Specialized Echocardiography Applications

Image Acquisition

3D imaging uses a complex multiarray transducer that simultaneously acquires ultrasound data from a 3D pyramidal volume. Rapid parallel image processing provides ultrasound images that can be viewed in real time in any orientation on the screen ( Fig. 4.2 ). These matrix array transducers typically include about 3000 piezoelectric elements with a transmission frequency of 2 to 4 MHz for transthoracic echocardiography (TTE) and 5 to 7 MHz for transesophageal echocardiography (TEE). Several approaches exist to the acquisition of echocardiographic data using a 3D matrix array transducer ( Table 4.1 ):

• ■

  Real-time narrow 3D section: A beat-by-beat view with a wider image plane than standard 2D imaging that can be rotated to view from different perspectives. It looks like a “thick” tomographic image.

• ■

  Real-time 3D-zoom volume-rendered images: A full-volume image of an enlarged area of interest that is rotated to show the structure of interest in a “surgical” view. These images are displayed with a perspective-type image similar to a photographic view from inside the heart.

• ■

  Full-volume gated acquisition volume-rendered images: Multiple-beat volumetric imaging stitches together narrow volumes of data over several cardiac cycles to provide a full volume of data that can be rotated and cropped to show the structures of interest.

• ■

  Simultaneous multiplane mode: This simultaneous display of two 2D image planes has the ability to adjust the rotation angle, tilt, and elevation of the second image plane.

• ■

  3D color Doppler imaging: This uses real-time or full-volume color Doppler data acquisition, but at frame rates lower than for imaging data.

Advantages of real-time narrow sector imaging are rapid image capture, familiar image planes, and evaluation of complex anatomy; however, only a narrow field of view is seen ( Fig. 4.3 ). With focused wide section or real-time “zoom” mode, the entire structure (e.g., the mitral valve) is included in the image, but spatial resolution and temporal resolution are poor, and the image must be rotated and gain carefully adjusted to display the internal cardiac anatomy ( Fig. 4.4 ). Full-volume gated images look similar to real-time zoom-mode images but have better spatial and temporal resolution. Full-volume images also can be analyzed after acquisition to provide additional views. Simultaneous multiplane imaging shows only a few (typically two) tomographic images but provides the highest temporal and spatial resolution within these image planes ( Fig. 4.5 ). The use of 3D color Doppler imaging is helpful for showing the spatial distribution of a flow disturbance, such as prosthetic paravalvular regurgitation or an intracardiac shunt, but it currently has very low temporal resolution.

During the acquisition of 3D images, transducer position is adjusted to optimize visualization of the structure of interest, for example, by imaging the mitral valve from the left atrial (LA) side on TEE imaging with the ultrasound beam perpendicular to the closed mitral leaflets. Next, gain and compression are set in the mid-range (about 50 units), and the time-gain compensation (TGC) curve is adjusted so the image is slightly “overgained” to avoid echo dropout appearing as “holes” in anatomic structures. With real-time imaging, transducer position and gain can be adjusted iteratively to improve image quality and to center the structure of interest in the image. My practice is to optimize position and gain on a zoomed real-time 3D view before the acquisition of a four-beat gated full-volume data set from the same transducer position. Postprocessing, gain, and compression then can be adjusted after image acquisition. With full-volume gated acquisitions, any change in heart position from beat to beat results in a vertical line across the image with misregistration of the image data on both sides of this “stitch” artifact. Causes of a stitch artifact include patient movement, respiratory motion, and an irregular heart rhythm.

Image Display

There are currently several types of 3D echocardiographic image displays, including:

• ■

  Volume-rendered 3D images

• ■

  Surface-rendered images

• ■

  Wireframe images

• ■

  Simultaneous display of multiple 2D images

• ■

  Graphic displays of 3D parameters versus time

In both the real-time 3D zoom mode and in full-volume imaging, the display is “cropped” to show different views of the interior structures of the heart. For example, the mitral valve can be viewed from the perspective of the LA; this provides a compelling view of prolapsing segments of the valve in patients with myxomatous mitral valve disease (see Fig. 12.29 ). The image can then be rotated and recropped to show a long-axis–type image of the mitral valve or to view the valve from the left ventricular (LV) side. Similarly, the aortic valve can be viewed en face from the perspective of the aorta, a view that correlates closely with the surgical view of valve anatomy, from the LV side of the valve or in a long-axis orientation. Real-time 3D images are cropped and rotated as the images are acquired. Full-volume gated acquisitions can be cropped and rotated during the exam but also can be reevaluated later because the full-volume data set is saved digitally.

