Ophthalmology science最新文献

{"title":"Targeted Defect-Mapping Microperimetry in Geographic Atrophy","authors":"Abera Saeed MChD ,&nbsp;Robyn H. Guymer MBBS, PhD ,&nbsp;Xavier Hadoux MEng, PhD ,&nbsp;Maxime Jannaud MEng ,&nbsp;Darvy Dang BOrth(Hons) ,&nbsp;Lauren A.B. Hodgson MPH ,&nbsp;Emily K. Glover OD ,&nbsp;Erin E. Gee BAppSc(MedRad) ,&nbsp;Peter van Wijngaarden MBBS(Hons), PhD ,&nbsp;Zhichao Wu BAppSc(Optom), PhD","doi":"10.1016/j.xops.2025.100856","DOIUrl":"10.1016/j.xops.2025.100856","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the effectiveness of targeted defect-mapping microperimetry (DMP) for capturing progressive visual function loss in eyes with a small extent of geographic atrophy (GA).</div></div><div><h3>Design</h3><div>Prospective longitudinal study.</div></div><div><h3>Participants</h3><div>Twenty-seven eyes from 25 participants with &lt;0.75 disc areas of GA, seen over 145 visits in total.</div></div><div><h3>Methods</h3><div>All participants underwent high-density targeted DMP testing of a 4° radius region-of-interest. Two tests were performed at each visit, and participants were reviewed at 3-monthly intervals up to a 24-month period. Spearman rank correlations were calculated to assess structure–function relationships between the proportion of locations missed (PLM; nonresponse to 10-decibel stimuli) on each test and GA extent in the corresponding region tested. Targeted DMP outcome measures were derived based on evaluating the PLM from all test locations, or only from points adjacent to locations that were repeatably nonresponding on both baseline tests. These DMP outcome measures, and GA extent in the central 8° radius region, were compared based on the coefficient of variation (CoV; representing performance for capturing longitudinal changes) and sample size requirements for trials powered to detect a ≥30% treatment effect.</div></div><div><h3>Main Outcome Measures</h3><div>Correlation coefficients, CoV, and sample size estimates.</div></div><div><h3>Results</h3><div>There was a strong and moderate structure–function correlation between the PLM on targeted DMP and GA extent in the corresponding region tested at the cross section and longitudinally, respectively (correlation coefficient = 0.88 and 0.57, respectively). Evaluating PLM from points ≤0.5° adjacent to repeatably nonresponding locations at baseline was the best-performing DMP outcome measure (CoV = 71%), which was comparable with evaluating GA extent in the central 8° region (CoV = 80%). Evaluating this DMP outcome measure reduced the sample size of a 24-month trial by 46% compared with evaluating GA extent.</div></div><div><h3>Conclusions</h3><div>Targeted DMP testing can provide a sensitive functional outcome measure for trials that can capture longitudinal changes as effectively as measuring structural changes in eyes with a small extent of GA. The high structure–function correlations observed suggests that beneficial treatment effects on GA growth would likely be accompanied by corresponding evidence of functional benefit captured by targeted DMP testing.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100856"},"PeriodicalIF":4.6,"publicationDate":"2025-06-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144722802","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Elevated Intraocular Pressure Immediately after Cataract Surgery and Future Risk of Primary Open-Angle Glaucoma in the IRIS® Registry (Intelligent Research in Sight)","authors":"Nayoon Gim BS ,&nbsp;Yu Jiang PhD ,&nbsp;Yelena Bagdasarova PhD ,&nbsp;Alina Ferguson BS ,&nbsp;Marian Blazes MD ,&nbsp;Aaron Y. Lee MD, MSCI ,&nbsp;Andrew Chen MD ,&nbsp;Cecilia S. Lee MD, MS ,&nbsp;Parisa Taravati MD ,&nbsp;IRIS® Registry Analytic Center Consortium","doi":"10.1016/j.xops.2025.100851","DOIUrl":"10.1016/j.xops.2025.100851","url":null,"abstract":"<div><h3>Objective</h3><div>This study evaluated associations between postoperative intraocular pressure (IOP) after cataract surgery and the future risk of developing primary open-angle glaucoma (POAG) in patients without prior glaucoma, glaucoma suspect, or ocular hypertension diagnoses.