The Place of Liver Elastography in Diagnosis of Alcohol-Related Liver Disease

Abstract

Harmful use of alcohol is associated with more than 200 diseases and types of injuries, the liver being one of the most important targets. Alcoholic liver disease (ALD) is the most frequent cause of severe chronic liver disease in Europe and worldwide. ALD can progress from alcoholic fatty liver to alcoholic steatohepatitis and alcoholic liver cirrhosis, the grade of fibrosis being the key prognostic factor for the severity of the diseases. This chapter will present the place of liver elastography in the noninvasive assessment of ALD. It will describe the data available in the literature regarding the different elastography techniques for liver stiffness assessment and also the potential of these techniques for screening ALD.

1. Introduction

Alcohol-related liver disease (ALD) is one of the major causes of liver injury worldwide, according to WHO [1]. More than 40% of the liver deaths are attributed to alcohol [2] and the indication for liver transplant in patients with ALD has significantly increased, being the top health burden. ALD is rarely detected at early stages, most of the patients being diagnosticated at the decompensation stage, when liver cirrhosis and its complications occur [3].

Diagnosis of ALD is suspected when alcohol consumption is >20 g/d in females and > 30 g/d in males and clinical and/or biological modifications suggestive of liver injury or extrahepatic manifestations of alcohol use disorder (AUD) occur [4]. Because a high proportion of patients with AUD do not express clinical symptoms or laboratory abnormalities, asymptomatic patients with harmful use of alcohol should undergo appropriate screening investigations [5].

ALD follows the typical progression of chronic liver diseases including alcoholic liver steatosis, steatohepatitis, fibrosis, and liver cirrhosis. Approximately 90% of heavy drinkers will develop liver steatosis and 5–10% liver cirrhosis in 5 years [6]. Liver cirrhosis is the main predictor of survival [7], so early recognition of fibrosis is the most important objective in this category of patients. On the other hand, another argument for early detection and diagnosis of patients with harmful use of alcohol is that the risk of developing liver disease decreases with abstinence [4]. ALD remains underestimated due to bad reporting of real alcohol consumption and a lack of specific investigations.

The main points to address in front of a patient with ALD from the hepatological point of view would be the evaluation of liver steatosis, inflammation, and fibrosis.

Liver biopsy remains the “gold standard” of diagnosis and staging for diffuse liver changes [4] since it is able to evaluate all points presented above; however, it is an invasive method, less likely accepted by patients, with a 7% rate of complications and sampling errors [8]. Noninvasive methods for evaluating steatosis and fibrosis in ALD gained a lot of interest lately, with many studies supporting their usefulness [9], but we still lack reliable noninvasive methods for grading liver inflammation. The main advantages of these noninvasive methods are the easy acceptability by patients, repeatability, and low costs. They consist of serum markers and elastography methods [9].

Noninvasive liver fibrosis evaluation in ALD by serum markers/biological scores can be performed by patented or non-patented serum biomarkers. Enhanced Liver Fibrosis (ELF™) and FibroTest (FT) are most commonly used as patented biomarkers. In a meta-analysis performed on nine studies, the ELF test showed good performance for the prediction of histological fibrosis stage [10]. A prospective study found that ELF and FT also had comparable diagnostic accuracy for ALD when it comes to AUROC, 0.92 for ELF and 0.9 for FT, and can rule out advanced fibrosis for ALD based on an ELF <10.5 or an FT value below 0.58 [11].

Nonpatented serum markers have been assessed in ALD for the diagnosis of liver fibrosis—age-platelet index, the aspartate transaminase (AST)-platelet-ratio index APRI [12], fibrosis-4 index-FIB-4 [13], and AST/alanine aminotransferase (ALT) ratio-AST/ALT [14]. A comparison of the performance of the different biological scores suggests that ELF and FT are better in the diagnosis of LF in ALD (Table 1) [9, 16].

Test	AUC	Se (%)	Sp (%)
APRI cut off 0.5 [15]	0.79	84	41
APRI ≥ 1 [11]	0.80	38	90
FIB-4 ≥ 3.25 [11]	0.85	58	91
AST/ALT ratio ≥ 1 [11]	0.76	85	46
Age-platelet index ≥ 6.0 [11]	0.81	65	85
ELF ≥ 10.5 [11]	0.92	79	91
FT ≥ 0.58 [11]	0.90	67	87

Table 1.

Comparison and performance of biological tests for the diagnosis of advanced fibrosis (F3) in studies with biopsy-proven ALD.

APRI – aspartate transaminase-platelet ratio index; FIB-4 – fibrosis 4 index; ELF-Enhanced Liver Fibrosis; FT – Fibrotest; AUC – area under the curve; Se – sensitivity; and Sp – specificity.

