Author + information
- Received May 20, 1999
- Revision received January 17, 2000
- Accepted March 2, 2000
- Published online October 1, 2000.
- Thomas M Binder, MDa,b,* (, )
- Raphael Rosenhek, MDa,b,
- Gerold Porenta, MD, PhDa,b,
- Gerald Maurer, MD, FACCa,b and
- Helmut Baumgartner, MD, FACCa,b
- ↵*Reprint requests and correspondence: Dr. Thomas Binder, Dept. of Cardiology, University of Vienna, Währingergürtel 18- 20, A-1090 Vienna, Austria
This study was performed to determine the feasibility, accuracy and reproducibility of real-time volumetric three-dimensional echocardiography (3-D echo) for the estimation of mitral valve area in patients with mitral valve stenosis.
Planimetry of the mitral valve area (MVA) by two-dimensional echocardiography (2-D echo) requires a favorable parasternal acoustic window and depends on operator skill. Transthoracic volumetric 3-D echo allows reconstruction of multiple 2-D planes in any desired orientation and is not limited to parasternal acquisition, and could thus enhance the accuracy and feasibility of calculating MVA.
In 48 patients with mitral stenosis (40 women; mean age 61 ± 13 years) MVA was determined by planimetry using volumetric 3-D echo and compared with measurements obtained by 2-D echo and Doppler pressure half-time (PHT). All measurements were performed by two independent observers. Volumetric data were acquired from an apical view.
Although 2-D echo allowed planimetry of the mitral valve in 43 of 48 patients (89%), calculation of the MVA was possible in all patients when 3-D echo was used. Mitral valve area by 3-D echo correlated well with MVA by 2-D echo (r = 0.93, mean difference, 0.09 ± 0.14 cm2) and by PHT (r = 0.87, mean difference, 0.16 ± 0.19 cm2). Interobserver variability was significantly less for 3-D echo than for 2-D echo (SD 0.08cm2 versus SD 0.23cm2, p < 0.001). Furthermore, it was much easier and faster to define the image plane with the smallest orifice area when 3-D echo was used.
Transthoracic real-time volumetric 3-D echo provides accurate and highly reproducible measurements of mitral valve area and can easily be performed from an apical approach.
Assessment of the severity of mitral valve stenosis requires accurate measurements of the mitral valve orifice area (MVA). Direct measurement of the MVA can be performed by planimetry using two-dimensional echocardiography (2-D echo) (1,2). Mitral valve area determined by planimetry reflects the anatomic orifice area and is largely independent of hemodynamic variables, left ventricular compliance and concomitant valvular disease. Thus, this technique should be superior to indirect methods of determining MVA, such as cardiac catheterization applying the Gorlin equation, Doppler calculations based on the pressure half-time (PHT) or the proximal isovelocity surface area (3–9). However, planimetry by 2-D echo requires significant experience and operator skill to define the correct image plane that displays the true mitral valve orifice. In addition, planimetry requires a parasternal short axis view of the mitral valve and is therefore limited to patients with favorable image quality from a parasternal window (10).
Real-time volumetric three-dimensional echocardiography (3-D echo) is a novel imaging technique that could enhance the ability to perform planimetry of the mitral valve. Three-dimensional echo utilizes a matrix array echo probe to scan a pyramidal volume in real time (11,12). This allows instant acquisition of a complete 3-D data set without complex post-processing procedures. In addition, real-time volumetric echo allows the reconstruction of free definable 2-D image planes from a single volume set independently of the acquisition window. Therefore, real-time volumetric echo could be used as a practical and accurate method for planimetry of mitral valve areas.
This study was performed to evaluate the feasibility, reproducibility and accuracy of 3-D echo for the assessment of MVA.
