c and SBP levels were associated with progression of no-mild and mild-moderate DR, and diabetes duration with no-to-mild and moderate-to-severe/proliferative DR. Lower HbA1c levels were associated with regression from mild-to-no and moderate-to-mild, and higher BMI with mild-to-no DR. CONCLUSIONS: Our results suggest a prolonged duration (∼8 years) in developing mild DR, with faster transitions (within a year) from mild or moderate states. Moderate/above DR greatly increases the probability of progression and death as compared to mild DR/below. HbA1c was associated with both progression as well as regression.

Geometric Correspondence-Based Multimodal Learning for Ophthalmic Image Analysis.

Color fundus photography (CFP) and Optical coherence tomography (OCT) images are two of the most widely used modalities in the clinical diagnosis and management of retinal diseases. Despite the widespread use of multimodal imaging in clinical practice, few methods for automated diagnosis of eye diseases utilize correlated and complementary information from multiple modalities effectively. This paper explores how to leverage the information from CFP and OCT images to improve the automated diagnosis of retinal diseases. We propose a novel multimodal learning method, named geometric correspondence-based multimodal learning network (GeCoM-Net), to achieve the fusion of CFP and OCT images. Specifically, inspired by clinical observations, we consider the geometric correspondence between the OCT slice and the CFP region to learn the correlated features of the two modalities for robust fusion. Furthermore, we design a new feature selection strategy to extract discriminative OCT representations by automatically selecting the important feature maps from OCT slices. Unlike the existing multimodal learning methods, GeCoM-Net is the first method that formulates the geometric relationships between the OCT slice and the corresponding region of the CFP image explicitly for CFP and OCT fusion. Experiments have been conducted on a large-scale private dataset and a publicly available dataset to evaluate the effectiveness of GeCoM-Net for diagnosing diabetic macular edema (DME), impaired visual acuity (VA) and glaucoma. The empirical results show that our method outperforms the current state-of-the-art multimodal learning methods by improving the AUROC score 0.4%, 1.9% and 2.9% for DME, VA and glaucoma detection, respectively.

Optical coherence tomography choroidal enhancement using generative deep learning.

Spectral-domain optical coherence tomography (SDOCT) is the gold standard of imaging the eye in clinics. Penetration depth with such devices is, however, limited and visualization of the choroid, which is essential for diagnosing chorioretinal disease, remains limited. Whereas swept-source OCT (SSOCT) devices allow for visualization of the choroid these instruments are expensive and availability in praxis is limited. We present an artificial intelligence (AI)-based solution to enhance the visualization of the choroid in OCT scans and allow for quantitative measurements of choroidal metrics using generative deep learning (DL). Synthetically enhanced SDOCT B-scans with improved choroidal visibility were generated, leveraging matching images to learn deep anatomical features during the training. Using a single-center tertiary eye care institution cohort comprising a total of 362 SDOCT-SSOCT paired subjects, we trained our model with 150,784 images from 410 healthy, 192 glaucoma, and 133 diabetic retinopathy eyes. An independent external test dataset of 37,376 images from 146 eyes was deployed to assess the authenticity and quality of the synthetically enhanced SDOCT images. Experts' ability to differentiate real versus synthetic images was poor (47.5% accuracy). Measurements of choroidal thickness, area, volume, and vascularity index, from the reference SSOCT and synthetically enhanced SDOCT, showed high Pearson's correlations of 0.97 [95% CI: 0.96-0.98], 0.97 [0.95-0.98], 0.95 [0.92-0.98], and 0.87 [0.83-0.91], with intra-class correlation values of 0.99 [0.98-0.99], 0.98 [0.98-0.99], and 0.95 [0.96-0.98], 0.93 [0.91-0.95], respectively. Thus, our DL generative model successfully generated realistic enhanced SDOCT data that is indistinguishable from SSOCT images providing improved visualization of the choroid. This technology enabled accurate measurements of choroidal metrics previously limited by the imaging depth constraints of SDOCT. The findings open new possibilities for utilizing affordable SDOCT devices in studying the choroid in both healthy and pathological conditions.

DRAC 2022: A public benchmark for diabetic retinopathy analysis on ultra-wide optical coherence tomography angiography images.

