| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Hepatobiliary Imaging |
1 Department of Radiology, Yamaguchi University School of Medicine, 1-1-1
Minamikogushi, Ube, Yamaguchi 755-8505, Japan.
2 GE Yokogawa Medical Systems, 67-4, Takakura-cho, Hachioji-shi, Tokyo 192-0033,
Japan.
Received November 13, 2003; accepted after revision April 20, 2004.
Address correspondence to K. Ito.
OBJECTIVE. The purpose of this study was to assess the value of multiarterial phase contrast-enhanced dynamic MRI of the whole liver obtained during a single breath-hold for small early enhancing hepatic lesions in patients with cirrhosis or chronic hepatitis, emphasizing the distinction between hypervascular hepatocellular carcinomas and pseudolesions.
MATERIALS AND METHODS. The study population included 40 patients with cirrhosis or chronic hepatitis who had small early enhancing hepatic lesions (a total of 70 lesions: 40 hepatocellular carcinomas, 30 pseudolesions). All patients underwent multiarterial phase contrast-enhanced dynamic MRI (six phases) of the whole liver during a single breath-hold.
RESULTS. Twenty-one (53%) of 40 hypervascular hepatocellular carcinomas showed rapid central washout after early enhancement of the lesion as well as peritumoral coronal enhancement, but these findings were not observed in any hypervascular pseudolesions (p < 0.001). In 19 hepatocellular carcinomas without rapid central washout, early enhancement of the lesion appeared at the second, third, or fourth phase (mean, 2.5 phases). In eight of these 19 hepatocellular carcinomas, lesion enhancement disappeared by the sixth phase. Conversely, in 30 hypervascular pseudolesions, early enhancement of the lesion appeared at the second, third, fourth, or fifth phase (mean, 3.0 phases). In 28 of these 30 pseudolesions, lesion enhancement continued until the sixth phase.
CONCLUSION. Rapid central washout after the early enhancement of the lesion and coronal enhancement surrounding the lesion are highly specific and diagnostic findings of small hypervascular hepatocellular carcinomas if present at multiarterial phase contrast-enhanced dynamic MRI. Hypervascular pseudolesions tend to show prolonged enhancement during the arterial phase compared with hypervascular hepatocellular carcinomas.
Multiphasic contrast-enhanced helical CT and MRI have been widely used as
accurate screening techniques for patients with cirrhosis and suspected
hepatocellular carcinomas
[1–4].
In multiphasic imaging, acquisition of three sets of images (arterial, portal,
and equilibrium phases) during separate breath-holds has been accepted as a
standard liver protocol. Recently, some investigators have reported that
double arterial phase dynamic helical CT or MRI using an MDCT scanner or
high-speed MRI system with the sensitivity encoding (SENSE) technique can
improve the detection rate of hypervascular hepatocellular carcinomas
[5,
6]. In clinical practice with
cirrhotic patients, however, we have sometimes encountered small nodular early
enhancing hepatic lesions on arterial phase contrast-enhanced dynamic CT or
MRI that resembled hepatocellular carcinomas but disappeared or decreased in
size during the clinical follow-up examinations. These lesions are considered
to be nonneoplastic hypervascular pseudolesions caused by small arterioportal
or other shunts including unknown causes and are often difficult to
differentiate from hypervascular hepatocellular carcinomas
[7–10].
Ueda et al. [11,
12] have reported that
single-level dynamic CT during hepatic arteriography, which increases the
temporal resolution of the images and allows detailed evaluation of lesion
hemodynamics, is a useful technique in the differentiation between
hypervascular hepatocellular carcinomas and pseudolesions on the basis of the
finding of coronal stain around the lesion and the required time for
visualization. Therefore, in IV contrast-enhanced dynamic MRI, multiarterial
phasic imaging with higher temporal resolution from early to late arterial
phase may be helpful for less invasively differentiating these two entities in
cirrhotic patients. The purpose of this preliminary study was to assess the
value of multiarterial phase contrast-enhanced dynamic MRI (six phases) of the
whole liver obtained during single breath-hold for small (
3 cm) early
enhancing hepatic lesions in patients with cirrhosis or chronic hepatitis,
emphasizing the distinction between hypervascular hepatocellular carcinomas
and hypervascular pseudolesions.
