One of the most significant developments in echocardiography over recent times has been the development and availability of three-dimensional (3D) echocardiography for practical clinical use. It has now become an integral aspect of both transthoracic and transesophageal cardiac assessment and provides complementary and incremental information over standard two-dimensional (2D) and M-mode echocardiographic techniques. Interchangeably referred to as 3D and four-dimensional (4D) imaging depending on vendor terminology, the additional dimension simply refers to 3D imaging in real time. High-quality 3D echocardiographic imaging is now routinely achieved in clinical practice, and 3D transesophageal assessment in particular has become an integral part of diagnosis and assessment of valvular heart disease.
3D echocardiography has revolutionized our morphologic and functional assessment of valvular heart disease. Progression in ultrasound and computer technologies has allowed us to achieve real-time direct visualization of valves and subvalvular apparatus from any orientation en face using a single-volume acquisition, thereby providing more realistic images with better reproducibility and improved scope for quantification.
A. Valvular assessment. Arguably, its most valuable contribution to valvular assessment has been our improved understanding of the complicated mitral valve, which can now be assessed from any spatial point of view. The 3D techniques pioneered for the mitral valve are now enabling better surgical planning for minimally invasive surgery and guiding of complex percutaneous procedures. Improved accuracy and reproducibility of distance and area measurements with 3D echocardiography means that it is now routinely employed for calculation of valve areas, annular dimensions, and preoperative planning.
B. Chamber size. Additionally, 3D full-volume technology has enabled us to more accurately assess cardiac chamber size. This has been particularly important in valvular conditions such as mitral and aortic regurgitation, where left ventricular enlargement may be an indication for surgery.
Common indications for valvular assessment with 3D echocardiography are listed in Table 19.1
The American Medical Association (AMA) has two Current Procedural Terminology (CPT) codes that can be used for 3D echocardiography. The ability to bill for 3D imaging and differentiation between the two billing codes is based upon performance of 3D reconstruction and whether the 3D reconstruction was performed online, on-cart, or offline on a remote workstation using postprocessing software. The two CPT codes are 76376 and 76377.
A. AMA descriptions for 3D CPT codes
: 3D rendering with interpretation and reporting of computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, or other tomographic
modality with image postprocessing under concurrent supervision, not requiring image postprocessing on an independent workstation.
TABLE 19.1 Common Indications for Valvular Assessment with Three-Dimensional Echocardiography
Aim of Three-Dimensional Echocardiographic Assessment
Assessment of valvular morphology
Determine congenital or acquired defects
Assessment of valvular stenosis or regurgitation
Determine mechanism and severity including prolapse/flail, perforation, prosthesis dehiscence, thrombus, pannus
Assessment of valvular masses
Determine etiology (e.g., infective endocarditis, tumors, postsurgical material)
Accurate measurements and sizing before percutaneous or surgical valve procedures
TAVR, valvuloplasty, edge-to-edge mitral valve clip, closure devices, valvular surgery
Evaluation for procedural success and complications
Assessment of associated factors
Ventricular volumes and function, left atrial and left atrial appendage thrombus
TAVR, transcatheter aortic valve replacement.
In this scenario, the echocardiographic technician or echocardiologist performs 3D rendering using image postprocessing; however, the provider performs this directly on the echo machine and not on a separate workstation. This method is felt to be less labor and time intensive, and reimbursements for this on-cart method have been accordingly set lower.
a. 2015 Medicare Reimbursement Tech (facility reimbursement): $21.65
b. 2015 Medicare Reimbursement Pro (professional reimbursement): $21.65
2. 76377: 3D rendering with interpretation and reporting of CT, MRI, ultrasound, or other tomographic modality with image postprocessing under concurrent supervision, requiring image postprocessing on an independent workstation.