</div></div><div><h3>Design</h3><div>Retrospective cohort study.</div></div><div><h3>Subjects</h3><div>1 912 101 individuals without prior glaucoma, glaucoma suspect, or ocular hypertension diagnoses who underwent their first cataract surgery in the American Academy of Ophthalmology IRIS® Registry (Intelligent Research in Sight).</div></div><div><h3>Methods</h3><div>The highest IOP recorded on postoperative days 0 to 2 was used for analysis. For Kaplan–Meier survival estimates and Cox proportional hazards model analysis, IOP was dichotomized into normal (≤21 mmHg) and high (&gt;21 mmHg) and assessed for associations with POAG. The stratified Cox model quantified the associations between IOP and the risk of POAG across different demographic groups. Additionally, postoperative IOP was divided into decile categories, and hazard ratios (HRs) of the risk of POAG were estimated for each, with the 40% to 60% IOP range as the reference, adjusting for demographic factors.</div></div><div><h3>Main Outcome Measures</h3><div>Cumulative probability of POAG diagnosis and HRs for POAG development.</div></div><div><h3>Results</h3><div>The median time to development of POAG was 682 days (interquartile range 191–1467 days). Kaplan–Meier estimates showed that the 4000-day cumulative probability of POAG diagnosis for the high IOP group was nearly double the normal group (3.4% vs. 1.7%, <em>P</em> &lt; 0.0001). The Cox proportional hazards model identified high postoperative IOP, older age, male sex, and Asian, Black, Native Hawaiian, and Other Pacific Islander races, as well as Hispanic ethnicity, as risk factors for POAG. In the stratified Cox analysis, high postoperative IOP was consistently associated with increased risk of POAG across demographic subgroups. The highest IOP decile was associated with increased risk of POAG (HR 2.42; 95% confidence intervals [CI] 2.26–2.58), while the lowest decile was not (HR 0.88, 95% CI 0.81–0.95). Similar trends were observed with risks of other types of glaucoma.</div></div><div><h3>Conclusions</h3><div>Elevated postoperative IOP after cataract surgery is a risk factor for future POAG development, independent of age, sex, race, and ethnicity.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100851"},"PeriodicalIF":3.2,"publicationDate":"2025-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144654562","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploiting Microperimetry as a Functional End Point in Healthy Aging and Different Stages of Age-Related Macular Degeneration","authors":"Johannes Schrittwieser MD ,&nbsp;Klaudia Birner MD ,&nbsp;Leonard M. Coulibaly MD ,&nbsp;Ursula Schmidt-Erfurth MD ,&nbsp;Gregor S. Reiter MD, PhD","doi":"10.1016/j.xops.2025.100850","DOIUrl":"10.1016/j.xops.2025.100850","url":null,"abstract":"<div><h3>Objective</h3><div>To compare functional parameters between healthy aged eyes and different stages of age-related macular degeneration (AMD) based on functional parameters in microperimetry (MP) in 2 commonly used MP devices.</div></div><div><h3>Design</h3><div>A prospective, cross-sectional study.</div></div><div><h3>Subjects</h3><div>From 80 eyes from 80 subjects, 14 400 stimuli points were included.</div></div><div><h3>Methods</h3><div>Subjects classified as healthy, intermediate AMD, neovascular AMD (nAMD), or geographic atrophy (GA) secondary to AMD were imaged with Spectralis HRA+OCT (Heidelberg Engineering) and underwent 2 consecutive examinations each, using the MP-3 (NIDEK) under photopic conditions and the MAIA (Centervue) under mesopic conditions. Pointwise sensitivity (PWS), mean sensitivity, range between highest and lowest PWS, fixation stability, and examination duration were compared between all 4 groups in both devices. Group comparison was performed using linear mixed-effects models and a discriminant analysis to find the parameters that best discriminated the respective AMD stage.</div></div><div><h3>Main Outcome Measures</h3><div>Pointwise sensitivity, range of the PWS, fixation metrics, and durations of the examinations.