By Dr. Lars Thorelius

March 23, 2006 -- Metastatic disease can be defined as the spread of viable malignant cells from the primary site of the tumor to other parts of the body, where they eventually stop and adhere to the tissue. At their new site they manage to acquire access to the bloodstream by the formation of new blood vessels, which supply the nutrients and oxygen necessary for the proliferation of new tumor cells. Each new independent tumor is generally referred to as a metastasis.

The cells released from the primary tumor may travel to other body parts through different mechanisms, including direct contact with the tumor surface, via lymph vessels and propagation through fluid such as ascites. Of course the bloodstream is one of the most important routes for the dispersion of metastases.

Different organs have different predispositions to be subject to metastases. And metastases from different primary tumors show different affinity to different target organs. Generally speaking, the lungs, liver, and bone marrow are highly susceptible to metastases, whereas metastases are less likely to be seen in the thyroid gland and spleen despite their great vascularity. The liver is susceptible to the spread of metastases from almost all kinds of malignant disease. Although virtually no organ or tissue in the body is guaranteed to be spared, the liver is quite frequently the organ where metastases first come to attention, and it is therefore almost invariably included in metastasis screening protocols.

In the Western world, contrast-enhanced CT is the most commonly used modality to screen for metastases. It is highly sensitive, especially since the introduction of helical CT and multidetector systems that are able to scan the entire liver in a few seconds, allowing for several scans during the liver's different circulatory phases. MRI is very sensitive as well, with the advantage of the availability of liver-specific contrast that accumulates in the cells of the liver parenchyma, rendering the metastases very conspicuous. MRI is obviously also used for screening purposes despite the fact that it is more time-consuming and may be subject to motion artifacts. According to different reports, these modalities detect more than 90% of liver metastases. CT/PET identifies the metabolic activity of metastases in addition to their position in the liver, but is to date very expensive and has not yet become a major modality in routine screening for liver metastases.

Throughout the world, plain unenhanced ultrasound (US) has been used for metastases screening for decades. Its advantages include low cost and wide availability even in developing countries. The major drawbacks of the traditional use of US are operator dependence, poor comparability with old exams, and very low sensitivity to metastases. Even in the most experienced hands, the detection rate is reported to be at best about 60% to 70%. Although these drawbacks seem to exclude unenhanced US as a justifiable screening modality in modern medicine, it remains in continuous use for just that purpose even in many Western countries. This is probably a matter of tradition and in many cases a lack of insight into the tremendous development of US and its possibilities in recent years.

Before the advent of helical CT and the availability of MRI, unenhanced US probably played a medically justifiable role for detecting metastasis despite its low sensitivity, but by today's standards the detection rate of unenhanced US is, in our opinion, no longer adequate and really should be discontinued for the purpose of finding metastases.

Contrast-enhanced ultrasound

With the introduction of second-generation US contrast agents (UCAs) and UCA-specific machine software, the prerequisites for metastasis detection improved dramatically, finally giving ultrasound the tools to compete at the same level as CT in routine screening, provided its use is limited to well-trained and experienced sonologists. According to several reports, the detection rates in skilled hands are similar to that of modern CT. Before UCAs were available, US was always compared to contrast-enhanced CT. This was a very uneven battle with a given outcome, but if CT had been deprived of the privilege of intravenous contrast, it would have come out short even to unenhanced US.

In our US department in Linköping, Sweden, contrast-enhanced US (CEUS) has been in abundant clinical use since the introduction of SonoVue (Bracco, Milan, Italy) in February 2002. SonoVue is still the only UCA registered for diagnostic use in the liver in Sweden although other UCAs are available that technically serve the same purpose.

SonoVue consists of microbubbles roughly the size of red blood cells containing sulphur hexafluoride encapsulated in a phospholipid monolayer (for details I refer to Bracco). Unlike first-generation microbubbles, second-generation ones do not burst when exposed to low-energy ultrasound. This very favorable capacity to withstand prolonged ultrasound exposure means that continuous imaging of the UCA-enhanced tissues can be maintained, permitting ample time for continuous scanning of the liver and other tissues. A key factor for the success of the exam is the power of insonation, which is referred to as the mechanical index (MI). The microbubbles burst when the MI reaches a critical level. Upon rupture of the shell, the gas component of the microbubble instantaneously dissolves and is ultimately exhaled. Of course, there is also a natural decay of the UCA concentration, and 10 to 15 minutes following the injection practically all of the SonoVue has disappeared and can no longer be seen.

