Three-dimensional (3D) transesophageal echocardiography (TEE) is subject to the same types of artifacts encountered on two-dimensional TEE. However, when displayed in a 3D format, some of the artifacts appear more “realistic,” whereas others are unique to image acquisition and postprocessing. Three-dimensional TEE is increasingly used in the setting of percutaneous catheter-based interventions and ablation procedures, and 3D artifacts caused by the metallic components of catheters and devices are particularly frequent. Knowledge of these artifacts is of paramount relevance to avoid misinterpretation of 3D images. Although artifacts and pitfalls on two-dimensional echocardiography are well described and classified, a systematic description of artifacts in 3D transesophageal echocardiographic images and how they affect 3D imaging is still absent. The aim of this review is to describe the most relevant artifacts on 3D TEE, with particular emphasis on those occurring during percutaneous interventions for structural heart disease and ablation procedures.
Three-dimensional (3D) transesophageal echocardiography (TEE) provides high-quality images of cardiac structures in a 3D format that appear identical to anatomic specimens. In percutaneous catheter-based interventions and ablation procedures, 3D TEE is increasingly used for its ability to intersect long segments of catheters, tips, and devices without excessive probe manipulations; moreover, in most cases, the entire scenario in which catheter-based interventions and electrophysiologic procedures take place can be shown in a single 3D view.
Three-dimensional TEE is subject to the same types of artifacts encountered on two-dimensional (2D) TEE. However, when displayed in a 3D format, some of the artifacts appear more “realistic” (i.e., dropout artifacts of the atrial septum or aortic leaflets resemble real holes), whereas others are unique to 3D image acquisition and postprocessing. In the setting of percutaneous interventions and ablation procedures, artifacts caused by the metallic components of catheters and devices are particularly frequent. Reverberations, shadowing, and blooming may alter 3D images of catheters or cover surrounding structures. Moreover, excessive undergain or overgain may significantly distort 3D images and reduce their quality. Knowledge of these artifacts is of paramount relevance to avoid misinterpretation of 3D images. Artifacts and pitfalls on 2D echocardiography are well described and classified. However, despite several exhaustive reviews of 3D echocardiography, a systematic description of artifacts on 3D TEE and how they affect images is still absent. The aim of this review is to describe the most relevant artifacts on 3D TEE, with particular emphasis on those occurring during percutaneous interventions for structural heart disease and ablation procedures. Table 1 summarizes these artifacts, their mechanisms, their impact on images, and how to avoid or minimize them.
| Type of artifact | Acquisition | Mechanism | Impact on images | How to avoid/minimize |
|---|---|---|---|---|
| Stitching artifacts | Full volume | Incorrect juxtaposition at the interface of sequential subvolumes (because of arrhythmias, breathing, probe/patient motion) | Strong demarcation between subvolumes leading to a “broken” image | Ask for breath holding and maintain still the position of probe |
| Dropout artifacts | Any | Poor echocardiographic signal strength due to weak echoes | These artifacts can be misdiagnosed as true holes/perforations | Maintain slightly oversetting the gain and surfaces perpendicular to ultrasound beam |
| Blurring artifacts | Any | Indistinct edges of structures due to the assembly of nonisotropic voxels | Thin structures (i.e., chordae tendineae) appear thicker than they actually are | Use perspectives in which the 3D image is created using mostly the axial resolution |
| Blooming artifacts | Any | Metallic structures when intersected by ultrasound produce fringes extending beyond the borders of the metallic devices/catheters | Metallic structures appear with irregular, thick edges | No suggestions |
| Railroad-shaped artifacts | Any | In large catheters with wide lumens, two surfaces are perpendicular to the ultrasound beam, producing strong echoes, while the other two are tangential, producing very weak echoes | Single catheter appears as two linear structures | Obtain the most favorable perspective (i.e., the “en face” view of the surface perpendicular to the beam), and increase compression to merge boundaries |
| Reverberations | Any | Multiple reflection of metallic component of catheters | Depending on the perspective and the position of catheter, reverberations may appear to lengthen the catheter | No suggestions |
| Shadowing | Any | Inability of ultrasound to pass through strong reflecting catheters/devices | Lack of tissue posterior to catheters/devices that may appear as a “tear” of cardiac structures | Cover the shadowing by properly rotating/angulating the volumetric data set |
