Expert Ophthalmologists Research Articles

Surgical ergonomics among ophthalmologists at Alsaim Eye Hospital, Wad Madani, Sudan: a clinical audit

BackgroundErgonomics is the design of efficiency and comfort in the working environment. Neglecting surgical ergonomics might lead to early retirement and decreased productivity. This study aimed to assess ophthalmologists’ surgical ergonomic practices and compared them to expert ophthalmologists’ recommendations published by the Royal College of Ophthalmologists.MethodsIn this audit cycle, we sent an anonymous online pretested questionnaire via Google Form to all ophthalmologists and residents at Alsaim Eye Hospital, aiming for total coverage. Based on the literature, we designed a questionnaire that inquired about the participants’ demographics, time spent in the operating room, prior reading about ergonomics, history of musculoskeletal (MSK) pain and injuries, history of exercise, and frequency of adherence to our standard. Descriptive statistics were generated by SPSS version 27, and the Chi-square test was used to check for possible associations.ResultsThe response rate was 51%. Most of the participants were residents 51.6%. Prominent respondents were females (74.2%) and within the 31–40-year age group. Almost half of them had no routine exercise. Nevertheless, most participants (83.9%) reported that they had not read about surgical ergonomics previously. 48.4% of the participants experienced work-related MSK pain. None of the operation room postures reached the targeted standard, with the highest value (77.3%) for the appropriate seating and the lowest (22.6%) for maintaining a straight back. When calculating the association between years of experience and adherence to the established operating postures, only maintaining proper gaze reached a statistical significance (P value of 0.04). MSK pain was associated with more years of experience (P value of 0.04).ConclusionWork-related MSK pain was prevalent among ophthalmologists, and their posture during surgery did not meet the standard criteria for ergonomic practice. Our recommendations include mandatory ergonomic training to the curriculum and conducting large-scale studies to review the situation in Sudan.

Large Language Models May Help Patients Understand Peer-Reviewed Scientific Articles About Ophthalmology: Development and Usability Study.

Adequate health literacy has been shown to be important for the general health of a population. To address this, it is recommended that patient-targeted medical information is written at a sixth-grade reading level. To make well-informed decisions about their health, patients may want to interact directly with peer-reviewed open access scientific articles. However, studies have shown that such text is often written with highly complex language above the levels that can be comprehended by the general population. Previously, we have published on the use of large language models (LLMs) in easing the readability of patient-targeted health information on the internet. In this study, we continue to explore the advantages of LLMs in patient education. This study aimed to explore the use of LLMs, specifically ChatGPT (OpenAI), to enhance the readability of peer-reviewed scientific articles in the field of ophthalmology. A total of 12 open access, peer-reviewed papers published by the senior authors of this study (ET and RA) were selected. Readability was assessed using the Flesch-Kincaid Grade Level and Simple Measure of Gobbledygook tests. ChatGPT 4.0 was asked "I will give you the text of a peer-reviewed scientific paper. Considering that the recommended readability of the text is 6th grade, can you simplify the following text so that a layperson reading this text can fully comprehend it? - Insert Manuscript Text -". Appropriateness was evaluated by the 2 uveitis-trained ophthalmologists. Statistical analysis was performed in Microsoft Excel. ChatGPT significantly lowered the readability and length of the selected papers from 15th to 7th grade (P<.001) while generating responses that were deemed appropriate by expert ophthalmologists. LLMs show promise in improving health literacy by enhancing the accessibility of peer-reviewed scientific articles and allowing the general population to interact directly with medical literature.

Attention and dilated convolutions inclusive deep-CNN with multiplexed texture features to diagnose Pathological and High Myopia.

