Accuracy of real-time shear wave elastography for assessing liver fibrosis in chronic hepatitis C: A pilot study


  • Potential conflict of interest: Nothing to report.

  • The ultrasound equipment (Aixplorer) was made available for the study by SuperSonic Imagine S.A. (Aix-en-Provence, France).


Real-time shear wave elastography (SWE) is a novel, noninvasive method to assess liver fibrosis by measuring liver stiffness. This single-center study was conducted to assess the accuracy of SWE in patients with chronic hepatitis C (CHC), in comparison with transient elastography (TE), by using liver biopsy (LB) as the reference standard. Consecutive patients with CHC scheduled for LB by referring physicians were studied. One hundred and twenty-one patients met inclusion criteria. On the same day, real-time SWE using the ultrasound (US) system, Aixplorer (SuperSonic Imagine S.A., Aix-en-Provence, France), TE using FibroScan (Echosens, Paris, France), and US-assisted LB were consecutively performed. Fibrosis was staged according to the METAVIR scoring system. Analyses of receiver operating characteristic (ROC) curve were performed to calculate optimal area under the ROC curve (AUROC) for F0-F1 versus F2-F4, F0- F2 versus F3-F4, and F0-F3 versus F4 for both real-time SWE and TE. Liver stiffness values increased in parallel with degree of liver fibrosis, both with SWE and TE. AUROCs were 0.92 (95% confidence interval [CI]: 0.85-0.96) for SWE and 0.84 (95% CI: 0.76-0.90) for TE (P = 0.002), 0.98 (95% CI: 0.94-1.00) for SWE and 0.96 (95% CI: 0.90-0.99) for TE (P = 0.14), and 0.98 (95% CI: 0.93-1.00) for SWE and 0.96 (95% CI: 0.91-0.99) for TE (P = 0.48), when comparing F0-F1 versus F2- F4, F0- F2 versus F3-F4, and F0 -F3 versus F4, respectively. Conclusion: The results of this study show that real-time SWE is more accurate than TE in assessing significant fibrosis (≥F2). With respect to TE, SWE has the advantage of imaging liver stiffness in real time while guided by a B-mode image. Thus, the region of measurement can be guided with both anatomical and tissue stiffness information. (HEPATOLOGY 2012;56:2125–2133)

In chronic hepatitis C (CHC), prognosis and management is driven largely by the extent of fibrosis.1, 2 Liver biopsy (LB) is still considered the gold standard in the evaluation of liver fibrosis, even though it is invasive, painful, costly, and with limitations in diagnostic use and accuracy. The accuracy of LB in assessing liver fibrosis is influenced by many factors, such as sampling error as well as intra- and interobserver variability.3, 4 Given these limitations, LB is not an ideal method for repeated assessment of disease progression. Following not only the progression, but also the regression of liver fibrosis over time could be of clinical significance, because research has demonstrated reduction in liver fibrosis with treatment, even in advanced stages.5, 6

These limitations of the LB have motivated research for noninvasive methods of measuring liver fibrosis. Transient elastography (TE) has emerged as the noninvasive test of reference and is entering clinical practice in Europe.7 TE is a noninvasive method that evaluates liver stiffness by measuring the velocity of elastic shear waves in the liver parenchyma generated by a mechanical push. Several studies have shown significant positive correlation between TE and stage of liver fibrosis.8-19 A key limitation of TE in clinical practice is the high rate of uninterpretable results, approximately 20% of cases according to the largest series to date.20

Elastography is commonly found on high-end ultrasound (US) equipment; however, most manufacturers use probe-induced or indigenous (i.e., respiratory or cardiac) displacements to generate qualitative images of strain. Because the stress over the medium is unknown, a quantifiable estimate of tissue stiffness cannot be obtained.21 Shear-wave elastography (SWE) is a new technique that is also based on shear waves implemented on a diagnostic US system. Like TE, SWE estimates the speed of a shear wave to provide a quantitative estimate of tissue stiffness. However, SWE has the advantage of being able to image liver stiffness in real time because the shear waves are generated by US pushes. Additionally, the SWE image is guided by a higher frame-rate B-mode image. This method could result in a more accurate score of fibrosis stage resulting from the SWE and B-mode image guidance.

The aim of our study was to evaluate the diagnostic accuracy of real-time SWE in the assessment of liver fibrosis in patients with CHC, in comparison with TE, by using the histologic METAVIR scoring system as the reference method.


