| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Full Paper |
Division of Hepatobiliary Pancreatic Medical Oncology, National Cancer Centre Hospital East, 6-5-1, Kashiwanoha, Kashiwa, Chiba, 277-8577, Japan
Correspondence: Jungi Furuse
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Early enhancement on dynamic CT and/or hepatic arteriography can differentiate hepatic tumours, especially hepatocellular carcinoma (HCC). Colour Doppler imaging, including power Doppler imaging, is also considered a useful means of detecting the tumour vascularity of HCCs because it is non-invasive and generates real-time images [12–15]. However, it cannot depict signals throughout the entire tumour, and it sometimes cannot clearly distinguish HCC from other hypervascular hepatic tumours. Detailed vascularization can be imaged in various hepatic tumours by contrast-enhanced ultrasound using harmonic based techniques [7–9]. Coded phase inversion harmonic imaging can depict fine vascular images in hepatic tumours at the early phase of contrast-enhanced ultrasound. To our knowledge, few reports have described vascular images of hepatic tumours. We considered that coded harmonic ultrasound imaging with Levovist (Schering AG, Berlin, Germany) could also show characteristic enhancement patterns and could thus differentially diagnose hepatic tumours. We therefore reviewed the findings of contrast-enhanced ultrasound of HCCs, intrahepatic cholangiocarcinomas, hepatic metastases, hepatic haemangiomas and focal nodular hyperplasias (FNHs).
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The HCC tumours were histologically diagnosed by an examination of specimens obtained using ultrasound guided biopsy using a 21-gauge needle (Sonopsy-C1; Hakko, Tokyo, Japan) in 23 lesions and by hepatic resection in 18 lesions. Tumour size was determined as the longest dimension on ultrasound images, and the range was from 1.4 cm to 7.1 cm in diameter (mean±standard deviation, 2.8±1.3 cm).
All intrahepatic cholangiocarcinomas were histologically diagnosed as adenocarcinoma from needle biopsy specimens (n=4) and by hepatic resection (n=1). Chest and abdominal CT and gastrointestinal examinations confirmed that these patients had no tumours in other organs. Tumours measured as the longest dimension on ultrasound images ranged from 4.0 cm to 6.7 cm in diameter (mean±standard deviation, 5.8±2.0 cm).
To precisely diagnose hepatic metastasis, only patients with colon carcinoma or gastric carcinoma who intended to undergo hepatectomy were examined and all tumours were histologically diagnosed as adenocarcinoma. Tumours measured as the longest dimension on ultrasound images ranged from 1.3 cm to 4.7 cm in diameter (mean±standard deviation, 3.1±1.5 cm).
Patients with hepatic haemangioma were selected for examination by contrast-enhanced ultrasound only when the following conditions were satisfied: chronic liver disease due to hepatitis B or C virus (n=2); malignant tumours in other organs (n=5); or liver tumours without typical findings of haemangioma by B-mode ultrasound (n=10). All haemangiomas were finally diagnosed by two-phase (hepatic arterial and venous phases) imaging at dynamic contrast-enhanced spiral CT and/or MRI. Imaging confirmed typical enhancement characteristics in all of them, namely peripheral nodular enhancement and subsequent filling during arterial and venous phase imaging, respectively.
Tumours measured as the longest dimension on ultrasound images ranged in size from 1.3 cm to 13.9 cm in diameter (mean±standard deviation, 4.7±3.9 cm).
FNH lesions were diagnosed using dynamic contrast-enhanced CT and MRI that indicated strong homogeneous enhancement at the arterial phase and/or a central scar. Tumours measured as the longest dimension on ultrasound images were 6.4 cm and 2.0 cm in diameter, respectively.
We diagnosed five benign hepatic nodules in liver cirrhosis and one focal fatty lesion in six other tumours, using ultrasound guided biopsy specimens. Tumours measured as the longest dimension on ultrasound images ranged in size from 1.1 cm to 2.2 cm in diameter (mean±standard deviation, 1.5±0.4 cm).
