{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,30]],"date-time":"2026-10-30T09:39:28Z","timestamp":1793353168863,"version":"4.5.0"},"reference-count":35,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2026,4,9]],"date-time":"2026-04-09T00:00:00Z","timestamp":1775692800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,4,10]],"date-time":"2026-04-10T00:00:00Z","timestamp":1775779200000},"content-version":"vor","delay-in-days":1,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Universit\u00e4tsklinikum Bonn"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Sci Rep"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal malignancies, with prognosis strongly influenced by the presence of lymph node metastasis (LNM). However, preoperative LNM assessment from computed tomography (CT) is limited by low sensitivity, high inter-observer variability, and substantial heterogeneity across imaging protocols. This retrospective multi-center study (546 patients from three institutions) introduces a privacy-preserving deep learning framework that integrates large-scale CT foundation model pre-training with heterogeneity-aware federated optimization to improve LNM detection in PDAC. A CT Vision Foundation Model, pre-trained on 148,000 volumetric CT scans using contrastive self-supervised learning, is fine-tuned to generate transferable 3D representations for patient-level LNM classification. To enable decentralized model training while mitigating inter-institutional variability, we extend federated aggregation to jointly account for label-distribution discrepancies and representation-level divergence across clients. The centralized model achieved a balanced accuracy of 0.601 and a diagnostic odds ratio (DOR) of 3.45, outperforming classical machine learning baselines and prior PDAC LNM approaches. Under federated settings, the proposed heterogeneity-aware strategy consistently outperformed standard FedAvg, recovering a substantial proportion of the centralized model\u2019s performance while preserving strict data privacy. In particular, it improved balanced accuracy by 12.6% over FedAvg and demonstrated superior discriminative ability across all participating cohorts. These findings indicate that combining foundation model pre-training with discrepancy-aware federated learning enhances generalization, robustness, and clinical relevance for multi-center PDAC LNM detection. The proposed framework offers a scalable and privacy-preserving pathway for deploying deep learning models across distributed healthcare systems.<\/jats:p>","DOI":"10.1038\/s41598-026-47631-2","type":"journal-article","created":{"date-parts":[[2026,4,9]],"date-time":"2026-04-09T11:17:23Z","timestamp":1775733443000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Federated CT foundation models for multi-center detection of lymph node metastasis in pancreatic cancer"],"prefix":"10.1038","volume":"16","author":[{"given":"Parinishtha","family":"Bhalla","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"David Due\u00f1as","family":"Gaviria","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Patrick","family":"Kupczyk","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ali Seif Amir","family":"Hosseini","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lena","family":"Conradi","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Uli","family":"Fehrenbach","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Matthaeus","family":"Felsenstein","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dou","family":"Ma","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Alexander","family":"Semaan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shadi","family":"Albarqouni","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,4,9]]},"reference":[{"key":"47631_CR1","doi-asserted-by":"publisher","first-page":"173","DOI":"10.4251\/wjgo.v12.i2.173","volume":"12","author":"P Sarantis","year":"2020","unstructured":"Sarantis, P., Koustas, E., Papadimitropoulou, A., Papavassiliou, A. G. & Karamouzis, M. V. Pancreatic ductal adenocarcinoma: Treatment hurdles, tumor microenvironment and immunotherapy. World J. Gastrointest. Oncol. 12, 173\u2013181 (2020).","journal-title":"World J. Gastrointest. Oncol."