Skip main navigation
×

Computed Tomography Aortic Valve Calcium Scoring in Patients With Aortic Stenosis

Originally publishedhttps://doi.org/10.1161/CIRCIMAGING.117.007146Circulation: Cardiovascular Imaging. 2018;11:e007146

    Abstract

    Background—

    Computed tomography aortic valve calcium scoring (CT-AVC) holds promise for the assessment of patients with aortic stenosis (AS). We sought to establish the clinical utility of CT-AVC in an international multicenter cohort of patients.

    Methods and Results—

    Patients with AS who underwent ECG-gated CT-AVC within 3 months of echocardiography were entered into an international, multicenter, observational registry. Optimal CT-AVC thresholds for diagnosing severe AS were determined in patients with concordant echocardiographic assessments, before being used to arbitrate disease severity in those with discordant measurements. In patients with long-term follow-up, we assessed whether CT-AVC thresholds predicted aortic valve replacement and death. In 918 patients from 8 centers (age, 77±10 years; 60% men; peak velocity, 3.88±0.90 m/s), 708 (77%) patients had concordant echocardiographic assessments, in whom CT-AVC provided excellent discrimination for severe AS (C statistic: women 0.92, men 0.89). Our optimal sex-specific CT-AVC thresholds (women 1377 Agatston unit and men 2062 Agatston unit) were nearly identical to those previously reported (women 1274 Agatston unit and men 2065 Agatston unit). Clinical outcomes were available in 215 patients (follow-up 1029 [126–2251] days). Sex-specific CT-AVC thresholds independently predicted aortic valve replacement and death (hazard ratio, 3.90 [95% confidence interval, 2.19–6.78]; P<0.001) after adjustment for age, sex, peak velocity, and aortic valve area. Among 210 (23%) patients with discordant echocardiographic assessments, there was considerable heterogeneity in CT-AVC scores, which again were an independent predictor of clinical outcomes (hazard ratio, 3.67 [95% confidence interval, 1.39–9.73]; P=0.010).

    Conclusions—

    Sex-specific CT-AVC thresholds accurately identify severe AS and provide powerful prognostic information. These findings support their integration into routine clinical practice.

    Clinical Trial Registration—

    URL: http://www.clinicaltrials.gov. Unique identifiers: NCT01358513, NCT02132026, NCT00338676, NCT00647088, NCT01679431.

    Introduction

    Aortic stenosis (AS) is a common and potentially fatal condition in which progressive fibrocalcific changes within the valve leaflets cause outflow tract obstruction. Severe symptomatic stenosis is an indication for aortic valve replacement (AVR), and timely referral is essential to prevent adverse clinical events.1 Stenosis severity is determined using echocardiography and commonly graded using the peak aortic jet velocity (Vmax), mean gradient, and aortic valve area (AVA).2 The latter is less flow dependent and can be indexed to body surface area 2 In the majority of patients, these measurements provide concordant assessments, and the severity of AS is clear. However, in around a quarter of cases, these measures are discordant, creating confusion as to the true severity of AS and difficulties in clinical decision making.3,4 An independent, complementary test that could be used to arbitrate the true severity of AS would, therefore, have major clinical utility and potentially improve patient care.5

    See Editorial by Vollema et al

    See Clinical Perspective

    In recent years, computed tomography aortic valve calcium scoring (CT-AVC) has emerged as a potential solution to this problem. Calcification is the predominant driver of AS68 and can be readily quantified using the same approach as coronary CT calcium scoring.9 Several studies have demonstrated an association between CT-AVC and hemodynamic measures of stenosis severity on echocardiography with women appearing to require less calcium to develop severe stenosis than men.1012 Recently, sex-specific CT-AVC thresholds have been proposed (women 1274 Agatston unit [AU] and men 2065 AU) to identify severe AS, as well as predicting disease progression and adverse clinical events.3,13 This has led to great interest in using CT-AVC as an alternative assessment of AS severity and as an umpire test in patients with discordant echocardiographic findings.14 Although CT calcium scoring has recently been recommended in the latest European Society of Cardiology guidelines (for patients with low flow and low ejection fraction with no demonstrable flow reserve on dobutamine stress echocardiography), it is not part of American Heart Association/American College of Cardiology guidelines, and before it enters routine clinical use, its widespread clinical applicability needs to be established.15

    The aim of this study was to investigate the clinical utility and generalizability of CT-AVC in an international multicenter registry incorporating a wide range of patient populations, different scanner vendors, and varied image analysis platforms. Specifically, we sought to determine the ability of CT-AVC to identify severe AS in those with concordant echocardiographic measures to establish the ability of this technique to predict clinical outcomes and to arbitrate disease severity in patients with discordant echocardiographic findings.

    Methods

    The data and analytic methods have been made available for purposes of reproducing the results or replicating the procedure. Eight international centers were invited to contribute clinical, CT and echocardiography data from patients with AS into a multicenter registry (Table I in the Data Supplement). Patients were required to have at least mild AS (peak aortic jet velocity, >2.5 m/s or mean gradient >10 mm Hg) and to have undergone ECG-gated CT calcium scoring within 3 months of the echocardiogram. Patients with established rheumatic heart disease, other forms of valvular heart disease of at least moderate severity, or an estimated glomerular filtration rate <30 mL/min per 1.73 m2 were excluded.

