| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Abdominal Radiology, New York University Medical Center, 560
First Ave., New York, NY 10016.
2 Department of Pathology, New York University Medical Center, New York, NY
10016.
3 Department of Transplant Surgery, New York University Medical Center, New
York, NY 10016.
Received November 1, 2001;
accepted after revision January 8, 2002.
Presented at the annual meeting of the American Roentgen Ray Society,
Atlanta, April—May 2002.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. The sonography reports of 200 patients with cirrhosis who underwent sonography and then underwent liver transplantation within 90 days were retrospectively reviewed for focal solid liver lesions. All focal solid masses detected on sonography were considered possible hepatocellular carcinomas. The sonographic findings were compared with thin-section explanted liver pathologic results.
RESULTS. Twenty-seven patients (13.5%) had hepatocellular carcinoma at explantation, including four patients with diffuse, multifocal tumors. Eight of the 39 lesions were detected on sonography for a patient sensitivity of 29.6% and a lesion sensitivity of 20.5%. Sonography revealed three (75%) of four hepatocellular carcinomas larger than 5 cm in diameter, one (50%) of two hepatocellular carcinomas with diameters of 3.1-5.0 cm, one (20%) of five hepatocellular carcinomas with diameters of 2.1-3.0 cm, three (13.6%) of 22 hepatocellular carcinomas with diameters of 1-2 cm, and no lesions with diameters smaller than 1 cm. Forty-two patients (21%) had a total of 126 dysplastic nodules including two patients with innumerable lesions. Sonography depicted only two dysplastic nodules, for a patient sensitivity of 4.8% and a lesion sensitivity of 1.6%. The overall specificity of sonography for either hepatocellular carcinomas or dysplastic nodules was 96%.
CONCLUSION. Sonography has low sensitivity but high specificity in revealing hepatocellular carcinomas and dysplastic nodules in patients with a cirrhotic liver requiring liver transplantation. In these patients, sonography should not be the sole imaging modality used for lesion detection before transplantation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although serum 
-fetoprotein levels are used clinically to screen for
hepatocellular carcinoma in patients at risk, the -fetoprotein test is
relatively insensitive in revealing small tumors, and -fetoprotein
levels do not increase in the presence of dysplastic nodules
[7,8,9].
Therefore, imaging has an important role to play.
The accuracy of sonography in the detection of hepatocellular carcinoma in the cirrhotic liver varies widely, with reported sensitivities ranging from 33% to 96% [10,11,12,13,14,15,16,17,18,19,20]. In part, these discrepancies can be attributed to differences in study design. Many studies to date have been limited by small study size, lack of pathologic correlation, and long intervals between imaging and pathologic confirmation. Furthermore, few studies have evaluated the accuracy of sonography in the detection of both hepatocellular carcinoma and dysplastic nodules. The purpose of our study was to evaluate the sensitivity and specificity of sonography in the detection of both hepatocellular carcinomas and dysplastic nodules in patients with a cirrhotic liver by correlating sonographic findings obtained before transplantation with the results found at thinsection whole-liver explantation pathology.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Sonographic Technique
All sonograms were obtained on one of three types of sonography units:
XP128 or Aspen (Acuson, Mountain View, CA) or AI 5200S (Acoustic Imaging,
Tempe, AZ) scanners using 2.5-, 3.5-, or 4-MHz transducers. Examinations were
performed by experienced technologists; hard-copy images were provided to the
radiologist to review. Additional real-time scanning was performed as
necessary by the attending radiologist. All studies were interpreted by
abdominal radiology faculty with experience in interpreting sonograms of the
cirrhotic liver. Examinations were limited to gray-scale assessment, without
the use of harmonic imaging or contrast agents. Color Doppler sonographic
characteristics were not uniformly assessed. All focal solid lesions were
interpreted as potential hepatocellular carcinomas and were described with
respect to size, location, and echotexture. Focal areas of heterogeneity were
not considered positive findings. Lesions described as simple cysts were not
included in the analysis.
Pathologic Analysis
Explanted livers were serially sectioned into 5- to 8-mm sections.
