| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
-Fetoprotein Levels
1 Department of Clinical Oncology, The Chinese University of Hong Kong, Hong
Kong, China.
2 Department of Diagnostic Radiology & Organ Imaging, Prince of Wales
Hospital, The Chinese University of Hong Kong, 30-32 Ngan Shing St., Sha Tin,
N.T., Hong Kong, China.
3 Department of Medicine and Therapeutics, The Chinese University of Hong Kong,
Hong Kong, China.
4 Department of Surgery, The Chinese University of Hong Kong, Hong Kong,
China.
Received August 15, 2003;
accepted after revision February 19, 2004.
Address correspondence to S. C. H. Yu
(simonyu{at}cuhk.edu.hk).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-fetoprotein levels is widely practiced. Negative
results on an abdominal sonogram generally indicate the absence of
hepatocellular carcinoma despite the elevation of -fetoprotein levels,
but the false-negative rate of abdominal sonography has not been established
prospectively.
SUBJECTS AND METHODS. In our screening program, we routinely
investigated patients with Lipiodol (iodized oil) CT when they presented with
-fetoprotein levels above 20 ng/mL or a focal lesion as depicted on
abdominal sonography. Lipiodol CT comprised a hepatic angiogram with injection
of Lipiodol selectively in the hepatic arteries, followed by an unenhanced CT
scan 10 days later. Positive findings on Lipiodol CT were confirmed
histologically by biopsy or surgical resection. We defined false-negative as
histologic diagnosis of hepatocellular carcinoma within 3 months of normal
findings on screening abdominal sonography.
RESULTS. One hundred three patients with elevated
-fetoprotein levels were investigated with Lipiodol CT within 2 months
of abdominal sonography. Of these, three of 70 patients with negative
abdominal sonography had histologically confirmed hepatocellular carcinoma.
Thus, abdominal sonography has a false-negative rate of 4.3%. Lipiodol CT is
associated with a significant false-positive rate of 43.7%. The sensitivity,
specificity, and positive predictive value of abdominal sonography for early
detection of hepatocellular carcinoma among hepatitis B virus carriers with
elevated -fetoprotein levels was 85.7%, 81.7%, and 54.5%,
respectively.
CONCLUSION. Negative results on a screening abdominal sonogram among
hepatitis B virus carriers with elevated -fetoprotein levels does not
rule out the presence of small hepatocellular carcinoma. Routine use of
Lipiodol CT as a supplementary screening tool is not recommended.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-fetoprotein levels
and abdominal sonography for screening patients with cirrhosis for
hepatocellular carcinoma in spite of the controversy that surrounds these
methods.
Abdominal sonography is considered to be the standard screening test and is
capable of detecting small hepatocellular carcinomas (< 3 cm in diameter).
The absence of lesions on abdominal sonograms is commonly assumed to imply the
absence of hepatocellular carcinoma at the time of screening
[2,
3]. However, limited
information is available on the false-negative rate of abdominal sonography in
hepatocellular carcinoma screening. To establish the false-negative rate of
abdominal sonography, researchers would have to investigate the screened
population with negative abdominal sonograms by an additional diagnostic
imaging technique such as hepatic angiography, CT, or MRI for visualization of
a lesion that can be subsequently confirmed at histology. Alternatively,
researchers could follow the patients with negative abdominal sonograms
closely with repeated abdominal sonography until a tumor becomes detectable.
Using the latter method, Maringhini et al.
[4] reported that 14 (6%) of
the 255 cirrhotic patients with negative results on initial screening
abdominal sonograms and an equivocal -fetoprotein level developed
hepatocellular carcinoma during the follow-up period of 4 years. In the same
study, the false-negative rate for detection of smaller tumors (< 5 cm) was
8.9%. We cannot determine, however, if the hepatocellular carcinomas were
present at the time of the initial negative abdominal sonograms. No
prospective study currently exists that determines the diagnostic utility of
screening abdominal sonography by additional imaging and histologic
confirmation.
Hepatitis B virus is the primary cause of hepatocellular carcinoma in China
and Southeast Asia [5,
6]. The incidence of
hepatocellular carcinoma in Hong Kong is 34.4 per 100,000, and more than 80%
of patients with hepatocellular carcinoma are chronic carriers of the
hepatitis B virus [7].
Screening the target population of hepatitis B virus carriers may help to
identify asymptomatic patients with small tumors and, therefore, improve their
chances of receiving curative treatment
[8]. In 1997, we initiated a
prospective screening study with serum -fetoprotein levels, abdominal
sonography, and hepatic angiography with Lipiodol (iodized oil, Guerbet) CT.
