Principles of Echocardiography
Physics and Instrumentation Sound is an energy form that travels through a medium as a series of alternating compressions and rarefactions of the molecules (Fig. 15–1). Sound is typically characterized by its wavelength, which is the distance between any two consecutive phases of the cycle (e.g., peak compression to peak compression), and by its frequency, which is the number of wavelengths per unit time [customarily expressed as cycles per second, or hertz (Hz)]. The velocity of sound is the product of wavelength and frequency; thus there is an inverse relationship between these two characteristics: the greater the frequency, the shorter the wavelength. Ultrasound is sonic energy with a frequency above the audible range of the human ear (greater than 20,000 Hz) and is useful for diagnostic imaging, since, like light, it can be directed as a beam that will obey the laws of reflection and refraction.12–14,17,18 Thus, an ultrasound beam will travel in a straight line through a homogeneous medium. If the beam meets an interface of different acoustic impedance, however, part of the energy will be reflected and the remaining attenuated signal will be transmitted. The reflected energy, or echo, is used to construct an image—in the case of echocardiography, an image of the heart (Fig. 15–2).
The most fundamental component of any echocardiographic instrument is the transducer, which is responsible for both transmitting and receiving the ultrasound signal. The transducer consists of electrodes and a piezoelectric crystal, whose ionic structure results in deformation of shape when exposed to an electric current.18 Thus, piezoelectric crystals are composed of synthetic materials, such as barium titanate, that, when exposed to electric current from the electrodes, alternately expand and contract to create sound waves. When subjected to the mechanical energy of sound returning from a reflecting surface, the same piezoelectric element changes shape, thereby generating an electrical signal detected by the electrodes (Fig. 15–3). Thus the transducer both produces and receives ultrasonic signals.
In the past, echographs have both transmitted and received signals of the same frequency. Recently, harmonic imaging has been implemented, in which ultrasound energy is transmitted at a baseline (fundamental) frequency but then received at a higher harmonic of that frequency (usually the first harmonic). Harmonic imaging is based upon the change in the ultrasound frequency of a transmitted wave induced by the interaction with a reflecting target. The sinusoidal waveform becomes peaked as it travels through tissue, thereby undergoing a change in frequency. Similarly if a sound signal strikes a contrast microbubble, the periodic expansion and contraction (resonation) of the bubble will change the frequency of the wave. This change in frequency generates harmonic signals with frequencies which are multiples of the transmitted signal. A simplistic analogy of this phenomenon would be the morphologic changes seen in ocean waves as they approach the shore and are affected by the rising ocean floor. The cresting of the waves and the changes in their height are analogous to the harmonic signals generated by the interaction of ultrasound and tissue.19 Of note, these signals take some time (and distance) to develop. Therefore (and in sharp distinction to reflected fundamental ultrasound energy), structures very close to the transducer do not generate much harmonic signal at all. Thus, near-field artifacts and reverberation artifacts are minimized with tissue harmonic imaging. Harmonic imaging has also been very useful in conjunction with intravascular echo contrast agents. The microbubbles in these agents resonate during ultrasound imaging, and demonstrate cyclic expansion and constriction: this resonance produces a large amount of harmonic energy.20 In contrast, myocardial tissue does not resonate to any appreciable degree. The net effect of harmonic imaging with echocardiographic contrast is a marked enhancement of the signal from the left ventricular (LV) cavity compared to that of the myocardium, and thereby an improvement in endocardial definition. As an imaging modality, ultrasound presents several unique technical difficulties. Sound energy is poorly transmitted through air and bone, and the ability to record adequate images is dependent upon a thoracic window that gives the interrogating beam adequate access to cardiac structures. The degree to which