We studied 48 consecutive patients (40 women, age 61 ± 13 years) referred to our outpatient clinic for evaluation of rheumatic mitral valve disease. Twenty-seven patients were in sinus rhythm and 21 in atrial fibrillation. Mitral stenosis was the predominant valvular lesion in all patients. Concomitant mild or moderate mitral regurgitation was present in 27 patients, and 6 patients had severe mitral regurgitation. Mild or moderate aortic stenosis (mean pressure gradient <20 mm Hg) was present in six patients. Severe aortic stenosis was present in one patient. Mild or moderate aortic regurgitation was present in 19 patients. None of the patients had severe aortic regurgitation. Surgical mitral valvulotomy had been performed in three patients and balloon valvuloplasty more than two years prior to the study in four patients. Mitral valve morphology was classified using the scoring proposed by Wilkins et al. (13). The mean score was 7.8 ± 2.1 (mitral valve mobility 1.8 ± 0.7, thickening 2.4 ± 0.6, subvalvular apparatus 1.9 ± 0.7 and calcification 1.7 ± 0.8).
Two-dimensional echo and continuous-wave Doppler were performed with a Sonos 5500 scanner using a 1.8–4.2 MHz broad-bandwidth phased array transducer (Hewlett-Packard, Andover, Massachusetts).
Short axis views of the mitral valve were obtained from the left parasternal window. Careful scanning from the left atrium towards the left ventricle and vice versa identified the smallest orifice of the mitral valve. Particular care was taken to optimize gain settings. Two independent experienced investigators performed planimetry of the mitral valve area by manual tracing of the contours using the diastolic frame with the largest opening amplitude of the mitral valve. The average of six consecutive measurements performed by both investigators was used.
Continuous-wave Doppler recordings of transmitral blood flow velocities were obtained in the apical four-chamber view. Mitral valve area was determined by dividing 220 by the PHT. Again, six consecutive measurements were averaged. Patients presenting with suspected diastolic dysfunction as defined by an E/A ratio of <1 or >2 were excluded from the analysis because it has been shown that the PHT is not reliable in these patients (14).
Real-time volumetric 3-D echocardiography
Real-time volumetric echocardiography was performed within 24 h of the 2-D echo study. A novel phased-array real-time volumetric 3-D imaging system was used. The system has previously been described in detail (11,12). In brief, the system uses a matrix phased-array transducer (2.5 MHz) with 16:1 (receive/transmit) parallel processing to scan a 65° pyramidal volume at a rate of 18–34 volumes/s. Images are displayed as two simultaneous intersecting orthogonal long-axis scans (B-mode scans) and two perpendicular short-axis scans (C-mode scans, Fig. 1). These C-mode scans can be steered through the 3-D space in increments of 1 mm depth and 1° angulations to generate views that cannot be obtained by conventional echo. Thus, C-mode scans allow the display of short-axis views of the mitral valve from an apical transducer position.
Image acquisition was performed from the apical window. Gain, time gain control and depth settings were adjusted to provide optimal images of the mitral valve. One complete cardiac cycle was stored in memory and on magneto-optical disc. Analysis was performed off-line. Planimetry was performed in short-axis views of the mitral valve using C-mode planes. The two B-mode sectors served as reference images to define the optimal C plane through the mitral orifice. The smallest orifice of the mitral valve at the greatest opening of the mitral valve during diastole was chosen for planimetry.
The slice thickness of the C plane was reduced to the smallest possible value, which still permitted visualization of the entire mitral valve circumference (slice thickness 0.6–1.2 mm).
Analysis of the 3-D data set was performed by two independent investigators without knowledge of the results of 2-D echo. The average value of six consecutive measurements was used. A typical image of the mitral valve orifice as imaged in the same patient by 2-D echo and 3-D echo is shown in Figure 2.
Influence of image plane variations on MVA measurements
The effect of inappropriate alignment of the image plane through the mitral valve on measurements of MVA was analyzed in two patients. Because one has to assume that this error is influenced by the geometry of the stenotic valve, we chose two distinct morphologies of mitral stenosis, which represent extremes (Fig. 3). One valve was characterized by prominent doming of the anterior mitral valve (type I) whereas the other had a funnel shape (type II). After definition of the optimal imaging plane (C plane) by real-time volumetric scanning and measurement of the reference MVA, x and y axis were varied in 3° increments over a range of −6° to +6° from the optimal plane. Thus, 24 image planes were obtained from which MVA were calculated. Afterwards, the depth where C planes were reconstructed was increased by 1 and 2 mm and x/y axis variations were repeated for each depth (Fig. 4).
