Artificial intelligence (AI) offers a promising solution for streamlining COVID-19 diagnoses; however, data privacy concerns hinder the collection of large-scale representative medical data. To address this, the Unified CT-COVID AI Diagnostic Initiative (UCADI) utilizes federated learning, enabling distributed model training and independent execution at each institution without data sharing. The federated learning model significantly outperformed local models and achieved comparable performance to expert radiologists. The model was further evaluated on hold-out and heterogeneous data, with visual explanations provided for its decisions. This study, based on 9,573 chest CT scans from 23 hospitals in China and the UK, demonstrates the potential of federated learning for privacy-preserving AI in digital health.

Xiang Bai, Hanchen Wang, Liya Ma, Yongchao Xu, Jiefeng Gan, Ziwei Fan, Fan Yang, Ke Ma, Jiehua Yang, Song Bai, Chang Shu, Xinyu Zou, Renhao Huang, Changzheng Zhang, Xiaowu Liu, Dandan Tu, Chuou Xu, Wenqing Zhang, Xi Wang, Anguo Chen, Yu Zeng, Dehua Yang, Ming-Wei Wang, Nagaraj Holalkere, Neil J. Halin, Ihab R. Kamel, Jia Wu, Xuehua Peng, Xiang Wang, Jianbo Shao, Pattanasak Mongkolwat, Jianjun Zhang, Weiyang Liu, Michael Roberts, Zhongzhao Teng, Lucian Beer, Lorena E. Sanchez, Evis Sala, Daniel L. Rubin, Adrian Weller, Joan Lasenby, Chuansheng Zheng, Jianming Wang, Zhen Li, Carola Schönlieb, Tian Xia