Lectures

Focal liver lesion: nonlinear contrast-enhanced ultrasound imaging

Introduction

Ultrasound (US) is the most widespread imaging modality for screening focal liver lesions. Noninvasive diagnosis (detection and characterization) of focal liver lesions is still carried out in many centers by computerized tomography (CT) and magnetic resonance (MR) using contrast agents. Hepatic lesions have in effect characteristic perfusion and enhancement patterns during the various vascular phases, which assist characterization and in many cases make definite diagnosis possible. Conventional and Doppler US are often limited in diagnosing this kind of lesion as color Doppler vascular information of the lesion is limited.

Recently, the detection and characterization of focal hepatic lesions by US has increased considerably, due to the development of new nonlinear (harmonic and nonlinear fundamental) technologies that optimize detection of contrast agents. Furthermore, small lesions may not be well defined by CT or MR imaging because real-time visualization of lesion vascularization is not possible. Because contrast-enhanced US (CEUS) is a dynamic analysis that allows this, small lesions may be detected that are sometimes missed with CT or MR imaging. Positron emission tomography (PET) is proving to be a molecular imaging modality with very high sensitivity, but not competitive as the advantages do not outweigh the high cost of the exam.

Indeed, color Doppler is a method that can separate echoes from blood and tissue. This is possible because of the relatively high velocity of blood in comparison to that of the surrounding tissue. This distinction is valid for flow in large vessels where the Doppler signals of blood flow may be separated from those of the tissue clutter using a wall filter. This fails for the parenchymal micro-vessel flow where blood velocity is similar or slower than that in tissue. In fact in the latter case the Doppler shift frequency coming from the moving tissue is comparable or higher (up to a thousand times) than that of the blood which perfuses it. Thus, when we analyze together tissue and the blood in it, wall filters eliminate both the flow and the clutter echoes. “Flash” artifact in color and “thump” artifact in spectral Doppler represent overwhelming signals from tissue movements which corrupt image information.

Ultrasound contrast media are suspensions containing air or gas microbubbles that increase reflection of the US beam used during the investigation. Contrast agents were previously used as echo-enhancers as they amplified blood reflectivity and, therefore, increased the signal-to-noise ratio for Doppler signals amplifying blood reflectivity and increasing the signal-to-noise ratio. This enhances the echogenicity of the blood but creates more artifacts due to blooming and tissue motion (Powers et al., 1997). Another, more efficient way to avoid artifacts would be to obtain a contrast B-mode image that suppresses Doppler artifacts without the use of a velocity-dependent filter. Contrast- enhanced ultrasound, defined as nonlinear or harmonic imaging, aims to provide such a method.

HARMONIC IMAGING: BASIC PRINCIPLES

Harmonics are additional frequencies that are multiples of the fundamental (transmitted) frequency and are commonly found in acoustics. Double and fourfold overtones (harmonics) give character to the different musical instruments. The same phenomenon occurs in ultrasound (Hamilton and Blackstock 1998). We need to distinguish two different modes of harmonic generation: those produced by transmission through the tissue and those produced by interaction with microbubbles . Tissue harmonics are derived from the nonlinear distortion of the incident beam as it passes through the tissues. The phenomenon occurs for US beams of a sufficiently high power. A well-known basic principle of US, the speed of propagation of an US beam, depends on the density of the tissue. The compression part of the US wave increases by fractions the density of the tissue, and thus this part of the wave travels marginally faster than the rarefaction part of the wave where the density is lower. Thus, over distance, the shape of the waveform becomes distorted and its angular components represent the overtones or harmonics. These can be filtered and used for imaging (Ward et al., 1997). The advantage is that they depend on a high acoustic pressure for their formation, and therefore are not produced as much by the weaker unwanted side and grating lobes or by reverberations than by the main lobe. Thus, tissue harmonic images are cleaner with fewer side lobe and reverberation artifacts.