Beat-to-beat variation of MVA
Beat-to-beat variation of the MVA was calculated in 15 consecutive patients. A 3-D data set of six beats was recorded from the apical position in each patient. For each beat, MVA was obtained by averaging six measurements. Beat-to-beat variation was analyzed by calculating the coefficient of variation and was compared with the intraobserver variability as calculated from six MVA measurements of one single beat by one observer.
All values are expressed as mean ± SD. Mitral valve area measured with the different methods were compared by analysis of variance. Significant differences between groups were assessed by two-tailed paired Student t test. Differences were considered statistically significant at p values <0.05. The correlation between MVA values obtained with the different methods was assessed by linear regression analysis. Agreement between MVA measurements was assessed as described by Bland and Altman (15).
The coefficient of variation was calculated for each individual patient to determine the beat-to-beat variation of the MVA and was defined as the standard deviation of the MVA of six beats divided by the mean value. From the results for individual patients the average coefficient of variation ± SD was obtained for the entire patient group. The coefficient of variation was also calculated for the RR intervals in each patient (mean of six beats) and for the entire patient group.
Using 2-D echo, planimetry was possible in 28 of 32 patients (87%). In two patients image quality was poor and the mitral valve orifice could not be properly visualized, precluding MVA planimetry. In contrast, real-time volumetric 3-D echo allowed the calculation of MVA in all patients. Acquisition of one volumetric data set and offline analysis, including definition of the image plane for planimetry, required approximately 5 min only.
Comparison of 2-D and 3-D echo
Regression analysis demonstrated a good correlation between MVA as obtained by 2-D echo and 3-D echo (r = 0.93; y = 0.90x + 0.21). However, MVAs obtained by 3-D echo were on average smaller than those determined by 2-D echo (1.19 ± 0.36 cm2 and 1.28 ± 0.35 cm2, respectively). This difference was statistically significant (p < 0.05; Fig. 5A). The corresponding Bland-Altman plot showing a mean difference of 0.09 ± 0.14 cm2 between the two methods is shown in Figure 5B. The agreement between 2-D and 3-D echo for patients with a high Block score (≥8) and the agreement in those with a low score was very similar (<8; 0.13 ± 0.11 cm2 and 0.14 ± 0.12 cm2, respectively).
Regression analysis showing the correlation between MVA obtained by two independent observers for both 2-D echo and 3-D echo is depicted in Figure 6. The correlation between the two observers was higher for 3-D echo (r = 0.97) than for 2-D echo (r = 0.83). The corresponding Bland-Altman plots (Fig. 6) demonstrate on average a good agreement between the two observers for both methods. However, the scatter was significantly less for 3-D echo than for 2-D echo (mean difference, 0.03 ± 0.08 cm2 and −0.06 ± 0.23 cm2, respectively; p < 0.0001).
Comparison of 2-D and 3-D planimetry with the PHT method
Diastolic dysfunction was suspected (E/A ratio >1 or >2) in 5 of the 43 patients (12%). These patients were excluded from further analysis. Regression analysis demonstrates a fair correlation between MVA by PHT and planimetry by both 2-D echo (r = 0.86, y = 0.82x + 0.17) and 3-D echo (r = 0.87, y = 0.87x + 0.02). Bland-Altman analysis demonstrated a good agreement between the PHT method and both 2-D echo (mean difference, 0.08 ± 0.21 cm2) and 3-D echo (mean difference, 0.15 ± 0.20 cm2). No statistical difference was found for the agreement between PHT method and 2-D echo and the agreement between PHT method and 3-D echo.
Influence of image plane variations on MVA measurements
The influence of minor image plane variations on MVA planimetry was calculated in two patients with different morphologies of mitral stenosis (Fig. 3).
