{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T18:50:42Z","timestamp":1791226242158,"version":"4.1.0"},"reference-count":32,"publisher":"Public Library of Science (PLoS)","issue":"10","license":[{"start":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T00:00:00Z","timestamp":1791158400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"National Eye Institute","award":["R01 EY021797"],"award-info":[{"award-number":["R01 EY021797"]}],"id":[{"id":"https:\/\/ror.org\/03wkg3b53","id-type":"ROR","asserted-by":"crossref"}]},{"name":"National Eye Institute","award":["R01 EY028557"],"award-info":[{"award-number":["R01 EY028557"]}],"id":[{"id":"https:\/\/ror.org\/03wkg3b53","id-type":"ROR","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000900","name":"California Institute for Regenerative Medicine","doi-asserted-by":"publisher","award":["TR2-01768"],"award-info":[{"award-number":["TR2-01768"]}],"id":[{"id":"10.13039\/100000900","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000900","name":"California Institute for Regenerative Medicine","doi-asserted-by":"publisher","award":["CLIN1-08686"],"award-info":[{"award-number":["CLIN1-08686"]}],"id":[{"id":"10.13039\/100000900","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000900","name":"California Institute for Regenerative Medicine","doi-asserted-by":"publisher","award":["CLIN2-11650"],"award-info":[{"award-number":["CLIN2-11650"]}],"id":[{"id":"10.13039\/100000900","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100000900","name":"California Institute for Regenerative Medicine","doi-asserted-by":"publisher","award":["TRAN1-14649"],"award-info":[{"award-number":["TRAN1-14649"]}],"id":[{"id":"10.13039\/100000900","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>\n                    Limbal Stem Cell Deficiency (LSCD) is a corneal disorder characterized by conjunctivalization of the cornea. Central corneal epithelial thickness, measured using Anterior Segment Optical Coherence Tomography (AS-OCT), is a biomarker for LSCD. Manual measurement of epithelial thickness remains the gold standard for diagnosis because existing automated commercial software is unreliable in severe cases. Standard machine learning approaches have limited applicability for automating this measurement due to the difficulty of distinguishing corneal epithelium from scar tissue, and imprecise clinical labels. The goal of this study is to develop a deep learning pipeline to automate measurement of central epithelial thickness in AS-OCT images of eyes with LSCD. To accomplish this goal, we developed AI-CE, a model based on Multi-scale U-Net (MSU-Net) trained to accurately delineate the epithelial layer. The primary dataset comprised of 134 patient eyes, including 47 healthy normal eyes (35.0%), 34 with Stage 1 (25.4%), 34 with Stage 2 (25.4%), and 19 with Stage 3 (14.2%) LSCD. Performance was evaluated by calculating the average mean absolute error between model segmentations and manually measured clinical assessments. AI-CE model performance was compared against a commercially available segmentation software to demonstrate its improvement over the current state-of-the-art. The proposed model achieved a mean absolute error of 5.10\n                    <jats:inline-formula id=\"pone.0358258.e001\">\n                      <jats:alternatives>\n                        <jats:graphic xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" id=\"pone.0358258.e001g\" mimetype=\"image\" position=\"anchor\" xlink:href=\"info:doi\/10.1371\/journal.pone.0358258.e001\" xlink:type=\"simple\"\/>\n                        <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" display=\"inline\" id=\"M1\">\n                          <mml:mrow>\n                            <mml:mi>\u03bc<\/mml:mi>\n                          <\/mml:mrow>\n                        <\/mml:math>\n                      <\/jats:alternatives>\n                    <\/jats:inline-formula>\n                    m, significantly improving (p\u2009&lt;\u20090.01) over clinical software (12.20\n                    <jats:inline-formula id=\"pone.0358258.e002\">\n                      <jats:alternatives>\n                        <jats:graphic xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" id=\"pone.0358258.e002g\" mimetype=\"image\" position=\"anchor\" xlink:href=\"info:doi\/10.1371\/journal.pone.0358258.e002\" xlink:type=\"simple\"\/>\n                        <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" display=\"inline\" id=\"M2\">\n                          <mml:mrow>\n                            <mml:mi>\u03bc<\/mml:mi>\n                          <\/mml:mrow>\n                        <\/mml:math>\n                      <\/jats:alternatives>\n                    <\/jats:inline-formula>\n                    m). This result was verified in a secondary retrospective dataset which similarly showed significant improvement over commercial software. We also demonstrate that clinical epithelial thickness values provided by AI-CE can distinguish between LSCD stages similarly to manual evaluations.