Quantification of Mitral Valve Anatomy by Three-Dimensional Transesophageal Echocardiography in Mitral Valve Prolapse Predicts Surgical Anatomy and the Complexity of Mitral Valve Repair


Three-dimensional (3D) transesophageal echocardiography (TEE) is more accurate than two-dimensional (2D) TEE in the qualitative assessment of mitral valve (MV) prolapse (MVP). However, the accuracy of 3D TEE in quantifying MV anatomy is less well studied, and its clinical relevance for MV repair is unknown.


The number of prolapsed segments, leaflet heights, and annular dimensions were assessed using 2D and 3D TEE and compared with surgical measurements in 50 patients (mean age, 61 ± 11 years) who underwent MV repair for mainly advanced MVP.


Three-dimensional TEE was more accurate (92%–100%) than 2D TEE (80%–96%) in identifying prolapsed segments. Three-dimensional TEE and intraoperative measurements of leaflet height did not differ significantly, while 2D TEE significantly overestimated the height of the posterior segment P1 and the anterior segment A2. Three-dimensional TEE quantitative MV measurements were related to surgical technique: patients with more complex MVP (one vs two to four vs five or more prolapsed segments) showed progressive enlargement of annular anteroposterior (31 ± 5 vs 34 ± 4 vs 37 ± 6 mm, respectively, P = .02) and commissural diameters (40 ± 6 vs 44 ± 5 vs 50 ± 10 mm, respectively, P = .04) and needed increasingly complex MV repair with larger annuloplasty bands (60 ± 13 vs 67 ± 9 vs 72 ± 10 mm, P = .02) and more neochordae (7 ± 3 vs 12 ± 5 vs 26 ± 6, P < .01).


Measurements of MV anatomy on 3D TEE are accurate compared with surgical measurements. Quantitative MV characteristics, as assessed by 3D TEE, determined the complexity of MV repair.

Surgical mitral valve (MV) repair provides excellent long-term results in patients with severe mitral regurgitation (MR) due to MV prolapse (MVP). Current American College of Cardiology and American Heart Association guidelines advocate MV repair in asymptomatic patients with severe MR and normal left ventricular systolic function if the likelihood of successful repair exceeds 90%. Successful MV repair is related mainly to surgical expertise and MV anatomy. Anatomic predictors of lower likelihood of repair are involvement of the anterior leaflet, bileaflet involvement, posterior leaflet height, and the extent of mitral annular disease.

Transesophageal echocardiography (TEE) is considered the standard of care for the assessment of MV anatomy and is critical for the stratification of patients with MVP, because it determines both the choice of the surgeon as well as the timing of surgery. Real-time three-dimensional (3D) TEE is more accurate than two-dimensional (2D) TEE in identifying prolapse of single MV segments and in patients with advanced MVP. Three-dimensional transesophageal echocardiographic quantitative measurements of annular dimensions, billowing volume, and height have been studied in the perioperative setting. Although all these studies were performed in comparison with qualitative surgical findings, only very few studies have tested the accuracy of 3D TEE in quantifying MV anatomy compared with direct surgical measurements. The accuracy of 3D TEE in measuring leaflet height, and the clinical relevance of quantitative MV assessment for MV repair using 3D TEE, are unknown.

In this study, we compared the accuracy of 2D and 3D TEE in quantifying MVP. In particular, we tested the accuracy of 2D and 3D TEE in measuring leaflet height and annular diameters against detailed surgical measurements. Furthermore, we studied whether qualitative and quantitative MV assessment on 3D TEE correlated with the surgical technique.


Study Population

From October 2009 to November 2010, we prospectively enrolled 50 consecutive patients with severe MR due to MVP who underwent elective surgical MV repair at Toronto General Hospital (Toronto, ON, Canada). For patients without previous TEE ( n = 17), study TEE was performed at our echocardiography laboratory. For patients with preoperative TEE performed at external institutions ( n = 33), TEE was performed in the operating room. All patients underwent postoperative 2D transthoracic echocardiography before hospital discharge to assess for residual MR. The study was approved by the Toronto General Hospital research ethics committee. All patients gave written informed consent.


TEE was performed using a Philips iE33 platform and X7-2t real-time 3D transesophageal echocardiographic probes (Philips Medical Systems, Andover, MA). Experienced cardiologists or cardiac anesthetists followed a specific MV protocol as part of a comprehensive transesophageal echocardiographic examination. Two 3D imaging modalities were used: 3D zoom mode (“live” acquisition over two heartbeats) and full-volume mode (four to seven stitched heartbeats), as previously described.