An MVA of 0.8 cm2 was found for the type I stenosis. Deviations of 6° from the optimal image plane in both the x- and y-axes resulted in a maximal relative measurement error of 63% (1.3 cm2). When the depth where the C plane was reconstructed was increased by 2 mm in relation to the optimal C plane, measurements of MVA showed a maximal relative error of 25% (1.0 cm2). When this change in depth was combined with deviations in the x and y angulation (6°), measurement errors of up to 88% (1.5 cm2) occurred.
The type II stenosis had a MVA of 1.15 cm2. Deviations of 6° in both the x and y axes led to relative measurement errors of 30% (1.5 cm2). When the C plane depth was positioned 2 mm parallel to the optimal C plane, measurements of MVA showed a maximal relative error of 13% (1.3 cm2). When the change in depth was combined with deviations in the x and y angulation (6°), measurement errors of up to 57% (1.7 cm2) occurred.
Beat-to-beat variation of MVA
The subgroup of patients in whom beat-to-beat variation was analyzed consisted of 14 patients in atrial fibrillation and 1 patient in sinus rhythm. The average coefficient of variation of RR intervals was 14 ± 9%. The mean coefficient of variation for the MVA as calculated for six different beats was lower than the intraobserver variability obtained from six measurements in one single beat (6.4 ± 4% vs. 9.0 ± 4.0%, p = ns). Thus, beat-to-beat variation was minimal and within the range of individual measurement errors.
This study demonstrates that accurate and reproducible measurements of mitral valve areas can be obtained with real-time 3-D echo. In contrast to 2-D echo, 3-D echo permits measurements from an apical window, and the interobserver variability for measurements with this technique appears to be significantly less than with 2-D echo.
Planimetry of the mitral valve
Planimetry of the mitral valve orifice is the only direct measurement of the MVA. Thus, it should be superior to indirect measurements such as cardiac catheterization using the Gorlin formula (4,16–18), Doppler echocardiography using the PHT (6,19–22), the continuity equation (23) or the proximal isovelocity surface area method (24), which all have significant limitations.
However, planimetry is presently limited to those patients in whom parasternal images of high quality can be obtained. The success rate has been reported to be as low as 75% depending on the study population (25).
Another major limitation of this method is the difficulty of defining the correct image plane that displays the true mitral valve orifice. The individual geometry (or morphology) of a stenotic mitral valve may range from the typical shape, with mobile doming of the anterior mitral valve leaflet, to a more funnel-like morphology. The active and passive motion of the mitral apparatus throughout diastole further complicates the definition of the correct image plane.
Our results show that depending on the geometry of the orifice, even a small deviation of 6° from the optimal plane can lead to an overestimation in MVA in the range of 63%. This error increases to 88% when this angle deviation is combined with a parallel shift of only 2 mm along the long axis. Thus, small changes of the transducer position on the chest wall and of its tilting and rotation result in significant changes of calculated orifice area. This is also the main reason that considerable experience and operator skill are necessary for the correct application of this method and why significant interobserver variability has been noted for these measurements.
Volumetric real-time 3-D echo
It has recently been demonstrated that 3-D echo provides accurate and reproducible measurements of MVA (26,27). However, in these reports 3-D echo was performed with a rotational scanning system. This technique requires electrocardiographic and respiratory gating, both complex post-processing procedures. Furthermore, it is time-consuming (in some instances more than 1 h) and susceptible to reconstruction artifacts (28). In many patients, the image quality that is required for appropriate reconstruction can only be achieved with transesophageal echocardiography. Thus, this method is not ready for clinical use.
Volumetric real-time echo is a novel imaging concept, which holds promise as a “break-through” technology for 3-D echo. Employing a matrix array echo probe, this technique allows instant (real-time) acquisition of a complete 3-D data set without complex post-processing. Several studies have already demonstrated the validity of real-time volumetric echo for the calculation of cardiac volume (11).
