CONCLUSION

CONCLUSION

Three-dimensional echocardiography is a safe, noninvasive

imaging modality that is complementary and

supplementary to 2D imaging and can be used to

assess cardiovascular function and anatomy in various

clinical settings. At present, available evidence suggests

that 3D echocardiography provides improved

accuracy and reproducibility over 2D methods for LV

volume and function calculation and the derivation of

mitral valve area in patients with mitral stenosis. Further

technological improvements and additional clinical

studies will broaden the list of appropriate applications

for this exciting new ultrasound modality.

FUTURE DIRECTIONS

FUTURE DIRECTIONS

Ongoing developments in 3D echocardiography

include technological innovations and expanding

clinical applications. Automated surface extraction

and quantification, single-heartbeat full-volume

acquisition, transesophageal RT3D imaging,

the ability to navigate within the 3D volume, and

stereoscopic visualization of 3D images are some

of the technological advances that can be expected

over the next several years. These will

further enhance the quality and clinical applications

of 3D echocardiography. In addition, standardized

and focused 3D protocols will be developed

and refined to optimize clinical application

of this technique.

Tagging and/or tracking the LV surface in real

time may provide new approaches to quantifying

myocardial mechanics, such as regional shape and

strain. This approach has great potential and will

complement and likely compare favorably with

the quantitative ability of cardiac MRI. The superior

temporal resolution of echocardiography

should offer unique advantages for this purpose.

In the future, combining the greater temporal

resolution of 3D echocardiography with the excellent

spatial resolution of MRI (or computed

tomography) may yield an imaging data set with

FUTURE DIRECTIONS

Ongoing developments in 3D echocardiography

include technological innovations and expanding

clinical applications. Automated surface extraction

and quantification, single-heartbeat full-volume

acquisition, transesophageal RT3D imaging,

the ability to navigate within the 3D volume, and

stereoscopic visualization of 3D images are some

of the technological advances that can be expected

over the next several years. These will

further enhance the quality and clinical applications

of 3D echocardiography. In addition, standardized

and focused 3D protocols will be developed

and refined to optimize clinical application

of this technique.

Tagging and/or tracking the LV surface in real

time may provide new approaches to quantifying

myocardial mechanics, such as regional shape and

strain. This approach has great potential and will

complement and likely compare favorably with

the quantitative ability of cardiac MRI. The superior

temporal resolution of echocardiography

should offer unique advantages for this purpose.

In the future, combining the greater temporal

resolution of 3D echocardiography with the excellent

spatial resolution of MRI (or computed

tomography) may yield an imaging data set with

Contrast Echocardiography

Contrast Echocardiography

The use of contrast with 3D echocardiography to

improve quantification of LV volumes offers several

advantages. The RT3D technique (single or

full volume) provides the most practical approach.

Triggering, although not essential, increases

the signal-to-noise ratio and thus is superior

to nontriggered imaging.38 Preliminary

clinical studies have shown promise with regard

to improved LV surface identification and volume

and ejection fraction measurement.129,130

Another evolving application of contrast 3D

echocardiography is in the evaluation of myocardial

perfusion. The ability to record the entire LV

and to quantify the full extent of hypoperfused

myocardium is a potential advantage of this approach.

131-133 However, the problem of microbubble

destruction, even with triggered imaging,

remains a challenge. This is especially true when

matrix array transducers are used, which results in

suboptimal myocardial opacification due to high

acoustic power. Further technological developments

should lead to improvements in all of these

areas and will contribute to more practical applications

of contrast 3D echocardiography.

Intraoperative Applications

Intraoperative Applications

The accuracy, feasibility, and value of 3D echocardiography

also have been demonstrated in the

intraoperative environment. Intraoperative 3D

echocardiography provides accurate and often

additional anatomic information compared with

2D transesophageal (TEE) imaging.123 In limited

studies examining 2D and 3D TEE intraoperative

evaluation of mitral valve prolapse anatomy, 3D

TEE evaluation provided complementary and additional

information compared with 2D TEE for

localization of prolapsed scallops (video clip

14).77,124 Intraoperative 3D TEE also has been

used to identify distortion and folding of the mitral

annulus as a cause of functional mitral stenosis or

worsening mitral regurgitation during beatingheart

surgery.125 Finally, intraoperative 3D TEE

has proven valuable in patients undergoing surgery

for congenital heart lesions. For example, the

superiority of intraoperative 3D TEE compared

with 2D has been demonstrated by its ability to

provide en face and oblique views of left atrioventricular

valve malformations in patients undergoing

reoperation for persistent regurgitant lesions

after previous repair of atrioventricular septal

defects.78,126

Intraoperative epicardial RT3D echocardiography

has been used to improve spatial orientation

and assess the extent of septal thickening, mitral

valve systolic anterior motion, and postsurgical LV

outflow tract patency in a patient with hypertrophic

cardiomyopathy undergoing septal myectomy,.