</div></div><div><h3>Results</h3><div>The groups exhibited significant differences in PWS (<em>P</em> &lt; 0.001) and mean sensitivity (<em>P</em> &lt; 0.001), with healthy eyes showing the highest and late stages of AMD showing the lowest sensitivity values. In addition, GA showed significantly greater fixation stability compared with nAMD in the MP-3 at 2° and 4° (<em>P</em> = 0.014 and <em>P</em> = 0.008, respectively). The examination duration in healthy patients was significantly shorter compared with patients with GA (<em>P</em> = 0.041) in MP-3. No significant differences in fixation stability and duration between groups were observed with the MAIA device. The range between the highest and lowest PWS was the most effective parameter for discrimination, with a classification accuracy of 52.5% and 50.6% in the MP-3 and MAIA, respectively.</div></div><div><h3>Conclusions</h3><div>Retinal sensitivity declines with disease progression in AMD in both mesopic and photopic background illumination. The lowest retinal sensitivity was observed in patients with GA. Background illumination should be considered when selecting an MP device for a clinical trial.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100850"},"PeriodicalIF":3.2,"publicationDate":"2025-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144549442","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Retinal Vascular Fingerprints as Novel Biomarkers for Primary Angle Closure Disease Progression","authors":"Jin Wang MD ,&nbsp;Shumei Han MD ,&nbsp;Dapeng Mou MD ,&nbsp;Xin Tang MD ,&nbsp;Danli Shi PhD ,&nbsp;Mingguang He MD ,&nbsp;Chunyan Guo ,&nbsp;Ningli Wang MD, PhD ,&nbsp;Ye Zhang MD","doi":"10.1016/j.xops.2025.100848","DOIUrl":"10.1016/j.xops.2025.100848","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate retinal vascular parameters across primary angle-closure disease (PACD) stages and explore their association with glaucomatous optic neuropathy (GON).</div></div><div><h3>Design</h3><div>A cross-sectional, hospital-based study.</div></div><div><h3>Participants</h3><div>We enrolled 638 eyes from 425 participants aged ≥40 years with PACD and further classified them into primary angle closure suspect (PACS), primary angle closure (PAC), and primary angle closure glaucoma (PACG) groups.</div></div><div><h3>Methods</h3><div>Retinal vascular parameters were measured using the Retinal-based Microvascular Health Assessment System and compared between 3 groups. A multivariable logistic mixed effects model was used to identify factors associated with the presence of GON.</div></div><div><h3>Main Outcome Measures</h3><div>Vessel caliber, tortuosity, complexity, and branching angle parameters.</div></div><div><h3>Results</h3><div>No significant differences in retinal vascular parameters were found between PACS and PAC groups. Eyes in PACG showed significant vascular changes compared to PACS (<em>P</em> &lt; 0.05). Elevated intraocular pressure (odds ratio [OR] = 2.44, <em>P</em> &lt; 0.001), reduced arteriolar curve tortuosity (OR = 0.12, <em>P</em> = 0.002), arteriolar fractal dimension (OR = 0.08, <em>P</em> = 0.027), arteriolar branching angle (OR = 0.16, <em>P</em> = 0.004), and asymmetry ratio (OR = 0.10, <em>P</em> &lt; 0.001 for artery and OR = 0.25, <em>P</em> = 0.023 for vein) were significantly associated with the presence of GON.</div></div><div><h3>Conclusions</h3><div>Retinal “vascular geometric fingerprints” show significant alterations in eyes with PACG compared to PACS and are independently associated with the presence of GON. These findings offer new insights into the vascular changes in GON, and longitudinal studies are needed to determine their prognostic value and clinical utility in managing PACD.