Imaging follow-up of endovascular repair of type B aortic dissection with dual-source, dual-energy CT and late delayed-phase scans.

PURPOSE: To evaluate the diagnostic performance of dual-energy (DE) computed tomography (CT) after thoracic endovascular aortic repair (TEVAR) of type B dissection, and to investigate the value of late delayed (LD) acquisition in endoleak detection and false lumen patency assessment. MATERIALS AND METHODS: Twenty-four patients with TEVAR for type B dissection underwent 53 tripe-phase CT examinations. Single-source unenhanced acquisition was followed by single-source arterial-phase and DE LD phase (300-s delay) imaging. Virtual noncontrast images were generated from DE acquisition. Two blinded radiologists retrospectively evaluated the cases in three reading sessions: session A (triphasic protocol), session B (virtual noncontrast and arterial phase), and session C (virtual noncontrast and arterial and LD phases). Endoleak detection accuracy during sessions B and C compared with session A (reference standard) was investigated. False lumen patency was assessed. Effective radiation dose was calculated. RESULTS: Session A revealed 37 endoleaks in 30 of 53 studies (56.6%). Session B revealed 31 of the 37 endoleaks, with one false-positive case, 83.8% sensitivity, 95.8% specificity, 79.3% negative predictive value, and 96.9% positive predictive value. Session C correctly depicted all 37 endoleaks, with one false-positive case, 100% sensitivity, 95.8% specificity, 100% negative predictive value, and 97.4% positive predictive value. Underestimation of false lumen patency was found in session B (P = .013). Virtual noncontrast imaging resulted in 17% radiation exposure reduction. CONCLUSIONS: Virtual noncontrast imaging can replace standard unenhanced images in follow-up after TEVAR of type B dissection, thus reducing radiation dose. Delayed-phase imaging is valuable in low-flow endoleaks detection and false lumen patency assessment.