| Gain | Any | Variation of gain may produce significant variation in the size of structures | Orifices may appear larger or smaller according to gain variation | No suggestions |
Stitching (Reconstruction) Artifacts
The full-volume data set acquisition in most instruments uses electrocardiographic gating to capture a large volumetric data set over sequential cardiac cycles by advancing sector by sector (or subvolume). A correct acquisition requires regular heart rhythm and breath-holding to obtain the highest 3D spatial and temporal resolution with a volume rate > 50 Hz. This is due to the inverse relationship between volume rate (i.e., the number of volumes scanned per second), volume width, and spatial resolution (high number of ultrasound lines). Because decreasing the volume width automatically increases volume rate and the number of ultrasound lines, narrow subvolumes have high temporal and spatial resolution. In full-volume acquisition, the final image is a summation of these narrow subvolumes. The large full volume maintains the same high temporal and spatial resolution ( Figure 1 , Video 1 ; available at www.onlinejase.com ). The price paid, however, is the lack of true real-time imaging (because the full image is unavailable until the last cycle is acquired) and the frequent occurrence of stitching artifacts. Stitching (or motion) artifacts are due to incorrect juxtaposition at the interface of the subvolumes. The positions of subvolumes may vary between sequential beats because of translation after respiration, different lengths of cardiac cycle, irregular heart rhythm, or any transducer or patient movement ( Figure 2 ; Videos 2 and 3 ; available at www.onlinejase.com ). Such artifacts not only affect comprehension of structures but also can result in difficulties in quantification. Thus, in any electrocardiographically gated full-volume acquisition, the patient must stay very still, suspending breathing for a few seconds, and the operator must stabilize the probe during the entire acquisition.
Dropout Artifacts
Dropout artifacts may be defined as a loss of ultrasound data on 3D surface images due to poor echocardiographic signal strength. They typically appear as a lack of tissue on 3D images. The most common dropout artifacts occur in the interatrial septum (IAS) and aortic leaflets and may be misinterpreted as real holes simulating atrial septal defects and leaflet perforations.
Dropout Artifacts of the IAS
The unique ability of 3D TEE is to image the surface of the IAS (the en face view). One of the easiest ways to obtain the IAS in an en face view is from the midesophageal 2D transesophageal bicaval view. From this position, the septum is nearly perpendicular to the ultrasound beam, and any 3D modality of acquisition (zoom mode, full-volume single beat, and full-volume multibeat) provides high-quality images of the left and right aspects of the septum.
Not infrequently, the area of the fossa ovalis may appear to have a lack of tissue ( Figure 3 A, Video 4 ; available at www.onlinejase.com ). This happens when the fossa ovalis is not completely perpendicular to the ultrasound beam ( Figure 3 B) and consequently reflects scattered, weak echoes, which create apparent holes in the tissue. These artifacts may be more evident in 3D than 2D format, because they also broaden in the third dimension. A second reason may be the thinness of the fossa ovalis compared with the surrounding structures, which may contribute to producing echocardiographic signals of low intensity. Fine tuning the gain setting and/or a slight change in probe position to optimize the biplane preview with the fossa ovalis perpendicular to the ultrasound beam may reduce these artifacts. However, dropouts usually disappear only when the gain is slightly overset and fine static noises cover the hole. This might result in poor-quality display. Borders of these apparent holes are rather indistinct, which may be one clue in distinguishing them from real atrial septal defects.
Dropout Artifacts of the Aortic Leaflets
Even in the most favorable conditions, acquisition of aortic valve images in diastole may present some dropout artifacts in the leaflet bodies, which look like real holes ( Figure 4 , Videos 5 and 6 ; available at www.onlinejase.com ).
Once again, these artifacts are due to the nearly parallel position the leaflet bodies assume in diastole relative to the direction of the ultrasound beam. Only weak, scattered echoes are generated from these portions of the valve, and they scarcely reach the transducers. In contrast, this phenomenon does not occur in systole, because the leaflet bodies become perpendicular to the ultrasound beam and are imaged in their entirety ( Figures 4 B and 4D). As in the case of the IAS, these artifacts are more evident in 3D format because they extend into the third dimension. This is true even when 2D data are derived from a 3D data set, as in multiplanar reconstruction: despite the same gain and compression, the artifacts are more evident in three than in two dimensions ( Figure 5 ).