In today's era, precise and timely diagnosis of ocular diseases are crucial as these disorders jeopardize millions of visions. Early detection and proactive management can minimize vision threatening complications from these disorders. High Myopia(HM) and Pathological Myopia(PM), are the globally prevalent ocular diseases, which can impair vision acuity and productivity across all age groups. Routine screening via computer aided diagnosis(CAD) is required to detect HM and PM early and halt their progression. Previous studies majorly relied only on deep convolutional neural network(deep-CNN) features and independent analysis of PM and HM. This work seeks to provide a holistic analysis of PM and HM pathology by creating a novel ternary classifier on colour fundus images to detect normal vision, PM and HM respectively. We built the classifier by integrating texture features generated using gray level co-occurrence matrix(GLCM) within the deep-CNN model. Deep-CNN model comprises of spatial attention(SA) to locate lesion, squeeze-excitation(SE) to model interdependent channel attention & atrous or dilated convolutions to capture salient multiscale features of the related disease. Investigations employing extensive ablation techniques have elucidated the significance of optic disc(OD) and retinal vessel(RV) in the fundus images and their respective alterations in the HM or PM fundus. On the diverse dataset of 3212 colour fundus images, our ternary classifier has achieved 5-fold cross validation mean accuracy of 0.9754(+/-0.014), test accuracy of 0.9767 on 645 test fundus images and the kappa score of 0.9622 indicating the clinical viability of our classifier. The performance of our classifier excelled that of other pertinent studies for both PM and HM colour fundus images. Our classifier's test findings, comprehensive aetiology analysis and class activation maps are validated by the expert ophthalmologists making it reliable to serve as a virtual doctor alleviating the concerns with the lack of skilled ophthalmologists and expensive optometry tools by facilitating remote diagnostics and telemedicine applications.

Detection of Diabetic Retinopathy (DR) using Fundus Images

Diabetic retinopathy (DR) is a leading cause of vision impairment and blindness among adults, resulting from prolonged diabetes. Early detection and timely intervention are crucial to prevent severe visual outcomes. Manual examinations of retinal images, are often subjective, time-consuming, and require expert ophthalmologists. Deep learning has shown remarkable success in image analysis across various medical fields. Its application in diagnosing diabetic retinopathy can process and evaluate images rapidly, facilitating quicker diagnoses and enabling timely treatment decisions. The proposed work highlights the use of CNN, a class of deep learning algorithms to identify relevant features from retinal images, such as micro aneurysms, hemorrhages, and exudates, which are crucial for diagnosing diabetic retinopathy. The proposed system detects the type of DR (like No DR, Mild, Moderate, Severe and Proliferative DR) based on CNN classification with accuracy of 94.35%. The large dataset of the fundus image from Kaggle is used for experimentation. The system is proven effective in detecting diabetic retinopathy

The use of artificial intelligence based chat bots in ophthalmology triage.

To evaluate AI-based chat bots ability to accurately answer common patient's questions in the field of ophthalmology. An experienced ophthalmologist curated a set of 20 representative questions and responses were sought from two AI generative models: OpenAI's ChatGPT and Google's Bard (Gemini Pro). Eight expert ophthalmologists from different sub-specialties assessed each response, blinded to the source, and ranked them by three metrics-accuracy, comprehensiveness, and clarity, on a 1-5 scale. For accuracy, ChatGPT scored a median of 4.0, whereas Bard scored a median of 3.0. In terms of comprehensiveness, ChatGPT achieved a median score of 4.5, compared to Bard which scored a median of 3.0. Regarding clarity, ChatGPT maintained a higher score with a median of 5.0, compared to Bard's median score of 4.0. All comparisons were statistically significant (p < 0.001). AI-based chat bots can provide relatively accurate and clear responses for addressing common ophthalmological inquiries. ChatGPT surpassed Bard in all measured metrics. While these AI models exhibit promise, further research is indicated to improve their performance and allow them to be used as a reliable medical tool.