ALT, alanine aminotransferase; ARFI, acoustic radiation force imaging; AUROC, area under the ROC curve; BMI, body mass index; CHC, chronic hepatitis C; CI, confidence interval; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR, interquartile range; kPa, kilopascal; LB, liver biopsy; LR, negative likelihood ratio; LR+, positive likelihood ratio; MRE, magnetic resonance elastography; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristic; ROI, region of interest; SD, standard deviation; SWE, shear wave elastography; TE, transient elastography; US, ultrasound.

Patients and Methods

Design, Overview, and Participants.

This was a single-center, cross-sectional study. From June 2010 through January 2012, all consecutive patients with confirmed CHC scheduled for liver biopsy at the Infectious Diseases Department of the Policlinico San Matteo at the University of Pavia (Pavia, Italy) were enrolled in the study. Inclusion criteria were the presence of hepatitis C virus (HCV) RNA in blood serum and, at least transiently, elevated serum alanine aminotransferase (ALT) levels. Patients with human immunodeficiency virus (HIV) coinfection and those under treatment were excluded from enrolment. Patient characteristics, epidemiological data, and biochemical tests were recorded. LB was performed on the same day as real-time SWE and TE as a day case procedure. Two physicians (G.F. and M.Z.), each of whom were blinded to the other's results, independently carried out the measurements. Real-time SWE measurements were all performed by G.F., and TE measurements were all performed by M.Z. The study protocol was approved by the institution's ethics committee. All participants gave their informed written consent.

This study was not sponsored by any real-time SWE manufacturer.

Real-Time SWE.

Real-time SWE studies were performed using the Aixplorer US system (SuperSonic Imagine S.A., Aix-en-Provence, France) with a convex broadband probe (SC6-1). In SWE, shear waves are created in tissue from the acoustic radiation force generated by focalized US pulses. A series of these push pulses creates plane shear waves, which propagate over a region of tissue. The speed of the shear wave is then estimated by a Doppler-like acquisition over a region.22 Finally, this shear-wave speed can then be used to calculate tissue stiffness by the formula E=ρc2, where E is tissue elasticity (in kilopascal; kPa), ρ is tissue density (kg/m3), and c is shear-wave velocity (m/s). Elasticity estimates are then color-coded creating a two-dimensional quantitative SWE image (kPa) of tissue stiffness, which is displayed in box form over a conventional B-mode image. The size and position of the SWE image is user adjustable, enabling a tradeoff in frame rate and extent of view. By placing a circular region of interest (ROI) in a SWE image, the mean and standard deviation (SD) of the elasticity within the ROI can be displayed. In this study, we used an SWE box size of 3.5 × 2.5 cm.

SWE measurements were performed on the right lobe of the liver, through intercostal spaces with the patient lying in the supine position with the right arm in maximal abduction.

The same intercostal space was used for both SWE measurements and LB, which was successively performed after SWE. The upper edge of the SWE box was placed 1.5-2.0 cm from Glisson's capsule in the liver and in an area of parenchyma free of large vessels. This placement of the SWE box aided in avoiding reverberation artifacts beneath Glisson's capsule and pulsations around larger vessels, which both could lead to erroneously elevated shear-wave speeds. Measurements of liver stiffness were obtained from the average of a circular ROI (2 cm in diameter), when scanning conditions permitted. The circular ROI was reduced in diameter, if limitations in viable signal within the SWE box prohibited a 2-cm diameter. Measurements were classified as failed when no or little signal was obtained in the SWE box for all acquisitions. The mean value of four consecutive measurements was used for statistical analyses. Because of temporal persistence, the displayed SWE image and measurements are the result of roughly three averaged frames in time. As a result, the average of four consecutive SWE acquisitions amounted to roughly 12 independent measurements to follow a similar protocol to that of TE.

The entire real-time SWE examination lasted approximately 5 minutes per patient.

