| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Internal Medicine II, Saarland University Hospital,
Kirrbergerstrasse, Bldg. 41, Homburg/Saar 66421, Germany.
2 Institute of Pathology, Mannheim, Germany.
Received March 3, 2006;
accepted after revision July 31, 2006.
Address correspondence to M. Friedrich-Rust
(mireen.friedrich.rust{at}uniklinikum-saarland.de).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Real-time elastography was performed in 79 patients with chronic viral hepatitis and known fibrosis stage and in 20 healthy volunteers. A specially developed program was used for quantification of tissue elasticity. Stepwise logistic regression analysis was performed to define an elasticity score using variables with high reproducibility in a preceding analysis of data from 16 different patients. In addition, aspartate transaminase-to-platelet ratio index (APRI) and routine laboratory values were included in the analysis.
RESULTS. The Spearman's correlation coefficient between the
elasticity scores obtained using real-time elastography and the histologic
fibrosis stage was 0.48, which is highly significant (p < 0.001).
The diagnostic accuracy expressed as areas under receiver operating
characteristic (ROC) curves were 0.75 for the diagnosis of significant
fibrosis (fibrosis stage according to METAVIR scoring system [F] 
F2),
0.73 for severe fibrosis (F F3), and 0.69 for cirrhosis. For a combined
elasticity-laboratory score, the areas under the ROC curves were 0.93, 0.95,
and 0.91, respectively.
DISCUSSION. Real-time elastography is a new and promising sonography-based noninvasive method for the assessment of liver fibrosis in patients with chronic viral hepatitis.
Keywords: elastography • fibrosis • hepatitis • liver biopsy • liver disease • real-time elastography • sonography
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Transient elastography (FibroScan) is performed with an ultrasound transducer probe mounted on the axis of a vibrator. A vibration transmitted from the vibrator toward the tissue induces an elastic shear wave that propagates through the tissue. These propagations are followed by pulse-echo sonographic acquisitions and their velocity, which is directly related to tissue stiffness, is measured: The harder the tissue, the faster the shear wave propagates [26, 27, 29, 33].
Real-time elastography is a new method for measurement of tissue elasticity integrated in a sonography machine (Hitachi EUB-8500 and EUB-900) and is technically different from transient elastography. With conventional ultrasound probes, echo signals before and under slight compression are compared and analyzed [34]. Because tissue elasticity cannot be measured directly from reflected ultrasound echoes, in previous studies using the elastography principle, methods analyzing the displacement of phases (e.g., cross-correlation method) were investigated [35-44]. However, these measurements were associated with strong aliasing. To overcome these restrictions, Hitachi Medical Systems developed real-time elastography based on the combined autocorrelation method and 3D tissue model for the determination of phase displacement in real time without aliasing [34, 45-47].
In recent studies, researchers have evaluated real-time elastography for the characterization and detection of focal lesions in the breast, thyroid gland, and prostate gland [34, 48-52]. Here, we present a proof-of-principle study to evaluate the value of realtime elastography for the assessment of liver fibrosis. Results of real-time elastography were compared with fibrosis stage obtained by assessing liver biopsy samples and with the APRI [25].
The aims of the study were, first, to establish a formula to obtain a first liver elasticity score for this new method (i.e., real-time elastography); second, to determine the accuracy of elasticity scores for correct prediction of liver fibrosis stage; third, to assess the diagnostic accuracy of predicting significant fibrosis (fibrosis stage according to METAVIR scoring system [F] 3 F2); and, fourth, to optimize the accuracy of fibrosis stage prediction by the combination of real-time elastography with simple laboratory parameters.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Because the mean progression rate of liver fibrosis has been estimated to be 0.133-0.154 fibrosis stages on the METAVIR scoring system per year [53], the five patients with a longer time interval between liver biopsy and real-time elastography were enrolled in the present study. The 59 patients who underwent a liver biopsy did not have overt signs of cirrhosis on sonography or MRI. However, in four of these 59 patients, fibrosis stage 4 (cirrhosis) was diagnosed on liver biopsy. None of the 59 patients received antiviral therapy between liver biopsy and real-time elastography.