We termed a hepatic angiogram with selective injection of Lipiodol in the
hepatic arteries, followed by an unenhanced CT scan 10 days later,
"Lipiodol CT." The gold standard for the diagnosis of
hepatocellular carcinoma is histology. We defined false-negative as histologic
diagnosis of hepatocellular carcinoma within 3 months of normal results on a
screening abdominal sonogram. On the basis of the data from this project, we
have prospectively defined the diagnostic utility of screening abdominal
sonography in hepatitis B virus carriers with elevated serum
-fetoprotein levels.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-fetoprotein levels who had at least one abdominal
sonogram and had undergone hepatic angiography with a post-Lipiodol CT scan
(Lipiodol CT) within 2 months of the abdominal sonogram. Recruitment for the
study was completed in March 2000 when the predetermined sample size of 1,018
subjects had been attained.
Screening Scheme
All enrolled patients had the following routine evaluations: medical
history, physical examination, complete blood counts, electrolytes,
creatinine, alkaline phosphatase, alanine transaminase, total bilirubin, total
protein, albumin, hepatitis B virus surface antigen, and hepatitis E antigen
and its antibody. The screening protocol is summarized in
Figure 1. Patients with serum
-fetoprotein levels above 20 ng/mL on two occasions at least 1 week
apart or focal lesions on abdominal sonograms were further evaluated with
Lipiodol CT. Tissue for histologic assessment was obtained from hypervascular
tumors either at surgery or percutaneous needle biopsy. Patients who declined
surgery or histologic assessment by biopsy were followed every 3 months with
repeated -fetoprotein level measurements and abdominal sonography. We
also followed patients with elevated -fetoprotein levels or focal
lesions but normal findings on Lipiodol CT every 3 months with repeated
-fetoprotein level measurements and abdominal sonography for 2 years
and then every 6 months thereafter. The repeated -fetoprotein level
measurements were always performed within 1-2 weeks after repeated abdominal
sonography. Any new lesion detected on follow-up abdominal sonography was
further investigated with repeated Lipiodol CT and, if indicated, histologic
confirmation. If both the enrollment -fetoprotein levels and abdominal
sonogram were normal, only serum -fetoprotein levels would be followed
up every 3 months. Elevated -fetoprotein level during follow-up was an
indication for repeated abdominal sonography.
|
Screening Tests
Alpha-fetoprotein level was measured by enzyme immunoassay method (MEIA,
Abbott Laboratories). Two consecutive serum levels above 20 ng/dL repeated at
least 1 week apart constituted a positive result. The selection of this cutoff
level was based on the reported sensitivity of -fetoprotein levels in
the two major screening studies of hepatitis B carriers
[9,
10]. Abdominal sonography was
performed by one of two designated radiologists, each of whom had more than 10
years of experience, using an electronic curvilinear 3.5-MHz real-time
transducer, scanning subcostally and intercostally with the patient in a
supine, and then left decubitus, position. A focal lesion was defined as a
well-defined solid nodule (mass) with hypoechoic, hyperechoic, or mixed
sonographic pattern. Results were categorized as positive, probable, or
negative. Positive abdominal sonograms meant definite presence of a focal
lesion, although probable abdominal sonograms implied the presence of an
atypical focal lesion. Negative abdominal sonograms meant that no solid lesion
could be detected.
Lipiodol CT comprised a hepatic angiogram with injection of Lipiodol
selectively administered in the hepatic arteries, followed by an unenhanced CT
scan 10 days later. Hepatic angiography involved access to the arterial system
via the femoral artery using the Seldinger technique. Celiac axis, superior
mesenteric artery, and selective hepatic artery angiograms were obtained as
contrast material was injected in the respective arteries. Angiographic images
were examined for evidence of hypervascular areas in the liver that would
suggest hepatocellular carcinoma. Lipiodol, an oily contrast medium that tends
to be selectively accumulated and retained by tumor tissue, was injected in
the hepatic arteries during the procedure. CT was performed on day 10 after
arterial Lipiodol administration to detect any Lipiodol retention. A positive
result on Lipiodol CT was defined as the presence of a hypervascular lesion
with corresponding Lipiodol retention. The result was considered nonspecific
if only one of the features was present. To ensure the consistency of the
procedure, a designated radiologist at the department of diagnostic imaging at
the Prince of Wales Hospital performed all the Lipiodol CT studies. A lesion
detected on Lipiodol CT that was confirmed to be nonmalignant on biopsy was
considered to be false-positive. Patients with negative results on Lipiodol CT
who subsequently developed histologically confirmed hepatocellular carcinoma
within 3 months or who died in the presence of a space-occupying lesion in the
liver associated with an -fetoprotein level above 1,000 ng/mL within 12
months of the procedure were considered to have a false-negative result.
