{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,30]],"date-time":"2026-09-30T01:54:24Z","timestamp":1790733264318,"version":"4.1.0"},"reference-count":43,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T00:00:00Z","timestamp":1775001600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T00:00:00Z","timestamp":1775088000000},"content-version":"vor","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100005760","name":"University of Gothenburg","doi-asserted-by":"crossref","id":[{"id":"10.13039\/501100005760","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Early identification of melanomas with high metastatic potential is crucial for treatment planning and survival prediction. We investigated whether weakly supervised WSI-based and multimodal learning approaches can identify metastatic risk in primary cutaneous melanoma beyond current clinicopathological predictors. A total of 426 routinely stained whole-slide images (WSIs), from primary melanomas (249 metastatic and 177 non-metastatic) together with corresponding clinicopathological features, were collected and divided to training and validation, and a hold-out test set. WSIs were divided into patches and encoded using Prov-GigaPath, while clinicopathological features were converted to text embeddings using BioMedBERT. We evaluated a weakly supervised transformer-based approach for slide-level metastatic risk prediction from WSIs, both alone and in combination with clinicopathological features, and compared performance to a model based on clinicopathological features alone. Both WSI-based models achieved strong performance (AUC\u2009=\u20090.887 and 0.883) and higher accuracy (0.847 for both) than the clinicopathological-based model (AUC\u2009=\u20090.849, accuracy\u2009=\u20090.753). This benefit was most pronounced in T2 tumors, where early risk stratification is clinically most relevant. These findings demonstrate that weakly supervised WSI-based models capture prognostically relevant information and highlight their potential for early metastatic risk stratification in primary melanoma.<\/jats:p>","DOI":"10.1038\/s41598-026-45588-w","type":"journal-article","created":{"date-parts":[[2026,4,1]],"date-time":"2026-04-01T11:51:21Z","timestamp":1775044281000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":2,"title":["Early detection of metastatic risk in primary cutaneous melanoma using weakly supervised learning"],"prefix":"10.1038","volume":"16","author":[{"given":"Filip","family":"Dahl\u00e9n","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ivan","family":"Shujski","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Filmon","family":"Yacob","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ida","family":"H\u00e4ggstr\u00f6m","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jovana","family":"Jovanovic","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Olga","family":"Dudina","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ilkka","family":"P\u00f6l\u00f6nen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Noora","family":"Neittaanm\u00e4ki","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,4,1]]},"reference":[{"key":"45588_CR1","first-page":"472","volume":"67","author":"J Gershenwald","year":"2017","unstructured":"Gershenwald, J. et al. Melanoma staging: Evidence-based changes in the american joint committee on cancer eighth edition cancer staging manual. CA Cancer J. Clin. 67, 472\u2013492 (2017).","journal-title":"CA Cancer J. Clin."},{"key":"45588_CR2","doi-asserted-by":"publisher","first-page":"374","DOI":"10.18632\/oncotarget.28591","volume":"15","author":"J Thompson","year":"2024","unstructured":"Thompson, J. & Williams, G. When does a melanoma metastasize? implications for management. Oncotarget 15, 374\u2013378 (2024).","journal-title":"Oncotarget"},{"key":"45588_CR3","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1111\/j.1755-148X.2009.00647.x","volume":"23","author":"L Ossowski","year":"2010","unstructured":"Ossowski, L. & Aguirre-Ghiso, J. Dormancy of metastatic melanoma. Pigment Cell. Melanoma Res. 23, 41\u201356 (2010).","journal-title":"Pigment Cell. Melanoma Res."},{"key":"45588_CR4","doi-asserted-by":"publisher","first-page":"13","DOI":"10.1111\/pcmr.12544","volume":"30","author":"N Adler","year":"2017","unstructured":"Adler, N., Haydon, A., McLean, C., Kelly, J. & Mar, V. Metastatic pathways in patients with cutaneous melanoma. Pigment Cell. Melanoma Res. 30, 13\u201327 (2017).","journal-title":"Pigment Cell. Melanoma Res."