On patient related factors and their impact on ultrasound-based shear wave elastography of the liver

(1)

This is the published version of a paper published in Ultrasound in Medicine and Biology.

Citation for the original published paper (version of record):

Byenfeldt, M., Elvin, A., Fransson, P. (2018)

Ultrasound in Medicine and Biology, 44(8): 1606-1615 https://doi.org/10.1016/j.ultrasmedbio.2018.03.031

Access to the published version may require subscription.

N.B. When citing this work, cite the original published paper.

Permanent link to this version:

http://urn.kb.se/resolve?urn=urn:nbn:se:umu:diva-150663

(2)

●

Original Contribution

ON PATIENT RELATED FACTORS AND THEIR IMPACT ON ULTRASOUND-BASED SHEAR WAVE ELASTOGRAPHY OF THE LIVER

Marie Byenfeldt,*Anders Elvin,^†and Per Fransson^*

* Department of Nursing, Umeå University, Umeå, Sweden; and^†Department of Molecular Medicine and Surgery, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden

(Received 13 August 2017; revised 23 March 2018; in final form 29 March 2018)

Abstract—The aim of the study was to investigate patient-related factors associated with either reliable or poorly reliable measurement results of ultrasound-based shear wave elastography (SWE) of the liver. A total of 188 pa- tients were analyzed prospectively with binary logistic regression using the interquartile range/median as cutoff to define two groups based on reliable and poorly reliable SWE results. SWE results correlated significantly with liver biopsy. Factors associated with reliable SWE results (i.e., no negative impact on measurements) were age, sex, cirrhosis, antiviral and/or cardiovascular medication, smoking habits and body mass index. Factors associ- ated with poorly reliable SWE results were increased skin-to-liver capsule distance (odds ratio= 3.08, 95% confidence interval: 1.70–5.60) and steatosis (odds ratio= 2.89, 95% confidence interval: 1.33–6.28). These findings indicate that the interquartile range/median as a quality parameter is useful in avoiding poorly reliable SWE results. How best to examine patients with increased skin-to-liver capsule distance is a matter of some controversy, as the in- cidences of obesity, diabetes and metabolic syndrome are increasing worldwide; however, our results indicate that reliable SWE results can be obtained in this group of patients by using ultrasound-based SWE. (E- mail:marie.byenfeldt@umu.se,Marie.byenfeldt@aleris.se) © 2018 The Author(s). Published by Elsevier Inc. on behalf of World Federation for Ultrasound in Medicine & Biology. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Key Words: Shear wave elastography, Liver fibrosis, Skin-to-liver capsule distance, Interquartile range, Quality parameter, Reliability, Steatosis, Liver biopsy.

INTRODUCTION

Since the introduction of the non-invasive ultrasound- based method of shear wave elastography (SWE), the number of patients undergoing liver biopsy has declined dramatically. Nevertheless, standardized examination pro- tocols for SWE have yet to be established, and uncertainties persist concerning how to perform reliable SWE examinations (Cosgrove et al. 2013); thus, there is a need to standardize these procedures. Several factors may affect SWE reliability (Dietrich et al. 2017), and little is known about how sex, body mass, patient positioning and examination technique influence the results.

Liver fibrosis is a progressive disease that can develop from chronic liver conditions, such as alcoholic steatohepatitis (ASH) (Canbay et al. 2016; Joshi-Barve et al.

2015) and non-alcoholic steatohepatitis (NASH) (Lee et al.

2017), as well as hepatitis B virus (HBV) (European Association for the Study of the Liver [EASL] 2017) and hepatitis C virus (HCV) (EASL 2014). Hepatitis may cause inflammation of the liver, which can lead to fibrosis, cirrhosis and, in the worst case, hepatocellular carcinoma (HCC) (EASL 2014, 2015, 2017). It is important to stage these patients so that treatment can be started when necessary because an estimated 22% of individuals with HCV progress to cirrhosis within 20 years (Freeman et al. 2001).

Several classification systems are available for staging liver fibrosis based on histologic findings, and the most com- monly used in Europe is the Metavir score table, where F0 represents normal liver tissue and F4 represents cirrhosis (Goodman 2007). Direct-acting antiviral agents (DAAs) are available that can leave patients virus free after successful treatment, and in Sweden, the consensus for treatment initiation is a cutoff Metavir≥F2. Reliable liver fibrosis staging is critical to ensure that the right patients receive treatment, because the cost for DAA treatment is substantial (Lagging et al. 2017). In addition, repeated follow-up examinations for previously detected fibrosis are

Address correspondence to: Marie Byenfeldt, Department of Nursing, Umeå University, Umeå, SE- 901 87 Sweden. E-mail:marie.byenfeldt@

umu.se,Marie.byenfeldt@aleris.se

https://doi.org/10.1016/j.ultrasmedbio.2018.03.031

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1606

(3)

important, and both invasive and non-invasive methods are available (Srinivasa Babu et al. 2016).

The current gold standard for diagnosing liver fibrosis is liver biopsy; however, some authors have viewed liver biopsy as an imperfect option (Barr et al. 2016). It is also an invasive method that involves tissue sample variability (Bedossa et al. 2003), even when performed by experienced physicians. Likewise, when expert patholo- gists interpret biopsy findings, the error rate is as high as 20% (Castera et al. 2010) for staging fibrosis, failure to recognize cirrhosis occurs in 20% of cases (Abdi et al.

1979; Afdhal 2003) and sampling errors arise because biopsies represent only 1/50,000th of the total liver mass (Lee 1994, 1–21). Inter-observer variation between two pa- thologists is 6%–10% among all cases with respect to staging (Regev et al. 2002). One in 1000 liver biopsies carries a risk for severe complications (Piccinino et al.

1986), with a mortality rate of approximately 1/100,000 (Bravo et al. 2001).

It would thus be advantageous if non-invasive methods could replace liver biopsy because such methods could be repeated frequently and without the need for post-procedure hospitalization. Blood marker tests, although non-invasive, exhibit low accuracy in discriminating among intermedi- ate stages of fibrosis, and several hepatic and extra- hepatic conditions may influence them (Schiavon et al.

