Ophthalmology science最新文献

{"title":"Cover","authors":"","doi":"10.1016/j.xops.2026.101175","DOIUrl":"10.1016/j.xops.2026.101175","url":null,"abstract":"","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"6 4","pages":"Article 101175"},"PeriodicalIF":4.6,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147702780","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":"Exploration of Morphological Characteristics in Diabetic Macular Ischemia Utilizing a Large Language Model–Assisted Knowledge Graph","authors":"Miyo Yoshida MD,&nbsp;Tomoaki Murakami MD, PhD,&nbsp;Kenji Ishihara MD, PhD,&nbsp;Yuki Mori MD, PhD,&nbsp;Akitaka Tsujikawa MD, PhD","doi":"10.1016/j.xops.2026.101075","DOIUrl":"10.1016/j.xops.2026.101075","url":null,"abstract":"<div><h3>Purpose</h3><div>To explore novel clinical terminologies in diabetic macular ischemia (DMI) using a large language model-assisted knowledge graph (KG) constructed from published literature and to validate the findings using clinical data.</div></div><div><h3>Design</h3><div>A review incorporating KG construction and subsequent exploration of clinical terminologies, validated in an observational cohort study.</div></div><div><h3>Participants</h3><div>Sixty-six original and review articles on DMI published between July 2008 and March 2025 were identified through PubMed, MEDLINE, and Embase. Validation was performed using data from 156 eyes of 156 patients with vision-threatening DMI.</div></div><div><h3>Methods</h3><div>Using generative pre-trained transformer 4, article texts were processed into entity–relation triplets. Entities were annotated with 13 predefined clinical properties and assembled into a KG using Neo4j. Community detection via the Leiden algorithm grouped related entities into subgraphs. Interpretation of subgraphs led to 2 novel clinical concepts: Disorganization of MIddle retinal Layers (DMIL), defined as structural disruption between the inner nuclear layer and outer plexiform layer; and degenerative DMI, defined as DMI with retinal neurodegenerative findings.</div></div><div><h3>Main Outcome Measures</h3><div>Characteristics of KG-derived subgraphs and definition of novel clinical terminologies.</div></div><div><h3>Results</h3><div>The final KG contained 2408 entities and 8133 relations. Simplified graphs composed of highly frequent entities revealed structured relationships among important terminologies; for example, disease concepts, imaging modalities, vascular parameters, and visual acuity (VA) in DMI. Community detection showed uneven distribution of entities between subgraphs. Interpretation of subgraphs divided by community detection led to the identification of <em>DMIL</em> and <em>degenerative DMI</em>. In the validation cohort, eyes with <em>DMIL</em> had significantly worse VA than those without (0.301 [0.064–0.523] vs 0.000 [–0.079 to 0.111]; <em>P</em> &lt; 0.001). <em>Degenerative DMI</em> was also significantly associated with both greater capillary nonperfusion and VA reduction (<em>P</em> &lt; 0.001 for both).</div></div><div><h3>Conclusions</h3><div>Large language model–assisted KG construction enables objective synthesis of clinical literature and facilitates the exploration of known and unknown clinical characteristics in DMI.</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":"6 3","pages":"Article 101075"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189733","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":"Discrepancies between Fundus Photography and Multimodal Imaging in Mapping of Choroidal Tumor Borders","authors":"Darvy Dang MOrth ,&nbsp;Meghna Burmi MD ,&nbsp;Xavier Hadoux PhD ,&nbsp;Daniel McKay MBBS ,&nbsp;Maxime Jannaud MEng ,&nbsp;Myra B. McGuinness PhD ,&nbsp;Peter van Wijngaarden PhD ,&nbsp;Roderick O’Day DMedSci","doi":"10.1016/j.xops.2025.101057","DOIUrl":"10.1016/j.xops.2025.101057","url":null,"abstract":"<div><h3>Purpose</h3><div>Accurate choroidal tumor border mapping is required for their management. We compared border mapping accuracy between unimodal assessment (color fundus photography [CFP] or scanning laser ophthalmoscopy [SLO]) against a multimodal assessment (CFP, SLO, and OCT) and identified tumor characteristics that affect performance.