For the additional 20 patients with chronic viral hepatitis and liver cirrhosis, no liver biopsy had been performed during the previous 2 years. In these patients, obvious clinical signs of liver cirrhosis were present at the time of presentation in the outpatient clinic. The definition of liver cirrhosis in these patients was based on results of sonography or MRI (liver surface nodularity, hypertrophy of segment I, signs of portal hypertension) or on historical histology results and on clinical and biochemical signs of cirrhosis (thrombocytopenia, low serum albumin, hyperbilirubinemia, endoscopic signs of portal hypertension, history of ascites).
All patients were enrolled consecutively for realtime elastography between January and August 2005 on the basis of the time of presentation in the hepatology outpatient clinic of Saarland University Hospital.
Blood parameters were obtained on the same day that real-time elastography was performed; for further description, see the Materials and Methods section, Blood Markers subheading.
The control group consisted of healthy adult volunteers with normal liver enzyme levels, negative anti-HCV antibodies, and negative HBsAg who did not have a history of relevant concomitant illness, such as heart, lung, or liver disease or neoplasia. All healthy volunteers did not take any medication or drugs and did not have an excessive daily alcohol intake (> 15 g/d).
In addition, for statistical reasons, real-time elastography was evaluated twice in 16 different patients with liver disease for a preceding analysis to assess the reproducibility of different variables characterizing elasticity. These two examinations were performed on the same day by two different operators to define variables with the lowest interobserver variability of real-time elastography. The two operators were blinded to each other's findings and to the clinical and histologic data of the patients.
The present study was performed in accordance with the ethical guidelines of the Declaration of Helsinki.
Liver Histology
Liver biopsy samples were taken via a right intercostal space from the
right liver lobe. First, sonography was performed to find the safest and best
accessible intercostal space from which to obtain a biopsy sample. After
disinfection and local anesthesia of the skin, intercostal space, peritoneum,
and liver capsule, liver biopsy was performed at the previously marked site
with the Hepafix set (B. Braun Melsungen AG) for percutaneous liver biopsy
using an 18-gauge needle (Menghini needle, B. Braun Melsungen) with an outer
diameter of 1.2 mm.
Local anesthesia and liver biopsy were performed without direct sonographic guidance. Liver biopsy specimens were fixed in formalin and embedded in paraffin. Four-micrometer-thick sections were stained with H and E stain and elastica van Gieson stain. An experienced pathologist blinded to the results of realtime elastography analyzed all the biopsy specimens.
Liver fibrosis stages were evaluated semiquantitatively according to the METAVIR scoring system [54]. Liver fibrosis was staged on an F0-F4 scale: F0, no fibrosis; F1, portal fibrosis without septa; F2, portal fibrosis with few septa; F3, numerous septa without cirrhosis; and F4, cirrhosis [54]. Steatosis was assessed according the number of hepatocytes with fatty degeneration: 0, none; 1, 1-10% of hepatocytes; 2, 11-33% hepatocytes; 3, 34-66% hepatocytes; and 4, 67-100% hepatocytes. The biopsies were judged as adequate if the number of portal tracts was at least six.
Blood Markers
The following blood parameters were determined in the same laboratory:
platelet count, aspartate aminotransaminase level (AST), alanine
aminotransaminase level (ALT), 
-glutamyl transpeptidase (GGT) level,
and total bilirubin level. Enzymatic activity was measured at 37°C,
according to International Federation of Clinical Chemistry standards.
The APRI index was calculated as follows: AST (x upper limit of normal range) x 100/platelet count (109/L). The upper limit of normal of AST for women was 35 IU/mL, and the upper limit of normal for men was 50 IU/mL [25].
