(Radiology. 1999;210:409-416.)
© RSNA, 1999
Improved Imaging of Liver Metastases with Stimulated Acoustic Emission in the Late Phase of Enhancement with the US Contrast Agent SH U 508A: Early Experience
Martin J. K. Blomley, FRCR1,
Thomas Albrecht, FRCR2,1,
David O. Cosgrove, FRCR1,
Nayna Patel, DMU1,
Vijay Jayaram, MD1,
Jenny Butler-Barnes, RGN1,
Robert J. Eckersley, PhD1,
Albrecht Bauer, MD2,2 and
Reinhard Schlief, MD2,2
1 Department of Imaging, Hammersmith Hospital, Du Cane Rd, London W12 0HS, England (M.J.K.B., T.A., D.O.C., N.P., V.J., J.B.B., R.J.E.)
2 Clinical Development Diagnostics, Schering, Berlin, Germany (A.B., R.S.).
 |
Abstract
|
|---|
PURPOSE: To see whether stimulated acoustic emission (SAE) in the liver parenchyma in the late phase of enhancement with SH U 508A increases the conspicuity of occult metastases at ultrasonography (US).
MATERIALS AND METHODS: Eighteen patients with known hypo- or hypervascular hepatic metastases underwent US after SH U 508A administration, after a delay of at least 5 minutes, to ensure decay of blood pool enhancement. In 16 patients with visible metastases, conspicuity was compared on registered SAE and gray-scale scans by two blinded readers and by using computerized analysis of relative gray-scale and color Doppler conspicuity scores inside and outside the lesion. In nine patients, areas suspected of being involved but without definite gray-scale masses were imaged in the same way. Paired sections were analyzed by two blinded readers looking for parenchymal color defects without corresponding gray-scale masses; nine control images from three healthy volunteers were also included.
RESULTS: Intense, transient parenchymal SAE was seen in all subjects. All metastases appeared as areas of reduced or absent signal. The conspicuity score was 80% for SAE versus 9% for gray-scale US (P < .001, Wilcoxon signed rank test). SAE-specific defects were seen in all patients but in none of the volunteers. Metastases seen on SAE but undetectable on gray-scale images were proved in three patients.
CONCLUSION: SAE with SH U 508A improves the conspicuity of metastases. SAE-specific defects may reveal isoechoic or subtle metastases.
Index terms: Liver neoplasms, metastases, 761.33 • Liver neoplasms, US, 761.12983, 761.12988, 761.33 • Ultrasound (US), contrast media, 761.12988 • Ultrasound (US), Doppler studies, 761.12983 • Ultrasound (US), technology, 761.12989
|
Introduction
|
|---|
Levovist (SH U 508A; Schering, Berlin, Germany) is, to our knowledge, the first microbubble agent to produce reliable and clinically useful systemic Doppler enhancement after intravenous injection at ultrasonography (US). It is given intravenously, as a typical dose of 2.5 g, and consists of galactose microaggregates with a small admixture of palmitic acid (0.1%). When mixed with water, agitated, and allowed to stand for 2 minutes, multiple small stabilized air microbubbles with a mean diameter of 2–3 mm are produced. These are sufficiently small and stable to pass through capillary beds repeatedly and thus produce useful systemic enhancement after intravenous injection. Extensive clinical trials with SH U 508A have demonstrated excellent tolerance and safety, with only minor adverse reactions reported (1). It is licensed for the improvement of suboptimal Doppler examinations in many countries (although not yet in the United States) as a blood pool agent, with a duration of useful Doppler enhancement of 1–5 minutes in main vessels (2).
An increase in the back scatter of blood is generally regarded as the major clinical value of SH U 508A and most other microbubble agents, but it has been recognized recently that the interaction between the ultrasound beam and microbubbles is extremely complex. At higher acoustic pressures, resonant signals at second, and other, harmonic frequencies start to appear. As sound pressures increase further, although still within levels permitted for diagnostic imaging, phenomena related to microbubble destruction or disruption occur (3). These include the production of transient ultrasonic signals, an effect referred to as stimulated acoustic emission (SAE). SAE can be visualized as intense transient enhancement on gray-scale images; this has been exploited by using intermittent imaging techniques with SH U 508A and other microbubbles (4–6). In color Doppler imaging, an additional effect can be seen. Here, frequency shifts are calculated from changes in the phase of sequential pulses of ultrasound. The sudden disappearance of a reflector, such as a microbubble, causes a large change of phase. This in combination with any emission signals actively produced by the microbubble as they are disrupted will be interpreted as a transient Doppler shift. This "pseudo-Doppler" effect produces intense signals at two-dimensional Doppler imaging (7). Known characteristics of SAE include the transient nature of the emissions and the dependence on high acoustic powers. Unlike conventional Doppler imaging, SAE does not rely on the movement of the target and is observed equally well from stationary microbubbles.
