Three-dimensional (3-D) echocardiography is an exciting development with the potential to have a major impact on the practice of perioperative transesophageal echocardiography (TEE). While imaging the heart with 3-D TEE is intuitive, operators experienced in two-dimensional (2-D) imaging often find difficultly orientating and manipulating the 3-D datasets. Therefore, to understand 3-D TEE, it is important to know how the images are formed, manipulated, and displayed. Currently, there are two commercially available 3-D TEE systems: reconstructive 3-D TEE from Siemens (with analysis performed on TomTec software) and (2) live 3-D TEE from Philips Medical Systems (with analysis performed on QLAB software). Because only real-time 3-D imaging has practical applicability in the operating room and intensive care unit (ICU), the emphasis in this chapter is on live 3-D from Philips. However, live 3-D imaging systems from other manufactures will shortly be available. The use of real-time 3-D imaging during cardiac surgery has been comprehensively reviewed.
Image formation
Reconstructive three-dimensional imaging
Reconstructive 3-D imaging is performed using a standard, multiplane TEE probe. Automatic steering is used to obtain a series of 2-D slices at different degrees of sector rotation from the same transducer position ( Figure 4-1 ). The operator determines the angle (3 or 5 degrees) that the transducer rotates between slices. An electrocardiogram (ECG) is used to trigger the acquisition of each 2-D slice; therefore, an irregular heartbeat, ectopy, or poor ECG tracing causes artifacts. Acquisition of the dataset takes 30 to 60 heartbeats. Off-line, the 2-D images are “stitched” together to form a single 3-D image ( Figure 4-2 ). Movements of the patient or TEE probe cause stitching artifacts. Stitching artifacts can be reduced by suspending ventilation during image acquisition and by holding the probe in a stable position. Because image acquisition and data analysis are time consuming, and because 3-D reconstruction must be performed off-line, the technique is not suited to perioperative use and is not discussed further.
Live three-dimensional imaging
Live 3D TEE provides instantaneous real-time imaging of the heart and therefore overcomes the practical limitations inherent in reconstructed 3-D imaging. Live 3D imaging is achieved due to an advance in transducer technology known as a matrix array. A standard, multiplane, phased array transducer provides only one plane of imaging at a time ( Figure 4-3 ). By contrast, a matrix array transducer functions in a similar way to a shower head, giving a 3-D “shower” of ultrasound in both the lateral and the elevation planes ( Figure 4-4 ). This is achieved by having a matrix of 2500 elements, which all function instantaneously to give a moving 3-D image in real time. Currently, only one matrix array TEE transducer is commercially available, the X7-2t from Philips. This transducer is similar in size to a standard, multiplane, phased array TEE transducer, so probe insertion is identical to that for standard 2-D imaging.
A standard 2-D sector has two axes: sector width (lateral plane or x-axis) and depth (height or y-axis). 3-D imaging also has elevational width (elevation plane or z-axis), which is the thickness of the sector ( Figure 4-5 ). The dimensions in the lateral and elevational planes, at a given depth, define the volume of the sector. Sector dimensions (lateral and elevational planes) are expressed in degrees. Somewhat confusingly, degrees also define sector rotation (0 to 180 degrees) for standard, multiplane, 2-D imaging.
Spatial (image quality) and temporal (frame rate or smoothness of motion) resolution are influenced by the sector width, elevational width, and depth. The best spatial and temporal resolutions are achieved by imaging with small 3-D volumes in the near field. However, small 3-D volumes display less anatomy and make image orientation more difficult. Frame rates with 3-D imaging vary depending on the imaging mode used (described later) but in general are lower than frame rates with standard 2-D imaging, which is typically more than 40 Hz. Very low frame rates confer a staggering quality to cardiac motion. 3-D imaging can be optimized for spatial or temporal resolution.
3-D datasets can be displayed on the screen in one of three ways: wireframe, surface rendering, or volume rendering ( Figure 4-6 ). Wireframe and surface rendering are most frequently used for quantifying ventricular volumes and function. For other 3-D TEE applications, volume rendering is used, which displays a virtual dissection of the structure of interest.
Three-dimensional imaging modes and image acquisition
Before starting 3-D imaging, it is important to ensure a good ECG trace is displayed and the 2-D image is optimized. The 2-D image can be optimized by activating the iScan function, adjusting the individual image controls, or both (see Chapters 1 and 2 ). The (TGC) and (LGC) controls should be moved to the center position before using the iScan function, as this allows greater control of these parameters during 3-D scanning.
xPlane mode
xPlane mode is an advanced form of 2-D imaging, in which two orthogonal 2-D images are displayed side by side. The sector rotation of the right-hand image can be adjusted by moving the trackball. In addition, the right-hand image can be adjusted in the elevation plane using the tilt function, as shown in Figure 4-7 . xPlane mode has a lower frame rate than standard 2-D imaging, particularly when color flow Doppler is used.
Three-dimensional live mode
Three-dimensional we provides a real-time, narrow-angle 3-D sector ( Figure 4-8 ). The elevational width can be set at either 1 degree (thin slice) or 30 degrees (thick slice). The maximum sector dimensions are 60 degrees (lateral plane) by 30 degrees (elevational plane). The frame rate is between 20 and 30 Hz. The 3-D image can be rotated to any orientation on the screen. However, because only a relatively thin 3-D wedge is displayed, physical movements of the probe, using standard TEE views, are required to fully image the heart. This mode is useful for rapid scanning of the heart and for helping to appropriately position the probe and adjust the sector rotation and imaging depth before using 3-D zoom and full-volume modes.
Three-dimensional zoom mode
Three-dimensional zoom mode provides a real-time magnified 3-D sector (see Figure 4-5 ) of up to 90 degrees (lateral plane) by 90 degrees (elevational plane). Selecting 3-D zoom displays two orthogonal 2-D images ( Figure 4-9 ), which are used for adjusting the size and position of the 3-D sector to focus on the structure of interest. The imaging plane (i.e., width or x-axis and height or y-axis) is adjusted on the left-hand 2-D image; the elevation plane (z-axis) is adjusted on the right-hand image. Selecting “acquire” stores the 3-D zoomed dataset.
Once the 3-D dataset has been stored, it can be cropped and orientated to view the structure of interest from the desired viewpoint, as described later. 3-D zoom mode can be used with the autocrop function turned on or off (explained later).
3-D zoom mode has a low frame rate (10 to 15 Hz with larger sector volumes) but has excellent spatial resolution. Thus, 3-D zoom mode provides detailed imaging of the structure and function of the entire mitral valve (MV) and is therefore the very useful during perioperative TEE. However, color flow Doppler cannot be used in this mode (color flow Doppler is only active in 3-D full-volume mode, as explained later).
Three-dimensional full-volume mode
3-D full-volume mode provides a large volume image of the heart, with sector dimensions of up to 90 degrees (lateral plane) by 90 degrees (elevational plane), at variable depth. Selecting “full volume” provides two orthogonal 2-D images, which display the sector dimensions. Selecting “full volume” again acquires the full-volume 3-D dataset over a set number of cardiac cycles. Unlike live 3-D and 3-D zoom modes, 3-D full-volume mode is not a real-time imaging modality. The full-volume 3-D dataset is acquired over four or seven cardiac cycles from subvolumes that are acquired during each cardiac cycle ( Figure 4-10 ). The subvolumes are stitched together (rendered), synchronized to one cardiac cycle, and then displayed on the screen. This 3-D dataset can be displayed in full (“blob view”; Figure 4-11 ) or with the autocropping 3-D sector ( Figure 4-12 ), which is more easily recognizable (as described later). Autocrop is active by default in 3-D full-volume mode.