In our department, CEUS has been incorporated as an efficient and accurate tool for routine characterization of focal liver lesions (FLLs) and detection of metastases. Patients with cancer of the colon, breast, lung, and carcinoids are most common in our practice. To date we have examined approximately 4,000 patients with CEUS, of which the liver accounts for about 85%. The majority of exams focused on the liver parenchyma and biliary tree receive UCA. This has become a natural routine and may very well be compared with the situation in the CT lab, where exams without contrast in most cases are of limited value.

Thanks to our routine use of CEUS for characterization of FLLs, such exams are very rare in the CT and MRI labs, with the resulting benefits of no radiation, high specificity, and low cost. CEUS is much more frequently used as a problem solver for CT and MRI than vice versa. However, to keep up an optimally high standard, no examiners work less than half time with ultrasound to keep up the training of the individual examiners. I believe that no serious ultrasound practice should be spread out on too many hands, since a continuously high individual exam volume is a prerequisite for results that can match those of other modalities.

Note: To view the clips in this discussion you must first download the DivX codec (as well as a Windows or other media player if you don't already have one), by using the links provided.

As usual, all clips in this article are samples from our PACS archive, which we began building in 2002 at the same time as we began scanning the different organs in a standardized manner. (A description of the standardization of the liver exam was given in the previous article in this series.) All clips in this article are samples from exams for illustration, and do not constitute the entire clip documentation of these exams.

When performing CEUS, unenhanced scans are routinely used for reference with the subsequent UCA-enhanced scans. Such unenhanced scans are usually referred to as "baseline" scans. In all CEUS clips a timer is used. We have chosen to inject the contrast five seconds into the timing by standard since examiners perform the injection themselves, and it would take a "third hand" to start the timer and inject simultaneously since one hand is obviously occupied with the transducer. In all clips the timer is located to the right in the image, as indicated in figure 1. The current MI is also displayed to the right as indicated in figure 2.

When working with CEUS, the timing of the injection and various circulatory events is critical to understanding the pathology. This is one of the reasons why CEUS must be stored and reviewed dynamically at full frame rate. The other reason is obvious: It is impossible to detect all metastases or other lesions in the entire liver during the bedside exam. Many metastases would have been missed in our practice if we did not have workstations for final soft-copy reading of the exams.

Mechanism behind characterization

For more information on the basics of the liver's circulatory phases and how to differentiate malignant from benign liver lesions with CEUS, I refer to the previous article about CEUS in this series. Basically, in the late phase after injection, benign lesions maintain their level of enhancement at the same level as the surrounding liver, while malignant lesions become less enhanced than the surrounding liver because of their principally different patterns of circulation. This is true for any kind of contrast medium for any modality, unless the contrast is actively targeted to be taken up into the liver cells, thus being liver-specific.

Figure 1. Bolus of 2.4 mL of SonoVue entering the liver

Figure 1: As in all clips, the UCA is injected five seconds into the timing of the timer on the right side of the image (>). Note how the contrast first appears in the liver artery (A), second in the portal vein (P), and last in the liver vein (V). The black tubular structure is the common bile duct. The accumulation of contrast in the parenchyma increases over time during the arterial and early portal phases. In this particular clip, the MI has not been high enough for adequate penetration to the deepest parts of the image. All video clips courtesy of Dr. Lars Thorelius.

Figure 2. The liver in the late phase

Figure 2: There is a homogenous contrast enhancement remaining in the normal liver parenchyma after more than three minutes following injection. As the MI (>) is deliberately increased, the microbubbles in the scan plane rupture. Following rapid decrease of MI back to normal CEUS level and turning of the transducer 90°, the parenchymal enhancement becomes conspicuous against the black "slice" of total bubble destruction. New microbubbles clearly enter the large vessels after the high MI burst.