Materials and Methods
Patient Population
MRI records and clinical MRI requests from our institution between January
and September 2002 were searched to identify patients with cirrhosis or
chronic hepatitis who underwent multiarterial phase contrast-enhanced dynamic
MRI of the liver. Among these, the following patients were included in this
study: those who underwent 3D multiarterial phase (six phases)
contrast-enhanced dynamic MRI of the whole liver with fat suppression as a
part of our routine liver MRI protocol, those who had small ( 3 cm) early
enhancing hepatic lesions on the multiarterial phase contrast-enhanced dynamic
MR images, and those who had reliable proof of hepatocellular carcinomas or
pseudolesions. Diagnosis of hepatocellular carcinoma was established by
percutaneous liver biopsy (n = 18) or by a combination of clinical
and radiologic criteria (n = 22), including increased tumor marker
levels, response to transcatheter arterial chemoembolization (e.g., dense
iodized oil accumulation), or interval progression in size during the
follow-up examinations performed more than 6 months later. Diagnosis of
hypervascular pseudolesions was determined when the lesion disappeared or
decreased in size for 6 months or more at serial imaging studies, accompanied
by nonvisualization on other MRI sequences, or when the corresponding lesion
was not detected on ferumoxides-enhanced MR images obtained during the
follow-up period. Hypovascular hepatocellular carcinomas that did not show
early enhancement during the multiarterial phase were not included in this
study even though pathologic confirmation of the lesion was obtained.
The following patients were excluded from this study: those who had early enhancing hepatic lesions without confirmation of hepatocellular carcinoma and with no follow-up imaging studies available; those who had early enhancing hepatic lesions that showed no change in size at serial follow-up imaging studies because differentiation between hypervascular pseudolesions and slow-growing hepatocellular carcinomas is difficult; and those who underwent therapeutic procedures for hepatocellular carcinomas, including radiofrequency ablation therapy, percutaneous alcohol injection therapy, or transcatheter chemoembolization within 2 months before the MRI examination because these procedures may affect imaging findings.
One radiologist experienced in abdominal MRI who served as a study coordinator reviewed clinical, radiologic, and pathologic reports and follow-up imaging studies and identified 40 patients (17 women, 23 men; age range, 51–75 years; mean age, 62 years) who satisfied the inclusion criteria. These patients had been referred for MRI for several reasons (e.g., to evaluate the severity of cirrhosis and portal hypertension, for the screening of hepatic lesions that were suspected from findings on other imaging techniques, or for the follow-up of previously suspected or treated hepatocellular carcinomas).
MRI Techniques
MRI was performed on a 1.5-T system (Signa CVi, GE Healthcare) using a
phased-array torso coil. As part of routine liver MRI examination,
multiarterial phase (six phases) contrast-enhanced dynamic MRI of the whole
liver was performed in all patients using 3D enhanced fast spoiled
gradient-echo sequence with fat suppression. The imaging parameters were as
follows: TR range/TE, 3–4/1.1; inversion time, 15 msec; flip angle,
15°; matrix, 256 x 128; section thickness, 10 mm; receiver
bandwidth, 125 kHz; and number of signals acquired, one half. A 0.55
rectangular field of view was used to reduce the number of phase-encoding
views. The k-space was zero-filled in the section-encoding direction to
decrease the increment between sections to 5 mm and to double the number of
reconstructed imaging sections. As a result, a total of 40 sections covering
the whole liver were acquired during 5 sec. Imaging was repeated six times
during a single breath-hold (
32 sec) for multiarterial phase scanning.
After obtaining unenhanced images, multiarterial phase (six phases)
contrast-enhanced dynamic MRI of the entire liver was performed after the IV
administration of 0.1 mmol/kg of gadopentetate dimeglumine followed by a 20-mL
bolus of normal saline by means of a power injector at a rate of 3 mL/sec
through a 22-gauge plastic IV catheter (patient's body weight range,
42–76 kg; mean body weight, 57.4 kg). The scanning delay was 10 sec
after the beginning of the contrast material injection for multiarterial phase
imaging. Although the 10-sec delay is earlier than usually advocated, we
intended to use first phase images as masked images for subtraction. (The
value of subtraction images is not assessed in this study.) A test bolus
injection or a bolus tracking technique was not used in this study because
multiarterial phase imaging was expected to obtain optimal arterial phase
scanning without such techniques which may be too troublesome to always
perform in routine clinical work. Patients who were not expected to suspend
respiration for 32 sec and underwent reduced multiarterial phase (4–5
phases) dynamic imaging were excluded from this study (n = 9).
