Imaging Based Methods of Liver Fibrosis Assessment in Viral Hepatitis: A Practical Approach

Liver fibrosis represents the repair mechanism in liver injury and is a feature of most chronic liver diseases. The degree of liver fibrosis in chronic viral hepatitis infections has major clinical implications and presence of advanced fibrosis or cirrhosis determines prognosis. Treatment initiation for viral hepatitis is indicated in most cases of advanced liver fibrosis and diagnosis of cirrhosis entails hepatology evaluation for specialized clinical care. Liver biopsy is an invasive technique and has been the standard of care of fibrosis assessment for years; however, it has several limitations and procedure related complications. Recently, several methods of noninvasive assessment of liver fibrosis have been developed which require either serologic testing or imaging of liver. Imaging based noninvasive techniques are reviewed here and their clinical use is described. Some of the imaging based tests are becoming widely available, and collectively they are shown to be superior to liver biopsy in important aspects. Clinical utilization of these methods requires understanding of performance and quality related parameters which can affect the results and provide wrong assessment of the extent of liver fibrosis. Familiarity with the strengths and weaknesses of each modality is needed to correctly interpret the results in appropriate clinical context.


Introduction
Liver fibrosis is a common pathway of liver injury for multiple chronic liver conditions. Chronic viral hepatitis, metabolic, autoimmune, and cholestatic liver diseases amongst others can lead to clinically significant fibrosis. Assessment of the severity of hepatic fibrosis is essential for determining the prognosis of patients with chronic viral hepatitis and estimating the urgency of antiviral therapy [1]. The severity of liver fibrosis and advanced liver disease dictates the need for further evaluations, such as hepatocellular carcinoma surveillance and esophageal varices screening. Although in several cases the diagnosis of advanced liver fibrosis or liver cirrhosis can be made clinically, the liver biopsy has been used as the conventional reference method for staging fibrosis and diagnosis of liver cirrhosis.
Liver biopsy offers histologic visualization of the extent of liver fibrosis and thus is considered the "best" standard for fibrosis staging and is universally accepted; however it is an invasive technique with associated morbidity. Minor complications are relatively frequent and about 25% of patients undergoing liver biopsy have pain in the right upper quadrant or right shoulder after the procedure [2]. Severe complications are uncommon; significant bleeding incidents range from 0.05% to 5.3% and mortality of less than 0.15% was seen in the largest series [3]. The accurate evaluation of fibrosis using liver biopsy is also complicated by sampling error and interobserver variation in staging, particularly when inadequate sampling occurs [4]. A study which included 124 patients with chronic Hepatitis C Virus (HCV) infection who underwent simultaneous laparoscopy guided biopsies of the right and left hepatic lobes showed that 33.1% of subjects were graded to have a difference of at least one stage between the two lobes [5]. Because liver biopsy has several limitations, namely, sampling error and  intra-and interobserver variability, the Area Under Receiving Operating Characteristic (AUROC) curve values of >0.9 for liver biopsy could not be attained in clinical studies for fibrosis staging [6], which is a characteristic for an ideal diagnostic test [7]. Rather than the conventional label of being the "gold" standard, liver biopsy is increasingly recognized as the "best" standard for liver fibrosis staging. In addition, performing liver biopsies on a large scale for clinical use may not be a reasonable, cost-effective, or practical approach. Currently, some of the noninvasive methods of liver fibrosis assessment have better diagnostic value than liver biopsy in chronic liver diseases and are safe to perform. As a result, the noninvasive methods are increasingly replacing invasive liver biopsy in clinical practice due to patient wariness of the morbidity associated liver biopsy and physician ease of clinical use with accuracy.
The noninvasive methods of liver fibrosis can be broadly divided into two main groups: serum biomarkers (which require blood sample collection and estimation and calculation of score) and the imaging based methods. In this review, we will focus on imaging based methods and describe the technical aspects, performance accuracy, meaningful utilization, and incorporation in clinical practice of now available elasticity based methods of liver fibrosis assessment (elastography).

Elastography Imaging
The techniques included can be divided into four main groups: (A) shear wave elastography technique based on mechanically generated impulse that includes Transient Elastography (FibroScan); (B) shear wave elastography techniques based on the acoustic beam that includes 2D shear wave elastography/point shear wave elastography/acoustic radiation force impulse; (C) real-time technique based elastography; and (D) magnetic resonance elastography. These are shown in Figure 1 and overview of comparison of various methods is described in Table 1. It is now recommended that every patient with viral hepatitis should have liver disease staging at least once by noninvasive methods [8]. Transient Elastography (FibroScan) is the most extensively used and validated method for fibrosis staging. Shear wave elastography/acoustic radiation force impulse seems promising; however its incorporation in clinical decision seems to be in early stage with only relatively recent pilot studies. Magnetic resonance elastography is not widely used given the described limitations despite the higher performance. Major advantage of these imaging methods is that these can be conveniently repeated over time in patients who are treated or those who stay on long term treatment for viral hepatitis. The LSM can be followed to monitor for therapy response and improvement of liver fibrosis. In addition, one can expect that the elastography improvement over time will predict improvement in clinical outcomes.