The 'MI dilemma'

In all exams intended to detect liver metastases, two major technical objectives are involved:

  1. Achieve strong contrast echo from the entire liver, including the deepest parts
  2. Preserve the microbubbles from the destruction of powerful insonation

These objectives are technically in complete opposition to one another, a fact that we may call the "MI dilemma" of CEUS. In unenhanced US, ultrasound penetration of the tissues improves with increasing MI without other significant negative consequences to image quality. Of course, the heating effect of ultrasound increases with increasing power of insonation, but all ultrasound equipment has the MI level set to be maximized to harmless levels to avoid thermal damage to the tissues. In practice the examiner need never think of MI as a variable parameter. (That said, care must always be taken not to expose fetuses, neonatal brains, or especially the eyes to exams of unnecessary duration. The eyes are particularly susceptible to thermal damage since the eyeball practically lacks cooling blood circulation. Ophthalmologic US use has its own very restrictive MI regulations to avoid coagulation of the eyeball.)

With CEUS, MI is introduced as a very real variable. While slightly too high an MI gradually destroys the microbubbles, too low an MI does not provide sufficient penetration to the deeper parts of the liver, yielding too little parenchymal enhancement to expose the comparative darkness of the washout of metastases. The MI setting is probably the CEUS parameter that requires most training to master intuitively, since there is no correspondence in unenhanced US to the adverse effects on the imaging of improper MI levels. Ultrasound machines to date do not set the MI level automatically for optimal balance between signal strength and hazard of microbubble destruction. Without overdramatizing the implications of MI, I must stress that it is a new variable that comes with a learning curve. The examiner must be able to keep continuous control of the MI value and change it as indicated by the image quality throughout all exams. However, in most liver exams there are some margins, and often the MI level can be limited to two: one for the shallow parts and one for the deeper parts of the liver.

With CEUS we follow predefined scanning patterns, but some scans usually have to be repeated with a different MI for optimal metastasis detectability at all depths of the liver. In such cases, the shallow areas are obviously scanned first with very low MI, and the deeper second with an increased MI level. Choosing the reversed order would destroy many microbubbles in the near field to the transducer while we focus our attention to the deeper areas of the liver, only to discover that another injection of UCA is needed for proper imaging of surface areas. With experience such mistakes are avoided. With time it actually becomes an "instinct" to keep the transducer off the patient while not actively scanning to preserve microbubbles. An analogy: When ironing you would not keep the iron on the same spot of clothing while you answer the phone. Instead, you would probably lift the iron off the garment if you were too distracted to keep ironing.

Figure 3a. Longitudinal baseline scan of left liver lobe
Figure 3b. CEUS of the same area at MI 0.14

Figure 3: The image plane is longitudinal while the transducer movement is transversal, in accordance with our nomenclature of the scanning phases. At MI 0.14 (figure 3b), there is good signal throughout the parenchyma. Not only is the large cyst evident with CEUS, but also a number of smaller ones.

Figure 4a. Transverse baseline scan of right liver lobe
Figure 4b. CEUS of same area at MI 0.13
Figure 4c. CEUS of same area at MI 0.22
Figure 4d. CEUS of same area at MI 0.27

Figure 4: Baseline scan shows a very deep liver with good sound penetration. At MI 0.13 there is excellent signal from the near field, but the deep half of the liver is not penetrated and cannot be seen. At MI 0.22 much more of the deeper areas are exposed except for the very deepest subdiaphragmal level. At MI 0.27 there is penetration all the way to the diaphragm at the expense of severe microbubble destruction in the shallow areas. In this particular scan, an MI of 0.22 was optimal for almost the entire depth of the liver.

Basis for detecting liver metastases

A very simplified description of normal versus malignant blood vessel architecture is useful for the understanding of the prevailing theories of why metastases and benign neoplasms can be differentiated by observing their respective circulatory patterns over time following a bolus injection of UCA. The proliferation of new blood vessels in normal organs is well-structured, with a tree-like buildup of large vessels branching out into the tissue and finally ending in the capillaries.

The capillaries constitute the ultimate target level for all circulation, and is the level at which the exchange of oxygen for carbon dioxide takes place, as well as the delivery of nutrients and disposal of waste products back into the blood. At this level the blood flow velocity is very low, virtually making the individual blood corpuscles come to a halt for a while. Hence, each unit of a bolus of any contrast agent injected intravenously will be caught in "congested lanes" of blood in the arterioli prior to the capillary level, waiting for its turn to enter the capillary.

