Data are lacking on the utility of real-time three-dimensional (3D) echocardiography (RT3DE) in congenital abnormalities of the atrioventricular (AV) valves. The purpose of this study was to determine whether transthoracic RT3DE is superior to combined transthoracic echocardiography and two-dimensional (2D) transesophageal echocardiography in determining mechanisms and sites of AV valve regurgitation in congenital heart disease.
Between January 2005 and November 2007, 48 consecutive patients were studied prior to AV valve repair (22 left AV valves and 26 tricuspid valves) using 2D transthoracic echocardiography, 2D transesophageal echocardiography, and transthoracic RT3DE. Ages ranged from 24 days to 30 years. The 2D data were reviewed by blinded observers, and the real-time 3D data by a separate observer. In all patients, surgical findings were documented by a surgical report, while in 40, video recordings were also available. Surgical findings were used as the reference standard for structural abnormalities; RT3DE was the reference standard for the site of AV valve regurgitation.
Compared with 2D echocardiography, RT3DE provided superior detail of the mural leaflet and anterior commissural abnormalities for the left AV valve. For the tricuspid valve, improved detection of leaflet abnormalities, prolapse of the anterior and posterior leaflets, and commissural pathology was observed by RT3DE. Apart from a central location, surgical saline testing correlated poorly with jet location on RT3DE.
RT3DE provides complementary information as to the mechanisms and sites of AV valve failure in congenital heart disease.
A precise understanding of the mechanisms of atrioventricular (AV) valve regurgitation is both desirable and essential prior to surgical repair. Although two-dimensional (2D) transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) with color Doppler have become the standard mode of investigation, they suffer from limitations, as each attempts to provide a three-dimensional (3D) mental image from a series of 2D slices. In levocardia and AV concordance, left-sided AV valve regurgitation may be seen in the setting of an AV septal defect or a morphologic mitral valve (MV). Left AV valve regurgitation is the major cause of late morbidity in patients after AV septal defect repair. As well, either congenital or acquired MV regurgitation can significantly affect outcomes. On the right side of the heart in levocardia and AV concordance, tricuspid valve (TV) regurgitation in association with congenital or acquired heart disease is associated with a poorer prognosis, with mortality being related to increasing severity of regurgitation.
Three-dimensional echocardiography using both rotational devices and, more recently, matrix-array technology has mainly been used in adults in the evaluation of valve morphology and function in patients with MVs or left AV valves in AV septal defects and TVs. Few data are available in those with congenital heart disease, in particular an understanding of the role that transthoracic 3D matrix-array technology plays in the evaluation of AV valve pathology and function.
Our objective was to compare real-time 3D echocardiography (RT3DE) with combined 2D TTE and 2D TEE with regard to the mechanisms and sites of AV valve regurgitation.
We consecutively enrolled all patients in our congenital program with planned surgical repair of an AV valve between January 2005 and December 2007 at the Stollery Children’s Hospital at the University of Alberta. Our program includes both children and adults with congenital heart disease. This retrospective study was approved by the Human Research Ethics Board at the University of Alberta.
All patients underwent complete 2D TTE and RT3DE 1 to 30 days prior to AV valve surgery. Of note, both studies were performed during the same echocardiographic examination. Prior to surgery, all but 2 patients underwent complete 2D transesophageal echocardiographic examinations. Two patients were considered too small (2.0 and 2.2 kg) for 2D TEE.
Complete 2D transthoracic echocardiographic and color Doppler studies were performed in all patients using a Philips iE33 (Philips Medical Systems, Bothell, WA) with an S5-1, S8-3, or S12-4 probe. Two-dimensional TTE included all standard parasternal and apical views, with an additional posterior to anterior sweep in the 4-chamber view during both image and color Doppler acquisition. The 2D transthoracic studies were performed by one of our experienced sonographers, using our standard protocol for AV valve evaluation.