Real-Time Elastography Principle
Real-time elastography is an imaging technique that can reveal the physical
property of tissue using conventional ultrasound probes. Although used for
clinical studies with sonography devices manufactured by different companies
[34,
48-52],
to our knowledge currently real-time elastography is commercially available
only in the Hitachi EUB-8500 and EUB-900 machines. The tissue elasticity
distribution can be calculated by the strain and stress of the examined
tissue.
In a first step, the amount of displacement of the reflected ultrasound echoes before and under compression are measured (stress field). In hard tissue, the amount of displacement of the reflected ultrasound echoes before and under compression is low, whereas in elastic or soft tissue, the amount of displacement is high because soft tissue can be compressed more than hard tissue. In addition, with the combined autocorrelation method, echo-frequency patterns of parallel ultrasound echoes are compared to detect possible lateral evasion of hardened tissue areas.
|
Elasticity Score
Patients and control subjects underwent real-time elastography (EUB-8500,
Hitachi Medical Systems; 9-MHz probe). The patients were examined in a supine
position with the right arm elevated above the head. Patients were instructed
to continue breathing as usual. Because each elastography image is obtained in
a few milliseconds, breathing did not cause any motion artifacts. The
examination was performed on the right lobe of the liver through the
intercostal space in all patients because, first, liver biopsy was also
performed on the right lobe; second, a subcostal approach cannot be achieved
in all patients; and, third, the compressions performed were micrometer
compressions only, which could be performed intercostally without having to
apply great pressure. An area was chosen where the liver tissue was at least 6
cm thick and was free of large blood vessels.
The examination was performed with a 9-MHz transducer because, similar to B-mode imaging, higher frequencies allow better analysis of areas close to the transducer, and assessment of real-time elastography is optimized by the manufacturer on superficial tissues. The measurement depth was between 20 and 50 mm (mean, 35 mm) with a 350-500 mm2 area of measurement (mean, 420 mm2). The results were considered reliable only if a pressure of 3-4 on a scale of 0-6 arbitrary units was applied for measurement (Fig. 1). Ten valid measurements were performed in each subject and recorded as colorcoded images. The entire examination lasted approximately 5-10 minutes per patient.
Because this study was, to our knowledge, the first application of real-time elastography for evaluating liver tissue elasticity, a new elasticity score had to be established. For each pixel, the value of the elasticity distribution was calculated on a scale of 0 to 1 (0 = maximum elasticity, 1 = minimum elasticity) from the color-coded bit-map image yielded by a computer program specially developed by our study group using Matlab (version 6, MathWorks). These results were summarized by descriptive statistics as mean, median, minimum, maximum, and the frequency of pixel values above 0.75 of a single measurement.
In a second step, further descriptive statistics were obtained taking into account images from all measurements with standardized pressure. In this way, we obtained a large number (> 250) of summarizing variables for characterization of real-time elastography using a professional computer program for statistical analysis (SPSS version 12.0, SPSS). To avoid over-fitting and to exclude unreliable variables, the final analysis was based on only 10 such summarizing variables that characterized elasticity with high reproducibility in a small preceding analysis of realtime elastography in 16 different patients. For the final analysis, we used stepwise multivariate logistic regression for prediction of significant fibrosis to define an elasticity score and also a combined elasticity-laboratory score with a mean value of approximately 100 and an SD of approximately 10 in patients with no or minor fibrosis.
Statistical Analysis
Because elasticity-laboratory scores were not normally distributed we used
the nonparametric Jonckheere-Terpstra test to compare these scores with
histologic fibrosis stage. The results are illustrated as the median and 25th-
to 75th-percentile values (box plot). The correlations between both scores and
the histologic fibrosis stage were also assessed using Spearman's correlation
coefficient.