    Three centers contributed data from 5 prospective AS clinical research studies: Edinburgh Heart Centre, Edinburgh, United Kingdom, Hôpital Bichat, Paris, France, and Institut Universitaire de Cardiologie et de Pneumologie, Québec City, Québec, Canada. Two centers (Paris and Quebec) had previously published CT-AVC thresholds,3 but here provided separate distinct populations of patients that did not overlap with their original cohort. The remaining 5 centers contributed data from patients in Europe and North America who were being considered for transcatheter aortic valve implantation and were undergoing CT scans as part of their work up (Table I in the Data Supplement). The University of Edinburgh coordinated the study. Informed consent was obtained for those patients who had taken part in research studies. In patients who had already undergone scans for clinical indications, patients were deidentified and anonymized. This study design was approved by the following institutional review boards: South East Scotland Research Ethics Committee and the Health Research Authority of the United Kingdom. Research was performed in accordance with the Declaration of Helsinki.

    Computed Tomography Aortic Valve Calcium Scoring

    All centers performed noncontrast CT scans gated from 75% to 80% of the R–R interval, with a tube current of 42 to 1312 A and a voltage of 120 kV. Imaging was performed on a range of different scanners (Table I in the Data Supplement). At the discretion of the attending clinician, some centers administered β-blockade to achieve a resting heart rate of ≤65 beats per minute (Amiens, France, and Edinburgh, United Kingdom).

    Image analysis was performed locally using a range of different software packages (Table I in the Data Supplement). At the initiation of the study, consensus was reached on the optimum method for calcium scoring, and this was then applied at each of the centers, ensuring consistency of approach. In brief, CT-AVC scores were quantified on contiguous 3-mm axial slices commencing at the base of the valve, with care taken to exclude calcium originating from extravalvular structures, such as the mitral valve annulus, the ascending aorta, and coronary arteries (Figure I in the Data Supplement). The total AVC in AU was calculated and subsequently indexed to the body surface area (AU/m2) or divided by the echocardiographic measurement of the left ventricular outflow tract area to estimate calcium density (AU/cm2).16

    Echocardiography

    Echocardiography was performed according to the European Association of Cardiovascular Imaging and American Society of Echocardiography guidelines with measurements of the Vmax (m/s), mean gradient (mm Hg), and AVA (cm2) made for each patient.2 AVA was calculated using the continuity equation and then indexed to body surface area (cm/m2). The aortic velocity time integral was multiplied by the left ventricular outflow tract area (derived from the left ventricular outflow tract diameter) and indexed to body surface area providing the stroke volume index (mL/m2). Where possible, left ventricular ejection fraction was calculated using Simpson biplane method. Disease severity was assessed according to latest the recommendations of the European Association of Cardiovascular Imaging and the American Society of Echocardiography using Vmax, AVA, stroke volume index, and left ventricular ejection fraction.2

    Echocardiographic data were defined as concordant when findings from both the Vmax and AVA provided the same assessment of disease severity (severe AS: AVA <1.0 cm2 Vmax ≥4.0 m/s; nonsevere AS: AVA ≥1.0 cm2 Vmax <4.0 m/s). Patients in whom the Vmax and AVA gave discordant assessments of AS severity were subcategorized as having normal flow or low flow. Discordant low-flow patients were further subdivided into those with a low (<50%, classical low-flow severe AS) or preserved (≥50%, paradoxical low-flow severe AS) ejection fraction.2 Discordant normal-flow patients were further subdivided into those with Vmax ≥4.0 m/s (and AVA ≥1.0 cm2) versus patients with Vmax <4.0 m/s (and AVA <1.0 cm2).17 The sex-specific CT-AVC thresholds were then applied to these discordant patients to arbitrate whether they had severe or nonsevere disease (Figure II in the Data Supplement).

    Prediction of Disease Progression and Adverse Outcomes

    The primary outcome was the time to first event of death or AVR post-CT calcium scoring. AVR included both open surgical procedures and transcatheter AVR. Decisions about whether to proceed to AVR were made according to international clinical guidelines,2,18 independent of CT-AVC and after multidisciplinary discussion. Patients in whom a decision to refer for AVR had already been made at the time of CT-AVC or who had CT imaging performed as part of the work up before transcatheter AVR or surgery were excluded from outcomes analysis.

    Statistical Analysis

    Continuous variables were expressed as mean±SD or median (interquartile range) as appropriate. CT-AVC data underwent square root transformation to achieve normality (√AU). Categorical data were presented as number and percentage. Correlations between continuous variables were assessed with linear regression analysis and either Pearson r or Spearman ρ. Parametric (unpaired Student t test) and nonparametric (Mann–Whitney U) tests were used as appropriate. One-way ANOVA was used to test for significant differences within multiple independent groups. In patients with concordant echocardiographic data, receiver operator curves were derived to assess CT-AVC thresholds and to identify the optimum thresholds for severe AS. Kaplan–Meier curves and Cox proportional hazards regression analyses were used to determine the ability of CT-AVC to predict adverse clinical events. Where appropriate, collinearity of variables was assessed before inclusion in the multivariable model. Two-sided significance was taken as P<0.05.

    Results

    Data were collated from 918 patients across 8 international centers (age, 77±10 years; 60% men; 6% bicuspid; Vmax, 3.88±0.90 m/s; Table 1) with 431 patients undergoing imaging within prospective clinical research studies and 487 patients being imaged as part of routine clinical care. The latter group included 366 patients being considered for transcatheter AVR who were excluded from outcome analyses.

