Chest radiographs are routinely used to verify the position of catheters, tubes, and other support devices. Existing image models usually output labels or segmentations, while report‑processing systems only structure text and lack image geometry. This work introduces an uncertainty‑aware clinical knowledge graph that models device instances, tip estimates, placement assessments, provenance, report events, and temporal links as separate but interconnected evidence. We evaluate the visual graph layer using saved predictions from the RANZCR CLiP test archive, comprising 30,083 studies from 3,255 patients across five non‑overlapping outer folds. The graph builder materializes 914,632 B7 evidence nodes and 884,549 typed relationships. All 118,647 B7 predicted‑device nodes retain tip covariance, placement probabilities, fragment provenance, and fragment counts, whereas the direct B2 baseline retains none of these fields. We also define typed data contracts, uncertainty representations, abstention rules, report‑image grounding, and longitudinal query mechanisms to extend the graph to report‑bearing cohorts. The reported graph‑materialization analysis is post‑hoc descriptive and does not establish report grounding, longitudinal performance, or clinical utility, but it demonstrates a reproducible, evidence‑preserving foundation for AI reasoning over chest X‑ray device assessments.
Review