The diagnostic performance of the elasticity score, APRI, and the combination of the elasticity score and laboratory values was assessed by receiver operating characteristic (ROC) curves. The ROC curve represents sensitivity versus 1 minus the specificity for all possible cutoff values for prediction of significant and severe fibrosis, respectively. The areas under the ROC curve (AUCs) and 95% CIs of AUC were calculated (SPSS version 12.0). Here, AUC values close to 1.0 indicate the highest diagnostic accuracy. Cutoff values defining prediction regions for each fibrosis stage were defined by a common optimization step maximizing the sum of the sensitivities in predicting the single stages. Finally, sensitivity, specificity, and positive and negative predictive values were calculated without further adjustments on the basis of the same data set using Matlab. Thereby, the whole study population was analyzed; the group of patients with proven liver cirrhosis was assigned to METAVIR fibrosis stage F4.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
The overall elasticity score was defined as a linear combination of descriptive statistics using only the upper half of the single images. The elasticity score is calculated with the following formula:
177 + 50 x log10 median(freq(pixel 0.75)) - 13,000
x min(min(pixel with values > 0))
where freq is frequency, and min is minimum. The elasticity scores ranged from 65 to 122.
Adding information from blood markers may improve the diagnostic accuracy of real-time elastography. Therefore, a logistic regression analysis was performed including a few routine laboratory parameters. The best accuracy could be achieved by combining the elasticity score with platelet count and GGT calculated by the following formula:
combined elasticity - laboratory score = 158 + 0.28 x elasticity score - 38 x log10 platelet count [103/mm3] + 0.3xlog10 GGT [x ULN]
where ULN is the upper limit of the normal range for GGT. The upper limit of normal of GGT for women was 39 IU/mL, and the upper limit of normal of GGT for men was 66 IU/mL. The combined elasticity-laboratory score ranged from 82 to 137.
Liver Histology
The mean number of histologic cores was two (range, 1-5). The mean number
of portal tracts was nine (range, 6-15). The mean length of liver biopsy
samples was 27.1 ± 14.6 mm.
Relationship Between Liver Elasticity and Liver Histology
Interestingly, all 10 elasticity variables used in the multivariate
logistic regression were significantly associated with the fibrosis stage and,
in all but one variable, the Jonckheere-Terpstra test resulted in a p
value below 0.0001. For comparison of histologic liver fibrosis stages and
real-time elastography elasticity scores, a high correlation of increasing
elasticity scores with increasing stage of fibrosis was observed
(Fig. 2A). The Spearman's
correlation coefficient between the elasticity scores and the histologic
fibrosis stages was highly significant, with a value of 0.48 (p <
0.001).
The AUCs, a measurement of the diagnostic accuracy of a test, were 0.75 for
the diagnosis of significant fibrosis (F F2), 0.73 for the diagnosis of
severe fibrosis (F F3), and 0.69 for the diagnosis of cirrhosis (F = F4)
(Fig. 3 and
Table 2). With regard to the
distribution of elasticity scores according to METAVIR fibrosis stages, the
cutoff values were determined as 100.1 for F F2, 102.5 for F F3, and
111.75 for F = F4 (Table 3).
Altogether, 80% of the patients with significant fibrosis (F F2) could be
correctly identified with the real-time elastography (sensitivity). In
patients with an elasticity score of less than 100.1, the presence of
significant fibrosis (F F2) could be excluded in 78.6% of cases (negative
predictive value) (Table 3).
Excluding liver biopsy length shorter than 25 mm did not improve the results.
Elasticity scores were not influenced by the degree of steatosis in the
liver.
|
|
Relationship Between APRI Index and Liver Histology
A high correlation of increasing elasticity scores with increasing stage of
fibrosis was observed (Fig.
2B). The Spearman's correlation coefficient between the APRI index
and the histologic fibrosis stages was highly significant, 0.77 (p
< 0.001). The AUCs for the APRI index were 0.87 for the diagnosis of
significant fibrosis (F F2), 0.88 for the diagnosis of severe fibrosis (F
F3), and 0.88 for the diagnosis of cirrhosis (F = F4)
(Fig. 3 and
Table 2).