Qualitative and Quantitative Image Analysis on 2D TEE

Two experienced echocardiographers (S.J., C.G.) blinded to the intraoperative findings analyzed the transesophageal echocardiographic studies several weeks after the operation. We used the eight-segment Carpentier nomenclature ( Figure 1 ) to localize prolapsed segments. Prolapse was defined as end-systolic displacement of the body of the MV leaflet ≥2 mm above the annular plane in either the commissural or long-axis view at end-systole. A leaflet was considered flail when the leaflet edge was pointing retrograde into the left atrium at end-systole, whether or not a ruptured chord was seen. Calcification of the mitral annulus had thick, highly echogenic annular deposits with distal shadowing. Leaflet height and anteroposterior and intercommissural annular diameters were measured in end-systole ( Figures 2 A– 2 C).

Figure 1

The eight-segment Carpentier MV nomenclature. (A) Schematic view of MV (adapted from Biaggi et al . ), (B) 3D transesophageal echocardiographic surgeon’s view, and (C) operative site, demonstrating side by side the eight MV segments as described by Carpentier et al . The oblique surgical perspective and annular distortion by suture suspension of the posterior annulus enlarges the appearance of the posterior segments (C) . A , Anterior; Ca , anterolateral commissural segment; Cp , posteromedial commissural segment; P , posterior.

Figure 2

Measurements on 2D TEE and 3D TEE and during surgery. Overview of measurements demonstrated on a schematic view of the MV (A) (adapted from Biaggi et al . ). (B–G) Example demonstration of measurements of segments A2 and P3. On 2D TEE, leaflet heights were measured in the long-axis view (segment A2) (B) and commissural view (segment P3) (C) . On 3D TEE, segments were first evaluated in predefined angled views (D) and later measured using MV quantification software (E) . During the operation, the surgeon used a Kelly clamp (and a ruler, not shown) to take the measurements (F,G) . A , Anterior; Ao , aortic valve; AP , anteroposterior annular diameter; AV , aortic valve; Ca , anterolateral commissural segment; COM , commissural annular diameter; Cp , posteromedial commissural segment; IT , intertrigonal.

Qualitative and Quantitative Image Analysis on 3D TEE

The 3D transesophageal echocardiographic loops were analyzed by a single observer blinded to the results of 2D TEE and the intraoperative findings (P.B.). Qualitative MV assessment ( Figure 2 D) involved our systematic approach showing the MV in standardized views at four different angles (“angled views”). Prolapse was assessed in end-systole and was defined as any movement of the leaflet body above the adjacent annulus. Flail was defined as described for the 2D images. Ruptured chordae were defined as independently mobile structures attached to the edge of a flail segment. Annular calcification was defined as a bright mass attached to the annulus independent of the adjacent leaflet. Quantitative MV assessment ( Figure 2 E) was performed offline, using MV quantification software (QLab version 7.1; Philips Medical Systems). The 3D intertrigonal distance was measured according to Suri et al.

Intraoperative Findings and Surgical Measurements

All patients were operated by a single surgeon (T.E.D.). Fibroelastic deficiency (FED) was defined as the presence of thin leaflets with single-segment prolapse and ruptured chord(s). Barlow’s disease (BD) was defined as prolapse of all segments with bulky, billowing leaflets. Between FED and BD, there is a wide spectrum of MV disease. To simplify this continuous disease spectrum, we divided our study population into groups with simple and advanced MVP. Simple MVP was considered to be prolapse of one segment only. Advanced MVP was defined as two or more prolapsed segments combined with the presence of at least one of the following findings: myxomatous changes (as visualized by the surgeon), excessive leaflet tissue, or relevant annular calcification.

A segment was considered to be prolapsed if the body of the segment fell behind the anterolateral commissure. In cases in which all segments prolapsed, the level of the annulus was used for reference. Annular calcification was identified visually and confirmed by detaching the leaflet and removing the calcium block during repair. A leaflet was defined as flail if there was a ruptured chord or if the chords were largely elongated during inspection. Leaflet height and the intertrigonal distance were measured using a caliper or a long Kelly clamp and a ruler in case of suboptimal exposure ( Figures 2 F and 2 G). The intertrigonal distance was considered the space between the right and left fibrous tissue trigone seen intraoperatively at the base of the anterior MV leaflet and adjacent to the MV commissures. The diameter of the mitral annulus was not measured, because the heart was flaccid due to cardioplegic arrest. The surgeon was aware of the qualitative 2D transesophageal echocardiographic findings on preoperative studies because they were part of the referral data on patients, and he was able to review 2D transesophageal echocardiographic images obtained perioperatively. However, the surgeon was blinded to the qualitative and quantitative 3D transesophageal echocardiographic MV assessment performed offline using the MV quantification software.

Analysis of Different Subgroups of Patients

To determine whether findings on 3D TEE correlated with the type of surgery performed, we divided the study population according to the number of prolapsed segments in three subgroups: simple MVP (one segment involved) and advanced MVP with two subgroups (two to four segments vs five or more segments involved). We compared findings on 3D TEE and the type of surgery performed among the subgroups, and we correlated the degree of disease with the extent of surgery.