This congestion of contrast in millions of arterioli and capillaries provides a high concentration of contrast throughout the organ parenchyma for some time even after the UCA concentration has decreased in the general arterial circulation of the body. When finally the next unit of blood "in line" for passage through the capillaries has a lower concentration of contrast, the general parenchymal contrast concentration will decrease. This decrease of contrast enhancement due to inflow of blood with less contrast is what we refer to as washout.

In contrast to normal vascular structures, the vessels in malignant lesions grow more or less without proper organization. As a result, many of the afferent arterial vessels do not ultimately end in capillaries. Instead, they frequently connect directly to efferent venous vessels, shunting the blood and contrast past any capillaries where the resistance to flow is higher. Although such shunts are still tiny enough to be microscopic, they have great effect on the average flow velocity through the lesion, which is considerably increased. The resulting high flow velocity causes noticeably earlier entrance of arterial blood with a decreasing amount of contrast into the lesion than into the surrounding normal liver parenchyma. As simple as it sounds, this sign of early washout with reference to the surrounding liver has proved to be a very accurate predictor of malignancy. Also, as far as is known, all FLLs are supplied only by the hepatic arteries, while the liver is supplied by the arteries and also the portal circulation. The liver parenchyma itself to the greatest extent depends upon the portal vein blood for its own supply of oxygen and nutrients. The combination of washout from the metastasis with the added enhancement of the surrounding liver from the portal circulation makes the washout in the metastasis even more conspicuous.

Microbubbles do not pass the walls of vessels or capillaries, a fact which makes UCAs true blood pool agents that truly represent only circulating blood. No microbubbles are able to diffuse into adjacent spaces to make them falsely look vascularized. Common contrast agents for CT and MRI to some extent do not share this positive property. This means that they may "contaminate" extravascular spaces during the time of high vessel concentration by leakage and create a false impression of blood circulation. For this reason, metastases sometimes become less conspicuous in CT late in the late phase than would otherwise have been the case. For UCAs, the duration of the late phase has no impact on the detectability of "washed out" metastases for as long as the microbubbles remain intact. But, on the other hand, microbubbles only last for an effective time of about five to six minutes in the liver, after which the enhancement of the surrounding liver rapidly decreases and the homogeneity of the enhancement deteriorates.

Since the definition of malignancy is washout of contrast in the late phase, dark focal spots of washout in the otherwise homogenously enhanced liver are actually what we look for when scanning for metastases. This task is much more complicated in the arterial phase, as the enhancement of highly vascularized metastases is rapidly visually mixed with the enhancement of the surrounding liver. The enhanced metastases are blended into the enhancement of the liver in less than 10 seconds, and it is of course too short a time for thorough scanning of the entire liver. Also, it is very uncommon for metastases not to become clearly dark within about two minutes after injection, and after four minutes virtually all metastases in our experience show a clear washout. However, it should be mentioned that when scanning for primary hepatocellular carcinomas (HCC) we actually do depend on the short arterial phase because many highly differentiated HCCs display no clear washout, but HCC detection is beyond the scope of this article.

The arterial phase of metastases is very unpredictable. All metastases enhance in the early arterial phase, but the enhancement intensity and homogeneity vary considerably from case to case. Some enhance intensely while others are poorly vascularized. Frequently, areas of necrosis in the metastases are evident. A common denominator for many metastasis is an initial intense enhancement of the periphery of the lesion, which can frequently be seen extending a few millimeters out from the evident metastasis mass. This peripheral enhancement is called rim enhancement. The rim enhancement is not specific enough, however, for true characterization as many metastases lack this circulatory behavior in our experience.

Figure 5a. Hemangioma and metastases, baseline
Figure 5b. Hemangioma and metastases, arterial phase
Figure 5c. Hemangioma and metastases, late phase

Figure 5: Case with multiple poorly vascularized metastases plus a high echogenic hemangioma. The scanner software allows for separation of baseline echo from UCA echo, and these can be mixed as desired as in these sequences. Figure 5b shows the arterial enhancement of the previously bright hemangioma, with its characteristic slow contrast filling from the periphery to the center. There is clear rim enhancement surrounding the metastases. In figure 5c the hemangioma preserves the contrast enhancement even slightly better than the surrounding liver, defining it as benign, while the metastases have washed out the small amount of contrast they accumulated during the arterial phase. The leftmost metastasis appears necrotic except for the rim zone.