We have recently assessed the reproducibility of SWE measurements in assessing liver elasticity in 42 healthy volunteers.23 Measurements were performed by two operators: an expert (operator 1) and a novice (operator 2). Intraobserver agreement showed intraclass correlation coefficient values of 0.95 (95% confidence interval [CI]: 0.93-0.98) and 0.93 (95% CI: 0.90-0.96) for operator 1 and operator 2, respectively. Interobserver agreement was 0.88 (95% CI: 0.82-0.94). The coefficient of variation between consecutive measurements ranged from 0.12 to 0.17. Healthy subjects showed real-time SWE values ranging from 4.92 (SD, 0.71) to 5.39 kPa (SD, 0.91).23


TE was carried out by using FibroScan (Echosens, Paris, France), a dedicated medical device that provides a quantifiable estimate of liver stiffness (kPa). The physician performing all examinations had the experience of at least 50 TE procedures, as previously recommended.10 Measurements of liver stiffness were performed on the right lobe of the liver through intercostal spaces on patients lying in the dorsal decubitus position with the right arm in maximal abduction, following the examination procedure previously described.8 Only patients with 10 validated measurements, a success rate of at least 60%, and interquartile range (IQR) of less than 30% of the median liver stiffness value were included. Values are expressed in kPa.

Liver Biopsy and Histology.

US-assisted percutaneous LB was performed by two experienced physicians (C.F. and G.M.) by using an intercostal approach. The same intercostal space, which was used for TE and SWE measurements, was used for the LB. A disposable 1.4-mm-diameter modified Menghini needle (Hepafix; Braun, Melsungen, Germany) was used. All biopsy specimens were fixed in formalin and embedded in paraffin. The length of each LB specimen (in millimeters) was recorded.

The specimens were read on site by a single expert liver pathologist (B.D.B.), blind to the results of both TE and real-time SWE results, but not to the patient's clinical and biochemical data. Liver fibrosis and necroinflammatory activity were evaluated semiquantitatively according to the METAVIR scoring system.24 Fibrosis was staged on a 5-point scale (from 0 to 4) according to the METAVIR scoring system (F0, absent; F1, enlarged fibrotic portal tract; F2, periportal or initial portal-portal septa, but intact architecture; F3, architectural distortion, but no obvious cirrhosis; and F4, cirrhosis).24 For the purpose of this study, F0 and F1 scores were grouped in the subsequent statistical analysis. Necroinflammatory activity was graded as follows: A0 = none; A1 = mild; A2 = moderate; and A3=severe. Steatosis was expressed as a percentage of fat in the hepatocytes and classified as absent (S0), <33% (S1), between 33% and 66% (S2), and >66% (S3).25

Statistical Analysis.

Descriptive statistics were produced for demographic, clinical, and laboratory characteristics for this study sample of patients. Shapiro-Wilk's test was used to test the normal distribution of quantitative variables. When quantitative variables were normally distributed, results were expressed as mean values and SD; otherwise, median and IQR (25th -75th percentile) were reported; qualitative variables were summarized as counts and percentages. Spearman's rank coefficient was used to test correlation between two study variables. Quantile regression was used for the multivariate model to assess the association between TE or real-time SWE and fibrosis, necroinflammation, steatosis, and biochemical tests. Kruskal-Wallis' one-way analysis of variance by ranks was used to test measurement differences between real-time SWE and TE among different METAVIR stages. A frequency distribution was obtained for choosing optimal cut-off values of real-time SWE and TE to maximize the sum of sensitivity and specificity for different fibrosis thresholds: F0-F1 versus F2- F4 (≥F2), F0-F2 versus F3-F4 (≥F3), and F0-F3 versus F4 (F = 4). The diagnostic performance of real-time SWE and TE, and their combinations, was assessed by using receiver operating characteristic (ROC) curves and the area under the ROC (AUROC) curve analysis.

Comparisons of AUROCs were done using the method described by DeLong et al. for correlated data.26 Agreement between METAVIR stages and real-time SWE and TE was illustrated with a contingency table, and weighted K was carried out. Data analysis was performed with the STATA statistical package (release 11, 2010; StataCorp LP, College Station, TX) and Medcalc (version 11.2; 2011 MedCalc Software bvba, Mariakerke, Belgium).

The study was conducted and written according to the Standards for Reporting of Diagnostic Accuracy.27



