{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,1]],"date-time":"2026-10-01T17:22:15Z","timestamp":1790875335979,"version":"4.1.0"},"reference-count":23,"publisher":"Oxford University Press (OUP)","issue":"2","license":[{"start":{"date-parts":[[2024,4,6]],"date-time":"2024-04-06T00:00:00Z","timestamp":1712361600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,7,16]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:sec>\n                  <jats:title>Background<\/jats:title>\n                  <jats:p>Artificial intelligence (AI) is reshaping healthcare, using machine and deep learning (DL) to enhance disease management. Dermatology has seen improved diagnostics, particularly in skin cancer detection, through the integration of AI. However, the potential of AI in automating immunofluorescence imaging for autoimmune bullous skin diseases (AIBDs) remains untapped. While direct immunofluorescence (DIF) supports diagnosis, its manual interpretation can hinder efficiency. The use of DL to classify DIF patterns automatically, including the intercellular (ICP) and linear pattern (LP), holds promise for improving the diagnosis of AIBDs.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Objectives<\/jats:title>\n                  <jats:p>To develop AI algorithms for automated classification of AIBD DIF patterns, such as ICP and LP, in order to enhance diagnostic accuracy, streamline disease management and improve patient outcomes through DL-driven immunofluorescence interpretation.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We collected immunofluorescence images from skin biopsies of patients suspected of having an AIBD between January 2022 and January 2024. Skin tissue was obtained via a 5-mm punch biopsy, prepared for DIF. Experienced dermatologists classified the images as ICP, LP or negative. To evaluate our DL approach, we divided the images into training (n = 436) and test sets (n = 93). We employed transfer learning with pretrained deep neural networks and conducted fivefold cross-validation to assess model performance. Our dataset\u2019s class imbalance was addressed using weighted loss and data augmentation strategies. The models were trained for 50 epochs using Pytorch, achieving an image size of 224 \u00d7 224 pixels for both convolutional neural networks (CNNs) and the Swin Transformer.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>Our study compared six CNNs and the Swin Transformer for AIBD image classification, with the Swin Transformer achieving the highest average validation accuracy (98.5%). On a separate test set, the best model attained an accuracy of 94.6%, demonstrating 95.3% sensitivity and 97.5% specificity across AIBD classes. Visualization with Grad-CAM (class activation mapping) highlighted the model\u2019s reliance on characteristic patterns for accurate classification.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusions<\/jats:title>\n                  <jats:p>The study highlighted the accuracy of CNNs in identifying DIF features. This approach aids automated analysis and reporting, offering reproducibility, speed, data handling and cost-efficiency. Integrating DL into skin immunofluorescence promises precise diagnostics and streamlined reporting in this branch of dermatology.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1093\/bjd\/ljae142","type":"journal-article","created":{"date-parts":[[2024,4,6]],"date-time":"2024-04-06T15:08:57Z","timestamp":1712416137000},"page":"261-266","source":"Crossref","is-referenced-by-count":21,"title":["A deep learning approach to direct immunofluorescence pattern recognition in autoimmune bullous diseases"],"prefix":"10.1093","volume":"191","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-1704-6799","authenticated-orcid":false,"given":"Niccol\u00f2","family":"Capurro","sequence":"first","affiliation":[{"name":"Section of Dermatology, Department of Health Sciences, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5827-5571","authenticated-orcid":false,"given":"Vito