2014). Hence, great interest exists in establishing non- invasive methods to diagnose liver fibrosis, such as SWE (Barr 2014), which can also predict significant liver fibrosis stage≥2 (Beland et al. 2014). These methods can also be repeated daily if necessary, and several options are available for detecting, monitoring and staging liver fibrosis.

Two such technologies are transient elastography (TE), using the FibroScan device, and ultrasound-based SWE.

Both technologies involve the creation of shear waves in the tissue while the speed of their propagation is measured and quantified. The applied force that generates the shear waves differs between the two methods; for TE, forces are mechanically generated through the skin surface into the liver, whereas for the ultrasound-based method, the source is acoustic radiation force impulses (ARFIs), also known as push pulses (Dietrich et al. 2017). TE using the FibroScan device is well established and was introduced in 2003. Results with this technology correlate well with the degree of fibrosis (Armstrong et al. 2013; Kettaneh et al.

2007). Although TE does not allow for B-mode imaging, it has been assumed to be less sensitive to boundary conditions, and the acquisition time is short and well adapted to mobile organs such as the liver (Sandrin et al. 2003).

Using ultrasound-based SWE with push pulse, the B-mode imaging capacity allows for measurements of the region of interest (ROI) set by the operator either as a focused point (pSWE) or within a volume (2-D SWE), depend- ing on the ultrasound device (Dietrich et al. 2017; Piscaglia

et al. 2016). Ultrasound-based SWE thus has a clear ad- vantage over TE because vessels and lesions can easily be avoided (Bamber et al. 2013; Barr 2014; Nightingale et al. 2002; Shiina et al. 2015). Results of SWE with push pulses correlate well with liver fibrosis. The mean diagnostic accuracy of ARFI, expressed as area under the receiver operating characteristic curve, is 0.87 for significant fibrosis (F≥ 2), 0.91 for severe fibrosis (F ≥ 3) and 0.93 for cirrhosis (Friedrich-Rust et al. 2009, 2012; Lupsor et al. 2009). In addition, intra- and inter-operator reliability is good to excellent (Bota et al. 2012; Hudson et al.

2013). An international multicenter study including 10 centers and 5 countries reported a highly significant correlation between liver fibrosis and ultrasound-based SWE results (Sporea et al. 2012).

To maintain reliable and valid measurement for SWE in the liver when using TE technology with the FibroScan device, the manufacturer, Echosens, specifies that two pa- rameters must be employed: the success rate (SR) and the reliability criteria for liver stiffness (Boursier et al. 2013).

SR is met if a minimum of 60% of at least 10 measurements are performed successfully. The reliability criteria for liver stiffness require that measurement results be con- sidered poorly reliable when the interquartile range (IQR)/

median is>0.30 with a liver stiffness median of ≥7.1 kPa (Boursier et al. 2013; Castera et al. 2010; Sandrin et al.

2003). Because SWE measurements do not have a normal distribution, the median value should be used. The IQR (the difference between the 75th and the 25th percen- tiles) is used as a distribution measurement for the median and expresses the distribution around the median (Dietrich et al. 2017).

The relationship between shear wave speed and shear modulus is represented by the equation CT= √µ/ρ, where CT= speed of shear wave propagation, µ = shear modulus andρ = density. The equation is for linear, isotropic and elastic solids, and two challenges lie in the relationship of generating shear waves within tissues in vivo and re- constructing CT-measured displacement fields. Common challenges with the two different techniques for applying the needed force can be (i) the ability to transmit enough energy through skin and subcutaneous fat to generate suf- ficient shear waves in the liver and (ii) the limitation of the distance between ribs for TE (Palmeri et al. 2008). Shear waves are transverse; the particle movements are across the direction of the waves and can be imagined as ripples on a water surface when disturbed. They are unlike ultrasound longitudinal waves, which are more rapidly attenuated in soft tissue and travel much more slowly (Bamber et al. 2013; Cosgrove et al. 2013). The positions or frictions at tissue boundaries are not known, and biological tissues are complicated and require assump- tions of a linearly elastic, homogeneous, isotropic, infinite and continuous medium. In practice, the force–deformation

(4)

relationship is usually both non-linear and time dependent; elasticity varies spatially and with direction (anisotropy); tissue boundaries and structure change the relationship between shear elastic modulus and shear wave speed because of shear wave propagation and shear wave scattering; and finally, tissue content may vary with scars, fluid collections or tumor boundaries (Bamber et al. 2013).

All of these factors together can result in artifacts and may contribute to effects on SWE measurement reliability.

Various influences on measurement reliability have been evaluated, such as breath-holding technique, patient positioning, post-prandial state and system-specific factors (Barr et al. 2016). There is also a large measurement bias that is depth dependent (Zhao et al. 2011). It is important to identify factors that may be affected by tissue and boundaries, such as body mass index (BMI), skin-to- liver capsule distance (SCD) and the presence of cirrhosis and/or steatosis in the liver, so that they can be consid- ered in a standardized examination protocol to increase the measurement reliability of ultrasound-based SWE.

Our aim was to investigate patient-related factors associated with either reliable or poorly reliable measurement results when performing ultrasound-based SWE of the liver, using the reliability criteria for liver stiffness (Boursier et al.

2013), which have not yet been applied in this context.

The investigated factors were age, sex, smoking habits, BMI, presence of cirrhosis and/or steatosis in the liver, use of cardiovascular and/or antiviral medication and SCD.

Another aim was to see how body constitution, as expressed by either BMI or SCD, would best associate with poorly reliable SWE results.

METHODS

This prospective study was approved by the Re- search Ethical Review Board in Umeå (No. 2015/355-

31). The research was performed in accordance with the World Medical Association Declaration of Helsinki, 2013. Between March 2014 and March 2017, we en- rolled a total of 261 consecutive patients with hepatitis B and C, ASH, NASH, primary biliary cholangitis (PBC) or human immunodeficiency virus (HIV) or having follow-up for methotrexate treatment, all examined for the first time using ultrasound-based SWE of the liver. A total of 73 patients were excluded as follows:

11 examinations not in accordance with the method description, 3 deaths shortly after the examination, 27 no responses to the consent letter, 25 not giving approval and 7 with wrong addresses. Thus, 188 patients (66 women and 122 men) were included after giving their written informed consent. All patients were recruited from a single radiology department in northern Sweden (Fig. 1).