Contrast agent
All patients were injected with the ultrasound contrast agent, Levovist, which is composed of 99.9% galactose and 0.1% palmitic acid. The agent (2.5 g) was shaken for about 10 s with 7 ml of sterile water to yield an opalescent suspension of air microbubbles. The suspension was equilibrated for a few minutes, then manually injected through a 20-gauge cannula placed into the antecubital vein as a bolus infusion (300 mg ml-1 of contrast agent; approximately 8 ml in total). The cannula was then flushed with an additional 10 ml of physiological saline.
Ultrasound examination
All ultrasound examinations were performed with LOGIQ 700 EXPERT Series (GE Medical Systems, Milwaukee, WI) and a 2–4 MHz curved-array wide-band transducer. Ultrasound images were continuously recorded on videotape at the beginning of the administration of the contrast agent. We examined tumours with coded harmonic ultrasound imaging, and observed images for 2 min from the beginning of the administration as the vascular phase. Vascular images were initially observed by continuous transmission for about 30 s after the hepatic artery was enhanced, followed by intermittent transmissions of 1 s or 2 s proceed for about 20 s. Thereafter, transmission was returned to the continuous mode until the end of the examination (Figure 1
). The mechanical index values were set at the default setting of MI of 0.6–0.8. The acoustic output was 100%. The frame rate was automatically set at 7 Hz to 9 Hz in the continuous mode. The mechanical index values and the acoustic output were not changed during intermittent transmission. While imaging tumours, we fixed the transducer at the position where the tumour was initially drawn in the maximal diameter. The transducer was repeatedly moved a little to sweep through the tumour during observation after enhancement.
|
Sensitivity was calculated as the number of lesions diagnosed by contrast-enhanced ultrasound divided by the total number of lesions diagnosed. Specificity was calculated as the number of negative lesions diagnosed by contrast-enhanced ultrasound divided by the total number of negative lesions diagnosed in each hepatic tumour. Results were statistically analyzed using the chi-square test to compare detection rates of the enhancement in each hepatic tumour. Differences with p values below 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
We examined the relationship between depth from the skin to the centre of the tumour and absence of enhancement effects in HCC, hepatic metastases, and haemangioma. In HCC, the depth ranged from 1.0 cm to 10.0 cm (average, 5.6 cm) in 38 tumours with enhancement and 3.0 cm, 10.0 cm, and 10.5 cm (average, 7.8 cm) in 3 tumours without enhancement. In hepatic metastases, the depth ranged from 3.0 cm to 10.0 cm (average; 5.8 cm) in 15 tumours with enhancement and 5.0 cm, 5.0 cm, and 7.0 cm (average, 5.7 cm) in 3 tumours without enhancement. In haemangioma, the depth ranged from 3.0 cm to 12.0 cm (average: 6.2 cm) in 15 tumours with enhancement and 2.0 cm, 4.0 cm, 7.0 cm, and 9.0 cm (average, 5.5 cm) in 4 tumours without enhancement. Tumour depth was not correlated with the absence of enhancement in these tumours.
Hepatocellular carcinoma
The most common enhancement pattern was tumour vessels with flow spreading into the tumour and/or homogeneously stained hyperechoic images. Tumour vessels were typically observed from outside to inside the tumour, with fine flow branching into the tumour. These types of vascular images were distinguished in 34 (82.9%) of the 41 lesions (Figure 2a, b). Perfusion flow images in very small vessels revealed homogeneously stained hyperechoic areas within the tumour (Figure 2c). Staining was similarly homogeneous in 32 (78.0%) lesions and intermittent imaging or sweeping through the tumour rendered them more obvious (Table 1). These findings of spreading tumour vessels and/or stained hyperechoic images indicated HCC with 82.9% sensitivity and 93.8% specificity.
|
|
) as fine images of tumour vessels only in the periphery of the tumour 30 s to 45 s after the injection of Levovist. A linear, homogeneously stained hyperechoic area appeared along the tumour border 60 s after the injection, but the central area of the tumour was not enhanced. However, one lesion had tumour vessels that flowed into the tumour, which was stained throughout (Table 1
), and dynamic CT as well as hepatic arteriography showed hypervascularity, indicating a diagnosis of HCC. However, hepatic resection finally confirmed that this tumour was intrahepatic cholangiocarcinoma. Peripheral enhancement was neither very sensitive (60.0%) nor specific (65.5%) in defining intrahepatic cholangiocarcinoma.
