| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Medical Imaging, Toronto General Hospital, University Health Network, 200 Elizabeth St., Toronto, Ontario, Canada M5G 2C4.
Received January 18, 2002;
accepted after revision August 27, 2002.
Partially supported by Berlex Canada.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. In our prospective study, 50 patients underwent routine sonography followed by postvascular Levovist-enhanced pulse inversion imaging of the liver. Thirty-six patients had malignant disease (28 invasive parenchymal tumors and eight intraductal tumors), and 14 had benign disease. The 36 malignancies included 29 cholangiocarcinomas, six invasive gallbladder carcinomas, and one colon metastasis. Fourteen patients had benign disease: benign strictures (n = 5), primary sclerosing cholangitis (n = 5), chronic Mirizzi's syndrome (n = 1), varicosities of the parabiliary venous plexus (n = 1), and inflammatory liver lesions (n = 2). Sonographic findings in all 50 patients were correlated with findings from other imaging modalities (n = 50) as well as surgical specimens (n = 20), core biopsies (n = 3), and both clinical and imaging follow-ups (n = 24).
RESULTS. Seventeen (61%) of the 28 invasive intraparenchymal malignancies were visualized on routine sonograms, whereas all 28 (100%) were visualized on enhanced sonograms (p < 0.01). In 15 (88%) of 17 patients in whom tumor was seen on routine sonograms, contrast-enhanced sonography showed further mass extent, increased conspicuity, or satellite nodules not visualized on the baseline image. All eight noninvasive intraductal malignancies were correctly identified and staged on the routine sonography. In one of these patients, hepatic invasion was prospectively overcalled on the enhanced image. Of the 14 benign lesions, three had inflammatory periductal abnormalities seen exclusively or to advantage on the enhanced study. Correct prediction of resectability in the 16 patients with malignant disease who underwent surgery improved from 11 (69%) of 16 on unenhanced sonography to 15 (94%) of 16 on enhanced sonography (p = 0.13).
CONCLUSION. Detection and staging of malignant hilar obstructions are improved by the use of Levovist in the postvascular phase of sonography compared with routine sonography.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The contrast agent Levovist (SHU 508A; Schering, Berlin, Germany) has been shown to persist in the normal hepatic parenchyma after a brief vascular phase. This liver-specific postvascular phase of enhancement has been used for improved visualization of hepatic malignancies. Several studies have shown that use of Levovist results in increased conspicuity of metastases as well as in the detection of more lesions compared with findings on unenhanced sonography [5, 6, 7]. We hypothesized that postvascular delayed sonograms obtained with Levovist would show increased conspicuity of hilar cholangiocarcinoma, improving sonographic detection and staging of this disease. In this preliminary study, we investigated the value of Levovist-enhanced sonography of the liver as compared with that of unenhanced sonography in patients presenting with hilar biliary obstruction.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Thirty-six of the 50 patients had a malignant hilar obstruction. Twenty-eight of the 36 had an invasive intraparenchymal tumor. Eight of the 36 had an intraductal tumor (five affecting extrahepatic hilar ducts and three involving the intrahepatic ducts only). Final diagnosis included 29 hilar cholangiocarcinomas, six invasive carcinomas of the gallbladder, and one colon metastasis. In 19 of the 36 patients with malignant disease, the diagnoses were confirmed at surgical resection (n = 16) or biopsy (n = 3). For all patients who had a surgical resection, detailed pathologic examination was performed on the surgical specimens, and the results were compared with the sonographic findings. In 17 of the 36 patients, the clinical and imaging features of overwhelming malignancy were such that palliative therapy was administered without tissue biopsy or with inconclusive biopsy results. Correlative imaging in the 36 patients with malignant disease included CT scans (n = 21 patients), MR images (n = 16), endoscopic retrograde cholangiopancreatograms (ERCP), or percutaneous transhepatic cholangiograms (n = 7), and positron emission tomographic scans (n = 1).
In 14 of the 50 patients, the cause of the hilar biliary obstruction was related to benign disease: five benign strictures, five primary sclerosing cholangitis, one case of chronic Mirizzi's syndrome, one case of varicosities of the parabiliary venous plexus, and two inflammatory liver lesions. Both inflammatory lesions were prospectively believed to represent malignant lesions on all imaging modalities. All of the patients with sclerosing cholangitis had clinical evidence of the disease. In all patients with benign disease, the lack of progression of symptoms during follow-up (range of follow-up periods, 6-16 months; mean, 12.8 months) and correlative imaging (MR imaging, n = 8; ERCP or percutaneous transhepatic cholangiography, n = 10; CT, n = 12); biopsy, (n = 1); and surgery, (n = 4) were used to exclude malignancy and confirm sonographic findings.