To our knowledge, SAE was first recognized with the investigational microbubble agent SH U 563A (Schering) (8–11), but it is now known that other microbubbles, including SH U 508A, also produce this response, although it is weaker (12,13) and has not hitherto been regarded as clinically important. We have, however, recently described (14) a liver-specific phase for SH U 508A. This can be seen as strong, very transient bands of wide-spectrum color (14). For the first few minutes after SH U 508A administration, the effect is seen in blood vessels; however, after blood pool clearance, it concentrates in the hepatic and splenic parenchyma. It is highlighted by using machine settings that suppress noise and artifacts (high pulse repetition frequency and low or medium color gain settings, well below the noise floor). It is short-lived, and once an area has been scanned for a few seconds the effect disappears. It is best seen with relatively high acoustic power settings (although still within accepted levels for diagnostic imaging) and shows a tendency to cluster at the level of the focal zone, where the power is maximal. These transient wide-frequency Doppler signals, within parenchyma, that are destroyed by the act of scanning cannot be explained on the basis of conventional back scatter or resonance and imply SAE as the mechanism. SAE is best seen with relatively high acoustic power settings (although still within accepted levels for diagnostic imaging) and shows a tendency to cluster at the depth of the focal zone, where power deposition is maximal (Fig 1).
View larger version (140K):
[in this window]
[in a new window]
|
Figure 1a. The SAE effect illustrated in the liver of a 33-year-old healthy male volunteer. Intravenous SH U 508A (2.5 g) had been administered 5 minutes before a was obtained. (a) Color Doppler has been turned on, producing a band of strong mosaic-like parenchymal signal (arrow) on the first color Doppler frame. (b) The frame obtained immediately after a, with a delay of 0.2 second, shows a reduction in the SAE (arrow), illustrating the transience of the effect. On the following frame (not shown), SAE had almost completely disappeared. Note that SAE is maximal at the level of the focal zone.
|
|
View larger version (142K):
[in this window]
[in a new window]
|
Figure 1b. The SAE effect illustrated in the liver of a 33-year-old healthy male volunteer. Intravenous SH U 508A (2.5 g) had been administered 5 minutes before a was obtained. (a) Color Doppler has been turned on, producing a band of strong mosaic-like parenchymal signal (arrow) on the first color Doppler frame. (b) The frame obtained immediately after a, with a delay of 0.2 second, shows a reduction in the SAE (arrow), illustrating the transience of the effect. On the following frame (not shown), SAE had almost completely disappeared. Note that SAE is maximal at the level of the focal zone.
|
|
Using this mode of imaging of the liver uptake of SH U 508A, we observed that focal liver lesions have little or no signal, particularly when imaging is performed relatively late (more than 5 minutes) after SH U 508A administration, when conventional vascular enhancement has faded. We therefore designed a study to try and exploit this liver-specific property by investigating the clinical value of SAE imaging in the late phase of enhancement, that is, more than 5 minutes after administration, first by determining whether the conspicuity of defined metastases was increased and second by assessing whether the detection of subtle or isoechoic lesions was improved.
|
MATERIALS AND METHODS
|
|---|
Patients with Biopsy-proved Malignancy
Eighteen patients (seven men, 11 women; mean age, 59 years; age range, 34–75 years) with biopsy-proved malignancies who were known to have hepatic metastases were examined. In 12, hepatic biopsy results had proved hepatic metastases. In six patients, malignancy had been proved with biopsy elsewhere in the body and hepatic metastases had been reported at computed tomography (CT), except in one patient who died of metastatic breast carcinoma before CT could be performed. The histologic diagnoses were neuroendocrine or carcinoid tumor (n = 5), adenocarcinoma of unknown primary site (n = 4), colorectal carcinoma (n = 3), squamous cell carcinoma of unknown primary site (n = 2), uterine leiomyosarcoma (n = 1), breast carcinoma (n = 1), gastric carcinoma (n = 1), and bronchial carcinoma (n = 1).
For 8 months (February through October 1997), we successfully elicited hepatic parenchymal SAE in all patients attending our department after the administration of SH U 508A. The patients in this study were all such patients examined during this period who had proved liver metastatic disease and in whom a standard protocol, described later, to acquire SAE images had been used.
All subjects (patients and volunteers) had given informed consent for the research use of SH U 508A (local research ethics committee numbers REC 96/4902 and 96/4946). In all subjects, SH U 508A was injected intravenously as a bolus in a dose of 2.5 g at a concentration of 300 mg/mL with a 10–20-mL normal saline flush by using an 18- or 20-gauge peripheral intravenous cannula. After vascular enhancement was over, and at least 5 minutes after the last SH U 508A injection (so that blood pool activity would be negligible), the liver was examined for SAE.