Relationship Between the Combined Elasticity-Laboratory Score and Liver Histology
The best diagnostic accuracy was obtained by combining the variables used
for the calculation of the elasticity score with platelet count and GGT. A
high correlation of increasing combined elasticity-laboratory scores with
increasing stage of fibrosis was observed
(Fig. 2C). Spearman's
correlation coefficient between these scores and the histologic fibrosis stage
was highly significant with 0.78 (p < 0.001). The AUCs for the
combined scores were 0.93 for the diagnosis of significant fibrosis (F
F2), 0.95 for the diagnosis of severe fibrosis (F F3), and 0.91 for the
diagnosis of cirrhosis (F = F4) (Table
2). With regard to the distribution of elasticity-laboratory
scores according to METAVIR fibrosis stages, the cutoff values were determined
to be 99.1 for F F2, 99.7 for F F3, and 114.7 for F = F4
(Table 4). The sensitivity for
the detection of significant fibrosis was 84.4%
(Table 4).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
F2) is of special interest because the presence
of fibrosis is an important parameter for evaluation of indications for
antiviral treatment. Therefore, a test with a high diagnostic accuracy for the
determination of significant fibrosis is of great therapeutic value. In
previous studies, estimation of liver fibrosis was performed with transient
elastography (Fibro-Scan) and was shown to have a high diagnostic accuracy for
the prediction of significant fibrosis
[26,
27,
29]. In the present study, we
have shown that real-time elastography can also be used as a sonography-based
method for noninvasive measurement of liver fibrosis. Patients with ascites were excluded from the study because real-time elastography, similar to FibroScan, works only with close contact to the liver. However, the development of ascites is a strong indicator for the presence of cirrhosis that makes noninvasive staging of fibrosis unnecessary. In the present study, patients with a body mass index (BMI) of up to 30 were enrolled, and real-time elastography was performed successfully in all included patients. Whether real-time elastography can be performed successfully in patients with a BMI > 30 needs to be addressed in future studies. The examination was performed through a right intercostal space, and patients did not need to hold their breath during the examination. Furthermore, no influence of the degree of steatosis in the liver on elasticity scores was observed in the present study. Similar results have been reported from studies analyzing transient elastography with FibroScan [55, 56].
In the present study, we found highly significant positive correlation
between elasticity scores obtained with real-time elastography and the METAVIR
fibrosis stage. An AUC value of 0.75 was obtained using realtime elastography
(elasticity score) for the diagnosis of significant fibrosis (F F2). In
comparison, three recent studies analyzing the noninvasive assessment of liver
fibrosis with FibroScan revealed AUCs between 0.75 and 0.84 for the diagnosis
of significant fibrosis (F F2)
[26,
27,
29]. Nevertheless, this is the
first evaluation of real-time elastography for the diagnosis of liver
fibrosis.
It is likely that the diagnostic accuracy of real-time elastography may be improved by further optimization of the images using different ultrasound probes, refined selection of the analyzed area of liver tissue, and a more refined statistical assessment of the elasticity images for a larger data set or a larger number of images for each patient. Especially the diagnosis of cirrhosis seems to be improvable if more reliable assessments of the variability between the single images are available. It is known that the accuracy of liver biopsy is limited because of significant intra- and interobserver variability and sampling errors [11-14]. Nevertheless, because this is the first study assessing real-time elastography for the detection of liver fibrosis, liver biopsy was used as a gold standard without further analysis of possible false staging of biopsy.
The diagnostic accuracies obtained for the APRI in the present study (AUC:
0.87 for F 3 F2, 0.88 for F F3, and 0.88 for F = F4) were
similar to those revealed in the large APRI evaluation study by Wai et al.
[25] (AUC: 0.80 and 0.88 for F
F2, and 0.89 and 0.94 for F = F4), but they were higher than in previous
studies comparing the APRI with the FibroTest or FibroScan (AUC: 0.74-0.75 for
F F2, 0.80-0.82 for F = F4)
[27,
57]. These differences might
be due to the relatively large group of patients without significant fibrosis
in the present study (including the healthy subjects) and in the study by Wai
et al. (50%) compared with the other studies (25%). Although the definition of
the upper limit of the normal range differs and limits the use of APRI
according to the literature, it did not seem to affect the results in our
study because only the correlation and AUC were determined, not specific
cutoffs.