</div></div><div><h3>Design</h3><div>A cross-sectional diagnostic accuracy study.</div></div><div><h3>Participants</h3><div>Sixty-four choroidal lesions (61% nevi, 39% melanomas; median basal diameter 5.65 mm, median thickness 1.85 mm) from 63 patients at tertiary ocular oncology clinics in Victoria, Australia. No separate control group was included.</div></div><div><h3>Methods</h3><div>Two ocular oncologists independently delineated lesion margins on CFP and SLO. Multimodal assessment was established by agreement. Agreement between unimodal and multimodal assessments was quantified using the 95th percentile Hausdorff Distance (HD95).</div></div><div><h3>Main Outcome Measures</h3><div>The HD95 in millimeters between unimodal and multimodal tumor borders. Dice coefficient summary statistics are also provided.</div></div><div><h3>Results</h3><div>Overall, unimodal CFP and SLO assessments had good agreement with multimodal assessments (median HD95 &lt;1 mm for each grader and device). However, HD95 was &gt;2 mm in 5% (grader 1) and 9% (grader 2) of CFP assessments and in 2% (grader 1) and 3% (grader 2) of SLO assessments. Nonpigmented and mixed-pigmented tumors showed significantly higher HD95 than pigmented lesions for most grader-modality pairs, particularly for grader 1 on CFP and SLO (<em>P</em> &lt; 0.05).</div></div><div><h3>Conclusions</h3><div>Choroidal tumor margin assessment was accurate on CFP and SLO as compared with a multimodal assessment that included enhanced-depth imaging OCT (EDI-OCT). However, the borders of a subset of tumors, especially those with reduced pigmentation, were inaccurately determined when using fundus photography alone. Incorporating EDI-OCT into choroidal tumor border mapping may reduce these discrepancies.</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":"6 3","pages":"Article 101057"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189244","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":"Cholinergic Improvement of Low-Flow Segmental Aqueous Humor Outflow in Humans","authors":"Alex S. Huang MD, PhD ,&nbsp;Mary Anne Garner PhD ,&nbsp;Mark E. Clark MEng ,&nbsp;Lindsay A. Rhodes MD ,&nbsp;Martin Kallab MD, PhD ,&nbsp;Hao F. Zhang PhD ,&nbsp;Robert N. Weinreb MD ,&nbsp;Clemens A. Strohmaier MD, PhD ,&nbsp;Christopher A. Girkin MD","doi":"10.1016/j.xops.2026.101070","DOIUrl":"10.1016/j.xops.2026.101070","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the influence of pharmacologic cholinergic stimulation on segmental aqueous humor outflow (AHO) in living human eyes of brain-dead organ donors.</div></div><div><h3>Design</h3><div>A prospective, nonrandomized, and interventional study.</div></div><div><h3>Subjects</h3><div>Brain-dead organ donors.</div></div><div><h3>Methods</h3><div>Aqueous angiography (AA) was performed using intracameral indocyanine green (ICG; 0.4%) in brain-dead organ donors to establish baseline segmental high-flow and low-flow (LF) AHO regions. The eyes were then treated using intracameral balanced salt solution (BSS; n = 5) or Miochol-E (n = 7; 10 mg/ml), which is US Food and Drug Administration-approved acetylcholine. Repeat AA was then performed using intracameral fluorescein (2%). Aqueous angiography images were acquired using a FLEX Spectralis (Heidelberg Engineering). Aqueous humor outflow signal intensity was quantified per quadrant or per eye in a masked fashion and compared across conditions.</div></div><div><h3>Main Outcome Measures</h3><div>Aqueous angiography fluorescent patterns, baseline AA signal intensity in various quadrants, and change in AA signal intensity after drug treatment or control.