UKnow-Net: Knowledge-Enhanced U-Net for Improved Retinal Vessel Segmentation

Retinal vessel segmentation plays a critical role in diagnosing and managing ophthalmic and systemic diseases, as abnormalities in retinal vasculature can indicate disease progression. Traditional manual segmentation by expert ophthalmologists is time-consuming, labor-intensive, and prone to variability, underscoring the need for automated methods. While deep learning approaches like U-Net have advanced retinal vessel segmentation, they often struggle to generalize across diverse datasets due to differences in image acquisition techniques, resolutions, and patient demographics. To address these challenges, I propose UKnow-Net, a knowledge-enhanced U-Net architecture designed to improve retinal vessel segmentation across multiple datasets. UKnow-Net employs a multi-step process involving knowledge distillation and enhancement techniques. First, I train four specialized teacher networks separately on four publicly available retinal vessel segmentation datasets—DRIVE, CHASE_DB1, DCA1, and CHUAC—allowing each to specialize in the unique features of its respective dataset. These teacher networks generate pseudo-labels representing their domain-specific knowledge. We then train a student network using the ensemble of pseudo-labels from all teacher networks, effectively distilling the collective expertise into a unified model capable of generalizing across different datasets. Experiments demonstrate that UKnow-Net outperforms traditional handcrafted networks (such as U-Net, UNet++, and Attention U-Net) and several state-of-the-art models in key performance metrics, including sensitivity, specificity, F1 score, and Intersection over Union (IoU). Specifically, our two variants, UKnowNet-A and UKnowNet-B, show well performance; UKnowNet-A, trained solely on pseudo-labels, achieved higher sensitivity across all datasets, indicating a superior ability to detect true positives, while UKnowNet-B, which combines pseudo-labels with ground truth annotations, achieved balanced precision and recall, leading to higher F1 scores and IoU metrics. The integration of pseudo-labels effectively transfers the collective expertise of the teacher networks to the student network, enhancing generalization and robustness. I aim to ensure fair comparison and reproducibility in future research by publicly sharing our source code and model weights.

Development and Validation of Claims-Based Algorithms for Conjunctivitis and Keratitis.

Ocular surface disorders have been reported among patients with various medical conditions, including atopic dermatitis (AD). Nonetheless, validated algorithms to identify conjunctivitis and keratitis in claims data are lacking. Develop validated, claims-based algorithms for conjunctivitis and keratitis among patients with AD using medical records. Patients with AD were identified in a claims database between March 2017 and November 2019. Among these patients, candidate algorithms were developed that included diagnosis codes for conjunctivitis or keratitis, alone and in combination with ophthalmic treatments. Among patients who met ≥ 1 candidate algorithms, a subset was randomly selected for medical record review. Additionally, records from a random sample of patients with AD were reviewed to assess sensitivity. Overall, 341 records were sought and 262 adjudicated by an expert ophthalmologist. The positive predictive value (PPV) of each algorithm was calculated and compared to a pre-specified threshold of ≥ 70%. For conjunctivitis, the final algorithm was ≥ 1 conjunctivitis diagnosis (PPV = 81%, 95% confidence interval [CI]: 73%-87%). For keratitis, the final algorithm combined the following 2 candidate algorithms: ≥ 1 keratitis diagnosis and ≥ 1 dispensing of a topical antibiotic or antibiotic-steroid combination (PPV = 91%); and ≥ 1 keratitis diagnosis and ≥ 1 dispensing of an ophthalmic corticosteroid, topical immune-modulator, or topical NSAID (PPV = 68%) for an overall PPV of 80% (95% CI: 55%-93%). The first claims-based algorithms to identify conjunctivitis and keratitis among AD patients were developed and validated. They are available for use in future studies, particularly to better understand conjunctivitis and keratitis occurrence among patients with AD.

Grading of diabetic retinopathy using a pre-segmenting deep learning classification model: Validation of an automated algorithm.

To validate the performance of autonomous diabetic retinopathy (DR) grading by comparing a human grader and a self-developed deep-learning (DL) algorithm with gold-standard evaluation. We included 500, 6-field retinal images graded by an expert ophthalmologist (gold standard) according to the International Clinical Diabetic Retinopathy Disease Severity Scale as represented with DR levels 0-4 (97, 100, 100, 103, 100, respectively). Weighted kappa was calculated to measure the DR classification agreement for (1) a certified human grader without, and (2) with assistance from a DL algorithm and (3) the DL operating autonomously. Using any DR (level 0 vs. 1-4) as a cutoff, we calculated sensitivity, specificity, as well as positive and negative predictive values (PPV and NPV). Finally, we assessed lesion discrepancies between Model 3 and the gold standard. As compared to the gold standard, weighted kappa for Models 1-3 was 0.88, 0.89 and 0.72, sensitivities were 95%, 94% and 78% and specificities were 82%, 84% and 81%. Extrapolating to a real-world DR prevalence of 23.8%, the PPV were 63%, 64% and 57% and the NPV were 98%, 98% and 92%. Discrepancies between the gold standard and Model 3 were mainly incorrect detection of artefacts (n = 49), missed microaneurysms (n = 26) and inconsistencies between the segmentation and classification (n = 51). While the autonomous DL algorithm for DR classification only performed on par with a human grader for some measures in a high-risk population, extrapolations to a real-world population demonstrated an excellent 92% NPV, which could make it clinically feasible to use autonomously to identify non-DR patients.

