Pancreas Part Segmentation under Federated Learning Paradigm

2025-09-27

AI Summary

Pancreatic head, body, and tail segmentation in MRI remains challenging due to high morphological variability, low soft-tissue contrast, substantial inter-center anatomical heterogeneity, and scarcity of annotated data. To address these issues, we propose the first federated learning framework specifically designed for pancreatic regional segmentation. Our method integrates U-Net, Attention U-Net, and Swin UNETR backbones, employs FedAvg and FedProx optimization strategies, and introduces an anatomy-aware loss function to enforce region-specific supervision. Evaluated on distributed T1-weighted and T2-weighted MRI data from seven clinical sites, the framework significantly improves cross-center generalizability and robustness, achieving clinically viable segmentation accuracy. This work represents the first successful application of federated learning to precise subregional pancreatic segmentation, establishing a novel paradigm for localized pancreatic disease diagnosis and personalized treatment planning.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We present the first federated learning (FL) approach for pancreas part(head, body and tail) segmentation in MRI, addressing a critical clinical challenge as a significant innovation. Pancreatic diseases exhibit marked regional heterogeneity cancers predominantly occur in the head region while chronic pancreatitis causes tissue loss in the tail, making accurate segmentation of the organ into head, body, and tail regions essential for precise diagnosis and treatment planning. This segmentation task remains exceptionally challenging in MRI due to variable morphology, poor soft-tissue contrast, and anatomical variations across patients. Our novel contribution tackles two fundamental challenges: first, the technical complexity of pancreas part delineation in MRI, and second the data scarcity problem that has hindered prior approaches. We introduce a privacy-preserving FL framework that enables collaborative model training across seven medical institutions without direct data sharing, leveraging a diverse dataset of 711 T1W and 726 T2W MRI scans. Our key innovations include: (1) a systematic evaluation of three state-of-the-art segmentation architectures (U-Net, Attention U-Net,Swin UNETR) paired with two FL algorithms (FedAvg, FedProx), revealing Attention U-Net with FedAvg as optimal for pancreatic heterogeneity, which was never been done before; (2) a novel anatomically-informed loss function prioritizing region-specific texture contrasts in MRI. Comprehensive evaluation demonstrates that our approach achieves clinically viable performance despite training on distributed, heterogeneous datasets.

Problem

• Segments pancreas into head, body, and tail regions in MRI scans
• Addresses data scarcity and privacy via federated learning across institutions
• Overcomes poor MRI contrast and anatomical variations for precise diagnosis

Innovation

• Federated learning for multi-institutional pancreas segmentation
• Attention U-Net with FedAvg optimized for pancreatic heterogeneity
• Anatomy-informed loss function prioritizing MRI texture contrasts