Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Federated prediction for scalable and privacy-preserved knowledge-based planning in radiotherapy

Jingyun Chen, David Horowitz, Yading Yuan

TL;DR

This work tackles the core challenge of data scarcity and heterogeneity in radiotherapy planning by introducing FedKBP+, a federated learning platform that supports both centralized and decentralized training for knowledge-based planning tasks. FedKBP+ unifies OAR delineation, tumor segmentation, and dose prediction under the KBP+ framework and demonstrates that federation can approach pooled-data performance while preserving privacy, with robust behavior under data heterogeneity and user participation variability. The study provides extensive empirical evidence across OpenKBP, BraTS, and PanSeg datasets, showing faster convergence and competitive accuracy for FedKBP+ compared to existing FL tools, and highlights the framework’s modular, scalable, and deployment-ready design. Overall, FedKBP+ offers a practical, privacy-preserving path to scalable, institutionally adaptive RT predictive modeling, with open-source potential for broader adoption.

Abstract

Background: Deep learning has potential to improve the efficiency and consistency of radiation therapy planning, but clinical adoption is hindered by the limited model generalizability due to data scarcity and heterogeneity among institutions. Although aggregating data from different institutions could alleviate this problem, data sharing is a practical challenge due to concerns about patient data privacy and other technical obstacles. Purpose: This work aims to address this dilemma by developing FedKBP+, a comprehensive federated learning (FL) platform for predictive tasks in real-world applications in radiotherapy treatment planning. Methods: We implemented a unified communication stack based on Google Remote Procedure Call (gRPC) to support communication between participants whether located on the same workstation or distributed across multiple workstations. In addition to supporting the centralized FL strategies commonly available in existing open-source frameworks, FedKBP+ also provides a fully decentralized FL model where participants directly exchange model weights to each other through Peer-to-Peer communication. We evaluated FedKBP+ on three predictive tasks using scale-attention network (SA-Net) as the predictive model. Conclusions: Our results demonstrate that FedKBP+ is highly effective, efficient and robust, showing great potential as a federated learning platform for radiation therapy.

Federated prediction for scalable and privacy-preserved knowledge-based planning in radiotherapy

TL;DR

This work tackles the core challenge of data scarcity and heterogeneity in radiotherapy planning by introducing FedKBP+, a federated learning platform that supports both centralized and decentralized training for knowledge-based planning tasks. FedKBP+ unifies OAR delineation, tumor segmentation, and dose prediction under the KBP+ framework and demonstrates that federation can approach pooled-data performance while preserving privacy, with robust behavior under data heterogeneity and user participation variability. The study provides extensive empirical evidence across OpenKBP, BraTS, and PanSeg datasets, showing faster convergence and competitive accuracy for FedKBP+ compared to existing FL tools, and highlights the framework’s modular, scalable, and deployment-ready design. Overall, FedKBP+ offers a practical, privacy-preserving path to scalable, institutionally adaptive RT predictive modeling, with open-source potential for broader adoption.

Abstract

Background: Deep learning has potential to improve the efficiency and consistency of radiation therapy planning, but clinical adoption is hindered by the limited model generalizability due to data scarcity and heterogeneity among institutions. Although aggregating data from different institutions could alleviate this problem, data sharing is a practical challenge due to concerns about patient data privacy and other technical obstacles. Purpose: This work aims to address this dilemma by developing FedKBP+, a comprehensive federated learning (FL) platform for predictive tasks in real-world applications in radiotherapy treatment planning. Methods: We implemented a unified communication stack based on Google Remote Procedure Call (gRPC) to support communication between participants whether located on the same workstation or distributed across multiple workstations. In addition to supporting the centralized FL strategies commonly available in existing open-source frameworks, FedKBP+ also provides a fully decentralized FL model where participants directly exchange model weights to each other through Peer-to-Peer communication. We evaluated FedKBP+ on three predictive tasks using scale-attention network (SA-Net) as the predictive model. Conclusions: Our results demonstrate that FedKBP+ is highly effective, efficient and robust, showing great potential as a federated learning platform for radiation therapy.

Paper Structure

This paper contains 24 sections, 3 equations, 15 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. Methods
  3. Preliminaries
  4. FL workflows in FedKBP+
  5. Predictive model
  6. Communication protocol
  7. Experiments
  8. Federated dose prediction
  9. Data
  10. FL settings
  11. Implementations
  12. Results
  13. Federated tumor target segmentation
  14. Data
  15. FL settings
  16. ...and 9 more sections

Figures (15)

  • Figure 1: Workflow of the three predictive tasks in the Knowledge-Based Planning+ (KBP+) framework
  • Figure 2: FedKBP+: the FL framework for the predictive models in KBP+
  • Figure 3: Centralized FL in FedKBP+ framework.
  • Figure 4: Decentralized FL in FedKBP+ framework.
  • Figure 5: Architecture of SA-Net(a) including ResSE block(b) and Scale Attention block(c).
  • ...and 10 more figures