Background
Echocardiography has evolved into the most predominant diagnostic imaging technique in cardiology. Over the last five decades the diagnostic capability of echocardiography has increased dramatically from M-mode to two-dimensional (2-D) imaging. Recent advances in ultrasound instrumentation and computer technology have led to three-dimensional (3-D) echocardiography, introducing a new era in cardiovascular imaging.1
Every imaging technique in cardiology aims at a complete visualisation and comprehensive assessment of cardiac morphology and pathology, as the heart is a complex geometric structure. Analysis of the heart in motion in all three or four (including time) dimensions can therefore further facilitate and enhance the diagnostic capabilities of echocardiography. Three-dimensional echocardiography is still in its evolution and at the phase of early adaptation with respect to its clinical application. It should complement current echocardiographic techniques by providing better understanding of the topographical aspects of pathology and refined definition of the spatial relationships of (intra)cardiac structures. Furthermore, it provides new measures not described by 2-D echocardiography and makes the existing measures more accurate.2
The assessment of patients with mitral valve disease is one of the most challenging and promising clinical applications of 3-D echocardiography. The 3-D anatomy of the mitral valve, as well as the feasibility, accuracy and incremental value of 3-D echocardiography in the evaluation of mitral valve disease, will be discussed.
3-D Reconstruction
There are two main approaches of 3-D reconstruction. The first is random or freehand scanning, which is based on free motion of the ultrasound transducer. Its position in space is located by an acoustic, electromagnetic or mechanical arm location device. A transthoracic approach is used in this mode of acquisition. The limitation of this method is that accurate endocardial border identification is not possible because of big spaces between imaging planes. The second approach is sequential scanning, where the ultrasound motion is predetermined in linear, fan-like or rotational ways.
A transthoracic or trans-oesophageal approach is used for this mode of data acquisition. Both the sequentially or randomly collected 2-D images are processed off-line by the computer using interpolation algorithms so that the gaps between individual images are filled and, finally, a volumetric 3-D data set is generated. There are several ways to present the data in the 3-D volumetric data set. The following are the most used.
Anyplane Echocardiography
This mode of presentation allows the examiner to generate 2-D tomographic images in any desired orientation that can be physically unobtainable by conventional 2-D echocardiography.
Volumetric Rendering Technique
This 3-D reconstruction creates images resembling the true anatomy of the heart. By choosing a cutting plane and reconstructing the image beyond this plane the heart can be opened as if by a surgeon. Different structures can be examined en face with increased perception of the anatomic relationships.
The ideal for 3-D reconstruction is realtime 3-D echocardiography (RT3D). The system uses a novel matrix phased-array transducer with parallel processing to scan a pyramidal volume. In a pyramidal volume, the images are displayed as anyplane or volume rendered images immediately. Second-generation matrix transducers recently became available for clinical studies.
Further advances in computer technology have enabled encoding of colour flow Doppler data together with greyscale imaging and 3-D presentation of Doppler flow events with surrounding cardiac anatomy.3
Functional Anatomy of the Mitral Valve
3-D Topography of the Mitral Valve
Viewing the mitral valve from the left atrium with 3-D echocardiography shows that the anterior and posterior mitral leaflets have several indentations dividing them into segments or scallops (see Figure 1). The anterior leaflet has three segments - A1 (anterolateral), A2 (middle) and A3 (posteromedial). Similarly, the posterior leaflet has three segments - P1 (anterolateral), P2 (middle) and P3 (posteromedial). The anterior and posterior leaflets are fused for 3mm to 8mm medially and laterally at the trigones and usually form distinct commissures (anterolateral and posteromedial).
The anterior leaflet comprises roughly two-thirds of the valve area, is approximately twofold longer than the posterior leaflet and is somewhat triangular. The posterior leaflet is more elongated and rectangular. The anterior leaflet is attached to the septum and fibrous annulus of the heart and it is relatively non-distensible. Although the anterior leaflet accounts for two-thirds of the mitral valve area, its attachment to the mitral annulus accounts for only approximately one-third of the mitral annular circumference. The anterior mitral leaflet spans the distance between the two fibrous trigones and is in direct continuity with the non-coronary aortic valve leaflet. The posterior mitral valve leaflet is attached to the posterior two-thirds of the mitral annulus, which runs along the free wall of the left ventricle (LV) and is primarily muscular with little fibrous tissue (explaining its tendency to distend and elongate).