The results of the APRI were superior to the real-time elastography
findings, although the highest diagnostic accuracy in the present study was
obtained by a mathematic combination of the elasticity score and two routine
laboratory values (platelet count and GGT) (i.e., combined
elasticity-laboratory score), achieving AUCs of 0.93 for the diagnosis of
significant fibrosis (F F2).
Recently, Castera et al.
[27] compared transient
elastography (FibroScan), FibroTest, APRI, and liver biopsy for the assessment
of liver fibrosis in a large number of patients. Interestingly, the authors
reported the best performance also by a combination of liver elasticity with
blood markers (FibroScan and FibroTest) with a comparable AUC value of 0.88
for the diagnosis of significant fibrosis (F 2).
In conclusion, real-time elastography is a new and promising sonography-based noninvasive method for the assessment of liver fibrosis in patients with chronic viral hepatitis. In combination with simple laboratory values, real-time elastography can further improve the discrimination of different fibrosis stages, which plays a decisive role in the management of patients with viral hepatitis.
Future studies on larger patient cohorts are necessary for improvement and also validation of the elasticity scores and their cutoffs obtained in the present study because the analyses of sensitivity and specificity in the present study were performed on the same data set as the calculation and optimization of the elasticity scores and cutoffs. Results of further studies are needed before real-time elastography can be introduced in clinical practice. In addition, the combination of real-time elastography with other blood tests such as FibroTest may further improve specificity and sensitivity for the noninvasive estimation of liver fibrosis. A head-to-head comparison of transient elastography (FibroScan) and real-time elastography would be of future interest as well.
Acknowledgments
We are grateful to Professor Klaus Remberger (Institute of Pathology,
Saarland University Hospital, Homburg/Saar, Germany) for providing the
histologic specimens and helpful discussions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Alam, K. Naito, J. Horiguchi, H. Fukuda, T. Tachikake, and K. Ito Accuracy of Sonographic Elastography in the Differential Diagnosis of Enlarged Cervical Lymph Nodes: Comparison with Conventional B-Mode Sonography Am. J. Roentgenol., August 1, 2008; 191(2): 604 - 610. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Gulizia, G. Ferraioli, and C. Filice Open Questions in the Assessment of Liver Fibrosis Using Real-Time Elastography Am. J. Roentgenol., June 1, 2008; 190(6): W370 - W371. [Full Text] [PDF]
|
|
|
|
|
|
A. Saftoiu, D. I. Gheonea, and T. Ciurea Reply Am. J. Roentgenol., June 1, 2008; 190(6): W372 - W373. [Full Text] [PDF]
|
|
|
|
|
|
C. G. Lall, A. M. Aisen, N. Bansal, and K. Sandrasegaran Nonalcoholic Fatty Liver Disease Am. J. Roentgenol., April 1, 2008; 190(4): 993 - 1002. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Friedrich-Rust, E. Herrmann, S. Zeuzem, and C. Sarrazin Reply to Real-Time Elastography in the Assessment of Liver Fibrosis Am. J. Roentgenol., February 1, 2008; 190(2): W164 - W164. [Full Text] [PDF]
|
|
|
|
|
|
A. Saftoiu, D. I. Gheonea, and T. Ciurea Hue Histogram Analysis of Real-Time Elastography Images for Noninvasive Assessment of Liver Fibrosis Am. J. Roentgenol., October 1, 2007; 189(4): W232 - W233. [Full Text] [PDF]
|
|
|
|
|
|
G. Ferraioli, R. Gulizia, and C. Filice Real-Time Elastography in the Assessment of Liver Fibrosis Am. J. Roentgenol., September 1, 2007; 189(3): W170 - W170. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
