Back to all papers

Addressing data heterogeneity in distributed medical imaging with heterosync learning.

October 24, 2025pubmed logopapers

Authors

Hu HT,Li MD,Lin XX,Cai MY,Liu S,Wu SH,Tong WJ,Ye FY,Hu JB,Ke WP,Chen LD,Yang H,Liu GJ,Wang HB,Lu MD,Huang QH,Kuang M,Wang W

Affiliations (10)

  • Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China.
  • School of Physics and Electronic Information, Guangxi Minzu University, Nanning, China.
  • Department of Medical Ultrasound, the First Affiliated Hospital of Guangxi Medical University, Nanning, China.
  • Department of Medical Ultrasonics, the Sixth Affiliated Hospital of Sun Yat-sen University (Guangdong Gastrointestinal Hospital), Guangzhou, China.
  • Research Center of Big Data and Artificial Intelligence for Medicine, the First Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China.
  • Center of Hepato-Pancreato-Biliary Surgery, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China.
  • School of Artificial Intelligence, Optics and Electronics (iOPEN), Northwestern Polytechnical University, Xi'an, China. [email protected].
  • Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. [email protected].
  • Center of Hepato-Pancreato-Biliary Surgery, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. [email protected].
  • Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. [email protected].

Abstract

Data heterogeneity critically limits distributed artificial intelligence (AI) in medical imaging. We propose HeteroSync Learning (HSL), a privacy-preserving framework that addresses heterogeneity through: (1) Shared Anchor Task (SAT) for cross-node representation alignment, and (2) an Auxiliary Learning Architecture coordinating SAT with local primary tasks. Validated via large-scale simulations (feature/label/quantity/combined heterogeneity) and a real-world multi-center thyroid cancer study, HSL outperforms local learning, 12 benchmark methods (FedAvg, FedProx, SplitAVG, FedRCL, FedCOME, etc.), and foundation models (e.g., CLIP) by better stability and up to 40% in area under the curve (AUC), matching central learning performance. HSL achieves 0.846 AUC on the out-of-distribution pediatric thyroid cancer data (outperforming others by 5.1-28.2%), demonstrating superior generalization. Visualizations confirm HSL successfully homogenizes heterogeneous distributions. This work provides an effective solution for distributed medical AI, enabling equitable collaboration across institutions and advancing healthcare AI democratization.

Topics

Artificial IntelligenceDiagnostic ImagingMachine LearningImage Processing, Computer-AssistedJournal Article

Ready to Sharpen Your Edge?

Join hundreds of your peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.

We respect your privacy. Unsubscribe at any time.