Parallel imaging methods such as SENSE were not applied in this study. The
other pulse sequences obtained during routine liver MRI included T2-weighted
fast spin-echo sequence with fat suppression (TR/TE, 2,500/74), T2-weighted
single-shot fast spin-echo sequence (infinite/67), T1-weighted opposed-phase
(150/2.2) and in-phase (150/4.8) gradient-echo sequences, and 3D enhanced fast
spoiled gradient-echo sequence with fat suppression for unenhanced images and
portal and equilibrium phase contrast-enhanced dynamic MR images. All imaging
was performed during a single breath-hold.
Image Interpretation
All examinations were reviewed retrospectively by two radiologists with
more than 10 years of experience in liver MRI without knowledge of the final
diagnosis; any interpretation discrepancies were resolved by consensus with
the participation of a third radiologist with 15 years' experience in liver
MRI. The reviewers knew that patients had cirrhosis or chronic hepatitis and
were at risk for hepatocellular carcinoma but did not know how many, if any,
hepatocellular carcinoma lesions were present in any patient. MR images were
evaluated for the presence of small early enhancing hepatic lesions ( 3
cm) suggestive of hypervascular hepato-cellular carcinoma or a hypervascular
pseudolesion on the multiarterial phase contrast-enhanced dynamic MR images.
The size and shape of early enhancing lesions were recorded. The shape of the
lesion was categorized as one of three types: round or oval, wedge-shaped, or
triangular. Hemodynamic changes (contrast enhancement pattern) of the lesion
through six arterial phases were evaluated. The focus was on the finding of
rapid central washout after early enhancement of the lesion and the presence
of coronal enhancement surrounding the lesion; these findings may correspond
to the appearances of hepatocellular carcinomas in the washout phase with
coronal stain observed at single-level dynamic CT during hepatic arteriography
[11,
12]. The reviewers also
recorded the following: in which phase aortic enhancement appeared, in which
phase the early enhancing lesion appeared and how long lesion enhancement
continued, and in which phase central washout of the lesion with coronal
enhancement was observed, if present. MR images were reviewed on a clinical
workstation (imageVINS, Yokogawa Electric). On this system, MR images of the
same slice level from each arterial phase (total six phases) can be displayed
simultaneously on the monitor using a multiwindow frame setting and always
reviewed on the same slice position with the paging mode. Thus, hemodynamic
change of the lesion through six arterial phases at the same level can be
readily analyzed. The chi-square test was used to compare the difference of
the frequency of rapid central washout with coronal enhancement between
hypervascular hepatocellular carcinomas and pseudolesions. Other statistical
analyses were performed by using the unpaired Student's t test.
Results
A total of 70 (40 hypervascular hepatocellular carcinomas, 30 hypervascular
pseudolesions) small ( 3 cm) early enhancing hepatic lesions were detected
in 40 patients during the multiarterial phase of contrast-enhanced dynamic
MRI. The shapes of the lesions were as follows: round or oval in 66 (94%),
wedge-shaped in three (5%), and triangular in one (1%). The average size of
the lesions was 11.3 mm, ranging from 3 to 30 mm. The mean diameter of
hypervascular hepatocellular carcinomas and hypervascular pseudolesions was
13.3 and 8.6 mm, respectively. Twenty-one (53%) (average size, 14.5 mm) of 40
hypervascular hepato-cellular carcinomas showed rapid central washout after
the early enhancement of the lesion as well as peritumoral coronal enhancement
(Figs. 1A,
1B,
1C,
1D,
1E,
1F and
2A,
2B,
2C,
2D,
2E,
2F), although these findings
were not observed in any hypervascular pseudolesions (p < 0.001).
Central washout of the lesions with coronal enhancement appeared at the third
(n = 4), fourth (n = 4), fifth (n = 7), or sixth
(n = 6) phase (mean, 4.7 phases). Enhancement of the aorta appeared
at the first or second phase (mean, 1.7 phases for hepatocellular carcinomas
and 1.7 phases for pseudolesion).