},{"key":"47631_CR2","doi-asserted-by":"publisher","first-page":"71","DOI":"10.1007\/s12029-025-01183-2","volume":"56","author":"PD Leiphrakpam","year":"2025","unstructured":"Leiphrakpam, P. D. et al. Trends in the global incidence of pancreatic cancer and a brief review of its histologic and molecular subtypes. J. Gastrointest. Cancer 56, 71 (2025).","journal-title":"J. Gastrointest. Cancer"},{"key":"47631_CR3","doi-asserted-by":"publisher","first-page":"367","DOI":"10.1016\/j.jamcollsurg.2015.02.019","volume":"221","author":"G Malleo","year":"2015","unstructured":"Malleo, G. et al. Reappraisal of nodal staging and study of lymph node station involvement in pancreaticoduodenectomy with the standard international study group of pancreatic surgery definition of lymphadenectomy for cancer. J. Am. Coll. Surg. 221, 367\u2013379 (2015).","journal-title":"J. Am. Coll. Surg."},{"key":"47631_CR4","doi-asserted-by":"publisher","DOI":"10.1016\/j.ejro.2024.100553","volume":"12","author":"C Bilreiro","year":"2024","unstructured":"Bilreiro, C., Andrade, L., Santiago, I., Marques, R. M. & Matos, C. Imaging of pancreatic ductal adenocarcinoma-an update for all stages of patient management. Eur. J. Radiol Open 12, 100553 (2024).","journal-title":"Eur. J. Radiol Open"},{"key":"47631_CR5","doi-asserted-by":"publisher","first-page":"229","DOI":"10.1016\/j.suronc.2014.10.005","volume":"23","author":"DS Tseng","year":"2014","unstructured":"Tseng, D. S. et al. Diagnostic accuracy of ct in assessing extra-regional lymphadenopathy in pancreatic and peri-ampullary cancer: a systematic review and meta-analysis. Surg. Oncol. 23, 229\u2013235 (2014).","journal-title":"Surg. Oncol."},{"key":"47631_CR6","doi-asserted-by":"publisher","first-page":"zrad125","DOI":"10.1093\/bjsopen\/zrad125","volume":"8","author":"R Goess","year":"2024","unstructured":"Goess, R. et al. Lymph node examination and survival in resected pancreatic ductal adenocarcinoma: retrospective study. BJS Open 8, zrad125 (2024).","journal-title":"BJS Open"},{"key":"47631_CR7","doi-asserted-by":"publisher","first-page":"436","DOI":"10.3390\/cancers16020436","volume":"16","author":"H Patel","year":"2024","unstructured":"Patel, H., Zanos, T. & Hewitt, D. B. Deep learning applications in pancreatic cancer. Cancers 16, 436 (2024).","journal-title":"Cancers"},{"key":"47631_CR8","doi-asserted-by":"publisher","first-page":"172","DOI":"10.1148\/radiol.220152","volume":"306","author":"P-T Chen","year":"2023","unstructured":"Chen, P.-T. et al. Pancreatic cancer detection on ct scans with deep learning: a nationwide population-based study. Radiology 306, 172\u2013182 (2023).","journal-title":"Radiology"},{"key":"47631_CR9","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1186\/s13244-025-02025-2","volume":"16","author":"L Zhu","year":"2025","unstructured":"Zhu, L. et al. A radiomics-based model for predicting lymph nodes metastasis of pancreatic ductal adenocarcinoma: a multicenter study. Insights Imaging 16, 141 (2025).","journal-title":"Insights Imaging"},{"key":"47631_CR10","doi-asserted-by":"publisher","first-page":"1342317","DOI":"10.3389\/fonc.2024.1342317","volume":"14","author":"Y Huang","year":"2024","unstructured":"Huang, Y. et al. Contrast-enhanced ct radiomics combined with multiple machine learning algorithms for preoperative identification of lymph node metastasis in pancreatic ductal adenocarcinoma. Front. Oncol. 14, 1342317 (2024).","journal-title":"Front. Oncol."},{"key":"47631_CR11","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1186\/s40644-020-0288-3","volume":"20","author":"K Li","year":"2020","unstructured":"Li, K. et al. Contrast-enhanced ct radiomics for predicting lymph node metastasis in pancreatic ductal adenocarcinoma: a pilot study. Cancer Imaging 20, 12 (2020).","journal-title":"Cancer Imaging"},{"key":"47631_CR12","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/JBHI.2026.3668217","volume":"7","author":"Z Ye","year":"2026","unstructured":"Ye, Z. et al. Rethinking multi-center semi-supervised breast cancer ultrasound image segmentation: An intermediate-domain perspective. IEEE J. Biomed. Health Inform. 7, 1. https:\/\/doi.org\/10.1109\/JBHI.2026.3668217 (2026).","journal-title":"IEEE J. Biomed. Health Inform."