Liver Ultrasound Elastography in Non-Alcoholic Fatty Liver Disease: A State-of-the-Art Summary

Non-alcoholic fatty liver disease (NAFLD) is a chronic disease which is currently the most common hepatic disorder affecting up to 38% of the general population with differences according to age, country, ethnicity and sex. Both genetic and acquired risk factors such as a high-calorie diet or high intake of saturated fats have been associated with obesity, diabetes and, finally, NAFLD. A liver biopsy has always been considered essential for the diagnosis of NAFLD; however, due to several limitations such as the potential occurrence of major complications, sampling variability and the poor repeatability in clinical practice, it is considered an imperfect option for the evaluation of liver fibrosis over time. For these reasons, a non-invasive assessment by serum biomarkers and the quantification of liver stiffness is becoming the new frontier in the management of patients with NAFLD and liver fibrosis. We present a state-of-the-art summary addressing the methods for the non-invasive evaluation of liver fibrosis in NAFLD patients, particularly the ultrasound-based techniques (transient elastography, ARFI techniques and strain elastography) and their optimal cut-off values for the staging of liver fibrosis.


Introduction
Non-alcoholic fatty liver disease (NAFLD) is the most common chronic liver disease worldwide [1].
The overall prevalence of NAFLD in the general population has been increased over time from 25% before 2005 to 38% in 2016 and later, and it is significantly higher (up to 80%) in the diabetic population [2][3][4].
NAFLD is histologically defined as the abnormal accumulation of fat (>5%) in the hepatocytes in the absence of secondary causes of fatty liver disease, such as significant alcohol consumption or viral infection. The severity of liver steatosis is graded as S0 when histological involvement is minimal (<5% of hepatocytes), S1 or mild (5-33% of hepatocytes), S2 or moderate (34-66%) and, finally, S3 or severe when >66% of liver cells show the intrahepatic accumulation of lipids [5].
NAFLD is a term which encompasses the simple and benign deposition of fat in hepatocytes to more progressive and aggressive NASH (non-alcoholic steatohepatitis) characterized by hepatitis and fibrosis up to cirrhosis and hepatocellular carcinoma (HCC) [6].
Despite the rising interest and gains in our understanding of NAFLD/NASH pathogenesis over the last two decades, there has been some dissatisfaction with the terminology Diagnostics 2023, 13, 1236 2 of 18 "non-alcoholic" because it overstresses the role of alcohol and plays down the role of metabolic risk factors in the development of the disease. Therefore, a name change from NAFLD to metabolic-associated fatty liver disease (MAFLD) has been proposed, even if it is still considered unsatisfactory because it gives more relevance to the metabolic risk factors without addressing the physiopathology of this liver disease. This is why some authors have suggested to use the term MAFLD with caution because changing the name without understanding its implications could have a negative impact on the field [8].
The diagnosis of NAFLD is based on the evidence of hepatic steatosis either by imaging or histology, whereas liver biopsy is always required for the diagnosis of NASH [9].
NASH is distinguished from isolated and benign hepatic steatosis (non-alcoholic fatty liver, NAFL) by the presence of hepatocellular injury, lobular inflammation and hepatocellular ballooning with or without liver fibrosis [10].
In 25% of patients, NAFLD progresses to NASH, which associates with an increased risk of cirrhosis and hepatocellular carcinoma [11]. In patients with NASH, the stage of fibrosis is the most important determinant of liver-related progression and mortality [12], and this is why the assessment of liver fibrosis is a key step for the optimal management of patients with NAFLD.
Liver biopsy has always been considered the gold standard for staging liver fibrosis; however, it is an invasive procedure associated with rare but potentially major complications [13], and sampling errors may easily occur due to the limited parenchymal tissue (1/50,000th of the total liver mass) [14].
Moreover, the error rate in staging fibrosis occurs in 20% of cases, cirrhosis is not correctly diagnosed in 20% of samples and the interobserver variability is about 10% [15][16][17].