One hundred and thirty-eight patients were eligible during the recruitment period. Six patients were excluded because of HIV coinfection and 11 because they were under antiviral therapy. As a result of the recruitment from our referring physicians, no patients with overt cirrhosis or ascites were in this series of patients. One hundred and twenty-one patients met the inclusion criteria. In a total of 3 patients, liver stiffness measurements failed with real-time SWE and were unreliable with TE. Real-time SWE measurements failed in 2 of these patients because of narrow intercostal spaces and obesity (body mass index [BMI] >32 kg/m2) in the other patient. In these 3 patients, TE measurements showed an IQR >30% in 2 patients because of narrow intercostal spaces and a success rate of less than 60% in the patient with obesity. An additional TE unreliable measurement resulting from IQR >30% was experienced again because of obesity. The characteristics of the 121 patients are summarized in Table 1. There were 87 men and 34 women. Mean length of LB specimens was 27 mm (SD, 8, range, 10-55), and length was >15 mm in 117 of 121 (96.7%) cases and >25 mm in 59 of 121 (48.8%) cases. In univariate analysis, real-time SWE and TE values showed a high correlation with degree of fibrosis and necroinflammation and a low correlation with degree of steatosis. Corresponding values for real-time SWE and TE, respectively, were as follows: (1) for liver fibrosis: r = 0.83 (P < 10−5) and r = 0.74 (P < 10−5); (2) for degree of necroinflammation: r = 0.59 (P < 10−5) and r = 0.57 (P < 10−5); and (3) for degree of steatosis: r = 0.26 (P = 0.07) and r = 0.24 (P = 0.01). Real-time SWE and TE values showed a moderate negative correlation with platelets (r = −0.51, P < 0.001; r = −0.50, P < 0.001) and prothrombin time (r = −0.49, P < 0.001; r = −0.52, P < 0.001). No correlation with other variables was found. Multivariate regression analysis, including METAVIR stage, METAVIR grade, steatosis, and platelets (prothrombin time was not included in the multivariate models because it correlated with platelets), confirmed the correlation, for both SWE and TE, with fibrosis stage (P < 0.001 for both), but not with all other variables. Within each fibrosis stage, no correlation was found between METAVIR activity grade and TE or SWE measurements (Fig. 1).

Table 1. Patient Demographics and Their Biochemical and Histological Data at Liver Biopsy Examination
Characteristicsn = 121
  1. SD values represent mean, and IQR values represent median.

  2. Abbreviation: AST, aspartate aminotransferase.

Sex, men (%)87 (71.9)
Age, years (SD; range)44.8 (11.9; 19-76)
BMI, kg/m2 (SD; range)25.4 (3.8; 17.1-39.0)
AST, IU/L (IQR; range)41 (27-81; 14-208)
ALT, IU/L (IQR; range)75 (40-126; 10-558)
Alkaline phosphatase, IU/L (SD; range)130.6 (54.3; 36-301)
Gamma-glutamyl transferase, IU/L (IQR; range)42.5 (25.9-87.5; 8-1,094)
Total bilirubin, μM/L (IQR; range)6.4 (4.8-9.3; 2.6-45.0)
Serum albumin, g/L (IQR; range)42 (40-43; 31-49)
Platelets count, 103/mm3 (IQR; range)210.5 (168-264; 93-437)
Prothrombin time, % (SD; range)94.4 (12.7; 64-121)
Fibrosis score (METAVIR) (%) 
 F0-F150 (41.3)
 F233 (27.3)
 F314 (11.6)
 F424 (19.8)
Steatosis (%) 
 S0 (0)77 (63.6)
 S1 (<33)32 (26.4)
 S2 (33-66)6 (5.0)
 S3 (>66)6 (5.0)
Activity grade (METAVIR) (%) 
 A03 (2.5)
 A151 (42.1)
 A250 (41.3)
 A317 (14.1)
Figure 1.

Box-and-whisker plots of (A) SWE and (B) TE values for each METAVIR stage in relation to degree of necroinflammatory acitivity. Liver stiffness values are reported on the y-axis, and METAVIR grade of necroinflammatory activity are reported on the x-axis. The central box represents values from the lower to upper quartile (25th-75th percentile). The line through each box represents the median. Error bars show minimum and maximum nonextreme values. A, necroinflammatory activity (METAVIR grade); F, fibrosis (METAVIR stage); o, extreme values; *, only 2 patients with A0 grade in F4 stage.

Liver Stiffness Assessment: Comparison of Real-Time SWE and TE.

Median values, IQR, range, number of outliers, and P values of measurements obtained for each fibrosis stage with SWE and TE are shown in Table 2. Figure 2 shows the ROCs for significant and severe fibrosis, as well as cirrhosis, and improvements in AUROCs by real-time SWE over TE for each level of fibrosis. For significant fibrosis (≥F2), a statistically significant improvement (P = 0.002) in AUROCs was observed between real-time SWE (0.92) and TE (0.84). The slight improvements for the AUROCs for severe fibrosis and cirrhosis were not significant (P = 0.14 and P = 0.48, respectively).