ultrasonic energy will be reflected when it strikes an interface of differential impedance depends on how perpendicular the interrogating beam is to the interface. When the ultrasound beam is directed to the interface, little or no sound energy will be reflected to the transducer. Therefore poor signal transmission, a nonorthogonal orientation of the ultrasound beam to the surface, and energy attenuation can cause failure to record signals from cardiac structures—a phenomenon referred to as echo dropout.21 Conversely, some structures may be such strong ultrasonic reflectors—being perpendicular to the beam or extremely dense—that sufficient energy returns to the transducer to be reflected and again transmitted into the field. This phenomenon can lead to reverberations, or the reproduction of the echoes of anatomic structures at multiple locations within the image.22 Also, background noise artifacts, or signals generated from the system rather than tissue, can also be encountered. Finally, targets lying on the periphery of the ultrasound beam may be recorded and displayed as if they were located along the central scan line (Fig. 15–4). This problem may be accentuated in the setting of very strong reflectors that result in the formation of side lobes.23 Thus beam-width problems associated with ultrasound may result in the depiction of targets in erroneous locations and create problems in interpreting the images.24
The construction of a cardiac image from ultrasound signals is based on computation of the distance between an anatomic structure and the transducer (Fig. 15–3). Thus, an ultrasound beam is produced by a handheld transducer positioned on the thorax and directed into the heart. This beam will travel in a straight line until it reaches an interface between structures of different acoustic impedance, such as blood and myocardium. At this point, some ultrasonic energy will be reflected (depending on the density of interface); some will be scattered and some will continue forward. The amplitude of the propagating signal will be attenuated because of the reduction in energy at the interface (Fig. 15–2). The reflected sound waves return to the transducer and form the basis of the echogram. Electronic circuitry within the echograph measures the time interval required for the transit of the ultrasound beam from the transducer to the interface and back again. Since the velocity of sound in soft tissue is constant (approximately 1540 m/s), the instrument can calculate the total distance traveled to and from the reflecting surface as the product of transit time and velocity of sound. Interface location is derived as one-half of the total transit distance, and a signal is depicted on an oscilloscope or video monitor at that point (Fig. 15–3). The amplitude of ultrasonic energy reflected from each target interface is represented by the brightness of the signal that is displayed. The 1D ultrasonic B- (or brightness) mode scan line resulting from a single transmitted beam is the cornerstone of echocardiographic imaging. In the most basic form of echocardiography, a single scan line produced by a piezoelectric crystal is passed through the heart (Fig. 15–5). At each structural interface, ultrasonic energy is reflected back and displayed at the appropriate distance as a signal, whose amplitude represents the acoustic impedance or density of the material encountered. These signals are subsequently displayed as dots, whose brightness is proportional to the amplitude of reflected ultrasonic energy. The distance from the transducer of these B-mode dots changes as the cardiac structures move during the cardiac cycle. Accordingly, if repetitive B-mode scan lines are produced and swept across the screen over time, the movement of the heart can be obtained as a time-motion (or M-mode) recording,25 providing dynamic rather than merely static cardiac images (Fig. 15–5). In clinical use, the piezoelectric crystal within the transducer is activated by alternating electric current to transmit at a rate of approximately 1000 pulses per second. This same crystal also receives the returning echo reflections and actually spends most of the time (>90 percent) in the "receive" rather than the "transmit" mode. Because the beam is confined to a single location and transmits ultrasound signals at the pulse rate of the transducer, M-mode echocardiography provides very high temporal resolution. Importantly, M mode is an excellent modality for timing cardiac events or recording high-velocity motion.