RT3DE was performed at the end of the 2D transthoracic echocardiographic examination, using either a matrix X3-1 or X7-2 transthoracic probe. Of note, at the time of the study, we did not have the matrix transesophageal echocardiographic probe. Full-volume data sets were acquired by either J.F.S. or K.T., depending on who was available. The protocol included full-volume analysis using the high–frame rate mode, with the narrowest sector size to provide an image with optimal valve detail. For all left AV valve patients, a data set was acquired from both the apical and parasternal long-axis view. In those with TV, pathologic images were obtained from an apical or slightly medial 4-chamber view. The acquisition was performed during a breath hold at end-expiration in patients aged > 3 years. For children aged < 3 years, the study was performed during quiet respiration after the administration of oral sedation with chloral hydrate, to minimize stitch artifacts. All real-time 3D data sets were obtained both in grayscale and with color Doppler imaging. The raw data sets were initially stored on a DVD and later transferred back to the iE33 for reconstruction. In all cases, later reconstruction of the valve (volume-rendering imaging) allowed a complete visualization of all segments of the left AV or TV.
Two-dimensional TEE was performed just prior to the operation, after the induction of anesthesia and endotracheal intubation. Probe size was dependent on the patient’s weight (S7-3t or SS7-2 OmniPlane probe for the Philips iE33). Two-dimensional transesophageal echocardiographic and color Doppler evaluation included a 4-chamber view with a sweep from posterior to anterior, as well as sweep from 0° to 120° to evaluate all aspects of the leaflets.
In all cases, a detailed surgical report was available for review. We had initially attempted to record all of the procedures on digital video, being successful in 40, but in 8 patients for technical reasons the camera was not available. A diagram ( Figure 1 ) was provided to the surgeon to complete after the operation. The surgeons used their reports as well as the video recordings in those for which they were available.
Reconstruction of 3D Data Sets
En face views of the left AV or TV from the left or right atrium and left or right ventricle were reconstructed to determine the precise anatomy of the valve leaflets and the mechanisms and sites of regurgitation. The mechanisms and sites of regurgitation were determined by obtaining an en face view of the valve with superimposition of the vena contracta(s). All 2D transthoracic and transesophageal images were analyzed by two experts in 2D echocardiography (J.D.D. for the left AV, M.R. for the TV) blinded to the surgical or real-time 3D findings. The 2D transthoracic and transesophageal images were reviewed in unison, because this represents the current practice prior to AV valve surgery. We did not design the study to compare 2D TTE and 2D TEE with each other. Three-dimensional data sets were reconstructed and analyzed by two experienced observers (J.F.S. for the left AV, K.T. for the TV). In 30 cases, 3D reconstruction was performed on the iE33 ultrasound machine using imported stored data from the original DVD ≥3 months after the date of RT3DE to minimize the influence of prior 2D and surgical findings. For the remaining 18 cases, 3D reconstruction was performed on the iE33 machine by a reviewer who was unaware of the 2D findings prior to the surgery. The quality of the 3D reconstructed images, on the basis of the clarity of the AV valve anatomy and the presence or absence of artifacts throughout the cardiac cycle, was rated as adequate or poor.
Definition of Terms Used for Analysis
Structural abnormality of the leaflet: This was defined as a leaflet with thickening, tears, or deformation (other than prolapse).
Leaflet definitions: Figure 1 demonstrates the accepted leaflet nomenclature used to describe the left-sided AV valve component in an AV septal defect, as well as for a normal MV and TV. In an AV septal defect, there are 3 left-sided leaflets: the superior and inferior bridging leaflets and a smaller mural leaflet. Of note, the mural leaflet has a similar embryologic derivation as the posterior leaflet of the MV; therefore, for simplicity, we refer both as the mural leaflet. For the TV, there is an anterior, a posterior, and a septal leaflet.
Commissural abnormality: This was defined by the line of coaptation between the leaflets, excluding the so-called cleft between the superior and inferior bridging leaflet, which was dealt with separately. The commissure was considered abnormal if it appeared to be irregular in nature or deficient. For those with AV septal defects, there was a superior bridging-mural leaflet or inferior bridging-mural leaflet commissure, while for those with normal MVs, it was the commissure at A1/P1 or A3/P3 (terms routinely used to arbitrarily describe the leaflets of the MV). For those with TVs, it was between the anterior-septal, anterior-posterior, or posterior-septal leaflets.