127 It also has been used to guide and monitor

off-pump atrial septal defect closure in a beatingheart

animal model.128 Finally, intraoperative epicardial

and postoperative transthoracic RT3D

echocardiography has been used to evaluate

changes in LV volume and function during cardiac

surgery in patients undergoing infarct exclusion

surgery for ischemic cardiomyopathy.69 In contrast

to 3D echocardiographic imaging, conventional

2D methods may not accurately quantify LV

volumes in patients with severe ischemic cardio

cardiomyopathy,

especially in the presence of significant

geometric changes due to LV aneurysm.

Congenital Heart Disease

Congenital Heart Disease

Clinical investigations examining the role of 3D echocardiography

in patients with congenital heart disease

have emphasized the unique perspective provided by

3D imaging and the versatility of the technique in

patients with simple defects or complex conditions

and in the postoperative state.115,116 Three-dimensional

echocardiography, using both reconstruction

methods and RT3D, has been used to detect several

forms of congenital heart disease. The ability to record

and analyze the entire cardiac structure and the ability

to display complex spatial relationships are potential

advantages of 3D imaging over 2D echocardiography.

In addition, the decreased examination time afforded

by RT3D echocardiography may reduce the need for

sedation in some children.116

In patients with atrial septal defects, 3D echocardiography

can record the size and shape of the defect. It

also can show the precise location of the defect and

the extent of residual surrounding tissue. In patients

with secundum atrial septal defects (Figure 11, video

clip 13), the extent of the retroaortic rim often determines

the feasibility of repair with percutaneous closure

devices. Three-dimensional echocardiography

also has been used after atrial septal defect closure to

evaluate the success of the procedure and identify the

origin of residual shunting.117 In patients with ventricular

septal defects, the ability to interrogate the entire

septum is frequently cited as an advantage of the 3D

technique.118,119 A novel application of 3D imaging in

patients with ventricular septal defects involves using

offline reconstruction to measure the shape and size of

the color flow jet, which allows for accurate measurement

of the magnitude of shunting in patients with

isolated ventricular septal defects.119

Various 3D echocardiographic techniques have

been used to evaluate RV and LV size and function in

patients with congenital heart disease. The approach

to the LV is similar to that described previously and

permits quantification of dimension, volume, mass,

and ejection fraction.120 Owing to the ellipsoidal shape

of the LV, the advantages of 3D over 2D echocardiographic

techniques are limited, because simple geometric

assumptions can be used to calculate LV volumes;

however, the RV’s asymmetrical shape

invalidates the simple geometric assumptions used for

LV volume calculations. In this case, the ability to

record and analyze the entire chamber rather than

relying on simplifying assumptions has proven

superior.48 In patients with congenital heart diseases

that involve RV pathology, 3D echocardiography correlates

well with MRI for the measurement of RV

volume.48,121,122

Three-dimensional echocardiography has been

successfully applied to the detection and assessment

of several anatomic defects. For example,

the circumferential extent and severity of discrete

subaortic membranes have been successfully visualized

with 3D echocardiography.119,120 With the

apical view, a unique en face image of the membrane

can be recorded, which permits analysis of

the effective orifice area and the dynamic nature

of the defect. Congenital malformations of the

mitral valve also have been assessed with 3D

echocardiography.80 The complex nature of these

defects can make a thorough anatomic evaluation

difficult. In such cases, the perspective provided

by 3D echocardiography can provide a complete

preoperative assessment of the extent and severity

of the valvular abnormality.

Valvular Heart Disease

Valvular Heart Disease

The recent widespread availability of RT3D echocardiography

obviates many of the practical limitations of

reconstructive 3D techniques and offers the potential

for greater clinical application for valvular heart disease

both in standard diagnostic evaluation and in

real-time guidance during surgical valve repair. This

technique is ideally suited for assessing valve function

given the nonplanar anatomy of the cardiac valves and

the associated anatomic and spatial alterations associated

with valvular heart disease.

