MammoFL: Mammographic Breast Density Estimation using Federated Learning

TL;DR

This paper demonstrates that federated learning enables effective and privacy-preserving mammographic breast density estimation across multiple institutions, achieving similar accuracy to centralized training.

Contribution

It introduces a federated learning approach for breast density estimation using neural networks trained on multi-institutional mammogram datasets.

Findings

Federated learning improves model generalization on unseen data.

Training on multi-institutional data enhances algorithm robustness.

Federated approach maintains patient privacy while achieving high accuracy.

Abstract

In this study, we automate quantitative mammographic breast density estimation with neural networks and show that this tool is a strong use case for federated learning on multi-institutional datasets. Our dataset included bilateral CC-view and MLO-view mammographic images from two separate institutions. Two U-Nets were separately trained on algorithm-generated labels to perform segmentation of the breast and dense tissue from these images and subsequently calculate breast percent density (PD). The networks were trained with federated learning and compared to three non-federated baselines, one trained on each single-institution dataset and one trained on the aggregated multi-institution dataset. We demonstrate that training on multi-institution datasets is critical to algorithm generalizability. We further show that federated learning on multi-institutional datasets improves model generalization to unseen data at nearly the same level as centralized training on multi-institutional datasets, indicating that federated learning can be applied to our method to improve algorithm generalizability while maintaining patient privacy.