Evaluation of aortic stenosis by cardiovascular magnetic resonance imaging: comparison with established routine clinical techniques.

OBJECTIVE To evaluate whether direct planimetry of aortic valve area (AVA) by cardiac magnetic resonance (CMR) imaging is a reliable tool for determining the severity of aortic stenosis compared with transthoracic echocardiography (TTE), transoesophageal echocardiography (TOE), and cardiac catheterisation. METHODS 44 symptomatic patients with severe aortic stenosis were studied. By cardiac catheterisation AVA was calculated by the Gorlin equation. AVA was measured with CMR from steady state free precession (true fast imaging with steady state precession) by planimetry. AVA was also determined from TOE images by planimetry and from TTE images by the continuity equation. RESULTS Bland-Altman analysis evaluating intraobserver and interobserver variability showed a very small bias for both (-0.016 and 0.019, respectively; n = 20). Bias and limits of agreement between CMR and TTE were 0.05 (-0.35, 0.44) cm2 (n = 37), between CMR and TOE 0.02 (-0.39, 0.42) cm2 (n = 32), and between CMR and cardiac catheterisation 0.09 (-0.30, 0.47) cm2 (n = 36). The sensitivity and specificity of CMR to detect AVA < or = 0.80 cm2 measured by cardiac catheterisation was 78% and 89%, of TOE 70% and 70%, and of TTE 74% and 67%, respectively. CONCLUSION CMR planimetry is highly reliable and reproducible. Further, CMR planimetry had the best sensitivity and specificity of all non-invasive methods for detecting severe aortic stenosis in comparison with cardiac catheterisation. Therefore, CMR planimetry of AVA with steady state free precession is a new powerful diagnostic tool, particularly for patients with uncertain or discrepant findings by other modalities.