LIVER ELASTOGRAPHY= EFSUMB GUIDELINES

LINK    DOWNLOAD  Ultrasound Med Biol. 2015 May;41(5):1161-79. doi: 10.1016/j.ultrasmedbio.2015.03.007. Epub 2015 Mar 20

....

POINT SHEAR WAVE SPEED MEASUREMENT (PSWSM) AND SHEAR WAVE SPEED IMAGING (SWSI)

Procedure

All technologies are implemented in a conventional US system under direct visualization using a curved array broadband transducer. A sample box is positioned on B-mode image of the liver and elastography measurements are obtained by pressing a button. Optimal conditions include:

 -Fasting;
 -Dorsal decubitus position, with the right arm elevated above the head for optimal intercostal access;  

 -Resting respiratory position (breath-hold without deep inspiration);  

 -ROI placement beneath Glisson’s capsule by 1.5-2.0 cm to avoid reverberation artifacts and increased subcapsular stiffness;

 - ROI placement to avoid large liver vessels;  

 -The median value of 5-10 measurements is considered with PSWSM, and the mean value of 4 measurements with SWSI.

Specific recommendations include:  

+For SWSI, the sample box size should be large enough to reduce the variation between measurements. This provides a cumulative value that is the average of stiffness at several points, thus being more representative of the heterogeneous stiffness in abnormal and normal livers.  

+For PSWSM, the ROI should be placed perpendicular to the center of the transducer surface as the angle of insonation may have a slight but significant influence on the result.

Results

Point Shear Wave Speed Measurement.

As of today, there are two techniques: Virtual Touch Tissue Quantification (VTTQ) technique that expresses the results in m/sec (Figure 3) and ElastPQ that gives the results in m/sec or in kPa (Figure 4). There are numerous reports of studies performed using the VTTQ technique, which has been commercially available since 2009, but only a few using ElastPQ, which was introduced in 2012. The reproducibility of the VTTQ technique is excellent, with an intraclass correlation coefficient ranging from 0.84 to 0.87 (Bota et al. 2012; Boursier et al. 2010; D’Onofrio et al. 2010; Guzman-Aroca et al. 2011). Operator training does not appear to be necessary (Boursier et al. 2010). Similarly, the ElastPQ technique is highly reproducible, with an interobserver agreement ranging from 0.83 for comparison of single measurements to 0.93 for the median value of 10 measurements (Ferraioli et al. 2014). In healthy volunteers, the values of PSWSM performed with VTTQ are available in several publications (D’Onofrio et al. 2010; Friedrich-Rust et al. 2009a; Goertz et al. 2012; Grgurevic et al. 2011; Kaminuma 2011; Karlas 2011; Kim 2010; Kircheis 2012; Osaki 2010; Piscaglia 2011; Rifai 2011; Rizzo et al. 2011; Son et al. 2012; Sporea et al. 2011; Takahashi et al. 2010). In all studies, the values were lower (, 1.2 m/sec) than in patients with chronic hepatitis. Food intake significantly increases the liver stiffness values (Goertz et al. 2012; Popescu et al. 2013). The median value of PSWSM obtained with ElastPQ in healthy volunteers is 3.5 kPa (Ling et al. 2013; Ferraioli et al. 2014).

Chronic viral hepatitis.

