{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,29]],"date-time":"2026-08-29T10:20:51Z","timestamp":1787998851469,"version":"build-2784847793"},"reference-count":30,"publisher":"BMJ","issue":"1","license":[{"start":{"date-parts":[[2025,10,8]],"date-time":"2025-10-08T00:00:00Z","timestamp":1759881600000},"content-version":"unspecified","delay-in-days":7,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100002347","name":"Federal Ministry of Education and Research","doi-asserted-by":"crossref","award":["16SV8616"],"award-info":[{"award-number":["16SV8616"]}],"id":[{"id":"10.13039\/501100002347","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["bmj.com"],"crossmark-restriction":true},"short-container-title":["BMJ Health Care Inform"],"accepted":{"date-parts":[[2025,9,29]]},"published-print":{"date-parts":[[2025,10]]},"abstract":"<jats:sec>\n                  <jats:title>Study objectives<\/jats:title>\n                  <jats:p>Chronic wounds represent a significant economic and personal burden. For their successful treatment, the causes must be known and treated. Wounds caused by pyoderma gangrenosum (PG), a rare inflammatory skin disease, are often misdiagnosed. This study, therefore, aims to develop a machine learning model capable of differentiating PG from other wound types, focusing on chronic leg wounds to address this diagnostic challenge.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Methods<\/jats:title>\n                  <jats:p>We used 3674 wound photographs from three specialised wound centres with the four most common types of foot and leg ulcers and the rare inflammatory differential diagnosis PG. The convolutional neural network classifier ConvNeXt \u2018B\u2019 was pretrained on LAION2B, ImageNet12k and ImageNet 1k before being trained and fine-tuned on an 85:15 train, validation split.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Results<\/jats:title>\n                  <jats:p>The model achieved an overall high accuracy in multiclass classification of the chronic wounds (unbalanced accuracy 90%, balanced accuracy 87%). The sensitivity for identifying PG was 94%, while the sensitivity forother chronic wound types was 97% for diabetic foot ulcers (DFU), 92% for venous leg ulcers (VLU), 78% for mixed leg ulcers and 74% for arterial leg ulcers.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Discussion<\/jats:title>\n                  <jats:p>The machine learning model effectively differentiates PG from the most common leg and foot ulcers and was very accurate for classifying DFU and VLU. A higher rate of misclassifications occurred for the other vascular ulcers, that is, mixed and arterial leg ulcers. This aligns with the challenges in clinical practice.<\/jats:p>\n               <\/jats:sec>\n               <jats:sec>\n                  <jats:title>Conclusion<\/jats:title>\n                  <jats:p>Despite the limited number of wound images, this novel multiclass wound classification model accurately identified PG and differentiated leg and foot ulcer subtypes, providing a foundation for a diagnostic support system.<\/jats:p>\n               <\/jats:sec>","DOI":"10.1136\/bmjhci-2024-101418","type":"journal-article","created":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T10:45:25Z","timestamp":1760093125000},"page":"e101418","update-policy":"https:\/\/doi.org\/10.1136\/crossmarkpolicy","source":"Crossref","is-referenced-by-count":2,"title":["Machine learning model to classify chronic leg wounds and identify pyoderma gangrenosum"],"prefix":"10.1136","volume":"32","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5176-8157","authenticated-orcid":false,"given":"Dorothee A","family":"Busch","sequence":"first","affiliation":[{"name":"Health Informatics Research Group, Osnabr\u00fcck University of Applied Sciences, Faculty of Business Management and Social Sciences, Osnabr\u00fcck, Germany"},{"name":"Department of Dermatology, Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg, University Hospital, Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mats L","family":"Richter","sequence":"additional","affiliation":[{"name":"Mila Quebec Artificial Intelligence Institute, Montreal, Qu\u00e9bec, Canada"},{"name":"Formerly: Institute of Cognitive Science, Osnabr\u00fcck University, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jens","family":"H\u00fcsers","sequence":"additional","affiliation":[{"name":"Health Informatics Research Group, Osnabr\u00fcck University of Applied Sciences, Faculty of Business Management and Social Sciences, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Mareike","family":"Przysucha","sequence":"additional","affiliation":[{"name":"Health Informatics Research Group, Osnabr\u00fcck University of Applied Sciences, Faculty of Business Management and Social Sciences, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Florian","family":"K\u00fccking","sequence":"additional","affiliation":[{"name":"Health Informatics Research Group, Osnabr\u00fcck University of Applied Sciences, Faculty of Business Management and Social Sciences, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jonathan