All of these issues make liver biopsy an imperfect option for evaluating liver fibrosis [18].
Therefore, several methods have been studied for the non-invasive assessment of liver fibrosis, including serum biomarkers and the quantification of liver stiffness (LS).
Among the different techniques for the non-invasive assessment of liver stiffness in chronic liver disease, ultrasound elastography has given the most important contribution to the non-invasive assessment of liver fibrosis in NAFLD patients and several manuscripts have been published in the last years. Therefore, we performed a review including the most recent literature addressing this hot topic in hepatology.

Ultrasound-Based Techniques for the Assessment of Liver Fibrosis: Principles and Systems
Non-invasive methods for the assessment of liver fibrosis are one of the fields that has most rapidly evolved in the last years due to the limitations of liver biopsy and the need for several re-evaluations of liver fibrosis over time.
Blood markers, as a non-invasive method for the staging of liver fibrosis, have low accuracy in discriminating among intermediate stages of fibrosis because several extrahepatic conditions may interfere or significantly influence them [14,19].
Additionally, the liver is a large parenchymal organ readily accessible to US scanning and elastography measurements [20] and, moreover, these methods can also be easily repeated without the risk of complications or need for post-procedure hospitalization [14]. This is why the interest in non-invasive ultrasound-based techniques has increased progressively as shown by the current medical literature.
These techniques, which differ from the physical approach used, include three major types of non-invasive methods: transient elastography (TE), which uses a mechanical external push; acoustic radiation force impulse (ARFI) techniques, which use an acoustic internal push; and, finally, the strain elastography (SE) based on the tissue deformation caused by pressing the body surface or by using the internal physiologic movements (breathing and heartbeat). The ARFI techniques can be divided into point shear wave elastography (p-SWE) and 2D-shear wave elastography (2D SWE) techniques [21]. TE and ARFI techniques measure the speed of shear waves in tissues [20]. Transient elastography (TE) is a 1D technique performed with the FibroScan (Echosens, Paris, France). Fibroscan has three types of probes (S, M, XL) which work with different ultrasound frequencies. The S probe, for the pediatric population, uses a frequency of 5.0 MHz for measurements at a depth of between 1.5 and 5.0 cm from the body surface. The M probe uses an ultrasound frequency of 3.5 MHz (depth from 2.5 to 6.5 cm from the skin) and, finally, the XL probe is able to reach a depth of 3.5-7.5 cm thanks to an ultrasound frequency of 2.5 MHz [21].
The ARFI techniques include the point shear wave elastography (p-SWE) ( Figure 1) and 2D-shear wave elastography (2-D SWE) (Figure 2) techniques. Both of these measure the speed of shear waves in tissues generated by a push pulse of a focused ultrasound beam. The shear wave speed expressed in m/s is then converted into kilopascals, the unit of Young's modulus E (3rv2, where r is the tissue density and v is the speed of the shear wave), assuming that the tissue is elastic, that the tissue density is a constant value (1000 kg/m) and that the elastic modulus is not influenced by the frequency and direction of the applied force [21,22]. tography (p-SWE) and 2D-shear wave elastography (2D SWE) techniques [21]. TE and ARFI techniques measure the speed of shear waves in tissues [20].
Transient elastography (TE) is a 1D technique performed with the FibroScan (Echosens, Paris, France). Fibroscan has three types of probes (S, M, XL) which work with different ultrasound frequencies. The S probe, for the pediatric population, uses a frequency of 5.0 MHz for measurements at a depth of between 1.5 and 5.0 cm from the body surface. The M probe uses an ultrasound frequency of 3.5 MHz (depth from 2.5 to 6.5 cm from the skin) and, finally, the XL probe is able to reach a depth of 3.5-7.5 cm thanks to an ultrasound frequency of 2.5 MHz [21].
