{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,10]],"date-time":"2026-10-10T15:06:45Z","timestamp":1791644805338,"version":"4.3.4"},"reference-count":33,"publisher":"Springer Science and Business Media LLC","issue":"4","license":[{"start":{"date-parts":[[2026,1,10]],"date-time":"2026-01-10T00:00:00Z","timestamp":1768003200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,1,10]],"date-time":"2026-01-10T00:00:00Z","timestamp":1768003200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Graefes Arch Clin Exp Ophthalmol"],"published-print":{"date-parts":[[2026,4]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec>\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>This study aims to develop an artificial intelligence (AI) model to assist ophthalmologists in distinguishing ocular surface squamous neoplasia (OSSN) from benign ocular surface lesions using ultrasound biomicroscopy (UBM) images.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>Data were retrospectively collected from 139 patients with biopsy-proven conjunctival lesions, including 201 UBM images of benign lesions (e.g.,pterygium, squamous papilloma) and 381 images of OSSN (e.g.,squamous cell carcinoma, conjunctival intraepithelial neoplasia). Patients with conjunctival pigmented lesions, melanoma, lymphoma, or those without a pathological diagnosis were excluded. UBM images were cropped to the anterior segment region and rescaled to a standard size of 300\u2009\u00d7\u2009200 pixels. Data augmentation techniques were applied to enhance the diversity of training images. A convolutional neural network was trained and tested using five-fold cross-validation. A heatmap was generated to illustrate the model\u2019s decision-making process. The AI model\u2019s performance was compared to that of three human experts with varying levels of experience. Additionally, univariate regression analysis was performed to assess the impact of patient-related factors (age, sex, race\/ethnicity, lesion location, and side) on model performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>\n                      Our AI model achieved an accuracy of 74.3\u2009\u00b1\u20093.9%, sensitivity of 75.0\u2009\u00b1\u20098.6%, specificity 73.0\u2009\u00b1\u200911.5%, precision of 83.3\u2009\u00b1\u20094.8%, F1 score (i.e., the harmonic mean of precision and recall) of 0.79\u2009\u00b1\u20090.06,and area under the receiver operating characteristic (AUROC) curve of 0.83\u2009\u00b1\u20090.03 in detection of OSSN. It significantly outperformed two ocular oncology fellows (\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.02 and 0.03, respectively) and demonstrated borderline significance compared to a senior ophthalmologist (\n                      <jats:italic>p<\/jats:italic>\n                      \u2009=\u20090.05). The heatmaps effectively highlighted the lesions, suggesting that echogenicity played a crucial role in the model\u2019s predictions. None of the patient-related factors significantly affected model performance (all\n                      <jats:italic>p<\/jats:italic>\n                      \u2009&gt;\u20090.1), supporting its equitable diagnostic capability across diverse patient groups.\n                    <\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec>\n                    <jats:title>Conclusion<\/jats:title>\n                    <jats:p>This study demonstrates the feasibility of using AI to differentiate OSSN from benign conjunctival lesions based on UBM images. The heatmap enhances model transparency, and the consistent performance across patient subgroups highlights its potential as a fair and valuable tool for clinical decision-making in ocular surface tumor evaluation.