The technique used was to scan in B-mode and then to switch into color velocity Doppler mode with the probe held still. A commercially available US system was used throughout (Sequoia 512; Acuson, Mountain View, Calif). A phased-array sector probe (model 4V2; Acuson) was used, with a color Doppler frequency of 2.5 MHz. Scanning for the shortest possible period and using low acoustic powers to reduce the amount of microbubble disruption, an area of interest was identified on a B-mode scan. With the focal zone set at, or just deep to, the level of interest, color Doppler was then switched on. We used maximal pulse repetition frequency and low to medium color gain settings well below the noise floor of the system so that little or no flow signals from blood vessels were displayed. The color acoustic power was set at the default (maximal) setting. The acoustic powers and color gain settings were recorded for each image used for analysis.
Shortly after switching to color Doppler, we froze the display and reviewed the images in the cine review, as the strongest effects were seen on the first and second frames. This US system allows images to be stored digitally with or without the color Doppler overlay. Paired images were thus acquired with and without the SAE information at the same time and with exact registration. These coregistered images (Fig 2) will be referred to as image pairs for the purposes of this article. All patient identifiers were removed from the images, but no other modifications were made; specifically, no changes were made to the image data. The time delay, rounded to the nearest minute, from injection and the time at which the image pair was obtained were recorded.
View larger version (140K):
[in this window]
[in a new window]
|
Figure 2a. Transverse section through the left lobe of the liver in a 53-year-old male patient with metastatic carcinoid tumor (arrow) in (a) gray-scale and (b) color SAE displays. Note that in b, SAE is confined to the liver parenchyma around the metastasis, increasing its conspicuity. Also note that there are several smaller defects (arrowheads) in the SAE inferiorly in b, which may represent further lesions.
|
|
View larger version (122K):
[in this window]
[in a new window]
|
Figure 2b. Transverse section through the left lobe of the liver in a 53-year-old male patient with metastatic carcinoid tumor (arrow) in (a) gray-scale and (b) color SAE displays. Note that in b, SAE is confined to the liver parenchyma around the metastasis, increasing its conspicuity. Also note that there are several smaller defects (arrowheads) in the SAE inferiorly in b, which may represent further lesions.
|
|
For the subjective analysis, the paired images, that is, the images with and the images without SAE data, were presented to two blinded observers (D.O.C., N.P.). The images were displayed on a computer screen so that the image pair could be shown with and without color overlay by clicking a mouse button. The observers both had many years of experience with Doppler US and were familiar with the SAE effect from previous work but were blinded to the subject details and findings.
Each observer performed the analysis by completing a written questionnaire in two stages. The observers were asked to view the image pairs as though they had been asked to rule out (or confirm) liver metastases and to report on the scans as though they had been asked to provide a clinical report.
Two evaluations were performed, first to see whether SAE increased the conspicuity of visible lesions and second to see whether SAE might reveal occult disease. The first evaluation was both subjective and objective; the second was only subjective. For the second evaluation, data from healthy volunteers were included as controls.
Healthy Volunteers
By following a similar method as that just described, we obtained control image pairs after the administration of 2.5 g of SH U 508A in three healthy volunteers (two men, one woman; aged 29, 33, and 45 years) who gave informed consent. These images were obtained as part of an ethically approved study (local research ethics committee number REC 96/4946). After completion of vascular enhancement, and at least 5 minutes after injection, three different areas of each liver were scanned. All subjects had normal livers at US, and none had any clinical or other evidence of liver disease. To avoid the need to examine large numbers of volunteers, three images pairs were obtained from different areas of the liver in each of the three subjects. In total, thus, nine image pairs in three healthy volunteers were obtained with and without SAE.
Effect of SAE Data on Lesion Conspicuity
The subjective part of the first evaluation was done by presenting the observers with image pairs showing one selected metastasis from each patient. The observers were asked to confirm that they could see the lesion on the B-mode image and to evaluate whether the conspicuity of the lesion was better, the same, or worse with the SAE display. In two of the 18 patients, no definite metastasis could be seen on B-mode images anywhere in the liver; these were therefore excluded from this phase of the evaluation. Thus, in total, images from 16 patients were evaluated. One metastasis was shown from each patient.
The objective part of the first evaluation was done by analyzing these same 16 image pairs digitally with the program MATLAB (Mathworks, Natick, Mass). This was done off-line by using a code specially written by one of us (R.J.E.). This allowed regions of interest (ROIs) to be drawn for a given area in an image pair and then calculation of color and gray-scale intensities for that ROI as percentages of pixels. By using the color data, color intensity was calculated as the percentage of color pixels within that ROI. Thus, if all pixels were color, the percentage of color pixels was 100%, and if no color signal was present, the percentage of color pixels was 0%. The mean gray-scale intensity for the same ROI was calculated by using the B-mode display. The gray-scale intensity was calculated on an 8-bit scale (ie, 0–256), and was converted to a percentage figure by expressing it as a percentage of the maximum score. Thus, a completely black ROI scored was 0% gray, and a saturated white ROI was 100%.