Table 2. Median Values, IQR, Range, Outliers, and P Values of Measurements Obtained for Each Fibrosis Stage With SWE and TE
METAVIR StageMethodF0-F1F2F3F4
  1. P values refer to differences between consecutive fibrosis stages (*F0-F1 versus F2; **F2 versus F3; ***F3 versus F4).

Median value, kPaSWE6.27.610.015.6
Number of outliersSWE1210
P valueSWE 0.0001*0.003**0.09***
TE 0.02*0.002**0.06***
Figure 2.

Comparison of ROC curves for real-time SWE and TE for different fibrosis thresholds. (A) F0-F1 versus F2- F4 (>F2), (B) F0-F2 versus F3-F4 (>F3), and (C) F0-F3 versus F4 (F = 4). In parentheses, 95% confidence intervals are shown. P values of differences between AUROCs are given.

Optimal cut-off values for the different levels of fibrosis were determined by analysis of ROCs for both real-time SWE and TE. Sensitivity, specificity, AUROCs, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR) of optimal cut-off values for each METAVIR stage are found in Table 3. With respect to TE, real-time SWE showed a higher sensitivity in assessing significant (≥F2) and advanced fibrosis (≥F3), whereas the sensitivity of TE in the assessment of cirrhosis (F = 4) was slightly higher with respect to that observed with real-time SWE. On the other hand, both TE and real-time SWE showed a high specificity in assessing liver cirrhosis.

Table 3. Resulting Clinical Performance of SWE and TE Using Optimal Measurement Cut-Off Values
ParameterMethod≥F2 (95% CI)≥F3 (95% CI)F = 4 (95% CI)
Cutoff in kPaSWE7.18.710.4
Sensitivity, %SWE90.0 (80.5-95.9)97.3 (85.8-99.9)87.5 (67.6-97.3)
TE69.6 (57.3-80.1)89.2 (74.6-97.0)91.7 (73.0-99.0)
Specificity, %SWE87.5 (74.8-95.3)95.1 (87.8-98.6)96.8 (91.0-99.3)
TE89.6 (77.3-96.5)88.8 (79.7-94.7)96.8 (90.9-99.3)
PPV, %SWE91.3 (82.0-96.7)90.0 (76.3-97.2)87.5 (67.6-97.3)
TE90.6 (79.3-96.9)78.6 (63.2-89.7)88.0 (68.8-97.5)
NPV, %SWE85.7 (72.8-94.1)98.7 (93.1-100)96.8 (91.0-99.3)
TE67.2 (54.3-78.4)94.7 (86.9-98.5)97.8 (92.4-99.7)
LR+SWE7.2 (6.3-8.2)19.7 (18.3-21.2)27.4 (23.5-32.0)
TE6.7 (5.6-8.0)7.9 (6.9-9.1)28.4 (25.0-32.2)
LRSWE0.11 (0.04-0.3)0.03 (0.003-0.2)0.13 (0.03-0.6)
TE0.34 (0.1-0.8)0.12 (0.04-0.4)0.09 (0.02-0.5)

Table 4 shows the concordance rate of real-time SWE and TE versus METAVIR stage. Overall, real-time SWE correctly classified 98 of 118 (83.1%) patients, whereas TE correctly classified 78 of 117 (66.7%) patients. Weighted K was higher for real-time SWE with respect to TE (0.90 versus 0.83), but this difference was not statistically significant (P = 0.272). Even though real-time SWE performed better than TE in F3 stage, both real-time SWE and TE showed a lower rate of correctly classified patients in this stage with respect to the others.

Table 4. Analysis of Concordance of SWE and TE Versus METAVIR Stage
 F0-F1F2F3F4TotalConcordance Rate (%)
F0-F1, kPa      
 SWE ≤7.1427004985.7
 TE ≤6.94319116467.2
F2, kPa      
 7.1 < SWE ≤ 8.7424012982.8
 6.9 < TE ≤ 8.036201154.5
F3, kPa      
 8.7 < SWE ≤ 10.4211121668.8
 8.0 < TE ≤ 11.627711741.2
F4, kPa      
 SWE > 10.4012212487.5
 TE > 11.6003222588.0
Cumulative concordance      
 SWEWeighted K = 0.9098/11883.1
 TEWeighted K = 0.8378/11766.7


In this study, the diagnostic accuracy of real-time SWE and TE in estimating liver fibrosis was compared against histology in patients with CHC. Real-time SWE measurements compared favorably to that of TE in assessing severe fibrosis and cirrhosis. Real-time SWE demonstrated a significant improvement in the identification of significant fibrosis, when compared to TE.