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1007\/s00417-025-07034-x","type":"journal-article","created":{"date-parts":[[2026,1,10]],"date-time":"2026-01-10T09:18:35Z","timestamp":1768036715000},"page":"1149-1157","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Utilizing artificial intelligence for the diagnosis of ocular surface squamous neoplasia with ultrasound biomicroscopy images"],"prefix":"10.1007","volume":"264","author":[{"given":"Kubra","family":"Serbest Ceylanoglu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhao","family":"Zhenyang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bernadete","family":"Ayres","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8465-130X","authenticated-orcid":false,"given":"Yike","family":"Li","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1593-1476","authenticated-orcid":false,"given":"Hakan","family":"Demirci","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,1,10]]},"reference":[{"key":"7034_CR1","doi-asserted-by":"publisher","first-page":"808","DOI":"10.1001\/ARCHOPHT.1997.01100150810025","volume":"115","author":"JA Shields","year":"1997","unstructured":"Shields JA, Shields CL, De Potter P (1997) Surgical management of conjunctival tumors. The 1994 Lynn B. McMahan lecture. Arch Ophthalmol 115:808\u2013815. https:\/\/doi.org\/10.1001\/ARCHOPHT.1997.01100150810025","journal-title":"Arch Ophthalmol"},{"key":"7034_CR2","doi-asserted-by":"publisher","first-page":"1287","DOI":"10.1097\/ICO.0000000000001651","volume":"37","author":"C Bellerive","year":"2018","unstructured":"Bellerive C, Berry JL, Polski A, Singh AD (2018) Conjunctival squamous neoplasia: staging and initial treatment. Cornea 37:1287\u20131291. https:\/\/doi.org\/10.1097\/ICO.0000000000001651","journal-title":"Cornea"},{"key":"7034_CR3","doi-asserted-by":"publisher","first-page":"429","DOI":"10.1016\/S0039-6257(05)80054-2","volume":"39","author":"GA Lee","year":"1995","unstructured":"Lee GA, Hirst LW (1995) Ocular surface squamous neoplasia. Surv Ophthalmol 39:429\u2013450. https:\/\/doi.org\/10.1016\/S0039-6257(05)80054-2","journal-title":"Surv Ophthalmol"},{"key":"7034_CR4","doi-asserted-by":"publisher","first-page":"98","DOI":"10.1136\/BJO.83.1.98","volume":"83","author":"M Tunc","year":"1999","unstructured":"Tunc M, Char DH, Crawford B, Miller T (1999) Intraepithelial and invasive squamous cell carcinoma of the conjunctiva: analysis of 60 cases. Br J Ophthalmol 83:98\u2013103. https:\/\/doi.org\/10.1136\/BJO.83.1.98","journal-title":"Br J Ophthalmol"},{"key":"7034_CR5","doi-asserted-by":"publisher","first-page":"244","DOI":"10.1016\/S0161-6420(90)32598-8","volume":"97","author":"CJ Pavlin","year":"1990","unstructured":"Pavlin CJ, Sherar MD, Foster FS (1990) Subsurface ultrasound microscopic imaging of the intact eye. Ophthalmology 97:244\u2013250. https:\/\/doi.org\/10.1016\/S0161-6420(90)32598-8","journal-title":"Ophthalmology"},{"key":"7034_CR6","doi-asserted-by":"publisher","first-page":"1297","DOI":"10.1016\/J.OPHTHA.2010.11.011","volume":"118","author":"C Bianciotto","year":"2011","unstructured":"Bianciotto C, Shields CL, Guzman JM et al (2011) Assessment of anterior segment tumors with ultrasound biomicroscopy versus anterior segment optical coherence tomography in 200 cases. Ophthalmology 118:1297\u20131302. https:\/\/doi.org\/10.1016\/J.OPHTHA.2010.11.011","journal-title":"Ophthalmology"},{"key":"7034_CR7","doi-asserted-by":"publisher","DOI":"10.1016\/J.PRETEYERES.2019.04.003","volume":"72","author":"DSW Ting","year":"2019","unstructured":"Ting DSW, Peng L, Varadarajan AV et al (2019) Deep learning in ophthalmology: the technical and clinical considerations. Prog Retin Eye Res 72:100759. https:\/\/doi.org\/10.1016\/J.PRETEYERES.2019.04.003","journal-title":"Prog Retin Eye Res"},{"key":"7034_CR8","doi-asserted-by":"publisher","first-page":"5200","DOI":"10.1167\/IOVS.16-19964","volume":"57","author":"MD