The present study shows that volumetric real-time 3-D echo also allows accurate assessment of MVA. The smaller interobserver variability of real-time volumetric echo compared with 2-D echo possibly reflects the greater accuracy of real-time volumetric echocardiography to define the image plane with the true orifice of the mitral valve.
Furthermore, 3-D echo allows closer evaluation of potential beat-to-beat variations of MVA in patients with mitral stenosis. As MVA can be measured repeatedly for one individual beat, it should be possible to separate variations of the MVA caused by measurement errors from a true beat-to-beat variation. The results of the present study suggest that beat-to-beat variation, if it exists, is very small and well below the intraobserver variability for such measurements.
In addition, our experience suggests that measurements of the MVA can be derived faster (within minutes) with 3-D echo than with 2-D echo. The ability to store a volumetric data set (representing a complete echocardiographic examination) also allows off-line calculation (or recalculation) of MVAs. Thus, real-time volumetric echo is not only an accurate and reliable but also a practical modality for the assessment of patients with mitral stenosis. Nevertheless, as with conventional echo, image quality and optimal instrument settings are of great importance.
Mitral valve areas calculated with real-time volumetric echo were slightly smaller than those derived with 2-D echo. This might be related to partial volume effects on the C-plane images, which display fewer details than conventional 2-D echo. However, image quality was sufficient to allow delineation of the orifice contour in all patients and will improve further within the near future.
Possibly, the fact that MVAs by 3-D echo were on average slightly smaller than those by 2-D echo simply reflects the greater potential of 3-D echo to detect the true anatomic orifice area and the tendency of 2-D echo to overestimate this area because of difficulties in defining the optimal imaging plane.
Mitral valve areas measurements obtained from 3-D echo were not compared with measurements derived from cardiac catheterization. However, as mentioned above, this method has several limitations and is no longer considered a gold standard, as it is not superior to the echocardiographic approach (29).
Volumetric real-time 3-D echo provides accurate and reproducible estimates of MVAs and appears to be superior to conventional echocardiographic techniques such as planimetry using 2-D echo or the PHT method. Measurements are simple and can be performed within a few minutes. In addition, planimetry using 3-D echo is not limited to the parasternal window. Foreseeable equipment refinements will enhance image quality and further improve accuracy.
- mitral valve area
- pressure half-time
- standard deviation
- Received May 20, 1999.
- Revision received January 17, 2000.
- Accepted March 2, 2000.
- American College of Cardiology
- Nichol P.M.,
- Gilbert B.W.,
- Kisslo J.A.
- Fredman C.S.,
- Pearson A.C.,
- Labovitz A.J.,
- Kern M.J.
- Hatle L.,
- Angelsen B.,
- Tromsdal A.
- Flachskampf F.A.,
- Weyman A.E.,
- Gillam L.,
- Liu C.M.,
- Abascal V.M.,
- Thomas J.D.
- Faletra F.,
- Pezzano A. Jr.,
- Fusco R.,
- et al.
- Kawahara T.,
- Yamagishi M.,
- Seo H.,
- et al.
- Shiota T.,
- Jones M.,
- Chikada M.,
- et al.
- Wilkins G.T.,
- Weyman A.E.,
- Abascal V.M.,
- Block P.C.,
- Palacios I.F.
- ↵Karp K, Teien D, Bjerle P, Eriksson P. Reassessment of valve area determinations in mitral stenosis by the pressure half-time method: impact of left ventricular stiffness and peak diastolic pressure difference. J Am Coll Cardiol 1989;1;13:594–9.
- Hammermeister K.E.,
- Murray J.A.,
- Blackmon J.R.
- Carabello B.A.
- Thomas J.D.,
- Weyman A.E.
- Derumeaux G.,
- Bonnemains T.,
- Remadi F.,
- Cribier A.,
- Letac B.
- Rifkin R.D.,
- Harper K.,
- Tighe D.
- Binder T.,
- Globits S.,
- Zangeneh M.,
- et al.
- Klarich K.W.,
- Rihal C.S.,
- Nishimura R.A.