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

AI-integrated ocular imaging for predicting cardiovascular disease: advancements and future outlook

Cardiovascular disease (CVD) remains the leading cause of death worldwide. Assessing of CVD risk plays an essential role in identifying individuals at higher risk and enables the implementation of targeted intervention strategies, leading to improved CVD prevalence reduction and patient survival rates. The ocular vasculature, particularly the retinal vasculature, has emerged as a potential means for CVD risk stratification due to its anatomical similarities and physiological characteristics shared with other vital organs, such as the brain and heart. The integration of artificial intelligence (AI) into ocular imaging has the potential to overcome limitations associated with traditional semi-automated image analysis, including inefficiency and manual measurement errors. Furthermore, AI techniques may uncover novel and subtle features that contribute to the identification of ocular biomarkers associated with CVD. This review provides a comprehensive overview of advancements made in AI-based ocular image analysis for predicting CVD, including the prediction of CVD risk factors, the replacement of traditional CVD biomarkers (e.g., CT-scan measured coronary artery calcium score), and the prediction of symptomatic CVD events. The review covers a range of ocular imaging modalities, including colour fundus photography, optical coherence tomography, and optical coherence tomography angiography, and other types of images like external eye images. Additionally, the review addresses the current limitations of AI research in this field and discusses the challenges associated with translating AI algorithms into clinical practice.

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.

Implementing and evaluating a fully functional AI-enabled model for chronic eye disease screening in a real clinical environment

BACKGROUND

Artificial intelligence (AI) has the potential to increase the affordability and accessibility of eye disease screening, especially with the recent approval of AI-based diabetic retinopathy (DR) screening programs in several countries.

METHODS

This study investigated the performance, feasibility, and user experience of a seamless hardware and software solution for screening chronic eye diseases in a real-world clinical environment in Germany. The solution integrated AI grading for DR, age-related macular degeneration (AMD), and glaucoma, along with specialist auditing and patient referral decision. The study comprised several components: (1) evaluating the entire system solution from recruitment to eye image capture and AI grading for DR, AMD, and glaucoma; (2) comparing specialist's grading results with AI grading results; (3) gathering user feedback on the solution.

RESULTS

A total of 231 patients were recruited, and their consent forms were obtained. The sensitivity, specificity, and area under the curve for DR grading were 100.00%, 80.10%, and 90.00%, respectively. For AMD grading, the values were 90.91%, 78.79%, and 85.00%, and for glaucoma grading, the values were 93.26%, 76.76%, and 85.00%. The analysis of all false positive cases across the three diseases and their comparison with the final referral decisions revealed that only 17 patients were falsely referred among the 231 patients. The efficacy analysis of the system demonstrated the effectiveness of the AI grading process in the study's testing environment. Clinical staff involved in using the system provided positive feedback on the disease screening process, particularly praising the seamless workflow from patient registration to image transmission and obtaining the final result. Results from a questionnaire completed by 12 participants indicated that most found the system easy, quick, and highly satisfactory. The study also revealed room for improvement in the AMD model, suggesting the need to enhance its training data. Furthermore, the performance of the glaucoma model grading could be improved by incorporating additional measures such as intraocular pressure.

CONCLUSIONS

The implementation of the AI-based approach for screening three chronic eye diseases proved effective in real-world settings, earning positive feedback on the usability of the integrated platform from both the screening staff and auditors. The auditing function has proven valuable for obtaining efficient second opinions from experts, pointing to its potential for enhancing remote screening capabilities.

TRIAL REGISTRATION

Institutional Review Board of the Hamburg Medical Chamber (Ethik-Kommission der Ärztekammer Hamburg): 2021-10574-BO-ff.

Validation of a deep learning system for the detection of diabetic retinopathy in Indigenous Australians

BACKGROUND/AIMS

Deep learning systems (DLSs) for diabetic retinopathy (DR) detection show promising results but can underperform in racial and ethnic minority groups, therefore external validation within these populations is critical for health equity. This study evaluates the performance of a DLS for DR detection among Indigenous Australians, an understudied ethnic group who suffer disproportionately from DR-related blindness.

METHODS

We performed a retrospective external validation study comparing the performance of a DLS against a retinal specialist for the detection of more-than-mild DR (mtmDR), vision-threatening DR (vtDR) and all-cause referable DR. The validation set consisted of 1682 consecutive, single-field, macula-centred retinal photographs from 864 patients with diabetes (mean age 54.9 years, 52.4% women) at an Indigenous primary care service in Perth, Australia. Three-person adjudication by a panel of specialists served as the reference standard.

RESULTS

For mtmDR detection, sensitivity of the DLS was superior to the retina specialist (98.0% (95% CI, 96.5 to 99.4) vs 87.1% (95% CI, 83.6 to 90.6), McNemar's test p<0.001) with a small reduction in specificity (95.1% (95% CI, 93.6 to 96.4) vs 97.0% (95% CI, 95.9 to 98.0), p=0.006). For vtDR, the DLS's sensitivity was again superior to the human grader (96.2% (95% CI, 93.4 to 98.6) vs 84.4% (95% CI, 79.7 to 89.2), p<0.001) with a slight drop in specificity (95.8% (95% CI, 94.6 to 96.9) vs 97.8% (95% CI, 96.9 to 98.6), p=0.002). For all-cause referable DR, there was a substantial increase in sensitivity (93.7% (95% CI, 91.8 to 95.5) vs 74.4% (95% CI, 71.1 to 77.5), p<0.001) and a smaller reduction in specificity (91.7% (95% CI, 90.0 to 93.3) vs 96.3% (95% CI, 95.2 to 97.4), p<0.001).

CONCLUSION

The DLS showed improved sensitivity and similar specificity compared with a retina specialist for DR detection. This demonstrates its potential to support DR screening among Indigenous Australians, an underserved population with a high burden of diabetic eye disease.

Advancing Glaucoma Care: Integrating Artificial Intelligence in Diagnosis, Management, and Progression Detection

Glaucoma, the leading cause of irreversible blindness worldwide, comprises a group of progressive optic neuropathies requiring early detection and lifelong treatment to preserve vision. Artificial intelligence (AI) technologies are now demonstrating transformative potential across the spectrum of clinical glaucoma care. This review summarizes current capabilities, future outlooks, and practical translation considerations. For enhanced screening, algorithms analyzing retinal photographs and machine learning models synthesizing risk factors can identify high-risk patients needing diagnostic workup and close follow-up. To augment definitive diagnosis, deep learning techniques detect characteristic glaucomatous patterns by interpreting results from optical coherence tomography, visual field testing, fundus photography, and other ocular imaging. AI-powered platforms also enable continuous monitoring, with algorithms that analyze longitudinal data alerting physicians about rapid disease progression. By integrating predictive analytics with patient-specific parameters, AI can also guide precision medicine for individualized glaucoma treatment selections. Advances in robotic surgery and computer-based guidance demonstrate AI's potential to improve surgical outcomes and surgical training. Beyond the clinic, AI chatbots and reminder systems could provide patient education and counseling to promote medication adherence. However, thoughtful approaches to clinical integration, usability, diversity, and ethical implications remain critical to successfully implementing these emerging technologies. This review highlights AI's vast capabilities to transform glaucoma care while summarizing key achievements, future prospects, and practical considerations to progress from bench to bedside.

Grand rounds in methodology: key considerations for implementing machine learning solutions in quality improvement initiatives

Machine learning (ML) solutions are increasingly entering healthcare. They are complex, sociotechnical systems that include data inputs, ML models, technical infrastructure and human interactions. They have promise for improving care across a wide range of clinical applications but if poorly implemented, they may disrupt clinical workflows, exacerbate inequities in care and harm patients. Many aspects of ML solutions are similar to other digital technologies, which have well-established approaches to implementation. However, ML applications present distinct implementation challenges, given that their predictions are often complex and difficult to understand, they can be influenced by biases in the data sets used to develop them, and their impacts on human behaviour are poorly understood. This manuscript summarises the current state of knowledge about implementing ML solutions in clinical care and offers practical guidance for implementation. We propose three overarching questions for potential users to consider when deploying ML solutions in clinical care: (1) Is a clinical or operational problem likely to be addressed by an ML solution? (2) How can an ML solution be evaluated to determine its readiness for deployment? (3) How can an ML solution be deployed and maintained optimally? The Quality Improvement community has an essential role to play in ensuring that ML solutions are translated into clinical practice safely, effectively, and ethically.

Exploring the matrix: knowledge, perceptions and prospects of artificial intelligence and machine learning in Nigerian healthcare

Background

Artificial intelligence technology can be applied in several aspects of healthcare delivery and its integration into the Nigerian healthcare value chain is expected to bring about new opportunities. This study aimed at assessing the knowledge and perception of healthcare professionals in Nigeria regarding the application of artificial intelligence and machine learning in the health sector.

Methods

A cross-sectional study was undertaken amongst healthcare professionals in Nigeria with the use of a questionnaire. Data were collected across the six geopolitical zones in the Country using a stratified multistage sampling method. Descriptive and inferential statistical analyses were undertaken for the data obtained.