Abr\u00e0moff","year":"2016","unstructured":"Abr\u00e0moff MD, Lou Y, Erginay A et al (2016) Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning. Invest Ophthalmol Vis Sci 57:5200\u20135206. https:\/\/doi.org\/10.1167\/IOVS.16-19964","journal-title":"Invest Ophthalmol Vis Sci"},{"key":"7034_CR9","doi-asserted-by":"publisher","first-page":"962","DOI":"10.1016\/J.OPHTHA.2017.02.008","volume":"124","author":"R Gargeya","year":"2017","unstructured":"Gargeya R, Leng T (2017) Automated identification of diabetic retinopathy using deep learning. Ophthalmology 124:962\u2013969. https:\/\/doi.org\/10.1016\/J.OPHTHA.2017.02.008","journal-title":"Ophthalmology"},{"key":"7034_CR10","doi-asserted-by":"publisher","first-page":"378","DOI":"10.1016\/J.JFO.2013.01.006","volume":"36","author":"JR F\u00e9nolland","year":"2013","unstructured":"F\u00e9nolland JR, Puech M, Baudouin C, Labb\u00e9 A (2013) Imaging of the iridocorneal angle in glaucoma. J Fr Ophtalmol 36:378\u2013383. https:\/\/doi.org\/10.1016\/J.JFO.2013.01.006","journal-title":"J Fr Ophtalmol"},{"key":"7034_CR11","doi-asserted-by":"publisher","first-page":"476","DOI":"10.3928\/15428877-20081101-02","volume":"39","author":"JPS Garcia","year":"2008","unstructured":"Garcia JPS, Rosen RB (2008) Anterior segment imaging: optical coherence tomography versus ultrasound biomicroscopy. Ophthalmic Surg Lasers Imaging 39:476\u2013484. https:\/\/doi.org\/10.3928\/15428877-20081101-02","journal-title":"Ophthalmic Surg Lasers Imaging"},{"key":"7034_CR12","doi-asserted-by":"publisher","first-page":"7","DOI":"10.1016\/j.ohc.2003.12.001","volume":"17","author":"H Ishikawa","year":"2004","unstructured":"Ishikawa H, Schuman JS (2004) Anterior segment imaging: ultrasound biomicroscopy. Ophthalmol Clin North Am 17:7\u201320. https:\/\/doi.org\/10.1016\/j.ohc.2003.12.001","journal-title":"Ophthalmol Clin North Am"},{"key":"7034_CR13","doi-asserted-by":"publisher","unstructured":"Chollet F (2017) Xception: deep learning with depthwise separable convolutions. Proc \u2013 30thIEEE Conf Comput Vis Pattern Recognit CVPR 2017 1800\u20131807. https:\/\/doi.org\/10.1109\/CVPR.2017.195","DOI":"10.1109\/CVPR.2017.195"},{"key":"7034_CR14","unstructured":"Keras Applications. Accessed March 29 (2025) https:\/\/keras.io\/api\/applications\/"},{"key":"7034_CR15","doi-asserted-by":"publisher","first-page":"15365","DOI":"10.1038\/S41598-022-19639-X","volume":"12","author":"Y Liu","year":"2022","unstructured":"Liu Y, Zhang L, Hao Z et al (2022) An Xception model based on residual attention mechanism for the classification of benign and malignant gastric ulcers. Sci Rep 12:15365. https:\/\/doi.org\/10.1038\/S41598-022-19639-X","journal-title":"Sci Rep"},{"issue":"19","key":"7034_CR16","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/AC8F11","volume":"67","author":"S Yi","year":"2022","unstructured":"Yi S, Wei Y, Luo X, Chen D (2022) Diagnosis of rectal cancer based on the Xception-MS network. Phys Med Biol 67(19):10.1088\/1361\u20136560\/ac8f11. https:\/\/doi.org\/10.1088\/1361-6560\/AC8F11","journal-title":"Phys Med Biol"},{"key":"7034_CR17","doi-asserted-by":"publisher","first-page":"1948","DOI":"10.1053\/J.GASTRO.2022.02.025","volume":"162","author":"N Cheng","year":"2022","unstructured":"Cheng N, Ren Y, Zhou J et al (2022) Deep learning-based classification of hepatocellular nodular lesions on whole-slide histopathologic images. Gastroenterology 162:1948-1961.e7. https:\/\/doi.org\/10.1053\/J.GASTRO.2022.02.025","journal-title":"Gastroenterology"},{"issue":"15","key":"7034_CR18","doi-asserted-by":"publisher","DOI":"10.3390\/DIAGNOSTICS13152583","volume":"13","author":"K Shaheed","year":"2023","unstructured":"Shaheed K, Abbas Q, Hussain A, Qureshi I (2023) Optimized Xception learning model and XGBoost classifier for detection of multiclass chest