Federated Distillation for Whole Slide Image via Gaussian-Mixture Feature Alignment and Curriculum Integration

Federated learning (FL) offers a promising framework for collaborative digital pathology by enabling model training across institutions. However, real-world deployments face heterogeneity arising from diverse multiple instance learning (MIL) architectures and heterogeneous feature extractors across institutions. We propose FedHD, a novel FL framework that performs local Gaussian-mixture feature alignment tailored for WSI analysis. Instead of exchanging model parameters, each client independently distills semantically rich synthetic feature representations aligned with the distribution of real WSIs. To preserve diagnostic diversity, FedHD adopts a one-to-one distillation strategy, generating a synthetic counterpart for each real slide to avoid over-compression. During federation, a curriculum-based integration strategy progressively incorporates cross-site synthetic features into local training once performance plateaus. Furthermore, an optional interpretation module reconstructs pseudo-patches from synthetic embeddings, enhancing transparency. FedHD is architecture-agnostic, privacy-preserving, and supports personalized yet collaborative training across diverse institutions. Experiments on TCGA-IDH, CAMELYON16, and CAMELYON17 show that FedHD consistently outperforms state-of-the-art federated and distillation baselines.

Medical AI for digital pathology often benefits from data across multiple hospitals, but hospitals usually cannot share raw patient images because of privacy concerns. They may also use different model designs, making standard collaborative training difficult. This paper proposes FedHD, a method that lets hospitals collaborate by sharing compact synthetic representations instead of raw whole slide images or model parameters. These representations preserve useful diagnostic patterns from local pathology slides while supporting different model architectures across institutions. Experiments on multiple pathology datasets show that FedHD improves classification performance compared with existing federated and distillation methods. The method also includes an optional visualization tool to help inspect the shared synthetic information.