Surgery and Histologic Confirmation of Hepatocellular Carcinoma
The hepatobiliary surgery team assessed all patients with positive results
on Lipiodol CT and decided on the likelihood of resectability. A radiologist
performed liver biopsy for all patients with positive Lipiodol CT but whose
lesions were deemed inoperable or unresectable by the surgeons. An 18-gauge
needle was used to obtain a sample from focal lesions under sonographic
guidance. Liver biopsy was performed within 3 weeks of Lipiodol CT while the
Lipiodol-retaining lesion could still be visualized on sonography. Patients
who declined liver biopsy were observed every 3 months with repeated
-fetoprotein levels and abdominal sonography. Patients who initially
declined were still offered liver biopsy if they subsequently changed their
minds when the lesion became visible on abdominal sonography. All biopsy
samples were examined by an experienced liver histopathologist, and histologic
confirmation of hepatocellular carcinoma was based on the microscopic features
of the biopsy tissue stained by H and E.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-fetoprotein levels and completion of Lipiodol CT within 2
months of abdominal sonography. Sixteen patients had positive, 17 had
probable, and 70 had negative abdominal sonographic examinations. The
characteristics of these patients are summarized in
Table 1. The median duration of
follow-up of all patients with elevated serum -fetoprotein levels who
had completed Lipiodol CT was 28 months.
|
Patients with Focal Lesion on Abdominal Sonography
We detected hypervascular lesions on hepatic angiography or
Lipiodol-retaining lesions on Lipiodol CT in 13 of 16 patients with positive
results on abdominal sonograms. Liver biopsy or surgical pathology results
confirmed hepatocellular carcinoma in 10 patients, and the histology results
from the other three patients showed hepatic adenoma or regenerative nodules.
None of the patients with negative Lipiodol CT had hepatocellular carcinoma.
Four of the 17 patients with probable focal lesions on abdominal sonography
had hypervascular lesions on hepatic angiograms that were confirmed to be
hepatocellular carcinoma. Among the 13 remaining patients who had a negative
Lipiodol CT initially and were followed with serial abdominal sonography and
-fetoprotein levels every 3 months, four developed hepatocellular
carcinoma. Histologic confirmation of hepatocellular carcinoma was obtained in
all patients with focal lesions on abdominal sonography with the exception of
three patients who declined liver biopsy. These three patients died from
enlarging tumor masses associated with rising -fetoprotein levels above
1,000 ng/mL. The period for these three patients from the initial abnormal
findings on abdominal sonography to death (thus confirmation of diagnosis of
hepatocellular carcinoma) was 12, 32, and 35 months, respectively. If we
consider a probable focal lesion to be a positive finding, the sensitivity,
specificity, and positive predictive value of abdominal sonography for early
detection of hepatocellular carcinoma among carriers of the hepatitis B virus
surface antigen with elevated -fetoprotein levels were 85.7%, 81.7%,
and 54.5%, respectively. Sensitivity, specificity, and positive predictive
value of Lipiodol CT in these 33 patients with positive or probable focal
lesions on abdominal sonograms were 77.8%, 80.0%, and 82.3%, respectively.
Patients with Negative Abdominal Sonography
The clinical outcomes of this group are summarized in
Figure 2. We confirmed
hepatocellular carcinoma by histology within 3 months from a negative initial
screening abdominal sonogram in three of the 70 carriers of the hepatitis B
virus surface antigen with elevated -fetoprotein levels. Thus, the
prospective false-negative rate was 4.3%. Benign histology was found in seven
patients with positive Lipiodol CT including five regenerative nodules, one
adenoma, and one hemangioma. Four patients had nonspecific findings on
Lipiodol CT, and none of these patients developed hepatocellular carcinoma.
Sensitivity, specificity, and positive predictive value of Lipiodol CT in
carriers of the hepatitis B virus surface antigen with elevated
-fetoprotein levels and negative abdominal sonograms were 100%, 83.4%,
and 21.4%, respectively.
|
Confirmed Hepatocellular Carcinoma
We have prospectively detected and confirmed hepatocellular carcinoma in 18
patients with focal lesions and three patients without focal lesions on
initial screening abdominal sonography. The mean maximum diameter of the
tumors detectable on abdominal sonography was 3.0 cm (range, 1.2-6.0 cm)
compared to 2.2 cm (range, 1.3-4.0 cm) in those not detected on abdominal
sonography. The ratio of solitary to multifocal lesions was 13:5 for patients
with sonographically detected tumors and 2:1 for patients with negative
abdominal sonograms. Resection of solitary lesions was successful in three
patients, and one patient died from liver failure in the postoperative period.