Transient Elastography (TE). TE or Vibration Controlled
Transient Elastography technique is exclusively used in FibroScan which was developed nearly 10 years ago by Echosens [9]. It is now the reference test and is the best validated technique for noninvasive evaluation of liver fibrosis. Its use in the United States was approved by the Food and Drug Administration (FDA) in 2013 for liver fibrosis assessment in HCV patients.

Technical
Aspects. TE uses mechanically generated elastic shear wave (SW) and follows its propagation through the liver to measure its speed in a single dimension. The speed of SW in the liver is directly related to its stiffness and the level of fibrosis (elastic modulus) [9]. A probe transducer is used which mechanically generates SW by vibration at the frequency of 50 Hz and is placed on the chest wall in the right ninth to eleventh intercostal space. SW is followed in the cylindrical Region of Interest (ROI) of 1 cm × 4 cm dimension, 2.5 or 6.5 cm below the skin surface to estimate the speed by use of a sensor. A reading is obtained each time button is pressed while probe is in a correct position and is called liver stiffness measurement (LSM), expressed in kilopascals (kPa). The LSMs range from 2.5 to 75 kPa, with mean value around 5 kPa [9]. Accurate result of an examination requires careful interpretation of data from at least 10 valid LSMs. A success rate (the ratio of valid measurements to the total number of measurements) of >60% and an interquartile range (IQR), which reflects variations among measurements, of <30% of the median LSM value (i.e., IQR/median LSM ≤30%) are considered as an adequate study [10]. Table 1 compares certain technical aspects in comparison to other elastography techniques.

Diagnostic Performance and Accuracy.
Since the introduction of TE in 2005 in France, its use in early European cohorts of HCV patients has clearly shown that TE can detect clinically significant fibrosis and cirrhosis with a good accuracy and reproducibility [11,12]. There have been several validation studies in HCV patients indicating that LSM strongly correlates with liver fibrosis on biopsy. Furthermore, in the largest multicenter independent study to date, TE outperformed all other noninvasive tests for the diagnosis of cirrhosis [13]. Similar findings were noted in multiple North American publications. In a recent large United States multicenter cohort study, TE was evaluated in patients with chronic type B and C viral hepatitis (903 patients), and the readings correlated strongly with the fibrosis stage on biopsy with AUROC of 0.89 and 0.92 for significant fibrosis (F ≥ 2) and F4, respectively, with good intra-and interoperator reproducibility (Interclass Correlation (ICC) of 0.95 and 0.99). Liver stiffness cutoff values in this study (kPa) were 8.4 for ≥F2 and 12.8 for F4 [14]. TE generated LSMs have been increasingly applied to other chronic liver diseases besides HCV that include Chronic Hepatitis B Virus (HBV) infection, Nonalcoholic Steatohepatitis (NASH), Alcoholic Liver Disease (ALD), and cholestatic liver diseases [15,16]. Different liver disease specific stiffness KPa cutoffs were described and are presented in Table 2. Similar to previously noted multiple systemic reviews and meta-analyses, recent large systematic review of published studies between 2001 and June 2011 across chronic liver diseases showed a strong diagnostic accuracy of TE compared to biopsy stages with an AUROC of 0.88, 0.92, and 0.94 for F ≥ 2, F ≥ 3, and F4, respectively [17].