Two ROIs of the same size and at the same depth were defined on the computer, one within and one outside the focal lesion. This was done with inspection of both the gray-scale image and SAE display, but the ROIs were drawn on the gray-scale image. As the images with and the images without the SAE data were exactly registered, the intensity of both the color and B-mode signals could then be calculated for identical regions. The absolute difference between the percentage of color pixels for the ROI outside the lesion and that within the lesion was calculated and was called "color conspicuity." Likewise, the absolute difference between the percentage of gray outside and that inside the lesion was also calculated to give the "gray conspicuity." For example, if the percentage of gray was 10% inside a lesion and 16% outside, the gray conspicuity score was 6%. Similarly, if the proportion of color pixels inside the lesion was 11% and that outside was 80%, the color conspicuity score was 69%. An example is shown in Figure 3. These two sets of values were compared by using a nonparametric test (Wilcoxon signed rank test) with the statistics package INSTAT (GraphPad Software, San Diego, Calif).
View larger version (124K):
[in this window]
[in a new window]
|
Figure 3a. Longitudinal section through the right lobe of the liver in a 65-year-old female patient with a metastatic islet cell tumor in (a, c) gray-scale and (b, d) color displays. The upper and lower edges of the metastasis are indicated with arrows in a. ROIs (red rectangles) have been drawn adjacent to (in a and b) and within (in c and d) this metastasis. The gray-scale and color data were quantified by using the program MATLAB to compare the differences in signal intensities between the lesion and the liver as a measure of lesion conspicuity. In this case, the color conspicuity score was 77%, and the gray conspicuity score was 3%.
|
|
View larger version (121K):
[in this window]
[in a new window]
|
Figure 3b. Longitudinal section through the right lobe of the liver in a 65-year-old female patient with a metastatic islet cell tumor in (a, c) gray-scale and (b, d) color displays. The upper and lower edges of the metastasis are indicated with arrows in a. ROIs (red rectangles) have been drawn adjacent to (in a and b) and within (in c and d) this metastasis. The gray-scale and color data were quantified by using the program MATLAB to compare the differences in signal intensities between the lesion and the liver as a measure of lesion conspicuity. In this case, the color conspicuity score was 77%, and the gray conspicuity score was 3%.
|
|
View larger version (121K):
[in this window]
[in a new window]
|
Figure 3c. Longitudinal section through the right lobe of the liver in a 65-year-old female patient with a metastatic islet cell tumor in (a, c) gray-scale and (b, d) color displays. The upper and lower edges of the metastasis are indicated with arrows in a. ROIs (red rectangles) have been drawn adjacent to (in a and b) and within (in c and d) this metastasis. The gray-scale and color data were quantified by using the program MATLAB to compare the differences in signal intensities between the lesion and the liver as a measure of lesion conspicuity. In this case, the color conspicuity score was 77%, and the gray conspicuity score was 3%.
|
|
View larger version (111K):
[in this window]
[in a new window]
|
Figure 3d. Longitudinal section through the right lobe of the liver in a 65-year-old female patient with a metastatic islet cell tumor in (a, c) gray-scale and (b, d) color displays. The upper and lower edges of the metastasis are indicated with arrows in a. ROIs (red rectangles) have been drawn adjacent to (in a and b) and within (in c and d) this metastasis. The gray-scale and color data were quantified by using the program MATLAB to compare the differences in signal intensities between the lesion and the liver as a measure of lesion conspicuity. In this case, the color conspicuity score was 77%, and the gray conspicuity score was 3%.
|
|
Detection of SAE-specific Defects
Eighteen image pairs were used, nine from the patients with malignancy and nine from the three healthy volunteers. All the patients in this part of the evaluation had undergone recent CT and/or angiographic imaging, which had been reported as showing extensive liver metastatic disease (apart from the one patient mentioned earlier who died before CT could be performed). We chose image pairs showing areas of the liver with no definite B-mode lesions but with multiple SAE color defects at the time of scanning. An example is shown in Figure 4. Two of the patients had no visible metastases on B-mode images, even though widespread metastatic disease had been seen at CT and liver biopsy results had been positive in both. (These were the two patients whose images could not be used for the evaluation of conspicuity.) In addition, we obtained nine image pairs from three healthy volunteers as control images. Thus, 18 image pairs were presented to the observers for the second evaluation. Of these, half were from patients with widespread malignancy and half were from healthy volunteers. The observers were not told which image pairs were from which group.