InceptionResNetV2 Based Eye Cataract Detection

An eye lens is a key source for visualizing things. Sometimes it may be cloudy, caused by cataracts, and disturb the visualization process. It is necessary to detect and treat the cataract in its early stages; otherwise, it may lead to blindness. For the early detection of eye cataracts, many solutions have been proposed in past years by experts. However, it requires an expert ophthalmologist or eye surgeon to detect cataracts in its early stages because early detection of cataracts is a challenging task. Now this is the era of digitalization and AI. Everyone is talking about AI. Many researchers use AI to propose a solution to this problem. This study also proposed an AI-based cataract detection solution. A binary class eye fundus images data set was used for this research work. The ”Cataract” data set is a famous dataset for detecting cataracts on KAGGLE. Some image augmentation techniques were used to enhance the data set proficiency. In this study, InceptionResNetV2 was applied with some hyper-parameter tuning. The proposed methodology achieved a training accuracy of 99% and a validation accuracy of 98%.

Relationship between Serum Zinc Level and Retinopathy in Type 1 Diabetic Patients

Abstract Background Diabetes mellitus is one of the common chronic metabolic diseases associated with various micro- and macro-vascular complications. Diabetic retinopathy (DR) is a microvascular disorder with various grades which leads to visual problems or even blindness. Nearly all patients with type 1 diabetes and more than 60% of those with type 2 have some degree of retinopathy after 20 years. Zinc is one of the micro-nutrients abundant in the human body and possess an important role in various metabolic reactions. Previous studies reported that decreased serum zinc levels was associated with certain diabetic microvascular complications. Zinc depletion has a major impact on the pathogenesis of diabetic complications. However, the correlation between zinc level and diabetic retinopathy needed more studies and especially in type 1diabetic patients which has not been studied at all. AIM/ Objectives Assess the correlation between serum zinc and diabetic retinopathy in type 1 diabetic patient. Subjects and Methods This observational study was conducted on 60 subjects with their ages ranging from 18-60, selected from diabetes outpatient clinics of Ain Shams University Hospitals for 7 months in a period from June 2022 to January 2023; 40 patients and 20 non diabetic as control divided into three groups: 20 Patients with Type 1 diabetes without diabetic retinopathy, 20 Patients with Type 1 diabetes with diabetic retinopathy and 20 apparently healthy individuals. All patients included in the study will be screened for diabetic retinopathy using fundus photography camera. Photos were assessed and graded by an expert ophthalmologist. Assessment of zinc level by automated chemistry analyzer will be done. Results The study shows that there was statistically significant negative correlation found between zinc level and duration of diabetes, HbA1c, LDL, TC, FPS, 2PP and serum creatinine level in all diabetic type 1 group. And in groups of patients with and without retinopathy. There was no statistically significant difference found between control group and DM type1 with and without retinopathy group regarding zinc level although there was numerical differences. Conclusion Positive relationship between decreased zinc level and diabetic retinopathy suggesting the contribution of this factors in the pathogenesis of DR through acting in a same pathway Further studies using larger numbers of patients are recommended to get benefit of giving zinc supplements to these patients

Multi-label deep learning for comprehensive optic nerve head segmentation through data of fundus images