AUROCs in differentiating no or mild fibrosis (F0-F1) from significant fibrosis (≥F2) were 0.84 and 0.92 for TE and real-time SWE, respectively (P = 0.002). The performance of TE in identifying severe fibrosis (≥F3) and cirrhosis (F = 4) is already quite high. No significant difference was observed between AUROCs of TE and real-time SWE for severe fibrosis (0.96 and 0.98, respectively) and cirrhosis (0.96 and 0.98, respectively). These findings suggest that real-time SWE can be used in the same way as TE is being used for the assessment of severe fibrosis and cirrhosis, with the benefit of improved assessment of significant fibrosis.

To the best of our knowledge, this is the first study aimed at comparing the noninvasive assessment of liver fibrosis by means of real-time SWE and TE measurements in patients with CHC undergoing liver biopsy in the same day. Bavu et al. compared the performance of the SWE method to TE in patients with CHC, but not to liver histology.28 A prototype system was used where an SWE image was retrospectively generated. The investigators reported similar SWE values for the lower fibrosis stages, but slightly higher values for F3 and F4, along with larger IQRs per stage, than that found in our series. The extra targeting information from the real-time SWE image, combined with a smaller sampling region of liver parenchyma, could explain the differences between our results at higher stages of fibrosis. The higher stages of fibrosis have been shown to have greater spatial heterogeneity.29

Another shear-wave–based elastography technique, referred to as acoustic radiation force imaging (ARFI), has been implemented on an US system. This approach differs from real-time SWE, in that it is limited to a quantitative estimate of liver stiffness at a single location. A small ROI with a measurement dimension of 0.5 × 1.0 cm2 is placed over a B-mode image to define the location from where a shear-wave estimate can then be performed by a user button push. Generally, the same protocol commonly used for TE is also used to obtain ARFI. Several pilot studies and recently some larger studies have appeared comparing ARFI with TE.30-32 In general, these studies report ARFI having similar performance to that of TE for the assessment of significant and severe fibrosis and cirrhosis, however with the advantage of being easier to use because of guidance with a B-mode image. One larger study reports a significant improvement of ARFI over TE for the detection of significant fibrosis; however, the performance of TE was unusually low (AUROC of 0.73).32 Real-time SWE has three advantages with respect to TE. First, real-time SWE, like ARFI, is integrated into a conventional diagnostic US system and therefore can make use of real-time B-mode imaging for the assessment of morphologic changes or detection of focal liver lesions (e.g., hepatocellular carcinoma). The use of the B-mode image for the guidance of SWE acquisitions is likely to also help to improve variability in stiffness measurements.30 Second, real-time SWE should benefit from improved separation of fibrosis stages as a result of the use of shear waves with greater bandwidths.28 The third advantage real-time SWE has over TE, as well as ARFI, is in providing a real-time quantitative map of liver tissue stiffness. The spatial heterogeneity of liver stiffness can be visualized, and the size of the region used for a measurement can be selectively placed or adjusted. As a result, physiological variations of liver fibrosis can be averaged to better represent the fibrosis state. The ROI for liver stiffness measurements can be adjusted in size and location to avoid artifacts, such as those arising around larger pulsating vessels. The real-time acquisition of real-time SWE enables user adjustment during acquisition for targeting a homogenous region of liver tissue. This also ensures that excessive liver motion is avoided during real-time SWE acquisitions. These advantages could be helpful to reduce the variability in liver stiffness measurements with real-time SWE.

In the study of Huwart et al.,33 magnetic resonance elasticity measurements showed a good performance for the prediction of liver fibrosis, with AUROCs of magnetic resonance elastography (MRE) significantly larger than those of TE (0.994 versus 0.985 for ≥F2, 0.998 versus 0.837 for ≥F3, and 0.960 versus 0.930 for F = 4). Even though accurate, MRE is a costly procedure and requires longer acquisition time with respect to real-time SWE. Moreover, claustrophobia and iron overload could hamper the results.

In this study, the correlation between liver stiffness values, assessed either with real-time SWE or TE, and fibrosis stage were not affected by steatosis or necroinflammation, which is in agreement with previous studies assessing liver elasticity with TE or MRE.8-10, 33, 34 However, most patients in our series, as observed in Huwart et al.'s series,33 had no-to-mild steatosis.