},{"key":"45588_CR5","doi-asserted-by":"publisher","first-page":"817510","DOI":"10.3389\/fonc.2022.817510","volume":"12","author":"H Huang","year":"2022","unstructured":"Huang, H., Fu, Z., Ji, J., Huang, J. & Long, X. Predictive values of pathological and clinical risk factors for positivity of sentinel lymph node biopsy in thin melanoma: A systematic review and meta-analysis. Front. Oncol. 12, 817510 (2022).","journal-title":"Front. Oncol."},{"key":"45588_CR6","doi-asserted-by":"publisher","first-page":"3622","DOI":"10.1200\/JCO.2001.19.16.3622","volume":"19","author":"C Balch","year":"2001","unstructured":"Balch, C. et al. Prognostic factors analysis of 17,600 melanoma patients: validation of the american joint committee on cancer melanoma staging system. J. Clin. Oncol. 19, 3622\u20133634 (2001).","journal-title":"J. Clin. Oncol."},{"key":"45588_CR7","doi-asserted-by":"publisher","first-page":"5574","DOI":"10.1002\/cam4.2474","volume":"8","author":"H Dillek\u00e5s","year":"2019","unstructured":"Dillek\u00e5s, H., Rogers, M. S. & Straume, O. Are 90% of deaths from cancer caused by metastases? Cancer Med. 8, 5574\u20135576 (2019).","journal-title":"Cancer Med."},{"key":"45588_CR8","doi-asserted-by":"publisher","first-page":"258","DOI":"10.1016\/j.jaad.2016.10.018","volume":"76","author":"SM Landow","year":"2017","unstructured":"Landow, S. M., Gjelsvik, A. & Weinstock, M. A. Mortality burden and prognosis of thin melanomas overall and by subcategory of thickness, SEER registry data, 1992\u20132013. J. Am. Acad. Dermatol. 76, 258\u2013263 (2017).","journal-title":"J. Am. Acad. Dermatol."},{"key":"45588_CR9","doi-asserted-by":"publisher","first-page":"1190","DOI":"10.1038\/jid.2014.452","volume":"135","author":"DC Whiteman","year":"2015","unstructured":"Whiteman, D. C., Baade, P. D. & Olsen, C. M. More people die from thin melanomas (1\u00a0mm) than from thick melanomas (>\u20094\u00a0mm) in Queensland, Australia. J. Invest. Dermatol. 135, 1190\u20131193 (2015).","journal-title":"J. Invest. Dermatol."},{"key":"45588_CR10","first-page":"06508","volume":"2407","author":"G Campanella","year":"2024","unstructured":"Campanella, G. et al. A clinical benchmark of public self-supervised pathology foundation models. arXiv 2407, 06508 (2024).","journal-title":"arXiv"},{"key":"45588_CR11","unstructured":"Zhang, H. et al. DINO: DETR with improved denoising anchor boxes for end-to-end object detection. (2022). arXiv 2203.03605."},{"key":"45588_CR12","first-page":"07193","volume":"2304","author":"M Oquab","year":"2024","unstructured":"Oquab, M. et al. DINOv2: learning robust visual features without supervision. arXiv 2304, 07193 (2024).","journal-title":"arXiv"},{"key":"45588_CR13","doi-asserted-by":"publisher","first-page":"2691","DOI":"10.1038\/s41591-025-03747-y","volume":"31","author":"S Yan","year":"2025","unstructured":"Yan, S. et al. A multimodal vision foundation model for clinical dermatology. Nat. Med. 31, 2691\u20132702 (2025).","journal-title":"Nat. Med."},{"key":"45588_CR14","doi-asserted-by":"crossref","unstructured":"Wang, Y. et al. Integrating Clinical Knowledge Graphs and Gradient-Based Neural Systems for Enhanced Melanoma Diagnosis via the Seven-Point Checklist. IEEE Trans. Neural Networks Learn. Syst. (2025).","DOI":"10.1109\/TNNLS.2025.3600443"},{"key":"45588_CR15","doi-asserted-by":"crossref","unstructured":"Li, X. et al. MSDUNet: A Model based on Feature Multi-Scale and Dual-input Dynamic Enhancement for Skin Lesion Segmentation. IEEE Trans. Med. Imaging (2025).","DOI":"10.1109\/TMI.2025.3549011"},{"key":"45588_CR16","doi-asserted-by":"publisher","first-page":"62","DOI":"10.1111\/joim.13030","volume":"288","author":"B Acs","year":"2020","unstructured":"Acs, B., Rantalainen, M. & Hartman, J. Artificial intelligence as the next step towards precision pathology. J. Intern. Med. 288, 62\u201368 (2020).","journal-title":"J. Intern. Med."},{"key":"45588_CR17","unstructured":"Dosovitskiy, A. et al. An image is worth 16\u00d716 words: transformers for image recognition at scale. arXiv 2010.11929 (2021)."},{"key":"45588_CR18","first-page":"2136","volume":"34","author":"Z Shao","year":"2021","unstructured":"Shao, Z. et al. TransMIL: transformer-based correlated multiple instance learning for whole-slide image classification. Adv. Neural Inf. Process. Syst. 34, 2136\u20132147 (2021).