Import and curation.

Getting data in shouldn't be a project. Radial speaks DICOM natively: point your systems at it, work the queue, and curate as cases land. The path is the same for one case or a decade of archive.

Getting data in

Radial includes its own DICOM service. Send studies to it directly from the TPS or the archive the way you'd send to any DICOM node, query and retrieve from an archive you already run, or import exported files. However it arrives, everything lands in the same staged import queue.

Working the queue

Staged studies are grouped into cases, and duplicates or re-exports are reconciled before anything counts: conflicting versions of a study become an explicit merge decision, not a silent overwrite. The queue shows what arrived, what's computing, and what's waiting on a person.

The curation pass

For each case, the curation cockpit walks the proposals: structure names against TG-263, and the cohort assignment with its evidence. Most decisions are one keystroke; anything below confidence stays flagged. Cases join their cohort as they land (assignment happens on import), and curation is what makes them trustworthy: every confirmation is stamped with who approved it, and any query can be scoped to curate-approved cases, or their not-yet-approved remainder, when you want approval to gate what you count.

Day one vs steady state

The first weeks are the heavy lift: the backlog imports, and your confirmations teach the dictionary your institution's vocabulary. Steady state is lighter; new cases arrive, mostly map themselves, and curation becomes a short review rather than a data-entry job.