In this scenario, the echocardiographic technician or echocardiologist performs 3D rendering using image postprocessing; however, the provider performs the reformatting work on a remote, independent workstation with dedicated postprocessing software and capabilities. This method is felt to be more labor and time intensive and typically attracts higher hospital relative value units. Additional setup costs may also be incurred for the remote workstation and postprocessing software. Hence, reimbursements for this offline method have been set at a higher rate. Importantly, in order to bill using this code, the provider must clearly state that 3D rendering was performed on a separate workstation; otherwise, it may be downcoded to 76376.
a. 2015 Medicare Reimbursement Tech (facility reimbursement): $61.53
b. 2015 Medicare Reimbursement Pro (professional reimbursement): $61.53
A. History. Early iterations of 3D echocardiography relied on acquisition of 2D images with respiratory and electrocardiographic gating, in addition to information regarding transducer position and orientation. These data were then reconstructed offline to create a 3D image from a 2D data set. This method was time-consuming and reliant on specialized processing systems, thereby relegating the technique to largely research-based purposes.
The ultimate success of 3D echocardiography technology centered on the progression from a sparse-array transducer to a full-matrix-array ultrasound transducer. This had previously been one of the most significant limiting factors in the roll out of this new echocardiographic application. Previously, the prohibitive size and limited number of crystal elements able to be incorporated into a reasonably sized 3D transducer limited spatial resolution and resulted in suboptimal image quality. These piezoelectric crystals both generate and receive ultrasound waves and are integral to ultrasound imaging. After substantial development and iterations of probes, new-generation transducers now incorporate upward of 3,000 tiny piezoelectric crystals, laser cut into equal squares and arranged in a matrix, thereby significantly enhancing image quality (Fig. 19.1
C. Cost. Each 3D-capable echocardiographic machine has its own vendor and, sometimes, model-specific matrix-array transducer. Given the sophisticated microtechnology incorporated into these transducers, 3D probes are substantially more expensive than the corresponding standard 2D probes. Ongoing optimization of 3D matrix design has resulted in a progressive reduction in probe size with continued improvement in image quality.
D. Image processing.
In addition to the transducer elements within the echocardiographic probe, newer probes also incorporate signal processing. This feature enables
real-time online rendering of 3D images with a time resolution of approximately 50 ms (similar to a frame rate of 25 volumes per second). This allows performance of harmonic imaging, with better tissue penetration and contrast resolution. This online 3D rendering technique has superseded the offline 3D reconstruction technique for transthoracic imaging. However, 3D transesophageal echocardiography (TEE) remains in evolution and still relies on some 3D reconstruction techniques to compensate for the limitations of the miniaturized transesophageal probe. Online real-time 3D image availability using TEE remains a priority, particularly for valvular heart disease, where immediate image visualization can enable on-the-spot procedural guidance and intraoperative assessment.
FIGURE 19.1 Full-matrix-array transducer, containing upward of 3,000 piezoelectric crystals cut into equal squares and arranged in a matrix, which both generate and receive ultrasound waves.
Current 3D echocardiographic rendering can be performed using three main methods (Fig. 19.2
1. Narrow-angle, real-time 3D imaging. The first method involves online, real-time image acquisition of a specific volumetric sector, typically narrow angle ranging from 30 to 50 degrees in size. This method is particularly well suited to focused examination of smaller structures such as cardiac valves and can be zoomed or magnified, depending on the targeted spatial resolution. By limiting the size of the field of view, frame rate and imaging resolution can be optimized.
2. Full-volume 3D imaging.
The second technique involves imaging of larger cardiac structures, whereby a full-volume data set needs to be acquired. Some newer platforms have introduced single-beat full-volume imaging, but traditionally this has been achieved by acquiring the data in four sectors during four consecutive heart beats, while maintaining the transducer in exactly the same position. The
four sectors of data are then combined to create a 3D data pyramid, although stitching artifact between these sectors can be problematic if there has been patient motion or variability in heart rate. Once the data pyramid is constructed, it can be cropped and manipulated on-cart or offline according to the structure of interest.
FIGURE 19.2 Ultrasound signals from the matrix-array transducer can be reconstructed to create a three-dimensional echocardiographic image. The narrower the image sector, the higher the frame rate and image resolution. (Redrawn with permission from Houck RC, Cooke JE, Gil EA. Live 3D echocardiography: a replacement for traditional 2D echocardiography? AJR Am J Roentgenol. 2006;187:1092-1106.)
