Background
Differentiating between mitral valve (MV) prolapse (MVP) and MV billowing (MVB) on two-dimensional echocardiography is challenging. The aim of this study was to test the hypothesis that color-coded models of maximal leaflet displacement from the annular plane into the atrium derived from three-dimensional transesophageal echocardiography would allow discrimination between these lesions.
Methods
Three-dimensional transesophageal echocardiographic imaging of the MV was performed in 50 patients with ( n = 38) and without ( n = 12) degenerative MV disease. Definitive diagnosis of MVP versus MVB was made using inspection of dynamic three-dimensional renderings and multiple two-dimensional cut planes extracted from three-dimensional data sets. This was used as a reference standard to test an alternative approach, wherein the color-coded parametric models were inspected for integrity of the coaptation line and location of the maximally displaced portion of the leaflet. Diagnostic interpretations of these models by two independent readers were compared with the reference standard.
Results
In all cases of MVP, the color-coded models depicted loss of integrity of the coaptation line and maximal leaflet displacement extending to the coaptation line. MVB was depicted by preserved leaflet apposition with maximal displacement away from the coaptation line. Interpretation of the 50 color-coded models by novice readers took 5 to 10 min and resulted in good agreement with the reference technique (κ = 0.81 and κ = 0.73 for the two readers).
Conclusions
Three-dimensional color-coded models provide a static display of MV leaflet displacement, allowing differentiation between MVP and MVB, without the need to inspect multiple planes and while taking into account the saddle shape of the mitral annulus.
Mitral valve (MV) prolapse (MVP), with a prevalence of about 2% to 3%, is the most common cause of mitral regurgitation (MR) in North America. Although in the clinical setting, the terms billowing and prolapse are often used interchangeably, these entities are not identical, and differentiation between them on two-dimensional (2D) echocardiography can be challenging. Although prolapse has been associated with endocarditis (up to four times more common), sudden death (<2% of cases during long-term follow-up), and stroke (<1% of cases), these complications are less frequently associated with billowing. Accordingly, the differentiation between prolapse and billowing is important for patient management and also for preoperative evaluation in patients referred for MV surgery.
From the surgeon’s perspective, MVP is defined when the free edge of the leaflet remains above the plane of the annulus at end-systole. It is usually due to chordal or papillary muscle elongation or rupture, resulting in disruption of leaflet coaptation. MV billowing (MVB), on the other hand, is defined when there is systolic protrusion of the body of the leaflet above the annulus plane with the free leaflet edge remaining at or below the annular plane during end-systole ( Figure 1 ). Using 2D transthoracic echocardiography, MVP has been defined for the long-axis view as end-systolic displacement of the body of the MV leaflet ≥2 mm above the mitral annular plane. This definition takes into consideration the saddle shape of the mitral annulus. In this view, the most commonly visualized scallops are A2 and P2, making it difficult to diagnose prolapse of the remaining nonvisualized scallops. Two-dimensional transthoracic echocardiographic transducer manipulations have been proposed to improve the visualization of the A1/P1 and A3/P3 scallops, but these views are often difficult to obtain and do not allow standardized measurements. Using 2D transesophageal echocardiographic (TEE) imaging, identification of mitral leaflet scallops is easier, but prolapse is still felt to be best identified in the long axis view.
Three-dimensional (3D) TEE imaging provides incremental benefit in the assessment of the extent and location of degenerative MV disease. Although the majority of prolapsed scallops can be seen on standard 2D TEE imaging, 3D TEE imaging provides a better anatomic appreciation of the size and anatomic location of the prolapsed scallop as well as information on the presence of clefts and fissures. Specifically, the surgical definition of prolapse and billowing (see above) can be applied by assessing multiple 2D cut planes extracted from the 3D TEE data sets. This definition effectively combines the long-axis and midcommissural views to assess leaflet position. The problem with this approach, however, is that this multiplane evaluation can be tedious and time-consuming. To circumvent this limitation, analysis software that generates color-coded parametric models of leaflet displacement from the annular plane into the atrium has been recently developed. because these color-coded models incorporate the saddle-shape of the mitral annulus and clearly depict the line of coaptation, we hypothesized that these parametric maps would be useful for differential diagnosis of MVP versus MVB. This study was designed to test this hypothesis and to generate a set of criteria that could be used as the basis for this objective diagnosis.
Methods
Patient Population
We retrospectively identified 38 nonconsecutive patients with degenerative MV disease confirmed by TEE evaluation of the MV and 12 patients with normal MVs (26 men, 24 women; mean age, 59 ± 15 years; range, 28–90 years). The study was approved by the institutional review board of the University of Chicago Medical Center.
