{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,3]],"date-time":"2025-10-03T00:30:41Z","timestamp":1759451441073,"version":"build-2065373602"},"reference-count":25,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,10,2]],"date-time":"2025-10-02T00:00:00Z","timestamp":1759363200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Med."],"abstract":"<jats:sec><jats:title>Background<\/jats:title><jats:p>Differentiating Actinic keratosis (AK) from Seborrheic keratosis (SK) can be challenging for dermatologists due to their visual similarities. This multi-center prospective study aims to investigate the efficacy of deep learning (DL) model in assisting dermatologists in accurately classifying AK from SK lesions.<\/jats:p><\/jats:sec><jats:sec><jats:title>Methods<\/jats:title><jats:p>A contrastive language-image pre-training (CLIP) model with ViT-B\/16 architecture was trained on an dataset of 2,307 patients and validated in three separate datasets of 386 (from Center A), 196 patients (from Center B and C) and 215 patients (from DermNet). Two dermatologists classified the lesions separately. Then they were showed the model\u2019s predictions and were requested to reclassify the results if needed. Area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the diagnostic performances of the DL model and the dermatologists before and after reclassification. The change in the dermatologists\u2019 classification decisions was also analyzed by net reclassification index (NRI) and total integrated discrimination index (IDI).<\/jats:p><\/jats:sec><jats:sec><jats:title>Results<\/jats:title><jats:p>The model\u2019s diagnostic performance in the training cohort and validation cohort 1, 2 and 3 showed an AUC of 0.85, 0.89, 0.84, and 0.89. For dermatologist 1, the diagnostic performance improved from 0.77 to 0.80 in the test cohort with NRI and IDI of 0.10 (<jats:italic>p<\/jats:italic>\u202f=\u202f0.006) and 0.14 (<jats:italic>p<\/jats:italic>\u202f&amp;lt;\u202f0.001). For dermatologist 2, the diagnostic performance increased from 0.69 to 0.79 with NRI and IDI of 0.19 (<jats:italic>p<\/jats:italic>\u202f&amp;lt;\u202f0.001) and 0.27 (<jats:italic>p<\/jats:italic>\u202f&amp;lt;\u202f0.001).<\/jats:p><\/jats:sec><jats:sec><jats:title>Conclusion<\/jats:title><jats:p>The DL model significantly improves dermatologists\u2019 accuracy in differentiating AK from SK, especially for less experienced ones. The DL model has the potential to reduce diagnostic subjectivity, aid early detection of precancerous lesions, and transform dermatological diagnostic and therapeutic practices.<\/jats:p><\/jats:sec>","DOI":"10.3389\/fmed.2025.1654813","type":"journal-article","created":{"date-parts":[[2025,10,2]],"date-time":"2025-10-02T05:30:52Z","timestamp":1759383052000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Enhancing dermatological diagnosis for differentiating actinic from seborrheic keratosis using deep learning model"],"prefix":"10.3389","volume":"12","author":[{"given":"Ying-Ying","family":"Ren","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Li-Hong","family":"Mei","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiang-Dong","family":"Liu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhe","family":"Quan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Gao","family":"Yang","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,10,2]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"17","DOI":"10.1016\/j.jmoldx.2022.09.008","article-title":"Molecular skin cancer diagnosis: promise and limitations","volume":"25","author":"Nataren","year":"2023","journal-title":"J Mol Diagn"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"1330491","DOI":"10.3389\/fmed.2024.1330491","article-title":"Advancements in elucidating the pathogenesis of actinic keratosis: present state and future prospects","volume":"11","author":"Wang","year":"2024","journal-title":"Front Med"},{"key":"ref3","doi-asserted-by":"publisher","first-page":"664","DOI":"10.1111\/j.1610-0387.2008.06788.x","article-title":"Seborrheic keratosis","volume":"6","author":"Hafner","year":"2008","journal-title":"J Dtsch Dermatol Ges"},{"key":"ref4","doi-asserted-by":"publisher","first-page":"e0112447","DOI":"10.1371\/journal.pone.0112447","article-title":"Automated detection of actinic keratoses in clinical photographs","volume":"10","author":"Hames","year":"2015","journal-title":"PLoS One"},{"key":"ref5","doi-asserted-by":"publisher","first-page":"e14817","DOI":"10.1111\/dth.14817","article-title":"The diagnostic value of p63, p16, and p53 immunohistochemistry in distinguishing seborrheic keratosis, actinic keratosis, and Bowen's disease","volume":"34","author":"Nasiri","year":"2021","journal-title":"Dermatol Ther"},{"key":"ref6","doi-asserted-by":"publisher","first-page":"e85270","DOI":"10.7554\/eLife.85270","article-title":"Single-cell sequencing highlights heterogeneity and malignant progression in actinic keratosis and cutaneous squamous cell carcinoma","volume":"12","author":"Zou","year":"2023","journal-title":"eLife"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"102470","DOI":"10.1016\/j.media.2022.102470","article-title":"Explainable artificial intelligence (XAI) in deep learning-based medical image analysis","volume":"79","author":"van der Velden","year":"2022","journal-title":"Med Image Anal"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"102863","DOI":"10.1016\/j.media.2023.102863","article-title":"A