M","family":"Mang","sequence":"additional","affiliation":[{"name":"Medical Center for Information and Communication Technology, Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg, University Hospital, Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Maurice","family":"Moelleken","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Venerology and Allergology, University Hospital Essen, Essen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joachim","family":"Dissemond","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Venerology and Allergology, University Hospital Essen, Essen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jan","family":"Heggemann","sequence":"additional","affiliation":[{"name":"Christian Hospital Melle, Niels-Stensen-Kliniken GmbH, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Guido","family":"Hafer","sequence":"additional","affiliation":[{"name":"Christian Hospital Melle, Niels-Stensen-Kliniken GmbH, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Carola","family":"Berking","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg, University Hospital, Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Cornelia","family":"Erfurt-Berge","sequence":"additional","affiliation":[{"name":"Department of Dermatology, Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg, University Hospital, Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ursula H","family":"H\u00fcbner","sequence":"additional","affiliation":[{"name":"Health Informatics Research Group, Osnabr\u00fcck University of Applied Sciences, Faculty of Business Management and Social Sciences, Osnabr\u00fcck, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"239","published-online":{"date-parts":[[2025,10,10]]},"reference":[{"key":"2025101003451457000_32.1.e101418.1","doi-asserted-by":"publisher","DOI":"10.1089\/wound.2023.0150"},{"key":"2025101003451457000_32.1.e101418.2","doi-asserted-by":"publisher","DOI":"10.1016\/j.mcna.2021.04.006"},{"key":"2025101003451457000_32.1.e101418.3","doi-asserted-by":"publisher","DOI":"10.1080\/07853890.2016.1231932"},{"key":"2025101003451457000_32.1.e101418.4","unstructured":"Quist SR . S1-Leitlinie Pyoderma gangrenosum. Dis Orphan 2020."},{"key":"2025101003451457000_32.1.e101418.5","doi-asserted-by":"crossref","DOI":"10.3390\/jcm11237103","article-title":"Computer-Assisted Differential Diagnosis of Pyoderma Gangrenosum and Venous Ulcers with Deep Neural Networks","volume":"11","author":"Birkner","year":"2022","journal-title":"J Clin Med"},{"key":"2025101003451457000_32.1.e101418.6","doi-asserted-by":"publisher","DOI":"10.1111\/bjd.16401"},{"key":"2025101003451457000_32.1.e101418.7","doi-asserted-by":"crossref","first-page":"50","DOI":"10.1038\/s41572-022-00377-3","article-title":"Chronic wounds","volume":"8","author":"Falanga","year":"2022","journal-title":"Nat Rev Dis Primers"},{"key":"2025101003451457000_32.1.e101418.8","doi-asserted-by":"crossref","first-page":"7043","DOI":"10.1038\/s41598-024-56626-w","article-title":"Integrated image and location analysis for wound classification: a deep learning approach","volume":"14","author":"Patel","year":"2024","journal-title":"Sci Rep"},{"key":"2025101003451457000_32.1.e101418.9","doi-asserted-by":"crossref","DOI":"10.1111\/iwj.14565","article-title":"Development of an explainable artificial intelligence model for Asian vascular wound images","volume":"21","author":"Lo","year":"2024","journal-title":"Int Wound J"},{"key":"2025101003451457000_32.1.e101418.10","doi-asserted-by":"crossref","DOI":"10.1038\/s41598-022-21813-0","article-title":"Multi-modal wound classification using wound image and location by deep neural network","volume":"12","author":"Anisuzzaman","year":"2022","journal-title":"Sci Rep"},{"key":"2025101003451457000_32.1.e101418.11","doi-asserted-by":"crossref","first-page":"1385","DOI":"10.3390\/bioengineering10121385","article-title":"Multi-Class Wound Classification via High and Low-Frequency Guidance Network","volume":"10","author":"Guo","year":"2023","journal-title":"Bioengineering (Basel)"},{"key":"2025101003451457000_32.1.e101418.12","doi-asserted-by":"crossref","first-page":"104536","DOI":"10.1016\/j.compbiomed.2021.104536","article-title":"Multiclass wound image classification using an ensemble deep CNN-based classifier","volume":"134","author":"Rostami","year":"2021","journal-title":"Comput Biol Med"},{"key":"2025101003451457000_32.1.e101418.13","first-page":"1406","article-title":"The Enlightening Role of Explainable Artificial Intelligence in Chronic Wound Classification","volume":"10","author":"Sarp","year":"2021","journal-title":"Electronics (Basel)"},{"key":"2025101003451457000_32.1.e101418.14","first-page":"63","article-title":"An Image Based Object Recognition System for Wound Detection and Classification of Diabetic Foot and Venous Leg Ulcers","volume":"294","author":"H\u00fcsers","year":"2022","journal-title":"Stud Health Technol Inform"},{"key":"2025101003451457000_32.1.e101418.15","unstructured":"Nilsson CA , Velic M . Classification of ulcer images using convolutional neural networks. 2018."