</div></div><div><h3>Results</h3><div>All eyes received initial baseline ICG AA. Segmental AHO was seen qualitatively and quantitatively across all quadrants (temporal: 18.4 ± 10.0; superior: 25.6 ± 16.5; nasal: 49.9 ± 28.6; and inferior: 25.9 ± 13.2 arbitrary units [AU], mean ± standard deviation). Consistent with prior reports, the greatest AHO was seen nasally (temporal vs nasal, <em>P</em> &lt; 0.001; superior vs. nasal, <em>P</em> = 0.012; and inferior vs nasal, <em>P</em> = 0.013). After BSS treatment, fluorescein AA showed very similar patterns to baseline ICG AA. After Miochol-E treatment, fluorescein AA showed qualitative AHO improvement in some baseline LF regions. After drug treatment, Miochol-E-treated eyes showed a greater quantitative increase in fluorescence intensity per quadrant (25.9 ± 31.7 AU; <em>P</em> &lt; 0.001) and per eye (25.9 ± 22.9; <em>P</em> = 0.024) compared to the BSS control condition.</div></div><div><h3>Conclusions</h3><div>Cholinergic stimulation enhances AHO in segmental regions as opposed to circumferentially around the limbus. In particular, focal improvement in baseline LF regions was observed, implying a segmental response to drug treatment. Understanding the biological and physiological basis underlying this responsiveness may lead to new pharmacological and surgical glaucoma treatments.</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":"6 3","pages":"Article 101070"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146079755","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":"Sex Differences and Age Distributions in Invasive Treatments for Chalazion and Hordeolum in Japan: A 9-Year Nationwide Claims Study","authors":"Masamitsu Kido MD, PhD ,&nbsp;Sayaka Kamada MD ,&nbsp;Tomo Suzuki MD, PhD ,&nbsp;Koji Kitazawa MD, PhD ,&nbsp;Yasufumi Tomioka MD, PhD ,&nbsp;Kengo Yoshii PhD ,&nbsp;Katsutoshi Shoda MD, PhD ,&nbsp;Chie Sotozono MD, PhD","doi":"10.1016/j.xops.2026.101067","DOIUrl":"10.1016/j.xops.2026.101067","url":null,"abstract":"<div><h3>Purpose</h3><div>To analyze nationwide trends and demographic patterns of invasive treatments for chalazion (ITC) and invasive treatments for hordeolum (ITH) in Japan.</div></div><div><h3>Design</h3><div>A retrospective population-based cohort study.</div></div><div><h3>Participants</h3><div>Patients undergoing ITC or ITH procedures from 2014 to 2022.</div></div><div><h3>Methods</h3><div>Using the National Database of Health Insurance Claims and Specific Health Checkups Open Data Japan, we calculated annual numbers and rates of ITC and ITH per 100 000 person-years. Sex-stratified and age-stratified demographic peak patterns were analyzed, and annual trends were evaluated using linear and Poisson regression models.</div></div><div><h3>Main Outcome Measures</h3><div>Procedure rates, demographic distributions, and annual trends of ITC and ITH.</div></div><div><h3>Results</h3><div>A total of 1 104 078 procedures (ITC: 465 379; ITH: 638 699) were recorded over the 9-year study period. The annual average rates were 40.9 for ITC and 56.2 for ITH per 100 000 person-years. Sex differences were prominent in younger groups: the overall female-to-male ratio was 1.1:1 for ITC and 1.0:1 for ITH, with a greater female predominance in youth (ITC: 2.0 at age 15–19 years; ITH: 1.7 at age 20–24 years). Invasive treatments for chalazion showed bimodal peaks in males (age 15–19 years: 40.9 procedures per 100 000 person-years; age 35–39 years: 56.0) and females (age 15–19 years: 80.6; age 30-34 years: 68.0). Invasive treatments for hordeolum showed bimodal peaks in males (age 10–14 years: 82.9; age 35-39 years: 72.1) and females (age 10–14 years: 102.4; age 30–34 years: 78.3). Both age-adjusted rates declined annually across all age groups and both sexes (<em>P</em> &lt; 0.0167), except for an increase in 0–4-year-old ITC cases (<em>P</em> &lt; 0.00088).</div></div><div><h3>Conclusions</h3><div>This study reveals sex-related and age-related demographic patterns and an overall decline in invasive eyelid procedures over the past decade and provides valuable insights into the epidemiology of common eyelid diseases and surgical trends in Japan.