Non-planarity of the Mitral Valve and Annulus
Based on the studies conducted in an effort to refine the diagnosis of mitral valve prolapse, the geometric shape of the mitral valve and its annulus was defined. It has been documented that the mitral annulus and leaflets are non-planar saddle-shaped structures which are equivalent with the so-called hyperbolic paraboloid - a geometric surface of which all sections parallel to one co-ordinate plane are hyperbolas and all sections parallel to another co-ordinate plane are parabolas (see Figure 2). There are two high points (peaks) lying anteriorly and posteriorly at the aortic insertion and posterior left ventricular wall and two low points (troughs) closest to the apex located medially and laterally (see Figure 3). According to Levine et al. the maximum deviation from planarity, i.e. the distance between the highest and lowest points of the mitral annulus, is on average 1.4cm ┬▒ 0.3cm.4 Regarding the leaflet-annular relations, in mediolateral view (four-chamber view in 2-D echocardiography) the leaflets can appear above the mitral annulus but in anterolateral view (parasternal long-axis view in 2-D echocardiography) they do not ascend the annulus. That is the reason for the misdiagnosis of superior leaflet displacement in otherwise normal individuals, in the 2-D echocardiography era. The leaflet-annular non-planarity is rational in two ways. Firstly, as the base of the LV decreases in circumference during systole but the leaflets do not contract, the mitral annular area can decrease in some way by folding, which is achieved by lowering of the distance between high and lower points of the annulus.5,6 Secondly, the saddle-shape provides a configuration capable of withstanding the stresses imposed by left ventricular pressure in systole.
Salgo et al. studied the effect of non-planarity on stress reduction.7 Two shape factors that have synergistic effect on stress reduction have been identified - leaflet billowing and annular non-planarity. The saddle-shape of the mitral annulus was preserved across three mammalian species (human, sheep and baboons) with an annular height commissural width ratio of approximately 15%. Their data suggest that nature conserves the saddle-shaped configuration of the annulus for a mechanical benefit.
Dynamics of the Mitral Annulus
There are two studies by Flachskampf et al. and Kaplan et al. which provide insights into the dynamics of the mitral annulus.5,6 According to these studies the mitral annular area was on average 5-6cm2/m2 corrected for body surface area (because of non-planarity an area projected into the least square plane was measured). This area decreased in systole by approximately 24%. The mechanism by which it was achieved was ellipticalisation due to reduced distance between two high points with increase in annular height and eccentricity and smaller amount by reduced distance between two low points (see Figure 4). At the same time the basic features of non-planarity (two high and two low points) were preserved throughout the cardiac cycle.
Mitral Stenosis
The severity of mitral stenosis is assessed mainly by estimation of the mitral valve area.8 Currently, the available 2-D echocardiographic and Doppler techniques have their well-known limitations. The advent of 3-D echocardiography has refined the assessment of mitral valve area to such an extent that it is now considered as a gold standard. Besides, 3-D echocardiography has introduced new indices which further improve the diagnosis of mitral stenosis. By means of 3-D echocardiography it is possible to assess the morphology of the mitral valve in balloon mitral valvuloplasty, which is important in the evaluation of the mechanism and success of the procedure.
Assessment of Mitral Valve Area
With a 3-D homogeneous data set using anyplane echocardiography it is possible to pinpoint the cut plane to the the tips of the mitral valve so the true anatomic valve area can be measured (see Figure 5). The advantage of this method is that, contrary to 2-D echocardiography, proper alignment of the cut plane is controlled in a 3-D data set.
It is important because errors due to malpositioning can be obviated. It has been shown that malpositioning errors can achieve up to 88% (1.5cm2) in the measurement of the mitral valve area that is not acceptable in the management of patients with mitral stenosis.9 In addition, assessment of the anatomic mitral valve area is advantageous because it is haemodynamically independent, contrary to effective mitral valve area (measured by pressure half-time and proximal isovelocity surface area (PISA) methods), which is haemodynamically influenced by associated abnormalities (aortic insufficiency and increased left ventricular stiffness).