Figure 6a. Baseline US of a metastasis
Figure 6b. Same metastasis from arterial phase to washout
Figure 6c. Late-phase detection of same metastasis among others

Figure 6. Highly vascularized metastasis showing typical washout in portal phase. It becomes very conspicuous in late-phase detection (6c). (Clip 6b has been cropped to downsize the file for faster download).

Figure 7a. Metastases of different appearances on baseline scan
Figure 7b. Same with CEUS in the arterial and portal phases
Figure 7c. Same with CEUS in the late phase

Figure 7: (7b and 7c cropped for faster download.) Mainly necrotic metastases. The one on top has a peripheral zone of viable tissue that displays clear washout in the portal phase. Care must be taken not to mistake metastases with only peripheral viable tissue for hemangiomas with slow peripheral enhancement. Hemangiomas by definition do not display washout. The initially bright metastasis in the lower part of the image is difficult to see with CEUS because of reflections of the surrounding UCA in the abundance of dense structures within its tissue, but it was readily detected on baseline. It is not uncommon to see metastases of this appearance following chemotherapy.

Based on the reliable washout in metastases in the late phase, we have decided to disregard the arterial and portal phases of CEUS when we examine livers for the purpose of finding metastases. We have chosen to begin the scanning 90 seconds after injection. (For a detailed description of our standardized liver scanning pattern I refer to the previous article covering liver exam technique.) Our experience of CEUS' sensitivity to metastases is very encouraging. Many times CEUS finds metastases that have not been found with CT, and in some cases also metastases missed with MRI have been visualized with CEUS.

Of course, CEUS occasionally misses metastases found with CT or MRI, but generally the three modalities are nowadays about equally efficient for revealing metastases, with a small advantage for MRI and CEUS over CT in the experience of our practice (we presently use 16-detector CT scanners). One explanation for CEUS' high sensitivity is the true dynamics of the modality. The liver can be scanned over and over again for several minutes, and for characterization purposes a single metastasis can be followed at 20 frames per second for minutes. These dynamic properties of CEUS are far superior to the dynamics of CT and MRI, which are limited to sequences of still images down to a few seconds apart. Still, some patients are simply not suitable for US, and no UCA imaginable can overcome that fact.

Figure 8a. Transverse scan of left part of liver from the epigastrium about 30 seconds after injection
Figure 8b. Same about 50 seconds after injection
Figure 8c. Same about 70 seconds after injection
Figure 8d. Same about 90 seconds after injection
Figure 8e. Same about 120 seconds after injection

Figure 8: This sequence explains why we wait 90 seconds after injection of UCA before beginning to scan for metastases. The clips show increasing washout with time in four to five metastases. In 8a there is no washout, and the metastases cannot be seen. Progressing through the series, one notes the increasing darkness of washout in the metastases that makes them increasingly conspicuous.

Figure 9a. Longitudinal baseline scan of left part of liver from the epigastrium
Figure 9b. Same about 15 seconds after injection
Figure 9c. Same about 30 seconds after injection
Figure 9d. Same about 45 seconds after injection
Figure 9e. Same about 60 seconds after injection
Figure 9f. Same about 80 seconds after injection

Figure 9: Another illustration of the increasing washout in metastases with time. In this case the washout is much faster than in figure 6, actually allowing detection very early after the arterial phase in 9b, where the circulation in the metastases and the liver arteries conceal one another.

Impact of CEUS for detecting metastases

Provided the patient is suitable for US, the effect of CEUS on liver metastases is often striking. It has given ultrasound the power to detect much smaller metastases than ever before, and we also reliably find the many larger metastases that are isoechoic with the liver on unenhanced US and therefore often missed.