Sonographic Technique
All patients initially underwent a routine physician-performed sonographic
examination, which consisted of a thorough gray-scale evaluation of the liver,
bile ducts, and porta hepatis as well as color Doppler sonographic assessment
of the hepatic vasculature, with all findings being recorded. Independent of
this routine sonography, the patients underwent Levovist-enhanced sonography
performed on a scanner (model 5000; ATL Ultrasound, Bothell, WA) equipped with
pulse inversion grayscale imaging software. Two of the investigators performed
all examinations with a convex array probe (Apogee 800 C5-2; ATL Ultrasound)
and maximal allowed mechanical index (1-1.3). The contrast-enhanced
examination consisted of two unenhanced continuous sweeps (90 frames) in the
orthogonal transverse and sagittal planes to cover the entire liver and porta
hepatis, followed by identical sweeps performed 4 min after injection of the
contrast agent (300 mg/mL of Levovist in three boluses of 4.5 mL each).
Contrast-enhanced sweeps were performed in the sagittal and transverse planes,
as on the unenhanced study, with a third performed at the discretion of the
investigators to repeat the sweep providing the most information. All
unenhanced and contrast-enhanced sweeps were stored on our PACS (picture
archiving and communication system) as 90-frame cine loops.
Data Analysis
Without knowledge of the results from other imaging tests, the two authors
who performed the sonography prospectively interpreted all the images by
consensus. The initial routine sonogram was analyzed first, and the findings
were recorded. The contrast-enhanced sweeps were then carefully analyzed and
compared with the unenhanced sweeps and the initial sonographic findings.
Unenhanced and contrast-enhanced sweeps were placed side by side and analyzed
frame by frame to ensure direct comparison at exactly the same anatomic
location. To avoid overcalling normal nonenhancing structures in the liver as
evidence of disease, we used the unenhanced images as a reference for the
interpretation of the contrast-enhanced sonograms. If a focal abnormality was
detected on contrast-enhanced sonograms, the lesion was subjectively assessed
for size, conspicuity, and border definition in comparison with its depiction
on unenhanced sonograms. Because the postvascular enhancement with Levovist
affects only the liver, extrahepatic tumor extent and adenopathy were assessed
primarily with the baseline sonograms.
The determination of resectability was made by clinicians on the basis of the results of all imaging examinations available to them. To determine the usefulness of contrast-enhanced sonography as a tool to predict resectability of hilar obstructive lesions if sonography alone had been used, we also performed a separate analysis on all of the patients who had surgery (n = 20). Sixteen of the patients had histologically proven malignancies, and four had benign disease. Patients with benign entities that were resected surgically included one patient with an inflammatory pseudotumor of the liver, one patient with inflammatory periductal changes due to chronic biliary obstruction, one patient with Mirizzi's syndrome, and one patient with a benign stricture. In the surgical group, retrospective analysis of resectability was determined by consensus of the three authors. The criteria for resectability were based on criteria developed by Jarnagin et al. [8] and are summarized in Appendix 1. Statistical analysis was performed using the exact form of the McNemar test.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
Of the eight patients with only an intraductal tumor, an intraductal irregularity (n = 2) or extrahepatic tumor (n = 5) was seen on the unenhanced sonogram with no evidence of intraparenchymal invasion on contrast-enhanced sonograms (Figs. 3A, 3B, 3C and 4A, 4B, 4C). In one patient, unenhanced sonography was correctly interpreted as showing an intraductal tumor, whereas contrast-enhanced sonography was falsely interpreted as revealing an intraductal tumor with parenchymal invasion. The surgical specimen showed the intraductal tumor only.
|
|
|
|
|
|
When a tumor was identified on unenhanced sonography, it appeared as a hypoechoic mass, with the exception of an intraductal polypoid lesion and a colon metastasis, both of which were echogenic. On contrast-enhanced sonography, all masses appeared hypoechoic compared with the enhanced liver parenchyma. None of the masses showed significant vascularity on the color Doppler sonographic assessment.
Patients with Benign Disease
Fourteen patients had benign lesions. Five patients had primary sclerosing
cholangitis, four of whom exhibited ductal irregularity and wall thickening on
both unenhanced and contrast-enhanced sonography with no evidence of an
intraparenchymal mass. In the fifth patient, extensive periductal tissue was
identified on the contrast-enhanced sonograms that had not been identified on
the unenhanced images (Fig.
