Prospective comparison of valve regurgitation quantitation by cardiac magnetic resonance imaging and transthoracic echocardiography.

BACKGROUND Both transthoracic echocardiography (TTE) and cardiac magnetic resonance (CMR) imaging allow quantification of chronic aortic regurgitation (AR) and mitral regurgitation (MR). We hypothesized that CMR measurement of regurgitant volume (RVol) is more reproducible than TTE. METHODS AND RESULTS TTE and CMR performed on the same day in 57 prospectively enrolled adults (31 with AR, 26 with MR) were measured by 2 independent physicians. TTE RVol(AR) was calculated as Doppler left ventricular outflow minus inflow stroke volume. RVol(MR) was calculated by both the proximal isovelocity surface area method and Doppler volume flow at 2 sites. CMR RVol(AR) was calculated by phase-contrast velocity mapping at the aortic sinuses and RVol(MR) as total left ventricular minus forward stroke volume. Intraobserver and interobserver variabilities were similar. For AR, the Bland-Altman mean interobserver difference in RVol was -0.7 mL (95% confidence interval [CI], -5 to 4) for CMR and -9 mL (95% CI, -53 to -36) for TTE. The Pearson correlation was higher (P=0.001) between CMR (0.99) than TTE readers (0.89). For MR, the Bland-Altman mean difference in RVol between observers was -4 mL (95% CI, -21 to 13) for CMR compared with 0.7 mL (95% CI, -30 to 32) for the proximal isovelocity surface area and -10 mL (95% CI, -76 to 56) for TTE volume flow at 2 sites. Correlation was similar for CMR (0.94) versus TTE readers (0.90 for the proximal isovelocity surface area). CONCLUSIONS Compared with TTE, CMR has lower intraobserver and interobserver variabilities for RVol(AR), suggesting CMR may be superior for serial measurements. Although RVol(MR) is similar by TTE and CMR, variability in measured RVol by both approaches suggests that caution is needed in clinical practice.