},{"key":"47631_CR13","doi-asserted-by":"publisher","first-page":"2996","DOI":"10.1007\/s11263-024-02314-1","volume":"133","author":"Z Zheng","year":"2025","unstructured":"Zheng, Z. et al. Deep attention learning for pre-operative lymph node metastasis prediction in pancreatic cancer via multi-object relationship modeling. Int. J. Comput. Vis. 133, 2996\u20133019 (2025).","journal-title":"Int. J. Comput. Vis."},{"key":"47631_CR14","doi-asserted-by":"crossref","unstructured":"Gaviria, D.\u00a0D. et al. Deep learning for lymph node metastasis detection in pancreatic ductal adenocarcinoma. In International Conference on Medical Imaging and Computer-Aided Diagnosis, 66\u201376 (Springer, 2024).","DOI":"10.1007\/978-981-96-3863-5_7"},{"key":"47631_CR15","unstructured":"McMahan, B., Moore, E., Ramage, D., Hampson, S. & y Arcas, B.\u00a0A. Communication-efficient learning of deep networks from decentralized data. In Artificial intelligence and statistics, 1273\u20131282 (PMLR, 2017)."},{"key":"47631_CR16","first-page":"1333","volume":"13","author":"Y Aono","year":"2017","unstructured":"Aono, Y. et al. Privacy-preserving deep learning via additively homomorphic encryption. IEEE Trans. Inf. Forensics Secur. 13, 1333\u20131345 (2017).","journal-title":"IEEE Trans. Inf. Forensics Secur."},{"key":"47631_CR17","doi-asserted-by":"publisher","first-page":"170","DOI":"10.1016\/j.ins.2021.12.102","volume":"589","author":"W Huang","year":"2022","unstructured":"Huang, W. et al. Fairness and accuracy in horizontal federated learning. Inf. Sci. 589, 170\u2013185 (2022).","journal-title":"Inf. Sci."},{"key":"47631_CR18","doi-asserted-by":"crossref","unstructured":"Truex, S. et al. A hybrid approach to privacy-preserving federated learning. In Proceedings of the 12th ACM workshop on artificial intelligence and security, 1\u201311 (2019).","DOI":"10.1145\/3338501.3357370"},{"key":"47631_CR19","doi-asserted-by":"publisher","first-page":"685","DOI":"10.1038\/s42256-022-00515-2","volume":"4","author":"CI Bercea","year":"2022","unstructured":"Bercea, C. I., Wiestler, B., Rueckert, D. & Albarqouni, S. Federated disentangled representation learning for unsupervised brain anomaly detection. Nat. Mach. Intell. 4, 685\u2013695 (2022).","journal-title":"Nat. Mach. Intell."},{"key":"47631_CR20","doi-asserted-by":"publisher","first-page":"1735","DOI":"10.1038\/s41591-021-01506-3","volume":"27","author":"I Dayan","year":"2021","unstructured":"Dayan, I. et al. Federated learning for predicting clinical outcomes in patients with covid-19. Nat. Med. 27, 1735\u20131743 (2021).","journal-title":"Nat. Med."},{"key":"47631_CR21","doi-asserted-by":"publisher","first-page":"119","DOI":"10.1038\/s41746-020-00323-1","volume":"3","author":"N Rieke","year":"2020","unstructured":"Rieke, N. et al. The future of digital health with federated learning. NPJ Digit. Med. 3, 119 (2020).","journal-title":"NPJ Digit. Med."},{"key":"47631_CR22","unstructured":"Zhao, Y. et al. Federated learning with non-iid data. arXiv preprint arXiv:1806.00582 (2018)."},{"key":"47631_CR23","unstructured":"Li, X., Jiang, M., Zhang, X., Kamp, M. & Dou, Q. Fedbn: Federated learning on non-iid features via local batch normalization. arXiv preprint arXiv:2102.07623 (2021)."},{"key":"47631_CR24","doi-asserted-by":"crossref","unstructured":"Zhang, J. et al. Fedala: Adaptive local aggregation for personalized federated learning. In Proceedings of the AAAI conference on artificial intelligence 37, 11237\u201311244 (2023).","DOI":"10.1609\/aaai.v37i9.26330"},{"key":"47631_CR25","unstructured":"He, J. et al. PeFoMed: Parameter efficient fine-tuning of multimodal large language models for medical imaging. arXiv preprint arXiv:2401.02797 (2024)."},{"key":"47631_CR26","doi-asserted-by":"publisher","first-page":"5179","DOI":"10.1109\/TMC.2023.3302410","volume":"23","author":"T Guo","year":"2023","unstructured":"Guo, T., Guo, S., Wang, J., Tang, X. & Xu, W. Promptfl: Let federated participants cooperatively learn prompts instead of models-federated learning in age of foundation model. IEEE Trans. Mob. Comput. 23, 5179\u20135194 (2023).","journal-title":"IEEE Trans. Mob. Comput."