In the first studies carried out by Kupferwasser et al. and Chen et al. the mitral valve area was assessed by anyplane 3-D trans-oesophageal echocardiography (TOE).10,11 The mitral valve area assessed by 3-D echocardiography was compared with the mitral valve area measured by 2-D methods (2-D planimetry and pressure half-time) and invasively assessed mitral valve area according to the Gorlin formula. Only Kasliwal et al. compared the mitral valve areas by 3-D echocardiography with the true mitral orifice measured directly at operation.12 The comparison achieved a high degree of agreement (r=0.95); thus, 3-D echocardiography can currently be considered as a new clinical standard in the assessment of the anatomic mitral valve area. 3-D echocardiography has also been shown as accurate in the assessment of mitral valve area using a transthoracic approach. Sugeng et al. confirmed that freehand 3-D transthoracic echocardiography, compared with 2-D planimetry, pressure half-time and PISA methods, was the most accurate when compared with invasively determined mitral valve areas according to the Gorlin formula.13 The most attractive is definitely RT3D echocardiography which allows online assessment of the mitral valve area. Images are displayed as two simultaneous intersecting orthogonal long-axis scans (B-mode scans) and two perpendicular short-axis scans (C-mode scans). These C-mode scans allow the display of short-axis views of the mitral valve from an apical transducer position. Binder et al. found that RT3D echocardiography compared with 2-D planimetry and pressure half-time methods allows accurate measurement of the mitral valve area from the transthoracic approach.9 It is evident in patients with an adequate acoustic window.
New Indices of Mitral Stenosis
The thickening of mitral leaflets in rheumatic mitral stenosis is a well-known phenomenon. Limbu et al. were the first who quantified the mitral valve volume in vivo in normal subjects and patients with rheumatic mitral stenosis (see Figure 6).14 The mitral valve volume in normal individuals was 4.5ml on average and 9ml in patients with mitral stenosis. When they divided patients with mitral stenosis into the sinus rhythm and atrial fibrillation groups, patients with atrial fibrillation had a propensity to have a larger mitral valve volume and were older than patients with sinus rhythm; an aetiologic relationship between atrial fibrillation and further enlargement of the mitral valve volume was speculated. Gilon et al. studied the hypothesis in vitro that the stenotic mitral valve influences the pressures and flows not only by cross-sectional area but also by 3-D geometry of the stenotic valve proximal to the orifice.15 With the use of 3-D echocardiography and stereolithography they constructed different shapes of mitral valve by laser polymerisation - domed, intermediate and flattened. Coefficient of contraction was calculated as effective area divided by anatomic orifice area. Coefficient of contraction decreased as the mitral valve was flattened. The study confirmed that variations in contraction coefficient (i.e. the 3-D geometry of the mitral valve) led to varying pressure gradients that were up to 40% higher for the flattest valves; therefore, doming valves permit a higher cardiac output than flat valves.