    Table 1. Patient Characteristics

    AllMenWomenConcordantDiscordant Low FlowDiscordant Normal Flow
    SevereNonsevereEF<50%EF≥50%Vmax≥ 4 m/s AVA≥1 cm2Vmax<4 m/s AVA<1 cm2
    Clinical91854837043627216593596
     Agey±SD77±1076±9.678±1080±972±1078±1077±1174±1179±8
     Males%601000517869578052
     Heightcm±SD165±12.0170±8.0159±12164±13169±8163±7167±9171±8163±9
     Weightcm±SD78±1782±1671±1675±1782±1769±1284±1784±1772±16
     Body surface aream2±SD1.88±0.251.95±0.231.76±0.241.83±0.241.94±0.251.76±0.191.96±0.231.99±0.231.79±0.23
     Body mass indexkg/m2±SD28±629±528±628±629±526±430±529±527±5
     Systolic blood pressuremm Hg±SD136±20136±20136±20134±19139±21128±21139±19137±18137±20
     Diastolic blood pressuremm Hg±SD72±1271±1271±1371±1274±1374±1475±1071±1071±11
     Heart ratebpm±SD69±1361±1371±1270±1267±1385±2273±1568±1267±10
     Possible symptoms%7065769147100756958
     Hypertension%777776778263818375
     Coronary artery disease%455137493750503444
     Ever smoked%324317244244503439
     Diabetes mellitus%282925302431282026
     Hyperlipidemia%656763645963685468
     Scan intervald [IQR]5 [1–25]4 [0–25]6 [1–26]7 [1–25]1 [0–10]8 [2–31]7 [2–26]4 [1–27]3 [0–23]
    Echocardiography
     Peak aortic jet velocitymm Hg±SD3.88±0.903.75±0.864.07±0.904.61±0.502.9±0.503.21±0.343.50±0.364.30±0.333.51±0.33
     Percentage ≥4 m/s%5145601000001000
     Mean gradientm/s±SD38±1936±1843±1954±1419±726±830±744±830±7
     Percentage ≥40 mm Hg%484258915.3012698
     Aortic valve areacm2±SD0.90±0.350.99±0.380.78±0.290.66±0.151.33±0.280.71±0.190.72±0.151.16±0.170.83±0.11
     Percentage ≤1.0 cm2%67588010001001000100
     Aortic valve area indexcm2/m2±SD0.48±0.180.50±0.190.44±0.170.36±0.080.69±0.150.38±0.100.37±0.080.59±0.130.47±0.07
     Percentage ≤0.6 cm2%77728399010010066100
     Bicuspid%78541203234
     LVOT diametercm±SD2.14±0.222.20±0.222.03±0.182.09±0.202.24±0.212.07±0.212.05±0.172.31±0.252.08±0.20
     LVOT areacm2±SD3.60±0.763.86±0.873.27±0.573.45±0.663.98±0.753.39±0.733.32±0.574.22±0.933.43±0.67
     LVOT VTIcm22.1±5.222.0±5.022.3±5.422±5.0423±514±2.9017±3.4729±7.622±4
     Indexed stroke volumemL/m2±SD42±1143±1141±1141±1046±925±529±560±1741±4
     Valsalva diametercm±SD3.32±0.463.43±0.443.15±0.453.29±0.433.40±0.523.13±0.683.25±0.453.61±0.413.21±0.48
     Tubular diametercm±SD3.05±0.573.13±0.552.94±0.573.02±0.553.2±0.562.74±0.362.81±0.593.30±0.482.91±0.60
     Ejection fraction%±SD61±8.560±961±861±860±938±858±861±1061±8
    Computed tomography
     Aortic valve calcium scoreAU [IQR]2055 [1054–3339]2108 [1125–3459]1988 [947–3077]2951 [2081–4144]895 [528–1577]1776 [1089–2590]2056 [1152–2786]2917 [1912–3752]1310 [763–2068]
     Aortic valve calcium indexAU/m2 [IQR]1088 [557–1810]1079 [557–1798]1105 [552–1835]1612 [1133–2330]468 [276–810]955 [546–1385]1000 [630–1409]1387 [900–1852]764 [481–1140]
     Aortic valve calcium densityAU/cm2 [IQR]580 [284–940]556 [275–930]616 [301–957]860 [599–1220]232 [133–282]526 [324–656]615 [406–817]690 [473–903]412 [236–630]
     Aortic valve calcium volumemm3 [IQR]1158 [594–2189]1319 [716–2311]950 [500–2061]1718 [938–2748]650 [356–1187]1232 [605–2065]1367 [609–1823]1069 [623–1811779 [450–1347]

    AU indicates Agatston unit; AVA, aortic valve area; AVC, aortic valve calcium score; EF, ejection fraction; IQR, interquartile range; LVOT, left ventricular outflow tract; and Vmax, peak velocity.

    CT-AVC correlated with the different echocardiographic measures of AS severity (Table II in the Data Supplement). When subdivided by sex, women had lower CT-AVC to achieve the same degree of hemodynamic obstruction than men even after indexing for body surface area or the left ventricular outflow tract area (AVC calcium density).

