Accurate assessment of the true mitral valve area in rheumatic mitral stenosis
- Correspondence to Professor Mark J Monaghan, Kings College Hospital, Department of Non-Invasive Cardiology, 2nd Floor, Hambleden Wing, Denmark Hill, London SE5 9RS, UK;
Rheumatic mitral valve stenosis (MVS) is still a common condition, especially in developing countries. Echocardiography has a primary role in its diagnosis, assessment of its severity and its functional implications. The indications for the quantification of MVS have recently changed. The European Association of Echocardiography/ American Society of Echocardiography guidelines in 20091 recommend using either the pressure gradient across the mitral valve (MV) by continuous wave Doppler, the pressure half time or MV planimetry. The recent European guidelines on the management of valve disease2 advise, as the method of choice, measurement of the MV area (MVA) by planimetry whenever feasible. Continuity equation for the estimation of valve area, proximal isovelocity surface area and the calculation of the mean transvalvular gradient, using Doppler velocities are suggested when additional assessment is needed.
It is worth questioning the reasons for this recent change in recommendations. The indirect methods of measuring the MVA have been tested in the past and have shown moderate reliability compared with cardiac catheterisation.3 ,4 However they are influenced by many variables such as heart rate, cardiac output, left ventricular systolic and diastolic functions, left ventricular and atrial compliance, left ventricular hypertrophy and concomitant valve disease.5–9 In these common clinical scenarios, Doppler-derived measurements of the MV area don't allow an accurate assessment of MVS severity.
Instead, MV planimetry is a direct and relatively hemodynamic-independent measurement of MVA. However in the presence of significant leaflet calcification, severe left atrial dilatation or distortion of the valve anatomy, it can be very challenging to obtain a perfect geometric alignment at the MV leaflet tips in a parasternal short-axis view using two-dimensional (2D) echocardiography. This clearly leads to an inaccurate measurement of the MVA.
Three-dimensional (3D) echocardiography enables the acquisition of a volumetric dataset incorporating the entire MV apparatus. Using 3D analysis software it is possible to perform a multiplane reconstruction from the dataset and generate a perfectly aligned cross-sectional image of the MV at the leaflet tips. Since the stenotic MV is almost like a funnel shaped structure, the valve area is at its minimum here and it represents the true anatomical MVA.
3D transthoracic echocardiography (TTE) planimetered valve area has been demonstrated to have a good correlation with invasive MVA measurements10–15 and it appears more reliable than the 2D planimetry used for measuring the true MV orifice. However, suboptimal image quality is an important limitation of 3D TTE, which may limit its application in patients with poor acoustic windows. 3D transoesophageal echocardiography (TOE) rarely suffers from poor image quality and generally allows excellent visualisation of the MV, the subvalvular apparatus and the commissural anatomy.16–20 As described above, with a 3D multiplane analysis it is possible to reliably measure the true anatomical MVA at the point of the smallest MV orifice during its maximum diastolic opening (figure 1).
The report by Sun-Yang Min et al 21 compares MVA measurements obtained by 2D transthoracic planimetry and 3D transoesophageal planimetry in a cohort of patients with rheumatic MVS.
The authors observed that the agreement between the MVA measurements using 2D planimetry and 3D TOE was good (ICC: 0.853, 95% limits of agreement: 0.784–0.902, p<0.001); however, 2D planimetry overestimated MVA by 0.19±0.2 cm2 compared with 3D TOE (p<0.001). The overestimation of the MVA by 2D TTE could be explained by a suboptimal alignment of the 2D echocardiographic imaging plane with the MV leaflet tips. As previously mentioned the funnel-like morphology of the stenotic MV may result in erroneously increased valve area measurements if the 2D short axis view is slightly above and/or not truly perpendicular to the plane of the leaflet tips. Obtaining correct alignment can become even more challenging in the presence of severely distorted MV anatomy or a much enlarged left atrium.
Sun-Yang Min et al assessed, in a multivariate linear regression analysis, some factors which could possibly influence the difference between MVA by 2D TTE and 3D TOE. The ejection fraction, left atrium dimension and Mα resulted in statistically significant determinants of the difference in MVA by 2D and 3D. Mα represents the angle between the line of the true MV tip and the line of the echocardiographic beam to the MV tip, measured in parasternal long axis view in early diastole, using TomTec software. The authors created a formula in order to correct the 2D TTE MV planimetry by LA dimension and Mα.
In the second validation cohort of 31 patients the corrected MVA showed an excellent correlation with the MVA by 3D TOE (confidence interval (ICC): 0.925, 95% limits of agreement: 0.850–0.962, p<0.001) and there was no significant difference between the corrected MVA and MVA by 3D TOE (−0.05±0.2 cm2, p=0.08). The cut-off values in predicting significant overestimation by 2D planimetry were: LA dimension ≥49 mm (78% sensitivity, 72% specificity) and Mα≥9.5° (56% sensitivity, 89% specificity).
These results are very encouraging. However the routine application of this formula is possibly limited because it is time consuming and requires the use of adjunctive software.
As mentioned before there is already good evidence for the routine use of 3D TTE in patients with MVS, this was also underlined in the latest EAE/ASE recommendations on the use of three-dimensional echocardiography. For the optimal acquisition of a mitral 3D dataset, the parasternal window is recommended, as the MV structure is closer to the transducer resulting in higher temporal and spatial resolution than when apical windows are utilised.
The x-plane is an alternative 2D visualisation mode that can also be very useful in this setting. It is based on the capability of acquiring two perpendicular scan planes simultaneously. The 3D matrix transducer technology facilitates multidirectional beam steering. In this setting it allows the positioning of the short-axis plane across the MV from a reference line displayed on the simultaneous long-axis view, thereby making it easier to ensure the short-axis is at the mitral leaflet tips. However, if the plane of the mitral orifice is very oblique, this methodology will still not allow correction in the same way that 3D multiplane reconstruction does.
Sun-Yang Min et al used a 3D TOE full volume mode for the dataset acquisition. The use of full volume provides relatively high temporal resolution. However it can result in stitching artefacts, mainly caused by translation of the heart during respiration or irregular R-R intervals during the volume acquisition time. Of course, many patients with mitral stenosis will be in atrial fibrillation resulting in an irregular heart rate which could have an adverse effect on the quality of the 3D datasets. The 3D zoom modality16–23 allows real-time acquisition of a MV dataset in one heartbeat without stitching artefacts. Therefore it should be the method of choice in patients, not capable of breath holding or in the presence of arrhythmia such as atrial fibrillation. The 3D zoom also eliminates the need for subsequent volume cropping besides providing a live 3D view during dataset acquisition (figure 2). Since all measurements are made by analysis of frozen still frames, the slightly lower volume rate obtained with this modality is not a disadvantage.
The interesting article by Sun-Yang Min et al underlines the importance of anatomically and geometrically corrected measurement of the MV area as the most reliable method to quantify MVS. As discussed earlier, 3D TOE planimetry is probably the most reliable method of accurate MVA assessment,16 however being an invasive procedure, is not routinely recommended.1 ,2 Therefore, it is important to determine a subgroup of patients who could benefit from further investigations by 3D TOE. The subgroup identified in this study includes patients with an enlarged left atrium and large Mα in which 2D planimetry has been shown to be less accurate.
This study is consistent with recently published recommendations22 on the utilisation of 3D echocardiography and highlights the increasingly important role that the technique is now playing in the evaluation of valvular heart disease.
Contributors SG and MJM are coauthors in the editorial.
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.