Results

Female participants (55.7%) were slightly higher in proportion compared to the male respondents (44.3%). Pharmacists accounted for 27.7% of the participants, and this was closely followed by medical doctors (24.5%) and nurses (19.3%). The majority of the respondents (57.2%) reported good knowledge regarding artificial intelligence and machine learning, about a third of the participants (32.2%) were of average knowledge, and 10.6% of the sample had poor knowledge. More than half of the respondents (57.8%) disagreed with the notion that the adoption of artificial intelligence in the Nigerian healthcare sector could result in job losses. Two-thirds of the participants (66.7%) were of the view that the integration of artificial intelligence in healthcare will augment human intelligence. Three-quarters (77%) of the respondents agreed that the use of machine learning in Nigerian healthcare could facilitate efficient service delivery.

Conclusion

This study provides novel insights regarding healthcare professionals' knowledge and perception with respect to the application of artificial intelligence and machine learning in healthcare. The emergent findings from this study can guide government and policymakers in decision-making as regards deployment of artificial intelligence and machine learning for healthcare delivery.

The performance of a deep learning system in assisting junior ophthalmologists in diagnosing 13 major fundus diseases: a prospective multi-center clinical trial

Artificial intelligence (AI)-based diagnostic systems have been reported to improve fundus disease screening in previous studies. This multicenter prospective self-controlled clinical trial aims to evaluate the diagnostic performance of a deep learning system (DLS) in assisting junior ophthalmologists in detecting 13 major fundus diseases. A total of 1493 fundus images from 748 patients were prospectively collected from five tertiary hospitals in China. Nine junior ophthalmologists were trained and annotated the images with or without the suggestions proposed by the DLS. The diagnostic performance was evaluated among three groups: DLS-assisted junior ophthalmologist group (test group), junior ophthalmologist group (control group) and DLS group. The diagnostic consistency was 84.9% (95%CI, 83.0% ~ 86.9%), 72.9% (95%CI, 70.3% ~ 75.6%) and 85.5% (95%CI, 83.5% ~ 87.4%) in the test group, control group and DLS group, respectively. With the help of the proposed DLS, the diagnostic consistency of junior ophthalmologists improved by approximately 12% (95% CI, 9.1% ~ 14.9%) with statistical significance (P < 0.001). For the detection of 13 diseases, the test group achieved significant higher sensitivities (72.2% ~ 100.0%) and comparable specificities (90.8% ~ 98.7%) comparing with the control group (sensitivities, 50% ~ 100%; specificities 96.7 ~ 99.8%). The DLS group presented similar performance to the test group in the detection of any fundus abnormality (sensitivity, 95.7%; specificity, 87.2%) and each of the 13 diseases (sensitivity, 83.3% ~ 100.0%; specificity, 89.0 ~ 98.0%). The proposed DLS provided a novel approach for the automatic detection of 13 major fundus diseases with high diagnostic consistency and assisted to improve the performance of junior ophthalmologists, resulting especially in reducing the risk of missed diagnoses. ClinicalTrials.gov NCT04723160.

Interpretable Detection of Diabetic Retinopathy, Retinal Vein Occlusion, Age-Related Macular Degeneration, and Other Fundus Conditions

Diabetic retinopathy (DR), retinal vein occlusion (RVO), and age-related macular degeneration (AMD) pose significant global health challenges, often resulting in vision impairment and blindness. Automatic detection of these conditions is crucial, particularly in underserved rural areas with limited access to ophthalmic services. Despite remarkable advancements in artificial intelligence, especially convolutional neural networks (CNNs), their complexity can make interpretation difficult. In this study, we curated a dataset consisting of 15,089 color fundus photographs (CFPs) obtained from 8110 patients who underwent fundus fluorescein angiography (FFA) examination. The primary objective was to construct integrated models that merge CNNs with an attention mechanism. These models were designed for a hierarchical multilabel classification task, focusing on the detection of DR, RVO, AMD, and other fundus conditions. Furthermore, our approach extended to the detailed classification of DR, RVO, and AMD according to their respective subclasses. We employed a methodology that entails the translation of diagnostic information obtained from FFA results into CFPs. Our investigation focused on evaluating the models' ability to achieve precise diagnoses solely based on CFPs. Remarkably, our models showcased improvements across diverse fundus conditions, with the ConvNeXt-base + attention model standing out for its exceptional performance. The ConvNeXt-base + attention model achieved remarkable metrics, including an area under the receiver operating characteristic curve (AUC) of 0.943, a referable F1 score of 0.870, and a Cohen's kappa of 0.778 for DR detection. For RVO, it attained an AUC of 0.960, a referable F1 score of 0.854, and a Cohen's kappa of 0.819. Furthermore, in AMD detection, the model achieved an AUC of 0.959, an F1 score of 0.727, and a Cohen's kappa of 0.686. Impressively, the model demonstrated proficiency in subclassifying RVO and AMD, showcasing commendable sensitivity and specificity. Moreover, our models enhanced interpretability by visualizing attention weights on fundus images, aiding in the identification of disease findings. These outcomes underscore the substantial impact of our models in advancing the detection of DR, RVO, and AMD, offering the potential for improved patient outcomes and positively influencing the healthcare landscape.

Deep learning-based optic disc classification is affected by optic-disc tilt

We aimed to determine the effect of optic disc tilt on deep learning-based optic disc classification. A total of 2507 fundus photographs were acquired from 2236 eyes of 1809 subjects (mean age of 46 years; 53% men). Among all photographs, 1010 (40.3%) had tilted optic discs. Image annotation was performed to label pathologic changes of the optic disc (normal, glaucomatous optic disc changes, disc swelling, and disc pallor). Deep learning-based classification modeling was implemented to develop optic-disc appearance classification models with the photographs of all subjects and those with and without tilted optic discs. Regardless of deep learning algorithms, the classification models showed better overall performance when developed based on data from subjects with non-tilted discs (AUC, 0.988 ± 0.002, 0.991 ± 0.003, and 0.986 ± 0.003 for VGG16, VGG19, and DenseNet121, respectively) than when developed based on data with tilted discs (AUC, 0.924 ± 0.046, 0.928 ± 0.017, and 0.935 ± 0.008). In classification of each pathologic change, non-tilted disc models had better sensitivity and specificity than the tilted disc models. The optic disc appearance classification models developed based all-subject data demonstrated lower accuracy in patients with the appearance of tilted discs than in those with non-tilted discs. Our findings suggested the need to identify and adjust for the effect of optic disc tilt on the optic disc classification algorithm in future development.

A newly developed deep learning-based system for automatic detection and classification of small bowel lesions during double-balloon enteroscopy examination

BACKGROUND

Double-balloon enteroscopy (DBE) is a standard method for diagnosing and treating small bowel disease. However, DBE may yield false-negative results due to oversight or inexperience. We aim to develop a computer-aided diagnostic (CAD) system for the automatic detection and classification of small bowel abnormalities in DBE.

DESIGN AND METHODS

A total of 5201 images were collected from Renmin Hospital of Wuhan University to construct a detection model for localizing lesions during DBE, and 3021 images were collected to construct a classification model for classifying lesions into four classes, protruding lesion, diverticulum, erosion & ulcer and angioectasia. The performance of the two models was evaluated using 1318 normal images and 915 abnormal images and 65 videos from independent patients and then compared with that of 8 endoscopists. The standard answer was the expert consensus.

RESULTS

For the image test set, the detection model achieved a sensitivity of 92% (843/915) and an area under the curve (AUC) of 0.947, and the classification model achieved an accuracy of 86%. For the video test set, the accuracy of the system was significantly better than that of the endoscopists (85% vs. 77 ± 6%, p < 0.01). For the video test set, the proposed system was superior to novices and comparable to experts.

CONCLUSIONS

We established a real-time CAD system for detecting and classifying small bowel lesions in DBE with favourable performance. ENDOANGEL-DBE has the potential to help endoscopists, especially novices, in clinical practice and may reduce the miss rate of small bowel lesions.

Insights into artificial intelligence in myopia management: from a data perspective

Given the high incidence and prevalence of myopia, the current healthcare system is struggling to handle the task of myopia management, which is worsened by home quarantine during the ongoing COVID-19 pandemic. The utilization of artificial intelligence (AI) in ophthalmology is thriving, yet not enough in myopia. AI can serve as a solution for the myopia pandemic, with application potential in early identification, risk stratification, progression prediction, and timely intervention. The datasets used for developing AI models are the foundation and determine the upper limit of performance. Data generated from clinical practice in managing myopia can be categorized into clinical data and imaging data, and different AI methods can be used for analysis. In this review, we comprehensively review the current application status of AI in myopia with an emphasis on data modalities used for developing AI models. We propose that establishing large public datasets with high quality, enhancing the model's capability of handling multimodal input, and exploring novel data modalities could be of great significance for the further application of AI for myopia.