As ultrasound technology advanced, multiple B-mode scan lines from different imaging angles were collected and displayed in proper alignment to create a 2D image. As opposed to B- or M-mode recordings, which are unidimensional (on an anteroposterior axis), 2D echocardiography provides additional information in either superoinferior or mediolateral directions. At present, M-mode recordings are derived from the 2D images rather than as a stand-alone signal. Several characteristics of sound energy are of fundamental importance in determining the quality of the images obtained. High-quality images require optimal resolution—that is, the ability to distinguish two individual objects separated in space. Short wavelengths yield excellent resolution in echo imaging, since the shorter the cycle length, the smaller the object that will reflect the signal and be detected by the echo scanner. Since wavelength is inversely related to frequency, transducers that emit a high-frequency signal (3.5 to 7.0 MHz or greater) yield high-resolution images. High-frequency signals also overcome a limitation of ultrasonic imaging associated with lateral resolution. Since ultrasonic beams diverge as they propagate away from the transducer, the width of the beam can become sufficiently great to encompass multiple targets and decrease resolution (Fig. 15–4). The degree of beam divergence is less with high-frequency sonic energy than with low-frequency signals. The smaller wavelengths associated with high-frequency signals, however, are subject to greater reflection and scattering, with substantially higher attenuation as the beam propagates through tissue. The resultant attenuation is greater and leads to decreased sensitivity. Therefore, in clinical practice, echocardiographic examinations are performed utilizing the highest-frequency transducer capable of obtaining signals from all potential targets within the ultrasound field.25 M-Mode Echocardiography The Standard M-Mode Examination Although largely supplanted by 2D imaging, M-mode echocardiography remains a useful part of a complete ultrasound examination. Figure 15–6A through D shows the typical views obtained when the transducer is placed at the left parasternal area and rocked through the heart from apex to base. Tissue typically reflects ultrasound at its surface (specular reflectors) and from internal inhomogenicity (backscatter), while blood is homogenous and does not produce reflections. Thus, blood is free of ultrasonic signals on the echocardiogram. At the mitral valve (MV) level (Fig. 15–6C), the cardiac structure seen closest to the transducer is the right ventricular (RV) free wall; it is followed by the RV cavity, the interventricular septum, the MV apparatus, and the LV posterior wall as the beam travels backward. At this level, MV excursion is well seen and is more easily recorded for the longer anterior leaflet. For the anterior leaflet, diastolic mitral opening is bipeaked (M-shaped), with maximal opening during early diastolic filling at the E point, a subsequent reclosure downslope to the F point, and a reopening with atrial contraction at the A point prior to valve closure at the C point26 (Fig. 15–7). The posterior leaflet manifests a mirror-image W-shaped pattern. When LV end-diastolic pressure is elevated, a shoulder ("B" bump) is often present between the A and C points27 (Fig. 15–8). If the transducer beam is directed inferolaterally from the MV level, the papillary muscles and LV apex will be imaged (Fig. 15–6A). With superior and medial angulation, the left atrium (LA), aortic valve (AoV), and aortic root are seen. The tricuspid valve (TV) can be imaged by angulating the transducer inferomedially and the pulmonic valve (PV) by angulating slightly superiorly and laterally.
Assessment of Systolic Function by M-Mode Echocardiography Measurements of the LV cavity dimension and wall thickness can be readily derived from M-mode recordings (Fig. 15–9) and are usually made according to the recommendations of the American Society of Echocardiography (ASE) at end diastole (the onset of the QRS complex) and end systole (the point of maximum upward motion of the LV posterior wall endocardium).28 These measurements should be made from leading edge to leading edge to avoid incorporating artifacts and reverberations; they are accurate if the beam is orthogonal to the long axis of the ventricle. By convention, left atrial (LA) dimension is measured at end systole and aortic root diameter is recorded at end diastole at the level of the base of the heart (Fig. 15–9). During systole, opening of the aortic leaflets appears as a parallelogram produced by motion of the right coronary and (usually) the noncoronary AoV cusps.29
The M-mode LV cavity dimensions can be used to estimate ventricular volumes and ejection fraction (EF) if desired, most simply by merely cubing the value (D3); but these calculations involve several assumptions regarding LV geometry that are not uniformly valid.30,31 In addition, the M-mode dimension may not be representative of the entire ventricle. The fractional shortening can also be determined.32 This value is often helpful in assessing systolic function, but it reflects the function of the LV in one chord and in one plane and can be misleading with asynchronous contraction [for example, left bundle branch block (LBBB)] or segmental dyssynergy.33 An additional M-mode marker of systolic function is E point–septal separation (EPSS), or the distance between the anterior MV leaflet at its most anterior opening excursion (the E point) and the interventricular septum. A value of 8 mm or greater is abnormal.34 The normal M-mode measurements are seen in Table 15–1. Two-Dimensional Echocardiography A number of technical approaches exist by which multiple individual B-mode scan lines can be rapidly transmitted, received, and displayed in appropriate spatial orientation to construct a 2D image of the heart. The initial approach simply utilized a linear array of 20 piezoelectric crystals placed side by side, each of which transmitted and received signals independently8 (Fig. 15–10A). The resulting scan lines were displayed simultaneously to yield rectangular images. Unfortunately, transducer size and interaction between the elements resulted in images of unsatisfactory quality.
