Reproducibility of Mitral Regurgitation Quantitation
Abstract & Commentary
By Michael H. Crawford, MD
Sources: Biner S, et al. Reproducibility of proximal isovelocity surface area, vena contracta, and regurgitant jet area for assessment of mitral regurgitation severity. JACC. 2010;3:235-243. Grayburn PA and Bhella P. Grading severity of mitral regurgitation by echocardiography: science or art? JACC. 2010;3:244-246.
The American Society of Echocardiography (ASE) has recommended that quantitative techniques be used to assess the severity of mitral regurgitation (MR) by echocardiography. Such measurements have been shown to accurately estimate the severity of MR with good reproducibility in research labs. These investigators, from Cedars Sinai Medical Center in Los Angeles, sought to assess interobserver agreement in the quantitation of MR in a typical clinical setting. The study focused on three measures of MR severity: color flow jet area, vena contracta (VC) width, and proximal isovelocity surface area (PISA) radius, which is the key measurement for the estimation of effective regurgitant orifice area (EROA). Web-based images from 16 consecutive patients were measured by 18 experienced cardiologists in 11 centers worldwide. Institutional variability was assessed by comparing the interobserver variability between those at the same institution to those at other institutions. The studies were all done at one center by the same technologist. Each echocardiogram was categorized as showing severe or non-severe MR for each measurement. Severe MR was defined as VC ≥ 7mm, EROA ≥ 40mm2, or jet area grade of 3 or 4. Interobserver agreement was classified as substantial if ≥ 80%; fair if between 60% and 79%; and poor if below 60%.
Results: Interobserver agreement between the semiquantitative jet area grade and the two echo Doppler measurements were about the same and only fair: Kappa = 0.32 for MR grade; 0.28 for VC; and 0.37 for PISA. Multivariate predictors of substantial agreement (≥ 80%) included the ability to identify a regurgitant orifice to measure VC and a central regurgitant jet for PISA (p = 0.02). The authors concluded that the quantitative measurements of MR severity recommended by the ASE are only modestly reliable, but the identification of a distinct regurgitant orifice area and a central MR jet predict substantial interobserver agreement.
The accurate assessment of MR severity is critical today since asymptomatic patients with severe MR who have a repairable valve are candidates for surgery, according to the latest guidelines. In the Acorn Clinical Trial of corrective mitral valve surgery for heart-failure patients with MR, echo studies that showed significant MR were submitted to a core lab. The core lab found that more than 40% did not have significant MR: 7% had none; 11% were mild; and 23% were moderate. If sites selected for a research study cannot get it right, is there hope for the rest of us? This study explores the reproducibility of the critical quantitative MR parameters in clinical practice at multiple sites, and found similar results. Interestingly, the variability in measurements was the same between individuals at the same institution and in comparison to readers at other institutions. This suggests that even institutional attempts to standardize readings are not always successful.
The ASE recommends an integrated approach where all qualitative and quantitative measures are considered to assess MR severity. This approach has never been fully tested because of the lack of a suitable gold standard, but studies evaluating multiple measures have shown considerable overlap between those ultimately graded as moderate or severe MR. Other measures considered included jet intensity, pulmonary artery pressure, pulmonary venous flow pattern, left atrial size, and left ventricular volume. Other factors to consider are the etiology of MR, valve morphology, jet characteristics (i.e., eccentric), dynamic changes during systole, and the image degradation of stop frame images. This complexity has prompted some to suggest that the grading of MR is more art than science. Although not critically evaluated, others have suggested that if the left atrium and left ventricle are not dilated, then the diagnosis of severe MR should be questioned.
There are some limitations to this study. There were more cardiologists than subjects in the study. A larger study would have been more convincing. Also, only patients graded as moderately severe to severe were considered. Inclusion of patients with mild MR would probably have improved reproducibility. Finally, there was no gold standard, so the accuracy of the severity estimation was not tested. The accompanying editorial concludes that the solution to this problem is improvement in "training, continuing education, and credentialing for sonographers and physicians involved in the increasingly complex patients with MR." Oh come on, can't we invent a new technological solution? MRI, CT, anyone?