Federated Continual Learning for Privacy-Preserving Hospital Imaging Classification

TL;DR

This paper introduces DP-FedEPC, a federated continual learning method for hospital imaging that preserves privacy while adapting to evolving data streams, addressing catastrophic forgetting without sharing raw data.

Contribution

It proposes a novel federated continual learning approach combining elastic weight consolidation, prototype rehearsal, and differential privacy tailored for healthcare imaging.

Findings

Effective in maintaining performance across evolving data streams.

Provides formal privacy guarantees for sensitive radiology data.

Reduces catastrophic forgetting in federated hospital settings.

Abstract

Deep learning models for radiology interpretation increasingly rely on multi-institutional data, yet privacy regulations and distribution shift across hospitals limit central data pooling. Federated learning (FL) allows hospitals to collaboratively train models without sharing raw images, but current FL algorithms typically assume a static data distribution. In practice, hospitals experience continual evolution in case mix, annotation protocols, and imaging devices, which leads to catastrophic forgetting when models are updated sequentially. Federated continual learning (FCL) aims to reconcile these challenges but existing methods either ignore the stringent privacy constraints of healthcare or rely on replay buffers and public surrogate datasets that are difficult to justify in clinical settings. We study FCL for chest radiography classification in a setting where hospitals are clients that receive temporally evolving streams of cases and labels. We introduce DP-FedEPC (Differentially Private Federated Elastic Prototype Consolidation), a method that combines elastic weight consolidation (EWC), prototype-based rehearsal, and client-side differential privacy within a standard FedAvg framework. EWC constrains updates along parameters deemed important for previous tasks, while a memory of latent prototypes preserves class structure without storing raw images. Differentially private stochastic gradient descent (DP-SGD) at each client adds calibrated Gaussian noise to clipped gradients, providing formal privacy guarantees for individual radiographs.