The harmonics are generated from within the tissue and only at the center portion of the main US beam. This results in higher tissue contrast. The disadvantage is that the harmonic echoes are some 20dB less intense than the fundamental echoes with the risk of reducing the signal-to-noise ratio . This is why harmonics are better handled by newer generation systems powered with a high dynamic range than those with older technology-designed scanners. When it works well, harmonic imaging is spectacularly useful in sharpening up images since it increases contrast  by eliminating noise,  especially in the evaluation of the fluid-filled structures such as the biliary tree, gallbladder and vascular structures (Ortega et al., 2001). In clinical practice, for most scanners the benefit in abdominal and general work is so great that most users now set the systems to default in a tissue harmonic mode; however, in our opinion, harmonic mode is more useful in the evaluation of fluid-filled structures positioned in the middle field because a penetration loss occurs in the far field with the poor echo intensity of deeper tissue second harmonics (Migaleddu et al., 2002).

Microbubble nonlinear (harmonic) responses can be produced in two different ways: microbubble oscillations or disruption . Microbubbles are air or gas bodies ranging from 1 to 7 μm in diameter, enclosed by some sort of membrane. Two important characteristics of microbubbles, stability and diameter, must be optimized in order to have efficient contrast media. Stability means that microbubbles of contrast medium must dwell in the blood stream long enough to allow the completion of an US investigation with contrast enhancement. Microbubbles must also be of a diameter necessary to obtain contrast medium that can pass through the pulmonary circulation and that, once in the systemic circulation, can act as a contrast-enhancing medium in all organs. Several cardiac and vascular US contrast agents are available on the American, Canadian, European or worldwide market; Echovist and Levovist (Schering), SonoVue (Bracco) are available in a number of European countries. Other agents that underwent or are still undergoing American regulatory approval include Optison (Amersham/GE Healthcare), Definity (BMS), Imagent (Alliance Pharmaceutical), Sonazoid (Amersham/GE Healthcare), Quantison (Quadrant Healthcare), SonoVue (Bracco), Biosphere (Ponit Biomedical), and AI-700 . Albunex and Echovist, used mainly in cardiac evaluations, are effectively one-pass-only agents and have been replaced by new-generation agents such as Optison, Levovist and SonoVue.

For radiology, the two most important are Levovist (Schering) and SonoVue (Bracco). Levovist is a first-generation contrast medium of air containing microbubbles made by shaking galactose micro-particles with water. The galactose micro-particles contain micro-defects, which force the attached air microbubbles into the required size. They are stabilized by a monomolecular layer of a surfactant, palmitic acid (Nanda and Carstensen, 1997). The second-generation contrast medium SonoVue uses an inert high-density gas (sulfur hexafluoride), which improves the longevity of the microbubbles because of its high molecular weight, which slows diffusion. Microbubble shell is constituted by a phospholipid layer which is similar to that of cell membranes (Schneider et al., 1995; Morel et al., 2000).The essence of their function as contrast agents is that microbubbles behave quite differently from solid or watery tissues in that they can be compressed and expanded much more readily. Thus, they change their diameter from two to tenfold, depending on the power of the incident US and, like all reactive processes, have a natural resonant frequency at which they respond most actively. It happens that the resonant frequency for 1-7 μm microbubbles lies in the 2-10 MHz range that is used for diagnostic imaging. This lucky coincidence explains the remarkable reflectivity of microbubbles which are many more times echogenic than comparable tissue elements such as red blood cells. (Cosgrove in Harmonic Imaging: tissues and microbubbles).

PHYSICAL AND TECHNOLOGICAL BACKGROUND

When an US beam encounters microbubbles present in an insonated field, different information is acquired depending on the pressure exerted by the US waves, which in turn is related to the mechanical index (MI) . The mechanical index is the tool which expresses the acoustic pressure of the US beam produced on insonated structures; it enables the power of the US beam to be quantified and is directly proportionate to the peak of maximum negative pressure at the point of focalization of the beam (P), and is inversely proportionate to the mean frequency of the US beam (V) in the following ratio:

M.I. = Formula

This quantity is related to the mechanical work that can be carried out on the microbubble during the negative cycle of US propagation (Apfel and Holland, 1991). With very low mechanical indexes reflection of ultrasounds occurs. For slightly higher mechanical indexes (up to 0.2) a nonlinear signal is emitted in relation to the resonance of the microbubbles that are not destroyed, and with a mechanical index greater than 1.2 the microbubbles break. This creates transitory echoes in second harmonics with a high signal intensity (Chin and Burns 1997).

