Author + information
- Seth Uretsky, MD and
- Steven D. Wolff, MD, PhD∗ ()
- ↵∗Carnegie Hill Radiology, 170 East 77th Street, New York, New York 10075
We thank Drs. Naqvi and Gotway for their comments, and offer the following point-by-point response:
1. Our study (1) included 5 (5%) patients with atrial fibrillation at imaging. Agreement between magnetic resonance imaging (MRI) and echocardiography was not adversely affected. Specifically, concordance between MRI and echocardiography was 40% in patients with atrial fibrillation and 37% in patients in sinus rhythm.
2. The “large” SDs of blood pressures merely reflects the range of blood pressures in our patients. A paired Student t test showed no statistically significant difference at the time of echocardiography and MRI. Furthermore, as noted in the Discussion section, the strong correlation shown in Figure 4A would not be possible if variability of afterload was an important consideration.
3. If aortic regurgitation were underestimated by MRI, it would lead to an overestimation of mitral regurgitant volume, not an underestimation.
4. We do not believe our study shows that pre-surgical left ventricular (LV) volumes predict recovery of LV end-diastolic volume (EDV) post-surgery. For example, in Table 4, patient 29 had mild mitral regurgitation (MR) and an end-diastolic diameter (EDD) of 6.2 cm, whereas patient 38 had severe MR and an EDD of 5.3 cm. Our study shows that MR volume (determined as the difference between LV stroke volume and forward flow) strongly correlates with the decrease in LVEDV following surgery (Figure 4A).
5. We do not believe studies showing reverse remodeling following percutaneous repair indicate that echocardiography selects appropriate surgical candidates. The bar graph shown in our Central Illustration confirms that echocardiography can be used to predict the degree of remodeling. However, it does a much poorer job than MRI. As a result, many patients are incorrectly classified as having severe MR, potentially leading to inappropriate surgical intervention (Figure 3).
6. We do not believe any amount of negative remodeling following surgery indicates an appropriate surgical candidate. MRI-derived ventricular volumes increase even in patients with mild MR (2). The fact that MRI detects negative remodeling following intervention does not indicate that the severity of MR is sufficient to warrant intervention.
7. Regarding the quality of the echocardiograms, as we described in our paper, blinded reviewers graded the echocardiograms as being of good to excellent quality. The echocardiograms were comprehensive and allowed an integrated approach for assessing MR severity. In those patients who underwent both transthoracic echocardiography and transesophageal echocardiography, the reviewers had access to both studies when determining MR severity. We believe the echocardiographic qualifications of our coauthors speak for themselves. The moderate amount of interobserver variability between the readers who read the echocardiograms is less than in other studies (3,4).
8. For MRI, end-diastolic and end-systolic ventricular traces included papillary muscles and trabeculations in the blood pool, as is often the case when they are manually drawn.
Please note: Dr. Wolff is the owner of NeoSoft, LLC and NeoCoil, LLC, medical device companies that make software and hardware for use with MRI. Dr. Uretsky has reported that he has no relationships relevant to the contents of this paper to disclose.
- American College of Cardiology Foundation
- Uretsky S.,
- Gillam L.,
- Lang R.,
- et al.
- Biner S.,
- Rafique A.,
- Rafii F.,
- et al.