Paolo","family":"Pastore","sequence":"additional","affiliation":[{"name":"MaLGa - DIBRIS, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Larbi","family":"Touijer","sequence":"additional","affiliation":[{"name":"MaLGa - DIBRIS, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Francesca","family":"Odone","sequence":"additional","affiliation":[{"name":"MaLGa - DIBRIS, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Emanuele","family":"Cozzani","sequence":"additional","affiliation":[{"name":"Section of Dermatology, Department of Health Sciences, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Giulia","family":"Gasparini","sequence":"additional","affiliation":[{"name":"Section of Dermatology, Department of Health Sciences, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aurora","family":"Parodi","sequence":"additional","affiliation":[{"name":"Section of Dermatology, Department of Health Sciences, University of Genoa , Genoa , Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"286","published-online":{"date-parts":[[2024,4,6]]},"reference":[{"key":"2024071611292652400_ljae142-B1","doi-asserted-by":"crossref","first-page":"1504","DOI":"10.1016\/j.jid.2020.02.026","article-title":"Artificial intelligence in dermatology: a primer","volume":"140","author":"Young","year":"2020","journal-title":"J Invest Dermatol"},{"key":"2024071611292652400_ljae142-B2","doi-asserted-by":"crossref","article-title":"Incorporating diagnostic prior with segmentation: a deep learning pipeline for the automatic classification of autoimmune bullous skin diseases","author":"Pastore","DOI":"10.1109\/ISBI53787.2023.10230494"},{"key":"2024071611292652400_ljae142-B3","doi-asserted-by":"crossref","first-page":"1111172","DOI":"10.3389\/fimmu.2023.1111172","article-title":"Computer-aided classification of indirect immunofluorescence patterns on esophagus and split skin for the detection of autoimmune dermatoses","volume":"14","author":"Hocke","year":"2023","journal-title":"Front Immunol"},{"key":"2024071611292652400_ljae142-B4","doi-asserted-by":"crossref","first-page":"185","DOI":"10.1111\/bjd.14408","article-title":"Laboratory diagnosis of pemphigus: direct immunofluorescence remains the gold standard","volume":"175","author":"Giurdanella","year":"2016","journal-title":"Br J Dermatol"},{"key":"2024071611292652400_ljae142-B5","doi-asserted-by":"crossref","first-page":"75","DOI":"10.1097\/DAD.0000000000001516","article-title":"Practical direct immunofluorescence","volume":"42","author":"Kim","year":"2020","journal-title":"Am J Dermatopathol"},{"key":"2024071611292652400_ljae142-B6","doi-asserted-by":"crossref","first-page":"1031","DOI":"10.1016\/j.jaci.2020.02.013","article-title":"Autoimmune bullous skin diseases, pemphigus and pemphigoid","volume":"145","author":"Egami","year":"2020","journal-title":"J Allergy Clin Immunol"},{"key":"2024071611292652400_ljae142-B7","doi-asserted-by":"crossref","first-page":"619","DOI":"10.1067\/mjd.2000.107488","article-title":"Sensitivity and specificity of clinical, histologic, and immunologic features in the diagnosis of paraneoplastic pemphigus","volume":"43","author":"Joly","year":"2000","journal-title":"J Am Acad Dermatol"},{"key":"2024071611292652400_ljae142-B8","doi-asserted-by":"crossref","first-page":"69","DOI":"10.1186\/s12880-022-00793-7","article-title":"Transfer learning for medical image classification: a literature review","volume":"22","author":"Kim","year":"2022","journal-title":"BMC Med Imaging"},{"key":"2024071611292652400_ljae142-B9","doi-asserted-by":"crossref","first-page":"230","DOI":"10.1016\/j.neucom.2021.08.159","article-title":"Transfer learning for medical images analyses: a survey","volume":"2022","author":"Yu","year":"2022","journal-title":"Neurocomputing"},{"key":"2024071611292652400_ljae142-B10","doi-asserted-by":"crossref","first-page":"6999","DOI":"10.1109\/TNNLS.2021.3084827","article-title":"A survey of convolutional neural networks: analysis, applications, and prospects","volume":"33","author":"Li","year":"2022","journal-title":"IEEE