The reliability criteria for liver stiffness cutoff were IQR/median>0.30 and a liver stiffness median ≥7.1 kPa, defined as poorly reliable measurement results (Boursier et al. 2013). Applying this cutoff divided the 188 patients into two groups: those with poorly reliable and those with reliable measurement results.

To study the impact on ultrasound-based SWE of patient-related factors (age, sex, smoking habits, BMI, SCD, presence of steatosis and/or cirrhosis in the liver, use of antiviral and/or cardiovascular medication [ATC codes: J05 A (DAA), C01, C07, C08, C09]), we used binary univariate and multivariate logistic regression to investigate associations with either group.

For smoking habits, BMI and SCD, data for 59 patients were analyzed, with the main aim of determining how body constitution associates with poorly reliable results. All ultrasound-based SWE measurement results were compared with liver biopsy findings, when available (n= 48).

: 11 not in accordance with method description,3 deaths shortly after the exam, 27 no- responses , 25 not giving approval, 7 with wrong adresses

IQR/median >0.30 with liver

stiffness IQR/median

Fig. 1. Flowchart for inclusion of the 188 patients. IQR= interquartile range.

1608 Ultrasound in Medicine & Biology Volume 44, Number 8, 2018

(5)

Point SWE liver examination

The ultrasound device used in the study was a Philips iU22 (Philips Healthcare, Bothell, WA, USA), applying the ElastPQ technique, pSWE and a C5 convex probe. The ElastPQ program uses ARFI as the applied force, with a focused radiation force impulse at a certain depth to generate shear waves that propagate laterally from the focus point. In the image, a fixed ROI of 0.5× 1.5 cm can be placed by moving a trackball, with the depth decided by the operator, and a single measurement can be performed by pressing a button. Tissue in the ROI is mechanically excited using short-duration pulses, which generate tissue displacements at the location. The displacements result in shear waves that propagate away laterally, and by tracking ultrasound beams at the time- to-peak displacement of each location, the speed of the shear waves can be measured and used as an indication of liver stiffness. The stiffer the liver, the faster the shear waves propagate through the tissue (Barr 2014; Palmeri et al. 2008; Shiina et al. 2015; Srinivasa Babu et al. 2016).

The shear wave speed can then be presented in meters per second or in kilopascals using Young’s modulus E= 3(vS2

xρ), where E is Young’s modulus, vSis the shear wave ve- locity andρ is the tissue density (Dietrich et al. 2017;

Goertz et al. 2012; Ling et al. 2013). Thus, in this study, the shear wave speed is expressed in kilopascals through Young’s modulus.

All patients had been fasting for 5 h. Because the liver becomes stiffer during physical activity, patients were re- quired to rest at least 10 min before the SWE examination (Gersak et al. 2016; Goertz et al. 2012; Popescu et al. 2013).

The respiratory phase is also important, and during measurement, patients had to hold their breath in a relaxed phase (Ling et al. 2013).

Measurements were performed using an intercostal approach, at liver segment VII/VIII (Goertz et al. 2010;

Liao et al. 2015), with the patient in supine position, the right arm elevated above the head (Barr et al. 2016). The measurements were performed at a depth of 3–5 cm below the liver capsule (Huang et al. 2014; Wang et al. 2014).

Whenever possible, the angle of the beam was kept at 0°, and the probe was kept perpendicular to the liver (Karlas et al. 2011). At least 10 measurements were performed, and the median result was used (Dietrich et al. 2017). The median result was then entered into the Metavir score table, which is graded from F0 to F4, with F0 representing normal liver tissue and F4 representing cirrhosis. With progress- ing fibrosis, the liver becomes stiffer and thus warrants a higher Metavir score (Barr 2014; Goodman 2007).

The presence of ascites and/or steatosis in the liver was confirmed or ruled out by means of standard B-mode ultrasound examination, comparing brightness to the right kidney for steatosis. The SCD was measured in the same image used for the SWE measurements.

Examination quality

The same ultrasound device and probe were used for all SWE liver examinations. The probe was tested regu- larly by a technician at the radiology department using the onboard program check. The patients were examined while wearing light clothing, and all weight and length measurements were taken using the same equipment. All examinations were performed by the same operator (M.B.).

Intrahepatic vessels were avoided with the help of power Doppler to identify fluid-filled vessels. The ultrasound device ElastPQ program displayed 00.00 for invalid measurements, and these were then not included in the examination. The median values for SWE liver measurement results from at least 10 measurements, expressed in kilopascals, were saved in a worksheet and printed as an image in the ultrasound device.

Statistical analyses

Statistical analysis was performed using SPSS Sta- tistics Version 23 (IBM, Armonk, NY, USA). Continuous variables were summarized as the mean± standard deviation or median (min–max) and categorical variables as the frequency (percentage). The Mann–Whitney test, St- udent’s t-test,χ²test and Fisher’s exact test were used as appropriate. Correlations between variables as well as SWE and liver biopsy results were analyzed using Spearman’s r. The data for liver stiffness were expressed as the median (IQR). The reliability criteria for liver stiffness (Boursier et al. 2013) were used as the cutoff; the group with poorly reliable measurement results was coded as 1, and the group with reliable measurement results was coded as 0.

Binary logistic stepwise regression analyses were used to assess factors independently associated with poorly reliable measurement results. Significant factors from the univariate analyses were then analyzed in a multivariate logistic regression model. Nagelkerke, Cox and Snell and Hosmer–Lemeshow tests were used to evaluate the model.

Ninety-five percent confidence intervals (CIs) were calculated for all statistical tests. A p value< 0.05 was re- garded as indicating statistical significance. All statistical tests were two-sided.

RESULTS

A total of 188 patients were included, and eight patient-related factors were tested for association with the reliable and poorly reliable measurement groups. The characteristics of all patients are provided inTable 1. Ascites was not present in the patients examined.

With the use of reliability criteria for liver stiffness (Boursier et al. 2013) as the cutoff, 52 of 188 patients (28%) were categorized into the poorly reliable measurement group and 136 of 188 (72%) were in the group with reliable measurement results.

(6)

On univariate regression analysis, among age, sex, presence of steatosis and/or cirrhosis in the liver, use of cardiovascular and/or antiviral medication and SCD, only sex and cirrhosis were not associated with poorly reliable results. Further analysis with the significant factors in the multivariate analysis revealed that the presence of steatosis in the liver and increased SCD were associated with poorly reliable results (odds ratio [OR]= 2.89 and 3.08, respectively; seeTable 2). For smoking habits, BMI and SCD, on multivariate analysis, only SCD was asso-

ciated with poorly reliable results (n= 59, OR = 6.12; see Table 3). The ultrasound-based SWE measurements exhibited a significant positive correlation (n= 48, r = 0.476, p= 0.001) with coexisting liver biopsy results (Fig. 2).

DISCUSSION

In this prospective study, eight patient-related factors were tested with respect to their impact on ultrasound- based SWE liver measurement reliability with pSWE in Table 1. Patient characteristics

Characteristic Reliable measurement (n= 136) Poorly reliable measurement (n= 52) p Value

Age (y) 45.5 (18.0–83.0)* 54.5 (22.0–71.0) 0.010^†

Sex 0.267^‡

Male 85 (62.5%) 37 (71.2%)

Female 51 (37.5%) 15 (28.8%)

Skin-to-liver capsule distance (cm) 1.8 (1.1–3.9) 2.3 (1.2–4.9) <0.001^†

Smoking habits^‖(n= 97) 0.930^‡

Non-smoker, former smoker n (%) 40 (29.4%) 20 (38.5%)

Daily smoker n (%) 25 (18.4%) 12 (23.1%)

Diagnosis 0.022^§

Primary biliary cholangitis 2 (1.5%) 0

Hepatitis C virus 72 (52.9%) 39 (75.0%)

Hepatitis B virus 45 (33.1%) 6 (11.5%)

Alcoholic steatohepatitis 4 (2.6%) 3 (5.8%)

Non-alcoholic steatohepatitis 3 (2.6%) 2 (3.8%)

Cryptogenic (of unknown cause) 9 (6.6%) 2 (3.8%)

Human immunodeficiency virus 1 (0.7%) 0

Disease duration (y)^‖(n= 136) 7.5 (0–58) 18.0 (0–54) <0.001^†

Presence of steatosis <0.001^‡

No steatosis 97 (71.3%) 18 (34.6%)

Steatosis 39 (28.7%) 34 (65.4%)

Presence of cirrhosis 0.095^§

No cirrhosis 130 (95.6%) 46 (88.4%)

Cirrhosis 6 (4.4%) 6 (11.5%)

Use of medication 0.018^§

No medication 104 (76.5%) 32 (61.5%)

Antiviral 6 (4.4%) 1 (1.9%)

Cardiovascular 26 (19.1%) 17 (32.7%)

Antiviral+ cardiovascular 0 2 (3.8%)

* Values are expressed as the median (min–max) or the number (%). P values< 0.05 indicate statistical significance and are in boldface.

†Mann–Whitney U-test.

‡χ²-test.

§Fisher’s exact test.

‖Data missing for some patients.

Table 2. Impact of patient-related factors on shear wave elastography (n= 188)

Factor β p Value* Wald Crude OR CI (95%) β pValue* Wald Adjusted OR 95% CI

Age (y) 0.03 0.01 6.00 1.03 1.00–1.05 0.01 0.44 0.27 1.01 0.98–1.04

Sex 0.39 0.27 1.23 1.48 0.74–2.96

Steatosis 1.55 <0.001 19.80 4.70 2.38–9.29 1.06 0.01 7.24 2.89 1.33–6.28

Medication^† 0.44 0.01 6.18 1.55 1.10–2.19 0.70 1.41

Cirrhosis 1.04 0.08 2.98 2.83 0.87–9.20

SCD (cm) 1.31 <0.001 22.04 3.70 2.14–6.37 1.12 <0.001 13.65 3.08 1.70–5.60

Nagelkerke R²= 28.4%

Cox and Snell R²= 19.7%

Hosmer–Lemeshow p= 0.73

CI= confidence interval; OR = odds ratio; SCD = skin-to-liver capsule distance.

* p Values< 0.05 indicate statistical significance and are in boldface.

†Antiviral, cardiovascular.

(7)

a total of 188 patients. The reliability criteria for liver stiffness (Boursier et al. 2013) were used as the cutoff in binary logistic regression analysis.

Ultrasound-based SWE exhibited a significant positive correlation with liver biopsy. In addition, on univariate analysis, sex, age, BMI, use of cardiovascular and/or antiviral medication, presence of cirrhosis in the liver and smoking habits were not associated with poorly reliable results. On multivariate analysis, presence of liver steatosis and increased SCD were associated with an increased likelihood of poorly reliable results. Body constitution appears to have been better measured with SCD than with BMI, as the latter exhibited no association with poorly re-

liable measurements. To our knowledge, this study is the first to use the reliability criteria for liver stiffness (Boursier et al. 2013) as a cutoff in binary logistic regression analysis of ultrasound-based SWE of the liver with push pulse as applied force.

The study results are relevant because obesity result- ing in steatosis and an increased amount of subcutaneous fat between the probe and liver constitutes a risk for diagnostic error. Thus, establishing a reliable ultrasound- based SWE method to identify patients with significant fibrosis is warranted to initiate treatment in those with hepatitis, thus avoiding progression to hepatocellular carcinoma.

Also, in non-alcoholic fatty liver disease (NAFLD), out- Table 3. Impact of BMI, SCD and smoking habits on shear wave elastography (n= 59)

Factor β p Value* Wald Adjusted OR 95% CI

Age (y) −0.01 0.61 0.26 0.99 0.95–1.03

Sex −0.86 0.30 1.07 0.42 0.08–2.16

Smoking habits^† 0.87 0.24 1.38 2.39 0.56–10.26

BMI (kg/m²) −0.02 0.84 0.04 0.98 0.80–1.20

SCD (cm) 1.81 0.01 6.33 6.12 1.50–25.10

Nagelkerke R²= 31.5%

Cox and Snell R²= 22.8%

Hosmer–Lemeshow p= 0.20

BMI= body mass index; CI = confidence interval; OR = odds ratio; SCD = skin-to-liver capsule distance.

* p values< 0.05 indicate statistical significance and are in boldface.

†Daily and former smoker versus non-smoker.