Early diagnosis and continuous monitoring of patients with eye diseases are critical in computer-aided detection (CAD) techniques. Semantic segmentation, a key component in computer vision, enables pixel-level classification and provides detailed information about objects within images. In this study, we present three U-Net models designed for multi-class semantic segmentation, leveraging the U-Net architecture with transfer learning. To generate ground truth for the HRF dataset, we combine two U-Net models, namely MSU-Net and BU-Net, to predict probability maps for the optic disc and cup regions. Binary masks are then derived from these probability maps to extract the optic disc and cup regions from retinal images. The dataset used in this study includes pre-existing blood vessels and manually annotated peripapillary atrophy zones (alpha and beta) provided by expert ophthalmologists. This comprehensive dataset, integrating existing blood vessels and expert-marked peripapillary atrophy zones, fulfills the study's objectives. The effectiveness of the proposed approach is validated by training nine pre-trained models on the HRF dataset comprising 45 retinal images, successfully segmenting the optic disc, cup, blood vessels, and peripapillary atrophy zones (alpha and beta). The results demonstrate 87.7 % pixel accuracy, 87 % Intersection over Union (IoU), 86.9 % F1 Score, 85 % mean IoU (mIoU), and 15 % model loss, significantly contributing to the early diagnosis and monitoring of glaucoma and optic nerve disorders.

Evaluating the effectiveness of large language models in patient education for conjunctivitis

AimsTo evaluate the quality of responses from large language models (LLMs) to patient-generated conjunctivitis questions.MethodsA two-phase, cross-sectional study was conducted at the Eye and ENT Hospital of Fudan University. In...

Predictive Role of Fluorescein Angiography in Retinopathy of Prematurity.

Fluorescein angiography (FA) has been a pivotal tool for studying the pathophysiology of retinopathy of prematurity (ROP) in vivo. We examined the course of ROP using FA to assess the predictive value of angiographic features. This is an observational retrospective cohort study of eyes screened for ROP with a binocular indirect ophthalmoscope and FA. RetCam fundus imaging and video digital fluorescein angiography were performed in the neonatal intensive care unit of Santa Maria Hospital of Perugia. The masked grading of the FA images was retrospectively conducted by two ROP expert ophthalmologists. A total of 80 eyes of 40 patients were included for this study. Among the angiographic features evaluated, leakage, shunts, and tangles were predictive of the development of treatment-requiring ROP (p < 0.05). FA can add to our understanding of the evolution of vascular abnormalities in the course of ROP and can help predict which eyes will go on to treatment.

Diabetic retinopathy prediction based on vision transformer and modified capsule network

Diabetic retinopathy is considered one of the most common diseases that can lead to blindness in the working age, and the chance of developing it increases as long as a person suffers from diabetes. Protecting the sight of the patient or decelerating the evolution of this disease depends on its early detection as well as identifying the exact levels of this pathology, which is done manually by ophthalmologists. This manual process is very consuming in terms of the time and experience of an expert ophthalmologist, which makes developing an automated method to aid in the diagnosis of diabetic retinopathy an essential and urgent need. In this paper, we aim to propose a new hybrid deep learning method based on a fine-tuning vision transformer and a modified capsule network for automatic diabetic retinopathy severity level prediction. The proposed approach consists of a new range of computer vision operations, including the power law transformation technique and the contrast-limiting adaptive histogram equalization technique in the preprocessing step. While the classification step builds up on a fine-tuning vision transformer, a modified capsule network, and a classification model combined with a classification model, The effectiveness of our approach was evaluated using four datasets, including the APTOS, Messidor-2, DDR, and EyePACS datasets, for the task of severity levels of diabetic retinopathy. We have attained excellent test accuracy scores on the four datasets, respectively: 88.18%, 87.78%, 80.36%, and 78.64%. Comparing our results with the state-of-the-art, we reached a better performance.

Large language models approach expert-level clinical knowledge and reasoning in ophthalmology: A head-to-head cross-sectional study.

