Fair evaluation of federated learning algorithms for automated breast density classification: The results of the 2022 ACR-NCI-NVIDIA federated learning challenge

Kendall Schmidt, Benjamin Bearce, Ken Chang, Laura Coombs, Keyvan Farahani, Marawan Elbatel, Kaouther Mouheb, Robert Marti, Ruipeng Zhang, Ya Zhang, Yanfeng Wang, Yaojun Hu, Haochao Ying, Yuyang Xu, Conrad Testagrose, Mutlu Demirer, Vikash Gupta, Ünal Akünal, Markus Bujotzek, Klaus H. Maier-Hein et al. (4 additional authors not shown)

Abstract: Highlights•The ACR-NCI-NVIDIA federated learning challenge attracted worldwide participation.•Participants trained algorithms across three simulated radiology facilities.•The simulated facility with the smallest dataset size showed worst performance.•The best federated algorithm scored comparably to a centrally trained model.•Use of state-of-the-art federated learning algorithms did not guarantee top ranks.