Prognosis and Longitudinal Follow-Up.
There is good evidence to support the association between TE measurements and survival. A French study examining 5-year survival in a large cohort of 1457 HCV patients demonstrated that the overall survival using TE for LSM ≤9.5 kPa was 96%;  [50] 200 All liver diseases 0.84 0.90 7.9 11.9 Degos et al. [13] 1839 Multicenter All liver diseases 0.76 0.90 5.6 12.9 Afdhal et al. [14] 907  [24]. United Stated FDA and FibroScan manufacturer agreement requires that an operator undergo a training course and perform a minimum of 10 cases under supervision of a proctor before getting certified to use the device independently.
Other Factors. Liver stiffness is not synonymous with fibrosis and any process that may change hepatic viscoelastic properties can affect LSM readings and lead to overestimation of liver stage. The most important LSM modifying factors include acute inflammation (elevated ALT) [25], extrahepatic cholestasis [26], passive cardiogenic congestion [27], and food intake [28]. Arena et al. showed that after a meal the liver stiffness returned to baseline levels within 120 min in all patients independent of fibrosis stage [28]. Patients without significant fibrosis but high levels of ALT could get LSM within the range for cirrhosis. This was also noted in a large Canadian multicenter study in which the authors reported that, for every 100 IU/mL increase in ALT, LSM increased by 1.1 kPa [29].  It is important to recognize that the clinical confounding variables in clinical decision-making are important (obesity, inflammation, and cholestasis). This entails that a physical examination is done including BMI estimation and assessment of presence of ascites; liver functional and biochemical testing (ALT, alkaline phosphatase) should be available at the time of LSM evaluation. Patient should be fasting for at least 2 hours prior to performing the test. LSMs are a section of the overall diagnosis process and must take into account all the disease clinical, laboratory, and radiological findings. Appreciating the TE intrinsic validities (positive and negative predictive value and AUROCs) and the pretest probability of advanced fibrosis would facilitate the decision-making process.  [30]. One suggested algorithm for staging of liver disease in HCV is shown in Figure 2. The current recommendations for treatment of HCV differ considerably between healthcare systems and insurance policies. TE can be used to prioritize patients for HCV therapy based on liver disease stage. TE should be combined with serologic noninvasive fibrosis assessment tests to increase the diagnostic accuracy for significant fibrosis and thus initiation of therapy. Once access to care for HCV widens to nonliver or infectious disease providers TE can be used to estimate the presence of cirrhosis and thus referral for specialized care can be suggested [31]. Studies have confirmed histologic improvement and thus clinical outcomes in patients who are cured of HCV infection after therapy [32,33]. Liver elasticity improves with HCV cure and can be measured as LSM by TE. Several studies showed improvement in LSM after HCV treatment [34,35]. However, care should be taken while interpreting the follow-up results of TE, as improvement in ALT with HCV treatment will have an impact on LSM. The clinical utility of improvement in LSM after SVR has a little clinical significance in patients without cirrhosis; however the determination of cirrhosis regression after SVR in HCV treated patients could be of important clinical implication. The study that evaluated the cirrhosis estimation showed sensitivity of 61% when the usual cutoff of 12 kPa for cirrhosis was used [36] in patients with cirrhosis who achieved SVR after treatment. This indicates that TE may not be a good tool to estimate cirrhosis regression in HCV patients after SVR, and the time interval between LSM assessments after obtaining SVR is not yet tested. TE showed good predictability, however, in regard to liver related outcomes in HCV patients. In patients who were untreated, a large study showed that gradually rising LSM by 1 kPa per year in HCV patients portends poor clinical outcomes [20]. Other studies were also able to stratify HCV patients based on LSM cutoff values as those with increased risk of long term liver related outcomes [18,19].

HCV including HIV-HCV
HBV. Overall TE performs better than the serologic noninvasive methods for diagnosing advanced liver fibrosis in HBV infected patients [37]. Chronic HBV infection has various phases of liver injury, inflammation, and viral replication during a lifetime and TE can be used effectively to estimate the extent of liver fibrosis. TE can yield falsely high readings in patients with elevated liver enzymes (ALT levels), either during active hepatitis phase or during reactivation flares and even 3-6 months after ALT normalization following an acute exacerbation of chronic HBV infection [38]. For HBeAg negative patients who are in immune inactive phase (inactive carriers) TE showed steady results [39] as compared with other methods and LS < 5-6 kPa suggested absence of or minimal fibrosis [40]. LSM readings higher than 12 kPa have good predictability to diagnose cirrhosis in HBV inactive carriers [41], while intermediate LSM values should be followed up with a liver biopsy for correct estimation of liver fibrosis [42]. For HBeAg negative patients in immune active phase and fluctuating or persistent borderline HBV viremia TE may be preferred over performing a liver biopsy to diagnose cirrhosis [43] or advanced fibrosis [42], respectively. Among HBeAg positive patients with normal or mildly elevated ALT, using TE can be helpful in estimation of liver fibrosis in adults to distinguish between immune tolerant phase and advance liver disease secondary to undiagnosed immune active phase of liver injury [44]. The estimation of liver fibrosis by TE can avoid liver biopsy in many patients and guide initiation of antiviral therapy in a timely manner in chronic HBV infection. TE should be considered to diagnose cirrhosis in patient and hence to start therapy with chronic HBV infection and clinical suspicion of advanced liver disease despite normal ALT. Long term antiviral therapy has shown resolution of liver fibrosis/regression of cirrhosis in HBV infected patients [45]. Several studies have shown that TE can be used for follow-up and documentation improvement in liver fibrosis while on HBV therapy [46,47]. The follow-up LSM readings should be compared with the baseline reading taken once ALT normalize after introduction of antivirals to avoid confounding effect of raised ALT on TE readings [48]. The prognostic utility of LSM obtained by TE in HBV patients in regard to liver related outcomes and survival is being evaluated. One study showed increased risk of HCC in HBV patients who were determined to have cirrhosis by TE assessment [49].