From an economical point of view, patented scores for ALD have higher costs than nonpatented but provide the best diagnostic performance of advanced liver fibrosis. Lifetime health costs in ALD are very high in decompensated stages, so noninvasive methods were proven to be cost-effective [17].

For liver steatosis assessment several methods can be used as noninvasive techniques such as ultrasound-based methods or magnetic resonance imaging (MRI)-based methods. In the following part of this chapter, ultrasound-based methods will be introduced. MRI methods use MRI-PDFF (proton density fat fraction) and a routinely used MRI scanner to identify liver steatosis. MRI sensitivity and specificity are 76.7–90.0% and 80.2–87.0%, respectively. It is not affected by the etiology of liver disease or other abnormalities such as inflammation, most seen in ALD or iron overload. It has several advantages such as the highest accuracy following liver biopsy for liver steatosis diagnosis, but the major disadvantages include the high cost, long time of examination, and the inability to be used in claustrophobic or overweight patients [18, 19].

Liver elastography by means of transient elastography compared to serum markers is superior when it comes to liver stiffness assessment [20], and in the following sections, the place of noninvasive ultrasound-based steatosis quantification methods and ultrasound-based elastography techniques in ALD are presented in detail.

In patients with suspected ALD (presence of alcohol use disorder-AUD, abnormal liver test or extrahepatic manifestations of AUD, and no other causes of chronic liver disease), noninvasive tests can be transferred into clinical practice for the detection of advanced fibrosis. Physical and biological approaches are complementary and both methods should be used starting from primary care to facilitate the early detection of ALD. First of all, patients should be routinely screened for AUD using AUDIT questionnaire [4]. Further, patients can be easily assessed by primary care using patented or nonpatented biological scores and in case of liver fibrosis presence, redirected to second-line assessment made by a liver specialist, to validate the results by elastography methods. All patients with AUD need to be referred to a specialized withdrawal center. Follow-up can be performed by primary care or in case of advanced liver fibrosis by liver clinic units for specific investigations [7, 16].

2. Alcoholic liver steatosis assessment by ultrasound-based methods

Hepatic steatosis is characterized by accumulation of fat-lipids, especially triglycerides in hepatocytes, and when is over 5%, it is considered pathological. It is usually asymptomatic and is mainly caused by alcohol use and metabolic factors. In patients with AUD, liver steatosis can be reversible with abstinence. Steatosis severity is associated with lobular inflammation and fibrosis [21]. Approximately 90% of patients with AUD will develop liver steatosis [6].

Liver biopsy remains the gold standard of steatosis assessment, but has many drawbacks, like potential complications, sampling error, invasive, and is difficult for patients to accept this method as a follow-up method [8]. Noninvasive methods were developed to easily assess patients at risk of developing liver steatosis.

Because fat accumulation alters liver imagistic appearance, B-mode ultrasound (US) is the first-line technique used for screening and assessment of fatty liver [9]. It is a safe method, available, accessible, repeatable, and cost-efficient, with a sensitivity between 60 and 94% and specificity between 88 and 95% in detecting steatosis [22]. However, it has a better performance in detecting severe liver steatosis as compared to mild steatosis, is operator-dependent, and cannot give information related to fibrosis presence [23]. Magnetic resonance imaging (MRI)-proton density fat fraction (PDFF) is considered the most specific and sensitive technique for liver steatosis assessment [24], but it is not appropriate as a point of care technique because it requires complex evaluation by specialized radiologists, has high costs, and is not available in all centers. Also, it is not possible in the case of obese, claustrophobic patients and with metallic devices implanted [24].

A novel noninvasive ultrasound-based elastographic parameter called CAP-controlled attenuation parameter has been developed for life’s steatosis assessment. It is based on vibration controlled transient elastography (VTCE) and is incorporated in FibroScan (Echosens, Paris, France) device and allows, in the same session, the evaluation of steatosis and fibrosis [25]. It is based on ultrasound attenuation, a physical characteristic of the propagation medium, which means the loss of energy when ultrasound spreads through this medium, and fat is known to be an attenuating medium [26]. CAP has been first developed on the M probe with a center frequency of 3.5 MHz [26], but when applied to overweight and obese patients the performance was impaired because of the thick subcutaneous fat layer. XL probe was then developed on 2.5 MHz and it measures to a depth of 7.5 cm [27]. The results are given in decibels per meter (dB/m) with a range from 100 to 400 dB/m. CAP is displayed only when liver stiffness measurements (LSMs) are valid, and it is recommended as a point-of-care technique for the detection of liver steatosis by the World Federation for Ultrasound in Medicine and Biology (WFUMB) [28].

