Last week, the Cancer AI Alliance (CAIA), a research collaboration among cancer centers, announced the development of the first scalable platform utilizing federated learning for cancer research.
According to the news release, the platform serves as the technological foundation for CAIA’s aim to save more lives by enabling researchers and clinicians to train AI models that learn from the millions of clinical data points from participating cancer centers.
CAIA explained: Participating cancer centers will implement federated learning technology at their facilities, each connecting to a central orchestration component of the platform. Using this architecture, AI models travel to each cancer center’s secure data to learn locally, generating a summary of their findings without any clinical data leaving the institutional firewalls. The insights gained from training the model on each center’s de-identified data are then combined centrally to improve the AI models' ability to detect patterns, maximizing the value of the collective knowledge base.
Researchers across the participating cancer centers are launching eight unique projects, CAIA described. The initial projects aim to address some of oncology's most persistent challenges, including predicting treatment response, identifying novel biomarkers, and analyzing trends in rare cancers. The projects utilize the federated learning platform and structured, de-identified data securely stored by participating cancer centers, which collectively provide a diverse and representative foundation of over 1 million patients for modeling and analysis.
The platform's true power, however, lies in its potential to scale up, the press announcement underscored. Over the next year, CAIA plans to enable dozens of research models and expand the alliance to include more participants.
“These updated AI models could significantly improve health outcomes for cancer patients by revealing trends across more diverse populations and rare cancers,” the announcement highlighted.
“Currently, it takes years to uncover new insights that can be translated into better treatments and care, but this platform will accelerate the pace of breakthrough discoveries by up to tenfold, reducing that time from years to months,” said Jeff Leek, Ph.D., VP and chief data officer at Fred Hutch, in a statement.
CAIA is made up of National Cancer Institute-designated cancer centers: Dana-Farber Cancer Institute, Fred Hutch Cancer Center, Memorial Sloan Kettering Cancer Center, and The Sidney Kimmel Comprehensive Cancer Center, along with the Whiting School of Engineering at Johns Hopkins. It receives financial and technological support from industry leaders like Amazon Web Services (AWS), Deloitte, Ai2 (Allen Institute for AI), Google, Microsoft, NVIDIA, and Slalom.