},{"key":"47631_CR27","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2025.127636","volume":"281","author":"H Zheng","year":"2025","unstructured":"Zheng, H., Chu, J., Li, Z., Ji, J. & Li, T. Accelerating federated learning with genetic algorithm enhancements. Expert Syst. Appl. 281, 127636 (2025).","journal-title":"Expert Syst. Appl."},{"key":"47631_CR28","unstructured":"Pai, S. et al. Vision foundation models for computed tomography. arXiv preprint arXiv:2501.09001 (2025)."},{"key":"47631_CR29","doi-asserted-by":"publisher","first-page":"1108","DOI":"10.1001\/jama.2024.16332","volume":"332","author":"G Perrotta","year":"2024","unstructured":"Perrotta, G. et al. Accuracy of clinical staging in early-stage pancreatic ductal adenocarcinoma. JAMA 332, 1108\u20131110 (2024).","journal-title":"JAMA"},{"key":"47631_CR30","unstructured":"Ye, R. et al. Feddisco: Federated learning with discrepancy-aware collaboration. In International Conference on Machine Learning, 39879\u201339902 (PMLR, 2023)."},{"key":"47631_CR31","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.230024","volume":"5","author":"J Wasserthal","year":"2023","unstructured":"Wasserthal, J. et al. Totalsegmentator: robust segmentation of 104 anatomic structures in ct images. Radiol. Artif. Intell. 5, e230024 (2023).","journal-title":"Radiol. Artif. Intell."},{"key":"47631_CR32","first-page":"429","volume":"2","author":"T Li","year":"2020","unstructured":"Li, T. et al. Federated optimization in heterogeneous networks. Proc. Mach. Learn. Syst. 2, 429\u2013450 (2020).","journal-title":"Proc. Mach. Learn. Syst."},{"key":"47631_CR33","unstructured":"Karimireddy, S.\u00a0P. et al. Scaffold: Stochastic controlled averaging for federated learning. In International conference on machine learning, 5132\u20135143 (PMLR, 2020)."},{"key":"47631_CR34","unstructured":"Cardoso, M.\u00a0J. et al. Monai: An open-source framework for deep learning in healthcare. arXiv preprint arXiv:2211.02701 (2022)."},{"key":"47631_CR35","doi-asserted-by":"publisher","DOI":"10.1016\/j.inffus.2024.102576","volume":"112","author":"W Huang","year":"2024","unstructured":"Huang, W. et al. Multimodal federated learning: Concept, methods, applications and future directions. Inf. Fusion 112, 102576 (2024).","journal-title":"Inf. Fusion"}],"container-title":["Scientific Reports"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-47631-2","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-47631-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-47631-2.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,10]],"date-time":"2026-04-10T10:05:12Z","timestamp":1775815512000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41598-026-47631-2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,9]]},"references-count":35,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2026,12]]}},"alternative-id":["47631"],"URL":"https:\/\/doi.org\/10.1038\/s41598-026-47631-2","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-8369921\/v1","asserted-by":"object"}]},"ISSN":["2045-2322"],"issn-type":[{"value":"2045-2322","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,9]]},"assertion":[{"value":"15 December 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"1 April 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"9 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":"All procedures involving human participants were performed in accordance with the ethical standards of the institutional and national research committees and with the 1964 Declaration of Helsinki and its later amendments. This retrospective study was approved by the Ethics Committee of the University of Bonn (approval number 462\/21), the Ethics Committee of the University of G\u00f6ttingen, and the Ethics Committee of Charit\u00e9-Universit\u00e4tsmedizin Berlin (EA1\/292\/16). Given the retrospective design and full anonymization of all imaging and clinical data before analysis, the requirement for informed consent was waived by all approving ethics committees. No identifiable patient information was accessed or shared during the study, and all data handling complied with relevant data protection regulations, including the GDPR.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical approval"}}],"article-number":"12051"}}