Surgery was not feasible for the remaining patients because of multifocal
disease (n = 6), poor liver function (n = 10), or patient
refusal (n = 2). Nonsurgical treatments were offered according to the
clinical judgment of the patient's oncologist. One patient with multifocal
disease received a course of combination chemotherapy and attained partial
response. The patient's disease remained stable for 18 months before one of
the lesions started to progress. Wedge excision of the lesion was performed,
and the patient remains well and disease-free. Other patients with inoperable
disease had progressive disease or died from hepatocellular carcinoma.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-fetoprotein levels. Several major prospective screening studies have
suggested that some hepatocellular carcinomas may not be detectable on
screening abdominal sonography
[10-12].
Only 58 (86.6%) of the 67 patients had confirmed hepatocellular carcinoma from
the screening series by Izzo et al.
[12] detectable on abdominal
sonography. The remaining patients were diagnosed on CT or MRI. Sherman et al.
[9] reported high sensitivity
associated with initial screening abdominal sonography, but three of 11
confirmed hepatocellular carcinomas at follow-up were not detectable on
abdominal sonography. Colombo et al.
[13] found 29 hepatocellular
carcinomas during follow-up screening of patients with cirrhosis. Three tumors
(10.3%) were diagnosed by methods other than abdominal sonography. In this
series of 103 patients with elevated -fetoprotein levels who were
evaluated using Lipiodol CT, we confirmed hepatocellular carcinoma in 21
patients and three (14.3%) did not have a detectable focal lesion on abdominal
sonography at the time of diagnosis.
The sensitivity of serum -fetoprotein levels in hepatocellular
carcinoma screening was reported to be 48.6% in a study that adopted a high
cutoff level above 500 ng/mL (reference range, < 10 ng/mL)
[4]. With a lower cutoff level
of 15-20 ng/mL in different studies in the sensitivity of -fetoprotein,
levels ranged from 64.3% to 96.9%
[9,
10,
14]. The false-negative rates
for -fetoprotein levels in these studies were commonly defined by the
number of cases with serum -fetoprotein levels below the cutoff value
in the presence of lesions on abdominal sonograms that were subsequently
confirmed to be hepatocellular carcinoma at histology. Alternatively, the gold
standard for hepatocellular carcinoma diagnosis was defined by other
diagnostic imaging techniques or subsequent evidence of tumor progression and
death.
Patients with elevated serum -fetoprotein levels may have small
hepatocellular carcinomas that are not detectable on abdominal sonography.
Colombo et al. [13] reported
the relative risk of hepatocellular carcinoma among patients with cirrhosis
who had persistently elevated serum -fetoprotein levels (> 20 ng/mL)
to be 14 times higher than patients with a normal level. A substantial portion
(29%) of patients with persistently elevated serum -fetoprotein levels
and negative abdominal sonograms at their initial screening developed
hepatocellular carcinoma during follow-up. In our previous study on the use of
hepatoma-specific -fetoprotein levels in the early detection of
hepatocellular carcinoma, we reported 15 patients with elevated
-fetoprotein levels (> 50 ng/mL) before evidence of tumor on
abdominal sonography [11]. The
elapsed time ranged from 1 to 18 months (median, 3.6 months) and the median
tumor size was 6 cm. Given the size of the tumors, it is most likely that
abdominal sonography had failed to detect the lesions at the time of the
initial elevation of -fetoprotein levels.
For the purpose of the present study, we deliberately selected a group of
patients with elevated serum -fetoprotein levels, so the proportion of
tumor-free patients included in this study population would not be too low. We
believe that such a selection scheme did not lead to any bias.
The sensitivity of abdominal sonography is operator-dependent, but accuracy
is consistent in experienced hands. The two radiologists in this screening
study were experienced members of our joint hepatoma clinic who had each
obtained more than 1,000 abdominal sonograms and accumulated over 10 years of
experience. It is probable that the smaller tumors were truly undetectable on
abdominal sonography. Ebara et al.