Mitral Valve. The 3D echocardiography technique

has contributed significantly to our understanding of

mitral valve function and anatomy. The mitral valve is

particularly suited to 3D assessment because of the

complex interrelationships among the valve, chordae,

papillary muscles, and myocardial walls. This technique

can provide important insight into mitral valve

structure, demonstrating the saddle shape of the mitral

annulus, with high points located anteriorly and low

points oriented in a mediolateral direction (Figure 10,

video clip 8). This has helped clarify the appropriate

diagnostic imaging planes from which mitral valve

prolapse should be diagnosed, thereby avoiding falsepositive

interpretations.64,65

In addition, 3D echocardiography has provided important

mechanistic insights into functional and ischemic

mitral regurgitation resulting from derangements

of the normal spatial relationships of the mitral valve

leaflets to its chordal attachments, papillary muscles,

and the LV.13,66 Distortion of the normal spatial relationship

between the LV and mitral valve apparatus

results in papillary muscle displacement and tethering

of the mitral leaflets, leading to incomplete closure of

the leaflets and mitral regurgitation (video clip 9). The

3D echocardiography technique has identified

changes in annular shape occurring with functional

mitral regurgitation.67-69 These mechanistic and anatomic

insights based on 3D analysis have provided the

basis for the development of new approaches to

treating ischemic mitral regurgitation.70-74

Three-dimensional echocardiography has been used

to define and localize mitral leaflet lesions in mitral

valve prolapse, endocarditis, and congenital mitral

abnormalities.75-80 This application has been particularly

important in guiding surgical repair (video clips

10 and 11).81-83

The RT3D approach has also demonstrated efficacy

in quantifying mitral regurgitation by using 3D guidance

to directly measure the proximal flow convergence

region.84,85 It has provided insight into how

premitral orifice geometry affects the calculation of

mitral valve area in mitral stenosis.86 Calculation of

mitral valve area by 3D echocardiography has been

demonstrated to be accurate, reproducible, and less

variable than conventional 2D methods 

and thus has been recommended as the firstline

method.92 In addition, 3D echocardiography has

been used for guidance during percutaneous mitral

valvuloplasty.93,94

Aortic Valve. Three-dimensional echocardiography

has been applied for anatomic assessment of the aortic

valve and root morphology and to calculate the valve

area in aortic stenosis.95-99 The technique has been

used to delineate aortic flow patterns100,101 and has

demonstrated feasibility and accuracy in quantifying

aortic regurgitation.102,103 Other applications have included

the detection and localization of aortic valve

vegetations, assessment of congenital outflow obstruction

abnormalities, and demonstration of morphological

changes in the valve after balloon dilation.103-107

Tricuspid and Pulmonary Valves. Compared with

the aortic and mitral valves, the tricuspid and pulmonary

valves have been less widely studied with 3D

echocardiography. This technique has demonstrated

anatomic changes with rheumatic and degenerative

tricuspid valve disease.108-110 and has accurately reconstructed

congenital tricuspid valve abnormalities, such

as atrioventricular canal defects.111,112 For the pulmonary

valve, 3D assessment has been limited to descriptive

case reports defining anatomic abnormalities associated

with pulmonary valve stenosis and

endocarditis

CLINICAL APPLICATIONS

CLINICAL APPLICATIONS

Chamber Quantification

Left Ventricle. LV chamber and mass quantification

have been studied extensively using 3D echocardiography

(Tables 2 and 3).20-38 Initial 3D methods to

measure LV volumes used reconstruction techniques

that, although more accurate and reproducible than

2D methods, required long acquisition and postprocessing

times.20,21,30,35,39-41 Moreover, the accuracy of

the volume calculations was highly dependent on

image quality. The introduction of real-time imaging

systems that use matrix phased-array transducers with

more processing elements has significantly improved

image quality. In addition, LV quantification algorithms

that can interface with 3D data sets obtained with

matrix phased-array transducers are now widely available

and are increasingly robust.

The wide-angle acquisition mode is often used to

acquire the entire LV volume, from which a detailed

analysis of global and regional wall motion can be

done. Images may be displayed with either orthogonal

long-axis views or multiple short-axis views. Currently,

data analysis is performed offline on a personal computer

with dedicated 3D software. Data also can be

analyzed online with software intrinsic to the ultrasound

machine. Because a data set comprises the

entire LV volume, multiple slices from different orientations

can be obtained from base to apex to evaluate

wall motion. If image quality is limited, then acquisition

can be combined with infusion of contrast to

improve the delineation of the endocardial border. An

advantage of a 3D data set over 2D is the ability to

manipulate the plane to align the true long axis and

minor axis of the LV, hence avoiding foreshortening

and oblique imaging planes (Figure 7). Once the LV

axes are appropriately aligned, LV volumes can be

calculated with a centroid-based algorithm that typically

uses 2 or 3 planes (Figure 8), thereby shortening

processing time. In addition, the LV volumes can be

segmented, which allows for regional LV function

assessment (Figure 9, video clip 5).

LV volume assessment by RT3D has been demonstrated

to be rapid, accurate, reproducible, and superior

to conventional 2D methods.42 The superiority of

the RT3D approach has been demonstrated in various

clinical situations, but its use is limited in patients with

a poor acoustic window. An alternative method of

calculating ventricular volumes from an RT3D cardiac

volume data set uses the disc summation method. This

technique may be advantageous in patients with asymmetrical

ventricles.22,43 LV volume and mass obtained

by RT3D echocardiography compare favorably with

those obtained with cardiac magnetic resonance imaging

(MRI) or radionuclide volumes.39,42,44 In addition,

RT3D echocardiography has demonstrated efficacy

and accuracy in assessing LV volumes in remodeled

ventricles after myocardial infarction and in assessing

global LV dyssynchrony (video clips 6 and 7).32,45

Preliminary clinical studies on the use of RT3D in

stress echocardiography confirm the feasibility of this

technique and report sensitivity and specificity comparable

to 2D stress imaging.46,47 An advantage of

RT3D stress imaging is the decreased imaging time;

standard views can be obtained with only 1 or 2 image

acquisitions. In preliminary clinical studies, average

acquisition times decreased from 65 to 28 seconds

with RT3D imaging.46,47

Right Ventricle. Assessment of right ventricle (RV)

function by 2D echocardiography is limited because of

the asymmetrical, pyramidal shape of the RV, which

does not conform to simple geometric assumptions. In

theory, direct visualization of the entire chamber

should be possible with 3D techniques, thereby overcoming

the inherent limitations of tomographic methods.