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"45588_CR19","doi-asserted-by":"crossref","unstructured":"Zeid, M. A. E., El-Bahnasy, K. & Abo-Youssef S. E. Multiclass colorectal cancer histology image classification using vision transformers. In Proc. ICICIS, 224\u2013230 (2021).","DOI":"10.1109\/ICICIS52592.2021.9694125"},{"key":"45588_CR20","first-page":"33863","volume":"13","author":"SJVK Yacob","year":"2023","unstructured":"Yacob, S. J. V. K. et al. Weakly supervised detection and classification of basal cell carcinoma using graph transformers on whole-slide images. Sci. Rep. 13, 33863 (2023).","journal-title":"Sci. Rep."},{"key":"45588_CR21","doi-asserted-by":"publisher","first-page":"123893","DOI":"10.1016\/j.eswa.2024.123893","volume":"250","author":"M Waqas","year":"2024","unstructured":"Waqas, M., Ahmed, S. U., Tahir, M. A., Wu, J. & Qureshi, R. Exploring multiple instance learning: a brief survey. Expert Syst. Appl. 250, 123893 (2024).","journal-title":"Expert Syst. Appl."},{"key":"45588_CR22","doi-asserted-by":"publisher","first-page":"1559","DOI":"10.3389\/fonc.2020.01559","volume":"10","author":"F De Logu","year":"2020","unstructured":"De Logu, F. et al. Recognition of cutaneous melanoma on digitized histopathological slides via artificial intelligence algorithms. Front. Oncol. 10, 1559 (2020).","journal-title":"Front. Oncol."},{"key":"45588_CR23","doi-asserted-by":"publisher","first-page":"79","DOI":"10.1016\/j.ejca.2019.04.021","volume":"115","author":"A Hekler","year":"2019","unstructured":"Hekler, A. et al. Pathologist-level classification of histopathological melanoma images with deep neural networks. Eur. J. Cancer. 115, 79\u201383 (2019).","journal-title":"Eur. J. Cancer"},{"key":"45588_CR24","doi-asserted-by":"publisher","first-page":"101150","DOI":"10.1016\/j.currproblcancer.2024.101150","volume":"53","author":"D Sun","year":"2024","unstructured":"Sun, D. et al. Artificial intelligence-based pathological application to predict regional lymph node metastasis in papillary thyroid cancer. Curr. Probl. Cancer. 53, 101150 (2024).","journal-title":"Curr. Probl. Cancer"},{"key":"45588_CR25","doi-asserted-by":"crossref","unstructured":"Tan, L. H. Y. et al. Colorectal cancer lymph node metastasis prediction with weakly supervised transformer-based multiple instance learning. Med. Biol. Eng. Comput. (2023).","DOI":"10.1007\/s11517-023-02799-x"},{"key":"45588_CR26","doi-asserted-by":"publisher","first-page":"555","DOI":"10.1038\/s41551-020-00682-w","volume":"5","author":"MY Lu","year":"2021","unstructured":"Lu, M. Y. et al. Data-efficient and weakly supervised computational pathology on whole-slide images. Nat. Biomed. Eng. 5, 555\u2013570 (2021).","journal-title":"Nat. Biomed. Eng."},{"key":"45588_CR27","doi-asserted-by":"publisher","first-page":"161","DOI":"10.1016\/j.ejca.2023.04.023","volume":"188","author":"P Jansen","year":"2023","unstructured":"Jansen, P. et al. Deep learning detection of melanoma metastases in lymph nodes. Eur. J. Cancer. 188, 161\u2013170 (2023).","journal-title":"Eur. J. Cancer"},{"key":"45588_CR28","doi-asserted-by":"publisher","first-page":"1418013","DOI":"10.3389\/fmed.2024.1418013","volume":"11","author":"J Siarov","year":"2024","unstructured":"Siarov, J. et al. Deep learning model shows pathologist-level detection of sentinel node metastasis of melanoma and intranodal nevi on whole-slide images. Front. Med. 11, 1418013 (2024).","journal-title":"Front. Med."},{"key":"45588_CR29","doi-asserted-by":"publisher","first-page":"227","DOI":"10.1016\/j.ejca.2021.05.026","volume":"154","author":"TJ Brinker","year":"2021","unstructured":"Brinker, T. J. et al. Deep learning approach to predict sentinel lymph node status directly from routine histology of primary melanoma tumours. Eur. J. Cancer. 154, 227\u2013234 (2021).","journal-title":"Eur. J. Cancer"},{"key":"45588_CR30","doi-asserted-by":"publisher","first-page":"13696","DOI":"10.1038\/s41598-022-13696-y","volume":"12","author":"RP Knuutila","year":"2022","unstructured":"Knuutila, R. P. et al. Identification of metastatic primary cutaneous squamous cell carcinoma using artificial intelligence analysis of whole-slide images. Sci. Rep. 12, 13696 (2022).","journal-title":"Sci. Rep."},{"key":"45588_CR31","doi-asserted-by":"publisher","first-page":"1126","DOI":"10.1158\/1078-0432.CCR-19-1495","volume":"26","author":"PM Kulkarni","year":"2020","unstructured":"Kulkarni, P. M. et al. Deep learning based on standard H&E images of primary melanoma tumors identifies patients at risk for visceral recurrence and death. Clin. Cancer Res. 26, 1126\u20131134 (2020).","journal-title":"Clin. Cancer Res."