Poor coaptation: This was defined as any area of the left AV valve or TV at which there was an obvious gap between the leaflet edges that could be observed during systole.
Prolapse: This was defined as leaflet tissue that extended past the annulus in systole. Using 2D TTE and 2D TEE, this was documented for the left AV from the long-axis view and for the TV from the 4-chamber image. With RT3DE, it was possible to use multiple imaging views for both the left AV and TV.
Cleft (for the left AV in patients with AV septal defects): This represented the space between the superior and inferior bridging leaflet, which may extend from the site that the two leaflets cross the interventricular septum, to the points of coaptation at their free edges.
TV leaflet immobility: This represented the septal leaflet that appeared to be tethered down to the septum by the supporting apparatus.
The reviewers recorded data on an information sheet that had a diagram of either the left AV or TV.
Structural Abnormalities of the Leaflets
For the left AV valve, abnormalities of the aortic and mural leaflets were documented in patients with morphologic MVs. For those with AV septal defects, abnormalities of the superior and inferior bridging and mural leaflets were noted. For the TV, anterior, posterior, and septal leaflet abnormalities were documented.
Mechanisms of Regurgitation
This was recorded as a result of (1) prolapse of the leaflet, (2) a commissural abnormality, (3) poor leaflet coaptation, (4) a primary or residual cleft (for left AV valve regurgitation in patients with AV septal defects), or (5) tricuspid septal leaflet tethering.
Location and Number of Regurgitant Jets
These were documented on the schematic diagram of either the left or right AV valve. The real-time 3D color Doppler findings were used as the reference standard for regurgitant jet assessment, because this is a sensitive technique for the detection of regurgitation, while surgical evaluation of an AV valve during saline testing is not under physiologic conditions.
For the left AV valve, there were very few patients with morphologic MVs, compared with those with AV septal defects. Therefore, for analysis, the combination of the left-sided component of superior bridging and inferior bridging leaflets was considered as one and assigned as equal to the anterior leaflet of the MV, with both being called the anterior leaflet. The cleft in the left AV valve was considered separately. The mural leaflets in AV septal defects or posterior leaflets in morphologic MVs were considered equally. As well, findings on 2D TTE and 2D TEE were combined for analysis, because the combination represents the current standard approach to AV valve assessment prior to surgery. The sensitivity, specificity, and accuracy of 2D and real-time 3D echocardiographic evaluation for leaflet structure and mechanisms of regurgitation were calculated using surgical findings as the reference standard. The sensitivity, specificity, and accuracy of 2D echocardiographic and surgical evaluation for the sites of the regurgitant jets were calculated with real-time 3D color Doppler as the reference standard. Color Doppler was chosen as the reference standard because it is a sensitive method to detect AV valve regurgitation and is obtained under physiologic conditions with the beating heart. Analyses were performed using Statcel 2003 (OMS, Saitama, Japan).
Between January 2005 and December 2007, 54 patients with planned AV valve repair were studied. Of these, the real-time 3D data sets of 6 (mean age, 4.8 years; mean weight, 16.6 kg) were poor and therefore excluded from analysis. The remaining 48 patients (22 undergoing left AV valve repair and 26 undergoing TV repair) were analyzed.
Patients undergoing left AV valve repair (12 male, 10 female) had a median age of 9.2 years (range, 7 months to 25 years) and median weight of 29.6 kg (range, 3.6-76.0 kg). Patients undergoing TV repair (17 male, 9 female) had a median age of 8.3 years (range, 1 month to 33 years) and a median weight of 23.5 kg (range, 2.2-78.0 kg). Cardiac diagnoses are summarized in Table 1 , and surgical findings are documented in Table 2 . The median frame rate of the real-time 3D data sets was 42 Hz (range, 22-69 Hz), with the higher frame rates being obtained from the transthoracic X7-2 real-time 3D probe.