Reproducibility of Measurements

The interobserver and intraobserver variability of 3D measurements was determined using 10 full-volume data sets analyzed by a second observer (M.M.). The same 10 patients were used for interobserver and intraobserver variability of the 2D measurements (S.J., C.G.).

Statistical Analysis

Continuous data are expressed as mean ± SD and categorical data as numbers and percentages. For each segment, we calculated sensitivity and specificity as well as accuracy (the sum of true-positive and true-negative results divided by the number of total segments) for the detection of MVP. The overall accuracy of 2D and 3D TEE for the detection of MVP for all segments was assessed as the percentage of patients with true-positive and true-negative findings for all eight segments. We used one-way analysis of variance with Duncan’s post hoc analysis (for normally distributed data) or the Kruskal-Wallis test (for data not normally distributed) for the comparison of measurements between 2D and 3D transesophageal echocardiographic and intraoperative findings, as appropriate. Categorical data were analyzed using χ 2 tests. P values < .05 were considered statistically significant.

Interobserver and intraobserver variability were assessed using multiple methods: Bland-Altman analysis, the correlation coefficient, the coefficient of variation, and the intraclass correlation coefficient. Statistical analyses were performed using SPSS releases 11.0 and 12.0 (SPSS, Inc., Chicago, IL).


Study Population

The study population consisted of 50 patients with a mean age of 61 ± 11 years ( Table 1 ). The average 3D full-volume frame rate was 29.6 ± 8.4 Hz. The quality of the 3D full-volume data sets was insufficient for MV quantification in four patients, but qualitative MVP assessment remained feasible using the 3D zoom data sets.

Table 1

Patient characteristics ( n = 50)

Variable Value
Age (y) 61 ± 11
Women 19 (38%)
Height (m) 1.68 ± 0.11
Weight (kg) 75 ± 16
BSA (m 2 ) 1.8 ± 0.2
Systolic BP at rest (mm Hg) 128 ± 17
Heart rate at rest (beats/min) 74 ± 13
Rhythm during TEE
Sinus rhythm 46 (92%)
Atrial fibrillation, n (%) 4 (8%)
Left ventricular ejection fraction (%) 62 ± 9
NYHA class
I 24 (48%)
II 19 (38%)
III 7 (14%)
Diabetes mellitus 4 (8%)
Arterial hypertension 23 (46%)
Atrial fibrillation 12 (24%)
Marfan disease 0 (0%)
Family history of MVP 2 (4%)

BP , Blood pressure; BSA , body surface area; NYHA , New York Heart Association.

Surgical Findings and Type of Surgery Performed

Twelve patients (24%) were found to have simple (one-segment) MVP, including three patients with FED and nine patients showing only mild myxomatous changes. Thirty-eight patients (76%) met our definition of advanced MVP, including six with BD. Bileaflet involvement was found in 30 patients (60%) ( Table 2 ). MVs were replaced in two patients with severe annular calcification and in one patient with extreme annular displacement. Of the 47 patients (94%) who underwent successful MV repair, all had insertion of Gore-Tex (W. L. Gore & Associates, Newark, DE) neochordae (mean, 15 ± 9 chords), and all underwent annuloplasty using Simplici-T (Medtronic, Inc., Minneapolis, MN) bands (mean length, 67 ± 11 mm). Twenty-two patients (44%) needed additional leaflet resection.

Table 2

Detection of MV pathology by 2D TEE and 3D TEE compared with surgical findings

Variable Surgical findings Sensitivity (%) Specificity (%) Accuracy (%)
Segment prolapse
P1 13 (26%) 85 100 84 97 84 98
P2 40 (80%) 100 100 80 90 96 98
P3 34 (68%) 76 100 88 81 80 94
A1 7 (14%) 100 100 88 95 90 96
A2 28 (56%) 79 100 100 100 88 100
A3 26 (52%) 73 100 96 96 84 98
Ca 7 (14%) 86 100 95 98 94 98
Cp 19 (38%) 53 95 97 90 80 92
Single-segment prolapse 12 (24%) 83 92 76 100 78 98
Anterior leaflet prolapse 34 (68%) 74 100 94 100 80 100
Bileaflet prolapse 30 (60%) 70 100 95 95 80 98
Prolapse of all eight segments 6 (12%) 100 100 95 98 96 98
Annular calcification 10 (20%) 70 56 88 98 84 90
Flail leaflet present 33 (66%) 85 94 71 88 80 92
Ruptured chord 31 (61%) 81 81 79 83 80 82

A , Anterior; Ca , anterolateral commissural segment; Cp , posteromedial commissural segment; P , posterior.