|
). However, vessels did not have flow that spread into any of these tumours and no staining within the tumour was evident. No enhancement was observed in the three other lesions. Peripheral enhancement indicating a diagnosis of hepatic metastasis was detected in contrast-enhanced ultrasound with 83.3% sensitivity and 76.4% specificity.
|
). However, pooling was observed at approximately 60 s or later in 13 (76.5%) of the 17 lesions (Figure 5c) as intensely stained hyperechoic areas that slowly spread into central area of the lesion. Such pooling was clarified by intermittent imaging or by sweeping through the tumour, and it was the most frequent type of enhancement in images of haemangiomas. None of the other four lesions were enhanced and all were hyperechoic in fundamental B-mode. These findings of pooling produced 76.5% sensitivity and 100% specificity in diagnosing haemangioma.
|
). In addition, vessels were radially located in the mass for 2 min or more (Figure 6c). In a larger lesion, perfusion flow images were revealed by 1-s intermittent transmission as diffuse vascularity within the tumour as in HCC.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
HCCs generally have an abundant blood supply from the hepatic arteries, and they are characterized as tumour vessels and/or tumour staining by hepatic arteriography and early enhancement on dynamic CT. The present study using contrast-enhanced and coded harmonic ultrasound imaging also found that tumour vessels and stain images are typical of HCC, even when less than 2 cm in diameter. Although these results should be further assessed using other novel modalities of ultrasound and in comparison studies with CT, MRI and arteriography, contrast-enhanced ultrasound can precisely diagnose many HCCs.
The present study selected only patients with one or two hepatic tumours because differentiating metastatic and primary hepatic tumours is very important for designing treatment strategies. Peripheral enhancement was the most frequent finding in hepatic metastasis and in intrahepatic cholangiocarcinoma. It appeared as fine images of tumour vessels only in the periphery of the tumour soon after injection of the contrast agent and as linear and homogeneously stained hyperechoic areas along tumour border starting from 60 s after the injection. Although intrahepatic cholangiocarcinoma and hepatic metastasis were indistinguishable, HCCs were distinguished from hepatic metastases in 49 (83.1%) of 59 lesions.
Pooling in the periphery of the tumour was the most characteristic feature of haemangioma in this study. This appeared as an intensely stained hyperechoic area in the periphery of the tumour. Hepatic tumours could be defined as haemangioma when this pooling was evident. Kim et al [16] reported that peripheral globular or rim-like enhancement with progressive centripetal fill-in was potentially useful for a specific diagnosis of haemangioma in contrast-enhanced ultrasound. This seems similar to the pooling appearance in the periphery. Detection rates for both were similar; that for peripheral globular enhancement was 70% and that of pooling was 76.5%. However, in this study, pooling did not appear in four lesions, all of which were visualized as haemangioma due to findings such as hyperechoic, rounded well-defined lesions in B-mode. We could diagnose most haemangiomas using B-mode and/or contrast-enhanced ultrasound.
A central scar and a haemodynamic profile of a centrifugal blood supply are key features of FNH [21, 22]. A central scar and strong homogeneous enhancement at the arterial phase are typically revealed on dynamic CT or MRI images. Furthermore, the centrifugal blood supply is a characteristic finding at hepatic arteriography [22]. In a few patients in the present study, a tortuous artery flowing into the centre of the mass and vessels observed as radial images in the mass were typically obvious for 2 min or more on contrast-enhanced ultrasound. These findings are similar to those of arteriography, and they are difficult to recognize using CT and MRI. These vascular images differed from those of HCC, and to distinguish between FNH and HCC should be useful. Our results should be further assessed in a larger number of patients with FNH.
We applied coded harmonic imaging, which is a grey scale imaging method of contrast ultrasound. Other contrast imaging modes such as pulse or phase inversion suppress only the fundamental signal, leaving both tissue harmonic and contrast harmonic signals displayed, while coded harmonic imaging expands coded technology to contrast imaging. Coded technology in coded harmonic imaging would increase both tissue harmonic and contrast harmonic signals, suppressing the fundamental signal at the first stage. Additionally, the coded harmonic imaging decoder further preferentially suppressed the non-moving harmonic signal from the scan line of each tissue specimen.