</div></div><div><h3>Financial Disclosure(s)</h3><div>The author(s) have no proprietary or commercial interest in any materials discussed in this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100848"},"PeriodicalIF":3.2,"publicationDate":"2025-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632091","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Individualized Estimation of Baseline Retinal Nerve Fiber Layer Thickness Using Conditional Variational Autoencoder","authors":"Ou Tan PhD ,&nbsp;Keke Liu MD ,&nbsp;Aiyin Chen MD ,&nbsp;Dongseok Choi PhD ,&nbsp;Jonathan C.H. Chan MD ,&nbsp;Bonnie N.K. Choy MD ,&nbsp;Kendrick C. Shih MD ,&nbsp;Jasper K.W. Wong MD ,&nbsp;Alex L.K. Ng MD ,&nbsp;Janice J.C. Cheung MD ,&nbsp;Michael Y. Ni MD ,&nbsp;Jimmy S.M. Lai MD ,&nbsp;Gabriel M. Leung MD ,&nbsp;Ian Y.H. Wong MD ,&nbsp;David Huang MD, PhD","doi":"10.1016/j.xops.2025.100849","DOIUrl":"10.1016/j.xops.2025.100849","url":null,"abstract":"<div><h3>Purpose</h3><div>Use generative deep learning (DL) models to estimate baseline reference nerve fiber layer thickness (NFLT) profiles, taking into account individual ocular characteristics.</div></div><div><h3>Design</h3><div>A cross-sectional study.</div></div><div><h3>Participants</h3><div>Six hundred eighty-six individuals from the Hong Kong FAMILY cohort and 75 individuals from the Casey Eye Institute (CEI) cohort.</div></div><div><h3>Methods</h3><div>Healthy eyes were selected from the Hong Kong FAMILY and CEI cohorts. Circumpapillary NFLT profiles and vascular patterns were measured by a spectral-domain OCT. Generative DL models were trained using the FAMILY data to reconstruct the individualized baseline NFLT, a customized normal reference based on each eye’s own vascular pattern, axial length (AL), spherical equivalent (SE) refractive error, disc size, and demographic information. Two DL models were developed. The MAG model used actual AL and SE, while the REG model estimated AL and SE using vascular patterns as input. For comparison, a multiple linear regression (MLR) was trained to estimate baseline NFLT using AL and demographic information. Fivefold cross-validation was used to assess performance.</div></div><div><h3>Main Outcome Measures</h3><div>The prediction error: root-mean-square of the difference between the actual NFLT profile and the predicted individualized baseline.</div></div><div><h3>Results</h3><div>A total of 1152 healthy eyes from 686 participants in the Hong Kong Family cohort were divided into 4 subgroups: high myopia (SE &lt;−6 diopters [D]), low myopia (SE = −6 D ∼ −1 D), emmetropia (SE = −1D∼1D), and hyperopia (SE &gt;1D). Compared with the population means, both DL models significantly reduced the prediction error for overall and quadrant NFLT and decreased the false-positive rate of identifying abnormal NFLT thinning in both myopia groups (from 13.0%-27.0% to 6.3%∼9.4%). Both DL models significantly reduced prediction error for the NFLT profiles compared with both the population mean and the MLR-adjusted NFLT. The reductions in prediction errors for NFLT profile and overall NFLT value were independently validated using the CEI data.</div></div><div><h3>Conclusions</h3><div>Generative DL models (a type of artificial intelligence) can construct individualized NFLT baseline profiles using the vascular pattern derived from the same OCT scans. The individualized baseline reduced the prediction error of the NFLT profile in healthy eyes and may improve the accuracy of identifying abnormal NFLT thinning, especially in myopic eyes.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100849"},"PeriodicalIF":4.6,"publicationDate":"2025-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757467","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Risk Analysis of Retinal Hemangioblastomas in Nonadvanced Stages of von Hippel–Lindau Syndrome Using Ultra-widefield Imaging: The ULTRA von Hippel–Lindau Study","authors":"Sebastiano Del Fabbro MD ,&nbsp;Maria Vittoria Cicinelli MD ,&nbsp;Rosangela Lattanzio MD ,&nbsp;Soufiane Bousyf MD ,&nbsp;Lorenza Bruno MD ,&nbsp;Alessio Antropoli MD ,&nbsp;Lorenzo Bianco MD ,&nbsp;Elena Bruschi MD ,&nbsp;Alessandro Arrigo MD ,&nbsp;Giovanni Pipitone MD ,&nbsp;Francesco Cei MD ,&nbsp;Lucia Salerno MD ,&nbsp;Alessandro Larcher MD ,&nbsp;Andrea Salonia MD ,&nbsp;Francesco Bandello MD ,&nbsp;Maurizio Battaglia Parodi MD ,&nbsp;the OSR VHL Program","doi":"10.1016/j.xops.2025.100846","DOIUrl":"10.1016/j.xops.2025.100846","url":null,"abstract":"<div><h3>Purpose</h3><div>To investigate the longitudinal progression, risk factors, and complications associated with retinal hemangioblastomas (RHs) in nonadvanced stages of von Hippel–Lindau (VHL) syndrome using ultra-widefield (UWF) imaging.