Fig. 2. Correlation of shear wave elastography measurement results with liver biopsy results, in Metavir score. SWE= shear wave elastography.

(8)

comes can include inflammation, fibrosis, cirrhosis and, in the worst case, hepatocellular carcinoma (Augustin et al.

2017).

With ARFI, it is possible to measure liver tissue displacements. In addition, ARFI can be performed in complex media, without significant artifacts originating from nearby structures, and therefore can directly correlate with liver stiffness (Nightingale et al. 2001).

The current results diverge in part from previous findings. Most of the participants were men, and HCV was the dominant cause of liver disease (59% of patients). Sex was not associated with poorly reliable results, although earlier studies had reported significantly higher values for men (Corpechot et al. 2006; Huang et al.

2014). The findings in general have been mixed. In one report on intra-class correlations for 186 patients, BMI, sex and age did not influence concordance (Fraquelli et al. 2016). Yet another analysis of data for 1031 participants found an association of older age, being male and BMI with a risk for failed and unreliable measurements (Bota et al. 2014).

To date, results with BMI also have varied. For example, in a study with 135 participants, BMI was significantly higher in the unreliable group (Maruyama et al.

2016), whereas it was not associated with less reliable results in another investigation (Huang et al. 2014). Fur- thermore, in a small study with healthy volunteers, BMI was not significantly correlated with liver elasticity (Fang et al. 2017). Why BMI did not associate with the poorly reliable measurement results in this study but SCD did is a somewhat confusing outcome that needs to be ad- dressed in future studies. Few reports have compared the two while also including a larger number of participants.

Recently published results for 55 patients, comparable to the 59 in the current work, indicated that SCD was more significant for increased measurement variability compared to BMI (Nadebaum et al. 2018). When using ultrasound-based SWE, SCD could therefore be a candi- date for replacing BMI to achieve greater accuracy because an increased layer of subcutaneous fat can have a tremen- dous impact on ultrasound imaging, both in general and also for SWE.

Another little-studied factor that affects boundaries is increased waist circumference, which together with increased BMI has exhibited a significant correlation with the standard deviation-to-mean stiffness value ratio in a study of 605 patients (Varbobitis et al. 2016). Increased SCD is also a problem for TE performed with FibroScan;

the manufacturer Echosens does not recommend measurements for SCD>3.5 cm with the XL probe. There is a broad consensus on not using the XL probe because measurements with it have been less reliable (Dietrich et al.

2017), which means that the M probe is the better choice with the limit of SCD<2.5 cm, according to Echosens.

In this study, the presence of steatosis in the liver was associated with poorly reliable results, which also was confirmed in a recent study using pSWE technology (Conti et al. 2017); thus, as for TE, there are problems with the accuracy of measurement results, with a risk for misclassification in the presence of obesity. Neverthe- less, poorly reliable results from other studies using pSWE and 2-D SWE exhibited no correlation with steatosis (Ferraioli et al. 2014a; Samir et al. 2015), which means that liver steatosis has a somewhat doubtful impact on reliability. It could possibly reflect the difficulties involved in performing SWE in obese patients, which is paired with the increased SCD because of increased subcutaneous fat.

Earlier studies confirmed that the rate of reliable measurements is significantly higher for more experienced operators when doing examinations in obese patients (Gradinaru-Tascau et al. 2013), indicating that training can help (Ferraioli et al. 2014b). Nevertheless, the presence of liver steatosis does give the operator an indication of concern with respect to reliability.

One limitation of this study is possibly the use of the reliability criteria for liver stiffness (Boursier et al. 2013) as a cutoff because other reports have indicated discrimi- nant cutoff values of 21% (Lucidarme et al. 2009).

Nevertheless, another study involving ultrasound-based SWE with push pulse revealed discordance with an IQR/

median≥30% (Bota et al. 2013). Therefore, even though TE and ultrasound-based SWE do not use the same applied force to create shear waves, we hypothesize that the reliability criteria for liver stiffness (Boursier et al. 2013) for TE are also valid for ultrasound-based SWE, which needs to be confirmed in future studies.

In a recent report, the standard deviation for the SWE mean was identified to be the strongest factor associated with the quality of shear wave propagation (Zhou et al.

2017). Studies have also found that technical failures are not machine dependent, and for high SWE reliability, ultrasound devices should not be used interchangeably (Woo et al. 2015). With the aid of reliability criteria for liver stiffness (Boursier et al. 2013), displayed on the ultrasound machine monitor during ultrasound-based SWE of the liver, poorly reliable measurement results can be avoided.

However, no guidelines for ultrasound-based SWE are cur- rently in place to describe how to achieve an IQR/

median ≤30%. One way is for the operator to delete uncertain measurements, including outliers, while performing the examination to maintain the IQR/median

≤30%. The IQR/median values are available on the monitor for some ultrasound devices, although not displayed in real time for all of them (Piscaglia et al. 2016).

Future feasibility solutions

Using the reliability criteria for liver stiffness (Boursier et al. 2013) as a cutoff in this study resulted in

(9)

approximately 28% (52/188) of SWE results being poorly reliable, primarily because of the increasing distance between the probe and liver capsule with increasing subcutaneous fat. Reducing the distance by applying pressure with the probe would be of great value; however, the rec- ommendation is to perform the examination with minimal scanning pressure (Dietrich et al. 2017) because probe pressure is thought to result in falsely high measurement values or greater measurement variance (Cui et al. 2013). Nev- ertheless, some authors have recommended applying force with the probe when necessary (Cosgrove et al. 2013), although this suggestion has not been confirmed in a controlled study. Another way of reducing the distance is to place the patient in a left decubitus position (Goertz et al.

2012), but one group has reported that SWE examinations performed in positions other than supine produce less accurate results (Liao et al. 2015). How to perform an accurate SWE examination in obese patients without negatively affecting measurement values will have to be determined in future studies.

CONCLUSIONS

Ultrasound-based SWE values exhibit a significant, positive correlation with liver biopsy results. When using the SWE method, application of the IQR/median as a quality parameter helps to avoid up to 28% of poor SWE results for patients with an increased layer of subcutaneous fat between the probe and liver capsule or with liver steatosis. These results also suggest that SCD should be used in preference to BMI to maintain a reliable SWE results. The patient-related factors of sex, age, smoking habits, use of antiviral and/or cardiovascular medication, BMI and presence of cirrhosis in the liver did not associate with poorly reliable SWE results.

