IP03 - Feasibility and accuracy of carbon dioxide Cone-Beam CT for post-operative evaluation in complex and standard EVAR

Objectives: To assess feasibility and diagnostic performance of carbon-dioxide cone-beam computed tomography (CO₂-CBCT) compared with conventional computed tomography angiography (CTA) for detecting technical complications after complex and standard aortic interventions.

Methods: We conducted a single-center retrospective study on patients undergoing standard and complex endovascular aortic repair (2023-2025). Only patients receiving CO2-CBCT and available computed tomography angiogram (CTA) performed within 30 days were selected. CO2-CBCT was performed with a ArtisPheno angiographer (Siemens) and Angiodroid automatic CO2 injector (Angiodroid Spa), using a 6F sheath positioned slightly above the celiac trunk.
Primary endpoints were CO2-CBCT image quality and accuracy in evaluating target vessel patency and endoleaks. Imaging quality was defined by the percentage of vessel lumen area occupied by CO2 on axial views, and was graded as excellent, high, moderate and low. Accuracy was determined by comparison of the results with pre-discharge CTA; the agreement between techniques was assessed using Cohen’s Kappa. Secondary endpoints were inter-observer agreement (Intra-class coefficient, ICC), and determination of optimal timing of CO2 injection relative to CBCT scan. Receiver Operating Characteristic (ROC) analysis was used to determine the optimal timing of injection.

Results: There were 70 patients, 40 treated by fenestrated EVAR, 10 branched EVAR, and 17 standard EVAR. Median timing of CO2 injection relative to CBCT acquisition was 0s (range: -2s; +1s), and CO2-CBCT image quality was overall classified as excellent in 42 (60%), high in 19 (27%), and moderate in 9 (13%).
Final CO2-CBCT demonstrated patency of all target vessels, and identified 23 endoleaks (33%; 17 type 2, 4 type 3, and 11 type 1a). Air trapping within the aneurysm sac without evidence of endoleaks, was detected in 6 (8%). The overall ICC for inter-observer agreement was 92%.
Pre-discharge CTA confirmed patency of all target vessels. Of the 23 endoleaks on CO2-CBCT, 16 (70%) were confirmed on CTA, while no endoleak was detected only on CTA. Comparing CO2-CBCT and CO2-CBCT, Cohen’s Kappa statistic was 100% for patency, type 3 endoleak, and type 1 endoleak; Cohen’s Kappa was 59% for type 2 endoleak. ROC analysis determined that the best timing for CO2 injection was 1.4s before CBCT acquisition. Posterior vessel origin angulation >90° was associated with suboptimal visualization (OR 2.4, 95%CI 1.1-6.5; P<.001).

Conclusions: CO2-CBCT can provide high imaging quality, adequate interobserver agreement, and comparable performance to CTA. The optimal timing of CO2 injection to optimize images quality is 1.4s before CBCT scan, while CO2-CBCT may be suboptimal for posteriorly oriented arteries.