},{"key":"45588_CR32","doi-asserted-by":"publisher","first-page":"847","DOI":"10.1093\/bjd\/ljae281","volume":"191","author":"K Lallas","year":"2024","unstructured":"Lallas, K. et al. Prediction of melanoma metastasis using dermatoscopy deep features: an international multicentre cohort study. Br. J. Dermatol. 191, 847\u2013848 (2024).","journal-title":"Br. J. Dermatol."},{"key":"45588_CR33","doi-asserted-by":"publisher","first-page":"155","DOI":"10.1016\/j.ejca.2022.12.021","volume":"182","author":"T Amaral","year":"2023","unstructured":"Amaral, T. et al. Identification of stage I\/II melanoma patients at high risk for recurrence using a model combining clinicopathologic factors with gene expression profiling. Eur. J. Cancer. 182, 155\u2013162 (2023).","journal-title":"Eur. J. Cancer"},{"key":"45588_CR34","doi-asserted-by":"publisher","first-page":"62","DOI":"10.1046\/j.1365-2133.2002.04867.x","volume":"147","author":"F Meier","year":"2002","unstructured":"Meier, F. et al. Metastatic pathways and time courses in the orderly progression of cutaneous melanoma. Br. J. Dermatol. 147, 62\u201370 (2002).","journal-title":"Br. J. Dermatol."},{"key":"45588_CR35","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1002\/jso.20079","volume":"86","author":"U Leiter","year":"2004","unstructured":"Leiter, U., Meier, F., Schittek, B. & Garbe, C. The natural course of cutaneous melanoma. J. Surg. Oncol. 86, 172\u2013178 (2004).","journal-title":"J. Surg. Oncol."},{"key":"45588_CR36","doi-asserted-by":"publisher","first-page":"e32955","DOI":"10.1371\/journal.pone.0032955","volume":"7","author":"L Mervic","year":"2012","unstructured":"Mervic, L. Time course and pattern of metastasis of cutaneous melanoma differ between men and women. PLoS One. 7, e32955 (2012).","journal-title":"PLoS One"},{"key":"45588_CR37","doi-asserted-by":"publisher","first-page":"605","DOI":"10.1067\/mjd.2000.107234","volume":"43","author":"MH Schmid-Wendtner","year":"2000","unstructured":"Schmid-Wendtner, M. H. et al. Late metastases of cutaneous melanoma: an analysis of 31 patients. J. Am. Acad. Dermatol. 43, 605\u2013609 (2000).","journal-title":"J. Am. Acad. Dermatol."},{"key":"45588_CR38","doi-asserted-by":"crossref","unstructured":"Zhai, X., Kolesnikov, A., Houlsby, N. & Beyer, L. Scaling vision transformers. (2022). arXiv 2106.04560.","DOI":"10.1109\/CVPR52688.2022.01179"},{"key":"45588_CR39","doi-asserted-by":"publisher","first-page":"27","DOI":"10.4103\/2153-3539.119005","volume":"4","author":"A Goode","year":"2013","unstructured":"Goode, A., Gilbert, B. & Heller, J. OpenSlide: a vendor-neutral software foundation for digital pathology. J. Pathol. Inf. 4, 27 (2013).","journal-title":"J. Pathol. Inf."},{"key":"45588_CR40","doi-asserted-by":"crossref","unstructured":"Xu, H. et al. A whole-slide foundation model for digital pathology from real-world data. Nature (2024).","DOI":"10.1038\/s41586-024-07441-w"},{"key":"45588_CR41","doi-asserted-by":"crossref","unstructured":"Ding, J. et al. LongNet: scaling transformers to 1,000,000,000 tokens. (2023). arXiv 2307.02486.","DOI":"10.14218\/JCTH.2022.00351"},{"key":"45588_CR42","unstructured":"Gu, Y. et al. Domain-specific language model pretraining for biomedical natural language processing. (2020). arXiv 2007.15779."},{"key":"45588_CR43","unstructured":"Vaswani, A. et al. Attention is all you need. arXiv 1706, 03762 (2017)."}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-45588-w","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-45588-w.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-45588-w.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,16]],"date-time":"2026-06-16T21:19:41Z","timestamp":1781644781000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-45588-w"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,1]]},"references-count":43,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["45588"],"URL":"https:\/\/doi.org\/10.1038\/s41598-026-45588-w","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-6297243\/v1","asserted-by":"object"}]},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,1]]},"assertion":[{"value":"24 March 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"19 March 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"1 April 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"The authors declare no competing interests.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}},{"value":"The Swedish Ethical Review Authority (Dnr 2023-06786-02) approved this study. Since all the material was anonymized, consent to participate was waived.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}}],"article-number":"11234"}}