The ARFI techniques include the point shear wave elastography (p-SWE) ( Figure 1) and 2D-shear wave elastography (2-D SWE) (Figure 2) techniques. Both of these measure the speed of shear waves in tissues generated by a push pulse of a focused ultrasound beam. The shear wave speed expressed in m/s is then converted into kilopascals, the unit of Young's modulus E (3rv2, where r is the tissue density and v is the speed of the shear wave), assuming that the tissue is elastic, that the tissue density is a constant value (1000 kg/m) and that the elastic modulus is not influenced by the frequency and direction of the applied force [21,22].
The stimulation is performed at a definite depth and generates shear waves that propagate perpendicularly to the axis of the push pulse. The shear wave velocity is measured in a definite region of interest (ROI) chosen by the operator during the B-mode examination. The speed of shear waves is assessed by measuring the time required to reach a specific point starting from an opposite point (lateral side) of the region of interest. The higher the shear wave velocity, the higher the tissue stiffness [22]. The shear wave velocity is measured in the region of interest (ROI) marked by the white line, which can be moved on the screen. Measurements may be expressed as m/s or KPa. Ten measurements should be obtained from 10 independent images, in the same location. The final result should be expressed as the median together with the IQR/M. IQR/M should be 30% of the 10 measurements expressed as kilopascals and 15% for measurements expressed as meters per second. Measurement should be taken at least 15-20 mm below liver capsule [23].
The 2D-SWE method is based on the combination of acoustic radiation force by focused ultrasonic beams and very high frame ultrasound imaging. The location of the regions of interest (ROIs) and their size can be chosen by the operator before obtaining the stiffness value [21]. A real-time 2D color map (elastogram) is superimposed on the B-mode image and most vendors provide a confidence map to confirm the quality of the recorded The shear wave velocity is measured in the region of interest (ROI) marked by the white line, which can be moved on the screen. Measurements may be expressed as m/s or KPa. Ten measurements should be obtained from 10 independent images, in the same location. The final result should be expressed as the median together with the IQR/M. IQR/M should be ≤30% of the 10 measurements expressed as kilopascals and ≤15% for measurements expressed as meters per second. Measurement should be taken at least 15-20 mm below liver capsule [23].
The stimulation is performed at a definite depth and generates shear waves that propagate perpendicularly to the axis of the push pulse. The shear wave velocity is measured in a definite region of interest (ROI) chosen by the operator during the B-mode examination. The speed of shear waves is assessed by measuring the time required to reach a specific point starting from an opposite point (lateral side) of the region of interest. The higher the shear wave velocity, the higher the tissue stiffness [22].
The 2D-SWE method is based on the combination of acoustic radiation force by focused ultrasonic beams and very high frame ultrasound imaging. The location of the regions of interest (ROIs) and their size can be chosen by the operator before obtaining the stiffness value [21]. A real-time 2D color map (elastogram) is superimposed on the B-mode image and most vendors provide a confidence map to confirm the quality of the recorded values. Three or five measurements are required if the US system has a map confirming that the area chosen for the assessment of liver fibrosis has high-quality shear waves.
values. Three or five measurements are required if the US system has a map confirming that the area chosen for the assessment of liver fibrosis has high-quality shear waves.
After recording the stiffness values, the average, the standard deviation, the minimum and maximum stiffness values can be obtained. The standard deviation within the ROI reports the variability of the pixel measurements within the ROI and it is not a measure of the quality of the measurement [21]. The color-coded confidence map (left) reflects the quality of shear waves. The real-time 2D quantitative color map, namely elastogram (right), is superimposed on the B-mode. Even if ten stiffness values obtained from 10 independent images in the same location are generally required for liver ultrasound elastography, when a quality assessment parameter (confidence map) is used, five measurements may be appropriate. The ROI should be positioned at least 15-20 mm below the liver capsule [23].
The strain elastography (SE) technique ( Figure 3) uses frame-to-frame differences (tissue deformation) with stress caused after pressing the body surface with the probe or using the internal physiologic motion.