To date, most studies that have applied 3D

echocardiographic techniques to the RV have involved

primarily rotational or freehand scanning methods

(Table 4); most of these series demonstrated improved

accuracy of RV function assessment.48-58 However,

these 3D data sets involved reconstruction from serial

2D images with the need for offline postprocessing,

thereby limiting their widespread clinical application.

The recent development and availability of RT3D

echocardiography has the potential to further improve

the ability to assess RV chamber size, volume, and

function.

Left Atrium. In a limited number of studies, left

atrial volume has been accurately quantified by 3D

echocardiography using both reconstructive and realtime

techniques. The 3D echocardiographic methods

correlate well with MRI59-62 and appear to have accuracy

comparable to 2D left atrial volume methods

Protocols

Protocols

A complete 3D echocardiographic study includes an

assessment of ventricular function, valvular morphology,

and hemodynamic status. Unlike 2D echocardiography,

in which standard views are described based on

the plane through which they pass, 3D echocardiography

is inherently volumetric. As such, it permits both

an external view of the heart and multiple internal

perspectives (through cropping).

Table 1 lists the components of a complete 3D

echocardiographic study. A general approach is to

describe cardiac structures using both the ultrasound

plane and the viewing perspective. Three

orthogonal planes are recommended: (1) the sagittal

plane, which corresponds to a vertical, longaxis

view of the heart; (2) the coronal plane,

which corresponds to a 4-chamber view; and (3)

the transverse plane, which corresponds to a

short-axis view (Figure 5). Each plane can be

viewed from 2 sides, which represent opposite

perspectives; for example, the transverse plane,

which represents the short-axis view, can be

visualized from the perspective of the apex or

base; the coronal plane can be viewed from above

or below; and the sagittal plane can be viewed

from the left or right. The choice of narrow-angle

or wide-angle imaging acquisition modes depends

on the cardiac structure to be examined. For

imaging of the ventricles, it is best to use a

wide-angle acquisition in the apical window (4-

chamber) so as to include the entire ventricle. For

smaller structures, such as the aortic valve, a

narrow-angle acquisition may be adequate.

As an alternative to a complete 3D study, 3D echocardiography

can be performed selectively as a complement

to a 2D study. Instead of a complete 3D

echocardiogram, a more focused 3D imaging study

may be appropriate in some cases. For example, in a

patient with mitral stenosis, the 3D portion of the

study may be limited to visualization and quantification

of the mitral orifice. Focused 3D imaging for LV

volume calculation, typically performed with an apical

4-chamber wide-angle acquisition, also can be used to

complement standard 2D imaging.

The ability to extract hemodynamic information

derived from 3D color Doppler ultrasonography is

currently being investigated. To capture and analyze

color flow imaging in 3 dimensions, the area of interest

should be obtained within the 3D data set, with the

angle of the ultrasound beam aligned as parallel as

possible to the direction of blood flow. Depth and

sector settings should be optimized for color Doppler

resolution. Extraneous flows can be cropped so that

only the area of interest is displayed. The color Doppler

flow patterns can be analyzed in multiple views to

provide a complete assessment of the color Doppler

data (Figure 6).

Technical Factors

Technical Factors

The technical aspects of acquiring a high-quality,

diagnostic 3D echocardiogram are similar to those

for 2D echocardiography. As with any new imaging

technique, a learning curve exists, and recognizing

and avoiding potential artifacts is critical.

Many of the artifacts are related to respiratory or

ECG gating and/or incorrect gain settings. The use

of optimal gain settings before acquisition is essential

for accurate diagnosis. Low gain settings

can artificially eliminate certain structures that

then will not be viewable during postprocessing.

Alternatively, using high gain settings can mask

structures and lead to significant misdiagnoses.

Therefore, overcompensating for the brightness

of the image using the time-gain compensations is

recommended, to allow the overall gain to be set

at midrange values. This maneuver will allow

maximum flexibility with postprocessing settings.

Most 3D echocardiographic systems use some

form of gating to obtain volumetric data. Gated

data sets are most challenging in patients with

arrhythmias or respiratory difficulties. The confounding

effects of the gating artifacts can be

minimized in different ways. For example, if the

gated system acquires sector slices in a sweeping

motion parallel to the reference image, then every

image viewed parallel to the reference image will

appear normal, whereas the gated artifacts will be

most noticeable when viewed from a plane orthogonal

to the reference image.