Large language models (LLMs) underlie remarkable recent advanced in natural language processing, and they are beginning to be applied in clinical contexts. We aimed to evaluate the clinical potential of state-of-the-art LLMs in ophthalmology using a more robust benchmark than raw examination scores. We trialled GPT-3.5 and GPT-4 on 347 ophthalmology questions before GPT-3.5, GPT-4, PaLM 2, LLaMA, expert ophthalmologists, and doctors in training were trialled on a mock examination of 87 questions. Performance was analysed with respect to question subject and type (first order recall and higher order reasoning). Masked ophthalmologists graded the accuracy, relevance, and overall preference of GPT-3.5 and GPT-4 responses to the same questions. The performance of GPT-4 (69%) was superior to GPT-3.5 (48%), LLaMA (32%), and PaLM 2 (56%). GPT-4 compared favourably with expert ophthalmologists (median 76%, range 64-90%), ophthalmology trainees (median 59%, range 57-63%), and unspecialised junior doctors (median 43%, range 41-44%). Low agreement between LLMs and doctors reflected idiosyncratic differences in knowledge and reasoning with overall consistency across subjects and types (p>0.05). All ophthalmologists preferred GPT-4 responses over GPT-3.5 and rated the accuracy and relevance of GPT-4 as higher (p<0.05). LLMs are approaching expert-level knowledge and reasoning skills in ophthalmology. In view of the comparable or superior performance to trainee-grade ophthalmologists and unspecialised junior doctors, state-of-the-art LLMs such as GPT-4 may provide useful medical advice and assistance where access to expert ophthalmologists is limited. Clinical benchmarks provide useful assays of LLM capabilities in healthcare before clinical trials can be designed and conducted.

SUPRACHOROIDAL SPACE INJECTION TECHNIQUE: Expert Panel Guidance.

To develop professional guidelines for best practices for suprachoroidal space (SCS) injection, an innovative technique for retinal therapeutic delivery, based on current published evidence and clinical experience. A panel of expert ophthalmologists reviewed current published evidence and clinical experience during a live working group meeting to define points of consensus and key clinical considerations to inform the development of guidelines for in-office SCS injection. Core consensus guidelines for in-office SCS injection were reached and reported by the expert panel. Current clinical evidence and physician experience supported SCS injection as a safe and effective method for delivering retinal and choroidal therapeutics. The panel established consensus on the rationale for SCS injection, including potential benefits relative to other intraocular delivery methods and current best practices in patient preparation, pre- and peri-injection management, SCS-specific injection techniques, and postinjection management and follow-up. These expert panel guidelines may support and promote standardization of SCS injection technique, with the goal of optimizing patient safety and outcomes. Some aspects of the procedure may reasonably be modified based on the clinical setting and physician judgment, as well as additional study.

Diagnostic Accuracy of Pediatrician-performed Digital Retinal Imaging with 3nethra neo for ROP Screening.

To evaluate the accuracy of pediatrician-performed wide-field digital retinal imaging (WFDRI) for diagnosing Retinopathy of prematurity (ROP), as compared to binocular indirect ophthalmoscopy (BIO) as the reference standard. Eligible infants undergoing ROP screening were enrolled consecutively. BIO was performed by trained ophthalmologists, followed by WFDRI (using "3nethra neo" camera) by a pediatrician. An expert pediatric ophthalmologist reviewed de-identified images for quality, presence, and severity of ROP. She was masked to the findings of BIO and the pediatrician. Diagnostic accuracy for detecting any ROP, ROP requiring treatment (Type 1), and ROP requiring referral (Type 1 or 2) were calculated for WFDRI, considering BIO as the reference standard. The analysis included 427 eyes. The sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic accuracy, and diagnostic odds ratio of WFDRI were 0.88 (95% CI: 0.81, 0.93), 0.89 (0.85, 0.92), 7.8 (5.7, 10.9), 0.14 (0.09, 0.21), 0.89 (0.85, 0.91), and 58.3 (31, 110) respectively for detection of 'any ROP'. For detecting ROP requiring treatment (Type 1), the sensitivity, specificity, NLR, and diagnostic accuracy were 0.90 (0.75, 0.97), 1.00 (0.99, 1.00), 0.11 (0.04, 0.27), and 0.99 (0.98, 1.00) respectively. For ROP requiring referral, the sensitivity, specificity, NLR, and diagnostic accuracy of pediatrician-performed WFDRI were 0.92 (0.80, 0.98), 1.00 (0.99, 1.00), 0.08 (0.03, 0.21), and 0.99 (0.98, 1.00) respectively. No serious adverse events were noted. The pediatrician and ophthalmologist had a near-perfect (k-1.00) and strong (k-0.88) agreement for ROP requiring treatment and any ROP, respectively. Pediatrician-performed WFDRI is feasible, safe, and has excellent diagnostic accuracy for identifying ROP requiring treatment.