},{"key":"2025101003451457000_32.1.e101418.16","doi-asserted-by":"crossref","first-page":"811028","DOI":"10.3389\/fbioe.2021.811028","article-title":"Classification of Diabetic Foot Ulcers Using Class Knowledge Banks","volume":"9","author":"Xu","year":"2022","journal-title":"Front Bioeng Biotechnol"},{"key":"2025101003451457000_32.1.e101418.17","doi-asserted-by":"crossref","first-page":"728","DOI":"10.1109\/TETCI.2018.2866254","article-title":"DFUNet: Convolutional Neural Networks for Diabetic Foot Ulcer Classification","volume":"4","author":"Goyal","year":"2020","journal-title":"IEEE Trans Emerg Top Comput Intell"},{"key":"2025101003451457000_32.1.e101418.18","doi-asserted-by":"crossref","first-page":"103616","DOI":"10.1016\/j.compbiomed.2020.103616","article-title":"Recognition of ischaemia and infection in diabetic foot ulcers: Dataset and techniques","volume":"117","author":"Goyal","year":"2020","journal-title":"Comput Biol Med"},{"key":"2025101003451457000_32.1.e101418.19","unstructured":"Hrsg.: bundesinstitut f\u00fcr arzneimittel und medizinprodukte (bfarm) im auftrag des bundesministeriums f\u00fcr gesundheit (bmg) unter beteiligung der arbeitsgruppe icd des kuratoriums f\u00fcr fragen der klassifikation im gesundheitswesen (kkg). icd-10-gm version 2021, systematisches verzeichnis, internationale statistische klassifikation der krankheiten und verwandter gesundheitsprobleme, 10. 2020. Available: www.bfarm.de \u2013 Kodiersysteme \u2013 Services - Downloads \u2013 ICD-10-GM \u2013 Version 2021"},{"key":"2025101003451457000_32.1.e101418.20","doi-asserted-by":"crossref","first-page":"e82","DOI":"10.3414\/ME17-02-0025","article-title":"MIRACUM: Medical Informatics in Research and Care in University Medicine","volume":"57","author":"Prokosch","year":"2018","journal-title":"Methods Inf Med"},{"key":"2025101003451457000_32.1.e101418.21","doi-asserted-by":"crossref","unstructured":"Liu Z , Mao H , Christoph CW , et al . ConvNeXt_CVPR22] A ConvNet for the 2020s. Cvpr 2022;11976\u201386.","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"2025101003451457000_32.1.e101418.22","doi-asserted-by":"crossref","unstructured":"Richter ML , Shenk JC , Byttner W , et al . Feature space saturation during training. British Machine Vision Conference 2021; 2021 doi:10.5244\/C.35.28 Available: https:\/\/bmva.org\/bmvc\/2021\/","DOI":"10.5244\/C.35.28"},{"key":"2025101003451457000_32.1.e101418.23","unstructured":"Ilharco G , Wortsman M , Carlini N , et al . OpenCLIP. Zenodo; 2021."},{"key":"2025101003451457000_32.1.e101418.24","first-page":"1","article-title":"LAION-5B: An open large-scale dataset for training next generation image-text models","volume":"35","author":"Schuhmann","year":"2022","journal-title":"Adv Neural Inf Process Syst"},{"key":"2025101003451457000_32.1.e101418.25","doi-asserted-by":"crossref","unstructured":"Richter ML , Pal C . Receptive field refinement for convolutional neural networks reliably improves predictive performance. arXiv Prepr arXiv 2022. doi:10.2139\/ssrn.4525468","DOI":"10.2139\/ssrn.4525468"},{"key":"2025101003451457000_32.1.e101418.26","doi-asserted-by":"crossref","unstructured":"Richter M . Receptive Field Analysis for Optimizing Convolutional Neural Network Architectures Without Training Receptive Field Analysis for Optimizing Convolutional Neural Network Architectures Without Training. 2022. doi:10.1007\/978-981-19-6153-3_10","DOI":"10.1007\/978-981-19-6153-3_10"},{"key":"2025101003451457000_32.1.e101418.27","unstructured":"Wightman R , Touvron H , J\u00e9gou H . ResNet strikes back: an improved training procedure in timm. 2021."},{"key":"2025101003451457000_32.1.e101418.28","doi-asserted-by":"crossref","unstructured":"Chattopadhyay A , Sarkar A , Howlader P , et al . Grad-CAM++: Generalized Gradient-based Visual Explanations for Deep Convolutional Networks. 2017.","DOI":"10.1109\/WACV.2018.00097"},{"key":"2025101003451457000_32.1.e101418.29","doi-asserted-by":"crossref","first-page":"2494","DOI":"10.3390\/diagnostics13152494","article-title":"An Overview of Clinical Examinations in the Evaluation and Assessment of Arterial and Venous Insufficiency Wounds","volume":"13","author":"Wang","year":"2023","journal-title":"Diagnostics (Basel)"},{"key":"2025101003451457000_32.1.e101418.30","doi-asserted-by":"crossref","first-page":"687","DOI":"10.1089\/wound.2021.0091","article-title":"Image-Based Artificial Intelligence in Wound Asses","volume":"11","author":"Anisuzzaman","year":"2022","journal-title":"Adv Wound Care"}],"container-title":["BMJ Health &amp; Care Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.1136\/bmjhci-2024-101418","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T10:45:31Z","timestamp":1760093131000},"score":1,"resource":{"primary":{"URL":"https:\/\/informatics.bmj.com\/lookup\/doi\/10.1136\/bmjhci-2024-101418"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10]]},"references-count":30,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,10,10]]},"published-print":{"date-parts":[[2025,10]]}},"alternative-id":["10.1136\/bmjhci-2024-101418"],"URL":"https:\/\/doi.org\/10.1136\/bmjhci-2024-101418","relation":{},"ISSN":["2632-1009"],"issn-type":[{"value":"2632-1009","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10]]}}}