We described a challenge named "DRAC - Diabetic Retinopathy Analysis Challenge" in conjunction with the 25th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2022). Within this challenge, we provided the DRAC datset, an ultra-wide optical coherence tomography angiography (UW-OCTA) dataset (1,103 images), addressing three primary clinical tasks: diabetic retinopathy (DR) lesion segmentation, image quality assessment, and DR grading. The scientific community responded positively to the challenge, with 11, 12, and 13 teams submitting different solutions for these three tasks, respectively. This paper presents a concise summary and analysis of the top-performing solutions and results across all challenge tasks. These solutions could provide practical guidance for developing accurate classification and segmentation models for image quality assessment and DR diagnosis using UW-OCTA images, potentially improving the diagnostic capabilities of healthcare professionals. The dataset has been released to support the development of computer-aided diagnostic systems for DR evaluation.

A deep learning system for predicting time to progression of diabetic retinopathy.

Diabetic retinopathy (DR) is the leading cause of preventable blindness worldwide. The risk of DR progression is highly variable among different individuals, making it difficult to predict risk and personalize screening intervals. We developed and validated a deep learning system (DeepDR Plus) to predict time to DR progression within 5 years solely from fundus images. First, we used 717,308 fundus images from 179,327 participants with diabetes to pretrain the system. Subsequently, we trained and validated the system with a multiethnic dataset comprising 118,868 images from 29,868 participants with diabetes. For predicting time to DR progression, the system achieved concordance indexes of 0.754-0.846 and integrated Brier scores of 0.153-0.241 for all times up to 5 years. Furthermore, we validated the system in real-world cohorts of participants with diabetes. The integration with clinical workflow could potentially extend the mean screening interval from 12 months to 31.97 months, and the percentage of participants recommended to be screened at 1-5 years was 30.62%, 20.00%, 19.63%, 11.85% and 17.89%, respectively, while delayed detection of progression to vision-threatening DR was 0.18%. Altogether, the DeepDR Plus system could predict individualized risk and time to DR progression over 5 years, potentially allowing personalized screening intervals.

A Competition for the Diagnosis of Myopic Maculopathy by Artificial Intelligence Algorithms.

Importance: Myopic maculopathy (MM) is a major cause of vision impairment globally. Artificial intelligence (AI) and deep learning (DL) algorithms for detecting MM from fundus images could potentially improve diagnosis and assist screening in a variety of health care settings. Objectives: To evaluate DL algorithms for MM classification and segmentation and compare their performance with that of ophthalmologists. Design, Setting, and Participants: The Myopic Maculopathy Analysis Challenge (MMAC) was an international competition to develop automated solutions for 3 tasks: (1) MM classification, (2) segmentation of MM plus lesions, and (3) spherical equivalent (SE) prediction. Participants were provided 3 subdatasets containing 2306, 294, and 2003 fundus images, respectively, with which to build algorithms. A group of 5 ophthalmologists evaluated the same test sets for tasks 1 and 2 to ascertain performance. Results from model ensembles, which combined outcomes from multiple algorithms submitted by MMAC participants, were compared with each individual submitted algorithm. This study was conducted from March 1, 2023, to March 30, 2024, and data were analyzed from January 15, 2024, to March 30, 2024. Exposure: DL algorithms submitted as part of the MMAC competition or ophthalmologist interpretation. Main Outcomes and Measures: MM classification was evaluated by quadratic-weighted κ (QWK), F1 score, sensitivity, and specificity. MM plus lesions segmentation was evaluated by dice similarity coefficient (DSC), and SE prediction was evaluated by R2 and mean absolute error (MAE). Results: The 3 tasks were completed by 7, 4, and 4 teams, respectively. MM classification algorithms achieved a QWK range of 0.866 to 0.901, an F1 score range of 0.675 to 0.781, a sensitivity range of 0.667 to 0.778, and a specificity range of 0.931 to 0.945. MM plus lesions segmentation algorithms achieved a DSC range of 0.664 to 0.687 for lacquer cracks (LC), 0.579 to 0.673 for choroidal neovascularization, and 0.768 to 0.841 for Fuchs spot (FS). SE prediction algorithms achieved an R2 range of 0.791 to 0.874 and an MAE range of 0.708 to 0.943. Model ensemble results achieved the best performance compared to each submitted algorithms, and the model ensemble outperformed ophthalmologists at MM classification in sensitivity (0.801; 95% CI, 0.764-0.840 vs 0.727; 95% CI, 0.684-0.768; P = .006) and specificity (0.946; 95% CI, 0.939-0.954 vs 0.933; 95% CI, 0.925-0.941; P = .009), LC segmentation (DSC, 0.698; 95% CI, 0.649-0.745 vs DSC, 0.570; 95% CI, 0.515-0.625; P < .001), and FS segmentation (DSC, 0.863; 95% CI, 0.831-0.888 vs DSC, 0.790; 95% CI, 0.742-0.830; P < .001). Conclusions and Relevance: In this diagnostic study, 15 AI models for MM classification and segmentation on a public dataset made available for the MMAC competition were validated and evaluated, with some models achieving better diagnostic performance than ophthalmologists.