Hepatocellular carcinomas and dysplastic nodules were identified grossly as
distinct from surrounding nodules in terms of size, texture, color, and degree
of bulging beyond the cut surface of the liver. Nodules were classified using
the International Working Party's terminology of nodular hepatocellular
lesions [21]: regenerative
nodule, dysplastic nodule—low grade, dysplastic nodule—high grade,
small hepatocellular carcinoma (
2 cm), or hepatocellular carcinoma (>2
cm). Low-grade dysplastic nodules were those that showed normal architecture
and either normal cytology or large-cell change. High-grade dysplastic nodules
contained one of the following features: diffuse small-cell change,
pseudogland formation, nodule-in-nodule lesions with small-cell change, iron
resistance in siderotic nodules, fatty change, clear-cell change, or
clustering of Mallory's bodies. Pathology reports before 1995 were revised to
reflect the new nomenclature. At pathology, the numbers, sizes, and anatomic
locations of all hepatocellular carcinomas and dysplastic nodules were
recorded. The presence and degree of liver cirrhosis were also recorded as
well as the cause of the patient's cirrhosis.
Radiology—Pathology Correlation
Because the study was conducted over the course of 9 years, not all
hard-copy sonograms were available for review. Therefore, data regarding
sonograms interpreted as showing negative findings for focal lesions were
obtained only from a review of the reports. However, all hard-copy sonograms
interpreted prospectively as showing positive findings were available and were
reviewed. The sonography reports and hard-copy images of patients with
reported lesions were retrospectively reviewed by a single radiologist who
correlated sonographic findings with pathology reports obtained from
evaluation of the explanted livers as well as from photomicrographs. All
lesions detected on sonography were matched with the pathologic descriptions
of lesion size and location. Four patients with diffuse multifocal
hepatocellular carcinomas and two patients with innumerable dysplastic nodules
were included in a separate category. Direct lesion correlation was difficult
in these patients because some pathology reports indicated only the presence
of a multifocal tumor. These patients were included for calculation of patient
sensitivity but could not be included in assessment for lesion sensitivity.
Patient and lesion sensitivities and specificities of sonographic detection of
hepatocellular carcinoma and dysplastic nodules were calculated for all other
patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Ten of 57 patients had at least one lesion detected on sonography, which is an overall patient sensitivity of 17.5%. Eight of 27 patients with pathologically confirmed hepatocellular carcinoma (single or multiple) had a lesion detected on sonography, resulting in a patient sensitivity of 29.6% (Figs. 1A,1B and 2A,2B). Two of 42 patients with dysplastic nodules had a lesion detected on sonography, resulting in a patient sensitivity for dysplastic nodules of 4.8% (Fig. 3A,3B,3C). On a lesion-by-lesion basis, (excluding the four patients with diffuse hepatocellular carcinoma), eight of 39 hepatocellular carcinomas were detected on sonography, which is a lesion sensitivity of 20.5% Two high-grade dysplastic nodules were detected on sonography, resulting in a lesion sensitivity for dysplastic nodules of 1.6%. Of the four patients with multifocal hepatocellular carcinoma, three had at least one lesion detected on sonography, and one had no lesions detected (Fig. 4A,4B). Neither of the two patients with multiple dysplastic nodules had nodules detected on sonography.
|
|
|
|
|
|
|
|
|
The size of hepatocellular carcinomas revealed on sonography ranged from 1.2 to 7.5 cm (mean, 4.1 cm). Sensitivity of sonography for hepatocellular carcinomas decreased with a decrease in the size of the lesion. Sonography revealed three (75%) of four hepatocellular carcinomas larger than 5 cm in diameter, one (50%) of two hepatocellular carcinomas with diameters of 3.1-5.0 cm, one (20%) of five hepatocellular carcinomas with diameters of 2.1-3.0 cm, three (13.6%) of 22 hepatocellular carcinomas with diameters of 1-2 cm, and no lesions with diameters smaller than 1 cm. All eight hepatocellular carcinomas detected were hypoechoic compared with the background liver parenchyma. The two dysplastic nodules detected were also hypoechoic and measured 1.5 and 1.0 cm, respectively.
