How accurate is cardiac magnetic resonance in quantitation of primary mitral regurgitation?
Abstract Background Since evaluation of mitral regurgitation (MR) in patients with mitral prolapse by echo may be challenging, cardiac MRI (CMR) has been used for identification of patients with severe valve disease. Little is known about potential factors influencing accuracy of this approach. Purpose To study the impact of left ventricular (LV) contouring methods and interobserver variability on quantitation of MR severity. Methods A total of 50 asymptomatic patients with echocardiographically moderate or severe MR were included to LV-REGURGE study and evaluated with CMR by two independent reviewers. LV volumes were calculated on SSFP short axis stack including or excluding intracavitary myocardium into/from the LV volume (“including” or “excluding” method, respectively). Aortic flow volumes were obtained using phase contrast method. Mitral regurgitant volume (RegVol) was calculated by subtracting forward aortic volume from LV stroke volume. Severe MR was diagnosed when RegVol >60mL and/or RegVol/LV stroke volume (RF) >40%. Agreement between MR grading into severe and non-severe, obtained by different tracing methods and different reviewers was assessed using kappa coefficient. Results Left ventricular and mitral regurgitant volumes obtained by “including” method were significantly larger as compared with “excluding” method, see table. “Including” method identified severe MR in 68% while “excluding” method in 57% of cases. Overall agreement between the methods was present in 84% of cases, meaning a substantial rate of agreement (κ 0.67). Within respective methods, interobserver agreement was present in 77% of cases, interpreted as a moderate rate of agreement (κ 0.48–0.54) Conclusion “Including” trabeculae and papillary muscles into the LV cavity volume substantially increases rate of recognizing severe MR, due to larger LV volumes, particularly LV ESV, comparing with the “excluding” method. Moreover, inter-observer variability in similar extent as tracing method seems to impact differentiation between severe and non-severe MR. Funding Acknowledgement Type of funding source: None