Acknowledgments—We express our sincere appreciation to Associate Professor Ewa Lundgren, M.D., for initiating the research conducted by Ph.D. student M.B., and to senior statistician Lars Söderström for excellent statistical support. Thanks also to Elira Maksuti, Ph.D., KTH Royal Institute of Technology, for technical support. We also thank all partici- pating patients.

REFERENCES

Abdi W, Millan JC, Mezey E. Sampling variability on percutaneous liver biopsy. Arch Intern Med 1979;139:667–669.

Afdhal NH. Diagnosing fibrosis in hepatitis C: Is the pendulum swinging from biopsy to blood tests? Hepatology 2003;37:972–

974.

Armstrong MJ, Corbett C, Hodson J, Marwah N, Parker R, Houlihan DD, Rowe IA, Hazlehurst JM, Brown R, Hubscher SG, Mutimer D.

Operator training requirements and diagnostic accuracy of FibroScan in routine clinical practice. Postgrad Med J 2013;89:685–692.

Augustin S, Graupera I, Caballeria J. Non-alcoholic fatty liver disease:

A poorly known pandemic. Med Clin (Barc) 2017;149:542–548.

Bamber J, Cosgrove D, Dietrich CF, Fromageau J, Bojunga J, Calliada F, Cantisani V, Correas JM, D’Onofrio M, Drakonaki EE, Fink M, Friedrich-Rust M, Gilja OH, Havre RF, Jenssen C, Klauser AS,

Ohlinger R, Saftoiu A, Schaefer F, Sporea I, Piscaglia F. EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography: Part 1. Basic principles and technology. Ultraschall Med 2013;34:169–184.

Barr RG. Elastography in clinical practice. Radiol Clin North Am 2014;

52:1145–1162.

Barr RG, Ferraioli G, Palmeri ML, Goodman ZD, Garcia-Tsao G, Rubin J, Garra B, Myers RP, Wilson SR, Rubens D, Levine D. Elastography assessment of liver fibrosis: Society of Radiologists in ultrasound consensus conference statement. Ultrasound Q 2016;32:94–107.

Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003;38:1449–1457.

Beland MD, Brown SF, Machan JT, Taliano RJ, Promrat K, Cronan JJ.

A pilot study estimating liver fibrosis with ultrasound shear-wave elastography: Does the cause of liver disease or location of measurement affect performance? AJR Am J Roentgenol 2014;203:

W267–W273.

Bota S, Sporea I, Sirli R, Popescu A, Danila M, Costachescu D. Intra- and interoperator reproducibility of acoustic radiation force impulse (ARFI) elastography—Preliminary results. Ultrasound Med Biol 2012;

38:1103–1108.

Bota S, Sporea I, Sirli R, Popescu A, Jurchis A. Factors which influence the accuracy of acoustic radiation force impulse (ARFI) elastography for the diagnosis of liver fibrosis in patients with chronic hepatitis C. Ultrasound Med Biol 2013;39:407–412.

Bota S, Sporea I, Sirli R, Popescu A, Danila M, Jurchis A, Gradinaru-Tascau O. Factors associated with the impossibility to obtain reliable liver stiffness measurements by means of acoustic radiation force impulse (ARFI) elastography—Analysis of a cohort of 1,031 subjects. Eur J Radiol 2014;83:268–272.

Boursier J, Zarski JP, de Ledinghen V, Rousselet MC, Sturm N, Lebail B, Fouchard-Hubert I, Gallois Y, Oberti F, Bertrais S, Cales P. De- termination of reliability criteria for liver stiffness evaluation by transient elastography. Hepatology 2013;57:1182–1191.

Bravo AA, Sheth SG, Chopra S. Liver biopsy. N Engl J Med 2001;344:

495–500.

Canbay A, Gerken G, Bechmann LP. Hepatic steatosis: Differential di- agnostics and current aspects]. Internist (Berl) 2016;57:879–892.

Castera L, Foucher J, Bernard PH, Carvalho F, Allaix D, Merrouche W, Couzigou P, de Ledinghen V. Pitfalls of liver stiffness measurement:

A 5-year prospective study of 13,369 examinations. Hepatology 2010;

51:828–835.

Conti F, Serra C, Vukotic R, Fiorini E, Felicani C, Mazzotta E, D’Errico A, Verucchi G, Lenzi M, Andreone P. Accuracy of elastography point quantification and steatosis influence on assessing liver fibrosis in patients with chronic hepatitis C. Liver Int 2017;37:187–195.

Corpechot C, El Naggar A, Poupon R. Gender and liver: Is the liver stiffness weaker in weaker sex? Hepatology 2006;44:513–514.

Cosgrove D, Piscaglia F, Bamber J, Bojunga J, Correas JM, Gilja OH, Klauser AS, Sporea I, Calliada F, Cantisani V, D’Onofrio M, Drakonaki EE, Fink M, Friedrich-Rust M, Fromageau J, Havre RF, Jenssen C, Ohlinger R, Saftoiu A, Schaefer F, Dietrich CF. EFSUMB guidelines and recommendations on the clinical use of ultrasound elastography: Part 2. Clinical applications. Ultraschall Med 2013;

34:238–253.

Cui XW, Friedrich-Rust M, De Molo C, Ignee A, Schreiber-Dietrich D, Dietrich CF. Liver elastography, comments on EFSUMB elastography guidelines 2013. World J Gastroenterol 2013;19:6329–6347.

Dietrich CF, Bamber J, Berzigotti A, Bota S, Cantisani V, Castera L, Cosgrove D, Ferraioli G, Friedrich-Rust M, Gilja OH, Goertz RS, Karlas T, de Knegt R, de Ledinghen V, Piscaglia F, Procopet B, Saftoiu A, Sidhu PS, Sporea I, Thiele M. EFSUMB guidelines and recommendations on the clinical use of liver ultrasound elastography, update 2017 (long version). Ultraschall Med 2017;38:

e16–e47.

European Association for the Study of the Liver (EASL). EASL 2017 clinical practice guidelines on the management of hepatitis B virus infection. J Hepatol 2017;67:370–398.