survey on deep learning for skin lesion segmentation","volume":"88","author":"Mirikharaji","year":"2023","journal-title":"Med Image Anal"},{"key":"ref9","doi-asserted-by":"publisher","first-page":"115","DOI":"10.1038\/nature21056","article-title":"Dermatologist-level classification of skin cancer with deep neural networks","volume":"542","author":"Esteva","year":"2017","journal-title":"Nature"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"1114362","DOI":"10.3389\/fmed.2023.1114362","article-title":"Classification of clinical skin lesions with double-branch networks","volume":"10","author":"Wang","year":"2023","journal-title":"Front Med"},{"key":"ref11","doi-asserted-by":"publisher","first-page":"965423","DOI":"10.3389\/fmed.2022.965423","article-title":"A deep learning-based approach toward differentiating scalp psoriasis and seborrheic dermatitis from dermoscopic images","volume":"9","author":"Yu","year":"2022","journal-title":"Front Med"},{"key":"ref12","doi-asserted-by":"publisher","first-page":"1507322","DOI":"10.3389\/fonc.2025.1507322","article-title":"Deep learning in assisting dermatologists in classifying basal cell carcinoma from seborrheic keratosis","volume":"15","author":"Mei","year":"2025","journal-title":"Front Oncol"},{"key":"ref13","doi-asserted-by":"publisher","first-page":"e58692","DOI":"10.7759\/cureus.58692","article-title":"Artificial intelligence-based distinction of actinic keratosis and seborrheic keratosis","volume":"16","author":"Reddy","year":"2024","journal-title":"Cureus"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"976235","DOI":"10.3389\/frai.2022.976235","article-title":"CLIP knows image aesthetics","volume":"5","author":"Hentschel","year":"2022","journal-title":"Front Artif Intell"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"e0298102","DOI":"10.1371\/journal.pone.0298102","article-title":"Brain tumor classification in VIT-B\/16 based on relative position encoding and residual MLP","volume":"19","author":"Hong","year":"2024","journal-title":"PLoS One"},{"key":"ref16","doi-asserted-by":"publisher","first-page":"491","DOI":"10.3390\/e26060491","article-title":"Correlations of cross-entropy loss in machine learning","volume":"26","author":"Connor","year":"2024","journal-title":"Entropy"},{"key":"ref17","doi-asserted-by":"publisher","first-page":"157","DOI":"10.1002\/sim.2929","article-title":"Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond","volume":"27","author":"Pencina","year":"2008","journal-title":"Stat Med"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"714","DOI":"10.3390\/bioengineering12070714","article-title":"Novel snapshot-based hyperspectral conversion for dermatological lesion detection via YOLO object detection models","volume":"12","author":"Huang","year":"2025","journal-title":"Bioengineering"},{"key":"ref19","doi-asserted-by":"publisher","first-page":"1331895","DOI":"10.3389\/fmed.2024.1331895","article-title":"Artificial intelligence and skin cancer","volume":"11","author":"Wei","year":"2024","journal-title":"Front Med"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"755","DOI":"10.3390\/bioengineering12070755","article-title":"Hyperspectral imaging for enhanced skin cancer classification using machine learning","volume":"12","author":"Lin","year":"2025","journal-title":"Bioengineering"},{"key":"ref21","doi-asserted-by":"publisher","first-page":"108","DOI":"10.3390\/cancers16010108","article-title":"Skinlesnet: classification of skin lesions and detection of melanoma cancer using a novel multi-layer deep convolutional neural network","volume":"16","author":"Azeem","year":"2023","journal-title":"Cancers (Basel)"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"1134","DOI":"10.3390\/jcm12031134","article-title":"Classification of skin cancer using novel hyperspectral imaging engineering via YOLOv5","volume":"12","author":"Huang","year":"2023","journal-title":"J Clin Med"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"3587","DOI":"10.1109\/JBHI.2025.3528196","article-title":"MedFILIP: medical fine-grained language-image pre-training","volume":"29","author":"Liang","year":"2025","journal-title":"IEEE J Biomed Health Inform"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"15887","DOI":"10.1038\/s41598-023-42847-y","article-title":"Improving CNNs classification with pathologist-based expertise: the renal cell carcinoma case study","volume":"13","author":"Ponzio","year":"2023","journal-title":"Sci Rep"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"411","DOI":"10.1007\/s00330-020-07099-8","article-title":"Radiologists with MRI-based radiomics aids to predict the pelvic lymph node metastasis in endometrial cancer: a multicenter study","volume":"31","author":"Yan","year":"2021","journal-title":"Eur Radiol"}],"container-title":["Frontiers in Medicine"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2025.1654813\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,2]],"date-time":"2025-10-02T05:30:53Z","timestamp":1759383053000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fmed.2025.1654813\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,2]]},"references-count":25,"alternative-id":["10.3389\/fmed.2025.1654813"],"URL":"https:\/\/doi.org\/10.3389\/fmed.2025.1654813","relation":{},"ISSN":["2296-858X"],"issn-type":[{"value":"2296-858X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,10,2]]},"article-number":"1654813"}}