3. Multiplanar reconstruction. Last, multiplanar reconstruction (MPR) is an accurate and reproducible technique for quantification of detailed measurements including dimensions, circumference, and areas. This is particularly useful for assessment of preprocedural planning whereby patient’s anatomical suitability can be established and prosthesis sizing can be determined. Different vendors have alternative iterations of this technique and reconstruction software, but primarily all involve volume segmentation of an image in three axes (x, y, and z). The positions of these axes are individually determined by the operator on the basis of the focus of interest, which is then demonstrated in a reconstructed 3D image. This can be performed on-cart or offline using postprocessing software on a remote workstation. Some aspects can be performed in real time, although more comprehensive quantitative analysis of measures such as length, area, and volume may require postprocessing. This technique is ideal for precise measurements of valve or annular areas, proximal isovelocity surface area (PISA), and vena contracta planimetry.
It must always be emphasized that 3D imaging will be suboptimal in any situation where 2D image quality is limited. In most cases, appropriate acoustic windows for 3D imaging are even more challenging to achieve owing to the larger and typically more cumbersome 3D probe.
B. Three-dimensional transesophageal echocardiography. Real-time 3D TEE was made available for routine clinical use in 2007. The utility of 3D transthoracic echocardiography (TTE) for imaging of the mitral valve was first confirmed by publication in 2008. 3D TEE has much higher spatial resolution than 3D TTE and, as such, has become the mainstay of valvular imaging owing to its superior anatomic detail. When imaging the mitral valve by 3D TEE, the established convention is to display the valve from its left atrial aspect (surgeon’s view), with the adjacent aortic valve orientated to the 12 o’clock position. This common logical approach has been advocated to create uniformity between studies and facilitate better communication between echocardiographers and surgeons. However, the ventricular aspect with a wide angle acquisition may be advantageous for assessment of chordae tendinae and papillary muscles, if these structures are not optimally visualized from a narrow-angled acquisition approach.
C. 3D color Doppler. Optimal real-time 3D color Doppler imaging remains a technical challenge, particularly for TEE imaging owing to lower frame rates. Currently, acceptable temporal resolution with single heart beat acquisitions requires a frame rate of at least 16 volumes per second.
D. Different vendors. Each echocardiographic platform vendor has specific 3D technical features for image acquisition, optimization, and postprocessing. Generally, all include options for full-volume acquisition, real-time narrow-angle imaging with zoom feature, MPR techniques, 3D color Doppler, and cropping tools. The nomenclature for these features is vendor specific, which can lead to confusion for users. However, overall, the available technical options between platforms and software remain relatively similar in scope and quality. Currently, some variations exist between vendors with respect to 3D color Doppler frame rate and the ability to acquire full-volume data sets in a single beat. However, this variation appears unlikely to persist for long, with further updates in hardware and software always in development.
The quality of 3D echocardiography is dependent on the intrinsic quality of the ultrasound images and is subject to the same imaging artifacts faced in 2D imaging. These include technical factors, which can be more pronounced in 3D imaging and are related to machine gain settings, as well as patient issues involving electrocardiogram gating, respiration, and poor acoustic windows. Undergaining and
overgaining can both impact the quality of 3D images, resulting in image dropout and blurriness, respectively. As outlined earlier, 3D multibeat acquisition techniques can be limited by stitch artifacts related to patient, respiratory, cardiac, or transducer motion. However, this is becoming less of an issue as more advanced and powerful 3D platforms are enabling single-beat acquisition of complete data sets for volume-rendered analysis, albeit at lower frame rates.
B. Frame rate. Currently, the relatively lower frame rates of 3D echocardiographic imaging detract from its spatial and temporal resolution. This means that some thinner and smaller structures may still be better delineated by 2D imaging at higher frame rates. Ongoing optimization of microprocessor technology is ameliorating this issue, by enabling further miniaturization of transducer size, along with faster data processing time.
C. Real-time imaging. Current 3D processors and matrix-array transducers are only able to perform real-time 3D imaging with relatively small (narrow-angle) fields of view. Ongoing development in this field will no doubt enable wider angle real-time imaging in the future. This will place less reliance on acquisition of full-volume data sets, which require time-consuming postprocessing, cropping, and analysis, without the clinical convenience and utility of real-time imaging, particularly for intraprocedural imaging.