Fang C, Konstantatou E, Romanos O, Yusuf GT, Quinlan DJ, Sidhu PS.

Reproducibility of 2-dimensional shear wave elastography assessment of the liver: A direct comparison with point shear wave

(10)

elastography in healthy volunteers. J Ultrasound Med 2017;36:1563–

1569.

Ferraioli G, Tinelli C, Lissandrin R, Zicchetti M, Dal Bello B, Filice G, Filice C. Point shear wave elastography method for assessing liver stiffness. World J Gastroenterol 2014a;20:4787–4796.

Ferraioli G, Tinelli C, Lissandrin R, Zicchetti M, Bernuzzi S, Salvaneschi L, Filice C. Ultrasound point shear wave elastography assessment of liver and spleen stiffness: Effect of training on repeatability of measurements. Eur Radiol 2014b;24:1283–1289.

Fraquelli M, Baccarin A, Casazza G, Conti CB, Giunta M, Massironi S, Invernizzi F, Donato MF, Maggioni M, Aghemo A, Conte D, Colombo M. Liver stiffness measurement reliability and main de- terminants of point shear-wave elastography in patients with chronic liver disease. Aliment Pharmacol Ther 2016;44:356–365.

Freeman AJ, Dore GJ, Law MG, Thorpe M, Von Overbeck J, Lloyd AR, Marinos G, Kaldor JM. Estimating progression to cirrhosis in chronic hepatitis C virus infection. Hepatology 2001;34:809–816.

Friedrich-Rust M, Wunder K, Kriener S, Sotoudeh F, Richter S, Bojunga J, Herrmann E, Poynard T, Dietrich CF, Vermehren J, Zeuzem S, Sarrazin C. Liver fibrosis in viral hepatitis: Noninvasive assessment with acoustic radiation force impulse imaging versus transient elastography. Radiology 2009;252:595–604.

Friedrich-Rust M, Nierhoff J, Lupsor M, Sporea I, Fierbinteanu-Braticevici C, Strobel D, Takahashi H, Yoneda M, Suda T, Zeuzem S, Herrmann E. Performance of acoustic radiation force impulse imaging for the staging of liver fibrosis: A pooled meta-analysis. J Viral Hepat 2012;

19:e212–e219.

Gersak MM, Sorantin E, Windhaber J, Dudea SM, Riccabona M. The influence of acute physical effort on liver stiffness estimation using Virtual Touch Quantification (VTQ): Preliminary results. Med Ultrason 2016;18:151–156.

Goertz RS, Zopf Y, Jugl V, Heide R, Janson C, Strobel D, Bernatik T, Haendl T. Measurement of liver elasticity with acoustic radiation force impulse (ARFI) technology: An alternative noninvasive method for staging liver fibrosis in viral hepatitis. Ultraschall Med 2010;31:

151–155.

Goertz RS, Egger C, Neurath MF, Strobel D. Impact of food intake, ultrasound transducer, breathing maneuvers and body position on acoustic radiation force impulse (ARFI) elastometry of the liver.

Ultraschall Med 2012;33:380–385.

Goodman ZD. Grading and staging systems for inflammation and fibrosis in chronic liver diseases. J Hepatol 2007;47:598–607.

Gradinaru-Tascau O, Sporea I, Bota S, Jurchis A, Popescu A, Popescu M, Sirli R, Szilaski M. Does experience play a role in the ability to perform liver stiffness measurements by means of supersonic shear imaging (SSI)? Med Ultrason 2013;15:180–183.

Huang Z, Zheng J, Zeng J, Wang X, Wu T, Zheng R. Normal liver stiffness in healthy adults assessed by real-time shear wave elastography and factors that influence this method. Ultrasound Med Biol 2014;

40:2549–2555.

Hudson JM, Milot L, Parry C, Williams R, Burns PN. Inter- and intra- operator reliability and repeatability of shear wave elastography in the liver: a study in healthy volunteers. Ultrasound Med Biol 2013;

39:950–955.

Joshi-Barve S, Kirpich I, Cave MC, Marsano LS, McClain CJ. Alco- holic, nonalcoholic, and toxicant-associated steatohepatitis: Mechanistic similarities and differences. Cell Mol Gastroenterol Hepatol 2015;

1:356–367.

Karlas T, Pfrepper C, Wiegand J, Wittekind C, Neuschulz M, Mossner J, Berg T, Troltzsch M, Keim V. Acoustic radiation force impulse imaging (ARFI) for non-invasive detection of liver fibrosis: Exam- ination standards and evaluation of interlobe differences in healthy subjects and chronic liver disease. Scand J Gastroenterol 2011;46:

1458–1467.

Kettaneh A, Marcellin P, Douvin C, Poupon R, Ziol M, Beaugrand M, de Ledinghen V. Features associated with success rate and performance of FibroScan measurements for the diagnosis of cirrhosis in HCV patients: A prospective study of 935 patients. J Hepatol 2007;

46:628–634.

Lagging M, Wejstal R, Norkrans G, Karlstrom O, Aleman S, Weiland O, Castedal M, Westin J. Treatment of hepatitis C virus infection:

Updated Swedish guidelines 2016. Infect Dis (Lond) 2017;49:561–

575.

Lee R. General principles. St. Louis, MO: Mosby-Year Book; 1994. p.

1–21.

Lee TY, Wu JC, Yu SH, Lin JT, Wu MS, Wu CY. The occurrence of hepatocellular carcinoma in different risk stratifications of clinical- ly non-cirrhotic non-alcoholic fatty liver disease. Int J Cancer 2017;

141:1307–1314.

Liao LY, Kuo KL, Chiang HS, Lin CZ, Lin YP, Lin CL. Acoustic radiation force impulse elastography of the liver in healthy patients:

Test location, reference range and influence of gender and body mass index. Ultrasound Med Biol 2015;41:698–704.

Ling W, Lu Q, Quan J, Ma L, Luo Y. Assessment of impact factors on shear wave based liver stiffness measurement. Eur J Radiol 2013;

82:335–341.

Lucidarme D, Foucher J, Le Bail B, Vergniol J, Castera L, Duburque C, Forzy G, Filoche B, Couzigou P, de Ledinghen V. Factors of accuracy of transient elastography (FibroScan) for the diagnosis of liver fibrosis in chronic hepatitis C. Hepatology 2009;49:1083–

1089.