A Deep Learning Model for Screening Computed Tomography Imaging for Thyroid Eye Disease and Compressive Optic Neuropathy

Purpose

Thyroid eye disease (TED) is an autoimmune condition with an array of clinical manifestations, which can be complicated by compressive optic neuropathy. It is important to identify patients with TED early to ensure close monitoring and treatment to prevent potential permanent disability or vision loss. Deep learning artificial intelligence (AI) algorithms have been utilized in ophthalmology and in other fields of medicine to detect disease. This study aims to introduce a deep learning model to evaluate orbital computed tomography (CT) images for the presence of TED and potential compressive optic neuropathy.

Design

Retrospective review and deep learning algorithm modeling.

Subjects

Patients with TED with dedicated orbital CT scans and with an examination by an oculoplastic surgeon over a 10-year period at a single academic institution. Patients with no TED and normal CTs were used as normal controls. Those with other diagnoses, such as tumors or other inflammatory processes, were excluded.

Methods

Orbital CTs were preprocessed and adopted for the Visual Geometry Group-16 network to distinguish patients with no TED, mild TED, and severe TED with compressive optic neuropathy. The primary model included training and testing of all 3 conditions. Binary model performance was also evaluated. An oculoplastic surgeon was also similarly tested with single and serial images for comparison.

Main Outcome Measures

Accuracy of deep learning model discernment of region of interest for CT scans to distinguish TED versus normal control, as well as TED with clinical signs of optic neuropathy.

Results

A total of 1187 photos from 141 patients were used to develop the AI model. The primary model trained on patients with no TED, mild TED, and severe TED had 89.5% accuracy (area under the curve: range, 0.96-0.99) in distinguishing patients with these clinical categories. In comparison, testing of an oculoplastic surgeon in these 3 categories showed decreased accuracy (70.0% accuracy in serial image testing).

Conclusions

The deep learning model developed in the study can accurately detect TED and further detect TED with clinical signs of optic neuropathy based on orbital CT. The model proved superior compared with human expert grading. With further optimization and validation, this TED deep learning model could help guide frontline health care providers in the detection of TED and help stratify the urgency of a referral to an oculoplastic surgeon and endocrinologist.

Financial Disclosures

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

A double-blinded study for quantifiable assessment of the diagnostic accuracy of AI tool "ADVEN-i" in identifying diseased fundus images including diabetic retinopathy on a retrospective data

PURPOSE

To quantifiably assess the diagnostic accuracy of Adven-I, a proprietary artificial intelligence (AI)-driven diagnostic system that automatically detects diseases from fundus images. The purpose is to quantify the performance of Adven-i in differentiating a nonreferable (within normal limits) image from a referable (diseased fundus) image and further segregating diabetic retinopathy (DR) from the rest of the abnormalities (non-DR) encompassing the wide spectrum of abnormal pathologies. The assessment is carried out in comparison to manual reading as the reference gold standard. Adven-i is the only AI system classifying retinal abnormalities into DR and non-DR classes separately, apart from predicting nonreferable fundus, while most existing systems classify fundus images into referable and nonreferable DR.

METHODS

The double-blinded study was conducted on retrospective data collected over the course of a year in the ophthalmology outpatient department (OPD) at a top Tier II eyecare hospital in Chandigarh, India. Three vitreoretina specialists who were blinded to one another read the images. The ground-truth was generated on the basis of majority agreement among the readers. An arbitrator's decision was regarded final if all three readers disagreed.

RESULTS

2261 fundus images were analyzed by Adven-i. The sensitivity and specificity of Adven-i in diagnosing images with abnormalities were 95.12% and 85.77%, respectively, and for segregating DR from rest of the retinal abnormalities were 91.87% and 85.12%, respectively.

CONCLUSIONS AND RELEVANCE

Adven-i shows definite promise in automated screening for early diagnosis of referable fundus images including DR. Adven-i can be adopted to scale for mass screening in resource-limited settings.

Population impact and cost-effectiveness of artificial intelligence-based diabetic retinopathy screening in people living with diabetes in Australia: a cost effectiveness analysis

Background

We aimed to evaluate the cost-effectiveness of an artificial intelligence-(AI) based diabetic retinopathy (DR) screening system in the primary care setting for both non-Indigenous and Indigenous people living with diabetes in Australia.

Methods

We performed a cost-effectiveness analysis between January 01, 2022 and August 01, 2023. A decision-analytic Markov model was constructed to simulate DR progression in a population of 1,197,818 non-Indigenous and 65,160 Indigenous Australians living with diabetes aged ≥20 years over 40 years. From a healthcare provider's perspective, we compared current practice to three primary care AI-based screening scenarios-(A) substitution of current manual grading, (B) scaling up to patient acceptance level, and (C) achieving universal screening. Study results were presented as incremental cost-effectiveness ratio (ICER), benefit-cost ratio (BCR), and net monetary benefits (NMB). A Willingness-to-pay (WTP) threshold of AU$50,000 per quality-adjusted life year (QALY) and a discount rate of 3.5% were adopted in this study.

Findings

With the status quo, the non-Indigenous diabetic population was projected to develop 96,269 blindness cases, resulting in AU$13,039.6 m spending on DR screening and treatment during 2020-2060. In comparison, all three intervention scenarios were effective and cost-saving. In particular, if a universal screening program was to be implemented (Scenario C), it would prevent 38,347 blindness cases, gain 172,090 QALYs and save AU$595.8 m, leading to a BCR of 3.96 and NMB of AU$9,200 m. Similar findings were also reported in the Indigenous population. With the status quo, 3,396 Indigenous individuals would develop blindness, which would cost the health system AU$796.0 m during 2020-2060. All three intervention scenarios were cost-saving for the Indigenous population. Notably, universal AI-based DR screening (Scenario C) would prevent 1,211 blindness cases and gain 9,800 QALYs in the Indigenous population, leading to a saving of AU$19.2 m with a BCR of 1.62 and NMB of AU$509 m.

Interpretation

Our findings suggest that implementing AI-based DR screening in primary care is highly effective and cost-saving in both Indigenous and non-Indigenous populations.

Funding

This project received grant funding from the Australian Government: the National Critical Research Infrastructure Initiative, Medical Research Future Fund (MRFAI00035) and the NHMRC Investigator Grant (APP1175405). The contents of the published material are solely the responsibility of the Administering Institution, a participating institution or individual authors and do not reflect the views of the NHMRC. This work was supported by the Global STEM Professorship Scheme (P0046113), the Fundamental Research Funds of the State Key Laboratory of Ophthalmology, Project of Investigation on Health Status of Employees in Financial Industry in Guangzhou, China (Z012014075). The Centre for Eye Research Australia receives Operational Infrastructure Support from the Victorian State Government. W.H. is supported by the Melbourne Research Scholarship established by the University of Melbourne. The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Artificial intelligence in retinal imaging: current status and future prospects

INTRODUCTION

The steadily growing and aging world population, in conjunction with continuously increasing prevalences of vision-threatening retinal diseases, is placing an increasing burden on the global healthcare system. The main challenges within retinology involve identifying the comparatively few patients requiring therapy within the large mass, the assurance of comprehensive screening for retinal disease and individualized therapy planning. In order to sustain high-quality ophthalmic care in the future, the incorporation of artificial intelligence (AI) technologies into our clinical practice represents a potential solution.

AREAS COVERED

This review sheds light onto already realized and promising future applications of AI techniques in retinal imaging. The main attention is directed at the application in diabetic retinopathy and age-related macular degeneration. The principles of use in disease screening, grading, therapeutic planning and prediction of future developments are explained based on the currently available literature.

EXPERT OPINION

The recent accomplishments of AI in retinal imaging indicate that its implementation into our daily practice is likely to fundamentally change the ophthalmic healthcare system and bring us one step closer to the goal of individualized treatment. However, it must be emphasized that the aim is to optimally support clinicians by gradually incorporating AI approaches, rather than replacing ophthalmologists.

Initiation of China Alliance of Research in High Myopia (CHARM): protocol for an AI-based multimodal high myopia research biobank

INTRODUCTION

High myopia is a pressing public health concern due to its increasing prevalence, younger trend and the high risk of blindness, particularly in East Asian countries, including China. The China Alliance of Research in High Myopia (CHARM) is a newly established consortium that includes more than 100 hospitals and institutions participating across the nation, aiming to promote collaboration and data sharing in the field of high myopia screening, classification, diagnosis and therapeutic development.

METHODS AND ANALYSIS

The CHARM project is an ongoing study, and its initiation is distinguished by its unprecedented scale, encompassing plans to involve over 100 000 Chinese patients. This initiative stands out not only for its extensive scope but also for its innovative application of artificial intelligence (AI) to assist in diagnosis and treatment decisions. The CHARM project has been carried out using a 'three-step' strategy. The first step involves the collection of basic information, refraction, axial length and fundus photographs from participants with high myopia. In the second step, we will collect multimodal imaging data to expand the scope of clinical information, for example, optical coherence tomography and ultra-widefield fundus images. In the final step, genetic testing will be conducted by incorporating patient family histories and blood samples. The majority of data collected by CHARM is in the form of images that will be used to detect and predict the progression of high myopia through the identification and quantification of biomarkers such as fundus tessellation, optic nerve head and vascular parameters.