2D and 3D TEE Imaging and Analysis
Two-dimensional and 3D TEE studies were performed using a Phillips iE33 ultrasound system (Philips Medical Systems, Andover, MA) with a fully sampled matrix-array transducer (probe model X7-2t). The 2D studies were performed according to a standard clinical protocol, which comprises a complete assessment of the MV, including 2D and color Doppler surveys in multiple views. Three-dimensional zoomed data sets of the MV were acquired from the midesophageal long-axis view (120°) over four beats. Gain settings were optimized before data acquisition using the narrow-angle mode. Biplane imaging was performed before zoomed acquisition to ensure the capture of the entire valve annulus, while both lateral elevation and box height were optimized. Three-dimensional data sets were analyzed offline on an Xcelera workstation (Philips Medical Systems) by an expert echocardiographer. The final image of the MV was presented in the surgeon’s view with the aortic valve in the 12 o’clock position. MR was quantified from 2D TEE images using vena contracta criteria on the basis of published guidelines.
3D TEE Multiplane Reference Standard
Visual inspection of dynamic 3D renderings of the MV and multiple 2D planes extracted from the 3D TEE data sets were used to diagnose leaflet segment prolapse and billowing. The results of this assessment were used as the reference standard for comparison with the volumetric modeling of the MV. This analysis was conducted offline using commercial 3D quantification software (QLAB version 9.0; Phillips Medical Systems). Initially, the end-systolic frame, defined as the frame immediately before aortic valve closure, was selected. Then, multiple parallel 2D cut planes were manually extracted from the 3D data set along the coaptation line ( Figure 2 ). Assessment of the leaflet edge position was accomplished by manually positioning the cut planes across the area of apparent excess leaflet tissue. Prolapse was diagnosed when the tip (or edge) of the leaflet was located in the left atrium (above the mitral annulus) at end-systole. Billowing was diagnosed if the tip (or edge) of the leaflet was located at or below the annular plane at end-systole, and only the body of the leaflet protruded into the left atrium ( Figure 2 ).
Volumetric Modeling of the MV
Color-coded parametric models of the MV were generated from 3D TEE data sets using commercial software (MVQ, QLAB version 9.0), as previously described. Briefly, four points were marked in two orthogonal planes (one plane being the midesophageal 120° view of the MV) to define the position of the MV in 3D space. These four points included the anterolateral and posteromedial hinge points of leaflet insertion and the anterior and posterior points. The annular perimeter was manually outlined by defining annular points in serial planes rotated around the axis perpendicular to the mitral annular plane. Finally, the leaflets were traced in consecutive parallel long-axis planes spanning the annulus from commissure to commissure to construct the volumetric MVQ model ( Figure 3 ). The final model was displayed as a color-coded surface representing a topographic map of the mitral leaflets ( Figure 3 , bottom ). The colors are coded to display maximal leaflet displacement into the left atrium from the annular plane at end-systole. The model was oriented with the aortic valve in the 12 o’clock position. The models were used to (1) visualize the prolapsed or billowing scallops, (2) define the spatial relationship between the prolapsed or billowing scallops relative to the coaptation line, and (3) assess the integrity of the coaptation line.
Data Analysis
Color-coded parametric models of the MV were reviewed by two echocardiographers with extensive experience with 3D imaging, who defined the diagnostic criteria for the categories normal MV, MVP, MVB, and combined MVB and MVP, using dynamic 3D rendering and the extracted 2D planes from the 3D TEE data set as the reference standard. MVP was diagnosed if a 2D cut plane across the excess leaflet tissue showed the mitral leaflet tip to be above the annulus at end-systole. MVB was diagnosed when a portion of the body of the leaflet protruded above the annular plane but with the leaflet tip remaining at or below the annulus at end-systole. Combined MVB and MVP was diagnosed if 3D renderings of the valve suggested both the presence of billowing segments and the above criteria for MVP were met. A normal valve showed evidence of neither MVB nor MVP.
Subsequently, the color-coded models were shown to two readers with intermediate experience in echocardiography (level 2 echocardiography training) and minimal experience interpreting 3D TEE images. After a brief instructional session, which included review of two images in each category combined with an explanation on how to interpret these images on the basis of the criteria defined by the experts, the readers independently reviewed the color-coded models of the 50 study patients and categorized each study using the above definitions. Each interpretation was then compared with the above reference standard on the basis of the visual inspection of multiple 2D planes extracted from the 3D data sets.