Using an intercostal space as a window, the strain image is obtained by the pressure exerted on the liver by cardiac movement. The patient should stop breathing in mid-inspiration/expiration during the acquisition. The main limitations are the somatic features because it can be difficult to obtain an optimal histogram in obese patients or in patients with shortness of breath [20]. The color-coded confidence map (left) reflects the quality of shear waves. The real-time 2D quantitative color map, namely elastogram (right), is superimposed on the B-mode. Even if ten stiffness values obtained from 10 independent images in the same location are generally required for liver ultrasound elastography, when a quality assessment parameter (confidence map) is used, five measurements may be appropriate. The ROI should be positioned at least 15-20 mm below the liver capsule [23].
After recording the stiffness values, the average, the standard deviation, the minimum and maximum stiffness values can be obtained. The standard deviation within the ROI reports the variability of the pixel measurements within the ROI and it is not a measure of the quality of the measurement [21].
The strain elastography (SE) technique ( Figure 3) uses frame-to-frame differences (tissue deformation) with stress caused after pressing the body surface with the probe or using the internal physiologic motion.
ROI reports the variability of the pixel measurements within the ROI and it is not a m ure of the quality of the measurement [21]. The color-coded confidence map (left) reflects the quality of shear waves. The real-time 2D qu tative color map, namely elastogram (right), is superimposed on the B-mode. Even if ten stif values obtained from 10 independent images in the same location are generally required for ultrasound elastography, when a quality assessment parameter (confidence map) is used, five m urements may be appropriate. The ROI should be positioned at least 15-20 mm below the capsule [23].
The strain elastography (SE) technique ( Figure 3) uses frame-to-frame differe (tissue deformation) with stress caused after pressing the body surface with the prob using the internal physiologic motion.
Using an intercostal space as a window, the strain image is obtained by the pres exerted on the liver by cardiac movement. The patient should stop breathing in mid spiration/expiration during the acquisition. The main limitations are the somatic feat because it can be difficult to obtain an optimal histogram in obese patients or in pat with shortness of breath [20]. Using an intercostal space as a window, the strain image is obtained by the pressure exerted on the liver by cardiac movement. The patient should stop breathing in midinspiration/expiration during the acquisition. The main limitations are the somatic features because it can be difficult to obtain an optimal histogram in obese patients or in patients with shortness of breath [20].

Liver Ultrasound Elastography in NAFLD Patients: Transient Elastography
Several studies are available on the performance of TE for the assessment of liver fibrosis in NAFLD patients. All studies including at least 50 patients and published during the last 10 years (between January 2012 and December 2023) addressing the topic of the use of TE in NAFLD patients are reported in Table 1. Three of them are meta-analyses published in 2017, 2021 and 2022, respectively.
The first meta-analysis was published by Xiao et al. in 2017, who addressed the interesting hot topic of the identification of the best method for diagnosing liver fibrosis in NAFLD patients and, particularly, the authors compared the performance of aspartate aminotransferase to the platelets ratio index (APRI), fibrosis-4 index (FIB-4), BARD score, NAFLD fibrosis score (NFS), FibroScan, shear wave elastography (SWE) and magnetic resonance elastography (MRE) [7].
Data were analyzed both for the M and XL probe (N = 13,046). Generally, transient elastography did not show a good accuracy in staging liver fibrosis in NAFLD patients because of a wide range in sensitivity and specificity and overlapping values for different stages of liver fibrosis (see Table 1).
The best range for the identification of advanced fibrosis was 7.6-9 kPa with 83% to 89% sensitivity and 77% to 78% specificity [7]. For the detection of liver fibrosis, the M probe showed a sensitivity and specificity of 91.9% and 55.5%, respectively, at a cut-off value of 5.8 kPa, whereas 6.65-7 kPa had a sensitivity and specificity of 80.1% and 68.3%, respectively [7]. Selvaraj et al. performed a new large meta-analysis in 2021 including 82 studies (14,609 patients) [24]. In total, 53 of them addressed the diagnostic performances of TE in the detection of the fibrosis stage. A cut-off value of 8.7 KPa was the best threshold for the identification of significant fibrosis (F3-F4). However, studies included in the meta-analysis used both M and XL probes, and a specific cut-off for different probes was not reported [24].
