Performance of Computed Tomographic Angiography–Based Aortic Valve Area for Assessment of Aortic Stenosis

Abstract

Background A total of 40% of patients with severe aortic stenosis (AS) have low‐gradient AS, raising uncertainty about AS severity. Aortic valve calcification, measured by computed tomography (CT), is guideline‐endorsed to aid in such cases. The performance of different CT‐derived aortic valve areas (AVAs) is less well studied. Methods and Results Consecutive adult patients with presumed moderate and severe AS based on echocardiography (AVA measured by continuity equation on echocardiography <1.5 cm 2 ) who underwent cardiac CT were identified retrospectively. AVAs, measured by direct planimetry on CT (AVA CT ) and by a hybrid approach (AVA measured in a hybrid manner with echocardiography and CT [AVA Hybrid ]), were measured. Sex‐specific aortic valve calcification thresholds (≥1200 Agatston units in women and ≥2000 Agatston units in men) were applied to adjudicate severe or nonsevere AS. A total of 215 patients (38.0% women; mean±SD age, 78±8 years) were included: normal flow, 59.5%; and low flow, 40.5%. Among the different thresholds for AVA CT and AVA Hybrid , diagnostic performance was the best for AVA CT <1.2 cm 2 (sensitivity, 85%; specificity, 26%; and accuracy, 72%), with no significant difference by flow status. The percentage of patients with correctly classified AS severity (correctly classified severe AS+correctly classified moderate AS) was as follows; AVA measured by continuity equation on echocardiography <1.0 cm 2 , 77%; AVA CT <1.2 cm 2 , 73%; AVA CT <1.0 cm 2 , 58%; AVA Hybrid <1.2 cm 2 , 59%; and AVA Hybrid <1.0 cm 2 , 45%. AVA CT cut points of 1.52 cm 2 for normal flow and 1.56 cm 2 for low flow, provided 95% specificity for excluding severe AS. Conclusions CT‐derived AVAs have poor discrimination for AS severity. Using an AVA CT <1.2‐cm 2 threshold to define severe AS can produce significant error. Larger AVA CT thresholds improve specificity.