Current 2D scanners utilize B-mode scan lines that are independently transmitted and received and are directed through a wedge-shaped sector of cardiac anatomy by means of mechanical or electrical beam steering (Fig. 15–10B to D). A variety of motorized devices are available that, by rapidly oscillating or rotating one or more ultrasonic crystals through space, can mechanically direct multiple scan lines through a sector arc of the cardiovascular system.9,10 The position of the beam in space is derived by determining the orientation of the piezoelectric crystal. Most current 2D scanners utilize a phased-array approach, where multiple ultrasonic crystals are employed in concert to create individual B-mode scan lines.11 The piezoelectric crystals are activated in a closely coordinated temporal sequence, so that the individual wavelets produced by each element merge to form a single beam whose direction is determined by the sequence of crystal firing (Fig. 15–11). Since the direction of the resultant beam is determined by the sequence of activation of the individual elements, the beam can be electrically swept throughout a 90-degree sector arc. Also, a firing sequence can be employed that results in dynamic focusing of the beam along its length to achieve minimal beam width and increased resolution. Phased-array 2D scanners employ small transducers without moving parts that could require repair; however, these systems are more costly.
Originally, echocardiographic data were displayed in analogue form on a standard oscilloscope, transferred to a video monitor by a television camera, and hard-copied onto videotape or paper. Currently, computerized analogue-to-digital scan conversion is standard, so that the polar signals of individual scan lines are converted to a series of numerical gray-level values for individual box-like picture elements (pixels) aligned along X-Y coordinates.35 The ability of a digital step-gradation technique to reproduce the continuous gradation of analog methods is a function of the density of pixels in the matrix and the shades of gray levels available. No loss of data is detected in current digitally converted images, and the digital format provides the opportunity for image processing, enhancement, and quantitation. More importantly, storage in digital format can avoid the image degradation inherent in videotape, provide random access and easy comparison of studies, enable rapid image transmission, and prevent deterioration with image copying and prolonged storage. Technology for fully digital echocardiography is now available, and fully digital acquisition and storage of echocardiograms will be commonplace in the near future, replacing analog videotape recordings. The Standard Two-Dimensional Examination To help standardize the 2D examination, the ASE has recommended that cardiac imaging be performed in three orthogonal planes: long-axis (from aortic root to the apex), short-axis (perpendicular to long axis), and four-chamber (traversing both ventricles and atria through the mitral and TVs)36 (Fig. 15–12). It is important to recognize that the long and short axes are those of the heart, not the body. These three planes can be visualized using four basic transducer positions: parasternal, apical, subcostal, and suprasternal37,38 (Fig. 15–13A,B, and C). In general, the long-axis plane is best imaged from parasternal, apical, and occasionally the suprasternal positions, while the short-axis plane is best imaged in the parasternal and subcostal positions. The four-chamber views are obtained from the apical and subcostal positions. The ASE recognizes that these basic positions and planes may be modified somewhat and recommends that an image obtained within 45 degrees of a basic orthogonal plane be identified with that orthogonal plane. Table 15–2 lists the standard transducer positions and transthoracic echocardiographic (TEE) views. Anatomic drawings of the various imaging planes are seen in Figs. 15–13, 15–14, 15–15, 15–16, 15–17, 15–18, 15–19, and 15–20.