|
|
|
|
|
|
|
|
|
|
|
|
In 19 hepatocellular carcinomas (average size, 11.9 mm) without rapid central washout, early enhancement of the lesion appeared at the second (n = 11), third (n = 6), or fourth (n = 2) phase (mean, 2.5 phases). Conversely, in 30 hypervascular pseudolesions, early enhancement of the lesion appeared at the second (n = 7), third (n = 16), fourth (n = 6), or fifth (n = 1) phase (mean, 3.0 phases). This difference was statistically significant (p < 0.02). In 19 hepatocellular carcinomas without rapid central washout, lesion enhancement continued until the third phase in four, the fourth phase in two, the fifth phase in two, and the sixth phase in 11 (mean, 5.1 phase). Conversely, in 30 hypervascular pseudolesions, lesion enhancement continued until the third phase in one, the fifth phase in one, and the sixth phase in 28 (mean, 5.9 phases). This difference was also statistically significant (p < 0.004). These results indicated that lesion enhancement of hypervascular pseudolesions tended to appear later and last longer than that of hepato-cellular carcinomas without rapid central washout (Figs. 3A, 3B, 3C, 3D, 3E, 3F, 4A, 4B, 4C, 4D, 4E, 4F and 5A, 5B, 5C, 5D, 5E, 5F), although considerable overlap between the two groups existed.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Discussion
Small, hypervascular hepatocellular carcinomas can be detected as early enhancing nodules during the arterial phase of contrast-enhanced dynamic MRI. In clinical practice, however, we have frequently encountered small early enhancing hepatic lesions on arterial-phase contrast-enhanced MR images that resemble hypervascular hepatocellular carcinomas but disappear or decrease in size on serial MRI studies and are considered to be hypervascular pseudolesions [13]. These lesions are being detected with increased frequency and often cause a diagnostic dilemma, particularly when small lesions show round or oval configuration in patients with cirrhosis or chronic hepatitis. Ueda et al. [11, 12] reported that hypervascular hepatocellular carcinomas can be differentiated from hypervascular pseudolesions by using single-level dynamic CT during hepatic arteriography on the basis of coronal enhancement surrounding the lesion, which was the specific finding for hypervascular hepatocellular carcinomas. However, disadvantages of single-level dynamic CT during hepatic arteriography include invasiveness and expensiveness, difficulty in visualizing a small lesion due to possible respiratory misregistration, and difficulty in evaluating the hemodynamics of multiple hepatic lesions.
In our study, rapid central washout after the early enhancement of the
lesion and peritumoral coronal enhancement were observed in 53% of
hypervascular hepatocellular carcinomas, although these findings were not seen
in any hypervascular pseudolesions. This fact suggests that rapid central
washout of the lesion with peritumoral coronal enhancement at multiarterial
phase contrast-enhanced dynamic MRI is a highly specific sign for
hypervascular hepatocellular carcinomas corresponding to the coronal stain in
the washout phase at single-level dynamic CT during hepatic arteriography.
Visualization of these findings on multiarterial phase contrast-enhanced
dynamic MRI may be attributed to the bolus effect of a small amount (0.2
mL/kg) of contrast material as well as the high temporal resolution during
arterial phase scanning. The combination of the bolus injection of a small
amount of contrast material and shorter imaging time (5 sec) with sequential
scanning (six times) can capture rapid wash-in and washout of contrast
material to the lesion and enables us to visualize early enhancement of the
lesion, rapid central washout of the lesion, and the subsequent coronal
enhancement surrounding the lesion. Conversely, in single arterial phase
dynamic MRI with longer data acquisition time ( 20 sec), staining of the
lesion may be integrated with washout of the contrast material to the
surrounding liver, masking the rapid central washout of the lesion with
coronal enhancement as a specific finding.
The ability of MDCT to scan the entire liver in less than 10 sec may allow
acquisition of four to five separate sets of arterial phase CT images of the
liver during a single breath-hold, similar to multiarterial phase
contrast-enhanced dynamic MRI. However, excessive radiation exposure from
multiarterial phase imaging with MDCT will be problematic in a clinical
practice. Additionally, the use of a large amount of contrast material (
2.0 mL/kg) at CT [14] requires
a longer injection time even though the contrast medium is injected at a high
flow rate of 5 mL/sec (e.g., 24-sec injection duration for a patient weighing
60 kg). In this situation, it is likely that continuous staining of the lesion
may be integrated with washout of contrast material from the lesion because of
the longer injection duration, masking the central washout with coronal
enhancement even though the scanning time of each phase is short.
Regarding 47% of hepatocellular carcinomas without the finding of rapid central washout, the enhancement pattern of these hepatocellular carcinomas was similar to that of hypervascular pseudolesions. However, lesion enhancement of hypervascular pseudolesions tended to appear later and last longer than that of hepatocellular carcinomas without rapid central washout with statistical significance, although these findings cannot be used to make a definitive diagnosis because of the substantial overlap. This finding may be caused by the difference in blood supply to the lesion between hypervascular hepatocellular carcinomas and hypervascular pseudolesions. Blood supply to hypervascular hepatocellular carcinomas is mainly via the hepatic artery, although hypervascular pseudolesions are presumably supplied by blood passing through arterioportal communications such as the peribiliary plexus and flowing into hepatic sinusoids via the portal venules [15].