Intra- and Inter-expert Validation of an Automatic Segmentation Method for Fluid Regions Associated with Central Serous Chorioretinopathy in OCT Images

Central Serous Chorioretinopathy (CSC) is a retinal disorder caused by the accumulation of fluid, resulting in vision distortion. The diagnosis of this disease is typically performed through Optical Coherence Tomography (OCT) imaging, which displays any fluid buildup between the retinal layers. Currently, these fluid regions are manually detected by visual inspection a time-consuming and subjective process that can be prone to errors. A series of six deep learning-based automatic segmentation architectural configurations of different levels of complexity were trained and compared in order to determine the best model intended for the automatic segmentation of CSC-related lesions in OCT images. The best performing models were then evaluated in an external validation study. Furthermore, an intra- and inter-expert analysis was conducted in order to compare the manual segmentation performed by expert ophthalmologists with the automatic segmentation provided by the models. Test results of the best performing configuration achieved a mean Dice of 0.868±0.056\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$0.868 \\mathop {\\pm } 0.056$$\\end{document} in the internal dataset. In the external validation set, these models achieved a level of agreement with human experts of up to 0.960 in terms of Kappa coefficient, contrasting with a value of 0.951 for agreement between human experts. Overall, the models reached a better agreement with either of the human experts than these experts with each other, suggesting that automatic segmentation models for the detection of CSC-related lesions in OCT imaging can be useful tools for assessing this disease, reducing the workload of manual inspection and leading to a more robust and objective diagnosis method.

AIROGS: Artificial Intelligence for Robust Glaucoma Screening Challenge.

The early detection of glaucoma is essential in preventing visual impairment. Artificial intelligence (AI) can be used to analyze color fundus photographs (CFPs) in a cost-effective manner, making glaucoma screening more accessible. While AI models for glaucoma screening from CFPs have shown promising results in laboratory settings, their performance decreases significantly in real-world scenarios due to the presence of out-of-distribution and low-quality images. To address this issue, we propose the Artificial Intelligence for Robust Glaucoma Screening (AIROGS) challenge. This challenge includes a large dataset of around 113,000 images from about 60,000 patients and 500 different screening centers, and encourages the development of algorithms that are robust to ungradable and unexpected input data. We evaluated solutions from 14 teams in this paper and found that the best teams performed similarly to a set of 20 expert ophthalmologists and optometrists. The highest-scoring team achieved an area under the receiver operating characteristic curve of 0.99 (95% CI: 0.98-0.99) for detecting ungradable images on-the-fly. Additionally, many of the algorithms showed robust performance when tested on three other publicly available datasets. These results demonstrate the feasibility of robust AI-enabled glaucoma screening.

Short-term duration of diabetic retinopathy as a predictor for development of diabetic kidney disease.

Diabetic retinopathy (DR) is a risk factor for diabetic kidney disease (DKD). Whether the duration, especially the short-term duration, of DR is associated with the development and progression of DKD remains unclear. A retrospective study and two-sample Mendelian randomization (MR) analysis were conducted. Kidney disease was defined by the urinary albumin-to-creatinine ratio (ACR) and the estimated glomerular filtration rate (eGFR). DR was diagnosed by an expert ophthalmologist by using a digital fundus camera. Binary and ordinal logistic regression analyses were performed. A restricted cubic spline was utilized to detect nonlinear associations. Summary statistics for DR- and DKD-associated single-nuclear polymorphisms (SNPs) were extracted from the FinnGen and the UK Biobank consortia. A total of 2674 patients with type 2 diabetes mellitus (T2DM) and type 2 diabetic kidney disease (T2DKD) were included. The prevalence and mean duration of DR increased with elevation of ACR and decline in eGFR. Renal function was significantly reduced in patients with DR in the fifth year of life. Binary and ordinal logistic regression showed that each 1-year increase in DR duration was associated with a 19% risk increase in the development of DKD, 16% in the elevation of ACR, and 21% in the decline of renal function. MR estimates indicated that DR was causally associated with DKD development, with an odds ratio of 2.89. DR and the duration of DR were independent risk factors for the development and progression of DKD. The short-term duration of DR may be associated with DKD development. DR had a statistically significant effect on DKD.