{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T18:05:09Z","timestamp":1780682709219,"version":"3.54.1"},"reference-count":43,"publisher":"Public Library of Science (PLoS)","issue":"6","license":[{"start":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T00:00:00Z","timestamp":1780617600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100019854","name":"Greater Copenhagen Health Science Partners","doi-asserted-by":"publisher","award":["CAG SCIN"],"award-info":[{"award-number":["CAG SCIN"]}],"id":[{"id":"10.13039\/501100019854","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Actinic keratosis (AK) is a pre-cancerous skin lesion typically caused by excessive exposure to ultraviolet light. Optical coherence tomography (OCT) can provide sub-surface information relevant for lesion classification, but manual analysis of images is time-consuming and subject to inter-observer variability; automated segmentation based on deep learning models can provide faster and more consistent results. However, structural changes, such as thickening and hyperkeratosis, complicate epidermal segmentation tasks. For this reason, we aimed to develop and optimize a U-Net model for the automated epidermal segmentation of AK lesions in OCT images, combining both accuracy and computational efficiency. Multiple configurations were evaluated by varying hyperparameters, including image sizes (256\u2009\u00d7\u2009256, 512\u2009\u00d7\u2009512, 1024\u2009\u00d7\u20091024, and 464\u2009\u00d7\u20091356 pixels), batch sizes (2, 4, 8, and 16), and training durations (50, 100, and 150 epochs). Model performance was evaluated quantitatively using several metrics including the Dice coefficient and Jaccard index, comparing automated epidermal segmentations against expert annotations. The optimal configuration, with an image resolution of 256\u2009\u00d7\u2009256 pixels and a batch size of 2 over 50 epochs, had a Dice score of 0.86 and a Jaccard index of 0.76. Higher resolutions and longer training increased computation without significantly enhancing the accuracy values and sometimes caused overfitting. These findings show that a simple U-Net architecture could achieve efficient and accurate epidermal segmentation in AK lesions, provided it is fine-tuned to enhance performance.<\/jats:p>","DOI":"10.1371\/journal.pone.0346059","type":"journal-article","created":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T17:42:06Z","timestamp":1780681326000},"page":"e0346059","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["A U-Net model for epidermal segmentation in optical coherence tomography images of actinic keratosis"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0001-4215-6818","authenticated-orcid":true,"given":"Theofanis","family":"Angelis","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9656-733X","authenticated-orcid":true,"given":"Peter A.","family":"Philipsen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vinzent K.","family":"Ortner","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gabriella","family":"Fredman","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Merete","family":"Haedersdal","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6334-0625","authenticated-orcid":true,"given":"Gavrielle R.","family":"Untracht","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,6,5]]},"reference":[{"key":"pone.0346059.ref001","doi-asserted-by":"crossref","first-page":"528","DOI":"10.1016\/j.jid.2021.07.172","article-title":"Genomic Progression of Precancerous Actinic Keratosis to Squamous Cell Carcinoma","volume":"142","author":"Y-S Kim","year":"2022","journal-title":"J Invest Dermatol"},{"issue":"5","key":"pone.0346059.ref002","first-page":"667","article-title":"Treatment options for actinic keratoses","volume":"76","author":"WJ McIntyre","year":"2007","journal-title":"Am Fam Physician"},{"issue":"8","key":"pone.0346059.ref003","doi-asserted-by":"crossref","first-page":"608","DOI":"10.1007\/s00403-024-03344-y","article-title":"Interobserver agreement on line-field confocal optical coherence tomography image markers in keratinocyte carcinomas and precursor lesions","volume":"316","author":"K Jacobsen","year":"2024","journal-title":"Arch Dermatol Res"},{"key":"pone.0346059.ref004","doi-asserted-by":"crossref","DOI":"10.5826\/dpc.1403S1a147S","article-title":"Clinical and Dermoscopic Diagnosis of Actinic Keratosis","volume":"14","author":"C Conforti","year":"2024","journal-title":"Dermatol Pract Concept"},{"issue":"12","key":"pone.0346059.ref005","article-title":"Dermoscopy, with and without visual inspection, for diagnosing melanoma in adults","volume":"12","author":"J Dinnes","year":"2018","journal-title":"Cochrane Database Syst Rev"},{"issue":"4","key":"pone.0346059.ref006","first-page":"389","article-title":"Histology: The gold standard for diagnosis?","volume":"64","author":"L-J Tseng","year":"2023","journal-title":"Can Vet J"},{"issue":"3","key":"pone.0346059.ref007","doi-asserted-by":"crossref","first-page":"172","DOI":"10.1016\/j.sder.2009.06.004","article-title":"Reflectance confocal microscopy--state-of-art and research overview","volume":"28","author":"R Hofmann-Wellenhof","year":"2009","journal-title":"Semin Cutan Med Surg"},{"issue":"4","key":"pone.0346059.ref008","doi-asserted-by":"crossref","first-page":"238","DOI":"10.14366\/usg.24015","article-title":"Role of high-frequency ultrasound in differentiating benign and malignant skin lesions","volume":"43","author":"Q Wang","year":"2024","journal-title":"Ultrasonography"},{"issue":"6","key":"pone.0346059.ref009","first-page":"1024","article-title":"European consensus-based interdisciplinary guideline for diagnosis, treatment and prevention of actinic keratoses and field cancerization","volume":"38","author":"L Kandolf","year":"2024","journal-title":"J Eur Acad Dermatol Venereol"},{"issue":"1","key":"pone.0346059.ref010","doi-asserted-by":"crossref","first-page":"89","DOI":"10.1111\/j.1600-0846.2007.00265.x","article-title":"Application of optical coherence tomography in non-invasive characterization of skin vascular lesions","volume":"14","author":"C Salvini","year":"2008","journal-title":"Skin Res Technol"},{"issue":"1","key":"pone.0346059.ref011","doi-asserted-by":"crossref","first-page":"2230","DOI":"10.1038\/s41598-023-47051-6","article-title":"Rapid measurement of epidermal thickness in OCT images of skin","volume":"14","author":"C-H Lin","year":"2024","journal-title":"Sci Rep"},{"key":"pone.0346059.ref012","doi-asserted-by":"crossref","first-page":"85","DOI":"10.1159\/000366541","article-title":"Reflectance confocal microscopy: hallmarks of keratinocyte cancer and its precursors","volume":"46","author":"TW Prow","year":"2015","journal-title":"Curr Probl Dermatol"},{"key":"pone.0346059.ref013","doi-asserted-by":"crossref","first-page":"738","DOI":"10.1016\/0190-9622(91)70113-G","article-title":"Vehicle-controlled study of masoprocol cream for actinic keratoses","volume":"24","author":"EA Olsen","year":"1991","journal-title":"J Am Acad Dermatol"},{"issue":"8","key":"pone.0346059.ref014","doi-asserted-by":"crossref","first-page":"1303","DOI":"10.1111\/jdv.13626","article-title":"Actinic keratosis: correlation between clinical and histological classification systems","volume":"30","author":"L Schmitz","year":"2016","journal-title":"J Eur Acad Dermatol Venereol"},{"issue":"4","key":"pone.0346059.ref015","doi-asserted-by":"crossref","first-page":"699","DOI":"10.1111\/bjd.17637","article-title":"From actinic keratosis to squamous cell carcinoma - answers to some open questions","volume":"180","author":"J Welzel","year":"2019","journal-title":"Br