For the XL probe, the cut-off proposed for the detection of advanced fibrosis ranged from 5.7 to 9.3 kPa. These data were obtained from three studies including 579 patients, with 75% sensitivity and 74% specificity [21].
The last meta-analysis available was published by Cao et al. in 2022 who found the following cut-off values for the identification of significant fibrosis (F2), advanced fibrosis (F3) and cirrhosis (F4): 7.8 KPa, 9.9 KPa and 13.2 KPa, respectively. The proposed cut-offs showed a wide range with overlapping values (see Table 1).
This study included 10,537 patients from 61 studies and found different liver stiffness measurement (LSM) values among different regions; for diagnosing stages ≥ F2 and F4, the mean cut-off values of European and American patients were 0.96 and 2.03 kPa higher than Asiatic ones [25].
It is noteworthy that most available studies included in our review showed heterogenous data because populations were studied by using the M or XL probe and final cut-offs were a combination of their results.
It has been reported that the use of the XL probe is associated with lower cut-off values (−2 KPa) in comparison to the M probe, and this suggests that different cut-off values should be defined. The most recent meta-analysis published by Cao et al. in 2022 showed that the cut-off value for the detection of significant fibrosis by using the M probe was 7.8 KPa, which is similar to the threshold always used for the staging of liver fibrosis in patients with chronic HCV infection [26], whereas, in line with the lower values obtained by using the XL probes, and reported for the first time in patients with chronic viral hepatitis, the use of 6.9 KPa as a threshold value can be useful for the diagnosis of significant fibrosis (F2 stage) in obese patients. The same authors proposed 12.2 and 13.2 KPa as cut-off values for the diagnosis in NAFLD patients.
We identified 24 studies analyzing only data from populations studied by M probe. The best sensitivity of 100% for the identification of liver cirrhosis was recorded by Attia et al. and Chang et al. and Chan et al. who proposed a cut-off value of 15 KPa. Populations of their studies came from Malaysia, Singapore and Germany, suggesting that this threshold may be used for Asian and Caucasian patients. The highest sensitivity was also recorded by Lee H.W. et al. in a Korean population; however, their best cut-off was lower (11 KPa). Only two studies studied the best cut-off for the identification of liver cirrhosis in obese patients who underwent a liver stiffness measurement by using the XL probe.
Wong et al. in 2019 used the same cut-off value for the identification of liver cirrhosis (15 KPa) and found that sensitivity was only 49% even if specificity was high (93%) (see Table 1) [27].
On the other hand, Attia et al. found that 11.9 KPa was the best cut-off for the identification of patients (sensitivity 100%) with a good specificity (93%) [28].
The impact of liver steatosis on the cut-off values irrespective of liver fibrosis is a challenging issue. Petta et al. in 2015 published the results of a multi-center study including NAFLD patients who underwent liver biopsy [29]. In total, 6.9 kPa and 8.4 kPa were the best cut-off values for F2 and F3 liver fibrosis. Interestingly, the authors showed that the presence of obesity and severe steatosis was associated with higher stiffness values, suggesting a potential overestimation of fibrosis due to the liver steatosis [29].
This observation was not confirmed by more recent studies that did not find a strict association between the severity of liver steatosis and liver stiffness measurements [30], and confirmed by Wong et al. who showed that BMI but not severity of steatosis increased liver stiffness and that the same LSM cut-offs could be used without further adjustment when M and XL probes were used according to the BMI subgroups [27].
