Background
Left ventricular outflow tract (LVOT) measurement is a critical step in the quantification of aortic valve area. The assumption of a circular morphology of the LVOT may induce some errors. The aim of this study was to assess the three-dimensional (3D) morphology of the LVOT and its impact on grading aortic stenosis severity.
Methods
Fifty-eight patients with aortic stenosis were studied retrospectively. LVOT dimensions were measured using 3D transesophageal echocardiography at three levels: at the hinge points (HP) of the aortic valve and at 4 and 8 mm proximal to the annular plane. Results were compared with standard two-dimensional echocardiographic measurements.
Results
Three-dimensional transesophageal echocardiography showed a funnel shape that was more circular at the HP and more elliptical at 4 and 8 mm proximal to the annular plane (circularity index = 0.92 vs 0.83 vs 0.76, P < .001). Cross-sectional area was smaller at the HP and larger at 4 and 8 mm from the annular plane (3.6 vs 3.9 vs 4.1 cm 2 , P = .001). The best correlation between two-dimensional and 3D transesophageal echocardiographic dimensions was at the HP (intraclass correlation coefficient = 0.75; 95% CI, 0.59–0.86). When the HP approach was selected, there was a reduction in the percentage of patients with low flow (from 41% to 29%).
Conclusions
A large portion of patients with aortic stenosis have funnel-shaped and elliptical LVOTs, a morphology that is more pronounced in the region farther from the annular plane. Two-dimensional LVOT measurement closer to the annular plane has the best correlation with 3D measurements. Measurement of the LVOT closer to the annular plane should be encouraged to reduce measurement errors.
Highlights
- •
Three-dimensional echocardiography in patients with AS shows a funnel-shaped morphology of the LVOT in >80% of patients.
- •
The size and circularity of the LVOT differ significantly on the basis of the point of measurement.
- •
More than 70% of patients have ellipse-shaped LVOTs at the HP of the aortic cusps.
- •
The best correlation between 2D TTE and 3D TEE dimensions is obtained when measurement occurs close to the annular plane.
Echocardiography is the standard imaging modality for the evaluation and management of patients with aortic stenosis (AS). Effective aortic valve area (AVA) and transvalvular gradients are the principal measures for grading stenosis severity. Left ventricular outflow tract (LVOT) measurement is a critical step in the quantification of AVA by the continuity equation. Current recommendations on echocardiography advise measuring the LVOT on the longitudinal long-axis plane, at mid-systole and at 4 to 8 mm proximal to the annular plane. More recently, some other investigators have suggested 2 mm as the recommended point for measuring the LVOT. Furthermore, the main limitation pointed out for this method is that the continuity equation assumes a circular morphology of the LVOT, whereas several studies have demonstrated its elliptical shape, and the frequent underestimation of AVA with this assumption compared with planimetered area. As a result, there are currently some concerns and controversies regarding the most appropriate method to measure the LVOT.
Three-dimensional (3D) transesophageal echocardiography (TEE) allows real-time volume rendering of images and offline volumetric quantification techniques. This method allows exploration of the morphology of the LVOT and measurement of its dimensions at any distance from the annular plane, without making geometric assumptions. Previous studies using this method confirmed the elliptical shape of the LVOT, but all of them were limited to the analysis of a single plane next to the insertion of the aortic valves. Nevertheless, the LVOT is a more complex structure that extends from the annulus of the aortic valve to 8 or 12 mm into the left ventricle. The present study was conducted to assess the morphology and size of the LVOT using 3D TEE along its 8 mm proximal to the annular plane and its impact on AVA estimated by the continuity equation.
Methods
A retrospective study was conducted, including 58 patients from a series of 62 consecutive patients referred to our echocardiography laboratory for preoperative evaluation between January 2010 and May 2012. Transthoracic echocardiography (TTE) and 3D TEE were performed as part of the protocol for patients who were candidates for transcatheter aortic valve implantation. Two-dimensional TTE and 3D TEE were performed with a time interval of <20 days between them. The local clinical research ethics committee approved the protocol, and informed consent was obtained for all patients. In case of poor image quality, patients were excluded from the analysis.
2D TTE
Two-dimensional TTE was performed using an iE33 ultrasound system equipped with an S5-1 phased-array transducer (Philips Medical Systems, Andover, MA). A systematic imaging protocol was performed by European Association of Cardiovascular Imaging–accredited cardiologists closely following current guidelines. The LVOT was measured by two independent observers on a zoomed longitudinal 2D image, in a mid-systolic frame, using the inner edge–to–inner edge technique and at the hinge points (HP) of the aortic valve. Referent LVOT measurement was performed 4 mm and 8 mm proximal to the annulus, in accordance with guideline recommendations. Each observer was free to choose any point as long as it was within 4 to 8 mm set in the LVOT. Doppler flow data were acquired from the LVOT, 2 to 4 mm below the annulus, in pulsed-wave mode from the apical five-chamber view. The position of the sample volume was moved distally in the LVOT, toward the aortic valve, until a much steeper rise was obtained, as recommended. Pulsed Doppler spectral strips of LVOT systolic flow were manually traced on the modal curve. Aortic valve flows were obtained from all accessible acoustic positions, including the right parasternal window, to ensure maximal transaortic velocity measurement. The effective AVA was estimated using the continuity equation, averaging measurements from three different cardiac cycles (10 cardiac cycles if the patient was in atrial fibrillation) and assuming a circular shape of the LVOT. Image processing was performed offline using an image management system (Xcelera; Philips Medical Systems).
