RADICAL: Advancing Imaging Research Without Moving the Data

Medical imaging research has traditionally depended on bringing data together. Researchers identify imaging studies across participating institutions, transfer them into a central repository and analyze the resulting collection.

That approach has supported countless scientific advances, but it carries real costs. Institutions may face resource constraints, competing operational priorities or restrictions on how clinical data can be shared. Even when successfully built, centralized repositories often capture only a fraction of the imaging generated in routine clinical practice.

What if researchers could bring the research to the data instead?

That question sits at the heart of RADICAL, “Research and Data Integration for Cancer Through Advanced Linkage”, an initiative the ACR led through its work in the Advanced Research Projects Agency for Health's Biomedical Data Fabric program.

Rather than requiring participating institutions to transfer large collections of imaging data to a central location, RADICAL tested whether several emerging research capabilities could work together in a real-world multisite study while allowing key research activities to occur within each institution's own environment.

Bringing Research to the Data

Two academic medical centers put the model to the test: The University of North Carolina at Chapel Hill and the University of Washington.

Using local ACR Connect implementations, each institution identified and processed its own imaging cohort, generating standardized imaging metadata and privacy-preserving identifiers on site, while directly identifying information remained under institutional control.

A shared technology framework supported this work. Local ACR Connect installations interfaced with institutional systems while ACR centralized research infrastructure coordinated approved workflows and downstream analyses (see Figure 1).