|Left AV valve regurgitation (n = 22)|
|Primum atrial septal defect (9 postoperative, 2 preoperative)||11|
|Complete AV septal defect (3 postoperative, 3 preoperative)||6|
|Patent ductus arteriosus||1|
|Aortic stenosis and insufficiency||1|
|Isolated mitral valve regurgitation||2|
|TV regurgitation (n = 26)|
|Hypoplastic left-heart syndrome||11|
|Tetralogy of Fallot||4|
|Ventricular septal defect||2|
|Transposition of great arteries, after Mustard operation||2|
|Atrial septal defect||1|
|Double-outlet right ventricle||1|
|Primum atrial septal defect (after primary repair)||1|
|Pulmonary atresia with intact ventricular septum||1|
|Left AV valve abnormality (n = 22)|
|Commissural abnormality||Between SBL and ML or A1 and P1||10|
|Between IBL and ML or A3 and P3||10|
|Prolapse||SBL, IBL, or AL ∗||6|
|Poor coaptation||At any position||6|
|TV abnormality (n = 26)|
|Commissural abnormality||Between ant and sept||14|
|Between post and sept||16|
|Between ant and post||16|
|Poor coaptation||At any position||17|
Table 3 summarizes the comparison of 2D echocardiography with RT3DE regarding leaflet abnormalities and mechanisms of left AV valve regurgitation. The sensitivity of RT3DE for leaflet abnormalities of the mural leaflet was higher than that of 2D echocardiography ( Figures 2 A and 2 B, [CR] ; view video clip online). For both the mural and anterior leaflets, RT3DE provided a greater degree of accuracy for detecting abnormalities of the leaflets. For commissural abnormalities between the superior bridging and mural leaflet, and between A1 and P1, RT3DE appeared to be more sensitive and very specific compared with 2D echocardiography. Although the sensitivities between 2D echocardiography and RT3DE were similar for the inferior bridging leaflet and the mural leaflet, or A3 and P3 commissural abnormalities, RT3DE was more specific and as well more accurate. Two-dimensional echocardiography was more sensitive for detecting abnormalities of coaptation than RT3DE. For other mechanisms of left AV valve regurgitation (prolapse and AV septal defect cleft), the accuracy of RT3DE and 2D echocardiography was similar. When all the parameters of the left AV valve were combined, RT3DE appeared to provide greater specificity and accuracy, but similar sensitivity to 2D echocardiography.
|Sensitivity (%)||Specificity (%)||Accuracy (%)|
|Left AV valve (n = 22)||2DE||RT3DE||2DE||RT3DE||2DE||RT3DE||P ∗|
|AL (n = 10)||30.0 (1.6-58.4)||40.0 (9.6-70.4)||69.0 (52.1-85.8)||93.1 (83.9-100.0)||59.0 (43.5-74.4)||79.5 (66.8-92.2)||.0001|
|ML (n = 4)||37.5 (4.0-71.0)||100.0 (100.0-100.0)||28.6 (4.9-52.2)||62.5 (38.8-86.2)||31.8 (12.4-51.3)||72.7 (54.1-91.3)||.01|
|SBL-ML or A1-P1 (n = 10)||30.0 (1.6-58.4)||50.0 (19.0-81.0)||20.0 (0.0-44.8)||100.0 (100.0-100.0)||25.0 (6.0-44.0)||75.0 (56.0-94.0)||.003|
|IBL-ML or A3-P3 (n = 10)||90.0 (71.4-100.0)||90.0 (71.4-100.0)||30.0 (1.6-58.4)||80.0 (55.2-100.0)||60.0 (38.5-81.5)||85.0 (69.4-100)||.22|
|Prolapse (n = 6)||53.6 (35.1-72.0)||39.3 (21.2-57.4)||90.9 (73.9-100.0)||100.0 (100.0-100.0)||64.1 (49.0-79.2)||56.4 (40.8-72.0)||1.0|
|Poor coaptation (n = 6)||66.7 (28.9-100.0)||16.7 (0.0-46.5)||60.0 (35.2-84.8)||80.0 (59.8-100.0)||61.9 (41.1-82.7)||61.9 (41.1-82.7)||1.0|
|Cleft AVSD (n = 17)||94.1 (82.9-100.0)||94.1 (82.9-100.0)||—||—||94.1 (82.9-100)||94.1 (82.9-100)||1.0|
|All parameters||59.6 (49.4-69.7)||59.8 (49.5-70.1)||53.9 (43.6-64.3)||85.7 (78.5-92.9)||56.7 (49.5-64.0)||73.0 (66.5-79.6)||<.0001|
Table 4 summarizes the results for the TV. RT3DE provided greater sensitivity and accuracy for detecting abnormalities of all the leaflets in comparison with 2D echocardiography. Commissural abnormalities between the anterior and septal and posterior and septal leaflet were detected by RT3DE more frequently than by 2D echocardiography. The detection of prolapse of the anterior and posterior leaflet, as well as poor coaptation as mechanisms of regurgitation, was also higher by RT3DE ( Figures 3 A and 3 B, Videos 3 and 4 ; view video clip online). As well, septal TV tethering was detected more frequently by RT3DE. Again, when combining all parameters, RT3DE provided greater sensitivity and accuracy compared with 2D echocardiography.