5A,
5B,
5C). These findings were
exactly as those seen on the patient's CT scan, and on both modalities the
findings were correctly predicted to represent benign disease.
|
|
|
In seven of the nine patients with other types of benign disease, no evidence of an invasive intraparenchymal mass was seen on either unenhanced or contrast-enhanced sonography. In one patient who had undergone a prior liver resection for cholangiocarcinoma, extensive periductal masses surrounding a dilated segmental duct were identified on contrast-enhanced sonography and MR imaging; the masses were incorrectly interpreted as recurrent cholangiocarcinoma on both modalities (Fig. 6A, 6B, 6C). The final histologic evaluation showed inflammatory and fibrotic tissue surrounding an anomalous, surgically obstructed bile duct. In another patient, all imaging modalities (MR imaging, CT, and unenhanced and contrast-enhanced sonography) showed multifocal hilar masses believed to represent a Klatskin's tumor. At pathology, surgical specimens showed extensive peribiliary inflammation of unknown origin without evidence of malignancy.
|
|
|
Analysis of Patients Who Underwent Surgery
Twenty patients of the 50 in our study underwent surgery, 16 patients with
malignancies and four with benign conditions. In patients with malignant
disease, resectability was correctly predicted in 11 (69%) of 16 using
unenhanced sonography and in 15 (94%) of 16 on Levovist-enhanced sonography.
The improvement in prediction of resectability was not statistically
significant (p = 0.13), likely due to the small number of patients in
this subgroup; however, there is a clear trend of improvement when
contrast-enhanced sonography was used.
Successful surgical resection was correctly predicted in all four patients with benign disease on both unenhanced and contrast-enhanced sonography. In the two patients with benign inflammatory liver masses associated with hilar biliary obstruction, the location and extent of disease were correctly identified on postvascular phase sonography. In these two patients, findings on all imaging modalities led to incorrect preoperative predictions of malignancy.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The liver-specific, postvascular phase of contrast enhancement of liver with Levovist has been well described in the literature [5, 6, 7, 10]. The microbubbles are selectively trapped in the hepatic parenchyma a few minutes after intravascular injection of the contrast agent. A high-mechanical-index sweep through the liver produces bright transient enhancement of the liver as the accumulated microbubbles are destroyed. The marked increase in the echogenicity of the liver parenchyma greatly improves the contrast between liver and nonhepatocytic components such as portal triads and liver masses.
Several studies have documented improved detection of metastases and hepatocellular carcinoma [5, 6, 7]. To our knowledge, cholangiocarcinoma has not routinely been evaluated with Levovist-enhanced sonography, although in a series of 46 patients with liver masses studied by Bertolotto et al. [10], two cases of peripheral cholangiocarcinoma were included.
Most (36/50) of our patients had a malignancy, a reflection of the surgical referral pattern in our institution. Levovist-enhanced sonography successfully depicted the obstructing lesion in all 36 patients with malignant hilar obstruction. In 11 (39%) of the 28 patients with an invasive intraparenchymal tumor, the mass was seen only on the contrast-enhanced examinations, a statistically significant improvement. In 15 (88%) of the 17 patients with invasive tumor that was seen on unenhanced sonography, further extension of the mass and increased conspicuity were noted on the enhanced sonography. The results of our study show that postvascular phase Levovist-enhanced sonography greatly improved the detection and determination of the extent of hilar malignant disease, which led to improved staging and resultant change in management in some patients. Failure of contrast-enhanced sonography to reveal a mass in patients with benign pathology increased our confidence level in recommending conservative management.
The main and the distal right and left hepatic ducts are extrahepatic, but their branches extend immediately into the liver. Malignant hilar cholangiocarcinoma has a propensity to grow axially along and into the ducts and invade the liver parenchyma. It is this invasive component that Levovist enhancement excellently depicts, particularly as the periductal supportive tissue becomes thinner with higher order branching. Frequently, on enhanced sonography, tumor resembling a grapelike cluster can be seen following the ducts into the liver. These small invasive foci are often not seen on unenhanced sonography due to the isoechoic nature of the lesions. Sonography has excellent spatial resolution, and our results suggest that Levovist enhancement improves the contrast resolution of the modality and therefore its sensitivity for tumor detection. We believe that Levovist-enhanced sonography probably offers no significant improvement in specificity over unenhanced sonography, because all nonhepatocytic tissues are nonenhancing, including both tumor and inflammatory or fibrotic foci (Figs. 5A, 5B, 5C and 6A, 6B, 6C).