</div></div><div><h3>Design</h3><div>Single-center longitudinal cohort study.</div></div><div><h3>Subjects</h3><div>Caucasian patients with genetically confirmed VHL syndrome.</div></div><div><h3>Methods</h3><div>Annual evaluations included dilated fundus examinations, UWF pseudocolor fundus retinal images, UWF fluorescein angiography, and OCT. Genetic analysis classified VHL mutations. Baseline RH counts and anatomical distributions were recorded as central (juxtapapillary, macular), peripheral, or both. Longitudinal follow-up tracked new RH formation and visual acuity (VA) values.</div></div><div><h3>Main Outcome Measures</h3><div>Cumulative incidence, incidence rate ratios (IRRs), and risk factors of new RHs assessed using mixed-effects negative binomial regression models. Hazard of recurrent RHs evaluated through Cox frailty models and longitudinal changes in VA.</div></div><div><h3>Results</h3><div>Among 78 eyes of 43 patients (mean age: 47.8 ± 13.6 years), 110 RHs were documented at baseline, with 3 (3%) centrally located, 35 (32%) peripherally, and 72 (65%) spanning both zones. von Hippel–Lindau variants were investigated in 37 patients: 19 had missense variants (51%), and 18 had presumed null alleles (49%), including nonsense (10 of 37; 27%), frameshift (1 of 37; 3%), splice site (1 of 37; 3%), and exon deletion mutations (6 of 37; 16%). Over a median follow-up of 31 months (interquartile range: 27–109), 35 (43%) eyes developed new RHs, with an incidence rate of 0.22 RHs per eye-year (95% confidence interval: 0.17–0.27). By the last available examination, 26 eyes (34%) remained disease-free, whereas17 (23%) showed no progression of existing RHs. Age reduced the IRR of new RHs by 4.2% annually (<em>P</em> = 0.003), whereas higher baseline tumor burden and vascular leakage increased the IRR significantly (<em>P</em> &lt; 0.001 and <em>P</em> = 0.03, respectively). Peripheral RHs were the strongest predictor of recurrence (hazard ratio = 16.4, <em>P</em> &lt; 0.001), whereas older age remained protective (hazard ratio = 0.96, <em>P</em> = 0.04). Visual acuity (logarithm of the minimum angle of resolution) worsened from 0.05 ± 0.2 (Snellen equivalent: 20/22) at baseline to 0.11 ± 0.3 (Snellen equivalent: 20/25) at the final visit.</div></div><div><h3>Conclusions</h3><div>Peripheral RHs and vascular leakage are significant risk factors for RH progression and recurrence in VHL syndrome. Although older age provides a protective effect, close monitoring of high-risk eyes is essential to enable timely intervention and preserve vision.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100846"},"PeriodicalIF":3.2,"publicationDate":"2025-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144595924","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Practical Guide to Evaluating Artificial Intelligence Imaging Models in Scientific Literature","authors":"Angela McCarthy ,&nbsp;Ives Valenzuela MD ,&nbsp;Royce W.S. Chen MD ,&nbsp;Lora R. Dagi Glass MD ,&nbsp;Kaveri Thakoor PhD","doi":"10.1016/j.xops.2025.100847","DOIUrl":"10.1016/j.xops.2025.100847","url":null,"abstract":"<div><h3>Objective</h3><div>Recent advances in artificial intelligence (AI) are revolutionizing ophthalmology by enhancing diagnostic accuracy, treatment planning, and patient management. However, a significant gap remains in practical guidance for ophthalmologists who lack AI expertise to effectively analyze these technologies and assess their readiness for integration into clinical practice. This paper aims to bridge this gap by demystifying AI model design and providing practical recommendations for evaluating AI imaging models in research publications.</div></div><div><h3>Design</h3><div>Educational review: synthesizing key considerations for evaluating AI papers in ophthalmology.</div></div><div><h3>Participants</h3><div>This paper draws on insights from an interdisciplinary team of ophthalmologists and AI experts with experience in developing and evaluating AI models for clinical applications.