ETHICS AND DISSEMINATION

The study has received approval from the Ethics Committee of Beijing Tongren Hospital (TREC2022-KY045). The establishment of CHARM represents an opportunity to create a collaborative platform for myopia experts and facilitate the dissemination of research findings to the global community through peer-reviewed publications and conference presentations. These insights can inform clinical decision-making and contribute to the development of new treatment modalities that may benefit patients worldwide.

TRIAL REGISTRATION NUMBER

ChiCTR2300071219.

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

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

From the diagnosis of infectious keratitis to discriminating fungal subtypes; a deep learning-based study

Infectious keratitis (IK) is a major cause of corneal opacity. IK can be caused by a variety of microorganisms. Typically, fungal ulcers carry the worst prognosis. Fungal cases can be subdivided into filamentous and yeasts, which shows fundamental differences. Delays in diagnosis or initiation of treatment increase the risk of ocular complications. Currently, the diagnosis of IK is mainly based on slit-lamp examination and corneal scrapings. Notably, these diagnostic methods have their drawbacks, including experience-dependency, tissue damage, and time consumption. Artificial intelligence (AI) is designed to mimic and enhance human decision-making. An increasing number of studies have utilized AI in the diagnosis of IK. In this paper, we propose to use AI to diagnose IK (model 1), differentiate between bacterial keratitis and fungal keratitis (model 2), and discriminate the filamentous type from the yeast type of fungal cases (model 3). Overall, 9329 slit-lamp photographs gathered from 977 patients were enrolled in the study. The models exhibited remarkable accuracy, with model 1 achieving 99.3%, model 2 at 84%, and model 3 reaching 77.5%. In conclusion, our study offers valuable support in the early identification of potential fungal and bacterial keratitis cases and helps enable timely management.

AI-based diagnosis of nuclear cataract from slit-lamp videos

In ophthalmology, the availability of many fundus photographs and optical coherence tomography images has spurred consideration of using artificial intelligence (AI) for diagnosing retinal and optic nerve disorders. However, AI application for diagnosing anterior segment eye conditions remains unfeasible due to limited standardized images and analysis models. We addressed this limitation by augmenting the quantity of standardized optical images using a video-recordable slit-lamp device. We then investigated whether our proposed machine learning (ML) AI algorithm could accurately diagnose cataracts from videos recorded with this device. We collected 206,574 cataract frames from 1812 cataract eye videos. Ophthalmologists graded the nuclear cataracts (NUCs) using the cataract grading scale of the World Health Organization. These gradings were used to train and validate an ML algorithm. A validation dataset was used to compare the NUC diagnosis and grading of AI and ophthalmologists. The results of individual cataract gradings were: NUC 0: area under the curve (AUC) = 0.967; NUC 1: AUC = 0.928; NUC 2: AUC = 0.923; and NUC 3: AUC = 0.949. Our ML-based cataract diagnostic model achieved performance comparable to a conventional device, presenting a promising and accurate auto diagnostic AI tool.

Teleophthalmology and retina: a review of current tools, pathways and services

Telemedicine, the use of telecommunication and information technology to deliver healthcare remotely, has evolved beyond recognition since its inception in the 1970s. Advances in telecommunication infrastructure, the advent of the Internet, exponential growth in computing power and associated computer-aided diagnosis, and medical imaging developments have created an environment where telemedicine is more accessible and capable than ever before, particularly in the field of ophthalmology. Ever-increasing global demand for ophthalmic services due to population growth and ageing together with insufficient supply of ophthalmologists requires new models of healthcare provision integrating telemedicine to meet present day challenges, with the recent COVID-19 pandemic providing the catalyst for the widespread adoption and acceptance of teleophthalmology. In this review we discuss the history, present and future application of telemedicine within the field of ophthalmology, and specifically retinal disease. We consider the strengths and limitations of teleophthalmology, its role in screening, community and hospital management of retinal disease, patient and clinician attitudes, and barriers to its adoption.

AI-Human Hybrid Workflow Enhances Teleophthalmology for the Detection of Diabetic Retinopathy

Objective

Detection of diabetic retinopathy (DR) outside of specialized eye care settings is an important means of access to vision-preserving health maintenance. Remote interpretation of fundus photographs acquired in a primary care or other nonophthalmic setting in a store-and-forward manner is a predominant paradigm of teleophthalmology screening programs. Artificial intelligence (AI)-based image interpretation offers an alternative means of DR detection. IDx-DR (Digital Diagnostics Inc) is a Food and Drug Administration-authorized autonomous testing device for DR. We evaluated the diagnostic performance of IDx-DR compared with human-based teleophthalmology over 2 and a half years. Additionally, we evaluated an AI-human hybrid workflow that combines AI-system evaluation with human expert-based assessment for referable cases.

Design

Prospective cohort study and retrospective analysis.

Participants

Diabetic patients ≥ 18 years old without a prior DR diagnosis or DR examination in the past year presenting for routine DR screening in a primary care clinic.

Methods

Macula-centered and optic nerve-centered fundus photographs were evaluated by an AI algorithm followed by consensus-based overreading by retina specialists at the Stanford Ophthalmic Reading Center. Detection of more-than-mild diabetic retinopathy (MTMDR) was compared with in-person examination by a retina specialist.

Main Outcome Measures

Sensitivity, specificity, accuracy, positive predictive value, and gradability achieved by the AI algorithm and retina specialists.

Results

The AI algorithm had higher sensitivity (95.5% sensitivity; 95% confidence interval [CI], 86.7%-100%) but lower specificity (60.3% specificity; 95% CI, 47.7%-72.9%) for detection of MTMDR compared with remote image interpretation by retina specialists (69.5% sensitivity; 95% CI, 50.7%-88.3%; 96.9% specificity; 95% CI, 93.5%-100%). Gradability of encounters was also lower for the AI algorithm (62.5%) compared with retina specialists (93.1%). A 2-step AI-human hybrid workflow in which the AI algorithm initially rendered an assessment followed by overread by a retina specialist of MTMDR-positive encounters resulted in a sensitivity of 95.5% (95% CI, 86.7%-100%) and a specificity of 98.2% (95% CI, 94.6%-100%). Similarly, a 2-step overread by retina specialists of AI-ungradable encounters improved gradability from 63.5% to 95.6% of encounters.

Conclusions

Implementation of an AI-human hybrid teleophthalmology workflow may both decrease reliance on human specialist effort and improve diagnostic accuracy.

Financial Disclosures

Proprietary or commercial disclosure may be found after the references.

Large language models and their impact in ophthalmology

The advent of generative artificial intelligence and large language models has ushered in transformative applications within medicine. Specifically in ophthalmology, large language models offer unique opportunities to revolutionise digital eye care, address clinical workflow inefficiencies, and enhance patient experiences across diverse global eye care landscapes. Yet alongside these prospects lie tangible and ethical challenges, encompassing data privacy, security, and the intricacies of embedding large language models into clinical routines. This Viewpoint highlights the promising applications of large language models in ophthalmology, while weighing up the practical and ethical barriers towards their real-world implementation. This Viewpoint seeks to stimulate broader discourse on the potential of large language models in ophthalmology and to galvanise both clinicians and researchers into tackling the prevailing challenges and optimising the benefits of large language models while curtailing the associated risks.

Risk Stratification for Diabetic Retinopathy Screening Order Using Deep Learning: A Multicenter Prospective Study

Purpose

Real-world evaluation of a deep learning model that prioritizes patients based on risk of progression to moderate or worse (MOD+) diabetic retinopathy (DR).

Methods

This nonrandomized, single-arm, prospective, interventional study included patients attending DR screening at four centers across Thailand from September 2019 to January 2020, with mild or no DR. Fundus photographs were input into the model, and patients were scheduled for their subsequent screening from September 2020 to January 2021 in order of predicted risk. Evaluation focused on model sensitivity, defined as correctly ranking patients that developed MOD+ within the first 50% of subsequent screens.

Results

We analyzed 1,757 patients, of which 52 (3.0%) developed MOD+. Using the model-proposed order, the model's sensitivity was 90.4%. Both the model-proposed order and mild/no DR plus HbA1c had significantly higher sensitivity than the random order (P < 0.001). Excluding one major (rural) site that had practical implementation challenges, the remaining sites included 567 patients and 15 (2.6%) developed MOD+. Here, the model-proposed order achieved 86.7% versus 73.3% for the ranking that used DR grade and hemoglobin A1c.

Conclusions

The model can help prioritize follow-up visits for the largest subgroups of DR patients (those with no or mild DR). Further research is needed to evaluate the impact on clinical management and outcomes.

Translational Relevance

Deep learning demonstrated potential for risk stratification in DR screening. However, real-world practicalities must be resolved to fully realize the benefit.

