Wei Zhang, Professor at Comprehensive Cancer Center of Wake Forest Baptist Medical Center, shared on LinkedIn:
“Our team at the Atrium Health Wake Forest Baptist Comprehensive Cancer Center is grateful to be funded for the second year and to work with our colleagues at Moffitt Cancer Center, Medical University of South Carolina, and University of Texas as part of the NCI FLAIMME project (Federated Learning-based AI in Multimodal Exploration), a consortium managed by the National Cancer Institute’s Center for Biomedical Informatics and Information Technology (CBIIT).
The project’s primary mission is to establish a secure infrastructure that allows NCI-designated cancer centers to co-develop and share advanced artificial intelligence (AI) models across multiple oncology research domains without compromising patient data privacy.
Key Focus Areas and Technology
- Federated Learning Infrastructure: Instead of centralizing sensitive patient data from different hospitals—which raises massive security, legal, and privacy challenges-FLAIMME uses the NVIDIA FLARE (Federated Learning Application Runtime Environment) framework. This setup sends the AI model to the data hosts to train locally and only brings back the optimized weights.
- Clinical Trial Matching via MedGemma: A primary focal point of the consortium involves four major NCI-designated cancer centers utilizing a federated version of MedGemma (a medical-specific AI model built on Google’s Gemma 3 architecture) to optimize and automate how patients are matched to clinical oncology trials across multiple sites.
- Diverse Oncology Use Cases: The project spans six distinct oncology use cases, tackling multi-site operational issues like data harmonization, model deployment timelines, infrastructure failure modes, and cross-institutional governance.
Thank you, Umit Topaloglu, for your vision and leadership!”
Other articles about AI on OncoDaily.