The MICCAI Federated Tumor Segmentation (FeTS) Challenge 2024: Efficient and Robust Aggregation Methods for Federated Learning

2025-12-05

AI Summary

This study addresses the poor model robustness and low communication efficiency in federated segmentation of glioma subregions (enhancing tumor—ET, tumor core—TC, whole tumor—WT) from multiparametric MRI. We propose a novel weight aggregation method based on a PID controller, dynamically regulating client weight updates within the BraTS multicenter federated learning framework to enhance training stability and accelerate convergence. Experiments demonstrate state-of-the-art performance: Dice similarity coefficients of 0.733 (ET), 0.761 (TC), and 0.751 (WT), with corresponding 95th-percentile Hausdorff distances of 33.922 mm, 33.623 mm, and 32.309 mm; the overall convergence score reaches 0.764—surpassing all prior top-performing methods in the BraTS Federated Challenge. The key contribution is the first integration of control-theoretic principles into federated weight aggregation, simultaneously improving segmentation accuracy, model robustness, and communication efficiency.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Segmentation
• Intelligent Robots: Behavior Learning & Control

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

We present the design and results of the MICCAI Federated Tumor Segmentation (FeTS) Challenge 2024, which focuses on federated learning (FL) for glioma sub-region segmentation in multi-parametric MRI and evaluates new weight aggregation methods aimed at improving robustness and efficiency. Six participating teams were evaluated using a standardized FL setup and a multi-institutional dataset derived from the BraTS glioma benchmark, consisting of 1,251 training cases, 219 validation cases, and 570 hidden test cases with segmentations for enhancing tumor (ET), tumor core (TC), and whole tumor (WT). Teams were ranked using a cumulative scoring system that considered both segmentation performance, measured by Dice Similarity Coefficient (DSC) and the 95th percentile Hausdorff Distance (HD95), and communication efficiency assessed through the convergence score. A PID-controller-based method achieved the top overall ranking, obtaining mean DSC values of 0.733, 0.761, and 0.751 for ET, TC, and WT, respectively, with corresponding HD95 values of 33.922 mm, 33.623 mm, and 32.309 mm, while also demonstrating the highest communication efficiency with a convergence score of 0.764. These findings advance the state of federated learning for medical imaging, surpassing top-performing methods from previous challenge iterations and highlighting PID controllers as effective mechanisms for stabilizing and optimizing weight aggregation in FL. The challenge code is available at https://github.com/FeTS-AI/Challenge.

Problem

• Develops efficient federated learning methods for glioma segmentation
• Evaluates robust weight aggregation to improve multi-institutional MRI analysis
• Enhances communication efficiency and segmentation accuracy in federated settings

Innovation

• PID-controller-based method for federated weight aggregation
• Multi-institutional dataset with 1,251 training cases for glioma segmentation
• Cumulative scoring system evaluating segmentation performance and communication efficiency