Our major objective in this preliminary study was not to test the diagnostic performance of Levovist-enhanced sonography against other imaging tests but rather to determine whether the technique had the potential to improve the performance of sonography used in the assessment of hilar biliary obstruction. Our protocol did not include correlative imaging, which was performed at the discretion of the referring clinician. In addition, only 20 patients underwent the gold standard of surgery. For both of these reasons, we could not compare the performance of contrast-enhanced sonography with that of CT or MR imaging. We did, however, confirm the results with all available clinical and imaging records and with follow-up examinations. Our analysis showed that two patients predicted to have unresectable disease on the basis of postvascular phase Levovist-enhanced sonography were found to have unresectable disease; surgery had been performed because CT findings suggested resectable disease.
Important principles for the accurate interpretation of contrast-enhanced sonograms are well exemplified by the two false-positive results in our study. Extensive benign fibrotic and inflammatory nodules surrounding a dilated segmental duct were mistaken for tumor on both sonography and MR imaging in a man who had undergone liver resection 6 years earlier (Fig. 6A, 6B, 6C). In another patient, multifocal periductal masses seen on all imaging modalities were incorrectly interpreted as representing cholangiocarcinoma. At pathology, a surgical specimen revealed benign peribiliary inflammatory masses. These cases taught us the importance of recognizing that only hepatocyte-containing tissue enhances; therefore, all other types of tissue appear hypoechoic compared with the enhancing liver parenchyma. These tissues include tumor, fibrous and inflammatory tissue, normal periductal connective tissue of the portal triads, and (occasionally) tissue of the portal vein and hepatic artery branches.
The lack of enhancement of normal connective tissue also led to overcalling the extent of tumor seen in a third patient on contrast-enhanced sonography. An intraductal tumor adjacent to the ascending left portal venous branch was interpreted as showing parenchymal invasion; however, the finding actually represented normal connective tissue in the portal triad. Direct analysis with the unenhanced scan made this obvious in retrospect. The error occurred early in our study (our seventh patient). Our result was concordant with a CT scan, and neither finding altered the tumor stage or treatment. The case of this patient reinforced our belief in the necessity of careful side-by-side analysis of unenhanced and contrast-enhanced cine loops to correctly estimate tumor extent and to correctly identify the normal liver vasculature and supporting stroma on the enhanced sonograms. We found that orthogonal sweeps were essential to confirming or discounting questionable lesions seen in one plane.
We acknowledge that there is a significant learning curve for the successful performance of Levovist-enhanced postvascular phase sonography. The person performing the examination must be able to cover, in a single sweep, as much of the liver as the direction of the sweep allows. Several practice sweeps are required to find the optimal patient and transducer positions, as well as the best sweep direction and speed. The practice sweeps also allow the patient to become familiar with the technique and with the requirement for a long suspended inspiration, which must be maintained for the entire sweep. Furthermore, to prevent bubble destruction, the sonographic mechanism must be frozen before the initiation of the sweep. The mechanical index must be maximal to produce good bubble destruction during the sweep, and the gain must be lowered to avoid a very bright and difficult-to-interpret image. The focal zone is best set at the level of the porta hepatis to clearly depict disease centered at this level of the liver. We did not measure the duration of the contrast-enhanced examination because study times became progressively shorter with increased proficiency. We believe that the contrast-enhanced study adds 30-45 min to the routine sonographic examination, including the interpretation of the findings, which can be performed after the patient has left the department.
The fact that both the unenhanced and contrast-enhanced sonograms were interpreted by the same authors at the same sitting may have introduced bias in the interpretations and is a limitation of the study methodology. However, because there was an obvious advantage in using the contrast-enhanced sonography for most patients, the additional information was shared with the clinicians, making temporally separate analyses impossible.
Our study has shown that following an unenhanced sonographic examination with a Levovist-enhanced sonographic examination significantly improved both the intrahepatic detection and staging of malignant biliary obstruction compared with using an unenhanced examination alone. Direct identification of cholangiocarcinoma augments information that previously had to be inferred from indirect evidence such as segmental intraductal dilatation. In our institution, our clinicians have adopted postvascular phase Levovist-enhanced sonography as a standard method of evaluating hilar obstructions. We are currently engaged in a prospective study comparing the performance of unenhanced and Levovist-enhanced sonography with the performances of MR imaging and MR cholangiography.
Although Levovist is not approved for medical imaging in the United States, phase III trials of several second-generation contrast agents are now being conducted. Some of these contrast agents offer the same postvascular enhancement of normal liver tissue that Levovist provides.
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J.-Y. Choi, M.-J. Kim, J. M. Lee, K. W. Kim, J. Y. Lee, J. K. Han, and B. I. Choi Hilar Cholangiocarcinoma: Role of Preoperative Imaging with Sonography, MDCT, MRI, and Direct Cholangiography Am. J. Roentgenol., November 1, 2008; 191(5): 1448 - 1457. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