CAP proved to have good accuracy for diagnosing steatosis in studies and meta-analysis including mixed cohorts [25] and especially in NAFLD [29]. In ALD, only one study is available from Thiele M. et al. [30], including 562 patients with ALD who underwent CAP, B-mode ultrasound, and liver biopsy. CAP proved to be superior to steatosis liver pattern by standard ultrasound. A CAP value over 290 dB/m ruled in any steatosis with 88% specificity, while CAP below 220 dB/m ruled out liver steatosis with 90% sensitivity. Moreover, CAP showed AUROCs of 0.77, 0.78, and 0.82 for mild, moderate, and severe steatosis, respectively. CAP had higher values for patients with ALD and metabolic syndrome (MetS) over imposed, with an average difference of 40 dB/m (302 ± 64 in patients with MetS vs. 262 ± 55 in patients without MetS; P < 0.001). The same was observed in patients with a BMI ≥30 kg/m² with a difference of 49 dB/m (311 ± 48 in obese patients versus 261 ± 57 in patients with BMI < 30; P < 0.001). In the same multicentric study, 293 patients were admitted for detoxification and CAP showed a decrease by 32 ± 47 dB/m, decreasing equally in patients with ALD with or without MetS, but did not significantly decrease in obese patients after detoxification. There was no evidence that CAP influences liver stiffness measurements by TE or vice-versa. Diagnostic accuracy of CAP seems to be lower in mild steatosis compared to other etiologies and optimal cut-offs, which varies in the different studies performed; variation is possibly explained by the pattern of alcohol consumption in the moment of investigations; hence, evaluation of CAP in ALD remains a challenge.

Several other ultrasound equipment developers designed new ultrasound-based steatosis quantification software embedded in ultrasound machines, based mainly also on the evaluation of the ultrasound beam attenuation. Such examples are Ultrasound-Guided Attenuation Parameter (UGAP) from General Electric Healthcare, Attenuation imaging (ATI) developed by Canon, Attenuation (ATT) from Hitachi, SSp.PLUS (Sound Speed Plane-wave UltraSound) and Att.PLUS (Attenuation Plane-wave UltraSound) from Supersonic Imagine, and TAI™ (Tissue Attenuation Image) & TSI™ (Tissue Scatter-distribution Image) from Samsung; all emerging techniques are under evaluation but with no data yet related to ALD.

3. Alcoholic-induced liver fibrosis evaluation by ultrasound-based elastography

Because the presence of liver fibrosis and liver cirrhosis is the main predictor of survival in patients with ALD, liver stiffness (LS) assessment is very important in high-risk patients [7]. Liver fibrosis assessment can be performed by biological and elastography methods in the detriment of liver biopsy. Direct comparison with serum markers showed a better performance of TE in patients with ALD [20] with an AUROCs >0.9 for F4 cirrhosis diagnosis. In this chapter, we will discuss only the ultrasound-based elastography methods.

Liver elastography methods became more and more reliable in the liver stiffness measurement (LSM), being supported by recently published guidelines [28, 31]. The methods are classified into shear wave elastography (SWE) and strain elastography (SE) (Figure 1). Both guidelines support that SWE is the best for clinical use in LSM.

Figure 1.

Classification of ultrasound-based elastography methods.

4. Ultrasound-based elastography techniques

Transient elastography (TE) (FibroScan, EchoSens, Paris, and France), the first elastography technique developed, is the most widely used method, and it is noninvasive, rapid, and reproducible, with lower sampling errors [9]. The most important published studies in ALD are listed in the table below, majority of them biopsy-proven (Table 2). These studies showed good performance in the diagnosis of liver cirrhosis with AUROCs from 0.87 [41] to 0.97 [40], but the cut-off values differ quite a bit, most likely due to the presence of inflammation in these patients, given by recent alcohol consumption and assessed by AST levels. Several studies performed by Mueller et al. [40, 45, 46] show that absolute alcohol withdrawal leads to a 13% reduction of LS after one week and even a reduction of 40% in alcohol consumption can lead to a 17% reduction of LS [24]. In another study, LS improved in almost 80% of patients admitted for alcohol detoxification due to the coexistence of inflammation seen by AST >100 U/ml [25]. Preliminary observational data on long-term abstinence, observed over a period of more than 5 years, show LS decreases by 50% and also LS again increases by 22% if alcohol consumption continues [46]. Table 3 resumes the data on alcohol abstinence/ relapse and LS improvement.

