From pretraining to privacy: federated ultrasound foundation model with self-supervised learning

doi:10.1038/s41746-025-02085-0

CSpace

	From pretraining to privacy: federated ultrasound foundation model with self-supervised learning
	Jiang, Yuncheng 1,2,3,4,5,6; Feng, Chun-Mei 7; Ren, Jinke 1,2; Wei, Jun 8; Zhang, Zixun 1,2; Hu, Yiwen 9; Liu, Yunbi 10; Sun, Rui 1,2; Tang, Xuemei 11; Du, Juan 12; Wan, Xiang 13; Xu, Yong 14; Du, Bo 15; Gao, Xin 16,17; Wang, Guangyu 18; Zhou, Shaohua 19,20; Cui, Shuguang 1,2; Li, Zhen 1,2
	2025-11-21
发表期刊	NPJ DIGITAL MEDICINE
ISSN	2398-6352
卷号	8 期号:1 页码:18
摘要	Ultrasound imaging is widely used in clinical diagnosis due to its non-invasive nature and real-time capabilities. However, traditional ultrasound diagnostics relies heavily on physician expertise and is often hampered by suboptimal image quality, leading to potential diagnostic errors. While artificial intelligence (AI) offers a promising solution to enhance clinical diagnosis by detecting abnormalities across various imaging modalities, existing AI methods for ultrasound face two major challenges. First, they typically require vast amounts of labeled medical data, raising serious concerns regarding patient privacy. Second, most models are designed for specific tasks, which restricts their broader clinical utility. To overcome these challenges, we present UltraFedFM, an innovative privacy-preserving ultrasound foundation model. UltraFedFM is collaboratively pre-trained using federated learning across 16 distributed medical institutions in 9 countries, leveraging a dataset of over 1 million ultrasound images covering 19 organs and 10 ultrasound modalities. This extensive and diverse data, combined with a secure training framework, enables UltraFedFM to exhibit strong generalization and diagnostic capabilities. It achieves an average area under the receiver operating characteristic curve (AUROC) of 0.927 for disease diagnosis and a dice similarity coefficient (DSC) of 0.878 for lesion segmentation. Notably, UltraFedFM surpasses the diagnostic accuracy of mid-level ultrasonographers (4-8 years of experience) and matches the performance of expert-level sonographers (10+ years of experience) in the joint diagnosis of 8 common systemic diseases.c These findings indicate that UltraFedFM can significantly enhance clinical diagnostics while safeguarding patient privacy, marking a significant advancement in AI-driven ultrasound imaging for future clinical applications.
DOI	10.1038/s41746-025-02085-0
收录类别	SCI
语种	英语
WOS研究方向	Health Care Sciences & Services ; Medical Informatics
WOS类目	Health Care Sciences & Services ; Medical Informatics
WOS记录号	WOS:001620847800004
出版者	NATURE PORTFOLIO
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/43072
专题	中国科学院计算技术研究所
通讯作者	Li, Zhen
作者单位	1.Chinese Univ Hong Kong, FNii Shenzhen, Shenzhen 518172, Peoples R China 2.Chinese Univ Hong Kong, Sch Sci & Engn, Shenzhen 518172, Peoples R China 3.Sichuan Univ, West China Hosp, Dept Gen Surg, Chengdu, Peoples R China 4.Sichuan Univ, Collaborat Innovat Ctr Biotherapy, Lab Gastr Canc, State Key Lab Biotherapy, Chengdu, Peoples R China 5.Sichuan Univ, West China Hosp, Canc Ctr, Chengdu, Peoples R China 6.Sichuan Univ, West China Hosp, Gastr Canc Ctr, Chengdu, Peoples R China 7.Univ Coll Dublin, Sch Comp Sci, Dublin, Ireland 8.Shenzhen Univ, Coll Comp Sci & Software Engn, Shenzhen 518060, Peoples R China 9.Shenzhen Univ, South China Hosp, Hlth Sci Ctr, Shenzhen 518111, Peoples R China 10.Nanjing Univ Posts & Telecommun, Sch Comp Sci, Nanjing 210023, Peoples R China 11.North Sichuan Med Coll, Affiliated Hosp, Nanchong 637000, Sichuan, Peoples R China 12.North Sichuan Med Coll, Nanchong 637000, Sichuan, Peoples R China 13.Shenzhen Res Inst Big Data, Shenzhen 518172, Peoples R China 14.Harbin Inst Technol, Biocomp Res Ctr, Shenzhen 518055, Peoples R China 15.Wuhan Univ, Sch Comp Sci, Wuhan 430072, Peoples R China 16.King Abdullah Univ Sci & Technol KAUST, Comp Sci Program, Comp Elect & Math Sci & Engn Div, Thuwal 239556900, Saudi Arabia 17.King Abdullah Univ Sci & Technol KAUST, Ctr Excellence Smart Hlth KCSH, Thuwal 239556900, Saudi Arabia 18.Beijing Univ Posts & Telecommun, Beijing 100876, Peoples R China 19.Univ Sci & Technol China, Suzhou Inst Adv Res, Sch Biomed Engn, Suzhou 215123, Peoples R China 20.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Jiang, Yuncheng,Feng, Chun-Mei,Ren, Jinke,et al. From pretraining to privacy: federated ultrasound foundation model with self-supervised learning[J]. NPJ DIGITAL MEDICINE,2025,8(1):18.
APA	Jiang, Yuncheng.,Feng, Chun-Mei.,Ren, Jinke.,Wei, Jun.,Zhang, Zixun.,...&Li, Zhen.(2025).From pretraining to privacy: federated ultrasound foundation model with self-supervised learning.NPJ DIGITAL MEDICINE,8(1),18.
MLA	Jiang, Yuncheng,et al."From pretraining to privacy: federated ultrasound foundation model with self-supervised learning".NPJ DIGITAL MEDICINE 8.1(2025):18.