Segmentation is the process by which anatomic

features are extracted from the raw ultrasound

data. Segmentation can be accomplished using

low-level techniques, such as edge detection (in

2D) and surface detection (in 3D) based on local

features, such as the spatial gradient in ultrasound

intensity. More sophisticated techniques attempt

to extract entire boundaries or surfaces at once

based on local features and the anticipated shape

of the overall structure. Compression is a mathematical

technique that can be applied to the

original digital image file to reduce the amount of

data, thereby decreasing storage requirements and

improving retrieval rates. A single-volume data set

from a typical RT3D echocardiographic system

consists of 64 64 512 bytes (approximately 2

MB), or more than 50 MB for a 1-second loop, a

load that can overwhelm storage systems. Compression

of the digital data files can reduce this

load by about 3:1. The motion-JPEG algorithm

currently used by DICOM (Digital Imaging and

Communications in Medicine) and applied to individual

2D slices could be expected to achieve an

approximate 20:1 compression. More advanced

algorithms, such as wavelets (JPEG-2000), potentially

can yield better results. The use of compression

algorithms can decrease the size of data files,

optimizing storage efficiency without sacrificing

image quality.

Real-Time 3D Acquisition Methods

Real-Time 3D Acquisition Methods

Several studies have demonstrated that 3D reconstruction

from serial 2D images provides accurate

anatomic information suitable for quantitative

analysis.4-8,10-15 However, this methodology is subject

to technical limitations during image acquisition

and requires significant offline data processing.

The development of RT3D echocardiographic

systems circumvents many of the disadvantages of

reconstructive methods. RT3D echocardiography

uses a transducer with ultrasound elements arranged

in a grid fashion (Figure 2). The earliest

devices, developed by von Ramm and colleagues,

3,16,17 used a sparse-array matrix transducer

transmitting at a frequency of 2.5 or 3.5

MHz. These transducers consisted of 256 nonsimultaneous

firing elements and acquired a pyramidal

volume data set measuring 60° 60° within

a single heartbeat. However, the resolution and

image quality of this first-generation sparse-array transducer

were relatively poor and often inferior to standard

2D images; frame rates were low; and the pyramidal

volume had a narrow sector angle of 60°,

resulting in an inability to accommodate larger ventricles.

Moreover, the images obtained with this system

were not volume-rendered online; instead, they consisted

of computer-generated 2D cut planes derived

from the 3D volume data set. These features limited

clinical application of this pioneering technology.

Current RT3D systems use matrix-array transducer

technology with a greater number of imaging

elements, typically containing more than 3000

imaging elements, compared with the 256 in the

sparse-array transducer. These current matrix-array

transducers offer improved resolution and are

rapidly becoming the primary technique for 3D

data acquisition in clinical and research practice.

However, recent improvements in transducer

technology have resulted in (1) a smaller transducer

footprint, (2) improved side-lobe suppression,

(3) greater sensitivity and penetration, and

(4) harmonic capabilities that may be used for

both gray-scale and contrast imaging. In addition,

these matrix-array transducers display either online

3D volume-rendered images or 2 to 3 simultaneous

orthogonal 2D imaging planes (ie, biplane

or triplane imaging).

RT3D systems generally have 3 acquisition

modes: real time (narrow), zoom (magnified), and

wide angle. The real-time mode displays a pyramidal

data set of approximately 50° 30° (Figure

3A, video clip 1).18 The zoom mode displays a

smaller, magnified pyramidal data set of 30° 30°

at a higher resolution (Figure 3B). The wide-angle

mode provides a pyramidal data set of approximately

90° 90°, which allows inclusion of a

larger cardiac volume (Figure 3C, video clip 2).

This wide-angle mode requires ECG gating, because

the wide-angle data set is compiled by

merging 4 narrower pyramidal scans obtained

over 4 consecutive heartbeats. To minimize reconstruction

artifacts, data should be acquired during

suspended respiration if possible. Although wideangle

data sets provide a larger pyramidal scan,

this is at the cost of lower resolution, which is

decreased compared with the narrow-angle 3D

mode.

Once a 3D data set is acquired, it must be sliced or

“cropped” to visualize the cardiac structures within

the pyramid (Figure 4). Multiple cropping methods

are available, but a common method displays 2 or 3

imaging planes simultaneously (video clips 3a and

3b). Each of these imaging planes can be manipulated

separately to appropriately align the cardiac

structures. Another cropping method involves a

single-slice plane that can be manually adjusted to

expose and display the cardiac structures of interest

METHODOLOGY

METHODOLOGY

Reconstruction Techniques

Early approaches to 3D echocardiography were

based on the principle that a 3D data set could be

reconstructed from a series of 2D images. In this

method, serial 2D images are obtained using either

freehand scanning or a mechanically driven transducer

that sequentially recorded images at predefined

intervals.4-8 With freehand scanning, a series

of images is obtained by manually tilting the transducer

along a fixed plane, and a spatial locator

attached to the transducer translates the 3D spatial

location onto a Cartesian coordinate system.

This approach has several practical limitations,

including the relative bulk of the acoustic spatial

locators, which makes transducer manipulation

difficult, and the need for a clear and direct path

between the acoustic locators and the transmitter.

