Sex-specific CT-derived aortic valve calcification predicts hemodynamically true severe aortic stenosis defined by invasive catheterization
T Yoshida, T Okuno, P J Santillan, T Kai, M Sekiguchi, Y Kobayashi, T Tanaka, R Oda, A Saito, T Shoji, D Miyahara, Y Sato, S Kuwata, M Izumo, Y J AkashiAbstract
Background
Sex-specific Agatston unit (AU) thresholds are recommended for the diagnosis of severe aortic stenosis (AS)1-3, and contrast-enhanced CT (CECT) derived calcium volume scores (CVS) have been proposed as an alternative4,5. However, the diagnostic performance of them has not yet been validated against invasively confirmed, hemodynamically true severe AS.
Purpose
We aimed to determine the prevalence of invasively confirmed true severe AS, to evaluate whether conventional sex-specific AU thresholds discriminate it, and to derive optimal cutoffs for AU and CVS.
Methods
Patients were enrolled from the prospective Invasive PPM Study between October 2024 and July 2025, designed to evaluate valve hemodynamics before and after transcatheter aortic valve replacement (TAVR) using right- and left-heart catheterization. Patients who underwent non-contrasted CT and CECT prior to TAVR were included. Hemodynamically true severe AS was defined as an invasively measured aortic valve area (AVA) ≤1.0 cm² calculated using the thermodilution-based Gorlin formula, whereas AVA >1.0 cm² was classified as non-severe AS. The AVC score was quantified using the Agatston method with sex-specific AU thresholds (>2000 in men and >1200 in women)1,3. CVS was calculated from CECT using a threshold of ≥ 850 HU.
Results
Of 100 enrolled patients, 85 were eligible for analysis. True severe AS was present in 51 (60%). Invasive AVA was lower in the true severe AS group than in the non-severe AS group (0.71 vs. 1.17; p < 0.01), as was echocardiography-derived AVA (0.55 vs. 0.70; p < 0.01). Both the AVC score and CVS were higher in true severe AS (AVC score: 2336.3 vs. 1654.1; CVS: 273.0 vs. 164.5; both p < 0.01). ROC analysis showed good discriminatory performance of the AVC score for identifying true severe AS, with an AUC of 0.76 in men and 0.85 in women. Using guideline-based sex-specific thresholds, sensitivity (Sn) and specificity (Sp) were 0.85 and 0.58 in men and 0.87 and 0.67 in women, respectively. The optimal cutoffs were 2272 AU for men (Sn 0.85, Sp 0.68; Figure A) and 1429 AU for women (Sn 0.74, Sp 0.93; Figure B). For CVS, AUC was 0.73 in men and 0.76 in women. The optimal cutoffs were 430 mm³ for men (Sn 0.50, Sp 0.90; Figure C) and 178 mm³ for women (Sn 0.61, Sp 0.93; Figure D). In men, there was no significant difference in AUC between the AVC score and CVS (p = 0.62), whereas in women, the AVC score demonstrated significantly superior discriminatory performance compared with CVS (p = 0.03).
Conclusions
Sex-specific AVC score provided robust discrimination of invasively-confirmed true severe AS. The optimal cutoffs were higher than the conventional thresholds in both sexes. Although its performance was similar to CVS in men, the AVC score significantly outperformed CVS in women. This is the first study to validate the diagnostic utility of AVC score against invasively-confirmed true severe AS, supporting their clinical relevance.Figure A, BFigure C, D