Surface-rendered images or wireframe displays are based on identifying the boundaries of a cardiac structure, either by using semiautomated methods or by tracing the boundaries on multiple 2D images. For example, the LV endocardial surface is shown as a 3D solid structure with contraction shown by a sequence of 3D volumes over the cardiac cycle, so the rendered volume appears to beat on the display screen ( Fig. 4.6 ). Alternatively, a wireframe-type display can be used. A graphic display also is helpful with time on the horizontal axis and with either LV volume or the position of each myocardial segment shown on the vertical axis.

The 3D echocardiographic data set also can be used for simultaneous display of multiple-image 2D planes ( Fig. 4.7 ). The ability to acquire LV images in multiple planes simultaneously speeds image acquisition during stress echocardiography, thus potentially improving diagnostic accuracy. In addition, the ability to “move through” a 3D data set in any 2D image plane allows better appreciation of cardiac anatomy in patients with complex structural heart disease and allows precise localization of abnormalities.

Examination Protocol

With TTE imaging, only limited 3D imaging is performed to complement a full 2D study, depending on the patient's diagnosis and the reason for the study. Examples of 3D imaging on transthoracic imaging include quantitation of LV volumes and ejection fraction in a patient with heart failure (see Fig. 9.4 ), 3D measurement of mitral orifice area in a patient with mitral stenosis, or 3D short-axis images of the aortic valve in a patient with calcific aortic valve disease (see Fig. 11.6 ). With TEE imaging, a systematic approach to 3D image acquisition and display is recommended, with additional views as needed depending on the specific pathology ( Table 4.2 ).

Recommendations for volume-rendered 3D image displays ( Fig. 4.8 ) are:

• ■

  Aortic valve: The right coronary cusp is located inferiorly (at the 6 o'clock position) for both aortic and LV views of the valve (see Fig. 3.21 ).

• ■

  Mitral valve: The aortic valve is located at the top of the image so the anterior mitral leaflet is superior to the posterior leaflet for both LA and LV views of the valve (see Fig. 3.23 ).

• ■

  LV : 3D TTE views of the LV are oriented like standard 2D images in either an apical four-chamber view (apex at the top of the image, LV on the right side of the screen) or a short-axis view.

• ■

  Right ventricle (RV): A four-chamber view or short-axis view is oriented with the LA superior (12 o'clock position).

• ■

  Pulmonic valve: The anterior valve cusp is located superiorly (12 o'clock position) for both the pulmonary artery and RV sides of valve.

• ■

  Tricuspid valve : The ventricular septum is placed inferiorly for both the right atrial (RA) and RV views of the valve.

• ■

  Interatrial septum: From the LA side, the right upper pulmonary view is shown in the 1 o'clock position. From the RA side, the superior vena cava is at the 11 o'clock position.

• ■

  LA appendage : This display shows the LA appendage en face from the LA perspective with the pulmonary veins shown superiorly or longitudinally.

Quantitation From 3D Images

In addition to volume-rendered images of each chamber and valve, surface-rendered images of the LV are derived from a gated full-volume acquisition with the transducer positioned at the LV apex (for TTE imaging) or in a TEE four-chamber view. A 2D image is used to ensure optimal positioning of the transducer with the entire LV included in the sector scan. Gain and transducer frequency are adjusted to optimize endocardial definition. Acquisition of the gated full-volume data set is guided by a split screen display of orthogonal views, and the patient is asked to suspend respiration to minimize stitch artifacts. Once the full-volume data is acquired, the LV apex and mitral annulus are used as landmarks to initiate the edge detection process. The operator then can adjust the automated tracings as needed to follow the endocardial border accurately. As for 2D measures of LV volumes, trabeculations and the papillary muscles are included in the LV chamber to avoid underestimation of LV volumes. The surface-rendered image data then is used for quantitation of:

• ■

  LV end-diastolic and end-systolic volumes

• ■

  LV ejection fraction

• ■

  LV regional wall motion

Each of these parameters can be displayed on a 3D perspective color-coded LV shape or as a graph over the cardiac cycle (see Fig. 8.7 ).