CLASSICA: Validating artificial intelligence in classifying cancer in real time during surgery

AIM

Treatment pathways for significant rectal polyps differ depending on the underlying pathology, but pre-excision profiling is imperfect. It has been demonstrated that differences in fluorescence perfusion signals following injection of indocyanine green (ICG) can be analysed mathematically and, with the assistance of artificial intelligence (AI), used to classify tumours endoscopically as benign or malignant. This study aims to validate this method of characterization across multiple clinical sites regarding its generalizability, usability and accuracy while developing clinical-grade software to enable it to become a useful method.

METHODS

The CLASSICA study is a prospective, unblinded multicentre European observational study aimed to validate the use of AI analysis of ICG fluorescence for intra-operative tissue characterization. Six hundred patients undergoing transanal endoscopic evaluation of significant rectal polyps and tumours will be enrolled in at least five clinical sites across the European Union over a 4-year period. Video recordings will be analysed regarding dynamic fluorescence patterns centrally as software is developed to enable analysis with automatic classification to happen locally. AI-based classification and subsequently guided intervention will be compared with the current standard of care including biopsies, final specimen pathology and patient outcomes.

DISCUSSION

CLASSICA will validate the use of AI in the analysis of ICG fluorescence for the purposes of classifying significant rectal polyps and tumours endoscopically. Follow-on studies will compare AI-guided targeted biopsy or, indeed, AI characterization alone with traditional biopsy and AI-guided local excision versus traditional excision with regard to marginal clearance and recurrence.

Three-year changes in sex judgment using color fundus parameters in elementary school students

PURPOSE

In a previous cross-sectional study, we reported that the sexes can be distinguished using known factors obtained from color fundus photography (CFP). However, it is not clear how sex differences in fundus parameters appear across the human lifespan. Therefore, we conducted a cohort study to investigate sex determination based on fundus parameters in elementary school students.

METHODS

This prospective observational longitudinal study investigated 109 right eyes of elementary school students over 4 years (age, 8.5 to 11.5 years). From each CFP, the tessellation fundus index was calculated as red/red + green + blue (R/[R+G+B]) using the mean value of red-green-blue intensity in eight locations around the optic disc and macular region. Optic disc area, ovality ratio, papillomacular angle, and retinal vessel angles and distances were quantified according to the data in our previous report. Using 54 fundus parameters, sex was predicted by L2 regularized binomial logistic regression for each grade.

RESULTS

The right eyes of 53 boys and 56 girls were analyzed. The discrimination accuracy rate significantly increased with age: 56.3% at 8.5 years, 46.1% at 9.5 years, 65.5% at 10.5 years and 73.1% at 11.5 years.

CONCLUSIONS

The accuracy of sex discrimination by fundus photography improved during a 3-year cohort study of elementary school students.

Accuracy of artificial intelligence model for infectious keratitis classification: a systematic review and meta-analysis

Background

Infectious keratitis (IK) is a sight-threatening condition requiring immediate definite treatment. The need for prompt treatment heavily depends on timely diagnosis. The diagnosis of IK, however, is challenged by the drawbacks of the current "gold standard." The poorly differentiated clinical features, the possibility of low microbial culture yield, and the duration for culture are the culprits of delayed IK treatment. Deep learning (DL) is a recent artificial intelligence (AI) advancement that has been demonstrated to be highly promising in making automated diagnosis in IK with high accuracy. However, its exact accuracy is not yet elucidated. This article is the first systematic review and meta-analysis that aims to assess the accuracy of available DL models to correctly classify IK based on etiology compared to the current gold standards.

Methods

A systematic search was carried out in PubMed, Google Scholars, Proquest, ScienceDirect, Cochrane and Scopus. The used keywords are: "Keratitis," "Corneal ulcer," "Corneal diseases," "Corneal lesions," "Artificial intelligence," "Deep learning," and "Machine learning." Studies including slit lamp photography of the cornea and validity study on DL performance were considered. The primary outcomes reviewed were the accuracy and classification capability of the AI machine learning/DL algorithm. We analyzed the extracted data with the MetaXL 5.2 Software.

Results

A total of eleven articles from 2002 to 2022 were included with a total dataset of 34,070 images. All studies used convolutional neural networks (CNNs), with ResNet and DenseNet models being the most used models across studies. Most AI models outperform the human counterparts with a pooled area under the curve (AUC) of 0.851 and accuracy of 96.6% in differentiating IK vs. non-IK and pooled AUC 0.895 and accuracy of 64.38% for classifying bacterial keratitis (BK) vs. fungal keratitis (FK).

Conclusion

This study demonstrated that DL algorithms have high potential in diagnosing and classifying IK with accuracy that, if not better, is comparable to trained corneal experts. However, various factors, such as the unique architecture of DL model, the problem with overfitting, image quality of the datasets, and the complex nature of IK itself, still hamper the universal applicability of DL in daily clinical practice.

Automated segmentation of ultra-widefield fluorescein angiography of diabetic retinopathy using deep learning

BACKGROUND/AIMS

Retinal capillary non-perfusion (NP) and neovascularisation (NV) are two of the most important angiographic changes in diabetic retinopathy (DR). This study investigated the feasibility of using deep learning (DL) models to automatically segment NP and NV on ultra-widefield fluorescein angiography (UWFA) images from patients with DR.

METHODS

Retrospective cross-sectional chart review study. In total, 951 UWFA images were collected from patients with severe non-proliferative DR (NPDR) or proliferative DR (PDR). Each image was segmented and labelled for NP, NV, disc, background and outside areas. Using the labelled images, DL models were trained and validated (80%) using convolutional neural networks (CNNs) for automated segmentation and tested (20%) on test sets. Accuracy of each model and each label were assessed.

RESULTS

The best accuracy from CNN models for each label was 0.8208, 0.8338, 0.9801, 0.9253 and 0.9766 for NP, NV, disc, background and outside areas, respectively. The best Intersection over Union for each label was 0.6806, 0.5675, 0.7107, 0.8551 and 0.924 and mean mean boundary F1 score (BF score) was 0.6702, 0.8742, 0.9092, 0.8103 and 0.9006, respectively.

CONCLUSIONS

DL models can detect NV and NP as well as disc and outer margins on UWFA with good performance. This automated segmentation of important UWFA features will aid physicians in DR clinics and in overcoming grader subjectivity.

Deep learning-based fundus image analysis for cardiovascular disease: a review

It is well established that the retina provides insights beyond the eye. Through observation of retinal microvascular changes, studies have shown that the retina contains information related to cardiovascular disease. Despite the tremendous efforts toward reducing the effects of cardiovascular diseases, they remain a global challenge and a significant public health concern. Conventionally, predicting the risk of cardiovascular disease involves the assessment of preclinical features, risk factors, or biomarkers. However, they are associated with cost implications, and tests to acquire predictive parameters are invasive. Artificial intelligence systems, particularly deep learning (DL) methods applied to fundus images have been generating significant interest as an adjunct assessment tool with the potential of enhancing efforts to prevent cardiovascular disease mortality. Risk factors such as age, gender, smoking status, hypertension, and diabetes can be predicted from fundus images using DL applications with comparable performance to human beings. A clinical change to incorporate DL systems for the analysis of fundus images as an equally good test over more expensive and invasive procedures may require conducting prospective clinical trials to mitigate all the possible ethical challenges and medicolegal implications. This review presents current evidence regarding the use of DL applications on fundus images to predict cardiovascular disease.

A practical guide to the implementation of AI in orthopaedic research - part 1: opportunities in clinical application and overcoming existing challenges

Artificial intelligence (AI) has the potential to transform medical research by improving disease diagnosis, clinical decision-making, and outcome prediction. Despite the rapid adoption of AI and machine learning (ML) in other domains and industry, deployment in medical research and clinical practice poses several challenges due to the inherent characteristics and barriers of the healthcare sector. Therefore, researchers aiming to perform AI-intensive studies require a fundamental understanding of the key concepts, biases, and clinical safety concerns associated with the use of AI. Through the analysis of large, multimodal datasets, AI has the potential to revolutionize orthopaedic research, with new insights regarding the optimal diagnosis and management of patients affected musculoskeletal injury and disease. The article is the first in a series introducing fundamental concepts and best practices to guide healthcare professionals and researcher interested in performing AI-intensive orthopaedic research studies. The vast potential of AI in orthopaedics is illustrated through examples involving disease- or injury-specific outcome prediction, medical image analysis, clinical decision support systems and digital twin technology. Furthermore, it is essential to address the role of human involvement in training unbiased, generalizable AI models, their explainability in high-risk clinical settings and the implementation of expert oversight and clinical safety measures for failure. In conclusion, the opportunities and challenges of AI in medicine are presented to ensure the safe and ethical deployment of AI models for orthopaedic research and clinical application. Level of evidence IV.

Artificial intelligence for detecting keratoconus

BACKGROUND

Keratoconus remains difficult to diagnose, especially in the early stages. It is a progressive disorder of the cornea that starts at a young age. Diagnosis is based on clinical examination and corneal imaging; though in the early stages, when there are no clinical signs, diagnosis depends on the interpretation of corneal imaging (e.g. topography and tomography) by trained cornea specialists. Using artificial intelligence (AI) to analyse the corneal images and detect cases of keratoconus could help prevent visual acuity loss and even corneal transplantation. However, a missed diagnosis in people seeking refractive surgery could lead to weakening of the cornea and keratoconus-like ectasia. There is a need for a reliable overview of the accuracy of AI for detecting keratoconus and the applicability of this automated method to the clinical setting.