3-D Echocardiography in Balloon Mitral Valvuloplasty
3-D echocardiography by volume rendering allows visualisation of the mitral valve en face either from the left atrium or the LV. Applebaum et al. evaluated the mechanism of balloon valvuloplasty by 3-D echocardiography.16 Volume rendered 3-D images enabled visualisation of commissural splitting and leaflet tears not seen with 2-D (see Figure 7 and 8). They found that balloon mitral valvuloplasty was more successful when complete splitting was achieved compared with partial splitting. Moreover, in 38% of patients in whom an increase of mitral regurgitation developed, tear was visualised by 3-D. Langerveld et al. conducted a similar study, where 3-D TOE enabled a better description of the mitral valvular anatomy following balloon mitral valvuloplasty, compared with 2-D echocardiography. In addition, significant relation of mitral valve volume before valvuloplasty to a successful procedure was found.17
Mitral Valve Prolapse
Echocardiography is the most utilised imaging modality for diagnosis of mitral valve prolapse. M-mode and 2-D echocardiography frequently lead to false-positive and false-negative diagnoses due to the non-planar leaflet-annular relationships of the mitral valve. Prolapse is generally defined as a displacement of a bodily part from its normal position or relations. By 3-D echocardiography it is possible to visualise the mitral valve en face from either the left atrium or the LV.18,19 In volume rendered images looking down in the left atrium, mitral valve prolapse is viewed as a convexity or bulge and often as a bright area compared with the rest of the mitral leaflet. Looking up in the LV mitral valve, prolapse appears as a spoon-like depression. In patients with mitral valve prolapse and mitral regurgitation a crack due to non-coaptation can be identified (see Figure 9). 3-D echocardiography allows accurate identification and quantification of the prolapse of individual scallops/segments of the mitral valve leaflets (see Figure 10). Two intra-operative studies, conducted by Ahmed et. al and Chauvel et al., confirmed that the topography of prolapsing scallops/segments shown by 3-D echocardiography was correct in 78% and 86%, respectively. Contrary to 2-D echocardiography, 3-D echocardiography allowed measurements of the area and width of the prolapsed portion of the leaflet as well as measurements of the circumference of the posterior part of the mitral annulus. This information could aid the surgeon in deciding the extent of valvular tissue resection.20,21
Mitral Regurgitation
Accurate evaluation of mitral regurgitation severity is a challenging task in clinical cardiology. The current 2-D echocardiographic methods used to quantify mitral regurgitation have their well-known limitations. 3-D echocardiography has the potential to improve the assessment of mitral regurgitation by facilitating the visualisation of complex mitral anatomy in three dimensions and providing more accurate quantification of regurgitant colour Doppler flow events.22-24 Investigators have aimed at validating 3-D echocardiography mainly in the measurements of regurgitant jet volumes, flow convergence surface area and anatomic regurgitant orifice area.
3-D Quantification of Regurgitant Jet Volume
The first studies used volume rendered greyscale imaging to visualise Doppler flow in three dimensions.25,26 The differentiation of regurgitant jets from surrounding cardiac structures was difficult. The volume of the mitral regurgitant jet was measured by a 'summation of discs' method.27 Later, De Simone and colleagues proposed a method of colour coding of regurgitant mitral jets derived from digital data.28-32 Jet volumes were calculated by segmentation with automatic selection of turbulence and high-velocity components or volume units (voxels) containing the selected Doppler data. This mode of 3-D colour flow imaging recognised different patterns of eccentric regurgitant jets not previously described, such as cylinder, tongue, spiral and spoon-like patterns. What is more important is that calculation of the jet volume was capable of accurately quantifying asymmetrical jets. Recently, Sugeng et al. improved the method of colour encoding of regurgitant jets together with greyscale imaging of the surrounding cardiac anatomy.33 This improved technique provided information on the origin and extent of the dehiscence in case of paravalvular leaks, as well as insight into the direction of the regurgitant jets.
Flow Convergence Zone and 3-D Echocardiography
The flow convergence or PISA method is based on the phenomenon that flow accelerates towards the regurgitant orifice and forms a series of concentric hemispheric shells of increasing velocity. According to the continuity concept, the flow rate is calculated by multiplying the isovelocity surface area and its corresponding aliasing velocity; thus, accurate measurement of the flow convergence surface area is the most important aspect for obtaining accurate flow rate. Conventional 2-D methods rely on assumptions that the isovelocity surface is hemispheric or hemielliptic. However, 3-D imaging has revealed that the morphology of the flow convergence region is more complex and unpredictable in shape.34 Since 3-D can display the entire flow convergence region en face viewing from the left atrium, a more accurate assessment of its area can be achieved without the need to make geometric assumptions. 3-D flow convergence-based methods have been shown to accurately predict flow rate.35,36
Regurgitant Orifice Area Measurement by 3-D Echocardiography
The regurgitant orifice area is a measure of valvular incompetence useful for assessment of mitral regurgitation. To date, measurement of the regurgitant orifice area was based on Doppler methods that calculated the effective orifice area. 3-D echocardiography provides an opportunity to visualise the regurgitant orifice and so anatomic regurgitant orifice area measurement. It has been shown that anatomic regurgitant orifice area measured directly by planimetry from 3-D volume rendered images correlates well with effective regurgitant orifice area calculated by the proximal convergence method.37-39
Functional Mitral Regurgitation
Functional mitral regurgitation is defined as an insufficiency of the structurally normal mitral valve developing as a consequence of regional or global left ventricular dysfunction. It is a complication of either chronic ischaemic heart disease or dilated or hypertrophic cardiomyopathy. Functional mitral regurgitation is associated with increased mortality independent of left ventricular dysfunction. The mechanisms that participate in the development of functional mitral regurgitation are related to the geometry of the mitral valve, mitral annulus and papillary muscles. Since the relation of these anatomical structures is explicitly 3-D, 3-D echocardiography provides the best mode to study their relationship.