</div></div><div><h3>Financial Disclosure(s)</h3><div>The authors have no proprietary or commercial interest in any materials discussed in this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"6 3","pages":"Article 101067"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146168317","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":"Association between Dupilumab and Ocular Surface Diseases in Children","authors":"Mohammad Ayoubi BS ,&nbsp;Taher K. Eleiwa MD, PhD ,&nbsp;J. Anthony Chacko MD ,&nbsp;John D. Pemberton DO ,&nbsp;Hajirah N. Saeed MD, MPH ,&nbsp;Paul H. Phillips MD ,&nbsp;Abdelrahman M. Elhusseiny MD, MSc","doi":"10.1016/j.xops.2026.101076","DOIUrl":"10.1016/j.xops.2026.101076","url":null,"abstract":"","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"6 3","pages":"Article 101076"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189734","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: Most et al: Can Multimodal Large Language Models Diagnose Diabetic Retinopathy from Fundus Photos? A Quantitative Evaluation","authors":"Lavinia Loss Henriques MD ,&nbsp;Carolina Pelegrini Barbosa Gracitelli MD, PhD ,&nbsp;Fernando Roberte Zanetti MD, MSc ,&nbsp;Ricardo Luz Leitão Guerra MD, MSc","doi":"10.1016/j.xops.2026.101063","DOIUrl":"10.1016/j.xops.2026.101063","url":null,"abstract":"","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"6 3","pages":"Article 101063"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189816","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":"Ultrastructural Findings of the Internal Limiting Membrane with and without Epiretinal Membrane: A Comparative Electron Microscopy Study","authors":"Thananop L. Pothikamjorn MD ,&nbsp;Thanapong Somkijrungroj MD ,&nbsp;Marisa Prasanpanich MD ,&nbsp;Nuntachai Surawatsatien MD, MSc ,&nbsp;Wasee Tulvatana MD, MSc","doi":"10.1016/j.xops.2026.101071","DOIUrl":"10.1016/j.xops.2026.101071","url":null,"abstract":"&lt;div&gt;&lt;h3&gt;Purpose&lt;/h3&gt;&lt;div&gt;To describe and compare the histopathological and ultrastructural findings of the internal limiting membrane (ILM) with and without epiretinal membrane (ERM), and investigate differences between idiopathic and secondary ERM, using light and transmission electron microscopy.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Design&lt;/h3&gt;&lt;div&gt;Cross-sectional analytical study, with pathologists and image assessors masked to clinical and imaging information.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Participants&lt;/h3&gt;&lt;div&gt;Consecutive participants undergoing pars plana vitrectomy (PPV) with peeling of ILM, ERM, or both were enrolled. One eye per participant was included.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Methods&lt;/h3&gt;&lt;div&gt;Specimens were collected intraoperatively following a standardized protocol and analyzed for cellular morphology, extracellular matrix composition, and ultrastructural alterations. Fisher exact test and unpaired &lt;em&gt;t&lt;/em&gt; tests were used for group comparisons. Univariate and multivariate regression models assessed associations between clinical and imaging characteristics and histopathological features, with significance set at &lt;em&gt;P&lt;/em&gt; &lt; 0.001.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Main Outcome Measures&lt;/h3&gt;&lt;div&gt;Cellular and extracellular matrix ultrastructural features and their associations with clinical characteristics and imaging findings.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Results&lt;/h3&gt;&lt;div&gt;A total of 107 eyes (90 ILM with ERM; 17 ILM without ERM) from 117 participants were analyzed. Mean age was 64.4 ± 10.9 years, and 59.8% were female. Internal limiting membrane with ERM exhibited cellular proliferation with fibrous astrocytes, fibroblasts, myofibroblasts, fibrocytes, and retinal pigment epithelium (RPE) cells, accompanied by extracellular matrix remodeling and newly formed collagen. Common ultrastructural features included ILM vacuolization, Müller cell fragments, electron-lucent bands, and disorganized collagen fibrils. Intraocular lens implantation (incidence rate ratio [IRR] = 0.333, &lt;em&gt;P&lt;/em&gt; = 0.012) and proliferative diabetic retinopathy (IRR = 0.190, &lt;em&gt;P&lt;/em&gt; = 0.002) were associated with reduced ILM vacuolization. Secondary ERM demonstrated significantly higher proportions of RPE cells (40.5% vs 2.8%, &lt;em&gt;P&lt;/em&gt; &lt; 0.001), intraretinal cysts (73.0% vs 31.8%, &lt;em&gt;P&lt;/em&gt; &lt; 0.001), and subfoveal ellipsoidal band loss (64.1% vs 15.9%, &lt;em&gt;P&lt;/em&gt; &lt; 0.001) compared to idiopathic ERM.