\n                  <\/jats:p>","DOI":"10.1371\/journal.pone.0358258","type":"journal-article","created":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T17:53:20Z","timestamp":1791222800000},"page":"e0358258","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Automated quantification of central corneal epithelial thickness for the diagnosis of limbal stem cell deficiency"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0002-0840-6235","authenticated-orcid":true,"given":"Micah","family":"Vinet","sequence":"first","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Jules Stein Eye Institute"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Department of Bioengineering"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Biomedical AI Research Lab"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7140-3937","authenticated-orcid":true,"given":"Nathan","family":"Siu","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Jules Stein Eye Institute"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Biomedical AI Research Lab"]},{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Medical Informatics Home Area, Graduate Programs in Bioscience"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1732-7643","authenticated-orcid":true,"given":"Clemence","family":"Bonnet","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Jules Stein Eye Institute"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-4084-7583","authenticated-orcid":true,"given":"Craig","family":"Choi","sequence":"additional","affiliation":[{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Department of Computational and Systems Biology, Graduate Programs in Bioscience"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-0245-5567","authenticated-orcid":true,"given":"Sarah","family":"Wu","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Biomedical AI Research Lab"]},{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Medical Informatics Home Area, Graduate Programs in Bioscience"]},{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Department of Computational and Systems Biology, Graduate Programs in Bioscience"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Celso","family":"Nguyen","sequence":"additional","affiliation":[{"name":"University of California","place":["Berkeley, California, United States of America"],"department":["Department of Molecular and Cell Biology"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Corey","family":"Arnold","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Department of Bioengineering"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Biomedical AI Research Lab"]},{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Medical Informatics Home Area, Graduate Programs in Bioscience"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Department of Radiology, David Geffen School of Medicine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"William","family":"Speier","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Department of Bioengineering"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Biomedical AI Research Lab"]},{"name":"UCLA","place":["Los Angeles, California, United States of America"],"department":["Medical Informatics Home Area, Graduate Programs in Bioscience"]},{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Department of Radiology, David Geffen School of Medicine"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sophie X.","family":"Deng","sequence":"additional","affiliation":[{"name":"University of California Los Angeles","place":["Los Angeles, California, United States of America"],"department":["Jules Stein Eye Institute"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,10,5]]},"reference":[{"issue":"3","key":"pone.0358258.ref001","doi-asserted-by":"crossref","first-page":"364","DOI":"10.1097\/ICO.0000000000001820","article-title":"Global consensus on definition, classification, diagnosis, and staging of limbal stem cell deficiency","volume":"38","author":"SX Deng","year":"2019","journal-title":"Cornea"},{"key":"pone.0358258.ref002","doi-asserted-by":"crossref","first-page":"123","DOI":"10.1016\/j.jtos.2021.12.005","article-title":"Biomarkers of in vivo limbal stem cell function","volume":"23","author":"Q Le","year":"2022","journal-title":"Ocul Surf"},{"key":"pone.0358258.ref003","doi-asserted-by":"crossref","first-page":"132","DOI":"10.1016\/j.ajo.2020.04.006","article-title":"Corneal epithelial thickness measured using anterior segment optical coherence tomography as a diagnostic parameter for limbal stem cell deficiency","volume":"216","author":"Q Liang","year":"2020","journal-title":"Am J Ophthalmol"},{"key":"pone.0358258.ref004","first-page":"174","article-title":"MobileNetV2 Based Diagnosis and Grading of Limbal Stem Cell Deficiency.","volume-title":"2022 IEEE BIBE","author":"P Liu","year":"2022"},{"key":"pone.0358258.ref005","doi-asserted-by":"crossref","first-page":"1270570","DOI":"10.3389\/fmed.2023.1270570","article-title":"Latent diffusion augmentation enhances deep learning analysis of neuro-morphology in limbal stem cell deficiency","volume":"10","author":"D Gibson","year":"2023","journal-title":"Front Med (Lausanne)"},{"key":"pone.0358258.ref006","unstructured":"U.S. Food and Drug Administration website. Optovue Inc. iVue, 510(k) Summary. 