Compared with 2D approaches, 3D quantitation of LV function avoids geometric assumptions and is more accurate and reproducible and thus is recommended when technically feasible (see Chapter 6 ). The 3D measures of LV mass, regional strain, curvature, and wall stress are more complicated and are currently investigational approaches.

Other quantitative measurements from 3D data sets are in evolution. Standard 3D volume-rendered image displays show the cutaway view of the heart as a solid structure using shading and lighting to provide the impression of a 3D perspective on a 2D viewing screen. This display is not conducive to quantitative measurements because only two of the three dimensions are shown. Advances in display and digital processing should alleviate this problem by allowing accurate measurement of distances and areas.

Potentially other 3D measurements have advantages compared with 2D measurements for nonplanar structures, such as a stenotic valve. For example, although experienced operators can accurately measure mitral valve area from 2D images aligned at the minimal orifice area in patients with rheumatic mitral stenosis, inexperienced operators show improved accuracy with 3D imaging, which reliably shows the stenotic orifice and is less dependent on transducer position or image plane positioning. For planimetry of mitral valve area, the 3D volume is acquired, and then a 2D plane is aligned at the minimal orifice with the valve opening traced in mid-diastole ( Fig. 4.9 ). In research applications, more complex structures, such as the mitral leaflets and annulus, can be reconstructed in 3D by tracing the structure in a series of 2D image planes within the 3D volumetric data set ( Fig. 4.10 ).

Clinical Utility

The clinical role of 3D echocardiography will continue to evolve as this technology matures. In addition to providing more detailed anatomic relationships and more accurate quantitation, 3D images are more intuitive than 2D images, thus allowing quicker appreciation of cardiac anatomy by more health care providers ( Table 4.3 ). Potentially, 3D echocardiography could be faster than 2D scanning and could reduce variability in image acquisition. However, because instrumentation is in development, 3D echocardiography is not yet a routine part of the clinical examination at all centers and typically is used to supplement the 2D study in selected patients, with imaging focused on a specific anatomic structure. The use of 3D imaging is more widespread for intraoperative and intraprocedural imaging because of the improved image quality and the additive value of the 3D perspective in these clinical settings (see Chapter 18 ).

The American Society of Echocardiography and European Association of Echocardiography guidelines recommend routine use of 3D imaging for:

• ■

  Quantitation of LV volumes and ejection fraction

• ■

  Evaluation of mitral valve anatomy (valve area in mitral stenosis)

• ■

  Guidance of transcatheter procedures

It is likely that other quantitative applications will become available in the near future, including quantitation of RV volumes and ejection fraction and 3D evaluation of aortic valve, outflow tract, and aortic sinus anatomy in adults with valvular aortic stenosis. Further studies are needed for other potential applications including 3D dyssynchrony, strain imaging, and the evaluation of prosthetic valves.

The use of 3D volume-rendered imaging has proved to be helpful in several clinical settings, both for facilitating communication with other physicians and for providing more detailed anatomic information about shape, size, and 3D anatomic relationships of structures. The benefits of 3D echocardiography for specific clinical settings include:

• ■

  Myxomatous mitral valve disease: Evaluation of the number and severity of prolapsed or flail segments and identification of chordal rupture is used for planning surgical repair (see Fig. 12.29 ).

• ■

  Atrial septal defects: Visualization of the location, size, and suitability is used for transcatheter closure (see Figs. 17.19 and 17.20 ).

• ■

  Transcatheter interventions: 3D imaging is used for guidance during procedures, evaluation of procedural results, and detection of complications (see Fig. 18.24 ).

The 3D color Doppler applications are challenging because of the low frame rates with this modality. Currently, 3D color Doppler is helpful in identifying the location of paravalvular regurgitation. Other potential clinical applications of 3D imaging, such as quantitation of valvular regurgitation based on 3D visualization of proximal jet geometry, require further validation.

Limitations

Although 3D imaging has greatly expanded the capability of echocardiography for the visualization of complex heart disease, this approach does have some limitations. Acquisition of 3D images can be time-consuming, particularly because 3D imaging currently serves as an adjunct, not a replacement, for 2D imaging. However, 3D imaging modalities likely will become more integrated into the standard clinical exam when the instrument interface allows effortless transitions between 2D and 3D imaging and more intuitive approaches to image manipulation. Current display formats attempt to show 3D images on 2D displays; this limitation should be resolved as 3D display systems become more widely available. As with all ultrasound modalities, the direction of the ultrasound beam relative to the structure of interest affects image quality; resolution is optimal in the axial direction for structures perpendicular to the ultrasound beam. In addition, ultrasound artifacts, such as shadowing, reverberations, and poor penetration affect the image, as with any ultrasound modality. Many patients with suboptimal 2D images also have poor 3D images. TEE 3D imaging tends to be much more useful than TTE 3D imaging. Finally, both spatial and temporal resolutions of 3D imaging are inferior to those of 2D imaging, so both modalities are needed for a full imaging study.