Today we use CEUS in our daily practice for preoperative screening for metastasis of the liver together with MRI of the pelvis in rectal cancer, and CEUS is used for the subsequent follow-up of these patients. We also use CEUS to monitor chemotherapy effect on a daily basis, since our workstations permit side-to-side comparison with previous exams. The workstations in symbiosis with our uncompromised standardization of scanning patterns gives US in our department a workflow that is quite similar to that of CT and MRI, with much of the diagnostic work being performed by workstation reading of cine loops. For the liver, we have chosen to scan the liver twice in the standardized manner described in the previous article, once without contrast and once beginning 90 seconds following contrast injection.

Figure 10a. Multitude of obvious metastases in right lobe on baseline, transverse scan
Figure 10b. Same, longitudinal scan
Figure 10c. Same, transverse scan with CEUS
Figure 10d. Same, longitudinal scan with CEUS
Figure 10e. Same as 10d, slow motion in workstation

Figure 10: Although there is no doubt about the existence of metastases on baseline, the conspicuous nature of the metastases on CEUS make them much clearer. Also note that many very small metastases down to a size of 3-4 mm are revealed by CEUS. At the workstation there is ample time to go through the exam for detailed evaluation and measurements, exemplified in 10e.

When many simple cysts are present in the liver, detecting metastases is more of a challenge. The cysts are always black, while the metastases are more or less dark in the late phase. Often, cysts are obvious by their benign appearance also with CEUS. However, careful comparison with the baseline scans and locally sometimes also early scans in the arterial phase may be needed to reveal metastases among cysts. In general, as with any modality, the presence of abundant cysts remains a problem for accurate detection in cases with a few small metastases.

Figure 11a. Baseline, abundant simple cysts and suspicion of metastases, right lobe transverse scan
Figure 11b. Same with CEUS about 30 seconds after injection
Figure 11c. Same with CEUS about 45 seconds after injection
Figure 11d. Same with CEUS about 60 seconds after injection
Figure 11e. Same with CEUS about 75 seconds after injection
Figure 11f. Same with CEUS about 90 seconds after injection
Figure 11g. Same with CEUS about 110 seconds after injection
Figure 11h. Longitudinal scan of same area

Figure 11: By comparing baseline scan 11a with CEUS late-phase scan 11g, it becomes evident that there are metastases among the cysts. Note the large partially projected metastasis to the right (left in image). In the consecutive series 11b-f of CEUS scans with contrast-specific software, the cysts are identically black throughout the contrast series, while the metastases become increasingly apparent with increasing washout.

Thanks to standardized scanning patterns we are able to monitor the results of chemotherapy by measuring reference metastases. We usually measure them in three planes. The three planes of measurement are also standardized so that we know in which sequence they are measured with reference to the screen image and the patient's body. Standardizing from the three different planes of the body would be impossible, since the transducer has very different positioning on the body in the different scans. Measurements are always made so that a lesion is measured as width by height on the monitor in the transverse scans, and by width on the monitor in the longitudinal planes. This way we always know what a series of three measurements represent and can easily repeat them for the next follow-up.

Figure 12a. Same case as in figure 11 at previous follow-up months earlier, baseline transverse scan
Figure 12b. Same with CEUS, transverse scan
Figure 12c. Same with CEUS, longitudinal scan

Figure 12: The previous follow-up of same case as figure 11, some months earlier. Progressive metastatic disease with increase of both number and size of metastases.

From time to time we find metastases with CEUS that could not be seen on baseline scans. They may have been either very small or isoechoic to the liver, both situations making them virtually impossible to detect without a UCA. This is of course the situation in which CEUS excels over US, and the situation is so common that unenhanced US alone really has no place in the search for metastases in modern medicine.

Figure 13a. Longitudinal baseline US scan of left liver lobe
Figure 13b. Transverse baseline US scan of right liver lobe
Figure 13c. Longitudinal CEUS scan of left liver lobe
Figure 13d. Transverse CEUS scan of right liver lobe

Figure 13: Woman with breast cancer. Despite good access to the liver no metastases are seen on baseline scans. Several black washed-out metastases are seen in the late phase of CEUS. 13d is cropped at bottom for file size reduction.

Figure 14a. Transverse baseline US scan of central left liver lobe
Figure 14b. Same, longitudinal US scan
Figure 14c. Same, transverse CEUS scan
Figure 14d. Same, longitudinal CEUS scan

Figure 14: Quite large metastasis cranially in liver segment 4. It is virtually black in the CEUS scans, but would very easily be missed on baseline scans because of its isoechogenicity.