View larger version (153K):
[in this window]
[in a new window]
|
Figure 4a. Oblique section through the right lobe of the liver in a 65-year-old male patient with metastatic gastric carcinoma in (a) gray-scale and (b) color SAE display. a shows diffuse heterogeneity of the liver but no definite focal lesion. In b, SAE (arrow) is seen at the level of the focal zone in the liver after SH U 508A administration. A defect (arrowhead in b) in the SAE is observed; this was interpreted as a possible metastasis. (c) Axial CT scan (abdominal soft-tissue window [level, 40 HU; width, 400 HU]) obtained after iopromide administration (300 mg/mL) the same day as a and b reveals multiple hepatic metastases (arrows), as well as long-standing hepatic and splenic calcified granulomata.
|
|
View larger version (141K):
[in this window]
[in a new window]
|
Figure 4b. Oblique section through the right lobe of the liver in a 65-year-old male patient with metastatic gastric carcinoma in (a) gray-scale and (b) color SAE display. a shows diffuse heterogeneity of the liver but no definite focal lesion. In b, SAE (arrow) is seen at the level of the focal zone in the liver after SH U 508A administration. A defect (arrowhead in b) in the SAE is observed; this was interpreted as a possible metastasis. (c) Axial CT scan (abdominal soft-tissue window [level, 40 HU; width, 400 HU]) obtained after iopromide administration (300 mg/mL) the same day as a and b reveals multiple hepatic metastases (arrows), as well as long-standing hepatic and splenic calcified granulomata.
|
|
View larger version (140K):
[in this window]
[in a new window]
|
Figure 4c. Oblique section through the right lobe of the liver in a 65-year-old male patient with metastatic gastric carcinoma in (a) gray-scale and (b) color SAE display. a shows diffuse heterogeneity of the liver but no definite focal lesion. In b, SAE (arrow) is seen at the level of the focal zone in the liver after SH U 508A administration. A defect (arrowhead in b) in the SAE is observed; this was interpreted as a possible metastasis. (c) Axial CT scan (abdominal soft-tissue window [level, 40 HU; width, 400 HU]) obtained after iopromide administration (300 mg/mL) the same day as a and b reveals multiple hepatic metastases (arrows), as well as long-standing hepatic and splenic calcified granulomata.
|
|
The observers were asked to look for definite SAE color defects in the liver parenchyma and for definite focal lesions on B-mode images. They were asked whether these matched and specifically whether they could see any SAE-specific defects. These were defined as focal areas of absent or reduced signal that were either within or largely surrounded by color signal and for which the defect could not be explained as a definite B-mode mass.
|
RESULTS
|
|---|
Strong SAE signals were obtained from the liver parenchyma in all cases.
The average color gain setting used was 47% (range, 21%–52%). The average acoustic power used was 1.66 (mechanical index values; range, 1.0–1.9). The mean time from injection and acquisition of the image pair was 6 minutes (range, 5–12 minutes).
Effect of SAE Data on Lesion Conspicuity
In all 16 image pairs, the two observers agreed that they could see a lesion. One observer reported improved conspicuity of the lesion with SAE in all 16 image pairs. The other reported improved conspicuity in 15 of the 16 pairs and no difference in one pair. When analyzed objectively, the percentage of color pixels was reduced in all the focal lesions. Color conspicuity averaged 80%, with a range of 61%–89%, while gray-scale conspicuity averaged 8.6%, with a range of 1%–26% (Fig 5). Thus, color conspicuity was always higher than gray conspicuity, with an average value nearly 10 times higher. The difference between the two sets of values was significant (P < .001, Wilcoxon signed rank test).
Detection of SAE-specific Defects
When the observers were presented with image pairs showing areas of liver that were either normal or without definite masses on B-mode images, they both reported no SAE abnormalities on the nine pairs of control images from the three healthy volunteers. Both observers detected SAE defects on every other of the nine image pairs from the nine patients with known metastases. In six of these nine image pairs, both observers considered that at least one of these defects was seen only in SAE mode. With three image pairs, only one of the two observers felt confident that the color defects were SAE specific. These three discordant cases were subsequently reviewed jointly with both observers. In two, there was no difference in the perception of SAE defects, but the observers differed in the confidence with which they perceived B-mode lesions. In one of the three cases, one observer had perceived two small SAE-specific defects that were not seen by the second observer.
In three of the nine image pairs from patients with known metastases, specific biopsy or angiographic proof was obtained that otherwise occult disease was being revealed by SAE. In all three cases, both observers reported SAE-specific defects. In one case, angiography revealed metastases in the right lobe, which was normal on conventional US images but which had shown SAE-specific defects (Fig 6). In the two cases in which B-mode US had failed to depict metastases, SAE-specific defects were seen (Fig 7). Liver biopsy in both cases revealed metastases.
View larger version (158K):
[in this window]
[in a new window]
|
Figure 6a. Longitudinal section through the right lobe of the liver in a 47-year-old male patient in (a) gray-scale and (b) SAE displays. The right lobe of the liver is normal in a, although a single lesion had been seen in segment IV on other views. In b, several color defects are seen within the right lobe. Both observers detected at least one SAE-specific lesion (arrow in b) when presented with the image pair of a and b. (c) Angiogram obtained during the parenchymal phase of selective hepatic arterial injection reveals multiple vascular deposits (arrows) of metastatic carcinoid throughout the liver, including the right lobe. Open hepatic biopsy confirmed metastatic carcinoid tumor.