Trans Neural Netw Learn Syst"},{"key":"2024071611292652400_ljae142-B11","doi-asserted-by":"crossref","first-page":"100216","DOI":"10.1016\/j.health.2023.100216","article-title":"A comprehensive review of deep neural networks for medical image processing: recent developments and future opportunities","volume":"4","author":"Mall","year":"2023","journal-title":"Healthc Anal (N Y"},{"key":"2024071611292652400_ljae142-B12","doi-asserted-by":"crossref","first-page":"226","DOI":"10.1007\/s10916-018-1088-1","article-title":"Medical image analysis using convolutional neural networks: a review","volume":"42","author":"Anwar","year":"2018","journal-title":"J Med Syst"},{"key":"2024071611292652400_ljae142-B13","doi-asserted-by":"crossref","first-page":"113","DOI":"10.1186\/s40537-019-0276-2","article-title":"Deep convolutional neural network based medical image classification for disease diagnosis","volume":"6","author":"Yadav","year":"2019","journal-title":"J Big Data"},{"key":"2024071611292652400_ljae142-B14","doi-asserted-by":"crossref","first-page":"100297","DOI":"10.1016\/j.imu.2020.100297","article-title":"Deep learning approaches to biomedical image segmentation","volume":"18","author":"Rizwan","year":"2020","journal-title":"Inform Med Unlocked"},{"key":"2024071611292652400_ljae142-B15","doi-asserted-by":"crossref","first-page":"59","DOI":"10.1016\/j.imed.2022.07.002","article-title":"Transformers in medical image analysis","volume":"3","author":"He","year":"2023","journal-title":"Intell Med"},{"key":"2024071611292652400_ljae142-B16","doi-asserted-by":"crossref","article-title":"Deep residual learning\u00a0for\u00a0image\u00a0recognition","author":"He","DOI":"10.1109\/CVPR.2016.90"},{"key":"2024071611292652400_ljae142-B17","author":"Huang"},{"key":"2024071611292652400_ljae142-B18","doi-asserted-by":"crossref","article-title":"Swin Transformer: hierarchical vision transformer using shifted windows","author":"Liu","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"2024071611292652400_ljae142-B19","article-title":"Deep learning","author":"Goodfellow"},{"key":"2024071611292652400_ljae142-B20","article-title":"Cross-entropy loss functions: theoretical analysis and applications","author":"Mao"},{"key":"2024071611292652400_ljae142-B21","first-page":"8024","article-title":"PyTorch: an imperative style, high-performance deep learning library","volume":"32","author":"Paszke","year":"2019","journal-title":"Adv Neural Inf Process Syst"},{"key":"2024071611292652400_ljae142-B22","doi-asserted-by":"crossref","article-title":"Towards better explanations of class activation mapping","author":"Jung","DOI":"10.1109\/ICCV48922.2021.00137"},{"key":"2024071611292652400_ljae142-B23","doi-asserted-by":"crossref","article-title":"Grad-CAM: visual explanations from deep networks via gradient-based localization","author":"Selvaraju","DOI":"10.1109\/ICCV.2017.74"}],"container-title":["British Journal of Dermatology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/academic.oup.com\/bjd\/advance-article-pdf\/doi\/10.1093\/bjd\/ljae142\/57671149\/ljae142.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bjd\/article-pdf\/191\/2\/261\/58554360\/ljae142.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/academic.oup.com\/bjd\/article-pdf\/191\/2\/261\/58554360\/ljae142.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,7,16]],"date-time":"2024-07-16T11:34:05Z","timestamp":1721129645000},"score":1,"resource":{"primary":{"URL":"https:\/\/academic.oup.com\/bjd\/article\/191\/2\/261\/7641799"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,4,6]]},"references-count":23,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2024,4,6]]},"published-print":{"date-parts":[[2024,7,16]]}},"URL":"https:\/\/doi.org\/10.1093\/bjd\/ljae142","relation":{},"ISSN":["0007-0963","1365-2133"],"issn-type":[{"value":"0007-0963","type":"print"},{"value":"1365-2133","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2024,8]]},"published":{"date-parts":[[2024,4,6]]}}}