Assessment of mitral valve stenosis by helical MDCT: comparison with transthoracic doppler echocardiography and cardiac catheterization.

Abstract

We evaluated the precision of helical MDCT for the quantification of mitral valve stenosis (MVS) compared with transthoracic echocardiography (TTE) and cardiac catheterization. A total of 28 patients with MVS of differing severity underwent an ECG-gated contrast-enhanced MDCT scan. The mitral valve area (MVA) was determined planimetrically by MDCT and was compared with Doppler TTE using the pressure half-time method and with cardiac catheterization using the Gorlin formula. Planimetry of the MVA with MDCT was feasible in all cases. The MVA on MDCT (1.88 ± 0.76 cm(2)) was significantly larger than that seen with TTE (1.74 ± 0.75 cm(2); p = 0.039) or cardiac catheterization (1.72 ± 0.67 cm(2); p = 0.037). The correlation between MDCT and TTE (r = 0.90; p < 0.001; limits of agreement, ± 0.65 cm(2)) and that between MDCT and cardiac catheterization (r = 0.86; p < 0.001; limits of agreement, ± 0.76 cm(2)) were good and similar to the correlation between TTE and cardiac catheterization (r = 0.88; p < 0.001; limits of agreement, ± 0.71 cm(2)). The best cutoff level for detecting moderate-to-severe stenosis at MDCT was an MVA of 1.70 cm(2), resulting in a sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of 73%, 88%, 82%, 80%, and 83%, respectively, with two false-positive and three false-negative results. The MVA planimetrically determined by MDCT is systematically larger than those calculated by Doppler TTE and cardiac catheterization. However, because of a good correlation between methods and adjustment for the systematic bias, MDCT may allow reliable quantification of MVS and effectual discrimination among severity grades, although discrepancies between methods remain in individual cases.