|
|
View larger version (139K):
[in this window]
[in a new window]
|
Figure 6b. Longitudinal section through the right lobe of the liver in a 47-year-old male patient in (a) gray-scale and (b) SAE displays. The right lobe of the liver is normal in a, although a single lesion had been seen in segment IV on other views. In b, several color defects are seen within the right lobe. Both observers detected at least one SAE-specific lesion (arrow in b) when presented with the image pair of a and b. (c) Angiogram obtained during the parenchymal phase of selective hepatic arterial injection reveals multiple vascular deposits (arrows) of metastatic carcinoid throughout the liver, including the right lobe. Open hepatic biopsy confirmed metastatic carcinoid tumor.
|
|
View larger version (125K):
[in this window]
[in a new window]
|
Figure 6c. Longitudinal section through the right lobe of the liver in a 47-year-old male patient in (a) gray-scale and (b) SAE displays. The right lobe of the liver is normal in a, although a single lesion had been seen in segment IV on other views. In b, several color defects are seen within the right lobe. Both observers detected at least one SAE-specific lesion (arrow in b) when presented with the image pair of a and b. (c) Angiogram obtained during the parenchymal phase of selective hepatic arterial injection reveals multiple vascular deposits (arrows) of metastatic carcinoid throughout the liver, including the right lobe. Open hepatic biopsy confirmed metastatic carcinoid tumor.
|
|
View larger version (162K):
[in this window]
[in a new window]
|
Figure 7a. Longitudinal section through the right lobe of the liver in a 47-year-old female patient. (a) While the B-mode image shows diffuse heterogeneity of the liver, it is hard to clearly define any discrete masses. (b) The SAE image, however, shows multiple defects; the largest of the defects is indicated by an arrow. Both observers detected SAE-specific defects when presented with the image pair of a and b. Metastatic squamous cell carcinoma was proved with hepatic biopsy.
|
|
View larger version (155K):
[in this window]
[in a new window]
|
Figure 7b. Longitudinal section through the right lobe of the liver in a 47-year-old female patient. (a) While the B-mode image shows diffuse heterogeneity of the liver, it is hard to clearly define any discrete masses. (b) The SAE image, however, shows multiple defects; the largest of the defects is indicated by an arrow. Both observers detected SAE-specific defects when presented with the image pair of a and b. Metastatic squamous cell carcinoma was proved with hepatic biopsy.
|
|
|
DISCUSSION
|
|---|
The accurate detection and assessment of metastatic disease is one of the most common and important problems faced in liver imaging today. Good imaging is increasingly important in planning the surgical, radiologic, and pharmacologic treatments that can now be directed at liver metastases (15,16). While most patients will undergo US in any case, many radiologists rely on CT and magnetic resonance (MR) imaging for the confident assessment of liver metastases. Although figures of 70%–85% for the detection of metastatic disease by using US are quoted, the sensitivity of US in the detection of an individual lesion is much harder to assess and is likely to be much lower (17,18). Difficulties with US include the fact that lesions may be missed for operator-dependent or perceptual reasons and that the differences in echogenicity between metastases and the surrounding liver may be low (so-called isoechoic lesions). A technique that could increase conspicuity—as contrast can be increased by using iodinated agents in CT—would be therefore of great value, particularly if the sensitivity of US could be improved to approach that of CT and MR imaging.
This study has demonstrated two main results. Our first main finding is that SAE after the administration of SH U 508A does increase the conspicuity of metastases, both subjectively and objectively. The increase in conspicuity may help in lesion delineation by increasing diagnostic confidence or by demonstrating focal abnormalities to clinicians. Perhaps more important, however, is the indirect implication of our observation. If the conspicuity of visible metastases is increased markedly, then the conspicuity of subtle or isoechoic lesions should also be increased, leading to an increase in the sensitivity of US to metastatic disease.
This is consistent with our second main finding: New lesions appear to be revealed by using this technique. We observed color defects in the absence of definite B-mode lesions (SAE-specific defects) in some patients with widespread liver metastatic disease. In six patients, both observers reported lesions on SAE displays that they both agreed were either completely occult or too subtle to be identified with confidence on gray-scale images. In three of these six patients, angiographic or biopsy proof was obtained that new metastases were being visualized only with the use of SAE. Our two observers did not always agree, however, and in three cases, only one noted SAE-specific defects. It is important, however, that neither reported SAE defects (with or without B-mode lesions) on control image pairs (ie, there were no false-positive findings), and the two observers mainly differed in the confidence with which they perceived gray-scale lesions. The more experienced of the two observers was more confident in identifying metastases on B-mode images and detected more color defects. This suggests that the use of SAE may be helpful in increasing diagnostic confidence for subtle or uncertain lesions or for less experienced observers.