 The range of cut-offs for each fibrosis stage is quite large with overlap between consecutive stages. The range of cut-offs for the fibrosis stage ranges from 1.13 to 1.55 m/sec for F.2; from 1.43 to 1.81 m/sec for F.3; and from 1.36 to 2.13 m/sec for F4. The largest series comprises more than 600 patients with mixed etiologies of chronic liver disease (Kircheis et al. 2012). Using TE as the reference method, the investigators obtained cut-off values of 1.32 m/sec for F2 and 1.62 m/sec for F4. Similar cut-offs were obtained in the meta-analysis of Friedrich-Rust et al. (2012a), in which nine studies were analyzed. Patients with chronic liver disease of several etiologies were included, and the cut-off values were 1.34, 1.55 and 1.80 m/sec, for significant fibrosis, severe fibrosis and cirrhosis, respectively. PSWSM showed accuracy similar to that of TE for the diagnosis of severe fibrosis, whereas a slightly but significantly higher diagnostic accuracy of TE with respect to PSWSM was found for the diagnosis of significant fibrosis and liver cirrhosis. In the more recent metaanalysis of Bota et al. (2013), in which thirteen studies were included, PSWSM showed a predictive value similar to TE for significant fibrosis and cirrhosis. In an international multicenter study comprising 1,095 patients (181 with chronic hepatitis B and 914 with chronic hepatitis C), the correlation of PSWSM with histological fibrosis was significantly higher in patients with chronic hepatitis C compared with those with chronic hepatitis B (r50.653 vs. r50.511, p50.007), whereas both groups showed similar PSWSM values for each fibrosis stage (Sporea et al. 2012). In the study of Rizzo et al. (2011), using the PSWSM cut-offs of 1.3 m/sec for the diagnosis of significant fibrosis (F $ 2), 1.7 m/s for severe fibrosis (F $ 3), and 2.0 m/sec for cirrhosis (F 5 4), the highest concordance was obtained for the diagnosis of mild fibrosis. TE may overestimate the fibrosis stage in cases with severe liver inflammation (Sagir et al. 2008; Arena et al. 2008). The same limitation has been observed in some studies with PSWSM (Takahashi et al. 2010; Yoon et al. 2012; Chen et al. 2012), but not in others (Friedrich-Rust et al. 2009a; Palmeri et al. 2011; Rizzo et al. 2011; Nishikawa et al. 2014).

 Figure 3. VTTQ technique in a healthy subject. Measurements of liver stiffness are given in m/sec; the sample box is shown.

Figure 4. ElastPQ in a patient with chronic hepatitis C of F4 Metavir stage on liver histology. The values of liver stiffness are expressed in kPa. Bottom left corner of the image: the stiffness is estimated and displayed by using a scale that ranges from soft to hard.

The grade of liver steatosis appears not to influence PSWSM (Friedrich-Rust et al. 2009a; Rizzo et al. 2011; Rifai et al. 2011). Measurement failure with PSWSM is reported in less than 3% of patients (Friedrich-Rust et al. 2012a). No invalid measurement occurred in the series of Rizzo et al. (2011) and in the study of Crespo et al. (2012). PSWSM provided valid results in all patients whereas TE failed in 11% of cases. In the series of Bota et al. (2014), reliable measurements were obtained in 93.3% of cases. Older age, higher BMI and male gender were associated with the risk of failed and unreliable measurements. A recent meta-analysis, which included either full papers or abstracts for a total of 36 studies, has shown that BMI has a significant influence for the diagnosis of significant fibrosis (F $ 2) (Nierhoff et al. 2013). In this meta-analysis, the diagnostic accuracy expressed as the area under the ROC curve was 0.84, 0.89 and 0.91 for the diagnosis of significant fibrosis, advanced fibrosis, and cirrhosis, respectively. Measurements are not limited by ascites because the US push beam, which generates the shear waves, propagates through fluids and appears not to be influenced by clinical and biochemical variables (Rizzo et al. 2011). The possibility to evaluate several areas of the liver parenchyma could be another advantage of PSWSM. Indeed, histological studies have shown that liver fibrosis is not homogeneously distributed within the liver, and can be missed when performing liver biopsy at one site only, thus leading to underestimation of liver fibrosis (Bedossa et al. 2003; Maharaj et al. 1986). It should be noted that D’Onofrio et al. (2010) reported significant differences between intercostal and subcostal scans, and Kaminuma et al. (2011) found that PSWSM were more reliable when performed in a deep portion of the right lobe. In their series of patients with chronic hepatitis, Toshima et al. (2011) obtained significantly higher values in the left lobe of the liver than the right lobe. It has been suggested that oscillation of the left liver by cardiac activity may interfere with stiffness measurements (Osaki et al. 2010). Karlas T et al. (2011) found that, in healthy individuals, the shear wave speed was higher in the left liver than in the right, but no difference in speed was observed in patients with advanced fibrosis and cirrhosis. The authors suggest that the absence of differences between the two sides could be a criterion for the diagnosis of advanced liver disease. Preliminary results of PSWSM using ElastPQ in 102 patients with chronic hepatitis C have shown that the accuracy of the method for staging liver fibrosis is similar to that of TE and the best cut-off value for significant fibrosis (F $ 2) is 5.7 kPa (Ferraioli et al. 2014). In a series of 291 patients with chronic hepatitis B, the AUROCs for significant fibrosis and cirrhosis were 0.94 and 0.89, respectively (Ma et al. 2014).