First-generation contrast agents exploit the high MI and corresponding microbubble rupture by using intermittent software imaging. Fragile microbubbles that contain air and have a galactose shell stabilized by palmitic acid (such as Levovist or first-generation contrast agent) result in limited stability and shortened time available for examination. They behave like soap bubbles and the greatest amount of high intensity information in response is obtained the moment they break. Second-generation contrast agents (SonoVue) are made up of microbubbles with notable elasticity, containing a gas (sulfur hexafluoride or similar), and are wrapped in a shell of phospholipids and/or albumin. This enables real-time continuous-mode software to be used, exploiting the information from the resonance of the microbubbles that are maintained in time, though their duration is not unlimited and they break spontaneously after 8–10 min. Microbubbles like any gas particle resist compression more than they respond to expansion so that they may respond asymmetrically to symmetrically transmitted US waves. This asymmetry may take either the form of a change in the bubbles’ shape whereby they become geometrically distorted or of an asymmetry in the temporal response.The latter  is perhaps more usual because the bubbles expand more quickly than they compress. In either case the result is that the returning wave contains different frequencies from those of the insonant wave and if these can be extracted, images favoring the contribution of the microbubbles can be created (Greis, 2004).

A variety of algorithms that have been marketed involve enhancement of the signal coming from the microbubbles and cancellation of signals from tissues. They favor a reading of the echoes with specific frequency of contrast agent resonance in second harmonics, and at the same time cancel the fundamental echoes coming from the tissues. In pulse inversion, also known as phase inversion, two pulses in rapid succession are emitted into the tissue; the second pulse undergoes a 180° phase change so to be a mirror image of the first. The fundamental response produced in the tissues which will be phase-inverted compared with the fundamental pulses is erased because the sum of two inverted pulse is zero. In this way the echoes from the microbubbles are enhanced so as to differentiate them from those coming from stationary tissues (Burns et al., 2000). A great spatial and details resolution (capability to identify macro and micro vessels) is achieved by intermittent imaging and specific software – Agent Detection Imaging (ADI by Acuson Sequoia, Siemens) - with a high MI and first-generation contrast agent. The ADI can be used for a short time either with static techniques or in real-time (Migaleddu et al., 2004).

Compared with the intermittent imaging characteristic of the reconstruction algorithms used in breaking first-generation contrast agents, the algorithms based on the use of a low mechanical index enable signals from the resonance of second-generation microbubbles to be detected in real-time; this permits enhancement of the contrast resolution of the vascular phenomena present in an organ or in the lesions it contains. These methods, however, show low spatial and details resolution, useful for detecting peri- and intra-lesional macro- and micro-circulation due to the low intensity of signal coming from microbubble resonance.

The observation that there is also a nonlinear response coming from the contrast agent in the proximity of the fundamental frequency has enabled a new algorithm to be developed, which permits the nonlinear fundamental echo to be detected. The advantage of detecting this nonlinear fundamental component is an increase in signal intensity that is greater than that of the second harmonics alone of the contrast agent. The identification of this nonlinear response requires the use of an algorithm capable of modulating the pulses so that fundamental responses are elicited from the tissue and from the contrast agent. Such an algorithm requires perfect pulse modulation both in phase and in amplitude. It has been marketed with the name cadence Contrast Pulse Sequencing (CPS by Acuson Sequoia,  Siemens) and it works in real-time with continuous observation, using second-generation contrast agents . Previous algorithms exploit only the nonlinear harmonic response of contrast agent. These require cancellation of the fundamental echoes coming from the tissue forcing examination to be done blindly (without an image) until the contrast agent arrives. Contrast Pulse Sequencing software has the immediate advantage of enabling observation of the fundamental signals coming from the tissue at the same time as signals coming from the nonlinear fundamental response of the contrast agent (Phillips and Gardner, 2004).