OBJECTIVES

To assess the diagnostic accuracy of artificial intelligence (AI) algorithms for detecting keratoconus in people presenting with refractive errors, especially those whose vision can no longer be fully corrected with glasses, those seeking corneal refractive surgery, and those suspected of having keratoconus. AI could help ophthalmologists, optometrists, and other eye care professionals to make decisions on referral to cornea specialists. Secondary objectives To assess the following potential causes of heterogeneity in diagnostic performance across studies. • Different AI algorithms (e.g. neural networks, decision trees, support vector machines) • Index test methodology (preprocessing techniques, core AI method, and postprocessing techniques) • Sources of input to train algorithms (topography and tomography images from Placido disc system, Scheimpflug system, slit-scanning system, or optical coherence tomography (OCT); number of training and testing cases/images; label/endpoint variable used for training) • Study setting • Study design • Ethnicity, or geographic area as its proxy • Different index test positivity criteria provided by the topography or tomography device • Reference standard, topography or tomography, one or two cornea specialists • Definition of keratoconus • Mean age of participants • Recruitment of participants • Severity of keratoconus (clinically manifest or subclinical) SEARCH METHODS: We searched CENTRAL (which contains the Cochrane Eyes and Vision Trials Register), Ovid MEDLINE, Ovid Embase, OpenGrey, the ISRCTN registry, ClinicalTrials.gov, and the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP). There were no date or language restrictions in the electronic searches for trials. We last searched the electronic databases on 29 November 2022.

SELECTION CRITERIA

We included cross-sectional and diagnostic case-control studies that investigated AI for the diagnosis of keratoconus using topography, tomography, or both. We included studies that diagnosed manifest keratoconus, subclinical keratoconus, or both. The reference standard was the interpretation of topography or tomography images by at least two cornea specialists.

DATA COLLECTION AND ANALYSIS

Two review authors independently extracted the study data and assessed the quality of studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) tool. When an article contained multiple AI algorithms, we selected the algorithm with the highest Youden's index. We assessed the certainty of evidence using the GRADE approach.

MAIN RESULTS

We included 63 studies, published between 1994 and 2022, that developed and investigated the accuracy of AI for the diagnosis of keratoconus. There were three different units of analysis in the studies: eyes, participants, and images. Forty-four studies analysed 23,771 eyes, four studies analysed 3843 participants, and 15 studies analysed 38,832 images. Fifty-four articles evaluated the detection of manifest keratoconus, defined as a cornea that showed any clinical sign of keratoconus. The accuracy of AI seems almost perfect, with a summary sensitivity of 98.6% (95% confidence interval (CI) 97.6% to 99.1%) and a summary specificity of 98.3% (95% CI 97.4% to 98.9%). However, accuracy varied across studies and the certainty of the evidence was low. Twenty-eight articles evaluated the detection of subclinical keratoconus, although the definition of subclinical varied. We grouped subclinical keratoconus, forme fruste, and very asymmetrical eyes together. The tests showed good accuracy, with a summary sensitivity of 90.0% (95% CI 84.5% to 93.8%) and a summary specificity of 95.5% (95% CI 91.9% to 97.5%). However, the certainty of the evidence was very low for sensitivity and low for specificity. In both groups, we graded most studies at high risk of bias, with high applicability concerns, in the domain of patient selection, since most were case-control studies. Moreover, we graded the certainty of evidence as low to very low due to selection bias, inconsistency, and imprecision. We could not explain the heterogeneity between the studies. The sensitivity analyses based on study design, AI algorithm, imaging technique (topography versus tomography), and data source (parameters versus images) showed no differences in the results.

AUTHORS' CONCLUSIONS

AI appears to be a promising triage tool in ophthalmologic practice for diagnosing keratoconus. Test accuracy was very high for manifest keratoconus and slightly lower for subclinical keratoconus, indicating a higher chance of missing a diagnosis in people without clinical signs. This could lead to progression of keratoconus or an erroneous indication for refractive surgery, which would worsen the disease. We are unable to draw clear and reliable conclusions due to the high risk of bias, the unexplained heterogeneity of the results, and high applicability concerns, all of which reduced our confidence in the evidence. Greater standardization in future research would increase the quality of studies and improve comparability between studies.

Glaucoma: now and beyond

The glaucomas are a group of conditions leading to irreversible sight loss and characterised by progressive loss of retinal ganglion cells. Although not always elevated, intraocular pressure is the only modifiable risk factor demonstrated by large clinical trials. It remains the leading cause of irreversible blindness, but timely treatment to lower intraocular pressure is effective at slowing the rate of vision loss from glaucoma. Methods for lowering intraocular pressure include laser treatments, topical medications, and surgery. Although modern surgical innovations aim to be less invasive, many have been introduced with little supporting evidence from randomised controlled trials. Many cases remain undiagnosed until the advanced stages of disease due to the limitations of screening and poor access to opportunistic case finding. Future research aims to generate evidence for intraocular pressure-independent neuroprotective treatments, personalised treatment through genetic risk profiling, and exploration of potential advanced cellular and gene therapies.

Trends in using deep learning algorithms in biomedical prediction systems

In the domain of using DL-based methods in medical and healthcare prediction systems, the utilization of state-of-the-art deep learning (DL) methodologies assumes paramount significance. DL has attained remarkable achievements across diverse domains, rendering its efficacy particularly noteworthy in this context. The integration of DL with health and medical prediction systems enables real-time analysis of vast and intricate datasets, yielding insights that significantly enhance healthcare outcomes and operational efficiency in the industry. This comprehensive literature review systematically investigates the latest DL solutions for the challenges encountered in medical healthcare, with a specific emphasis on DL applications in the medical domain. By categorizing cutting-edge DL approaches into distinct categories, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), generative adversarial networks (GANs), long short-term memory (LSTM) models, support vector machine (SVM), and hybrid models, this study delves into their underlying principles, merits, limitations, methodologies, simulation environments, and datasets. Notably, the majority of the scrutinized articles were published in 2022, underscoring the contemporaneous nature of the research. Moreover, this review accentuates the forefront advancements in DL techniques and their practical applications within the realm of medical prediction systems, while simultaneously addressing the challenges that hinder the widespread implementation of DL in image segmentation within the medical healthcare domains. These discerned insights serve as compelling impetuses for future studies aimed at the progressive advancement of using DL-based methods in medical and health prediction systems. The evaluation metrics employed across the reviewed articles encompass a broad spectrum of features, encompassing accuracy, precision, specificity, F-score, adoptability, adaptability, and scalability.

Prediction of Central Visual Field Measures From Macular OCT Volume Scans With Deep Learning

Purpose

Predict central 10° global and local visual field (VF) measurements from macular optical coherence tomography (OCT) volume scans with deep learning (DL).

Methods

This study included 1121 OCT volume scans and 10-2 VFs from 289 eyes (257 patients). Macular scans were used to estimate 10-2 VF mean deviation (MD), threshold sensitivities (TS), and total deviation (TD) values at 68 locations. A three-dimensional (3D) convolutional neural network based on the 3D DenseNet121 architecture was used for prediction. We compared DL predictions to those from baseline linear models. We carried out 10-fold stratified cross-validation to optimize generalizability. The performance of the DL and baseline models was compared based on correlations between ground truth and predicted VF measures and mean absolute error (MAE; ground truth - predicted values).

Results

Average (SD) MD was -9.3 (7.7) dB. Average (SD) correlations between predicted and ground truth MD and MD MAE were 0.74 (0.09) and 3.5 (0.4) dB, respectively. Estimation accuracy deteriorated with worsening MD. Average (SD) Pearson correlations between predicted and ground truth TS and MAEs for DL and baseline model were 0.71 (0.05) and 0.52 (0.05) (P < 0.001) and 6.5 (0.6) and 7.5 (0.5) dB (P < 0.001), respectively. For TD, correlation (SD) and MAE (SD) for DL and baseline models were 0.69 (0.02) and 0.48 (0.05) (P < 0.001) and 6.1 (0.5) and 7.8 (0.5) dB (P < 0.001), respectively.

Conclusions

Macular OCT volume scans can be used to predict global central VF parameters with clinically relevant accuracy.

Translational Relevance

Macular OCT imaging may be used to confirm and supplement central VF findings using deep learning.