    Patients With Concordant Echocardiography

    Performance of CT-AVC Thresholds

    Overall, 708 (77%) patients had concordant echocardiographic assessments of their disease severity (defined using the Vmax and AVA). Patients with concordant-severe disease (n=437) had AVC scores that were more than double patients with concordant nonsevere disease (n=272; P<0.001). In these concordant patients, we determined that the optimum CT-AVC thresholds for severe AS were 1377 AU for women and 2062 AU for men (Figure 1). These thresholds had a sensitivity of 87% and specificity of 84% in women and a sensitivity of 80% and a specificity of 82% in men. This provided a high degree of discrimination for identifying severe AS (Figure 2). Overall the C-statistic for CT-AVC in thediagnosis of severe AS were 0.92 in women and 0.89 in men (Table 2). CT-AVC performed similarly well when echocardiographic disease severity was defined using mean gradient ≥40 mm Hg and AVA index ≤0.6 cm2 (C statistic: 0.93 women, 0.92 men; Table 3). Although the CT-AVC density also performed well, our optimum value in women (420 AU/cm2) was different to that previously proposed (292 AU/cm2). On this basis, the subsequent outcomes analyses were performed using values for CT-AVC only.

    Table 2. Computed Tomography Aortic Valve Calcium Scoring (CT-AVC) Thresholds for Severe Aortic Stenosis in Patients With Concordant Echocardiographic Measures

    ThresholdSensitivity (%)Specificity (%)
    Aortic valve calcium (AU)
     Women
      AUC=0.92Clavel et al312748981
    Optimum13778784
    Most specific26465196
    Most sensitive7749562
     Men
      AUC=0.89Clavel et al320658082
    Optimum20628082
    Most specific39053695
    Most sensitive11969558
    Aortic valve calcium index, AU/m2
     Women
      AUC=0.92Clavel et al36379277
    Optimum7848883
    Most specific15124995
    Most sensitive5009568
     Men
      AUC=0.90Clavel et al310678185
    Optimum10588184
    Most specific18144595
    Most sensitive6599569
    Aortic valve calcium density, AU/cm2
     Women
      AUC=0.93Clavel et al32929675
    Optimum4208888
    Most specific7066196
    Most sensitive3079577
     Men
      AUC=0.91Clavel et al34768881
    Optimum5278384
    Most specific9095295
    Most sensitive3769573

    The peak velocity and aortic valve area were used to adjudicate stenosis severity. Clavel et al refers to thresholds proposed in reference 3. AU indicates Agatston unit; and AUC, area under curve.

    Table 3. Computed Tomography Aortic Valve Calcium Scoring (CT-AVC) Thresholds for Severe Aortic Stenosis in Patients With Concordant Echocardiographic Measures

    ThresholdSensitivity (%)Specificity (%)
    Aortic valve calcium, AU
     Women
      AUC=0.92Clavel et al312748981
    Optimum13778883
     Men
      AUC=0.89Clavel et al320658187
    Optimum20628187
    Aortic valve calcium index, AU/m2
     Women
      AUC=0.92Clavel et al36379376
    Optimum7848980
     Men
      AUC=0.90Clavel et al310678288
    Optimum10588388
    Aortic valve calcium density, AU/cm2
     Women
      AUC=0.93Clavel et al32929677
    Optimum4209088
     Men
      AUC=0.91Clavel et al34769088
    Optimum5278589

    The mean gradient and indexed aortic valve area were used to adjudicate stenosis severity. Clavel et al refers to thresholds proposed in reference3. AU indicates Agatston unit; and AUC, area under curve.

    Figure 1.

    Figure 1. Receiver-operator characteristic curves for women (A) and men (B) assessing the accuracy of computed tomography aortic valve calcium scoring against echocardiography in patients with concordant echocardiographic measures (n=708). The dashed lines represent 95% sensitivity (horizontal) and 95% specificity (vertical). AUC indicates area under the curve.

    Figure 2.

    Figure 2. Computed tomography aortic valve calcium scores (CT-AVC) in patients with concordant echocardiographic measurements. To stratify for sex, CT-AVC scores were divided by the respective sex-specific thresholds for severe aortic stenosis (1377 Agatston unit [AU] in women and 2062 AU in men). A score above the dotted line represents CT-AVC above the sex-specific threshold and therefore severe aortic valve calcification. A CT-AVC score below the dotted line represents CT-AVC below the sex-specific threshold and nonsevere calcification. Patients with concordant-severe aortic stenosis, as defined by echocardiography, had CT-AVC ratios that were more than double those with concordant nonsevere disease. Box and whiskers plot, error bars are from the 10th to 90th centile and the horizontal line represents the median value.

    Prediction of Adverse Outcomes

    Clinical outcome data were available in 215 (23%) patients after a median of 1029 (126–2251) days. Compared with the cohort as whole, these patients were more likely to be men and were younger (73±9 versus 79±10 years). They had less severe AS as measured using peak aortic jet velocity (3.4 versus 4.0 m/s), AVA (1.10 versus 0.83 cm2), or aortic valve calcium score (median 2295 versus 1366 AU; all P<0.001). In total, 79 patients were adjudicated to have reached the primary clinical end point: with 59 having undergone AVR and 20 patients having died. Collinearity (variance inflation factors <2) did not preclude outcomes analysis. On univariate analysis, when examined as a continuous variable, the CT-AVC (√AU) predicted adverse events (hazard ratio [HR] per square root increase, 1.03 [95% confidence intervals (CI), 1.02–1.05]; P<0.001). Moreover CT-AVC remained an independent predictor of clinical events after adjustment for traditional predictors age, sex, Vmax, and AVA (HR per 1-unit square root increase, 1.04 [95% CI, 1.02–1.06]; P<0.001).