disease from X-ray images. Diagnostics 13(15):2583. https:\/\/doi.org\/10.3390\/DIAGNOSTICS13152583","journal-title":"Diagnostics"},{"key":"7034_CR19","doi-asserted-by":"publisher","first-page":"1","DOI":"10.3390\/S20247253","volume":"20","author":"X Duan","year":"2020","unstructured":"Duan X, Gou M, Liu N, Wang W, Qin C (2020) High-capacity image steganography based on improved Xception. Sens (Basel) 20:1\u201316. https:\/\/doi.org\/10.3390\/S20247253","journal-title":"Sens (Basel)"},{"key":"#cr-split#-7034_CR20.1","unstructured":"Kingma DP, Ba J (2015) Adam: A Method for Stochastic Optimization. In: International Conference on Learning Representations."},{"key":"#cr-split#-7034_CR20.2","unstructured":"2015. Accessed April 22, 2019. http:\/\/arxiv.org\/abs\/1412.6980"},{"key":"7034_CR21","doi-asserted-by":"publisher","unstructured":"Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D (2017) Grad-CAM: visual explanations from deep networks via Gradient-Based localization. Proc IEEE Int Conf Comput Vis 618\u2013626. https:\/\/doi.org\/10.1109\/ICCV.2017.74","DOI":"10.1109\/ICCV.2017.74"},{"key":"7034_CR22","doi-asserted-by":"publisher","DOI":"10.2196\/51706","volume":"26","author":"B Chen","year":"2024","unstructured":"Chen B, Li Y, Sun Y et al (2024) A 3D and explainable artificial intelligence model for evaluation of chronic otitis media based on temporal bone computed tomography: model development, validation, and clinical application. J Med Internet Res 26:e51706. https:\/\/doi.org\/10.2196\/51706","journal-title":"J Med Internet Res"},{"key":"7034_CR23","doi-asserted-by":"publisher","DOI":"10.3390\/BIOENGINEERING12030282","volume":"12","author":"Y Yu","year":"2025","unstructured":"Yu Y, Jia H, Zhang L et al (2025) Deep multi-modal skin-imaging-based information-switching network for skin lesion recognition. Bioengineering 12:282. https:\/\/doi.org\/10.3390\/BIOENGINEERING12030282","journal-title":"Bioengineering"},{"key":"7034_CR24","doi-asserted-by":"publisher","first-page":"375","DOI":"10.1080\/01676830.2020.1812094","volume":"40","author":"VH Vasanthapuram","year":"2021","unstructured":"Vasanthapuram VH, Saha P, Mohamed A, Naik MN (2021) Ultrasound biomicroscopic features of the normal lower eyelid. Orbit 40:375\u2013380. https:\/\/doi.org\/10.1080\/01676830.2020.1812094","journal-title":"Orbit"},{"key":"7034_CR25","doi-asserted-by":"publisher","first-page":"209","DOI":"10.1007\/S40477-018-0301-X","volume":"21","author":"B Kuzmanovi\u0107 Elabjer","year":"2018","unstructured":"Kuzmanovi\u0107 Elabjer B, Bu\u0161i\u0107 M, Mileti\u0107 D, Bjelo\u0161 M, Vukojevi\u0107 N, Bosnar D (2018) Possible role of standardized echography complementing ultrasound biomicroscopy in tumors of the anterior eye segment: a study in a series of 13 patients. J Ultrasound 21:209. https:\/\/doi.org\/10.1007\/S40477-018-0301-X","journal-title":"J Ultrasound"},{"key":"7034_CR26","doi-asserted-by":"publisher","DOI":"10.1016\/J.CMPB.2021.106086","volume":"205","author":"TK Yoo","year":"2021","unstructured":"Yoo TK, Choi JY, Kim HK, Ryu IH, Kim JK (2021) Adopting low-shot deep learning for the detection of conjunctival melanoma using ocular surface images. Comput Methods Programs Biomed 205:106086. https:\/\/doi.org\/10.1016\/J.CMPB.2021.106086","journal-title":"Comput Methods Programs Biomed"},{"key":"7034_CR27","doi-asserted-by":"publisher","DOI":"10.1155\/2022\/3942110","volume":"2022","author":"S Zhu","year":"2022","unstructured":"Zhu S, Fang X, Qian Y et al (2022) Pterygium screening and lesion area segmentation based on deep learning. J Healthc Eng 3942110. https:\/\/doi.org\/10.1155\/2022\/3942110","journal-title":"J Healthc Eng"},{"key":"7034_CR28","doi-asserted-by":"publisher","DOI":"10.3390\/DIAGNOSTICS12030639","author":"SR Abdani","year":"2022","unstructured":"Abdani