Andrey Fedorov, PhD: Imaging Data Commons: Data + Tools + Compute to Power Your AI Explorations

This talk was presented at a AI4C (Artificial Intelligence for Clinicians) Seminar Series Event.

Rapid advances in technology open the possibility to fundamentally change medical imaging research. Ever-increasing capabilities of Graphical Processing Units (GPU) defied processing performance growth expectations. This, in turn, precipitated revolution in Artificial Intelligence, enabling automation of medical image analysis tasks that remained unsolved for decades. Advancements in virtualization and cloud computing provide unprecedented computational power without extensive investment into on-premises infrastructure. Open source development of innovative AI tools became the norm, streamlining access to the latest advances in AI, their evaluation and refinement.

National Cancer Institute Imaging Data Commons (IDC) is a cloud-based environment containing ~80TB of public cancer imaging data, opening new opportunities to leverage these technology advances for the development of robust cancer image analysis tools. IDC supports life-cycle of imaging data from curation into standard representation and harmonization of metadata on ingestion, to integration with cloud-based tools for seamless visualization and exploration, application of analysis workflows, and continuous enrichment of the images with annotations and analysis results. I will give an overview of IDC focusing on its capabilities, opportunities and examples of enabling AI-driven analysis, and the potential to catalyze collaboration and discovery in medical imaging research.

Andrey Fedorov is a researcher at Brigham and Women’s Hospital (BWH) and Associate Professor in Radiology at Harvard Medical School. Andrey is one of the leads of the team tasked with building National Cancer Institute Imaging Data Commons (IDC). A computer scientist by training, Andrey spent the past 15 years at the BWH Surgical Planning Lab working on translation and evaluation of image computing tools in clinical research applications. He is dedicated to developing infrastructure and best practices to help imaging researchers improve transparency of their studies, simplify data sharing and make their analyses more easily accessible and reproducible by others.