Predischarge 2D transthoracic echocardiography in the 47 patients with MV repair showed trivial, mild, and mild to moderate MR in 35 (75%), 11 (23%), and one (2%), respectively.

Qualitative MV Assessment by TEE

Although 3D TEE had sensitivity of 100% for all segments except the posterior commissure (95%), the sensitivity of 2D TEE was generally lower ( Table 2 ). The accuracy of 3D TEE for detecting anterior and bileaflet prolapse was 100% and 98%, respectively, and was higher compared with that of 2D TEE. The overall accuracy in detecting prolapsed segments (correct description of all eight segments in a given patient) was 50% for 2D TEE and 86% for 3D TEE. The lower overall accuracy for both 2D and 3D TEE was caused mainly by the lower accuracy of the recognition of prolapse in the posteromedial half of the valve (segments P3 and A3) and, in particular, in the posteromedial commissural segment (Cp).

Quantitative MV Assessment by TEE

Three-dimensional transesophageal echocardiographic measurements of segment heights did not differ significantly compared with intraoperative measurements ( Figure 3 ). However, 2D TEE significantly overestimated leaflet heights compared with surgical measurements by an average of 1.9 ± 0.7 mm (segment P1) and 2.1 ± 1.0 mm (segment A2). Two-dimensional TEE overestimated the anteroposterior annular diameter by 3.1 mm and underestimated the commissural annular diameter by 1.4 mm ( P < .01 for both measurements; Table 3 ) compared with 3D TEE.

Figure 3

Comparison of leaflet height among 2D TEE, 3D TEE, and surgery. Abbreviations as in Figure 1 . A , Anterior; P , posterior.

Table 3

Comparison of annular measurements between 2D and 3D TEE and surgical inspection

Measurement Distance (mm) r Bland-Altman bias
2D TEE 3D TEE Surgery
Annulus, anteroposterior 37.2 ± 5.3 34.3 ± 5.1 0.91 3.1 ± 6.4
Annulus, commissural 43.0 ± 7.1 44.6 ± 7.8 0.81 −1.4 ± 6.4
Intertrigonal 26.3 ± 3.8 26.9 ± 5.4 0.84 0.7 ± 5.9

Data are expressed as mean ± SD.

P < .001.

P < .01.

Relationship Between Degree of MVP and Type of Surgery

Patients with single-segment prolapse were older than those with five or more prolapsed segments (67 ± 8 vs 57 ± 13 years, P = .05), and none had relevant annular calcification. Compared with patients with single-segment prolapse, patients with advanced MVP had significantly longer P1 and P3 segments and larger annular dimensions but were less likely to have flail leaflets ( Table 4 ).

Table 4

Comparison of 3D transesophageal echocardiographic findings and surgery performed in three subgroups with increasing disease severity

Number of prolapsed segments
Simple MVP Advanced MVP
Finding 1 ( n = 12) 2–4 ( n = 25) ≥5 ( n = 13) P
Height of segment P1 (mm) 9.7 ± 1.8 11.2 ± 2.1 13.2 ± 4.1 .02
Height of segment P2 (mm) 17.3 ± 7.3 18.2 ± 5.7 19.3 ± 6.5 .79
Height of segment P3 (mm) 8.6 ± 2.5 11.0 ± 2.8 12.4 ± 3.5 .02
Height of segment A2 (mm) 20.7 ± 2.9 24.3 ± 5.3 24.6 ± 4.5 .06
Anteroposterior diameter (mm) 31.4 ± 4.5 34.1 ± 3.7 37.4 ± 6.4 .02
Commissural diameter (mm) 39.8 ± 5.9 44.2 ± 5.0 50.1 ± 10.3 .04
Intertrigonal distance (mm) 24.6 ± 3.3 25.5 ± 2.2 29.6 ± 5.2 .03
Flail 9 (75%) 19 (76%) 5 (38%) .05
Surgery performed
Leaflet resection 4 (33.3%) 12 (48%) 6 (46.2%) .69
Resection type
Triangular 4 (100%) 10 (83%) 1 (17%) .02
Quadrangular 0 (0%) 2 (17%) 5 (83%) .02
Simplici-T band size (mm) 59.7 ± 12.8 67.2 ± 9.4 72.2 ± 9.7 .02
Number of artificial chordae inserted 6.9 ± 2.8 12.0 ± 5.3 25.6 ± 5.7 <.01
MV replacement 0 (0%) 2 (8%) 1 (8%) .60

Only gold members can continue reading. Log In or Register to continue

Jun 7, 2018 | Posted by in CARDIOLOGY | Comments Off on Quantification of Mitral Valve Anatomy by Three-Dimensional Transesophageal Echocardiography in Mitral Valve Prolapse Predicts Surgical Anatomy and the Complexity of Mitral Valve Repair

Full access? Get Clinical Tree

Get Clinical Tree app for offline access