There are currently no standardized three-dimensional (3D) transesophageal echocardiographic (TEE) views of the interatrial septum and atrial septal defects (ASDs). Without a standardized approach, it is difficult to ascertain the important anatomic relationships (such as the location of the aortic rim of an ASD), to perform relevant measurements (such as the size of an ASD or the size of its rims), or to guide the deployment of catheters and devices during atrial septal closure.
Using a 3D TEE matrix-array transducer, 706 TEE studies were performed over a 14-month period. The purpose of the study was to develop a standardized protocol for anatomically correct orientation of 3D TEE images of the interatrial septum and ASDs.
Among 706 TEE studies, there were 23 patients with ASDs, representing 3.3% of the study population. Eighteen patients had secundum ASDs, two had primum ASDs, and three had sinus venosus ASDs of the superior vena cava type. A protocol for properly orienting 3D TEE images of the interatrial septum and ASDs was developed. When the images are acquired at an angle of 0°, the septum is properly oriented by the tilt-up-then-left maneuver. The initial 3D TEE image in first tilted up to reveal the right atrial side of the septum. Then the image is tilted 180° around its vertical axis to reveal the left atrial side of the septum; the aortic rim is on the left, the superior vena cava on the top, and the right-sided pulmonary vein ostia on the right side of the screen. For acquisitions at a higher angle, the rotate-left-in-z-axis maneuver is used. The image is first tilted up to reveal the right atrial side of the septum, as in the tilt-up-then-left maneuver. The image is then rotated counterclockwise in the z axis until the superior vena cave is at 12 o’clock. Finally, the image is tilted 180° around its vertical axis to reveal the left atrial side of the septum.
The use of standardized tilt-up-then-left and rotate-left-in-z-axis maneuvers enhances the diagnosis of ASDs, ascertains the important anatomic relationships of ASDs to surrounding structures, and facilitates communication between echocardiographers obtaining 3D TEE images and interventional cardiologists or cardiac surgeons performing ASD closures.
Three-dimensional (3D) transesophageal echocardiographic (TEE) imaging has been revolutionized by the introduction of a 3D-TEE probe with a matrix-array transducer with 3,000 elements. This is approximately a 50-fold increase in the number of imaging elements compared with a standard two-dimensional transesophageal echocardiographic probe, which typically has 64 elements. The basic principles and history of 3D echocardiography have been presented in detail elsewhere.
What is the major difference between two-dimensional (2D) and 3D TEE image acquisition? Briefly, a 2D TEE probe acquires a sector image whose dimensions are as follows: width of up to 90° in the lateral (azimuth) direction, depth of up to approximately 16 cm in the axial direction, and thickness (elevation) that is negligible. The 3D TEE matrix-array probe expands this concept by acquiring not just one but a series of 2D sector images along the elevation axis to create a 3D pyramidal data set referred to as a frustum. By convention, the lateral (azimuth) direction is encoded in red, the elevation direction in green, and the depth direction in blue.
Aside from the three axes of the pyramidal data set itself, there are also three axes of the 2D image used to display the pyramidal set on the screen. The horizontal axis runs from the left to the right edge of the screen. The vertical axis runs from the top to the bottom of the screen. Additionally, the axis that is perpendicular to the computer monitor on the ultrasound system and comes out of the screen toward the user at the time of image review is referred to as the z axis. Rotations along the z axis move in clockwise or counterclockwise direction and allow for placing 3D images into conventional views (such as the surgical view of the mitral valve or the anatomic orientation of the interatrial septum).
There are currently no standardized 3D TEE views of the interatrial septum and atrial septal defects (ASDs). Without a standardized approach, it is difficult to ascertain the important anatomic relationships (such as the location of the aortic rim of an ASD), to perform relevant measurements (such as the size of the ASD or the size of its rims), or to guide the deployment of catheters and devices during atrial septal closure.
The purpose of this study was to develop a standardized protocol for the placement of 3D TEE images in anatomically correct orientation.
Three-dimensional TEE studies were recorded with a commercially available ultrasound system (Philips iE33; Philips Medical Systems, Andover, MA) using a matrix-array 3D TEE probe (X7-2 t; Philips Medical Systems). Over a 14-month period, 3D TEE studies were performed in 706 individuals.