[15] followed the natural
history of 22 small hepatocellular carcinomas without specific treatment and
described the changes in the sonographic patterns of the tumors with repeated
abdominal sonography (mean interval of 2.5 months). At the time tumor size was
increasing, some of the nodules changed their sonographic pattern from
hypoechoic to isoechoic with a hypoechoic rim. Unless physicians had prior
knowledge of their location, these isoechoic lesions would have been difficult
to detect. The mean diameter of the tumors at the isoechoic stage was 2.5 cm,
which is compatible with the size of the tumors detected in our study. The
tumors then began to exhibit hyperechoic or mixed patterns on abdominal
sonography when the mean diameter reached 3 cm and became detectable. During
this critical period of progression of tumor size from 2 to 3 cm, it may not
be the size but rather the sonographic pattern that makes the lesions
apparent.
|
|
|
We chose Lipiodol CT as the alternative imaging technique because of its reported high accuracy from earlier studies [16, 17]. This technique depends on the hypervascularity of hepatocellular carcinoma and is independent of tissue density. Size and vascularity are the main determining factors for visualization of the tumor. However, more recent reports have shown limitations of Lipiodol CT in detection of small hepatocellular carcinomas. Bizollon et al. [18] studied the diagnostic utility of Lipiodol CT by comparing the test results of 72 cirrhotic patients awaiting liver transplantation with their respective explanted livers. Lipiodol CT was able to detect only six of the 24 small hepatocellular carcinomas from 14 patients. In addition, three of the nine abnormal hypervascular lesions on Lipiodol CT were shown to be benign (one hemangioma and two regenerative nodules). In our study of 70 patients with negative abdominal sonography, only three of the 10 patients with hypervascular lesions on Lipiodol CT had hepatocellular carcinoma confirmed at biopsy. The other seven patients had false-positive findings that were confirmed to be regenerative nodules, adenomas, or hemangiomas on biopsy. The positive predictive value of Lipiodol CT in patients with negative abdominal sonograms was only 30%. Lipiodol CT did not appear to be superior to screening abdominal sonography for early diagnosis of hepatocellular carcinoma (Fig. 3). If we consider the probable lesions to represent positive results, the sensitivity, specificity, and positive predictive value of Lipiodol CT as a screening tool for the 103 patients was 85.7%, 82.9%, and 56.3%, respectively, which is not significantly different from that of screening abdominal sonography (85.7%, 81.7%, and 54.5%, respectively). Therefore, the role of Lipiodol CT as a screening tool supplementary to abdominal sonography is limited.
|
A major limitation in the methodology of this study was the absence of a reliable gold standard that could provide the true false-negative rate, such as systematic sectioning of explanted livers. In our study, histologic diagnosis of hepatocellular carcinoma within 3 months was adopted as the gold standard. A period of 3 months could be a source of error because new tumor lesions may arise in this interval.
Another limitation in this study was the use of Lipiodol CT as the alternative imaging technique. Lipiodol CT relies on hypervascularity to detect hepatocellular carcinoma and, therefore, hypovascular tumors, which are rare among grade I hepatocellular carcinomas, may be missed. This study may potentially be improved if CT and MRI are used as well. However, we believe that such a limitation of Lipiodol CT did not significantly affect the outcome of this study because Lipiodol CT was not used as a gold standard.
In summary, our study has prospectively confirmed the false-negative rate
of abdominal sonography among carriers of the hepatitis B virus surface
antigen who have elevated -fetoprotein levels to be 4.3%. A negative
abdominal sonogram does not rule out the presence of a small hepatocellular
carcinoma that may be detectable on Lipiodol CT. However, the routine use of
Lipiodol CT as a screening tool is not recommended.
Acknowledgments
We thank the Hong Kong Cancer Fund for supporting this screening
project.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. M. Bittle and F. S. Chew Imaging of Hepatocellular Carcinoma: Self-Assessment Module Am. J. Roentgenol., June 1, 2006; 186(6_Supplement_1): S431 - S433. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Bittle and F. S. Chew Radiological Reasoning: Incidentally Discovered Liver Mass Am. J. Roentgenol., June 1, 2006; 186(6_Supplement_1): S434 - S441. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. S.K. Mok, W. Yeo, S. Yu, P. Lai, H. L.Y. Chan, A. T.C. Chan, J. W.Y. Lau, H. Wong, N. Leung, E. P. Hui, et al. An Intensive Surveillance Program Detected a High Incidence of Hepatocellular Carcinoma Among Hepatitis B Virus Carriers With Abnormal Alpha-Fetoprotein Levels or Abdominal Ultrasonography Results J. Clin. Oncol., November 1, 2005; 23(31): 8041 - 8047. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