Novel agents for contrast-enhanced ultrasound have been developed such as DMP115 (Definity), NC100100 (Sonazoid), and BR1 (SonoVue), and they have been applied to abdominal examinations, particularly for focal hepatic lesions [23–25]. DMP115 consists of stabilized microbubbles of perfluoropropane, and Maruyama et al [23] reported that it could reveal fine tumour vessels at the arterial phase corresponding to findings on angiography. On the other hand, Levovist, which is an air-based contrast agent, has a weak harmonic response compared with such newer agents. Dill-Macky et al [20] reported that Levovist-enhanced pulse-inversion interval-delay flash and post-vascular phase images are helpful in the differential diagnosis of focal hepatic lesions, but it is less helpful in terms of the vascular image interpretation of focal hepatic masses.
We focused on the vascular phase of various types of hepatic tumours using coded harmonic imaging with Levovist enhancement to determine whether characteristic findings could be established that correspond to tumour vascularity. The results showed that continuous or intermittent transmission of 1 s or 2 s could image tumour vascularity particularly in HCC and haemangioma. Precise observation during the 2 min immediately after the administration using coded harmonic imaging is also useful for evaluating vascularity in some hepatic tumours.
The present study has some limitations that should be recognized with respect to the differential diagnosis of hepatic tumours. We attempted to diagnose hepatic tumours at the vascular phase of contrast-enhanced ultrasound, but late phase images would also be required to more precisely diagnose these tumours. We did not examine them because a second injection of contrast agent would be needed to examine the late phase. In this study, when physicians analysed vascular images and diagnosed tumours on them, they were blind to information such as the patient's name and the diagnosis. However, blind reading has some limitations. For example, tumours could be diagnosed from ultrasound grey scale findings. We calculated the sensitivity and specificity in the diagnosis, but these figures may not be reliable as the population was skewed with many HCCs compared with other hepatic tumours. We could not examine many patients with cholangiocarcinoma and FNH. Whether contrast-enhanced ultrasound would be useful in differential diagnosis between these tumours and HCC requires further assessment. Furthermore, it remains difficult to distinguish intrahepatic cholangiocarcinoma from metastatic tumours in some patients.
|
Acknowledgments |
|---|
Received for publication September 23, 2002. Revision received February 17, 2003. Accepted for publication March 17, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  O. Catalano, V. Migaleddu, E. Quaia, and G. Caruso Terminology for Contrast-Enhanced Sonography: A Practical Glossary J. Ultrasound Med., June 1, 2007; 26(6): 717 - 730. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Iijima, S. Suzuki, F. Moriyasu, K. Yanagisawa, M. Yamada, T. Miyahara, M. Yokoi, T. Horibe, N. Ito, D. Kakizaki, et al. Visualization of the Drainage Veins With Contrast-Enhanced Sonography Was Useful in Diagnosis of Small Focal Nodular Hyperplasia J. Ultrasound Med., June 1, 2006; 25(6): 799 - 803. [Full Text] [PDF]
|
|
|
|
|
|
H.-X. Xu, M.-D. Lu, G.-J. Liu, X.-Y. Xie, Z.-F. Xu, Y.-L. Zheng, and J.-Y. Liang Imaging of Peripheral Cholangiocarcinoma With Low-Mechanical Index Contrast-Enhanced Sonography and SonoVue: Initial Experience J. Ultrasound Med., January 1, 2006; 25(1): 23 - 33. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Catalano, A. Nunziata, R. Lobianco, and A. Siani Real-Time Harmonic Contrast Material-specific US of Focal Liver Lesions RadioGraphics, March 1, 2005; 25(2): 333 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. V. Sahani and S. P. Kalva Imaging the Liver Oncologist, July 1, 2004; 9(4): 385 - 397. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Nagase, J. Furuse, H. Ishii, and M. Yoshino Evaluation of Contrast Enhancement Patterns in Pancreatic Tumors by Coded Harmonic Sonographic Imaging With a Microbubble Contrast Agent J. Ultrasound Med., August 1, 2003; 22(8): 789 - 795. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