Particularly, these authors found that liver stiffness assessed with the XL probe was 2.3 KPa lower than values obtained with the M probe; however, patients with a BMI ≥ 30 kg/m 2 had similar liver stiffness values regardless of the probe used, suggesting that when M and XL probes are used in subgroups of patients by BMI (<30 vs. ≥30 kg/m 2 ), identical stiffness measurements are obtained with similar accuracy and diagnostic performance [27]. However, this limitation could be overcome by using the XL probe, which generally gives cut-off values 1.5-2 KPa lower than those found by using the M probe. These results were suggested by several observations reported in the literature showing that values obtained with M and XL probes have a comparable diagnostic accuracy, even if the stiffness values obtained by using XL probes are lower [31][32][33]. However, even if the probe does not affect the liver stiffness measurement [30], the use of the XL probe can improve the reliability of TE on the condition that specific cut-off values should be considered when XL probes are used [33].
The cut-off values proposed by different authors for the identification of significant fibrosis (≥F2) were highly variable and showed a wide range of sensitivity.
In 2017, Lee M.S. et al. enrolled 94 patients with biopsy-proved NAFLD and found that 7.4 and 8.0 KPa as cut-off values for the identification of liver fibrosis F2 and ≥F3 by using the M probe had the following sensitivities: 62.5%, and 82.6%, respectively; specificities were 91.7% and 84.9%, respectively [34].
On the other hand, by using the M probe and a cut-off value of 8.  [30].
In 2016, Imajo et al. enrolled 142 patients studied by using the M probe and with biopsy-proven NAFLD and found that 11 KPa and 11.4 KPa were the cut-off values for the correct staging of intermediate liver fibrosis [36]. Essentially

Liver Ultrasound Elastography in NAFLD Patients: Point Shear Wave Elastography (pSWE)
Most studies dealing with the role of pSWE in the management of patients with liver disease included patients with chronic viral hepatitis; however, only few studies have addressed the utility of point shear wave elastography in the assessment of liver fibrosis in patients with NAFLD. Table 2 shows all of the studies published in the last 10 years.
Before this period of time, only two studies addressed this topic in a very small cohort of patients. In 2010, Osaki et al. addressed the usefulness of pSWE in the management of non-alcoholic steatohepatitis. Only twenty-six patients were included in the study, of which twenty-three had NASH, whereas three patients were the controls [71]. The authors suggested that a cut-off value of 1.47 m/s gave a significant contribution for distinguishing stages 3 and 4 from stages 0 and 1 with excellent sensitivity and specificity (100% and 75%, respectively). In 2011, Friedrich-Rust et al. published the first meta-analysis addressing the performance of pSWE for the staging of liver fibrosis and included a small group of NAFLD patients (n = 77) from four studies [72]. The overall diagnostic accuracy was optimal for the identification of cirrhosis (0.94 (0.81, 1.00)) in comparison with the diagnostic accuracy of ≥F2 or ≥F3 (0.86 (0.75, 0.96) and 0.86 (0.58, 1.00), respectively).
After this publication, more studies have been published addressing this topic. As shown in Table 2, cut-off values for the identification of significant fibrosis are highly variable; however, a cut-off value of 14.2 KPa reported by Argalia et al. [35] showed both a sensitivity and specificity of 100%. It is noteworthy that more than one region may be used for the study of liver stiffness. For this purpose, Attia et al. confirmed that both segments 6 and 8 had similar cut-off values, sensibility and specificity [28]. Most parts of the cut-off values obtained in the study realized in the last 10 years included Caucasian populations. Therefore, more studies are needed to confirm that similar thresholds may be used in non-Caucasian people.
Two meta-analyses have been published in 2015 and 2018 by Liu et al. [73] and Jiang et al. [74], respectively; both of them addressed the reliability of pSWE in staging liver fibrosis in NAFLD patients and found that sensibility and specificity were high enough to consider pSWE a good method for the identification of patients with significant fibrosis and cirrhosis.