Five patients whose sonograms raised suspicion of the presence of five solid lesions were found to have no lesion at pathology. An additional patient had a hyperechoic lesion detected on sonography that was subsequently found to be a hemangioma at pathology. Therefore, six of 200 patients had a false-positive diagnosis of hepatocellular carcinomas that was based on sonography, a patient-by-patient specificity of 96%.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Imaging surveillance of patients with cirrhosis may be performed with sonography, CT, or MR imaging, with the reported sensitivities varying depending on the technique, patient population, study design, and correlation between pathologic and imaging findings [8, 10,11,12,13,14,15,16,17,18,19,20, 28,29,30]. The advantages of sonography include its widespread availability, low cost, and lack of invasiveness. The sensitivity of sonography for the detection of hepatocellular carcinoma has previously been evaluated in studies with whole-liver explantation correlations. However, the patient populations, number of patients and lesions included, knowledge of lesions at the time of the examination, and time interval between the imaging and transplantation vary from study to study. In a retrospective study, Miller et al. [13] found a sonographic sensitivity for hepatocellular carcinoma of 81% in 36 patients; however, in this study, many of the patients were known to have hepatocellular carcinoma at the time of imaging and the exact number of lesions detected at pathology was not indicated. Dodd et al. [19] reported a lesion detection sensitivity of 42% for 69 hepatocellular carcinoma nodules in 28 patients with end-stage cirrhosis evaluated prospectively before transplantation. In this study, the time from sonography to transplantation ranged from 1 to 343 days (mean, 63 days). Shapiro et al. [14] reported a lesion sensitivity of 51% and patient detection sensitivity of 67% for 40 lesions in 21 patients. This study was limited to a retrospective review of patients with a known tumor, and the time interval from sonography to transplantation was not included.
These earlier studies did not evaluate the sensitivity of sonography for the detection of dysplastic nodules, a precursor lesion of hepatocellular carcinoma. Detection of these nodules has been addressed only in more recent studies with relatively small numbers of lesions. Rode et al. [17] reported a sensitivity of 46.2% (6/13) for detection of hepatocellular carcinomas, 33.3% (2/6) for "borderline nodules," and 2% (1/50) for macroregenerative (dysplastic) nodules in a study of 46 patients. In a study by Kim et al. [18], six of 18 hepatocellular carcinomas and zero of 20 dysplastic nodules were detected in 52 patients with advanced cirrhosis, which is a sensitivity of 33% and 0%, respectively.
In our large study of 200 patients with cirrhosis, all sonograms were obtained within 90 days (mean, 22 days) of transplantation. We chose the 90-day time period because of the rapid doubling time of some hepatocellular carcinomas [31]. As we previously noted, earlier studies included time intervals of 300 days or more [19], which may have falsely lowered sensitivity. In other series, patients were not known to have hepatocellular carcinoma at the time of sonography, which may have falsely increased sensitivity. We also included an analysis of sonographic sensitivity for dysplastic nodules, the prevalence of which in our patients was 21%, a finding that is in agreement with other series [32]. To our knowledge, our series represents the largest study to date to evaluate the sensitivity of sonography for detection of both hepatocellular carcinomas and dysplastic nodules in patients with a cirrhotic liver with wholeliver explantation pathologic correlation.
The results of our study support the conclusion that sonography is not sensitive for the detection of both hepatocellular carcinomas and dysplastic nodules in patients with advanced cirrhosis. The patient and lesion sensitivities for detection of hepatocellular carcinomas were 29.6% and 20.5%, respectively. The patient and lesion sensitivities for detection of dysplastic nodules were 4.9% and 1.6%, respectively. This limited sensitivity of sonography for detection of hepatocellular carcinomas and dysplastic nodules in the cirrhotic liver may result from the presence of fibrosis, fatty infiltration (altering background liver echogenicity), and nonneoplastic regenerative nodules. The alterations in background hepatic parenchymal echogenicity make infiltrating tumors particularly difficult to detect. In our series, sonography was limited for both the detection and characterization of lesions. We found no features to distinguish between the sonographic appearance of hepatocellular carcinoma and dysplastic nodules; both were hypoechoic. However, we should note that we detected only two dysplastic nodules in our series.
We found that detection of hepatocellular carcinoma on sonography was dependent on lesion size. We detected three (75%) of four hepatocellular carcinomas with diameters larger than 5 cm and only three (13.6%) of 22 lesions with diameters ranging from 1-2 cm. No lesions with diameters of less than 1 cm were detected. Dodd et al. [19] also found detection rates increased for larger tumors. Most dysplastic nodules in our series were smaller than 1 cm in diameter, which likely explains the substantially lower sensitivity of sonography for revealing dysplastic nodules compared with revealing hepatocellular carcinomas.