2D Shear Wave Elastography (SWE).
Cleared by the FDA in December 2014, SWE technique is based on the combination of SW induced by multiple focused acoustic beams and a very rapid acquisition of ultrasound images (up to 20,000 images per second). In contrast to TE, an acoustic SW is generated and evaluated in SWE which can be coupled to the standard liver ultrasound examination for stiffness measurements. Multiple ultrasound systems that have SWE capabilities are commercially available in the United States, including those by Supersonic Imagine, SA (Aix-en-Provence, France; Ultrafast), and Logiq E9 ultrasound system (GE Healthcare, WI, USA).

Technical Aspects.
This application is able to capture the propagation of the resulting SW in real time and in a large area of liver parenchyma, developing real-time colorcoded elasticity imaging. Quantitative measurements can be performed in the color window by placing one or more ROI inside the sample box. The ROI of SWE is fan-shaped and larger than other modalities (up to 50 mm × 50 mm).
Results are given in m/s or kPa [58]. During the B-mode sonographic examination of the liver, a standardized ROI box is positioned in a predetermined anatomical site within the liver parenchyma for evaluation. The SW is generated from a focused ultrasound beam in the vicinity of the designated ROI. The velocity of the wave propagation, expressed in meters per second (m/s) or kilopascals (KPa), is calculated and allows LSM. The measurements are not limited by the presence of ascites as the ultrasound beam generating the SW propagates through fluids. The stiffer the liver is, the higher the recorded SW velocity is [59].

Diagnostic Performance and Accuracy.
As the application of SWE to liver fibrosis is a relatively newer approach, there are only a few studies evaluating its accuracy in liver diseases (Table 3). Poynard et al. in the largest liver SWE study to date (422 patients) showed the performance of SWE for the diagnosis of cirrhosis was similar to those of TE except in patients with ascites [52]. The better performance of SWE in patients with ascites may not be important clinically.
Another recent single center prospective study with 349 consecutive patients with chronic liver diseases compared the liver stiffness measurements obtained by SWE and TE with liver biopsy [53].  diagnosis of cirrhosis. SWE had a higher accuracy than TE for the diagnosis of severe fibrosis (⩾F3) but no statistically significant difference was observed for the diagnosis of mild fibrosis and cirrhosis [53]. The initial United States study was a pilot program that was recently published in September 2014. In this prospective study of 50 patients with chronic liver disease, the SWE measurements were compared to liver biopsy with a good accuracy for advanced fibrosis [60]. SWE is a new approach that has been applied to liver stiffness measurement. However its accuracy and performance in liver diseases need further validation in larger multicenter studies.

Technical Limitations.
With SWE only a few diseases specific LSMs are reported. The bulk of the studies have been performed in patients with HCV; therefore these cutoffs may not be applicable to other liver conditions. Only small series of patients with NAFLD have been reported and few studies addressing the confounding factors have been published. Another challenge in interpreting and obtaining data is a lack of standardization.  [64]. In addition, the authors performed a subgroup analysis; the Body Mass Index (BMI), the degree of liver inflammation, and the underlying etiology of the liver disease did not influence the accuracy of the MRE [64]. Similar results were noted in another meta-analysis in which 11 MRE studies (982 patients) and 15 ARFI studies (2,128 patients) were selected and modalities were compared [65]. Authors concluded that the MRE is more accurate than ARFI particularly in diagnosing early stages of hepatic fibrosis; AUROCs for MRE staging fibrosis were 0.94, 0.97, 0.96, and 0.97 for F1-F4, respectively, whereas AUROCs for ARFI staging were 0.82, 0.85, 0.94, and 0.94 for F1-F4, respectively [65]. Excitingly, MRE consistently has been reported as a highly accurate, noninvasive technique for the diagnosis and staging of liver fibrosis and is likely superior to US based techniques across all published studies. This has been noted across most of the liver diseases.

Technical Limitations.
As a new technique, many MRE related studies may not have captured all the potential confounding factors. High hepatic iron content may explain the failure rate noted in early MRE evaluations (signal-tonoise limitations). Patient related factors (inadequate breath hold/claustrophobia) and the cost may be additional limitations to the overall acceptance.

Conclusion
There is an increasing evidence for the diagnostic and prognostic utility of noninvasive methods of estimation of liver fibrosis and cirrhosis in patients with viral hepatitis. It is estimated that the use of noninvasive imaging based methods to assess liver fibrosis will increase tremendously in the near future. Accurate clinical use and understanding of reported findings will help in patient care and reduce the number of and thus morbidity associated with liver biopsies performed. In addition, imaging based methods have greater diagnostic utility than liver biopsy and can effectively be for follow-up assessments and they might have a role in prognostication of clinical liver related complications and mortality.