Volume ultrasound competes with multiplanar CT and MRI

Despite its widespread use as a tool that provides rapid diagnosis at a relatively low cost without the need for bulky equipment or ionizing radiation, ultrasound faces mounting competition from modalities such as CT and MRI, which combine short acquisition times with the ability to rapidly generate multiplanar and 3D images. That may change with volume ultrasound, a technique that lets clinicians and sonographers scan the patient and rapidly analyze data from a volume of interest.

As with CT and MRI, the acquired volumes may be interrogated to produce images in any arbitrary plane, as well as provide static and dynamic 3D renderings. These capabilities, in combination with new quantification techniques, promise to spark renewed interest in ultrasound by radiologists, clinicians, and sonographers.

In conventional ultrasound, the operator acquires a series of 2D static images and real-time clips to evaluate a region of interest. At minimum, structures are viewed in two orthogonal planes (usually sagittal and transverse), although other intermediate planes are often imaged as well. However, 2D images and clips do not permit examination of structures in planes that cannot be directly interrogated by the ultrasound beam, nor do they provide 3D representations of anatomy.

Volumetric ultrasound has the potential to scan large anatomic areas in seconds, producing a data set that can then be manipulated offline to produce multiplanar and 3D images. Depending on the lab setup, staffing, and other considerations, the volume postprocessing may be performed by a specially trained sonographer or other personnel.

VOLUME ACQUISITION

Several vendors offer volumetric ultrasound functions on their equipment. Generally, this capability is limited to top-of-the-line ultrasound platforms, although volume ultrasound will probably become available on less expensive equipment as time goes on. Currently, volumetric ultrasound data may be acquired in three different ways:

• Freehand acquisition. In this method, the operator uses a conventional transducer to sweep a 2D ultrasound plane through a volume of interest, usually by sliding it across the patient's skin or by angling it from side to side. The ultrasound platform essentially stacks the acquired images to produce a volume. This technique offers the advantage of not requiring specialized probes, but, because it does not provide spatial reference information, no measurements can be made from the acquired data. Some researchers and manufacturers have attempted to rectify this shortcoming with electromagnetic or other position sensors that can precisely localize the transducer in space, but this approach is cumbersome and difficult to implement in clinical practice.

• Mechanical acquisition. This technique employs specially constructed probes with embedded electric motors that translate or rotate the transducer to scan through a volume. Unlike freehand-acquired data, the resultant volumes may be quantified. However, these probes tend to be bulky, and they are limited by the speed at which the scanning plane can be changed mechanically.

• Electronic acquisition. This new approach uses fully sampled matrix-array transducers that operate completely electronically. Because they do not have any moving parts, the ultrasound scanning plane can effectively be angled or moved instantaneously. Therefore, these probes are ideally suited to applications that require high frame rates, such as true real-time 3D or live volume sonography.

3D VISUALIZATION

Planar displays, whether they take the form of a viewbox, a cathode ray tube monitor, or an LCD, are well-suited for displaying conventional 2D static or motion images. Volume ultrasound poses unique challenges, which ultrasound vendors have met in several ways:

• Multiplanar displays. In this approach, the plane in which the region of interest was scanned (the acquisition plane) is shown along with two planes that are generated from the data set. The three planes are typically orthogonal to each other, but the radiologist may view any arbitrary plane by adjusting controls on the

display software. This permits plane interrogation that would not be possible in vivo. For example, true coronal imaging of the aorta may not be feasible if the bowel or other structures obscure the aorta from a vantage point in the flank. With multiplanar displays, if the aorta is visible using an anterior approach, the 3D software can generate a coronal image.

In one variant of the multiplanar view, the ultrasound system displays real-time ultrasound in two orthogonal planes on the screen simultaneously. This capacity is available only with the newer electronic matrix transducers because the rapid switching of planes required for this technique is impossible with mechanical transducers.

• Automatic slicing. This is a variant of the multiplanar display in which the software takes the acquired volume and produces a series of parallel slices. By manipulating interactive controls, the operator can adjust the number and thickness of the slices to suit the region of interest.

The goal is to produce static displays that are similar to image displays that have been used in ultrasound since the days of articulated arm scanners. However, like the stacked image displays that are familiar to radiologists who interpret CT and MRI, the plane of a section may be set to any angle needed.

