TRACE: Evidence-Linked Computational Representation of Coronary Quantification from CCTA to Clinical Reports
Abstract
Coronary computed tomography angiography (CCTA) quantification requires anatomical and measurement context for interpretation and reuse. The Traceable Representation of Anatomy and Coronary Evidence (TRACE) links named anatomy, centreline geometry, lumen profiles, lesions, supporting sections, provenance and quality-related states. Evaluation used a 90-study institutional archive and 1000 ImageCAS studies. Clinical comparisons distinguished 30 studies using original program output-state eligibility from 16 three-source studies using a human-reviewed retained-lesion register. The separate ImageCAS geometry pathway processed binary masks under original, one-voxel erosion, dilation and opening conditions. All 4000 ImageCAS runs completed. Median total-branch-length changes were 1.6% and 1.9% after erosion and dilation, while local diameter and category outputs were more sensitive. Clinical rank correspondence depended on the endpoint workflow: Spearman ρ was 0.408 against hospital CCTA in the original-output subset (101 territory pairs), and 0.814 against hospital CCTA (64 pairs) and 0.617 against hospital angiography (60 pairs) in the human-reviewed register subset. At the report-defined ≥ 50% stenosis target, the corresponding areas under the receiver operating characteristic curve were 0.767 (95% CI 0.587–0.966), 0.943 (0.850–1.000) and 0.825 (0.706–0.934), respectively (diagnostically eligible n = 101, 60 and 56 after excluding raw non-assessable program territories). TRACE provides an evidence-linked representation layer that makes coronary findings and their endpoint construction inspectable. The results distinguish stable global geometry from sensitive local measurements and source-dependent report correspondence.
Ethics Statement
The retrospective analysis of institutional clinical data was approved by the Ethics Committee of the First Affiliated Hospital with Nanjing Medical University (Jiangsu Province Hospital), Nanjing, China (approval no. 2025-SR-1242; 7 January 2026). The study was conducted in accordance with the Declaration of Helsinki. The requirement for informed consent was waived by the approving ethics committee. Results are reported without directly identifying patient information.