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Conclusions&lt;/h3&gt;&lt;div&gt;This study provides a comparative ultrastructural assessment of ILM and ERM, revealing distinct differences between idiopathic and secondary ERM. These findings suggest differing underlying biological processes that merit future mechanistic investigation. According to fibrocellular changes reported, macular disruption observed on OCT in older eyes may indicate susceptibility to secondary ERM, raising the possibility, yet to be validated, that ILM peeling during PPV for other indications could offer prophylactic benefit.&lt;/div&gt;&lt;/div&gt;&lt;div&gt;&lt;h3&gt;Financial Disclosure(s)&lt;","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"6 3","pages":"Article 101071"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189213","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":"Automated Nonperfusion Quantification in Diabetic Retinopathy on Ultra-Widefield Swept-Source OCT Angiography","authors":"Tai Yong Loh MBBS ,&nbsp;Juling Sia MBBS ,&nbsp;Wei Hing Seah MBBS ,&nbsp;Lingyi Zhuang MBBS ,&nbsp;Wenjun Song MD ,&nbsp;Yue Qiu ,&nbsp;Xiaofeng Shen M.Eng ,&nbsp;Zhongqing Yu M.Eng ,&nbsp;Ryan Tan MBBS ,&nbsp;Nuo Tang MBBS ,&nbsp;Yusra Asad MBBS ,&nbsp;Colin Ming Hui Goh MBBS ,&nbsp;Charmayne Xinyi Ang MBBS ,&nbsp;Celyn Chng MBBS ,&nbsp;Peiqi Lo MBBS ,&nbsp;Pavan Paniharam MBBS ,&nbsp;Ser Koon Goh MBBS ,&nbsp;Hnin Hnin Oo MBBS ,&nbsp;Min Wang MD, PhD ,&nbsp;Rupesh Agrawal MD ,&nbsp;Sandy Wenting Zhou MD","doi":"10.1016/j.xops.2025.101059","DOIUrl":"10.1016/j.xops.2025.101059","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the performance of a customized deep learning algorithm for automated segmentation of nonperfusion area (NPA) on ultra-widefield swept-source OCTA (UWF SS-OCTA) and its utility in diabetic retinopathy (DR) severity assessment.</div></div><div><h3>Design</h3><div>Cross-sectional study.</div></div><div><h3>Subjects</h3><div>A total of 180 eyes from 122 participants representing all grades of DR severity.</div></div><div><h3>Methods</h3><div>We developed a convolutional neural network based on a multiscale U-Net backbone with squeeze-and-excitation attention for segmentation of NPAs on en face SS-OCTA all-retinal-layer images from 3 scan patterns: 6 × 6 mm, 12 × 12 mm, and 29 × 24 mm. Ground-truth annotations of NPAs and nongradable area (NGA) on en face OCTA images were generated by 2 independent graders and adjudicated by a vitreoretinal specialist. A corresponding en face structural OCT image was incorporated to distinguish true NPAs from shadow artifacts. Segmentation outputs included NPA, NGA, and shadow artifacts. Pixel-level accuracy was assessed with the F1 score. Nonperfusion index (NPI) was defined as NPA/gradable area. The level of agreement between human-labeled and algorithm-predicted NPI was analyzed using Bland–Altman analysis.</div></div><div><h3>Main Outcome Measures</h3><div>Algorithm F1 score and NPI.</div></div><div><h3>Results</h3><div>The algorithm for NPA segmentation achieved a mean F1 score of 0.82 ± 0.01 in 6 × 6 mm, 0.84 ± 0.03 in 12 × 12 mm, and 0.83 ± 0.02 in 29 × 24 mm, with no significant difference across fields of view (<em>P</em> = 0.12). Algorithm-derived NPI strongly agreed with expert grading (intraclass correlation coefficient &gt;0.979). Both human- and algorithm-derived NPI increased progressively with increased DR severity in all scan patterns demonstrated by the Kruskal–Wallis test (6 × 6 mm: human: <em>P</em> = 0.02; algorithm: <em>P</em> = 0.03; 12 × 12 mm: algorithm <em>P</em> &lt; 0.001; human <em>P</em> &lt; 0.001; 29 × 24 mm: algorithm: <em>P</em> &lt; 0.001; human: <em>P</em> &lt; 0.001) with the largest magnitude of increase in 29 × 24 mm scans. The algorithm for foveal avascular zone segmentation also achieved a mean F1 score of 0.88 ± 0.05 for 6 × 6 mm images and 0.85 ± 0.05 for 12 × 12 mm images.