2017. www.accessdata.fda.gov\/cdrh_docs\/pdf16\/K163475.pdf"},{"issue":"11","key":"pone.0358258.ref007","doi-asserted-by":"crossref","first-page":"1353","DOI":"10.1097\/ICO.0000000000003012","article-title":"Corneal Epithelial thickness mapping in the diagnosis of ocular surface disorders involving the corneal epithelium: a comparative study","volume":"41","author":"A Levy","year":"2022","journal-title":"Cornea"},{"issue":"13","key":"pone.0358258.ref008","doi-asserted-by":"crossref","first-page":"3830","DOI":"10.1049\/ipr2.12900","article-title":"Uncertainty\u2010aware iterative learning for noisy\u2010labeled medical image segmentation","volume":"17","author":"P Hao","year":"2023","journal-title":"IET Image Process"},{"key":"pone.0358258.ref009","doi-asserted-by":"crossref","first-page":"36827","DOI":"10.1109\/ACCESS.2021.3062380","article-title":"Medical image segmentation with limited supervision: a review of deep network models","volume":"9","author":"J Peng","year":"2021","journal-title":"IEEE Access"},{"issue":"17","key":"pone.0358258.ref010","doi-asserted-by":"crossref","first-page":"3771","DOI":"10.3390\/math11173771","article-title":"Enhancing medical image segmentation: ground truth optimization through evaluating uncertainty in expert annotations","volume":"11","author":"G Athanasiou","year":"2023","journal-title":"Mathematics"},{"key":"pone.0358258.ref011","doi-asserted-by":"crossref","unstructured":"Theodoridis J, Hofmann J, Maucher J, Schilling A. Trapped in texture bias? A large scale comparison of deep instance segmentation. In: Lecture Notes in Computer Science; 2022. pp. 609\u201327.","DOI":"10.1007\/978-3-031-20074-8_35"},{"issue":"9","key":"pone.0358258.ref012","doi-asserted-by":"crossref","first-page":"095017","DOI":"10.1088\/1361-6560\/aabd19","article-title":"Superpixel-based and boundary-sensitive convolutional neural network for automated liver segmentation","volume":"63","author":"W Qin","year":"2018","journal-title":"Phys Med Biol"},{"issue":"1","key":"pone.0358258.ref013","doi-asserted-by":"crossref","first-page":"17345","DOI":"10.1038\/s41598-024-67554-0","article-title":"Boundary guidance network for medical image segmentation","volume":"14","author":"R Xu","year":"2024","journal-title":"Sci Rep"},{"issue":"1","key":"pone.0358258.ref014","doi-asserted-by":"crossref","first-page":"9924","DOI":"10.1038\/s41598-025-94437-9","article-title":"A novel framework for segmentation of small targets in medical images","volume":"15","author":"L Zhao","year":"2025","journal-title":"Sci Rep"},{"issue":"5","key":"pone.0358258.ref015","doi-asserted-by":"crossref","first-page":"6289","DOI":"10.1109\/TPAMI.2022.3211171","article-title":"Small-object sensitive segmentation using across feature map attention","volume":"45","author":"S Sang","year":"2023","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"pone.0358258.ref016","doi-asserted-by":"crossref","unstructured":"Konrad PM, Popa AA, et al. Challenges in Deep Learning-Based Small Organ Segmentation: A Benchmarking Perspective for Medical Research with Limited Datasets. 2026. doi: 10.48550\/arXiv.2509.05892","DOI":"10.2139\/ssrn.5529533"},{"key":"pone.0358258.ref017","first-page":"234","article-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation","volume-title":"Lecture Notes in Computer Science","author":"O Ronneberger","year":"2015"},{"issue":"2","key":"pone.0358258.ref018","doi-asserted-by":"crossref","first-page":"622","DOI":"10.1364\/BOE.10.000622","article-title":"CorneaNet: Fast segmentation of cornea OCT scans of healthy and keratoconic eyes using deep learning","volume":"10","author":"VA Dos Santos","year":"2019","journal-title":"Biomed Opt Express"},{"issue":"9","key":"pone.0358258.ref019","doi-asserted-by":"crossref","first-page":"1231","DOI":"10.1136\/bjophthalmol-2019-315723","article-title":"Deep learning algorithms to isolate and quantify the structures of the anterior segment in optical coherence tomography images","volume":"105","author":"TH Pham","year":"2021","journal-title":"Br J Ophthalmol"},{"issue":"8","key":"pone.0358258.ref020","doi-asserted-by":"crossref","first-page":"3968","DOI":"10.1364\/BOE.493065","article-title":"Comprehensive