</div></div><div><h3>Methods</h3><div>A structured framework was developed based on expert discussions and a review of key methodological considerations in AI research.</div></div><div><h3>Main Outcome Measures</h3><div>A stepwise approach to evaluating AI models in ophthalmology, providing clinicians with practical strategies for assessing AI research.</div></div><div><h3>Results</h3><div>This guide offers broad recommendations applicable across ophthalmology and medicine.</div></div><div><h3>Conclusions</h3><div>As the landscape of health care continues to evolve, proactive engagement with AI will empower clinicians to lead the way in innovation while concurrently prioritizing patient safety and quality of care.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100847"},"PeriodicalIF":4.6,"publicationDate":"2025-06-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144722803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"OpenAI o1 Large Language Model Outperforms GPT-4o, Gemini 1.5 Flash, and Human Test Takers on Ophthalmology Board–Style Questions","authors":"Ryan Shean BA ,&nbsp;Tathya Shah BS ,&nbsp;Sina Sobhani BS ,&nbsp;Alan Tang BS ,&nbsp;Ali Setayesh BA ,&nbsp;Kyle Bolo MD ,&nbsp;Van Nguyen MD ,&nbsp;Benjamin Xu MD, PhD","doi":"10.1016/j.xops.2025.100844","DOIUrl":"10.1016/j.xops.2025.100844","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate and compare the performance of human test takers and three artificial intelligence (AI) models—OpenAI o1, ChatGPT-4o, and Gemini 1.5 Flash—on ophthalmology board–style questions, focusing on overall accuracy and performance stratified by ophthalmic subspecialty and cognitive complexity level.</div></div><div><h3>Design</h3><div>A cross-sectional study.</div></div><div><h3>Subjects</h3><div>Five hundred questions sourced from the <em>Basic and Clinical Science Course (BCSC)</em> and <em>EyeQuiz</em> question banks.</div></div><div><h3>Methods</h3><div>Three large language models interpreted the questions using standardized prompting procedures. Subanalysis was performed, stratifying the questions by subspecialty and complexity defined by the Buckwalter taxonomic schema. Statistical analysis, including the analysis of variance and McNemar test, was conducted to assess performance differences.</div></div><div><h3>Main Outcome Measures</h3><div>Accuracy of responses for each model and human test takers, stratified by subspecialty and cognitive complexity.</div></div><div><h3>Results</h3><div>OpenAI o1 achieved the highest overall accuracy (423/500, 84.6%), significantly outperforming GPT-4o (331/500, 66.2%; <em>P</em> &lt; 0.001) and Gemini (301/500, 60.2%; <em>P</em> &lt; 0.001). o1 demonstrated superior performance on both <em>BCSC</em> (228/250, 91.2%) and <em>EyeQuiz</em> (195/250, 78.0%) questions compared with GPT-4o (<em>BCSC</em>: 183/250, 73.2%; <em>EyeQuiz</em>: 148/250, 59.2%) and Gemini (<em>BCSC</em>: 163/250, 65.2%; <em>EyeQuiz</em>: 137/250, 54.8%). On <em>BCSC</em> questions, human performance was lower (64.5%) than Gemini 1.5 Flash (65.2%), GPT-4o (73.2%), and OpenAI o1 (91.2%) (<em>P</em> &lt; 0.001). OpenAI o1 outperformed other models in each of the nine ophthalmic subfields and three cognitive complexity levels.</div></div><div><h3>Conclusions</h3><div>OpenAI o1 outperformed GPT-4o, Gemini, and human test takers in answering ophthalmology board–style questions from two question banks and across three complexity levels. These findings highlight advances in AI technology and OpenAI o1’s growing potential as an adjunct in ophthalmic education and care.