In other studies, in which an influence of necroinflammatory activity on TE measurements have been found,35-37 the inclusion criteria were different from that applied in our study. In fact, in their series patients with hepatitis B35, 36 as well as other different etiologies37 were included. Unlike with the findings of other studies,35-37 we did not find a correlation between ALT values and TE results. On the other hand, only 25% of patients in our series had ALT values >3 times the normal upper limit.

In our series, SWE correctly classified 83.1% of patients, whereas TE correctly classified 66.7%. On the other hand, it has been suggested that most discordant results between TE and histology were caused by histology measurement failures.8-10, 38 Even though still considered the benchmark for validation of noninvasive techniques aimed at assessing degree of liver fibrosis, the accuracy of LB examination is challenged by sampling errors and intra- and interobserver variability.3, 4 Moreover, the METAVIR scoring system does not take into account quantitative changes in liver collagen, but is largely an assessment of architectural changes in a small sample of the liver.29 Although distribution of fibrosis in the liver is heterogeneous, histological staging is based on a biopsy specimen that represents, at most, 1/50,000th of the total liver mass.39

In CHC, prognosis and management strongly rely on degree of liver fibrosis. In patients with CHC, assessment of the stage of fibrosis is a crucial issue preceding therapy. Treatment should be initiated promptly in patients with advanced fibrosis (METAVIR score F3-F4) and should be strongly considered in patients with significant fibrosis (METAVIR score F2).1, 2 In this study, real-time SWE improved the ability to identify significant fibrosis, in comparison with TE. Whereas in the United States TE is not U.S. Food and Drug Administration approved,1 the recent guidelines for management of hepatitis C infection of the European Association for the Study of the Liver have indicated that noninvasive methods can currently be used, instead of LB, in patients with CHC to assess liver disease severity before therapy at a safe level of predictability.2 In fact, TE has been approved and entered clinical use in France for the assessment of fibrosis in naïve patients with CHC without comorbidities.40

This study has limitations. First, our series of patients were composed of a variable distribution of patients for the different stages of fibrosis, particularly for F3. Second, our study included a cohort of European patients with low prevalence of obesity, which is the major technical limitation reported for TE, therefore the applicability in the general population of the results we have obtained with real-time SWE is limited. Third, the analysis was carried out in a relatively small number of patients, and it would be critical to validate these results in larger studies.

In conclusion, the results of this study show that real-time SWE was more accurate than TE in assessing significant fibrosis (≥F2). With respect to TE, SWE has the advantage of imaging liver stiffness in real time while guided by a B-mode image. As a result, liver stiffness measurements benefit from the guidance of both anatomical and tissue stiffness information. This enables the identification of artifacts in the SWE image, such as pulsating vessels or reverberation artifact or lack of signal or motion. As opposed to TE in difficult patients, the operator can search for an acoustic window in real time, where a sufficient SWE signal can be obtained for measurement. Real-time SWE is subject to the same limitations encountered with conventional US imaging modes, where, largely, user dependencies and patient body habitus can influence liver stiffness measurements. Further studies in larger patient populations are needed to confirm these results and the values of the thresholds of real-time SWE for the different fibrosis stages.


The authors thank Matthew Bruce, Assistant Director of Ultrasound at SuperSonic Imagine, for his technical support and Ms. Livia Astroni, Ms. Natali Calabrese, Mr. Filippo Cuda, Mr. Lorenzo Guioli, Ms. Maura Marchisoni, Ms. Giampiera Nava, Ms. Loredana Pavesi, Ms. Barbara Ricci, nurses in the outpatient ward of the infectious diseases department, for their valuable help in complying with the study protocol.


The members of the Liver Fibrosis Study Group are: Elisabetta Above, M.D., Giorgio Barbarini, M.D., Enrico Brunetti, M.D., Willy Calderon, M.D., Marta Di Gregorio, M.D., Raffaella Lissandrin, M.D., Serena Ludovisi, M.D., Laura Maiocchi, M.D., Antonello Malfitano, M.D., Giuseppe Michelone, M.D., Mario Mondelli, M.D., Savino F.A. Patruno, M.D., Alessandro F. Perretti, M.D., Gianluigi Poma, M.D., Paolo Sacchi, M.D., Marco Zaramella, M.D., Fondazione IRCCS Policlinico San Matteo, University of Pavia, Pavia, Italy.