D. Cost. The current cost of high-end echocardiographic machines with 3D capability, along with 3D probes, currently makes them more expensive relative to 2D-only platforms. Reimbursement for 3D acquisitions may negate some of this cost over time, but the setup costs or leasing plans may remain too prohibitive for some small centers or practices. Ideally, if a center could secure at least one machine with 3D echocardiography capability, this could be prioritized for valvular cases where arguably the most clinical benefit can be derived.
VII. STANDARD VIEWS
A. Imaging planes. Unlike conventional 2D echocardiographic imaging, 3D imaging is not governed by the same defined imaging planes. That being said, for real-time imaging using a narrow-angle field of view and 3D zoom mode, it is preferable to have the ultrasound beam perpendicular to the structure of interest. The resultant en face perspective (surgeon’s view) allows assessment of the valve of interest in its entirety.
B. Transthoracic. Typically, images are initially optimized in two dimensions in the standard apical and parasternal acoustic windows. Activation of 3D mode imaging is then performed, with prespecification of the desired image field size (narrow or wide angle). A full-volume data set can be acquired for on-cart or offline manipulation, although typically this is used for larger structures including cardiac chambers for volume quantification. Valvular imaging is usually best performed with narrow-angle, real-time imaging to optimize frame rate. The images can then be manipulated or cropped to demonstrate the structure of interest in optimum orientation. Once adequately displayed, the 3D image can then be zoomed and acquired.
C. Transesophageal. 3D TEE imaging is also based on 2D images taken in the conventional planes. For valvular imaging, these are typically midesophageal views acquired at 0, 45, 60, 90, and 120 degrees. Once these images are optimized, the operator then switches to 3D imaging mode. Typically, 3D TEE, narrow-angle, real-time imaging is preferred as the valvular structures are relatively small, and this allows the frame rate to be enhanced. Once in 3D mode, the images can be cropped, rotated, and manipulated into nonconventional, off-axis orientations to adequately align the structure of interest en face and in its entirety.
VIII. CLINICAL APPLICATIONS—3D TRANSTHORACIC AND TRANSESOPHAGEAL ECHOCARDIOGRAPHY
In theory, 3D echocardiography can be employed for assessment of all valves and paravalvular structures; however, most incremental benefit has been achieved by imaging the mitral valve with 3D TEE.
A. Mitral valve
2D echocardiographic assessment of the mitral valve is complicated and multifaceted, perhaps reflecting that there is no good single method for quantification of either mitral stenosis or regurgitation severity. It requires identification of the pathology from multiple separate views, with subsequent extrapolation and mental reconstruction of the identified abnormality in three dimensions. Potentially, this type of imaging can also result in suboptimal visualization of the abnormality if it is not apparent on one of the standard 2D imaging planes.
Now for the first time with echocardiography, 3D imaging allows us to examine the mitral valve and subvalvular apparatus in its entirety and in any orientation relative to its surrounding structures, from both the atrial (so-called surgical view) and ventricular aspects. This obvious superiority in technique has been repeatedly recognized with both 3D TTE and TEE imaging. Thanks to 3D imaging, we now have a better appreciation of the saddle shape of the normal mitral valve annulus and asymmetric or symmetric alterations in this geometry, which can occur with different disease states such as ischemic or dilated cardiomyopathy, respectively. Greater anatomic detail also allows us to more precisely identify and localize valvular abnormalities, such as prolapse and flail of specific scallops, without rotational artifacts. Regions of prolapse are demonstrated as “bulging” segments into the left atrium (Fig. 19.3
). Volumetric reconstruction can then color-code the region of prolapse to differentiate it from the normal surrounding leaflet. This technique is particularly useful for commissural pathology, which can be difficult to appreciate on 2D imaging. Ultimately, more accurate localization of pathology should result in better preoperative planning and potential interventional strategies. This can also influence the selection of surgeon and surgical technique, depending on the complexity of the lesion and the likelihood of successful valve repair.
By incorporating 3D color Doppler, we also have a better understanding of mitral valve dynamics. Not only will this likely influence procedural success, but, importantly, it may also impact upon the rates of successful valve repair able to be achieved.
FIGURE 19.3 Three-dimensional transesophageal echocardiography en face reconstruction of the mitral valve demonstrating prolapse of the lateral and to a lesser degree the middle posterior leaflet scallops (P1 & P2) (arrow).