Three-dimensional TEE images were obtained in the following four modalities: (1) biplane imaging (a side-by-side display of a pair of 2D TEE images that are 90° apart), (2) full-volume imaging (the frustum is automatically subdivided by the ultrasound system in several slices; each slice is acquired over one cardiac cycle; individual slices are stitched together and displayed in a delayed nonlive fashion), (3) narrow-angle live 3D imaging (live imaging of a system-selected frustum segment measuring approximately 60° × 30° in lateral x elevation axes and having a full depth in the depth axis), and (4) wide-angle 3D zoom imaging (live imaging of a user-selected frustum segment measuring up to 85° × 85° in lateral x elevation axes but with a system-limited slice thickness in the depth axis).
Wide-angle 3D zoom appears to be the most useful 3D modality for visualizing the anatomy of the interatrial septum because it provides instantaneous (live) images of almost the entire interatrial septum. Although full-volume 3D imaging provides a wider view of the interatrial septum than the 3D zoom, it is not instantaneous and often suffers from stitching artifacts (misalignment of image slices obtained in separate cardiac cycles). Narrow-angle live 3D imaging is used extensively during percutaneous atrial septal closure procedures to visualize the tips and trajectories of various catheters and devices used in closing defects.
With reference to the manipulation of the 3D image on the ultrasound system, the word “tilt” is used in this report to refer to image movements around either the horizontal or the vertical axis of the image and the word “rotation” to refer to image movements in the z axis.
Because the raw 3D TEE images are often nonintuitive, a standardized approach for image acquisition and display of the interatrial septum and ASDs is needed. In this article, we propose two simple maneuvers for rapid orientation of 3D TEE images of the interatrial septum in proper anatomic orientations.
Among 706 3D TEE studies performed, we identified 23 patients with ASDs (3.3% of the study population). There were 18 patients with secundum ASDs, two with primum ASDs, and three with sinus venosus ASDs of the superior vena cava (SVC) type. After a trial-and-error optimization, we developed a simple protocol for the placement of 3D TEE images of the interatrial septum in anatomically correct orientation. We then used this protocol for imaging a variety of ASDs.
Preparations for 3D TEE imaging start with the acquisition of good 2D TEE images of the interatrial septum. In principle, the interatrial septum can be imaged at any 2D angle. Image acquisition at 0° and 90° is described first. Then the impact of image acquisition at intermediate 2D angles on the subsequent 3D TEE images is discussed. All image acquisition described in this article were performed on 3D TEE systems from Philips Medical Systems. However, the general principles should apply to 3D TEE systems that are being developed by other manufacturers.
Image Acquisition at 0° (Tilt-Up-Then-Left [TUPLE] Maneuver)
After obtaining a good midesophageal view of the interatrial septum at 0° in 2D mode, the 3D zoom mode is selected. Initially, a pair of biplane images appears on the screen. The left image represents the lateral ( azimuth or red ) plane, while the right image represents the elevation ( green ) plane. In each image, there is a region-of-interest selection box; the user can select the sector width in azimuth or elevation planes by changing the left and right borders of the box using the track ball. Additionally, by moving the selection box up and down the screen, the user determines which portion of the depth ( blue ) plane will be displayed in the subsequent 3D view.
After the region of interest is selected, the 3D zoom view is obtained. The image of the interatrial septum appears in the same orientation as the 2D image at 0° ( Figure 1 A); that is, the interatrial septum is seen in its long axis rather than en face. We refer to this initial image as the “opening scene.”
In the next step, the image is manipulated to obtain the en face view of the interatrial septum from either the left or the right atrial perspective. This view, which is unobtainable in real time by any imaging technique other than 3D echocardiography, demonstrates the interatrial septum the way an anatomist or a surgeon would view it.
One way to obtain the en face view of the interatrial septum from the opening scene 3D zoom image is to tilt the image down around its horizontal axis. Although this maneuver indeed provides an en face view of the interatrial septum from the left atrial perspective, it places the superior rim of the interatrial septum at the bottom of the screen. In essence, it creates an upside-down image of the interatrial septum.
Instead, we propose the TUPLE maneuver to provide more intuitive en face images of the interatrial septum. In this maneuver, the initial 3D image of the interatrial septum seen in Figure 1 A is first tilted up along its horizontal axis to reveal an en face view of the interatrial septum from the right atrial perspective. The superior portion of the interatrial septum is now on the top of the screen and the anterior portion on the right side of the screen ( Figure 1 B).
The SVC and the ascending aorta are the most important landmarks in this view; by identifying and properly orienting these two vessels, one ensures proper orientation of the interatrial septum. The SVC is at the top of the screen, and the aortic valve and the ascending aorta are on the right side of the screen. The fossa ovalis, located in the middle of the image, is another important landmark, because its visualization is essential for guiding transseptal puncture during various percutaneous cardiac interventions.