Save our Sight (SOS): a collective call-to-action for enhanced retinal care across health systems in high income countries

With a growing aging population, the prevalence of age-related eye disease and associated eye care is expected to increase. The anticipated growth in demand, coupled with recent medical advances that have transformed eye care for people living with retinal diseases, particularly neovascular age-related macular degeneration (nAMD) and diabetic eye disease, has presented an opportunity for health systems to proactively manage the expected burden of these diseases. To do so, we must take collective action to address existing and anticipated capacity limitations by designing and implementing sustainable strategies that enable health systems to provide an optimal standard of care. Sufficient capacity will enable us to streamline and personalize the patient experience, reduce treatment burden, enable more equitable access to care and ensure optimal health outcomes. Through a multi-modal approach that gathered unbiased perspectives from clinical experts and patient advocates from eight high-income countries, substantiated perspectives with evidence from the published literature and validated findings with the broader eye care community, we have exposed capacity challenges that are motivating the community to take action and advocate for change. Herein, we propose a collective call-to-action for the future management of retinal diseases and potential strategies to achieve better health outcomes for individuals at-risk of, or living with, retinal disease.

Deep Learning Performance of Ultra-Widefield Fundus Imaging for Screening Retinal Lesions in Rural Locales

Importance

Retinal diseases are the leading cause of irreversible blindness worldwide, and timely detection contributes to prevention of permanent vision loss, especially for patients in rural areas with limited medical resources. Deep learning systems (DLSs) based on fundus images with a 45° field of view have been extensively applied in population screening, while the feasibility of using ultra-widefield (UWF) fundus image-based DLSs to detect retinal lesions in patients in rural areas warrants exploration.

Objective

To explore the performance of a DLS for multiple retinal lesion screening using UWF fundus images from patients in rural areas.

Design, Setting, and Participants

In this diagnostic study, a previously developed DLS based on UWF fundus images was used to screen for 5 retinal lesions (retinal exudates or drusen, glaucomatous optic neuropathy, retinal hemorrhage, lattice degeneration or retinal breaks, and retinal detachment) in 24 villages of Yangxi County, China, between November 17, 2020, and March 30, 2021.

Interventions

The captured images were analyzed by the DLS and ophthalmologists.

Main Outcomes and Measures

The performance of the DLS in rural screening was compared with that of the internal validation in the previous model development stage. The image quality, lesion proportion, and complexity of lesion composition were compared between the model development stage and the rural screening stage.

Results

A total of 6222 eyes in 3149 participants (1685 women [53.5%]; mean [SD] age, 70.9 [9.1] years) were screened. The DLS achieved a mean (SD) area under the receiver operating characteristic curve (AUC) of 0.918 (0.021) (95% CI, 0.892-0.944) for detecting 5 retinal lesions in the entire data set when applied for patients in rural areas, which was lower than that reported at the model development stage (AUC, 0.998 [0.002] [95% CI, 0.995-1.000]; P < .001). Compared with the fundus images in the model development stage, the fundus images in this rural screening study had an increased frequency of poor quality (13.8% [860 of 6222] vs 0%), increased variation in lesion proportions (0.1% [6 of 6222]-36.5% [2271 of 6222] vs 14.0% [2793 of 19 891]-21.3% [3433 of 16 138]), and an increased complexity of lesion composition.

Conclusions and Relevance

This diagnostic study suggests that the DLS exhibited excellent performance using UWF fundus images as a screening tool for 5 retinal lesions in patients in a rural setting. However, poor image quality, diverse lesion proportions, and a complex set of lesions may have reduced the performance of the DLS; these factors in targeted screening scenarios should be taken into consideration in the model development stage to ensure good performance.

Artificial intelligence in retinal disease: clinical application, challenges, and future directions

Retinal diseases are a leading cause of blindness in developed countries, accounting for the largest share of visually impaired children, working-age adults (inherited retinal disease), and elderly individuals (age-related macular degeneration). These conditions need specialised clinicians to interpret multimodal retinal imaging, with diagnosis and intervention potentially delayed. With an increasing and ageing population, this is becoming a global health priority. One solution is the development of artificial intelligence (AI) software to facilitate rapid data processing. Herein, we review research offering decision support for the diagnosis, classification, monitoring, and treatment of retinal disease using AI. We have prioritised diabetic retinopathy, age-related macular degeneration, inherited retinal disease, and retinopathy of prematurity. There is cautious optimism that these algorithms will be integrated into routine clinical practice to facilitate access to vision-saving treatments, improve efficiency of healthcare systems, and assist clinicians in processing the ever-increasing volume of multimodal data, thereby also liberating time for doctor-patient interaction and co-development of personalised management plans.

Harnessing the Power of Artificial Intelligence in Dermatology: A Comprehensive Commentary

This special article provides a comprehensive commentary on the significant role of artificial intelligence (AI) in the field of dermatology. It explores the potential of AI in various aspects of dermatologic practice, including diagnosis, treatment planning, research and patient management. The article discusses the current state of AI in dermatology, its challenges and the ethical considerations surrounding its implementation. It highlights the transformative impact of AI on dermatologic care and offers insights into the future directions of AI in the field.

Validating the Generalizability of Ophthalmic Artificial Intelligence Models on Real-World Clinical Data

Purpose

This study aims to investigate generalizability of deep learning (DL) models trained on commonly used public fundus images to an instance of real-world data (RWD) for glaucoma diagnosis.

Methods

We used Illinois Eye and Ear Infirmary fundus data set as an instance of RWD in addition to six publicly available fundus data sets. We compared the performance of DL-trained models on public data and RWD for glaucoma classification and optic disc (OD) segmentation tasks. For each task, we created models trained on each data set, respectively, and each model was tested on both data sets. We further examined each model's decision-making process and learned embeddings for the glaucoma classification task.

Results

Using public data for the test set, public-trained models outperformed RWD-trained models in OD segmentation and glaucoma classification with a mean intersection over union of 96.3% and mean area under the receiver operating characteristic curve of 95.0%, respectively. Using the RWD test set, the performance of public models decreased by 8.0% and 18.4% to 85.6% and 76.6% for OD segmentation and glaucoma classification tasks, respectively. RWD models outperformed public models on RWD test sets by 2.0% and 9.5%, respectively, in OD segmentation and glaucoma classification tasks.

Conclusions

DL models trained on commonly used public data have limited ability to generalize to RWD for classifying glaucoma. They perform similarly to RWD models for OD segmentation.

Translational Relevance

RWD is a potential solution for improving generalizability of DL models and enabling clinical translations in the care of prevalent blinding ophthalmic conditions, such as glaucoma.

Ensemble neural network model for detecting thyroid eye disease using external photographs

PURPOSE

To describe an artificial intelligence platform that detects thyroid eye disease (TED).

DESIGN

Development of a deep learning model.

METHODS

1944 photographs from a clinical database were used to train a deep learning model. 344 additional images ('test set') were used to calculate performance metrics. Receiver operating characteristic, precision-recall curves and heatmaps were generated. From the test set, 50 images were randomly selected ('survey set') and used to compare model performance with ophthalmologist performance. 222 images obtained from a separate clinical database were used to assess model recall and to quantitate model performance with respect to disease stage and grade.

RESULTS

The model achieved test set accuracy of 89.2%, specificity 86.9%, recall 93.4%, precision 79.7% and an F1 score of 86.0%. Heatmaps demonstrated that the model identified pixels corresponding to clinical features of TED. On the survey set, the ensemble model achieved accuracy, specificity, recall, precision and F1 score of 86%, 84%, 89%, 77% and 82%, respectively. 27 ophthalmologists achieved mean performance of 75%, 82%, 63%, 72% and 66%, respectively. On the second test set, the model achieved recall of 91.9%, with higher recall for moderate to severe (98.2%, n=55) and active disease (98.3%, n=60), as compared with mild (86.8%, n=68) or stable disease (85.7%, n=63).

CONCLUSIONS

The deep learning classifier is a novel approach to identify TED and is a first step in the development of tools to improve diagnostic accuracy and lower barriers to specialist evaluation.

Clinician-Driven AI: Code-Free Self-Training on Public Data for Diabetic Retinopathy Referral

Importance

Democratizing artificial intelligence (AI) enables model development by clinicians with a lack of coding expertise, powerful computing resources, and large, well-labeled data sets.

Objective

To determine whether resource-constrained clinicians can use self-training via automated machine learning (ML) and public data sets to design high-performing diabetic retinopathy classification models.

Design, Setting, and Participants

This diagnostic quality improvement study was conducted from January 1, 2021, to December 31, 2021. A self-training method without coding was used on 2 public data sets with retinal images from patients in France (Messidor-2 [n = 1748]) and the UK and US (EyePACS [n = 58 689]) and externally validated on 1 data set with retinal images from patients of a private Egyptian medical retina clinic (Egypt [n = 210]). An AI model was trained to classify referable diabetic retinopathy as an exemplar use case. Messidor-2 images were assigned adjudicated labels available on Kaggle; 4 images were deemed ungradable and excluded, leaving 1744 images. A total of 300 images randomly selected from the EyePACS data set were independently relabeled by 3 blinded retina specialists using the International Classification of Diabetic Retinopathy protocol for diabetic retinopathy grade and diabetic macular edema presence; 19 images were deemed ungradable, leaving 281 images. Data analysis was performed from February 1 to February 28, 2021.

Exposures

Using public data sets, a teacher model was trained with labeled images using supervised learning. Next, the resulting predictions, termed pseudolabels, were used on an unlabeled public data set. Finally, a student model was trained with the existing labeled images and the additional pseudolabeled images.