assessment of the anterior segment in refraction corrected OCT based on multitask learning","volume":"14","author":"K Li","year":"2023","journal-title":"Biomed Opt Express"},{"key":"pone.0358258.ref021","doi-asserted-by":"crossref","first-page":"105851","DOI":"10.1016\/j.bspc.2023.105851","article-title":"Automatic simultaneous ciliary muscle segmentation and biomarker extraction in AS-OCT images using deep learning-based approaches","volume":"90","author":"E Goyanes","year":"2024","journal-title":"Biomed Signal Process Control"},{"key":"pone.0358258.ref022","doi-asserted-by":"crossref","first-page":"108143","DOI":"10.1016\/j.compbiomed.2024.108143","article-title":"SeqCorr-EUNet: A sequence correction dual-flow network for segmentation and quantification of anterior segment OCT image","volume":"171","author":"J Fang","year":"2024","journal-title":"Comput Biol Med"},{"issue":"3","key":"pone.0358258.ref023","doi-asserted-by":"crossref","first-page":"19","DOI":"10.1167\/tvst.14.3.19","article-title":"Corneal layer segmentation in healthy and pathological eyes: A joint super-resolution generative adversarial network and adaptive graph theory approach","volume":"14","author":"KY Win","year":"2025","journal-title":"Transl Vis Sci Technol"},{"issue":"11","key":"pone.0358258.ref024","doi-asserted-by":"crossref","first-page":"24","DOI":"10.1167\/tvst.9.11.24","article-title":"Pathological-corneas layer segmentation and thickness measurement in OCT images","volume":"9","author":"A Elsawy","year":"2020","journal-title":"Transl Vis Sci Technol"},{"issue":"5","key":"pone.0358258.ref025","doi-asserted-by":"crossref","first-page":"708","DOI":"10.1007\/s10384-025-01242-y","article-title":"Corneal irregularity of the anterior and posterior surface in patients with limbal stem cell deficiency evaluated with anterior-segment optical coherence tomography","volume":"69","author":"H Goto","year":"2025","journal-title":"Jpn J Ophthalmol"},{"key":"pone.0358258.ref026","doi-asserted-by":"crossref","first-page":"182","DOI":"10.1016\/j.ajo.2025.02.019","article-title":"Detection of ocular surface squamous neoplasia using artificial intelligence with anterior segment optical coherence tomography","volume":"273","author":"JA Greenfield","year":"2025","journal-title":"Am J Ophthalmol"},{"issue":"4","key":"pone.0358258.ref027","doi-asserted-by":"crossref","first-page":"422","DOI":"10.1097\/ICO.0000000000003682","article-title":"Comparison of automated and manual measurement of corneal epithelial thickness in eyes with limbal stem cell deficiency","volume":"44","author":"T Kittipibul","year":"2025","journal-title":"Cornea"},{"key":"pone.0358258.ref028","doi-asserted-by":"crossref","DOI":"10.1609\/aaai.v38i7.28490","article-title":"DiffusionEdge: Diffusion Probabilistic Model for Crisp Edge Detection.","volume-title":"2024 AAAI Conference on Artificial Intelligence","author":"Y Ye","year":"2024"},{"key":"pone.0358258.ref029","doi-asserted-by":"crossref","first-page":"639930","DOI":"10.3389\/fgene.2021.639930","article-title":"MSU-Net: Multi-Scale U-Net for 2D Medical Image Segmentation","volume":"12","author":"R Su","year":"2021","journal-title":"Front Genet"},{"issue":"4","key":"pone.0358258.ref030","doi-asserted-by":"crossref","first-page":"1256","DOI":"10.3390\/jcm14041256","article-title":"Corneal epithelial thickness maps in eyes with mild and moderate keratoconus","volume":"14","author":"P Mlyniuk","year":"2023","journal-title":"J Clin Med"},{"issue":"5","key":"pone.0358258.ref031","doi-asserted-by":"crossref","first-page":"198","DOI":"10.1097\/ICL.0000000000001267","article-title":"Post-orthokeratology changes in the corneal epithelium induced by an orthokeratology lens: An anterior-segment optical coherence tomography assessment","volume":"52","author":"K Nemoto","year":"2016","journal-title":"Eye Contact Lens"},{"issue":"3","key":"pone.0358258.ref032","doi-asserted-by":"crossref","first-page":"232","DOI":"10.1016\/j.clae.2020.02.015","article-title":"Redistribution of the corneal epithelium after overnight wear of orthokeratology contact lenses for myopia reduction","volume":"43","author":"J Zhang","year":"2020","journal-title":"Contact Lens Anterior Eye"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0358258","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,5]],"date-time":"2026-10-05T17:53:27Z","timestamp":1791222807000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0358258"}},"subtitle":[],"editor":[{"given":"Michele","family":"Madigan","sequence":"first","affiliation":[{"name":"Save Sight Institute","place":["AUSTRALIA"]}],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,10,5]]},"references-count":32,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,10,5]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0358258","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,5]]}}}