SR, Zulkifley MA, Shahrimin MI, Zulkifley NH (2022) Computer-assisted pterygium screening system: a review. Diagnostics. https:\/\/doi.org\/10.3390\/DIAGNOSTICS12030639","journal-title":"Diagnostics"},{"key":"7034_CR29","doi-asserted-by":"publisher","first-page":"182","DOI":"10.1016\/J.AJO.2025.02.019","volume":"273","author":"JA Greenfield","year":"2025","unstructured":"Greenfield JA, Scherer R, Alba D et al (2025) Detection of Ocular Surface Squamous Neoplasia Using Artificial Intelligence With Anterior Segment Optical Coherence Tomography. Am J Ophthalmol 273:182\u2013191. https:\/\/doi.org\/10.1016\/J.AJO.2025.02.019","journal-title":"Am J Ophthalmol"},{"key":"7034_CR30","doi-asserted-by":"publisher","first-page":"632","DOI":"10.1136\/BJO.86.6.632","volume":"86","author":"DH Char","year":"2002","unstructured":"Char DH, Kundert G, Bove R, Crawford JB (2002) 20\u00a0MHz high frequency ultrasound assessment of scleral and intraocular conjunctival squamous cell carcinoma. Br J Ophthalmol 86:632\u2013635. https:\/\/doi.org\/10.1136\/BJO.86.6.632","journal-title":"Br J Ophthalmol"},{"key":"7034_CR31","doi-asserted-by":"publisher","first-page":"168","DOI":"10.1001\/ARCHOPHT.121.2.168","volume":"121","author":"PT Finger","year":"2003","unstructured":"Finger PT, Tran HV, Turbin RE et al (2003) High-frequency ultrasonographic evaluation of conjunctival intraepithelial neoplasia and squamous cell carcinoma. Arch Ophthalmol 121:168\u2013172. https:\/\/doi.org\/10.1001\/ARCHOPHT.121.2.168","journal-title":"Arch Ophthalmol"},{"key":"7034_CR32","doi-asserted-by":"publisher","first-page":"17","DOI":"10.1016\/j.ajo.2019.05.017","volume":"206","author":"CL Karp","year":"2019","unstructured":"Karp CL, Mercado C, Venkateswaran N, Ruggeri M et al (2019) Use of High-Resolution optical cohernece tomography in the surgical management of ocular surface squamous neoplasia:a pilot study. Am J Ophthalmol 206:17\u201331. https:\/\/doi.org\/10.1016\/j.ajo.2019.05.017","journal-title":"Am J Ophthalmol"}],"container-title":["Graefe's Archive for Clinical and Experimental Ophthalmology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00417-025-07034-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00417-025-07034-x","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00417-025-07034-x.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,19]],"date-time":"2026-03-19T08:37:30Z","timestamp":1773909450000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00417-025-07034-x"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,10]]},"references-count":33,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2026,4]]}},"alternative-id":["7034"],"URL":"https:\/\/doi.org\/10.1007\/s00417-025-07034-x","relation":{},"ISSN":["0721-832X","1435-702X"],"issn-type":[{"value":"0721-832X","type":"print"},{"value":"1435-702X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,10]]},"assertion":[{"value":"5 August 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"28 October 2025","order":2,"name":"revised","label":"Revised","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"11 November 2025","order":3,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"10 January 2026","order":4,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"No conflicting relationship exists for any author.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of interest"}},{"value":"This study adhered the tenets of the Declaration of Helsinki and received approval the University of Michigan\u2019s ethics committee(#HUM00046408).","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical approval"}},{"value":"Due to its retrospective design, this article does not contain any new studies with human participants performed by any of the authors.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Informed consent"}}]}}