Main Outcomes and Measures

The analyzed metrics for the models included the area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, specificity, and F1 score. The Fisher exact test was performed, and 2-tailed P values were calculated for failure case analysis.

Results

For the internal validation data sets, AUROC values for performance ranged from 0.886 to 0.939 for the teacher model and from 0.916 to 0.951 for the student model. For external validation of automated ML model performance, AUROC values and accuracy were 0.964 and 93.3% for the teacher model, 0.950 and 96.7% for the student model, and 0.890 and 94.3% for the manually coded bespoke model, respectively.

Conclusions and Relevance

These findings suggest that self-training using automated ML is an effective method to increase both model performance and generalizability while decreasing the need for costly expert labeling. This approach advances the democratization of AI by enabling clinicians without coding expertise or access to large, well-labeled private data sets to develop their own AI models.

A Prediction Model for Sight-Threatening Diabetic Retinopathy Based on Plasma Adipokines among Patients with Mild Diabetic Retinopathy

Background

Accumulating evidence has suggested a link between adipokines and diabetic retinopathy (DR). This study is aimed at investigating the risk factors for sight-threatening DR (STDR) and establishing a prognostic model for predicting STDR among a high-risk population of patients with type 2 diabetes mellitus (T2DM).

Methods

Plasma concentrations of adipokines were determined by enzyme-linked immunosorbent assay. In the case-control set, principal component analysis (PCA) was performed to select optimal predictive cytokines for STDR, involving severe nonproliferative DR (NPDR) and proliferative DR. Support vector machine (SVM) was used to examine the possible combination of baseline plasma adipokines to discriminate the patients with mild NPDR who will later develop STDR. An individual prospective cohort with a follow-up period of 3 years was used for the external validation.

Results

In both training and testing sets, involving 306 patients with T2DM, median levels of plasma adiponectin (APN), leptin, and fatty acid-binding protein 4 (FABP4) were significantly higher in the STDR group than those in mild NPDR. Except for adipsin, the other three adipokines, FABP4, APN, and leptin, were selected by PCA and integrated into SVM. The accuracy of the multivariate SVM classification model was acceptable in both the training set (AUC = 0.81, sensitivity = 71%, and specificity = 91%) and the testing set (AUC = 0.77, sensitivity = 61%, and specificity = 92%). 110 T2DM patients with mild NPDR, the high-risk population of STDR, were enrolled for external validation. Based on the SVM, the risk of each patient was calculated. More STDR occurred in the high-risk group than in the low-risk group, which were grouped by the median value of APN, FABP4, and leptin, respectively. The model was validated in an individual cohort using SVM with the AUC, sensitivity, and specificity reaching 0.77, 64%, and 91%, respectively.

Conclusions

Adiponectin, leptin, and FABP4 were demonstrated to be associated with the severity of DR and maybe good predictors for STDR, suggesting that adipokines may play an important role in the pathophysiology of DR development.

Uncertainty-inspired open set learning for retinal anomaly identification

Failure to recognize samples from the classes unseen during training is a major limitation of artificial intelligence in the real-world implementation for recognition and classification of retinal anomalies. We establish an uncertainty-inspired open set (UIOS) model, which is trained with fundus images of 9 retinal conditions. Besides assessing the probability of each category, UIOS also calculates an uncertainty score to express its confidence. Our UIOS model with thresholding strategy achieves an F1 score of 99.55%, 97.01% and 91.91% for the internal testing set, external target categories (TC)-JSIEC dataset and TC-unseen testing set, respectively, compared to the F1 score of 92.20%, 80.69% and 64.74% by the standard AI model. Furthermore, UIOS correctly predicts high uncertainty scores, which would prompt the need for a manual check in the datasets of non-target categories retinal diseases, low-quality fundus images, and non-fundus images. UIOS provides a robust method for real-world screening of retinal anomalies.

Perspectives of Latinx Patients with Diabetes on Teleophthalmology, Artificial Intelligence-Based Image Interpretation, and Virtual Care: A Qualitative Study

Background

Latinx populations in the United States bear a disproportionate burden of diabetic eye disease. Teleophthalmology with and without artificial intelligence (AI)-based image interpretation are validated methods for diabetic eye screening, but limited literature exists on patient perspectives. This study aimed at understanding the perspectives of Latinx patients with diabetes on teleophthalmology, AI-based image interpretation, and general virtual care to prevent avoidable blindness in this population.

Methods

We conducted semi-structured, individual interviews with 20 Latinx patients with diabetes at an urban, federally qualified health center in Madison, WI. Interviews were transcribed verbatim, professionally translated from Spanish to English, and analyzed using both inductive open coding and deductive coding.

Results

Most participants had no prior experience with teleophthalmology but did have experience with virtual care. Participants expressed a preference for teleophthalmology compared with traditional in-person dilated eye exams but were willing to obtain whichever method of screening was recommended by their primary care clinician. They also strongly preferred having human physician oversight in image review compared with having images interpreted solely using AI. Many participants preferred in-person clinic visits to virtual health care due to the ability to have a more thorough physical exam, as well as for improved non-verbal communication with their clinician.

Discussion

Leveraging primary care providers' recommendations, human oversight of AI-based image interpretation, and improving communication may enhance acceptance and utilization of teleophthalmology, AI, and virtual care by Latinx patients.

Conclusions

Understanding Latinx patient perspectives may contribute toward the development of more effective telemedicine interventions to enhance health equity in Latinx communities.

Machine learning-based prediction of symptomatic intracerebral hemorrhage after intravenous thrombolysis for stroke: a large multicenter study

Background

This study aimed to compare the performance of different machine learning models in predicting symptomatic intracranial hemorrhage (sICH) after thrombolysis treatment for ischemic stroke.

Methods

This multicenter study utilized the Shenyang Stroke Emergency Map database, comprising 8,924 acute ischemic stroke patients from 29 comprehensive hospitals who underwent thrombolysis between January 2019 and December 2021. An independent testing cohort was further established, including 1,921 patients from the First People's Hospital of Shenyang. The structured dataset encompassed 15 variables, including clinical and therapeutic metrics. The primary outcome was the sICH occurrence post-thrombolysis. Models were developed using an 80/20 split for training and internal validation. Performance was assessed using machine learning classifiers, including logistic regression with lasso regularization, support vector machine (SVM), random forest, gradient-boosted decision tree (GBDT), and multilayer perceptron (MLP). The model boasting the highest area under the curve (AUC) was specifically employed to highlight feature importance.

Results

Baseline characteristics were compared between the training cohort (n = 6,369) and the external validation cohort (n = 1,921), with the sICH incidence being slightly higher in the training cohort (1.6%) compared to the validation cohort (1.1%). Among the evaluated models, the logistic regression with lasso regularization achieved the highest AUC of 0.87 (95% confidence interval [CI]: 0.79-0.95; p < 0.001), followed by the MLP model with an AUC of 0.766 (95% CI: 0.637-0.894; p = 0.04). The reference model and SVM showed AUCs of 0.575 and 0.582, respectively, while the random forest and GBDT models performed less optimally with AUCs of 0.536 and 0.436, respectively. Decision curve analysis revealed net benefits primarily for the SVM and MLP models. Feature importance from the logistic regression model emphasized anticoagulation therapy as the most significant negative predictor (coefficient: -2.0833) and recombinant tissue plasminogen activator as the principal positive predictor (coefficient: 0.5082).

Conclusion

After a comprehensive evaluation, the MLP model is recommended due to its superior ability to predict the risk of symptomatic hemorrhage post-thrombolysis in ischemic stroke patients. Based on decision curve analysis, the MLP-based model was chosen and demonstrated enhanced discriminative ability compared to the reference. This model serves as a valuable tool for clinicians, aiding in treatment planning and ensuring more precise forecasting of patient outcomes.

Deep learning for precision medicine: Guiding laser therapy in ischemic retinal diseases

In this issue of Cell Reports Medicine, Zhao and colleagues1 report a multi-tasking artificial intelligence system that can assist the whole process of fundus fluorescein angiography (FFA) imaging and reduce the reliance on retinal specialists in FFA examination.

Artificial intelligence in diabetes management: Advancements, opportunities, and challenges

The increasing prevalence of diabetes, high avoidable morbidity and mortality due to diabetes and diabetic complications, and related substantial economic burden make diabetes a significant health challenge worldwide. A shortage of diabetes specialists, uneven distribution of medical resources, low adherence to medications, and improper self-management contribute to poor glycemic control in patients with diabetes. Recent advancements in digital health technologies, especially artificial intelligence (AI), provide a significant opportunity to achieve better efficiency in diabetes care, which may diminish the increase in diabetes-related health-care expenditures. Here, we review the recent progress in the application of AI in the management of diabetes and then discuss the opportunities and challenges of AI application in clinical practice. Furthermore, we explore the possibility of combining and expanding upon existing digital health technologies to develop an AI-assisted digital health-care ecosystem that includes the prevention and management of diabetes.