Background-Accurate quantification of aortic valve stenosis (AVS) is needed for relevant management decisions. However, transthoracic Doppler echocardiography (TTE) remains inconclusive in a significant number of patients. Previous studies demonstrated the usefulness of phase-contrast cardiovascular magnetic resonance (PC-CMR) in noninvasive AVS evaluation. We hypothesized that semiautomated analysis of aortic hemodynamics from PC-CMR might provide reproducible and accurate evaluation of aortic valve area (AVA), aortic velocities, and gradients in agreement with TTE. Methods and Results-We studied 53 AVS patients (AVA TTE =0.87±0.44 cm 2 ) and 21 controls (AVA TTE =2.96±0.59 cm 2 ) who had TTE and PC-CMR of aortic valve and left ventricular outflow tract on the same day. PC-CMR data analysis included left ventricular outflow tract and aortic valve segmentation, and extraction of velocities, gradients, and flow rates. Three AVA measures were performed: AVA CMR1 based on Hakki formula, AVA CMR2 based on continuity equation, AVA CMR3 simplified continuity equation=left ventricular outflow tract peak flow rate/aortic peak velocity. Our analysis was reproducible, as reflected by low interoperator variability (<4.56±4.40%). Comparison of PC-CMR and TTE aortic peak velocities and mean gradients resulted in good agreement (r=0.92 with mean bias=−29±62 cm/s and r=0.86 with mean bias=−12±15 mm Hg, respectively). Although good agreement was found between TTE and continuity equationbased CMR-AVA (r>0.94 and mean bias=−0.01±0.38 cm 2 for AVA CMR2 , −0.09±0.28 cm 2 for AVA CMR3 ), AVA CMR1 values were lower than AVA TTE especially for higher AVA (mean bias=−0.45±0.52 cm 2 ). Besides, ability of PC-CMR to detect severe AVS, defined by TTE, provided the best results for continuity equation-based methods (accuracy >94%).
Conclusions-Our
Defrance et al Phase-Contrast Evaluation of Aortic Valve Stenosis 605on phase-contrast (PC) imaging demonstrated the ability of velocity-encoded sequences to characterize AVS by evaluating: peak transvalvular velocity, 12 AVA using the continuity equation, [5][6][7][8][9][10] or Hakki formula 7,11 derived from the Gorlin equation. The evaluation of AVA and hemodynamic parameters in the aforementioned PC-CMR studies was mostly based on manual delineation [5][6][7][8]10,11 of the aortic valve and the LVOT, which is subjective and time-consuming, rendering the usefulness of CMR in clinical routine for AVS evaluation limited. In addition, among these studies, some used the estimation of LVOT area from its diameter. [5][6][7][8] Although this strategy complies with TTE measurements, it does not take full advantage of PC-CMR data, which when combined with an accurate segmentation enable a direct estimation of the LVOT stroke volume (SV LVOT ) while taking into account the shape of the outflow tract.
Clinical Perspective on p 612Accordingly, our first objective was to design a semiautomated analysis of LVOT and aortic valve PC-CMR data, which would enable an accurate evaluation of AVA, mean and maximal...