Model utility and explainability in federated learning - A case study in healthcare using fundus oculi datasets

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: Up to 9.97% improvement compared to isolated training

• Primary Outcome: Model utility improvement with federated learning methods

• Nnt Or Sensitivity: High agreement between model explanations and ophthalmologist annotations for vertical cup-to-disc ratio detection

• Confidence Interval: Not reported

Clinical Application

High feasibility given widespread use of fundus imaging in ophthalmology practice Highly relevant to Australian practice where glaucoma screening is covered under Medicare and federated learning could enhance diagnostic capabilities across distributed healthcare networks while maintaining patient privacy Patients undergoing fundus photography for glaucoma screening and diagnosis

Abstract

OBJECTIVE: Introduce a case study for Federated Learning (FL) in healthcare, addressing challenges posed by patient privacy and limited large-scale datasets. Our goal is to assess the features learned by FL methods in a simulated, diverse setting that emphasizes realistic data heterogeneity, and to analyze the learned representations for their medical relevance using both local and global explainability techniques. METHODS: Six fundus oculi datasets were combined to simulate a diverse federated learning environment, representing heterogeneous data conditions. We evaluated three established FL methods against centrally trained models, assessing both predictive performance and the learned representations. Specifically, explainability techniques were employed to examine the features learned by the models, and local explanations were evaluated against attention maps annotated by ophthalmologists. Robustness against common biases in fundus datasets was also assessed. RESULTS: Our study found improvements in model utility (up to 9.97%) with FL methods compared to isolated training. Analysis of learned representations revealed that federated models predominantly learn the vertical cup-to-disc ratio, a crucial feature for glaucoma diagnosis, and demonstrated robustness against common biases. High agreement was observed between local explanations and ophthalmologist-annotated attention maps. CONCLUSION: This study demonstrates the benefits of FL systems in a healthcare scenario, providing a case study for evaluating federated systems beyond idealized benchmarks. Our findings highlight the potential of FL to not only improve model utility in privacy-sensitive medical domains but also to learn medically relevant features instead of spurious correlations.

References

1. 1.Penzel, N., Scheliga, D., Oppermann, H., Mäder, P., Haueisen, J., Denzler, J., & Seeland, M. (2026). Model utility and explainability in federated learning - A case study in healthcare using fundus oculi datasets. Journal of Biomedical Informatics, 105010. https://doi.org/10.1016/j.jbi.2026.105010