• Shaded surface displays. Whereas multiplanar views and automatic slicing provide 2D representations of anatomy, shaded displays attempt to represent 3D structures based on volumetric ultrasound data. Because these images must be viewed on the same flat monitors as 2D sonograms, software engineers use gray-scale and color shading techniques to give an illusion of depth. To a large degree, the fidelity of the illusion is dependent on the extent to which adjacent structures differ in echogenicity, which explains why the presence of fluid within or adjacent to the area of interest is so conducive to 3D ultrasound imaging.

In effect, fluid acts as a negative contrast medium that permits the visualization software to map margins more precisely. This advantage is particularly evident in certain clinical applications such as fetal and vascular imaging, as well as in the evaluation of fluid-filled organs such as the gallbladder and the urinary bladder.

• Live volume imaging. Live volume imaging produces 3D images that are updated in real-time. Electronic matrix-array ultrasound transducers are particularly suited to this technique, since they are capable of generating volumetric data at high rates. This approach has already had an impact in echocardiography by permitting cardiologists to obtain views of the heart not previously attainable. In some instances, structural abnormalities or alterations of valve or wall motion are more conspicuous with this technique. Live volume imaging is also being widely used in obstetrics.

CLINICAL APPLICATIONS

It is not surprising that 3D ultrasound has already gained a strong foothold in obstetric imaging: The fetus is complex, and the ability to visualize it in 3D is clearly beneficial. Moreover, the amniotic fluid that surrounds the fetus provides an excellent environment for volumetric imaging.

• Quantification of volume. Clinicians have long used conventional ultrasound to calculate the volume of fluid-containing structures. However, these values are only estimates, because they are based on linear measurements and make geometric assumptions that are not always accurate. With volumetric ultrasound, it is possible to directly measure volumes, such as the capacity of the urinary bladder both prior to and after voiding. This technique may also prove useful in other organs, such as the gallbladder, as it can assess contractility after a meal or administration of a pharmacologic agent.

Direct estimation of the volume of neoplasms using volumetric ultrasound will probably see initial application in cases in which the lesion under consideration protrudes into a fluid-filled organ or structure. However, even lesions within solid organs such as the liver can be segmented and their volume estimated using volumetric analysis software (Figure 1).

Moreover, as volumetric capabilities migrate to higher frequency linear-array transducers, volume techniques may prove to be helpful in superficial organs such as the breast and thyroid gland. For example, analysis of surface contour may be a valuable adjunct to assessment of internal architecture in the appraisal of indeterminate breast masses and thyroid nodules (Figure 2).

• Guidance for intervention. Although conventional ultrasound usually suffices to guide placement of needles and other instruments during interventional procedures, 2D displays can be misleading. For example, a biopsy needle may appear to be completely surrounded by liver parenchyma when imaged from one vantage point, when in reality it

passes very close to the capsule. Live multiplanar and live volume imaging can help by letting the radiologist view the instrument's path from multiple perspectives simultaneously (Figure 3).

This is particularly important when accessing lesions that are located close to vital structures such as the diaphragm and major blood vessels, especially during procedures such as radiofrequency ablation in which the instrument "tip" is spatially complex.

• Other clinical applications. Clinicians are becoming increasingly convinced of the ability of 3D sonography to shorten acquisition times or to increase diagnostic confidence in normal and abnormal cases. Volume sonography is also proving its worth in the evaluation of the nongravid uterus, particularly because of its ability to depict the organ coronally no matter what its anatomic orientation, making it possible for the first time to distinguish various types of congenital fusion anomalies, such as bicornuate uterus and uterus didelphys.

Evaluation of vascular structures is another ideal application for volumetric ultrasound, as flowing blood provides excellent contrast with the enclosing vessel walls. Aneurysms, varices, plaques, and dissections may be viewed in arbitrary planes and accurately measured and characterized. Volumetric imaging, in conjunction with color or power Doppler sonography, also affords 3D views of vessels and stents that could not previously be obtained without CT or MRI (Figure 4).

While its ability to provide views not achievable with conventional techniques will clearly be valuable in many clinical applications, the potential effects on workflow in many ultrasound labs may turn out to be even greater in the long run.

Dr. Tessler is a professor of radiology and chief of body imaging at the University of Alabama at Birmingham. Mr. Brown is senior director of technical and clinical marketing for Philips Medical Systems, Ultrasound.