ODELIA was featured in the session “Novel approaches for trustworthy artificial intelligence (AI) in cancer diagnostics and treatment” at the recent European Congress of Radiology 2026 in Vienna, AT.
A highlight of the session was the ODELIA presentation on swarm learning, an innovative approach enabling collaborative AI training without requiring data exchange between institutions.
Dr. Oliver Lester Saldanha presented ODELIA as a pan-European initiative that implements open-access Swarm Learning to democratize distributed AI training for breast cancer detection across eight international hospitals. By allowing models to travel while keeping raw patient data securely within each institution, this decentralized workflow provides a privacy-preserving alternative to commercial solutions. This collaborative approach has proven effective for high-quality model training, facilitating trustworthy medical AI without the need for direct data exchange. Beyond ODELIA, the presentation highlights a broader ecosystem of decentralized projects, including DECADE for colorectal cancer, SWAG for synthetic oncology data, and Surgery Swarm for surgical video analysis. Supporting these initiatives is MediSwarm, an open-source clinical platform designed to provide a secure, reproducible environment for real-world hospital deployment.
To watch the session online for free, simply log in via ESR Connect with your ESR MyUserArea credentials or create a free account. Then select the session you want to watch.
