Abstract / Summary
To compare bronchial artery diameters measured by computed tomography angiography (CTA) and digital subtraction angiography (DSA) in patients with hemoptysis, and to assess whether active bleeding during imaging affects the measurement discrepancy. This single-center retrospective study was conducted at The First Affiliated Hospital of Chongqing Medical University, Chongqing, China. Patients with hemoptysis who underwent both bronchial artery CTA and subsequent DSA (before embolization) between June 2024 and June 2025 were enrolled, yielding 81 patients with 154 bronchial arteries available for paired analysis. To account for clustering of multiple arteries within the same patient, a linear mixed-effects model with a random intercept per patient was used. Subgroup analyses were stratified by the presence of hemoptysis during CTA and DSA. Inter-observer reproducibility was assessed using the intraclass correlation coefficient (ICC). A threshold-based agreement analysis was performed using a ≥ 2 mm cutoff for clinical relevance. CTA-derived diameters (2.34 ± 0.90 mm) were larger than DSA-derived diameters (2.18 ± 0.95 mm). The vessel-level mean CTA–DSA difference was 0.16 mm (95% limits of agreement [LoA] − 0.73 to + 1.04 mm), and the patient-level mean difference was 0.14 mm (95% LoA − 0.55 to + 0.83 mm). Measurements showed a strong positive correlation (Spearman’s ρ = 0.83, P < 0.001). Inter-observer reproducibility was good (ICC for CTA: 0.854; for DSA: 0.823). Using DSA ≥ 2 mm as the reference, CTA at the same ≥ 2 mm cutoff had a vessel-level sensitivity of 92.0% and specificity of 61.2%; at the patient level, sensitivity was 86.7% and specificity was 80.4%.In the LMM, none of the examined variables reached statistical significance, including bleeding status at CTA (no bleeding vs. bleeding: β = −0.024, 95% CI − 0.294 to 0.247, P = 0.863) and bleeding status at DSA (no bleeding vs. bleeding: β = 0.057, 95% CI − 0.175 to 0.288, P = 0.628). In exploratory vessel-level analyses, the median CTA–DSA difference was 0.00 mm (IQR − 0.20 to 0.25) when hemoptysis occurred only during CTA and 0.20 mm (IQR 0.10 to 0.50) when no bleeding occurred during either examination; the corresponding patient-level median differences were 0.00 mm and 0.33 mm. Group C (hemoptysis only during DSA) contained only one vessel/patient, precluding statistical inference. Between-group comparisons excluding Group C did not reach statistical significance (Kruskal–Wallis P = 0.154; all Dunn-adjusted P > 0.05). CTA shows a small mean systematic positive difference compared with DSA and remains useful for preprocedural assessment; however, individual-level discrepancies may be clinically relevant near the 2-mm threshold. The finding of a reduced systematic difference when hemoptysis occurs only during CTA is hypothesis-generating, not confirmatory, and transient vasoconstriction remains speculative.