Monitoring disease progression and prognosis.

Very few studies regarding disease progression and prognosis have been published with conflicting results. In the cohort of Vermehren et al. (2012), the diagnostic accuracy of PSWSM of the liver and the spleen for the prediction of esophageal varices was not signifi- cantly different from that of TE and the Fibrotest; however, the AUROCs of all methods were fairly low, ranging from 0.50 to 0.58. In the series of Morishita et al. (2013), a cutoff value of 2.39 m/s had a sensitivity of 81% and a specificity of 82% for detecting high-risk esophageal varices. In a recent study, a spleen stiffness value ,3.3 m/s ruled out the presence of high-risk varices in patients with compensated or decompensated liver cirrhosis (negative predictive value, 99.4%). Regardless of the etiologies of liver disease, spleen stiffness was highly accurate for the detection of esophageal varices (Takuma et al. 2013).

 Nonalcoholic fatty liver disease (NAFLD).

Only few studies in small series of patients are available (Yoneda et al. 2010; Osaki et al. 2010; Palmeri et al. 2011; Friedrich-Rust et al. 2012b; Fierbinteanu Braticevici et al. 2013). In 172 patients diagnosed with NAFLD, a cutoff of 4.24 kPa distinguished low (fibrosis stage 0–2) from high (fibrosis stage 3–4) fibrosis stages with a sensitivity of 90% and a specificity of 90% (AUROC 0.90) (Palmeri et al. 2011). In a study on 61 patients with NAFLD/ NASH, the paired comparison of diagnostic accuracies between TE and PSWSM for the diagnosis of significant fibrosis, severe fibrosis and liver cirrhosis were similar (Friedrich-Rust et al. 2012). In 64 patients with histologically proven NAFLD, the diagnostic performance of PSWSM in predicting significant fibrosis and cirrhosis had an AUROC of 0.94 and 0.98, respectively (Fierbinteanu Braticevici et al. 2013).

Shear Wave Speed Imaging. SWSI expresses the results in m/sec or kPa (Figure 5). Ferraioli et al. (2012b) found ICCs of 0.95 and 0.93 for expert and novice operators when comparing measurements performed on the same day, and 0.84 and 0.65 for measurements performed on different days. The interobserver agreement was 0.88. These results have been confirmed in the recently published study of Hudson et al. (2013). Like conventional US, SWSI technology may be user dependent, so it is recommended that at least 50 supervised scans and measurements should be performed by a novice to obtain consistent measurements (Ferraioli et al. 2012b). Values ranging from 2.6 to 6.2 kPa have been reported for histologically proven normal livers (Suh et al. 2013). In patients with chronic hepatitis C, cut-off values for SWSI are reported in two publications based on 4 (Ferraioli et al. 2012c) or 5 (Bavu et al. 2011) measurements from an intercostal space. Bavu et al. (2011) evaluated 113 patients with chronic hepatitis C, comparing the results to those obtained with TE; liver biopsy was not performed. The results showed a good agreement between fibrosis staging and elasticity assessment. SWSI showed a higher accuracy in assessing mild and intermediate stages of fibrosis. The diagnostic accuracy of SWSI in the assessment of liver fibrosis in patients with chronic hepatitis C was evaluated in a pilot study on 121 patients (Ferraioli et al. 2012c). The optimal cut-off values of SWSI were 7.1 kPa for significant fibrosis (F $ 2), 8.7 kPa for advanced fibrosis (F $ 3), and 10.4 kPa for cirrhosis (F54). Areas under the ROC curves were 0.92 for F $ 2; 0.98 for F $ 3 and 0.98 for F54. A better performance of SWSI compared to TE has also been observed in 226 patients with chronic hepatitis B (Leung et al. 2013). In a study that evaluated liver fibrosis in a cohort of 422 patients without a gold standard, Poynard et al. (2013) report that the applicability of SWSI is lower than that of TE whereas the performance of the two methods is similar. In the same study, the applicability of SWSI was higher than that of TE in patients with ascites. It has been reported that stiffness values are not correlated with liver steatosis (Ferraioli et al. 2012c; Suh et al. 2013) or with necro-inflammation (Ferraioli et al. 2012c).