Tissue Doppler Strain and Strain Rate

Doppler blood flow velocity measurements are based on backscatter of low-amplitude, high-velocity signals from moving blood cells ( Fig. 4.11 ). In contrast, Doppler tissue velocity measurements are based on the high-amplitude, low-velocity signals reflected from the myocardium. Thus these signals are easily separated by adjusting the gain, wall filters, and velocity scale of the Doppler spectral or color display.

Tissue Doppler velocity recording at a specific intracardiac site is analogous to pulsed Doppler blood flow velocity recordings. Tissue velocity measurements depend on a parallel alignment between the ultrasound beam and the direction of myocardial motion; in other words, motion is measured only in the direction toward and away from the transducer. For example, a component in evaluation of diastolic function is the tissue Doppler signal recorded in the apical four-chamber view with a 2-mm sample volume positioned about 1 cm apical from the septal side of the mitral annulus ( Fig. 4.12 ). The spectral display is recorded at a velocity range of ±0.2 m/s, using very low gain and wall filter setting. The Doppler velocities show systolic motion of the myocardium toward the apex, corresponding to the apical motion of the annulus in systole seen on 2D imaging. In diastole, an early diastolic motion away from the apex (E′) occurs, corresponding to the early phase of diastolic filling, and a late diastolic motion away from the apex (A′) occurs, corresponding to the atrial phase of ventricular filling.

Strain rate imaging is based on the difference in tissue Doppler velocity (V) between sample volumes divided by the distance (D) between them ( Fig. 4.13 ). This measures the rate of change in myocardial length, normalized to the original length. Strain rate (SR) then is:

The units of strain are seconds ⁻¹ (or /s) because the velocity measured in centimeters per second is divided by the distance in centimeters. Typically, strain rate is measured in the apical-base direction, in the apical four-chamber view with three sample volumes placed in the septal or lateral wall myocardium about 12 mm apart. The tissue Doppler mean velocity curves are examined to ensure a clear signal without excessive noise, lack of aliasing, and avoidance of blood pool signals (see Fig. 4.12 ). The instrument calculates strain rate from these velocity curves for each time point and displays strain rate in seconds ⁻¹ as a function of time. The stain rate curve looks like a vertical mirror image of the velocity curve because myocardial shortening is a negative strain and myocardial lengthening is a positive strain. Strain rate provides data on relative timing of myocardial motion and peak systolic and diastolic strain rates. Peak systolic strain rate is a measure of ventricular contractile function that is insensitive to changes in loading conditions.

Strain is a measure of deformation of a material, defined as the difference between the final length ( l ) and the original length ( l _o ), divided by the original length. Thus strain can be thought of as the percentage change in length:

Strain can be estimated from the tissue Doppler strain rate by integrating the curve over time.

Thus strain is analogous to ejection fraction (i.e., change in volume normalized to initial volume) with the advantage that spatial localization and temporal localization are possible. In fact, a graph of strain over the cardiac cycle ( Fig. 4.14 ) looks similar to a ventricular volume curve. Because strain is relative to the baseline length, end-diastole is considered zero strain. During systole strain decreases rapidly until end-systole is reached. Isovolumetric relaxation and contraction result in a slight flattening of the curve just before and after systole. In diastole, a rapid increase in strain occurs during the early phase of diastolic filling (E) , followed by a plateau during diastasis and then another increase with atrial contraction (A) back to the baseline at end-diastole. Peak systolic strain is a measure of regional ventricular function. However, like ejection fraction, strain varies with preload.

Accurate measurements of Doppler strain rate and strain require careful attention to technical aspects of data recording. The sample volumes must fit within the myocardium at an adequate distance from each other. In addition, velocity is measured only in the direction toward and away from the transducer. Signal quality is enhanced by the use of harmonic imaging, an adequate pulse repetition frequency, a high frame rate, and by tracking the sample volume to the ventricular wall. The Suggested Reading section provides further details about data acquisition and interpretation.