For electromagnetic spatial locator systems, an

additional problem is the potential for interference

of the electromagnetic field by ferromagnetic

material in close proximity to the transducer

(eg, material in hospital beds and medical equipment).

9

An alternative to freehand scanning is the use of a

mechanized transducer to obtain serial images at set

intervals in a parallel fashion or by pivoting around

a fixed axis in a rotational, fanlike manner. Because

the intervals and angles between the 2D images are

defined, a 3D coordinate system can be derived from

the 2D images in which the volume is more uniformly

sampled than with the freehand scanning

approach.

More recently, the use of a transesophageal or

transthoracic multiplane probe has emerged as a

readily available method to obtain rotational images

at defined interval angles around a fixed axis.10-15

Typically, images are collected over a 180-degree

rotation at set intervals. To minimize reconstruction

artifacts, sequential images are gated to both electrocardiography

(ECG) and respiration. Acquisition

of a complete data set typically takes 1 to 5 minutes,

depending on respiratory and heart rates and the

predefined spatial intervals. During cardiac surgery,

respiration can be suspended during acquisition to

minimize the effects of respiratory motion.

The quality of 3D reconstructions from 2D

images depends on a number of factors, including

the intrinsic quality of the ultrasound images, the

number (or density) of the 2D images used to

reconstruct the 3D image, the ability to limit

motion artifact, and adequate ECG and respiratory

gating. In general, the greater the number of

images obtained (ie, the smaller the space intervals

between images), the better the 3D reconstruction.

However, increasing the number of

images also lengthens the acquisition time, which

can potentially introduce motion artifact. Consequently,

the optimal number of images necessary

for 3D reconstruction depends on the cardiac

structure being examined and the resolution required.

For example, 4 to 6 serial images are

usually adequate for volume reconstructions of

the left ventricle (LV), whereas more images are

often needed to visualize more complex, rapidly

moving structures, such as mitral and aortic

valves.

Once the 2D images have been obtained, they

are processed offline with customized or commercially

available software. The cardiac structures

are manually or semiautomatically traced to the

3D spatial coordinates to reconstruct a 3D image.

BACKGROUND

BACKGROUND

Attempts to record and display ultrasound images in

3D format were first reported in the 1960s. One of

the earliest studies described the acquisition of a

series of parallel scans of a human orbit to reconstruct

3D anatomy.1 Despite the limited technology

of the day, these initial studies demonstrated that

complex anatomic structures were ideally displayed

using 3D techniques. Concerns about image quality

and the computational power needed for storage

and reconstruction greatly limited the early application

of this methodology.

More than a decade later, investigators began to

obtain 3D ultrasound images of the heart.2 Through

the careful tracking of a transducer, a sequence of

2-dimensional (2D) echocardiograms could be recorded,

aligned, and reconstructed into a 3D data

set. This methodology was limited by the need for

offline data processing to create and display the 3D

images. In the early 1990s, von Ramm and colleagues3

developed the first real-time 3D (RT3D)

echocardiographic scanner, capable of acquiring

volumetric data at frame rates sufficient to depict

cardiac motion. More recently, further improvements

in design and engineering have led to the

commercialization of RT3D echocardiography. This

methodology has evolved quickly, and different

versions of RT3D imaging are currently available on

several platforms.

3D Echocardiography: A Review of the Current Status and Future Directions

Ultrasound technology has improved markedly in

the past 10 to 15 years, prompting echocardiographers

to extend its use in studying cardiac

structure and function. New ultrasound equipment

and techniques offer superior image quality, greater

accuracy, and expanding capabilities. As a result,

more and improved imaging modalities are available

for evaluating cardiac anatomy, ventricular function,

blood flow velocity, and valvular diseases. Threedimensional

(3D) echocardiography offers the ability

to improve and expand the diagnostic capabilities

of cardiac ultrasound. However, as with any

emerging technology, the enthusiasm to embrace a

new technique must be tempered by a critical

appraisal of the evidence supporting its use. It is

essential to assess the limitations as well as the

unique capabilities it provides. Cardiac imaging

should be safe, accurate, versatile, comprehensive,

and cost-effective, while providing important clinical

information. Criteria for appropriate utilization

should be based on current evidence and updated as

new information becomes available.

To justify the use of a new 3D modality, its unique

contribution to clinical practice must be critically

analyzed. In this article we review the status of 3D

echocardiography, examine the evidence for its use

in various clinical situations, and propose guidelines

for appropriate application of this technique based

on available evidence.

The Future of 3D-Echo

The Future of 3D-Echo

With the rapid advances in digital image processing 3Dimaging

is probably just at the beginning of its evolution with

a number of innovations already approaching. The integration

of 3D-systems into conventional scanners and operator

friendly applications will reduce the time and effort required

to obtain 3D-images. Improvements in 2-dimensional

imaging and 3-dimensional reconstruction software will lead

to enhanced image quality. Novel ways of image representation

such as stereoscopy, holography or the generation of

physical 3D-models (Figure 6) could enhance our perception

of cardiac structures [17, 18].

