Procedural planning for transcatheter aortic valve replacement (TAVR) is routinely performed using contrast computed tomography (CT) in patients with severe aortic stenosis (AS). Despite its potential, little investigation has been done into the possibility of aortic valve (AV) calcification scoring in contrastenhanced CT. This is valuable because contrastenhanced CT has better dynamic contrast and spatial resolution than noncontrast CT, allowing for better descriptors of calcific lesions on the leaflets. We developed a new method that automatically finds a threshold to minimize the false positive rate (FPR), defined as the ratio of falsely labeled calcific pixels to the total number of calcific pixels (Figure 1A). We also introduce a scheme for describing the geometric and relative concentration of calcification via intensity as well as radial and longitudinal maps (Figure 1B). The method was evaluated on a retrospective cohort of 201 AS patients (Hamilton General Hospital, Hamilton, Ontario, Canada) between 2020 and 2022(sex: 45%/55% female/male; AV morphology: 21%/79% bicuspid/tricuspid; AS severity: 30%/70% moderate/severe). The protocol was approved by the Institutional Research Ethics Board. Scores produced using the proposed method were compared with the standard scores produced from noncontrast images using the calcium scoring application syngo. via (Siemens Healthineers). We also compared the method against techniques based on either fixed HU threshold (650) or a mean HU attenuation at the ascending aorta (þ2 SD) using the same field of view and region of interest. Correlations and comparisons between continuous variables were evaluated using Spearman’s rank test and Bland-Altman analysis of difference. Interobserver variability was assessed on a random subgroup of 49 patients using the intraclass correlation coefficient (2-way random agreement) showing excellent reproducibility (intraclass correlation coefficient: 0.95 [95% CI: 0.94-0.96]; raters ¼ 2). The FPR method showed the strongest correlation with noncontrast volume and Agatston scores (R ¼ 0.86 [FPR] vs R ¼ 0.84 [mean LAþ2 SD] vs R ¼ 0.68 [650 HU]; P< 0.001; N ¼ 201) evaluated using the standard 130-HU technique. 1, 2 The FPR method had the best correlation with hemodynamic determinants of AS severity compared with other evaluated contrast methods (indexed aortic valve area: R ¼ 0.32; mean aortic pressure gradient: R ¼ 0.43; peak aortic jet velocity: R ¼ 0.41; P< 0.001; N ¼ 201). The method also demonstrates a stronger correlation with the pressure gradient and velocity when compared with noncontrast volume or Agatston scores (Figure 1C). Finally, we examined the differences between volume scores measured using contrast vs noncontrast images for all methods and found that contrast images yielded significantly lower volume estimates than those for noncontrast images (Figure 1D). All methods exhibited a proportional underestimation compared with noncontrast methods, which increased at larger calcium volumes. This effect was determined by the Bland-Altman level-of-agreement analysis (for FPR, mean difference: 1,957 [limits of agreement: À407 to 4,322]; for mean LAþ2 SD, mean difference: 1,638 [limits of agreement: À570 to 3,845]; for HU 650 mean difference: 2,104 [limits of agreement: À682 to 4,890]) followed by simple linear regression to quantify the significance and proportionality of the bias