We used fixed scanning delays in our multiarterial phase dynamic MRI protocol without a bolus tracking technique. Our results showed that enhancement of the aorta appeared at the first or second phase in all our patients, indicating the small variation in the circulation time of patients in this study. However, for meticulous comparison, it will be necessary to analyze only the patient group with aortic enhancement at the first phase examined by using a timing bolus sequence. The use of a bolus tracking technique may allow a more precise evaluation of lesion hemodynamics even in patients with poor cardiac function. With our current technique, we use a 10-mm section thickness with a 5-mm overlap and a relatively low spatial resolution of 256 x 128 matrix to shorten the imaging time. The combined use of the SENSE technique will improve spatial resolution, allow multiarterial phase imaging with higher temporal resolution and thinner section thickness, and improve the sensitivity of findings (rapid central washout and coronal enhancement) suggestive of hypervascular hepatocellular carcinomas.
In this study, we did not evaluate whether multiarterial phase imaging can improve the detection rate of hypervascular hepatocellular carcinomas and can be useful for the diagnosis of hypovascular hepatocellular carcinomas. Additionally, the role of the portal and equilibrium phases or T2-weighted sequences for the distinction of hypervascular hepatocellular carcinomas and pseudolesions was not profoundly evaluated. Our study was focused on determining the value of multiarterial phase contrast-enhanced dynamic MRI of the liver for differentiating hypervascular hepatocellular carcinomas and pseudolesions on the basis of their temporal hemodynamic changes. Although the addition of equilibrium phase imaging may contribute to improving the detection of hypovascular hepatocellular carcinomas [16–18], further evaluation is necessary for the topics described.
This study is limited by the fact that pseudolesions were determined on the basis of follow-up MRI findings, not by means of histologic confirmation. However, pathologic proof of small hypervascular pseudolesions presumably due to an arterioportal shunt would not be practical in the clinical setting. On the basis of the lesion or decreasing in size at follow-up, we believe that it is unlikely that hypervascular hepatocellular carcinomas were incorrectly categorized as pseudolesions, although the possibility of a slow-growing hepatocellular carcinoma cannot completely be excluded. Additionally, we used several exclusion criteria in this study, preventing adequate evaluation of MRI findings for a subset of hepatocellular carcinomas including slow-growing hepatocellular carcinomas. A future study of these lesions will be necessary.
In conclusion, multiarterial phase contrast-enhanced dynamic MRI of the whole liver has a potential to evaluate transitional hemodynamics of hepatic lesions during the six arterial phases because of excellent temporal resolution. Rapid central washout after the early enhancement of the lesion and coronal enhancement surrounding the lesion are highly specific and diagnostic findings that distinguish small hypervascular hepatocellular carcinomas and hypervascular pseudolesions if present. Hypervascular pseudolesions tend to show prolonged enhancement during the arterial phases compared with hypervascular hepatocellular carcinomas.
References
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. H. Choi, J. M. Lee, N. C. Yu, K.-S. Suh, J.-J. Jang, S. H. Kim, and B. I. Choi Hepatocellular Carcinoma in Liver Transplantation Candidates: Detection with Gadobenate Dimeglumine-Enhanced MRI Am. J. Roentgenol., August 1, 2008; 191(2): 529 - 536. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R Moucari, P-E Rautou, D Cazals-Hatem, A Geara, C Bureau, Y Consigny, C Francoz, M-H Denninger, V Vilgrain, J Belghiti, et al. Hepatocellular carcinoma in Budd-Chiari syndrome: characteristics and risk factors Gut, June 1, 2008; 57(6): 828 - 835. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Willatt, H. K. Hussain, S. Adusumilli, and J. A. Marrero MR Imaging of Hepatocellular Carcinoma in the Cirrhotic Liver: Challenges and Controversies Radiology, May 1, 2008; 247(2): 311 - 330. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Choi, D. G. Mitchell, S. K. Verma, D. Bergin, V. J. Navarro, A. B. Malliah, C. McGowan, H.-W. L. Hann, and S. K. Herrine Hepatocellular Carcinoma with Indeterminate or False-Negative Findings at Initial MR Imaging: Effect on Eligibility for Curative Treatment Initial Observations Radiology, September 1, 2007; 244(3): 776 - 783. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||