There are several important limitations of this study. The first is the relatively small number of patients. Although the number of patients was adequate to demonstrate that conspicuity is increased by using the technique, larger numbers are needed to show whether sensitivity is improved. The second is that, in many cases, the focal lesion imaged and that in which biopsy was performed were not identical, because the biopsy was not performed at the same time as the contrast material–enhanced study. In the cases of the SAE-specific defects, targeted biopsy would be difficult because of the transience of the effect. A third limitation is that conspicuity is a complex perceptual phenomenon, and the techniques we used to assess it are imperfect. For example, our observers were not completely blinded, in that they were aware of the obvious differences between color and gray-scale images and were aware that all images in the first part of the analysis showed a metastasis. Lesion conspicuity might have been better assessed by using real-time data (eg, videotape analysis), but this would not have enabled us to compare registered, simultaneously obtained color and B-mode information. While our off-line measurement technique, using the color conspicuity score and gray conspicuity score, does have the merit of adding some objectivity, it undoubtedly has limitations. Since the ultrasound beam destroys the microbubbles, accurate measurement of the magnitude of SAE is in itself inherently prone to error. Gray conspicuity scores will be strongly influenced by gain and dynamic range, and the choice of ROI will introduce some subjectivity.
Despite these limitations, the magnitude of the differences in conspicuity, which was much higher in every case when color SAE data were used, combined with the subjective assessment data is evidence of a real benefit.
Our findings of SAE in the liver parenchyma that persisted after clearance from the general vasculature demonstrate that SH U 508A has a liver-specific phase that was not previously recognized. Whether the microbubbles are simply retained in the sinusoidal spaces or whether Kuppfer cells are implicated—as has been demonstrated with the encapsulated microbubble SH U 563A (8)—is not known and is the subject of ongoing research.
This effect could be used in several ways. SAE could be used as an adjunct to a "conventional" contrast-enhanced US examination after the blood pool phase of enhancement is complete. It might thus be used, for example, after a contrast-enhanced US study of a breast carcinoma as part of a search for liver metastases, without requiring any additional contrast material administration; this could have economic advantages. Most important, it exploits an agent already widely available in many countries as a licensed contrast agent. Several points should, however, be noted for any future studies in which it was planned to use SAE to survey the liver in this way.
First, although no SAE defects were seen in the three control subjects, the false-positive rate of this technique in a substantial series has not been investigated.
Second, SAE is a transient effect, as the SH U 508A is inactivated by the act of insonation. It may thus not work well if a particular region of the liver has been studied shortly after SH U 508A administration, for example, as part of a Doppler study. The magnitude of the effect is also influenced by technical factors, such as tissue depth (it does not work well at depths of more than 10–12 cm, depending on the Doppler frequency) and focal zone settings: We have addressed many of these issues in separate research (19). The effect of diffuse liver disease may also prove important. It is not clear exactly how much the presence of parenchymal disease, such as cirrhosis, might alter SAE, although in a separate study (19) we have observed very similar amounts of SAE in patients with and in patients without cirrhosis. If indeed the effect worked well in patients with cirrhosis, another important application could be the detection of hepatocellular carcinoma in the follow-up of patients with cirrhosis.
In conclusion, SH U 508A shows delayed enhancement in the normal liver parenchyma that can be detected by using SAE, and this can be used to increase the conspicuity of liver metastases. In this pilot study, we also observed that it could make occult lesions visible and did not demonstrate any false-positive findings in a small control group. It addresses two major limitations of US: the relatively low intrinsic contrast differences between many normal and abnormal areas of the liver and perceptual problems in appreciating these differences. Further studies are required to accurately evaluate the sensitivity and specificity of this technique and to determine its place in the work-up of patients with suspected metastatic disease of the liver.
|
Acknowledgments
|
|---|
The authors thank Damian Sell, BA, for help in the production of the illustrations for this article and Lippincott Williams & Wilkins for permission to reproduce Figures 6a and 6b. The Levovist used in this study was provided by Schering, Berlin, Germany.
|
Footnotes
|
|---|
Current address: Klinikum Benjamin Franklin, Berlin, Germany. 
Supported in part by Schering, Berlin, Germany. M.J.K.B. supported by the United Kingdom Medical Research Council.
Address reprint requests to M.J.K.B.
From the 1997 RSNA scientific assembly.
Abbreviations: ROI = region of interest
SAE = stimulated acoustic emission
Author contributions: Guarantor of integrity of entire study, M.J.K.B.; study concepts, M.J.K.B., D.O.C., T.A.; study design, M.J.K.B.; definition of intellectual content, M.J.K.B., T.A., D.O.C.; literature research, M.J.K.B.; clinical studies, M.J.K.B., T.A., V.J., J.B.B.; data acquisition, M.J.K.B., T.A., V.J., R.J.E., J.B.B.; data analysis, M.J.K.B., R.J.E., D.O.C., N.P.; statistical analysis, M.J.K.B.; manuscript preparation and review, M.J.K.B., D.O.C., T.A., R.J.E., A.B., R.S.; manuscript editing, M.J.K.B., T.A., D.O.C., N.P., V.J., J.B.B., R.J.E., A.B., R.S.