</div></div><div><h3>Conclusions</h3><div>This deep learning algorithm was validated on single-scan UWF SS-OCTA for automated NPA segmentation and quantification. It demonstrates high accuracy and scalability across multiple scan sizes, supporting its potential integration into objective DR OCTA biomarker analysis.</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":"6 3","pages":"Article 101059"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146189243","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":"Convolutional Graph Isomorphism Network to Detect Glaucomatous Visual Field Defects","authors":"Douglas R. da Costa MD ,&nbsp;Dániel Unyi ,&nbsp;Rafael Scherer MD, PhD ,&nbsp;Rohit Muralidhar ,&nbsp;Mario Luiz Ribeiro Monteiro MD, PhD ,&nbsp;Felipe A. Medeiros MD, PhD","doi":"10.1016/j.xops.2025.101041","DOIUrl":"10.1016/j.xops.2025.101041","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the performance of a deep learning (DL) model based on graph isomorphism networks (GINs) for detecting glaucomatous visual field defects on 24-2 standard automated perimetry (SAP) and to compare it against traditional diagnostic criteria, a dense neural network (NN) model, and a convolutional neural network (CNN) model.</div></div><div><h3>Design</h3><div>A cross-sectional retrospective study. Participants: 1874 reliable SAP tests (Humphrey Field Analyzer, Carl-Zeiss Meditec Inc.) from 1009 eyes of 676 patients.</div></div><div><h3>Methods</h3><div>Standard automated perimetry tests were classified as normal or abnormal due to glaucomatous damage by two glaucoma specialists, with adjudication by a third. A GIN architecture was developed to classify tests using full 54-point spatial SAP data modeled as graphs, with node features comprising sensitivity, total deviation, and pattern deviation values. The dataset was split at the patient level (60% training/validation, 40% testing). The GIN model’s diagnostic performance was compared to the Anderson criteria, the glaucoma hemifield test/pattern standard deviation (GHT/PSD) criteria, a fully connected dense NN, and a CNN model.</div></div><div><h3>Main Outcome Measures</h3><div>Area under the receiver operating characteristic curve (AUC), precision–recall curve, sensitivity at 95% specificity, F1-score, repeatability, and model explainability.</div></div><div><h3>Results</h3><div>Among the 1874 SAP tests, 70.0% were graded as abnormal. The GIN model achieved an AUC of 0.982, significantly outperforming the Anderson criteria (AUC: 0.906, <em>P</em> &lt; 0.001), GHT/PSD (AUC: 0.936, <em>P</em> = 0.006), the NN model (AUC: 0.941, <em>P</em> = 0.007), and the CNN model (AUC: 0.941, <em>P</em> = 0.027). At 95% specificity, the GIN model reached the highest sensitivity of 94.1%, surpassing the NN model (88.3%), CNN model (92.0%), GHT/PSD (90.1%), and Anderson criteria (85.1%). The GIN model also achieved the highest average precision (0.952) among evaluated criteria. Explainability analysis using GraphNOSE demonstrated that the GIN model emphasized clinically relevant regions associated with glaucomatous loss, offering interpretability advantages over conventional DL approaches.</div></div><div><h3>Conclusions</h3><div>By modeling SAP as a graph and incorporating spatial relationships among test points, the GIN model provided superior diagnostic performance and interpretability relative to traditional criteria and standard NNs. This graph-based approach offers a promising tool for accurate and explainable detection of glaucomatous visual field defects in clinical practice.</div></div><div><h3>Financial Disclosures</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":"6 3","pages":"Article 101041"},"PeriodicalIF":4.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146079753","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}