    Similar results were obtained when CT-AVC was considered as a categorical variable (severe versus nonsevere). On univariate analysis, severe CT-AVC was associated with an increase in the likelihood of AVR and death (HR, 3.6 [95% CI, 2.20– 5.74]; P<0.001). In addition, severe CT-AVC was the only independent predictor of AVR and death on multivariable analysis after adjustment for age, sex, Vmax ≥4 m/s, and AVA <1 cm2 (HR, 3.90 [95% CI, 2.19–6.78]; P<0.001; Figure 3).

    Figure 3.

    Figure 3. Sex-specific thresholds (1377 Agatston unit [AU] in women and 2062 AU in men) for severe calcification using computed tomography aortic valve calcification (CT-AVC) predict death and aortic valve replacement in patients with aortic stenosis. A, Kaplan–Meier Curves demonstrating event-free survival using the sex-specific thresholds for severe CT-AVC. Severe calcification was associated with an adverse prognosis (log-rank P=0.002) compared with patients with nonsevere calcification. B, Forrest plot for multivariable analysis. Sex-specific CT-AVC thresholds emerged as the only independent predictor of aortic valve replacement and death. Patients with a severe CT-AVC had a 3- to 4-fold increase in these events. AVA indicates aortic valve area; CI, confidence interval; HR, hazard ratio; and Vmax, peak aortic jet velocity.

    Patients With Discordant Echocardiography

    A total of 210 (23%) patients in this study had discordant echocardiographic assessments of disease severity: 79 with low flow and 131 with normal flow (Table 1; Figure 4; Figure II in the Data Supplement). Considerable heterogeneity in CT-AVC scores was observed in these patients with 102 (49%) having severe calcification and 108 (51%) having nonsevere calcification. This heterogeneity persisted to differing extents in each of the subgroups examined, with overall significantly different CT-AVC scores (P<0.001) between the subgroups (Figure 4). Within the low-flow category, 4 patients did not have an ejection fraction available. Severe calcification was present in 33 patients (56%) with paradoxical low-flow and 8 (50%) with classical low-flow AS (Figure 4). Among discordant normal-flow patients, severe calcification was observed in 26 (74%) subjects with a high Vmax >4.0 m/s (AVA >1.0 cm2) and 33 (34%) patients with a low Vmax <4.0 m/s (AVA <1.0 cm2).

    Figure 4.

    Figure 4. Computed tomography aortic valve calcium scores (CT-AVC) in discordant patients. A, To stratify for sex, CT-AVC scores were divided by the respective sex-specific thresholds for severe aortic stenosis (1377 Agatston unit [AU] in women and 2062 AU in men). A score above the dotted line represents CT-AVC above the sex-specific threshold and therefore severe aortic valve calcification. A CT-AVC score below the dotted line represents CT-AVC below the sex-specific threshold and nonsevere calcification. Considerable heterogeneity in disease severity as assessed by CT-AVC was observed in all 4 subgroups of patients with discordant echocardiographic findings. Box and whiskers plot, error bars are from the 10th to 90th centile and the horizontal line represents the median value. B, Kaplan-Meier Curves demonstrating event free survival using sex-specific thresholds for severe CT-AVC amongst patients with discordant echocardiographic measurements. Severe calcification was associated with an adverse prognosis (log-rank P=0.007) compared with patients with nonsevere calcification. AVA indicates aortic valve area; EF, ejection fraction; and Vmax, peak aortic jet velocity.

    Within the discordant subgroup, outcomes were available in 41 patients in whom 17 patients were adjudicated to have reached the primary end point. On univariate analysis, severe CT-AVC was associated with an increase in the likelihood of AVR and death (HR, 3.67 [95% CI, 1.39–9.73]; P=0.010). On multivariable analysis, severe CT-AVC was once more an independent predictor of adverse outcomes (HR, 3.31 [95% CI, 1.10–9.94]; P=0.03) after adjustment for age, sex, and Vmax ≥ 4 m/s.

    Performance of Previously Published CT-AVC Thresholds

    We repeated our analyses using the previously published CT-AVC thresholds, (women, 1274 AU and men, 2065 AU), which were nearly identical to the present thresholds.3 In women, the threshold of 1274 AU had a sensitivity of 89% and specificity 81%. Similarly, in men the threshold of 2065 AU had a sensitivity of 80% and a specificity of 82% (Table 2).

    These thresholds conferred similar risk prediction. On univariate analysis, severe calcification predicted AVR and death (HR, 3.51 [95% CI, 2.18–5.67]; P<0.001) and was the only independent predictor of AVR and death on multivariable analysis after adjustment for age, sex, peak velocity ≥4 m/s, and AVA <1 cm2 (HR, 3.80 [95% CI, 2.16–6.69]; P<0.001).

    When applied to the discordant population, the published thresholds reclassified 4 patients so that numbers with severe calcification within each subgroup were as follows: 34 patients (54%) with paradoxical low flow; 8 (50%) with classical low-flow AS; 27 (77%) with a high Vmax >4.0 m/s (AVA >1.0 cm2); and 35 (36%) patients with a low Vmax <4.0 m/s (AVA <1.0 cm2). With respect to clinical outcomes, severe calcification was again associated with a 3-fold increase in the likelihood of AVR and death on univariate analysis (HR, 3.37 [95% CI, 1.27–8.95]; P=0.015) and multivariable analysis (HR, 3.06 [95% CI, 1.00–9.32]; P=0.05) after adjustment for, age, sex, and Vmax ≥4 m/s.