Received November 18, 1997;
revision requested February 5, 1998; revision received July 13, 1998;
accepted September 11, 1998.
|
References
|
|---|
-
Nanda N, Carstensen EL. Echo-enhancing agents: safety. In: Nanda N, Schlief R, Goldberg BB, eds. Advances in echo imaging using contrast enhancement. 2nd ed. Lancaster, England: Kluwer Academic, 1997; 115-131.
-
Schlief R. Developments in echo-enhancing agents. Clin Radiol 1996; 51(suppl 1):5-7.
-
de Jong N. Physics of microbubble scattering. In: Nanda N, Schlief R, Goldberg BB, eds. Advances in echo imaging using contrast enhancement. 2nd ed. Lancaster, England: Kluwer Academic, 1997; 39-64.
-
Porter TR, Shouping L, Kricsfield D, Armbruster RW. Detection of myocardial perfusion in multiple echocardiographic windows with one intravenous injection of microbubbles using transient response second harmonic imaging. J Am Coll Cardiol 1997; 29:791-799.[Abstract]
-
Kamiyama N, Moriyasu F, Kono Y, Mine Y, Nada T, Yamazaki N. Investigation of the "FLASH echo" signal associated with an ultrasound contrast agent (abstr). Radiology 1996; 201(P):158.
-
Moriyasu F, Kono Y, Nada T, Matsumura T, Suginoshita Y, Kobayashi K. FLASH echo (passive cavitation) imaging of the liver by using US contrast agents and intermittent scanning sequence (abstr). Radiology 1996; 201(P):196.
-
Burns PN, Fritzsch T, Weitschies W, Uhlendorf V, Hope-Simpson D, Powers JE. Pseudo-Doppler shifts from stationary tissue due to the stimulated emission of ultrasound from a new microsphere contrast agent (abstr). Radiology 1995; 197(P):402.[Free Full Text]
-
Fritzsch T, Heldmann D, Reinhardt M. The potential of a novel contrast medium. In: Goldberg G, eds. Ultrasound contrast agents. London, England: Dunitz, 1997; 169-176.
-
Bauer A, Mahler M, Urbank A, Zomack M, Schlief R, Niendorf HP. Microvascular imaging: results from a phase I study of the novel polymeric contrast agent SHU 563 A. In: Nanda N, Schlief R, Goldberg BB, eds. Advances in echo imaging using contrast enhancement. 2nd ed. Lancaster, England: Kluwer Academic, 1997; 685-690.
-
Uhlendorf V, Hoffman C. Non-linear acoustic response of coated microbubbles in diagnostic ultrasound Proceedings of the IEEE International Ultrasonics Symposium. Piscataway, NJ: IEEE, 1994; 1559-1562.
-
Bauer A, Zomack M, Lohse D, Hilgenfeldt S, Urbank A, Schlief R. Acoustic response of microbubbles: acoustic emission and localization imaging (abstr). Radiology 1997; 205(P):278.[Abstract/Free Full Text]
-
Bauer A, Schlief R, Zomack M, Urbank A, Niendorf HP. Acoustically stimulated microbubbles in diagnostic ultrasound: properties and implications for diagnostic use. In: Nanda N, Schlief R, Goldberg BB, eds. Advances in echo imaging using contrast enhancement. 2nd ed. Lancaster, England: Kluwer Academic, 1997; 669-684.
-
Tieman K, Becher H, Bimmel D, Schlief R, Nanda N. Stimulated acoustic emission: non back scatter contrast effect of microbubbles seen with harmonic power Doppler imaging. Echocardiography 1997; 14:65-69.[Medline]
-
Blomley MJK, Albrecht T, Cosgrove DO, Jayaram V, Butler-Barnes J, Eckersley RJ. Stimulated acoustic emission in liver parenchyma with Levovist. Lancet 1998; 351(9102):568.
-
Baker ME, Pelley R. Hepatic metastases: basic principles and implications for radiologists. Radiology 1995; 197:329-337.[Free Full Text]
-
Tranberg K, Bengmark S. Metastatic tumours of the liver. In: Blumgart LH, ed. Surgery of the liver and biliary tract. 2nd ed. Edinburgh, Scotland: Churchill Livingstone, 1995; 4:1385-1398.
-
Cosgrove DO, Bolondi L. Malignant liver disease. In: Cosgrove D, Meire H, Dewbury K, eds. Abdominal and general ultrasound. Vol 1. Edinburgh, Scotland: Churchill Livingstone, 1993; 271-293.
-
Carter R, Hemingway D, Pickard R, Poon FW, MacKillop JA, McArdle CS. A prospective study of six methods for detection of hepatic colorectal metastases. Ann Royal Coll Surg Engl 1996; 78:27-30.
-
Blomley MJK, Albrecht T, Cosgrove DO, et al. The use of stimulated acoustic emission to image a late liver-specific phase of Levovist: an investigation in normal volunteers and in patients with and without liver disease. Ultrasound Med Biol ; (in press).
TOP
Abstract
Introduction
) and color SAE mode (
) information.