Further advances in real-time 3-dimensional imaging and

the advent of real-time 3D-color Doppler will generate 3-dimensional

images directly on the screen during the investigation,

making the technique as simple as using Doppler or Mmode

features [19]. And, finally, other echocardiographic

modalities such as contrast echocardiography, tissue Doppler

imaging or intracardiac ultrasound could benefit from a

3-dimensional display format [20].

Some of these developments could soon become available
while others are still experimental. Despite present limitations
3-dimensional imaging could soon be considered as a
milestone in echocardiography along with M-mode, 2-dimensional
imaging or Doppler techniques.

Limitations

Limitations

However, despite the potential of 3D-echo to visualize cardiac

structures and perform volume computations this technique

has not gained wide spread acceptance to date. This

might be related to several factors: (1) 3D can only visualize

what is also seen on the two dimensional image, thus, an experienced

echocardiographer will obtain similar information

from a conventional examination without the need for costly

instrumentation and long post-processing times, (2) operator

experience with the reconstruction and interpretation of 3Dimages

is necessary, (3) 3D-image quality greatly depends on

the quality of the two-dimensional image and the ability to

obtain a motion and artifact free 3D-data set, (4) three-dimensional

imaging only creates a “virtual sense of depth” on a

flat (2-dimensional) screen. And finally, manual endocardial

contour tracing is still required to obtain 3D-volumes.

Some of these limitations will certainly be overcome with

newer techniques and growing experience with 3D-echo.

3D-Echocardiography as Teaching and Research Tool

3D-Echocardiography as Teaching and Research Tool

Spatial representation of cardiac structures greatly enhances

the understanding of cardiac function and pathology. Thus,

three-dimensional images could assist in the teaching

of echocardiography where a significant amount of spatial

understanding is required. An example of such an application

is a system which couples 3D-echo with a virtual reality heart

model [16]. The system allows standardized echocardiographic

views to be selected on the virtual heart and displayed

from the 3D-dataset to provide a correlation between

anatomy of the heart and echocardiographic image planes.

Potential Applications of 3D-Echocardiography

Potential Applications of 3D-Echocardiography

The potential applications of 3D-echo can be categorized

into 3 major areas: (1) Interpretation of morphology and

pathology, (2) Quantification of volumes and function,

(3) 3D-echocardiography as a teaching tool.

Interpretation of Morphology and Pathology

The clinical potential of 3D-echocardiography has been thoroughly

explored. Our own experience and that of others have

clearly demonstrated that the anatomy (Figure 1) and pathology

of the heart and the great vessels can often be displayed

[8, 9] in a more comprehensive format. Even fairly small

structures such as coronary arteries, a paravalvular leak or

small masses and vegetations can be visualized [7, 10].

Our findings also show that this technique can be applied

in numerous settings. For example, in valvular heart disease

(Figure 2), to determine the size of infectious vegetations, to

determine the mitral valve area in mitral stenosis (Figure 3),

for complex congenital malformations, or aortic dissection

(Figure 4). Furthermore, it has also been shown that jets can

be reconstructed from color Doppler information to assist in

the quantification of valvular lesions [11].

Quantification of Ventricular Volumes and Function

3D-echo has been applied to derive quantitative measurements

of volume, mass and dimensions of the left and right

ventricles and also other cardiac lesions, such as atrial and

ventricular septal defects [12]. While quantification of

ventricular volumes with two-dimensional imaging requires

geometric assumptions, measurement obtained with

3D-echo represents true volumes. Several studies have

shown 3D-echo to be superior to 2D-echocardiography for

both left and right ventricular volumes [13]. The processrequires acquisition of a 3-dimensional data set and manual

endocardial contour tracing. Several calculations including

volumes (throughout the cardiac cycle), global and regional

ejection fractions can be computed (Figure 5). The endocardial

surface of the ventricular cavity can be displayed from

multiple angles in a dynamic mode. Since the process of

manual endocardial border tracing is still time-consuming,

semi-automated contour detection algorithms are now being

developed. In addition, there is experimental evidence that

contrast opacification of the left ventricle could further

enhance the applicability of 3D-volume computation [14].

The advent of real time volumetric scanning will certainly

enhance the applicability of 3D-volume computation [15].

Principles of 3D-Echocardiography

Principles of 3D-Echocardiography

Data Acquisition

Three-dimensional echocardiography requires the collection

of a volumetric data set where each image (cut plane) is defined

with respect to its exact position in space [4]. Most systems

currently rely on sequential collection of image planes.

With this technique it is necessary to use ECG and respiratory

triggering or breath hold acquisition to account for motion

artifacts caused by respiration and to permit alignment of the

images in the time domain.

Newer developments use a transthoracic probe technology

with volumetric scanning capabilities, which allows simultaneous

acquisition of an entire 3D-data set. As a result,

data acquisition is less time consuming and less susceptible to

artifacts [5]. 3D-reconstructions have also been applied to

the color Doppler information allowing a three dimensional

representation of jets superimposed on the 3D-grayscale image.