Evaluating the outcome of screening for glaucoma using colour fundus photography-based referral criteria in a teleophthalmology screening programme for diabetic retinopathy.

AIMS: To evaluate the effectiveness of glaucoma screening using glaucoma suspect (GS) referral criteria assessed on colour fundus photographs in Singapore's Integrated Diabetic Retinopathy Programme (SiDRP). METHODS: A case-control study. This study included diabetic subjects who were referred from SiDRP with and without GS between January 2017 and December 2018 and reviewed at Singapore National Eye Centre. The GS referral criteria were based on the presence of a vertical cup-to-disc ratio (VCDR) of ≥0.65 and other GS features. The final glaucoma diagnosis confirmed from electronic medical records was retrospectively matched with GS status. The sensitivity, specificity and positive predictive value (PPV) of the test were evaluated. RESULTS: Of 5023 patients (2625 with GS and 2398 without GS) reviewed for glaucoma, 451 (9.0%, 95% CI 8.2% to 9.8%) were confirmed as glaucoma. The average follow-up time was 21.5±10.2 months. Using our current GS referral criteria, the sensitivity, specificity and PPV were 81.6% (95% CI 77.7% to 85.1%), 50.6% (95% CI 49.2% to 52.1%) and 14.0% (95% CI 13.4% to 14.7%), respectively, resulting in 2257 false positive cases. Increasing the VCDR cut-off for referral to ≥0.80, the specificity increased to 93.9% (95% CI 93.1% to 94.5%) but the sensitivity decreased to 11.3% (95% CI 8.5% to 14.6%), with a PPV of 15.4% (95% CI 12.0% to 19.4%). CONCLUSIONS: Opportunistic screening for glaucoma in a lower VCDR group could result in a high number of unnecessary referrals. If healthcare infrastructures are limited, targeting case findings on a larger VCDR group with high specificity will still be beneficial.

Synthetic artificial intelligence using generative adversarial network for retinal imaging in detection of age-related macular degeneration.

Introduction: Age-related macular degeneration (AMD) is one of the leading causes of vision impairment globally and early detection is crucial to prevent vision loss. However, the screening of AMD is resource dependent and demands experienced healthcare providers. Recently, deep learning (DL) systems have shown the potential for effective detection of various eye diseases from retinal fundus images, but the development of such robust systems requires a large amount of datasets, which could be limited by prevalence of the disease and privacy of patient. As in the case of AMD, the advanced phenotype is often scarce for conducting DL analysis, which may be tackled via generating synthetic images using Generative Adversarial Networks (GANs). This study aims to develop GAN-synthesized fundus photos with AMD lesions, and to assess the realness of these images with an objective scale. Methods: To build our GAN models, a total of 125,012 fundus photos were used from a real-world non-AMD phenotypical dataset. StyleGAN2 and human-in-the-loop (HITL) method were then applied to synthesize fundus images with AMD features. To objectively assess the quality of the synthesized images, we proposed a novel realness scale based on the frequency of the broken vessels observed in the fundus photos. Four residents conducted two rounds of gradings on 300 images to distinguish real from synthetic images, based on their subjective impression and the objective scale respectively. Results and discussion: The introduction of HITL training increased the percentage of synthetic images with AMD lesions, despite the limited number of AMD images in the initial training dataset. Qualitatively, the synthesized images have been proven to be robust in that our residents had limited ability to distinguish real from synthetic ones, as evidenced by an overall accuracy of 0.66 (95% CI: 0.61-0.66) and Cohen's kappa of 0.320. For the non-referable AMD classes (no or early AMD), the accuracy was only 0.51. With the objective scale, the overall accuracy improved to 0.72. In conclusion, GAN models built with HITL training are capable of producing realistic-looking fundus images that could fool human experts, while our objective realness scale based on broken vessels can help identifying the synthetic fundus photos.