There are recognized limitations to our study, in part because of its retrospective nature. The hard-copy images of sonograms prospectively interpreted as showing negative findings were not reviewed; these data were obtained from reports. Therefore, it could not be verified that the findings of these studies were truly negative. Because sonography is a real-time imaging modality, it is unlikely that any lesion would have been detected in retrospect that was not documented at the time of scanning. For each lesion detected on sonography, the size and location of the lesion were described in the sonography report and the hard-copy images were reviewed. This procedure allowed correlation with pathology reports and photomicrographs. Not all patients were scanned by an attending radiologist, which may have lowered sensitivity results. However, in most high-volume sonography practices, it is impractical for every patient to be examined by a radiologist; the technologists perform the imaging. Therefore, the study reflects the circumstances most likely to be encountered in standard clinical practice.
An additional limitation of our study is the selection bias introduced
because only patients who underwent liver transplantation were included in our
analysis. At our institution, inclusion criteria for liver transplantation
generally include one hepatocellular carcinoma lesion smaller than 5 cm or up
to three lesions smaller than 3 cm and no venous invasion
[6]. Therefore, patients with
more or larger lesions had already been excluded, resulting in a study
population selection tilted toward patients with smaller and fewer lesions.
Furthermore, our patient cohort included only patients with Child-Pugh class B
or C cirrhosis, and our findings may not apply to patients with less severe
cirrhosis. Also, we did not evaluate the sensitivity of sonography combined
with -fetoprotein levels for hepatocellular carcinoma detection because
-fetoprotein levels were not available in all patients.
This study was carried out over the course of 9 years, and recent technologic advances have improved sonography. For example, we did not address the added value of harmonic imaging or sonographic contrast agents in revealing focal liver lesions. With these new technologies, increased detection of hepatocellular carcinomas and dysplastic nodules detection is likely [33,34,35].
In conclusion, our results show that sonography is not sensitive for the detection of hepatocellular carcinoma and dysplastic nodules in patients with a cirrhotic liver. We cannot endorse the routine use of gray-scale sonography as the sole modality for detecting lesions before liver transplantation. Finally, and solid lesion detected on sonography in a patient with cirrhosis severe enough to require transplantation should be considered hepatocellular carcinoma until proven otherwise and needs further characterization with CT, MR imaging, or biopsy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
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]
|
|
|
|
|
|
A. Giorgio, G. Ferraioli, L. Tarantino, G. de Stefano, V. Scala, F. Scarano, C. Coppola, and L. Del Viscovo Contrast-Enhanced Sonographic Appearance of Hepatocellular Carcinoma in Patients with Cirrhosis: Comparison with Contrast-Enhanced Helical CT Appearance Am. J. Roentgenol., November 1, 2004; 183(5): 1319 - 1326. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Matsumoto, T. Hirata, E. Ogawa, T. Fukuse, H. Ueda, T. Koyama, T. Nakamura, and H. Wada Ultrasonographic evaluation of small nodules in the peripheral lung during video-assisted thoracic surgery (VATS) Eur. J. Cardiothorac. Surg., September 1, 2004; 26(3): 469 - 473. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E Caturelli, L Solmi, M Anti, S Fusilli, P Roselli, A Andriulli, F Fornari, C Del Vecchio Blanco, and I de Sio Ultrasound guided fine needle biopsy of early hepatocellular carcinoma complicating liver cirrhosis: a multicentre study Gut, September 1, 2004; 53(9): 1356 - 1362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. E. Steingruber, A. Mallouhi, B. V. Czermak, P. Waldenberger, E. Gassner, F. Offner, A. Chemelli, A. Koenigsrainer, W. Vogel, and W. R. Jaschke Pretransplantation Evaluation of the Cirrhotic Liver with Explantation Correlation: Accuracy of CT Arterioportography and Digital Subtraction Hepatic Angiography in Revealing Hepatocellular Carcinoma Am. J. Roentgenol., July 1, 2003; 181(1): 99 - 108. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gupta, S. Bent, and J. Kohlwes Test Characteristics of {alpha}-Fetoprotein for Detecting Hepatocellular Carcinoma in Patients with Hepatitis C: A Systematic Review and Critical Analysis Ann Intern Med, July 1, 2003; 139(1): 46 - 50. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
6-cm)
hypoechoic solid mass (cursors) in left lobe.