Image acquisition can be performed from both a transthoracic

and a transoesophageal approach (TEE).

Transthoracic 3D-Echo

Three-dimensional transthoracic imaging can be performed

with mechanical steering devices, which are attached to

standard transducers. These devices steer the transducer motion

causing incremental changes in the scan plane either by

rotating, shifting or fanning the probe. In addition, various

locating systems (ie, acoustic or electromagnetic) have been

used effectively. The advantage of this technique is that freely

definable image planes can be chosen allowing for more flexibility.

Others have proposed a rapid (6 seconds) acquisition technique

that collects apical tomograms (within 6 sec) using an

internal continuously rotating transthoracic transducer [6].

Volumetric real-time echocardiography is a recently developed

technique based on the design of an ultrasound

transducer with a matrix array that instantaneously acquires

the image contained in a pyramidal volume. Volumetric realtime

echocardiography is a novel imaging concept, which

holds promise as a “break-through” technology for 3D-echo.

Employing a matrix array echo probe this technique allows

instant (real-time) acquisition of a complete 3-dimensionaldata

set without complex post-processing. Several studies

have already demonstrated the validity of real-time volumetric

echocardiography for the calculation of cardiac volumes

[5]. In addition, real-time volumetric echocardiography allows

the reconstruction of freely definable 2D-image planes

from a single volume set independently of the acquisition

window.

Transoesophageal 3D-Echo

First attempts to acquire a 3-dimensional data set from the

oesophagus were made with a specially designed probe

(echo-CT, lobster tail probe, Tom Tec). This probe was capable

of acquiring parallel data sets by passing a transducer

along the oesophagus [7]. Newer technologies, however, use

multiplane TEE probes that acquire sequential images at different

transducer rotation points (0–180o). 3D-echocardio-simply by mounting the steering device onto the TEE probe.
Data Post-Processing and Representation
Post-processing of the data for sequentially acquired images
is performed off-line using dedicated software. Varying
amounts of user interaction are required to define the region
of interest, view, cut plane, rendering algorithm, filter, magnification
and thresholds. The systems provide a variety of
3D-tools for advanced image processing. Multiplanar, 3Dreconstructions
(volume rendering) as well as wire frame
(surface rendering) display formats can be chosen and measurements
(distance, area, angle, volume) can be performed.
Recent advances in computing capabilities such as parallel
processing have greatly reduced the time necessary for data
manipulation. Three-dimensional reconstruction can now
be achieved within seconds and viewed from different angels
in a dynamic format. It is even possible to “electronically”
dissect the heart to visualize otherwise concealed structures.
Volumetric scanning allows instantaneous (real-time) display
of multiple views (multiplanar) using a split screen. In
addition, prototype systems have demonstrated the feasibility
of near real-time 3D-reconstruction, which permits almost
simultaneous display of 3D-images during the examination.

introduction

The interpretation of echocardiographic images requires a

complex mental integration of multiple image planes for

a true understanding of anatomic and pathologic structures.

The representation of images in a 3-dimensional format

more closely resembles reality and could therefore enhance

image interpretation. In addition, 3-dimensional imaging allows

direct calculation of volumes and is, thus, more accurate

than current models relying on geometric assumptions.

First attempts to incorporate multiple views to form a

3-dimensional image were made in the seventies. But, because

of technical limitations (eg lack of processing power,

relatively poor image quality, difficulties in image plane alignment)

this technique was limited to an experimental setting.

The advent of transoesophageal echocardiography together

with newer imaging probes and enhanced image

processing capabilities have now led to a remarkable progress

in the field of 3-dimensional imaging.

Numerous applications of three-dimensional echocardiography

(3D-echo) have been proposed. For example, improvements

in image interpretation with 3D-echo could be

of value in the decision making and planning of cardiac surgery,

and in the diagnosis of complex cardiac lesions [1]. In

addition, 3-dimensional imaging allows quantitative parameters

such as valve areas, the size of defects (atrial septal defect,

ventricular septal defect) or volumes to be obtained [2, 3].

With new developments that allow system integration of 3Dscanning,

rapid or even near real time 3D-reconstruction and

measurements, 3D-echo is now on the verge of becoming an

integral part of an echo examination.

Three-Dimensional Echocardiography – Principles and Promises

Three-Dimensional Echocardiography –
Principles and Promises
Three-dimensional echocardiography facilitates spatial recognition of intracardiac structures, potentially enhancing diagnostic
confidence of conventional echocardiography. In addition, 3D-echocardiography allows exact computation of cardiac
volumes and could serve as a teaching tool in cardiology. Over the last decade, significant developments in 3D-echocardiography
have been made. Refinements in instrumentation, data acquisition, post processing and computation speed together
with improvements in 2D-image quality could now allow three-dimensional echocardiography to play an important role in
clinical echocardiography. This review focuses on the methodology, current status, potential clinical applications and future
direction of 3D-echocardiography