    Discussion

    In a multicenter international registry, we present the largest study to date simultaneously investigating both CT-AVC and echocardiography in patients with AS. We have demonstrated that sex-specific thresholds for CT-AVC are highly reproducible across different patient populations and demonstrate excellent discrimination for detecting severe AS. Moreover, we have demonstrated for the first time that CT-AVC thresholds are a powerful independent predictor of adverse clinical events, including among the subgroup of patients with discordant echocardiography. Given that CT-AVC is widely available, requires no contrast, and involves low radiation exposure, we think it should be used as a complementary imaging test alongside echocardiography, particularly in the high proportion of AS patients with discordant echocardiographic measurements and uncertain disease severity.

    The identification of severe AS is essential because in those patients with symptoms, this diagnosis often triggers major cardiac surgery and AVR. CT-AVC holds promise as an alternative assessment of disease severity to complement echocardiography with the advantage that it is independent of loading conditions and hemodynamic influences. CT-AVC thresholds for severe AS have recently been proposed from a derivation cohort comprising 451 patients, demonstrating good agreement with echocardiography and the prediction of clinical events in the same population.13,19 However, these thresholds had not previously been validated in an independent multicenter cohort, and it was unclear whether they were reproducible or generalizable in different patient populations or influenced by variations in imaging technology and analysis software. The primary objective of this study was to validate CT-AVC in AS in an independent international multicenter cohort and to investigate its widespread clinical applicability. We included patients from a variety of clinical settings, across 5 countries, imaged on an array of different scanners, and analyzed locally using a spectrum of different analysis software. Moreover, we tested CT-AVC against all 4 of the recommended echocardiographic measurements of disease severity in widespread clinical practice.

    By design, we included patients with a broad range of disease severity, recruited from multiple international centers population, and had imaging on a range of scanners from different vendors. Despite this heterogeneity, our thresholds (women 1377 AU and men 2062 AU) for severe disease were nearly identical to those originally proposed (women 1274 AU and men 2065 AU) and performed similarly well. Our data, therefore, confirm both the reproducibility and generalizability of these CT-AVC thresholds for severe disease and their clinical utility as an alternative assessment of disease severity. Moreover, CT-AVC assessments using the Agatston method is relatively quick, widely available on commercial software and familiar to reporting physicians.

    Of primary importance is the ability of CT-AVC to predict clinical outcomes. Indeed, assessing how quickly patients with AS are likely to proceed to AVR and whether they are at risk of death are 2 major objectives in clinical practice. Echocardiographic estimation of valvular calcification has been overlooked as a measure of disease severity despite conferring powerful risk-prediction6 because of difficulties in assessment and poor reproducibility. CT-AVC facilitates accurate, reproducible calcium quantification,9 and in a large subgroup of our population in whom prospective outcome data were available, we have confirmed the powerful prognostic information that CT-AVC provides independent of standard echocardiographic and clinical variables. Indeed CT-AVC was associated with an ≈4-fold increase in AVR or death and emerged as the sole predictor of these events on multivariable analysis. This is despite the fact that decisions to refer for AVR were based on the echocardiographic assessments not CT-AVC scores, the results of which were unavailable to clinicians.

    How then might CT-AVC be used in clinical practice? Consistent with previous studies, almost a quarter of our patients (n=210; 23%) had discordant echocardiographic measurements. In these patients, the severity of stenosis remained in question, and we explored whether CT-AVC might act as an arbitrator or umpire test. Our results suggest that there is substantial heterogeneity in disease severity in these discordant patients that persisted within each of the different subgroups examined. Given the lack of a gold standard assessment of disease severity in discordant patients, it is important to note that CT-AVC provided similarly powerful and independent prognostic information in these subjects as it did across the wider AS population (HR, 3–4 for AVR or death). This is the first time that this has been demonstrated and highlights the particular value that CT-AVC holds as both an arbitrator of disease severity and prognostic marker in this challenging and common cohort of patients (Figure 5). Indeed we think that CT-AVC should be used as a robust and powerful guide to patient management in patients where diagnostic and prognostic uncertainty exists, either in combination or even in substitution to dobutamine stress echocardiography, which is not always easy to perform or interpret. Indeed, unlike advanced echocardiographic techniques that are more dependent on operator technique and expertise, the simplicity, generalizability, and reproducibility of CT-AVC lends itself to rapid application in the clinic. Importantly, we still think that echocardiography should remain the first-line diagnostic assessment in AS and that CT-AVC should only be used when the echocardiographic measures of severity are discordant.

    Figure 5.

    Figure 5. Patients with aortic stenosis (AS) should initially be assessed with echocardiography. In a quarter of patients, there is discordance in echocardiographic assessments of disease severity, causing diagnostic uncertainty. Computed tomography aortic valve calcium score (CT-AVC) provides an alternative and complementary assessment of disease severity. Sex-specific thresholds (men 2062 Agatston unit [AU], women 1377 AU) are highly reproducible across different patient populations and provide excellent discrimination for severe AS vs echocardiography (C statistic: women 0.92, men 0.89). CT-AVC also provides powerful prognostic information and risk stratification. Patients with severe calcification have a 3- to 4-fold increase in aortic valve replacement and death compared with those with nonsevere calcification (hazards ratio, 3.90 [95% confidence interval, 2.19–6.78]; P<0.001) after adjustment for age, sex, peak aortic jet velocity, and aortic valve area).