Figure 15a. Transverse baseline US scan of right liver lobe
Figure 15b. Same with CEUS
Figure 15b. Same with CEUS, slow motion

Figure 15: Obese patient with a small number of small metastases that cannot possibly be detected on baseline scans. Metastases of this size are often a challenge also for CT and MRI. Slow-motion CEUS in workstation reveals metastases, the two largest being marked with yellow arrows (> <).

Not only does normal liver parenchyma become homogenously enhanced with contrast, but cirrhotic and fibrotic livers, which may be very irregular on baseline scans, are also to a very great extent rendered homogenous by the UCA. The basis for this is probably the benign nature of such liver changes, which do not cause significant shunting, thus very little or no washout. The CEUS homogeneity in such altered liver parenchymas gives CEUS a potent position for metastasis detection in such livers. CEUS is in fact an important problem solver for CT and MRI, which often have problems finding metastases in such livers.

Figure 16a. Baseline of liver with bizarre steatotic and fibrotic parenchyma
Figure 16b. Same with CEUS

Figure 16: Severe fatty change and fibrotic strands in liver after years of chemotherapy for breast cancer. With CEUS, the irregularities of the liver are replaced by a fairly homogenous UCA enhancement. Several dark washed-out metastases are evident against the enhanced liver parenchyma.

Figure 17a. A metastasis is to be biopsied, but cannot be accurately identified
Figure 17b. With CEUS the metastasis is found with ease

Figure 17: CEUS is frequently used as a tool for guidance to metastases that are difficult to see on baseline.

Differential diagnosis

A few conditions may mimic the washout of metastases. Such are rare, but are encountered occasionally. The most metastasis-like condition that we see is probably when several small fungal abscesses are in the liver. Mainly this kind of lesion is seen in patients with leukemia or other diseases with compromised immune system. Small fungal abscesses are frequently not liquefied, and cannot be separated from metastases by appearance. Cytology is needed for diagnosis.

Figure 18a. Transverse baseline scan
Figure 18b. Same with CEUS

Figure 18: Multiple fungal abscesses in a patient with leukemia. Cytology proved fungal abscesses.

Other conditions that may rarely mimic metastases are remnants of inflammatory conditions such as abscesses, very aged focal nodular hyperplasias (FNH) with a high fibrotic content of poor vascularization, and granulomas. However, in patients with a known malignancy, metastases are by far the most common liver lesions with late-phase washout that are found when CEUS is performed. With CEUS benign lesions such as typical FNH and hemangiomas are not detected in late-phase scanning, but this has no negative consequences for the treatment of the patients.

CEUS is here to stay

From our experience, we conclude that CEUS is a powerful, fast, and efficient modality for detecting metastases. It is comparable to modern CT and MRI, and is in our practice often used as a problem solver for those modalities. In early 2004, the CEUS expert committee of the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB) recommended CEUS to be used in all cases in which US is involved in detecting liver metastases. The experience in our own clinic is completely in line with those recommendations, and we firmly believe that the use of unenhanced US for ruling out liver metastases has no relevance and must be discontinued. Furthermore, the development of CEUS has just begun.

On the horizon are the possibilities of molecular CEUS with tumor-targeted microbubbles, full-frame rate 4D acquisitions, accurate quantification of tissue perfusion, and more advanced implementations of automatic settings for CEUS in the ultrasound machines. We look forward to a rapidly expanding role for CEUS in the years to come.

By Dr. Lars Thorelius contributing writer
March 23, 2006

Related Reading

Contrast-enhanced US useful for differentiating focal liver lesions, January 19, 2006

Contrast-enhanced ultrasound aids in grading liver disease, December 1, 2005

Ultrasound contrast enhances liver imaging, May 11, 2005

Exam technique critical in liver ultrasound, October, 22, 2004

Second-generation microbubbles enhance ultrasound's clinical utility, August 27, 2002

Copyright © 2006


To read this and get access to all of the exclusive content on create a free account or sign-in now.

Member Sign In:
MemberID or Email Address:  
Do you have a password?
No, I want a free membership.
Yes, I have a password:  
Forgot your password?
Sign in using your social networking account:
Sign in using your social networking