Lupsor M, Badea R, Stefanescu H, Sparchez Z, Branda H, Serban A, Maniu A. Performance of a new elastographic method (ARFI technology) compared to unidimensional transient elastography in the noninvasive assessment of chronic hepatitis C: Preliminary results.

J Gastrointestin Liver Dis 2009;18:303–310.

Maruyama H, Kobayashi K, Kiyono S, Sekimoto T, Kanda T, Yokosuka O. Two-dimensional shear wave elastography with propagation- based reliability assessment for grading hepatic fibrosis and portal hypertension. J Hepatobiliary Pancreat Sci 2016;23:595–602.

Nadebaum DP, Nicoll AJ, Sood S, Gorelik A, Gibson RN. Variability of liver shear wave measurements using a new ultrasound elastographic technique. J Ultrasound Med 2018;37:647–656.

Nightingale K, Soo MS, Nightingale R, Trahey G. Acoustic radiation force impulse imaging: In vivo demonstration of clinical feasibility. Ultrasound Med Biol 2002;28:227–235.

Nightingale KR, Palmeri ML, Nightingale RW, Trahey GE. On the feasibility of remote palpation using acoustic radiation force. J Acoust Soc Am 2001;110:625–634.

Palmeri ML, Wang MH, Dahl JJ, Frinkley KD, Nightingale KR. Quan- tifying hepatic shear modulus in vivo using acoustic radiation force.

Ultrasound Med Biol 2008;34:546–558.

Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications follow- ing percutaneous liver biopsy: A multicentre retrospective study on 68,276 biopsies. J Hepatol 1986;2:165–173.

Piscaglia F, Salvatore V, Mulazzani L, Cantisani V, Schiavone C. Ul- trasound shear wave elastography for liver disease: A critical appraisal of the many actors on the stage. Ultraschall Med 2016;37:1–5.

Popescu A, Bota S, Sporea I, Sirli R, Danila M, Racean S, Suseanu D, Gradinaru O, Ivascu Siegfried C. The influence of food intake on liver stiffness values assessed by acoustic radiation force impulse elastography—Preliminary results. Ultrasound Med Biol 2013;39:

579–584.

Regev A, Berho M, Jeffers LJ, Milikowski C, Molina EG, Pyrsopoulos NT, Feng ZZ, Reddy KR, Schiff ER. Sampling error and intraobserver variation in liver biopsy in patients with chronic HCV infection. Am J Gastroenterol 2002;97:2614–2618.

Samir AE, Dhyani M, Vij A, Bhan AK, Halpern EF, Mendez-Navarro J, Corey KE, Chung RT. Shear-wave elastography for the estimation of liver fibrosis in chronic liver disease: Determining accuracy and ideal site for measurement. Radiology 2015;274:888–896.

Sandrin L, Fourquet B, Hasquenoph JM, Yon S, Fournier C, Mal F, Christidis C, Ziol M, Poulet B, Kazemi F, Beaugrand M, Palau R.

Transient elastography: A new noninvasive method for assessment of hepatic fibrosis. Ultrasound Med Biol 2003;29:1705–1713.

Schiavon LD, Narciso-Schiavon JL, de Carvalho-Filho RJ. Non- invasive diagnosis of liver fibrosis in chronic hepatitis C. World J Gastroenterol 2014;20:2854–2866.

Shiina T, Nightingale KR, Palmeri ML, Hall TJ, Bamber JC, Barr RG, Castera L, Choi BI, Chou YH, Cosgrove D, Dietrich CF, Ding H, Amy D, Farrokh A, Ferraioli G, Filice C, Friedrich-Rust M, Nakashima K, Schafer F, Sporea I, Suzuki S, Wilson S, Kudo M.

(11)

WFUMB guidelines and recommendations for clinical use of ultrasound elastography: Part 1. Basic principles and terminology.

Sporea I, Bota S, Peck-Radosavljevic M, Sirli R, Tanaka H, Iijima H, Badea R, Lupsor M, Fierbinteanu-Braticevici C, Petrisor A, Saito H, Ebinuma H, Friedrich-Rust M, Sarrazin C, Takahashi H, Ono N, Piscaglia F, Borghi A, D’Onofrio M, Gallotti A, Ferlitsch A, Popescu A, Danila M. Acoustic radiation force impulse elastography for fibrosis evaluation in patients with chronic hepatitis C: An international multicenter study. Eur J Radiol 2012;81:4112–4118.

Srinivasa Babu A, Wells ML, Teytelboym OM, Mackey JE, Miller FH, Yeh BM, Ehman RL, Venkatesh SK. Elastography in chronic liver disease: Modalities, techniques, limitations, and future directions.

Radiographics 2016;36:1987–2006.

Varbobitis IC, Siakavellas SI, Koutsounas IS, Karagiannakis DS, Ioannidou P, Papageorgiou MV, Pavlopoulou ID, Schizas D, Bamias G, Vlachogiannakos I, Ladas SD, Papatheodoridis GV. Reliability and applicability of two-dimensional shear-wave elastography for the eval-

uation of liver stiffness. Eur J Gastroenterol Hepatol 2016;28:1204–

1209.

Wang CZ, Zheng J, Huang ZP, Xiao Y, Song D, Zeng J, Zheng HR, Zheng RQ. Influence of measurement depth on the stiffness assessment of healthy liver with real-time shear wave elastography.

Woo H, Lee JY, Yoon JH, Kim W, Cho B, Choi BI. Comparison of the reliability of acoustic radiation force impulse imaging and supersonic shear imaging in measurement of liver stiffness. Radiology 2015;

277:881–886.

Zhao H, Song P, Urban MW, Kinnick RR, Yin M, Greenleaf JF, Chen S. Bias observed in time-of-flight shear wave speed measurements using radiation force of a focused ultrasound beam. Ultrasound Med Biol 2011;37:1884–1892.

Zhou BG, Wang D, Ren WW, Li XL, He YP, Liu BJ, Wang Q, Chen SG, Alizad A, Xu HX. Value of shear wave arrival time contour display in shear wave elastography for breast masses diagnosis. Sci Rep 2017;7:7036.