    Our study has some limitations. In the prospective clinical studies, patients with a wide range of AS disease severity were recruited; however, those patients undergoing CT scans for clinical indications by definition had more severe disease. Nonetheless, it is in precisely these patients that we anticipate CT-AVC scoring is most likely to be used. It is also important to note that while agreement with echocardiography was generally excellent in a small subgroup of patients, there was clear disagreement between the CT-AVC thresholds and concordant echocardiographic measures of severity. Further work is required to determine the prognostic ramifications of unexpectedly high (or low) calcium burdens in these patients and whether they have evidence of noncalcific valve thickening.

    For the outcomes analysis, we had to exclude patients who were undergoing CT evaluation for aortic valve intervention because they would inevitably have AVR. Nevertheless, we were still able to assess clinical outcomes in patients recruited in to the prospective cohort studies. Indeed, our outcome cohort ultimately comprised 210 patients (considerably larger than for many comparable echocardiographic studies) and for the first time confirmed the prognostic importance of CT-AVC in patients with discordant echocardiography. Finally, we did not undertake core laboratory analyses of the CT-AVC because our intention instead was to assess the generalizability of CT-AVC as performed and analyzed in individual centers. Moreover, despite this lack of centralized standardization, CT-AVC continued to demonstrate highly consistent and reproducible thresholds, thereby underlining the clinical utility of this approach.

    Conclusions

    In this large multicenter international registry of nearly a thousand patients with AS, we have established the excellent reproducibility of sex-specific AVC thresholds and confirmed both their accuracy for severe AS and independent prognostic capability. On this basis, we think that CT-AVC is now ready for widespread clinical use as a complementary imaging test alongside echocardiography in patients with AS. We anticipate that its primary application would be to identify patients with severe AS in whom echocardiography provides conflicting readouts. These findings provide strong support to the integration of CT-AVC in routine practice for the management of AS.

    Footnotes

    *Drs Messika Zeitoun and Dweck contributed equally to this work.

    The Data Supplement is available at http://circimaging.ahajournals.org/lookup/suppl/doi:10.1161/CIRCIMAGING.117.007146/-/DC1.

    Correspondence to Tania Pawade, MD, PhD, University of Edinburgh, Chancellor’s Bldg, 49 Little France Crescent, EH16 4SB, Edinburgh, United Kingdom. E-mail