3D TEE
Three-dimensional TEE was performed using the iE33 ultrasound system equipped with a matrix-array X7-2t transducer (Philips Medical Systems). Sedation with intravenous propofol (1 mg/kg followed by 0.5 mg/kg every 3–5 min) was used to minimize distress. Real-time 3D imaging of a pyramidal 60° × 30° data set of the aortic valve and LVOT was obtained. Settings were optimized using narrow-angled acquisition mode to ensure a volume rate up to 25 Hz. QLAB (3DQ module) (Philips Medical Systems) was used for postprocessing of images. In a mid-systolic frame (defined as maximal aortic valve opening on 3D transesophageal volume-rendered images), two orthogonal planes with their lines parallel to the longitudinal edge were set through the LVOT. A third plane was traced perpendicular to the previous two planes. This plane was reoriented as needed to ensure the most circular orifice area of the slot. Using the multislice tool of the software, we drew different cuts of the LVOT. First, we selected the minimum number of planes for the software tool, four planes. We adjusted the first plane to coincide with the HP of the aortic sigmoid cusps. Next, we adjusted the distance between the different planes to 4 mm. In this way, we obtained four cuts of the LVOT: at the HP of the aortic leaflets, at 4 mm proximal to the annular plane, and also at 8 mm proximal to the annular plane. The fourth plane, the farthest from the aortic valve, usually passes through the proximal part of the anterior mitral leaflet, so it was not considered for LVOT analysis ( Figure 1 ). A subsequent selection of the appropriate picture allowed us to measure the diameters (anteroposterior and medial-lateral) and cross-sectional area (CSA) by planimetry of the LVOT ( Figure 2 ). LVOT morphology was evaluated by means of the circularity index dividing anteroposterior (short) by medial-lateral (large) diameters. A circularity index of 1 would represent a perfect circle, while a progressively lower value would represent a more ellipsoid geometry. The 3D assessment of AVA was made using the continuity equation and combining the Doppler velocity-time integral (VTI), measured in the LVOT, with the LVOT CSA determined at the three levels.
Data were reanalyzed in the whole cohort of patients with the same images and videos, and 2D transthoracic echocardiographic LVOT dimensions were remeasured by the same cardiologist and by a second observer in a blinded fashion, with a time interval of 2 to 7 days. A different observer blinded to the results of 2D TTE performed 3D image processing offline.
Statistical Analysis
Normal distribution of continuous variables was assessed using the Kolmogorov-Smirnov test (Lilliefors correction). Variable magnitudes are described as proportions, mean ± SD, or median (interquartile range) as appropriate. Differences between variables were assessed using the paired t test, analysis of variance, or the Mann-Whitney U test as appropriate. Differences between proportions were assessed using the χ 2 test. Agreement between observers and imaging modalities was assessed using the intraclass correlation coefficient (ICC). Two-tailed P values <.05 were considered to indicate statistical significance. Analysis was performed using SPSS version 15.0 (SPSS, Chicago, IL).
Results
Four patients were excluded from the analysis because of inadequate images. The demographic and echocardiographic characteristics of the 58 patients included are depicted in Table 1 . AVA was <1 cm 2 in 50 patients and between 1 and 1.3 cm 2 in the remaining eight patients.
