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 cm2) who underwent cardiac CT were identified retrospectively. AVAs, measured by direct planimetry on CT (AVACT) and by a hybrid approach (AVA measured in a hybrid manner with echo-cardiography and CT [AVAHybrid]), 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 AVACT and AVAHybrid, diagnostic performance was the best for AVACT <1.2 cm2 (sensitivity, 85%; specificity, 26%; and ac-curacy, 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 cm2, 77%; AVACT <1.2 cm2, 73%; AVACT <1.0 cm2, 58%; AVAHybrid <1.2 cm2, 59%; and AVAHybrid <1.0 cm2, 45%. AVACT cut points of 1.52 cm2 for normal flow and 1.56 cm2 for low flow, provided 95% specificity for excluding severe AS. CONCLUSIONS: CT-derived AVAs have poor discrimination for AS severity. Using an AVACT <1.2-cm2 threshold to define severe AS can produce significant error. Larger AVACT thresholds improve specificity.