    References

    • 1. Braunwald E. On the natural history of severe aortic stenosis.J Am Coll Cardiol. 1990; 15:1018–1020.CrossrefMedlineGoogle Scholar
    • 2. Baumgartner H, Hung J, Bermejo J, Chambers JB, Edvardsen T, Goldstein S, Lancellotti P, LeFevre M, Miller F, Otto CM. Recommendations on the echocardiographic assessment of aortic valve stenosis: a focused update from the European Association of Cardiovascular Imaging and the American Society of Echocardiography.J Am Soc Echocardiogr. 2017; 30:372–392. doi: 10.1016/j.echo.2017.02.009.CrossrefMedlineGoogle Scholar
    • 3. Clavel MA, Messika-Zeitoun D, Pibarot P, Aggarwal SR, Malouf J, Araoz PA, Michelena HI, Cueff C, Larose E, Capoulade R, Vahanian A, Enriquez-Sarano M. The complex nature of discordant severe calcified aortic valve disease grading: new insights from combined Doppler echocardiographic and computed tomographic study.J Am Coll Cardiol. 2013; 62:2329–2338. doi: 10.1016/j.jacc.2013.08.1621.CrossrefMedlineGoogle Scholar
    • 4. Dweck MR, Chin C, Newby DE. Small valve area with low-gradient aortic stenosis: beware the hard hearted.J Am Coll Cardiol. 2013; 62:2339–2340. doi: 10.1016/j.jacc.2013.08.1620.CrossrefMedlineGoogle Scholar
    • 5. Chin CW, Pawade TA, Newby DE, Dweck MR. Risk stratification in patients with aortic stenosis using novel imaging approaches.Circ Cardiovasc Imaging. 2015; 8:e003421. doi: 10.1161/CIRCIMAGING.115.003421.LinkGoogle Scholar
    • 6. Rosenhek R, Binder T, Porenta G, Lang I, Christ G, Schemper M, Maurer G, Baumgartner H. Predictors of outcome in severe, asymptomatic aortic stenosis.N Engl J Med. 2000; 343:611–617. doi: 10.1056/NEJM200008313430903.CrossrefMedlineGoogle Scholar
    • 7. Pawade TA, Newby DE, Dweck MR. Calcification in aortic stenosis: the skeleton key.J Am Coll Cardiol. 2015; 66:561–577. doi: 10.1016/j.jacc.2015.05.066.CrossrefMedlineGoogle Scholar
    • 8. Dweck MR, Jones C, Joshi NV, Fletcher AM, Richardson H, White A, Marsden M, Pessotto R, Clark JC, Wallace WA, Salter DM, McKillop G, van Beek EJ, Boon NA, Rudd JH, Newby DE. Assessment of valvular calcification and inflammation by positron emission tomography in patients with aortic stenosis.Circulation. 2012; 125:76–86. doi: 10.1161/CIRCULATIONAHA.111.051052.LinkGoogle Scholar
    • 9. Messika-Zeitoun D, Aubry MC, Detaint D, Bielak LF, Peyser PA, Sheedy PF, Turner ST, Breen JF, Scott C, Tajik AJ, Enriquez-Sarano M. Evaluation and clinical implications of aortic valve calcification measured by electron-beam computed tomography.Circulation. 2004; 110:356–362. doi: 10.1161/01.CIR.0000135469.82545.D0.LinkGoogle Scholar
    • 10. Cowell SJ, Newby DE, Burton J, White A, Northridge DB, Boon NA, Reid J. Aortic valve calcification on computed tomography predicts the severity of aortic stenosis.Clin Radiol. 2003; 58:712–716.CrossrefMedlineGoogle Scholar
    • 11. Koos R, Mahnken AH, Sinha AM, Wildberger JE, Hoffmann R, Kühl HP. Aortic valve calcification as a marker for aortic stenosis severity: assessment on 16-MDCT.AJR Am J Roentgenol. 2004; 183:1813–1818. doi: 10.2214/ajr.183.6.01831813.CrossrefMedlineGoogle Scholar
    • 12. Cueff C, Serfaty JM, Cimadevilla C, Laissy JP, Himbert D, Tubach F, Duval X, Iung B, Enriquez-Sarano M, Vahanian A, Messika-Zeitoun D. Measurement of aortic valve calcification using multislice computed tomography: correlation with haemodynamic severity of aortic stenosis and clinical implication for patients with low ejection fraction.Heart. 2011; 97:721–726. doi: 10.1136/hrt.2010.198853.CrossrefMedlineGoogle Scholar
    • 13. Clavel MA, Pibarot P, Messika-Zeitoun D, Capoulade R, Malouf J, Aggarval S, Araoz PA, Michelena HI, Cueff C, Larose E, Miller JD, Vahanian A, Enriquez-Sarano M. Impact of aortic valve calcification, as measured by MDCT, on survival in patients with aortic stenosis: results of an international registry study.J Am Coll Cardiol. 2014; 64:1202–1213. doi: 10.1016/j.jacc.2014.05.066.CrossrefMedlineGoogle Scholar
    • 14. Glasziou P, Irwig L, Deeks JJ. When should a new test become the current reference standard?Ann Intern Med. 2008; 149:816–822.CrossrefMedlineGoogle Scholar
    • 15. Baumgartner H, Falk V, Bax JJ, De Bonis M, Hamm C, Holm PJ, Iung B, Lancellotti P, Lansac E, Rodriguez Muñoz D, Rosenhek R, Sjögren J, Tornos Mas P, Vahanian A, Walther T, Wendler O, Windecker S, Zamorano JL; ESC Scientific Document Group. 2017 ESC/EACTS Guidelines for the management of valvular heart disease.Eur Heart J. 2017; 38:2739–2791. doi: 10.1093/eurheartj/ehx391.CrossrefMedlineGoogle Scholar
    • 16. Clavel MA, Malouf J, Messika-Zeitoun D, Araoz PA, Michelena HI, Enriquez-Sarano M. Aortic valve area calculation in aortic stenosis by CT and Doppler echocardiography.JACC Cardiovasc Imaging. 2015; 8:248–257. doi: 10.1016/j.jcmg.2015.01.009.CrossrefMedlineGoogle Scholar
    • 17. Baumgartner H, Hung J, Bermejo J, Chambers JB, Evangelista A, Griffin BP, Iung B, Otto CM, Pellikka PA, Quiñones M; American Society of Echocardiography; European Association of Echocardiography. Echocardiographic assessment of valve stenosis: EAE/ASE recommendations for clinical practice.J Am Soc Echocardiogr. 2009; 22:1–23; quiz 101. doi: 10.1016/j.echo.2008.11.029.CrossrefMedlineGoogle Scholar
    • 18. Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP, Guyton RA, O’Gara PT, Ruiz CE, Skubas NJ, Sorajja P, Sundt TM, Thomas JD; American College of Cardiology/American Heart Association Task Force on Practice Guidelines. 2014 AHA/ACC guideline for the management of patients with valvular heart disease: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.J Am Coll Cardiol. 2014; 63:2438–2488. doi: 10.1016/j.jacc.2014.02.537.CrossrefMedlineGoogle Scholar
    • 19. Nguyen V, Cimadevilla C, Estellat C, Codogno I, Huart V, Benessiano J, Duval X, Pibarot P, Clavel MA, Enriquez-Sarano M, Vahanian A, Messika-Zeitoun D. Haemodynamic and anatomic progression of aortic stenosis.Heart. 2015; 101:943–947. doi: 10.1136/heartjnl-2014-307154.CrossrefMedlineGoogle Scholar

    CLINICAL PERSPECTIVE

    Echocardiography is the gold standard investigation for diagnosing and grading aortic stenosis disease severity. However, in ≈25% patients, the individual measures of disease severity are discordant. This creates clinical uncertainty, particularly with respect to the timing of surgery. To help resolve these difficult but common clinical scenarios, computed tomography aortic valve calcium scoring has been proposed as a novel load-independent arbiter of severe aortic stenosis that incorporates sex-specific thresholds (women 1274 Agatston unit and men 2065 Agatston unit). In a large international multicenter cohort, we derived similar sex-specific thresholds (women 1377 Agatston unit and men 2062 Agatston unit) that emerged as the most powerful predictor of adverse clinical events, both in the cohort as a whole and when confined to patients with discordant echocardiographic findings. On this basis, we think that computed tomography aortic valve calcium scoring should be used as an adjunct to echocardiography in determining disease severity and guiding clinical decision making in patients with aortic stenosis, particularly when echocardiographic measurements are discordant.