As opposed to other types of cardiac imaging, which are well standardized, the echocardiographic examination is iterative and largely determined by the anatomic characteristics of the patient and manual manipulation of the transducer by the operator. Of paramount importance is the identification of a thoracic site (window) that enables transmission of the ultrasound signal to the heart. In actual practice, the echocardiographic examination is performed with the operator either to the patient's left or right. The patient is in the left lateral decubitus position for most of the examination, with the head of the bed elevated 20 to 30 degrees. Alternate positioning may be employed for individual patients and views. Use of a thick foam rubber mattress (made expressly for echocardiography) that has a removable section under the area of the cardiac apex may facilitate the examination. The examination customarily begins with the transducer in the left parasternal position in the long-axis view (Fig. 15–14). This provides excellent images of the LV, aorta, LA, and the mitral and aortic valves. By angling the beam slightly rightward and inferiorly (RV inflow view), the right atrium, RV, and TV are visualized (Fig. 15–15). If the beam is turned slightly leftward and rotated clockwise from the standard parasternal long-axis view, the RV outflow tract, PV, and main pulmonary artery (PA) appear (RV outflow view). A 90-degree clockwise turn of the transducer produces the parasternal short-axis view. Slight axial angulation of the transducer enables visualization of the LV at various levels of the short axis, including the papillary muscle, mitral leaflets, and AoV (Fig. 15–16). With angulation toward the base, the LA, right heart structures, main PA, and occasionally the LA appendage are also recorded. The apical views are best acquired with the patient in a steep left lateral decubitus position and the transducer at the point of the apical impulse. The four-chamber view is obtained by turning the transducer so that both ventricles, atrioventricular valves, and atria are visualized (Fig. 15–17). In this view, the septal, apical, and lateral walls of the LV are visualized. Slight superior angulation of the transducer will add the AoV and proximal ascending aorta to the echocardiographic image (apical five-chamber view). From the four-chamber view, 90 degrees of counterclockwise transducer rotation will produce the apical two-chamber view (Fig. 15–18A and B). This imaging plane demonstrates the LA and the inferior, apical, and anterior wall segments of the LV (the right heart structures are absent). If the transducer is rotated slightly back toward the four-chamber plane, a three-chamber view similar to the parasternal long-axis view is produced (Fig. 15–18C) and provides images of the posterior, apical, and anteroseptal LV wall segments as well as the LA, aorta, and mitral and aortic valves. To facilitate subcostal imaging, the patient is moved into a supine position. The subcostal four-chamber view is much like the apical four-chamber view (Fig. 15–19), but because the ultrasound beam is now more perpendicular to the interventricular and interatrial septa, subcostal imaging is often helpful in the examination of these structures. A 90-degree rotation of the transducer will record a subcostal short-axis view. The transducer can also be angled to image the RV outflow and PA as well as the inferior vena cava (Fig. 15–19). The long-axis suprasternal imaging plane is shown in Fig. 15–20. In adult echocardiography, the LV is usually not visualized satisfactorily from the suprasternal position, but these imaging planes are well suited for examination of the thoracic aorta, PA, and great vessels. Normal values for 2D echocardiographic measurements are shown in Table 15–3. Three-Dimensional Echocardiography Several approaches exist to obtaining 3D echocardiographic images. The simplest approach is to merely move the transducer through a defined space and align the tomographic slices appropriately. A variety of spatial locator devices can be attached to the transducer to provide spatial orientation. This enables the acquisition of data from many transducer positions. Images obtained in this way can be strikingly accurate, but they require computer reconstruction and therefore cannot be displayed in "real time." Several years ago, a probe with two orthogonally positioned crystal arrays was applied in conjunction with rapid parallel signal processing to achieve real-time 3D volumetric imaging. A pyramid-shaped ultrasound beam is produced that can often encompass the entire heart from one transducer location and acquire an entire data set in a single cardiac cycle (Fig. 15–20C). The resultant 3D data sets from any approach can be displayed as 2D tomographic cuts with 3D spatial orientation, as wire runs, or with surface rendering. This type of real-time 3D imaging has evolved considerably, and new software advances have improved tissue rendering and endocardial border definition (Fig. 15–20D). 3D images have been particularly of value in providing accurate quantitation, in assessing congenital heart disease and in evaluating structures of complex geometry such as the RV.39,40 Assessment of Systolic Function by Two-Dimensional Echocardiography Because 2D echocardiography enables visualization of the entire LV perimeter in multiple planes, it is significantly superior to M-mode approaches for the measurement of cardiac chamber volumes and EF.41–44 Numerous algorithms have been applied to calculate LV volumes by echocardiography (Fig. 15–21). Most such algorithms have assumed that the LV conforms to the shape of a prolate ellipsoid and calculated volume by diameter-length or area-length formulas.42,45 Multiple studies comparing LV volume calculated by area-length methods to those obtained by other techniques have yielded good correlations, with the best results obtained utilizing biplane apical views.45,46 Other algorithms have assumed an LV cavity configuration that is a combination of geometric shapes, such as a cylinder-cone or a cylinder-hemiellipse.45, |