Statistical Analysis
Continuous variables, such as age and left ventricular ejection fraction, are expressed as mean ± SD. The comparisons of color-coded parametric model based interpretations with the reference standard were performed using κ statistics of agreement between categorical variables. The calculated κ coefficients were judged as follows: 0 to 0.20, low; 0.21 to 0.40, moderate; 0.41 to 0.60, substantial; 0.61 to 0.80, good; and >0.80, excellent.
Results
Baseline characteristics of the patient population are shown in Table 1 . The reference standard technique based on visual inspection of dynamic 3D renderings and multiple 2D planes extracted from the 3D TEE data sets identified 12 normal valves, 18 cases of MVB, and 20 cases of MVP, of which nine were identified as MVP alone and 11 as cases of combined MVP and MVB. Ten patients had multisegment billowing (three or more scallops), five patients had multisegment prolapse with billowing (three or more scallops), and one patient had multisegment prolapse with no billowing. For the remaining 22 patients (prolapse, n = 14; billowing, n = 8) individual scallops were counted, and the results are summarized in Table 2 . All patients with prolapse (20 of 20) had moderate or severe MR defined by vena contracta width ( Table 3 ). Only 11% (two of 18) of the billowing valves depicted moderate to severe (vena contracta width, 0.45–0.59 cm) or severe MR (vena contracta width ≥ 0.7 cm). Annular sizes were significantly larger in valves with prolapse compared with those with billowing (MV area, 18.7 ± 5.7 vs 13.7 ± 3.7 cm 2 , P = .003; Table 3 ).
| Prolapse ( n = 20) | Billowing ( n = 18) | Normal MV ( n = 12) | |
|---|---|---|---|
| Age (y) | 59 ± 11 | 56 ± 17 | 63 ± 18 |
| Men | 70 | 22 | 67 |
| Comorbidities | |||
| Coronary artery disease | 2 (10%) | 3 (17%) | 4 (33%) |
| History of congestive heart failure | 7 (35%) | 4 (22%) ∗ | 1 (8%) |
| Atrial fibrillation | 11 (55%) | 10 (56%) | 5 (42%) |
| Left ventricular function | |||
| Left ventricular ejection fraction (%) | 57 ± 11 | 58 ± 10 | 62 ± 9 |
∗ Etiologies of congestive heart failure in these four patients included diastolic dysfunction, coronary artery disease, severe MR (this patient had surgery for severe annular dilatation, no prolapse), and rapid atrial fibrillation.
| Segment | Anterior | Posterior | ||||
|---|---|---|---|---|---|---|
| A1 | A2 | A3 | P1 | P2 | P3 | |
| Prolapse ∗ | 0 | 4 | 3 | 3 | 10 | 1 |
| Billowing | 0 | 2 | 0 | 3 | 5 | 2 |
| Characteristic | Prolapse ( n = 20) | Billowing ( n = 18) | Normal MV ( n = 12) | P value (prolapse vs billowing) |
|---|---|---|---|---|
| MR severity by vena contracta width (cm) | ||||
| Mild (<0.3) | 0 (0%) | 12 (67%) | — | |
| Moderate (0.3–0.44) | 0 (0%) | 4 (22%) | — | |
| Moderate to severe (0.45–0.69) | 11 (55%) | 1 (5%) | — | |
| Severe (≥0.7) | 9 (45%) | 1 (5%) | — | |
| Average | 0.68 ± 0.17 | 0.27 ± 0.21 | — | <.0001 |
| Mitral annular dimensions and area (end-systole) | ||||
| Anteroposterior (cm) | 3.8 ± 0.8 | 3.2 ± 0.6 | 2.1 ± 0.3 | .012 |
| Transverse (cm) | 5.4 ± 0.8 | 4.7 ± 0.7 | 3.8 ± 0.6 | .004 |
| Annular area (cm 2 ) | 18.7 ± 5.7 | 13.7 ± 3.7 | 8.2 ± 2.3 | .003 |
| Surgical information | ||||
| MV surgery | 20 (100%) | 2 (11%) | 0 | <.001 |
Figure 4 shows an example of images obtained in a patient with a normal MV, including a 3D rendered en face view of the MV from the left atrial perspective (“surgeon’s view”), a 2D cut plane that depicts the leaflet coaptation line and leaflet edge position with respect to the annular plane, and the corresponding color-coded parametric model depicting intact coaptation line and no leaflet displacement above the annular plane (no red/orange coloring of the leaflets).