| Variable | Value |
|---|---|
| Age (y) | 74.2 ± 8.4 |
| Women | 28 (48%) |
| Weight (kg) | 73.9 ± 15.2 |
| Height (cm) | 160.3 ± 8.5 |
| Body surface area (m 2 ) | 1.77 ± 0.19 |
| Atrial fibrillation | 11 (19%) |
| Interventricular septal thickness (mm) | 14.6 ± 2.9 |
| LV posterior wall thickness (mm) | 12.9 ± 2.1 |
| LV diastolic diameter (mm) | 44.5 ± 8.3 |
| LV systolic diameter (mm) | 28.5 ± 9.4 |
| LVEF (%) | 58.7 ± 14.8 |
| Left atrial volume index (mL/m 2 ) | 39.5 ± 18.9 |
| Peak gradient (mm Hg) | 69.8 ± 27.1 |
| Mean gradient (mm Hg) | 42.1 ± 17.3 |
| SVi (mL/m 2 ) | 38.1 ± 13.2 |
| AVA, 2D TTE (cm 2 ) | 0.72 ± 0.24 |
| AVAi, 2D TTE (cm 2 /m 2 ) | 0.41 ± 0.12 |
Three-Dimensional Transesophageal Echocardiographic Morphology of the LVOT
LVOT morphology assessed by 3D TEE showed a funnel shape ( Figure 1 ). LVOT CSA was smaller at the level of the HP of the aortic valve and larger when measured at 4 and 8 mm proximal to the annular plane (3.76 ± 0.9 vs 4.05 ± 1.1 vs 4.46 ± 1.3 cm 2 , P < .001). Only 11 patients (18.9%) had CSAs of the LVOT that were larger at the HP of the aortic valve than measured at 8 mm proximal to the annular plane. These patients had increases of interventricular septal thickness (16.2 ± 4.2 vs 14.2 ± 2.2 mm, P < .001) and interventricular septum/left ventricular posterior wall ratio (1.3 ± 0.4 vs 1.1 ± 0.2, P < .01) and higher VTIs at the LVOT (27.2 ± 9.7 vs. 20.2 ± 4.5 cm, P < .05). CSA of the LVOT was more circular at the level of the HP of the aortic valve and more elliptical when measured at 8 mm proximal to the annular plane ( Figure 3 A). In fact, 30% of patients showed circular shapes at the HP, but only 3% did so at 8 mm ( Figure 3 B). LVOT anteroposterior diameter dimension was similar when measured by 3D TEE at any level. In contrast, LVOT lateral diameter increased from the HP of the aortic valve as measures were made more distal to the annulus (23.1 ± 3.3 vs 24.9 ± 3.3 vs 27.3 ± 3.9 mm, P < .001; Table 2 ).
| Variable | 2D TTE | 3D TEE | |||
|---|---|---|---|---|---|
| LVOT | HP | HP | 4 mm | 8 mm | |
| APD (mm) | 19.8 ± 2.3 ∗ | 20.8 ± 2.3 † | 20.8 ± 2.7 | 20.1 ± 3.7 † | 20.4 ± 3.6 † |
| LD (mm) | — | — | 23.1 ± 3.3 | 24.9 ± 3.3 ∗ | 27.3 ± 3.9 ∗ |
| CIn | — | — | 0.91 ± 0.09 | 0.81 ± 0.1 ∗ | 0.75 ± 0.09 ∗ |
| CSA (cm 2 ) | 3.18 ± 0.8 ∗ | 3.44 ± 0.8 ∗ | 3.76 ± 0.9 | 4.05 ± 1.1 ∗ | 4.46 ± 1.3 ∗ |
| SVi (mL/m 2 ) | 38.1 ± 13.2 ∗ | 41.2 ± 14.1 ∗ | 45.1 ± 17.2 | 47.8 ± 15.9 ∗ | 51.9 ± 17.1 ∗ |
| AVA (cm 2 ) | 0.72 ± 0.24 ∗ | 0.79 ± 0.29 ∗ | 0.86 ± 0.32 | 0.92 ± 0.32 ∗ | 1.0 ± 0.35 ∗ |
| AVA < 1 cm 2 | 50 (86%) ∗ | 48 (83%) ∗ | 42 (72%) | 34 (58%) ∗ | 32 (55%) ∗ |
| SVi < 35 ml/m 2 | 24 (41%) ∗ | 17 (29%) ∗ | 15 (25%) | 12 (21%) ∗ | 8 (14%) ∗ |
∗ P < .01 compared with 3D transesophageal measurement at HP.
Two-Dimensional TTE: Annular Plane versus LVOT
LVOT anteroposterior diameter was larger when measured at the HP than when measured more distal to the annular plane at the reference point (20.8 ± 2.3 vs 19.8 ± 2.3 mm, P < .01). Also, the values of the derived calculations were higher when diameter was measured at the HP: stroke volume index (SVi) (41.2 ± 14 vs 38.1 ± 13.2 mL/m 2 , P < .01) and AVA (0.79 ± 0.29 vs 0.72 ± 0.24 cm 2 , P < .001) ( Table 2 ).
2D versus 3D Measurements
The best correlation between 2D TTE and 3D TEE was at the HP level (ICC = 0.75; 95% CI, 0.59 to 0.86; P < .001), whereas at the reference LVOT, the correlation with 3D TEE was very poor (ICC = 0.35; 95% CI, −0.21 to 0.57; P = NS).
Reclassification of AS Severity
With TTE, when HP was selected, the number of patients with AVA < 1 cm 2 was reduced from 50 (86%) to 48 (83%) ( P < .01), and the number of patients with low flow (SVi < 35 mL/m 2 ) decreased from 24 (41%) to 17 (29%) ( P < .001) ( Figure 4 )
