Principles of Three-Dimensional Ultrasound

Fully Sampled Matrix Array Transducers

Considerable advancements in hardware and software, involving microelectronic techniques, image formation algorithms, and digital processing, have led to the development of fully sampled matrix array transducers, which enabled the volumetric 3DE acquisition with good imaging quality within a short acquisition time ( Fig. 5.1 ). At present, matrix array transducers are composed of nearly 3000 piezoelectric elements (as opposed to only 128 elements in a conventional 2DE phased-array transducer), with operating frequencies ranging from 2 to 4 MHz for TTE and from 5 to 7 MHz and TEE imaging. The piezoelectric elements are arranged in rows and columns to form of a rectangular grid (i.e., matrix configuration), individually connected and simultaneously active (fully sampled). The electronically controlled phased firing of the piezoelectric elements enable to generate a scan line that propagates radially (y, axial direction) and can be steered both laterally (x, azimuthal direction) and in elevation (z, vertical direction) to acquire a pyramidal volumetric data set (see Fig. 5.1 A and B ).

Currently matrix-array probes are available for both TTE and TEE imaging and, in addition to the conventional 2D-Doppler imaging, they enable three different acquisition modalities: multiplane 2DE imaging, real-time (or live) 3DE imaging, and multibeat ECG-gated 3DE imaging, all three with/without color flow information ( Chapter 10 , “3D Image Acquisition”).

Previous 3DE equipment could only acquire and display in real-time volumetric data sets of a relatively small size (about 30° × 50°). These pyramids were sufficiently large to allow a partial display of the ventricles or of the valvular structures; the larger volumes needed to encompass the whole structure required at least four smaller component volumes acquired over a series of consecutive cardiac cycles to yield a 90° × 90° image (see Fig. 5.1C ). However, the technology evolved and the current 3DE systems have the capability of acquiring and displaying single-beat volumes as 90° × 90° pyramids in real time (see Fig 5.1A ), with improved temporal and spatial resolution (even though significantly lower than by multi-beat acquisition).

A major technological breakthrough that allowed manufacturers to develop fully sampled matrix array transducers was the electronics miniaturization and microbeamforming. Several miniaturized circuit boards have been incorporated into the matrix-array transducer, allowing partial beamforming to be performed in the transducer itself, reducing both power consumption and the size of the connecting cable. In addition, more advanced crystal manufacturing processes (such as the PureWave crystal technology), by increasing the efficiency of transduction and of conversion process between electrical power and ultrasound energy, helped reduce heat production.

Multibeat ECG-Gated Acquisition

This technique uses a number of ECG-gated subvolumes acquired from consecutive cardiac cycles (see Fig. 5.1C ), stitched together in position and size, to increase the size of the pyramidal volume maintaining the volume rate (>30 Hz). Nevertheless, the pyramidal volume should be optimized to the smaller volume capable to encompass the cardiac structure of interest, for the highest spatial resolution. The main limitation of gated imaging is the occurrence of stitching artifacts, shown as subvolume malalignments in the pyramidal imaging, which may impede proper image interpretation or quantification. Those can occur by transducer movement, cardiac translation motion due to respiration, or change in cardiac cycle length (arrhythmias; Fig. 5.2 and Video 5.1 ).

Real-Time Zoom Acquisition

Real-time zoom mode is ideal for the study of a restricted structure of interest, with high spatial and temporal resolution. The operator can adjust the area of acquisition with the minimum lateral and elevation width, and the ultrasound system automatically crops the adjacent structures to provide a real-time display of the structure of interest. If the structure of interest is large, the temporal resolution can be further increased by using a multibeat zoom acquisition. The main disadvantage of real-time zoom is the limited ability to show the anatomic relationships between the structure of interest and surrounding structures. However, it is of great use to study the mitral valve, particularly by TEE, and whenever a large ECG-gated acquisition is not feasible (arrhythmias, patient unable to breath hold, etc.; Fig. 5.3 and Video 5.2, Video 5.3 ).

Parallel Receive Beamforming

Beamforming is a technique used to process signals so that directionally or spatially selected signals can be sent or received from sensor arrays. Parallel receive beamforming uses a high-end imaging engine, wherein the system transmits one wide beam and receives multiple narrow beams in parallel, and requires a platform with higher volume/information rate. The transmit beamformer transmits timed pulses thousands of times per second, and the receive beamformer generates multiple beams through parallel and real-time processing of echo. The number of receive beams generated in parallel by the receive beamformer determines the maximum information rate the imaging system can achieve. In this way, the volume rate/temporal resolution ratio is increased by a factor equal to the number of received beams. Nevertheless, the finite speed of sound through tissue imposes a physical limit on the maximum number of pulses that can be fired per second, as it takes hundreds of microseconds for a round trip pulse to reach the deepest depth, and then to propagate back to the acoustic array. Conversely, the size of the volume, target volume rate, and lateral resolution determines the total number of beams needed per second. For example, to obtain a full volume (90° × 90°), 16-cm depth pyramidal volume at 25 volumes/s, the system needs to receive 100,000–200,000 beams/s, depending on the lateral resolution. Since the emission rate is around 5000 transmitted pulses, the system needs to form 20 to 40 beams in parallel for each transmit pulse, and even higher for real-time imaging of motion. However, there are limits to the increase in the number of parallel beams, since it leads to an increase in size, cost, and power consumption of the beamforming electronics, as well as deterioration of the signal-to-noise ratio and contrast resolution. To overcome this problem, manufactures have been developing innovative technologies for beamforming, image forming, and processing (nSIGHT imaging, Philips Healthcare; XDclear and cSound, GE Healthcare; coherent volume formation, Siemens Healthcare; etc.), which allow us to increase the spatial resolution, clarity, and overall quality of the 3DE images, and to reduce noise.

Similar to what occurs with conventional 2DE imaging, the use of 3D color Doppler further impacts the spatial and temporal resolution of the 3DE images. The use of 3D color Doppler in real time is limited by lower frame rates and by the smaller volume size. Conversely, the multibeat 3D color full-volume can be affected by stitching artifacts. The TEE approach for 3D color acquisition provides higher spatial resolution and superior image quality than TTE. The limited spatial and temporal resolution of 3D color Doppler used to be a major limitation, but it has greatly improved with the advancing technology, and color Doppler volumes can be currently obtained with up to 40 voxels/s ( Fig. 5.4 and Video 5.4 ).

Finally, the image quality of the 3DE data sets is also affected by the point spread function. The point spread function describes the imaging system response to a point input, which is represented as a single pixel. The degree of spreading (blurring) of any point object varies according to the dimension employed. In current 3DE systems, the degree of spreading (blurring) is around 0.5 mm in the axial (y) dimension, 2.5 mm in the lateral (x) dimension, and 3 mm in the elevation (z) dimension. As a result, higher resolution images should be expected when acquired using the axial dimension (i.e., parasternal views), rather than the elevation dimension (i.e., apical views), and when acquiring the structure of interest in the center of the pyramidal volume sector, rather than in an eccentric position.

As a consequence, to obtain effective 3DE images, it is pivotal that the appropriate acquisition window and modality are selected, and that acquisition settings are optimized according to the desired structure of interest.