Reference	Number of patients (n)	Elastography method	Cut-off values for different fibrosis stage
			F2	F3	F4
Anastasiou 2010 [32]	14 patients	TE	>7.15 kPa		>12.5 kPa
Bardou-Jacquet 2013 [33]	8 patients	TE	>7.15 kPa		>17 kPa
Boursier 2009 [34]	106 patients	TE	>7.15 kPa	>9.5 kPa	>17.3 kPa
de Ledinghen 2012 [35]	34 patients	TE	>7.15 kPa	>9.5 kPa	>12.5 kPa
Fernandez 2012 [36]	139 patients	TE	>7.15 kPa	>10.5 kPa	>18 kPa
Janssens 2010 [37]	49 patients	TE	>7.15 kPa		>21.1 kPa
Lannerstedt 2013 [38]	16 patients	TE	>7.15 kPa	>9.5 kPa	>12.5 kPa
Lemoine 2008 [39]	48 patients	TE	>7.15 kPa		>34.9 kPa
Mueller 2010 [40]	101 patients	TE	>7.15 kPa	>8 kPa	>11.5 kPa
Nahon 2008 [41]	147 patients	TE	>7.15 kPa		>22.6 kPa
Nguyen-Khac 2008 [20]	103 patients	TE	>7.15 kPa	>11 kPa	>19.5 kPa
Muller 2015 [42]	364 patients	TE	>6 kPa	>8 kPa	>12.5 kPa
Voican 2017 [43]	188 patients	TE	—	>13 kPa	>20.8 kPa
Kim 2009 [44]	45 patients	TE	>7.15 kPa	>9.5 kPa	>25.8 kPa

Table 2.

Elastography in ALD patients performed by TE.

Reference	Patients (n)	Time	Mean LS Before/after	LS improvement
Mueller 2020 [45]	45- reduction of alcohol consumption and treatment with Nalmefene	12 weeks	10.5 kPa–8.7 kPa	−13%
Mueller 2010 [40]	50- detoxification	5 days	20.1 kPa–16.5 kPa	−17%
Mueller 2020 [46]	23- abstinence; 23-relapse	5.7 years; 5.3 years	20.5–10.5 kPa; 14.8 kPa–18.1 kPa	−48%; +22%

Table 3.

Alcohol abstinence/relapse and liver stiffness improvement.

TE is followed by other ultrasound-based methods, such as point Share Wave Elastography (pSWE), Two-Dimensional Share Wave Elastography (2D-SWE), or Time-Harmonic Elastography embedded in ultrasound systems [47, 48, 49, 50, 51, 52, 53]. There are few studies that show data on the performance of pSWE or 2D-SWE in ALD (Table 4), with a small number of included patients, and in some studies, data show a wide range of values.

Reference	Number of patients (n)	Elastography technique	Cut-off values for different fibrosis stage
			F2	F3	F4
Thiele 2016 [54]	199 patients	2D-SWE	>10.2 kPa	—	>16.4 kPa
Kiani 2016 [55]	69 patients	pSWE	>1.63 m/s	>1.84 m/s	>1.94 m/s
Zhang 2015 [56]	112 patients	pSWE	>1.27 m/s	>1.40 m/s	>1.65 m/s
Cho Y 2020 [57]	251 patients	pSWE	>1.46 m/s	>1.47 m/s	>1.66 m/s

Table 4.

Elastography in ALD patients performed by 2D-SWE and pSWE.

An important aspect of liver elastography in alcohol-induced liver fibrosis, compared to the rest of liver fibrosis etiologies is the presence of inflammation given by the levels of AST. In alcoholic liver disease, AST levels are typically higher as compared to ALT [58]. Although in cirrhotic stages, liver transaminases normalize, if alcohol consumption is continued, AST may be continuously increased. The presence of steatohepatitis with AST >100 U/ml will increase liver stiffness in patients with ALD, so it was proposed to assess the presence of advanced fibrosis when AST decreases below <100 U/ml [40]. For that, an algorithm was developed for inflammation-adapted cut-off values in ALD [42], based on a multicentric study that included over 2000 patients with biopsy-proven HCV and ALD. In the absence of inflammation given by elevated transaminases, cut-off values for ALD and HCV were similar. The cut-off values increased exponentially in relation to median AST. After the formula was applied there was an improved agreement of the AST cut-off values with the histological stage for both HCV and ALD, so using inflammation-adapted cut-off values avoid repetitive assessment of LS in ALD.

In a recent meta-analysis [43], it was proved that in addition to AST, bilirubin can have a significant effect on LS assessment in ALD. Bilirubin was independently associated with the presence of asymptomatic and non-severe steatohepatitis on histological features.

From an economical perspective, lifetime health care costs associated with ALD in advanced stages are very high, so noninvasive elastography methods for the diagnosis of advanced alcohol liver fibrosis were proven to be cost-effective [17] and may be used also for screening.

5. Conclusion

Because alcohol-related liver diseases are increasing, there is an unmet need for the identification and evaluation of patients at risk. Noninvasive elastography methods allow screening, diagnosis, and follow-up of liver steatosis and liver fibrosis in patients with ALD, with good accuracy and performance.

Conflict of interest

Alina Popescu has received speaker fees from Philips, General Electric Healthcare.

Camelia Foncea has nothing to declare.