In the next step, the image of the interatrial septum is tilted to the left around the vertical axis by approximately 180° until the en face view of the interatrial septum from the left atrial perspective is obtained. In this view, the superior rim of the interatrial septum remains at the top of the screen; the anterior (aortic) rim is now on the left side and the ostia of the right-sided pulmonary veins on the right side of the monitor ( Figure 1 C). The imaging of a normal interatrial septum using the TUPLE maneuver is illustrated in Video 1 ( view video clip online).
Image Acquisition at 90° (TUPLE Plus Rotate-Left-in-z-Axis [ROLZ] Maneuver)
Using the same principle of initial imaging (selection of region of interest in the biplane midesophageal view), the opening scene 3D zoom image is obtained. Again, this initial 3D image has the same orientation as the equivalent 2D image, which represents the bicaval view.
In the first step, the image is tilted up to reveal the right atrial side of the interatrial septum. In comparison with the image obtained at 0°, the SVC now appears on the right side of the screen. In effect, changing the 2D angle at time of image acquisition leads to 3D zoom image rotation in the z axis. Therefore, to orient this 3D image to the proper anatomic position, one must rotate the image to the left (counterclockwise) by 90° in the z axis.
Once the image of interatrial septum from the right atrial perspective is properly oriented, it is tilted to the left around its vertical axis to obtain the view of the interatrial septum from the left atrial perspective. We refer to this manipulation of the image as the TUPLE-plus-ROLZ maneuver. It is illustrated in Video 2 ( view video clip online).
Imaging at Intermediate Angles
As a general rule, the larger the 2D angle used at the time of image acquisition, the more rotation of the 3D image in the counterclockwise direction in the z axis will be needed. Images that were acquired at 0° in two dimensions will require no rotation in the z axis of the 3D images, images acquired at 75° will require a 75° counterclockwise rotation in the z axis, images acquired at 100° will require a 100° counterclockwise rotation in the z axis, and so forth. The impact of image acquisition at various 2D angles is shown in Figure 2 .
The TUPLE and ROLZ maneuvers were successful in obtaining diagnostic images in all 23 patients with ASDs. On average, the TUPLE maneuver with or without the ROLZ maneuver takes 1 to 2 minutes to perform.
3D TEE Imaging of Secundum ASDs
In our series, there were 18 patients with secundum ASDs, 14 of whom were women (72%; all secundum ASDs). Using the TUPLE and ROLZ maneuvers, one places a secundum ASD in the proper anatomic orientation. Figure 3 and Video 3 ( view video clip online) demonstrate imaging of a secundum ASD using the TUPLE maneuver and a 0° acquisition angle.
Complete resorption of the septum primum over the fossa ovalis leads to the classic appearance of the secundum ASD ( Video 4 ; view video clip online). Three-dimensional TEE imaging clearly demonstrates that these ASDs are frequently not perfectly circular, as has been commonly assumed in 2D echocardiography, but often oval or irregular in shape. In addition, what is assumed to be an ASD’s diameter on 2D TEE imaging is often a geometric chord rather than the true ASD diameter. (A geometric chord is a line that connects two points of a circle but does not pass the circle’s center, as its diameter does.)
With 3D TEE imaging, one can also appreciate that the size of an ASD varies throughout the cardiac cycle, being maximal during ventricular systole and minimal during atrial systole. In our laboratory, we report the size of the ASD at the time of its maximal diameters. The precise measurement of ASD size helps avoid inappropriate patient selection for percutaneous ASD closure.
A properly oriented 3D TEE image of a secundum ASD allows for evaluation of the size of the ASD tissue rims and their relationship to the aortic valve and the ascending aorta ( Figure 4 ). Knowledge of the defect size and the size of tissue rims is of utmost importance for percutaneous device closure of secundum ASDs. Percutaneous closure of a secundum ASD is usually considered feasible is the largest ASD diameter is <35 mm, its aortic rim is ≥3 mm, and other rims are >7 mm. Using the TUPLE and ROLZ maneuvers, the location and the size of the each rim in general and the aortic rim in particular can usually be determined.
At present, no measurements of ASD size can be obtained directly from 3D TEE images. Instead, the size of an ASD can be determined (1) semiquantitatively by superimposing a rectangular grid of known dimensions over the 3D TEE ASD image ( Figure 5 A) and (2) quantitatively by tracing the outlines and diameters of the ASD using offline software for multiplanar reconstruction.