Liver Ultrasound Elastography in NAFLD Patients: 2D-Shear Wave Elastography
Together with MRE, 2D-SWE has been currently considered the method with the highest accuracy for staging liver fibrosis in NAFLD patients [7].
The 2D-SWE has shown a sensitivity and specificity of 90% and 93%, respectively, for the identification of advanced liver fibrosis [7]. However, only a few studies are available in the literature addressing the clinical utility of 2D-SWE for the assessment of liver stiffness in NAFLD patients. We found only 10 studies which are shown in Table 3.
The latest guidelines recommended a potential role for 2D-SWE to rule out advanced fibrosis and for the selection of patients who deserve further assessment [21].
This recommendation was based on the availability of three studies which showed a good performance in patients with advanced fibrosis.
The authors found that, using the same cut-off values for the 2D-SWE and TE for advanced liver fibrosis, the accuracy of this method was good, and the inclusion of 2D-SWE in a three-step strategy (FIB4 +TE+2D-SWE) strongly decreased the need for liver biopsy to <5% of patients who require the invasive approach for the correct classification of liver fibrosis.

Liver Ultrasound Elastography in NAFLD Patients: Strain Elastography
The role of strain elastography in the staging of liver fibrosis in NAFLD patients is very limited. In 2001, Ogino et al. studied 107 patients and assessed the diagnostic performance of the strain elastography in staging liver fibrosis in comparison with SWE in biopsy-proven NAFLD.
The diagnostic performance of the strain elastography measured by the area under the curve was 0.75 for F2, 0.80 for F3 and 0.85 for F4, whereas the AUROC for SWE was 0.88 for F2, 0.87 for F3 and 0.92 for F4. Therefore, the results showed that strain elastography was inferior to SWE in terms of the ability to classify liver fibrosis in patients with NAFLD [84].
On the other hand, Ochi et al. in 2012 assessed the liver stiffness in 181 patients with real-time tissue elastography in patients with non-alcoholic fatty liver diseases and compared the results with the histological stage. The cut-off values were 2.47 for F1, 2.67 for F2, 3.02 for F3 and 3.36 for F4 with a diagnostic accuracy between 82.6% and 96.0% in all stages [85].
The authors concluded that strain elastography reliably identifies the early stage of fibrosis in NAFLD patients and that, therefore, it is a useful tool for the management of NAFLD patients.
In 2015, Kobayashi et al. published a meta-analysis to study the overall accuracy of strain elastography in the staging of liver fibrosis in patients affected by different liver diseases [86]. The analysis included 15 studies and 1626 patients and showed that, compared with transient elastography and ARFI imaging, the accuracy of strain elastography was similar for evaluating significant liver fibrosis, but less accurate for the identification of patients with cirrhosis [86]. General results were confirmed in a subgroup analysis including only patients with NAFLD.
Finally, general experience in the literature has shown that strain elastography is a tool which has not been playing a key role in the clinical management of patients with NAFLD. However, recently, some authors have proposed a new concept of "combinational elastography" based on the assumption that the combination of strain and shear wave imaging may increase the relevance of each single ultrasound-based elastography and, therefore, improve their accuracy in the correct classification of liver fibrosis [87].

Conclusions
Ultrasound elastography has become the most important non-invasive tool for the assessment of liver fibrosis in patients with NAFLD. It includes different techniques with different diagnostic performances. Strain elastography seems to have the lowest diagnostic accuracy when it is used alone. Transient elastography and 2D-shear wave elastography have shown good accuracy in diagnosing significant fibrosis; however, their sensibility and specificity are not optimal for detecting low-grade fibrosis yet. Future studies are needed to explain the role of the operator experience on the accuracy of liver ultrasound elastography in detecting intermediate stage liver fibrosis and the impact of the severity of liver steatosis and/or somatic features (obesity or overweight) on the diagnostic performances of the different ultrasound elastography techniques.

Conflicts of Interest:
The authors declare no conflict of interest.