Limitations 

- SWSI accuracy has only been assessed in the right lobe through intercostal access. Interlobe variations of liver stiffness have been reported with PSWSM. Body habitus (obesity, narrow intercostal spaces) may hamper the results.

 - Because of the frequency-dependency of the elasticity properties of tissue, great care and consideration must be used when comparing quantitative results among these techniques.
 - Results in kilopascals are not comparable between SWSI, PSWSM and TE.

 - The majority of the studies has been performed in patients with chronic hepatitis C, therefore these cutoffs may not be applicable to other viral etiologies or to NAFLD. Only small series of patients with NAFLD have been studied, therefore the cut-offs in these patients need further assessment.

 - Readings may be higher in patients with ALT levels greater than five times the upper limit of normal; thus, the effect of inflammation should be taken into account, and the results should always be evaluated in the clinical setting. As with TE, it is likely that congestive heart failure, and feeding will be associated with a stiffer liver.

Recommendations

PSWSM and SWSI can be used to assess the severity of liver fibrosis in patients with chronic viral hepatitis, best evidenced in patients with hepatitis C. Nonetheless, the evidence that is available is still limited, particularly for SWSI. Like TE, PSWSM and SWSI are more accurate in detecting cirrhosis than significant fibrosis.

Figure 5. SWSI technique in a patient with decompensated liver cirrhosis. Measurements of liver stiffness are given in kPa. The mean value along with the minimum and maximum values and the standard deviation are shown.

RECOMMENDATIONS

- Is elastography useful in the evaluation of diffuse liver disease?

Liver elastography is useful for the evaluation of diffuse liver diseases. The level of evidence is high for TE, moderate for PSWSM, and still low for SWSI and SE. Some methods have been used for more than ten years while others have been introduced more recently, resulting in large variability in the number of published manuscripts on different techniques. The majority of studies have evaluated patients with viral chronic hepatitis and results obtained in this setting may not be applicable to other clinical situations as the critical cut-offs are strongly dependent on the etiology. Values with shear wave-based elastography and with strain techniques vary between manufacturers. Thus, the cutoffs are both system and etiology dependent. Elastography is capable of distinguishing significant fibrosis (F2 or greater) from non-significant (F0 - F1) fibrosis. However, more data are needed to confirm its use to distinguish between consecutive stages of early fibrosis. It is also important to note that each method may provide different values expressed in different units (meters per second, kilopascals) or indices. Several confounding factors have been identified, such as liver inflammation, liver congestion and biliary obstruction. Elastography results should be interpreted in the full clinical context of the patient, taking into account the method used to obtain the results. Elastography can be used for follow-up of patients with chronic liver diseases.

- Is the method reproducible?

Generally, the reproducibility of elastography techniques is good. However, manufacturer recommendations should be followed. Dedicated training is required for all elastography methods.

- What is its accuracy in a range of pathologies?

The accuracy of elastography methods improves with the severity of fibrosis. The most studied etiology is chronic viral hepatitis. The body of evidence is highly dependent on the method for other etiologies.

 - What are the limitations?

Obesity is a common limitation of all ultrasoundbased elastography methods. Other limitations are narrow intercostal spaces and, for transient elastography, the presence of ascites. Most methods show higher values when the levels of aminotransferases are elevated. Some manufacturers do not recommend the use of liver elastography in pregnancy.

- To what extent can elastography reduce the use of liver biopsies?

In some countries, where liver elastography is used in clinical practice, the number of liver biopsies has decreased significantly. When elastography results are consistent with other clinical findings, liver biopsy may be avoided.

 - Can elastography provide additional information for focal liver lesions? 

Currently, the body of evidence concerning the use of elastography in focal liver lesions is not strong enough to recommend its use in clinical practice. These recommendations are based on the international literature and on the findings of the WFUMB expert group.

WFUMB Guidelines for Ultrasound Elastography - Liver . G. FERRAIOLI et al. Ultrasound in Medicine and Biology Volume 41, Number 5, 2015.