</div></div><div><h3>Financial Disclosure(s)</h3><div>The author(s) have no proprietary or commercial interest in any materials discussed in this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100844"},"PeriodicalIF":3.2,"publicationDate":"2025-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144569780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Retinograd-AI: An Open-Source Automated Fundus Autofluorescence Retinal Image Gradability Assessment for Inherited Retinal Diseases","authors":"Gunjan Naik PhD ,&nbsp;Saoud Al-Khuzaei MD, PhD ,&nbsp;Ismail Moghul PhD ,&nbsp;Thales A.C. de Guimaraes PhD, MD ,&nbsp;Sagnik Sen MD ,&nbsp;Malena Daich Varela MD, PhD ,&nbsp;Yichen Liu MSci ,&nbsp;Pallavi Bagga PhD ,&nbsp;Vincent Rocco PGCERT ,&nbsp;Dun Jack Fu MD, PhD ,&nbsp;Mariya Moosajee MD, PhD ,&nbsp;Savita Madhusudhan MD ,&nbsp;Andrew R. Webster MD ,&nbsp;Samantha De Silva MD, PhD ,&nbsp;Praveen J. Patel MD ,&nbsp;Omar A. Mahroo MD, PhD ,&nbsp;Susan M. Downes MD ,&nbsp;Michel Michaelides MD ,&nbsp;Konstantinos Balaskas MD ,&nbsp;Nikolas Pontikos PhD ,&nbsp;William Woof A. PhD","doi":"10.1016/j.xops.2025.100845","DOIUrl":"10.1016/j.xops.2025.100845","url":null,"abstract":"<div><h3>Purpose</h3><div>To develop an automated system for assessing the quality of fundus autofluorescence (FAF) images in patients with inherited retinal diseases (IRDs).</div></div><div><h3>Design</h3><div>A retrospective study of imaging data.</div></div><div><h3>Participants</h3><div>Patients with a confirmed molecular diagnosis of IRD who have undergone FAF imaging at Moorfields Eye Hospital.</div></div><div><h3>Methods</h3><div>A dataset of 2445 FAF images from patients with IRD was marked by 3 expert graders as either gradable (acceptable quality) or ungradable (poor quality), following a strict grading protocol. This dataset was used to train an artificial intelligence (AI) algorithm, Retinograd-AI, which was then applied to predict the gradability label of our entire dataset of 136 631 FAF images.</div></div><div><h3>Main Outcome Measures</h3><div>Fundus autofluorescence gradability of FAF images as predicted and validated against human assessment.</div></div><div><h3>Results</h3><div>Retinograd-AI achieves 91% accuracy on our held-out dataset of 133 images with an area under the receiver operator characteristic curve of 0.94, indicating high performance in distinguishing between gradable and ungradable images. Applying Retinograd-AI to our entire dataset, a small but significant positive association of gradability with age was found (ß = 0.002, <em>P</em> &lt; 0.001). Excluding X-linked conditions, 77.1% of images were rated as gradable in men and 82.3% in women (odds ratio = 1.43, <em>P</em> &lt; 0.001). By genotype, from the 30 most common genetic diagnoses in our dataset, the highest proportion of gradable images was in patients with disease-causing variants in <em>PRPH2</em> (93.1%), while the lowest was in <em>RDH12</em> (27.1%). Applying Retinograd-AI to filter images improved the accuracy of a gene prediction classifier from 33.8% to 68.9%. Retinograd-AI is open-sourced and available at <span><span>https://github.com/Eye2Gene/retinograd-ai</span><svg><path></path></svg></span>.</div></div><div><h3>Conclusions</h3><div>Retinograd-AI is an open-source AI model for automated retinal image quality assessment of FAF images in IRDs. Automated gradability assessment through Retinograd-AI enables large-scale analysis of retinal images and the development of robust analysis pipelines. Quality assessment is essential for the deployment of AI algorithms, such as Eye2Gene, into clinical settings. Due to the diverse nature of IRD pathologies, Retinograd-AI will be extended to other conditions, either in its current form or through transfer learning and fine-tuning.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100845"},"PeriodicalIF":3.2,"publicationDate":"2025-06-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144687389","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Re: Patel et al: Oral Antithrombotic Medication Is Associated with Improved Visual Acuity Outcomes in Eyes with Neovascular Age-Related Macular Degeneration","authors":"Pierre-Henry Gabrielle MD, PhD,&nbsp;Catherine Creuzot-Garcher MD, PhD","doi":"10.1016/j.xops.2025.100840","DOIUrl":"10.1016/j.xops.2025.100840","url":null,"abstract":"","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 6","pages":"Article 100840"},"PeriodicalIF":3.2,"publicationDate":"2025-06-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144569781","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}