J Dermatol"},{"issue":"4","key":"pone.0346059.ref016","doi-asserted-by":"crossref","first-page":"217","DOI":"10.3390\/diagnostics10040217","article-title":"AK-DL: shallow neural network for diagnosing actinic keratosis","volume":"10","author":"L Wang","year":"2020","journal-title":"Diagnostics"},{"issue":"3","key":"pone.0346059.ref017","article-title":"Photographic Quantification of Actinic Keratoses Using Human- and Artificial Intelligence-Based Assessment: A Pilot Study","volume":"146","author":"Y Li","year":"2026","journal-title":"J Invest Dermatol"},{"key":"pone.0346059.ref018","first-page":"234","article-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation","author":"O Ronneberger","year":"2015","journal-title":"Lecture Notes in Computer Science. Springer International Publishing"},{"issue":"1","key":"pone.0346059.ref019","doi-asserted-by":"crossref","DOI":"10.1049\/ipr2.70019","article-title":"Review of U-Net and its variants in medical image segmentation","volume":"19","author":"W Jiangtao","year":"2025","journal-title":"IET Image Process"},{"issue":"10","key":"pone.0346059.ref020","doi-asserted-by":"crossref","first-page":"1034","DOI":"10.3390\/bioengineering11101034","article-title":"Advances in medical image segmentation approaches","volume":"11","author":"Y Xu","year":"2024","journal-title":"Bioengineering"},{"issue":"10","key":"pone.0346059.ref021","doi-asserted-by":"crossref","DOI":"10.1136\/bmjopen-2022-061012","article-title":"The Copenhagen Actinic Keratosis Study (COAKS). A decentralised clinical trial to evaluate tolerability, safety and efficacy of daily field-directed topical treatment with cytosolic phospholipase A2 inhibitor, AVX001, in participants with actinic keratosis: protocol for a randomised controlled phase I\/IIa trial","volume":"12","author":"VK Ortner","year":"2022","journal-title":"BMJ Open"},{"issue":"2","key":"pone.0346059.ref022","doi-asserted-by":"crossref","first-page":"30","DOI":"10.3390\/computers8020030","article-title":"A Study of Image Upsampling and Downsampling Filters","volume":"8","author":"D Dumitrescu","year":"2019","journal-title":"Computers"},{"key":"pone.0346059.ref023","unstructured":"GeeksforGeeks. CNN: introduction to padding. https:\/\/www.geeksforgeeks.org\/machine-learning\/cnn-introduction-to-padding\/ 2023. Accessed 2023 December 13."},{"key":"pone.0346059.ref024","doi-asserted-by":"crossref","first-page":"106","DOI":"10.58496\/BJN\/2025\/009","article-title":"Applying Keras-based deep learning for network security analysis","volume":"2025","author":"S Yousef","year":"2025","journal-title":"Babylonian J Netw"},{"key":"pone.0346059.ref025","unstructured":"Iqbal H. U-net visualization tool. https:\/\/github.com\/HarisIqbal88\/UNet-Visualization 2018."},{"issue":"1","key":"pone.0346059.ref026","doi-asserted-by":"crossref","first-page":"210","DOI":"10.1186\/s13104-022-06096-y","article-title":"Towards a guideline for evaluation metrics in medical image segmentation","volume":"15","author":"D M\u00fcller","year":"2022","journal-title":"BMC Res Notes"},{"issue":"9","key":"pone.0346059.ref027","first-page":"2395","article-title":"Comparative assessment of regression models","volume":"8","author":"AV Tatachar","year":"2021","journal-title":"Int Res J Eng Technol"},{"issue":"6","key":"pone.0346059.ref028","first-page":"599","article-title":"Classification model evaluation metrics","volume":"12","author":"\u017d Vujovi\u0107","year":"2021","journal-title":"Int J Adv Comput Sci Appl"},{"issue":"4","key":"pone.0346059.ref029","doi-asserted-by":"crossref","first-page":"716","DOI":"10.1109\/42.363096","article-title":"Morphometric analysis of white matter lesions in MR images","volume":"13","author":"AP Zijdenbos","year":"1994","journal-title":"IEEE Trans Med Imaging"},{"key":"pone.0346059.ref030","first-page":"1","article-title":"Quality