|Sensitivity (%)||Specificity (%)||Accuracy (%)|
|AL (n = 9)||0.0 (0.0-0.0)||88.9 (68.4-100.0)||82.4 (64.2-100.0)||58.8 (35.4-82.2)||53.8 (34.7-73.0)||69.2 (51.5-87.0)||.45|
|PL (n = 11)||9.1 (0.0-26.1)||81.8 (59.0-100.0)||93.3 (80.7-100.0)||93.3 (80.7-100)||57.7 (38.7-76.7)||88.5 (76.2-100)||.01|
|SL (n = 18)||38.9 (16.4-61.4)||94.4 (83.9-100.0)||28.6 (0.0-62.0)||71.4 (38.0-100.0)||36.0 (17.2-54.8)||88.0 (75.3-100)||.06|
|Between AL-SL (n = 14)||50.0 (23.8-76.2)||78.6 (57.1-100.0)||91.7 (76.0-100.0)||91.7 (76.0-100.0)||69.2 (51.5-87.0)||84.6 (70.7-98.5)||.29|
|Between PL-SL (n = 16)||37.5 (13.8-61.2)||87.5 (71.3-100.0)||70.0 (41.6-98.4)||70.0 (41.6-98.4)||50.0 (30.8-69.2)||80.8 (65.6-95.9)||.04|
|Between AL-PL (n = 16)||43.8 (19.4-68.1)||56.3 (31.9-80.6)||60.0 (29.6-90.4)||70.0 (41.6-98.4)||50.0 (30.8-69.2)||61.5 (42.8-80.2)||.39|
|Prolapse of AL (n = 16)||56.3 (31.9-80.6)||81.3 (62.1-100.0)||80.0 (55.2-100.0)||70.0 (41.6-98.4)||65.4 (47.1-83.7)||76.9 (60.7-93.1)||.55|
|Prolapse of PL (n = 8)||37.5 (4.0-71.0)||87.5 (64.6-100.0)||61.1 (38.6-83.6)||88.9 (74.4-100.0)||53.8 (34.7-73.0)||88.5 (76.2-100)||.006|
|Poor coaptation (n = 17)||29.4 (7.8-51.1)||82.4 (64.2-100.0)||44.4 (12.0-76.9)||77.8 (50.6-100.0)||34.6 (16.3-52.9)||80.8 (65.6-95.9)||.004|
|Tethered SL (n = 11)||54.5 (25.1-84.0)||72.7 (46.4-99.0)||73.3 (51.0-95.7)||73.3 (51.0-95.7)||65.4 (47.1-83.7)||73.1 (56.0-90.1)||.75|
|All parameters||39.4 (31.1-47.7)||81.8 (75.2-88.4)||72.4 (64.7-80.2)||78.7 (71.6-85.9)||55.6 (49.5-61.6)||80.3 (75.5-85.2)||<.0001|