Mitral Annular Geometry
Flachskampf et al. and Kaplan et al. have shown that functional mitral regurgitation is associated with annular dilation and its reduced cyclic variation by using 3-D echocardiography.5,6 Compared with normal subjects, the annulus in patients with functional mitral regurgitation is larger and has greater mitral annular area, longer perimeter, reduced annular height and eccentricity and increased distance between high points of the mitral annulus.
Leaflet Geometry
Using RT3D echocardiography, Kwan et al. studied the difference of mitral valve deformation between ischaemic and dilated cardiomyopathy with significant functional mitral regurgitation.40 Mitral valve tethering has been found to be the strongest determinant of mitral regurgitation severity and the pattern of mitral valve deformation was asymmetrical in ischaemic heart disease, whereas it was symmetrical in dilated cardiomyopathy.
Papillary Muscle Geometry
Several studies have been performed in an effort to elucidate a 3-D papillary muscle-mitral relationship.41-44 The results of these studies show that medial and posterior shift of the ischaemic medial papillary muscle, measured by 3-D reconstruction, is particularly related to the development of functional mitral regurgitation.
Mitral Valve Repair and 3-D Echocardiography
Mitral valve repair has become more common in the last decade, accounting for half of the mitral valve procedures.45 Various techniques have been proposed for valve reconstructions.46-48 The decision of 'how to operate' depends on the underlying pathology of the mitral valve diagnosed by pre-operative echocardiography. Conventional 2-D echocardiography is a useful guide for accurate surgical analysis; however, in complex valvular pathologies some spatial relations and different structural features can be perceived erroneously even by experienced echocardiographers. The technique of repair is consistently modified in the operating room by close examination of the mitral valve, although the surgeon is challenged by limited time, operating field and non-physiological condition of the heart being devoid of blood. 3-D echocardiography has the potential to overcome these difficulties by showing the heart in the 'surgeon's view' and even in a more physiological state as in operation.
There are few published data regarding the feasibility of 3-D echocardiography in the operating theatre. Abraham et al. demonstrated that, in 25% of cases, 3-D echocardiography can detect new morphologic findings (mainly valve fenestrations) not seen with 2-D TOE. In one patient, 3-D TOE resulted in a decision to perform valve repair instead of replacement. As previously mentioned, 3-D TOE has been proven as accurate in identifying the location of the prolapsing segment and quantifying the amount of the prolapsed tissue by measuring the area or the width. This information could aid the surgeon in deciding the extent of mitral valve resection.49
Conclusion
3-D echocardiography allows visualisation of the heart differently to 2-D echocardiography, as it looks at the heart in true reality. The assessment of the morphology, function and pathology of the heart, and particularly the mitral valve apparatus by 3-D echocardiography, becomes more accurate. Compared with 2-D echocardiography, 3-D echocardiography offers advantages for the morphologic and quantitative assessment of mitral valve stenosis, prolapse and regurgitation. It appears that 3-D echocardiography has the potential for planning operations and assessing interventional or surgical results. Furthermore, 3-D echocardiography provides new quantitative indices unobtainable by conventional 2-D imaging. Both technical improvement and larger studies will enhance the clinical applicability of 3-D echocardiography in the near future.