Quantification in Deep Convolutional Neural Networks for Skin Lesion Segmentation using Fuzzy Uncertainty Measurement.","author":"Q Lin","year":"2022","journal-title":"2022 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)"},{"issue":"19","key":"pone.0346059.ref031","doi-asserted-by":"crossref","first-page":"4861","DOI":"10.3390\/cancers15194861","article-title":"Semantic segmentation for actinic keratosis detection","volume":"15","author":"P Derekas","year":"2023","journal-title":"Cancers"},{"issue":"1","key":"pone.0346059.ref032","first-page":"9262","article-title":"ARCUNet for skin lesion segmentation","volume":"15","author":"T Soni","year":"2025","journal-title":"Scientific Reports"},{"key":"pone.0346059.ref033","article-title":"Retinal image segmentation with small datasets","author":"N Ndipenoch","year":"2023","journal-title":"arXiv"},{"key":"pone.0346059.ref034","first-page":"220","article-title":"Automatic segmentation of epidermis in OCT images","volume":"7","author":"R Del Amor","year":"2020","journal-title":"Front Med (Lausanne)"},{"issue":"19","key":"pone.0346059.ref035","doi-asserted-by":"crossref","first-page":"11079","DOI":"10.3390\/ijms231911079","article-title":"Segmentation of laser-induced injury OCT images","volume":"23","author":"T Gao","year":"2022","journal-title":"Int J Mol Sci"},{"issue":"1","key":"pone.0346059.ref036","doi-asserted-by":"crossref","first-page":"6","DOI":"10.17691\/stm2025.17.1.01","article-title":"Segmentation of 3D OCT images using U-Net","volume":"17","author":"VA Shishkova","year":"2025","journal-title":"Sovremennye Tekhnologii v Meditsine"},{"issue":"14","key":"pone.0346059.ref037","first-page":"1752","article-title":"AKtransU-Net for actinic keratosis detection","volume":"15","author":"P Derekas","year":"2025","journal-title":"Diagnostics (Basel)"},{"key":"pone.0346059.ref038","first-page":"104059","article-title":"CNN-based system for actinic keratosis detection","volume":"79","author":"P Spyridonos","year":"2023","journal-title":"Biomed Signal Process Control"},{"issue":"1","key":"pone.0346059.ref039","first-page":"015002","article-title":"Deep-learning approach for automated thickness measurement of epithelial tissue and scab using optical coherence tomography","volume":"27","author":"Y Ji","year":"2022","journal-title":"J Biomed Opt"},{"issue":"10","key":"pone.0346059.ref040","first-page":"5723","article-title":"LS-Net for epidermal segmentation in OCT imaging","volume":"15","author":"J Liao","year":"2024","journal-title":"Biomed Opt Express"},{"issue":"24","key":"pone.0346059.ref041","first-page":"7499","article-title":"Automatic assessment of actinic keratosis stage","volume":"13","author":"K Korecka","year":"2024","journal-title":"J Clin Med"},{"issue":"11","key":"pone.0346059.ref042","doi-asserted-by":"crossref","first-page":"1359","DOI":"10.1111\/ddg.15194","article-title":"AI-based PRO score assessment from LC-OCT","volume":"21","author":"JR Thamm","year":"2023","journal-title":"J Dtsch Dermatol Ges"},{"issue":"12","key":"pone.0346059.ref043","first-page":"2268","article-title":"Line-field confocal OCT for skin imaging","volume":"13","author":"F Latriglia","year":"2023","journal-title":"Life (Basel)"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0346059","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,5]],"date-time":"2026-06-05T17:42:19Z","timestamp":1780681339000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0346059"}},"subtitle":[],"editor":[{"given":"Xiaohui","family":"Zhang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,6,5]]},"references-count":43,"journal-issue":{"issue":"6","published-online":{"date-parts":[[2026,6,5]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0346059","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,5]]}}}