Structure-name mapping and TG-263.

TG-263 is the AAPM's standard nomenclature for radiotherapy structures: the shared vocabulary that makes Parotid_L from one clinic comparable to LT PAROTID from another. Radial maps every imported structure to it, and shows its work.

Evidence-based, not a black box

Every ROI gets a predicted TG-263 name with a confidence and the reason it matched. Geometry counts as evidence alongside the name. A contour sitting a uniform 5 mm outside the cord reads as a PRV expansion, whatever the planner called it. The matching is deterministic: the same structure, the same dictionary, the same answer.

High confidence confirms; the rest queue

High-confidence matches confirm automatically, recorded like any hand confirmation. Everything else lands in a review queue built for one-keystroke decisions, and everything below confidence stays flagged for audit until a person has looked. Nothing files itself silently.

A dictionary that learns your institution

Confirm a local variant once and it becomes part of your deployment's dictionary. Radial will offer to backfill every existing structure that matches, and the next import of that variant maps itself. The learning is local, per deployment: your confirmations improve your institution's dictionary and stay inside it.

Why mapping is the foundation

Every downstream number depends on this layer. A parotid dose metric is only as good as the certainty that the structure really is the parotid. That is why mapping runs first in curation, and why the audit re-checks this layer every time you run it.