How far is noninvasive assessment of liver fibrosis from replacing liver biopsy in hepatitis C?
- File Under liver biopsy, noninvasive tests for fibrosis
How far is noninvasive assessment of liver fibrosis from replacing liver biopsy in hepatitis C?
Article first published online: 10 JAN 2012
DOI: 10.1111/j.1365-2893.2011.01518.x
Journal of Viral Hepatitis
Special Issue: How to Optimize Treatment of Hepatitis C
Volume 19, Issue Supplement s1, pages 18–32, January 2012
Download PDF
Download PDF
Abstract
Summary. Chronic hepatitis C represents a major cause of progressive liver disease that can eventually evolve into cirrhosis and its end-stage complications. Formation and accumulation of fibrosis in the liver is the common pathway that leads to evolutive liver disease. Precise staging of liver fibrosis is essential for patient management in clinical practice because the presence of bridging fibrosis represents a strong indication for antiviral therapy, while cirrhosis requires a specific follow-up. Liver biopsy has always represented the standard of reference for assessment of hepatic fibrosis, but it has limitations: it is invasive, costly and prone to sampling errors. Recently, blood markers and instrumental methods have been proposed for the noninvasive assessment of liver fibrosis in hepatitis C. However, international guidelines do not recommend the widespread use of noninvasive methods for liver fibrosis in clinical practice. This is because of, in some cases, unsatisfactory accuracy and incomplete validation of others. Some studies suggest that the effectiveness of noninvasive methods for assessing liver fibrosis may increase when they are combined, and a number of sequential and synchronous algorithms have been proposed for this purpose, with the aim of reducing rather than substituting liver biopsies. This may represent a rational and reliable approach for implementing noninvasive assessment of liver fibrosis in clinical practice. It could allow more comprehensive first-line screening of liver fibrosis in hepatitis C than would be feasible with liver biopsy alone.
Table 3. The role of liver biopsy across the main international guidelines
In 2007, the Asian Pacific Association for the Study of the Liver released a consensus statement about the management of CHC [30]. Overall, treatment is indicated in those patients with histological stage of F1 or above on liver biopsy. Patients with HCV genotype 2 or 3 can be treated regardless of stage. A liver biopsy is not mandatory before initiating therapy, especially if the subject is infected with HCV genotype 2 or 3. However, a liver biopsy before commencing therapy might provide information on prognosis.
Epidemiology and Natural History of Chronic Hepatitis C: the Impact of Liver Fibrosis
Chronic hepatitis C (CHC) is a major public health concern that affects about 200 million individuals worldwide and has a greater prevalence in Western countries [1]. Natural history studies indicate that approximately 10–20% of chronically infected patients with hepatitis C will develop severe liver cirrhosis and 1–5% will develop hepatocellular carcinoma (HCC) within two to three decades [2]. The event with the greatest impact on the natural history of CHC is the progressive formation and accumulation of liver fibrosis. This accumulation of fibrillar extracellular matrix (ECM) components is the hallmark of the natural history of the disease and significantly influences prognosis [2,3]. Indeed, there is a clear association between mild or no fibrosis at diagnosis and low risk of developing cirrhosis (25–30%) over the next 20 years. On the other hand, 100% of individuals with portal fibrosis develop cirrhosis, but this process takes 18–20 years. However, individuals with septal fibrosis develop cirrhosis sooner (8–10 years) [4]. Thus, early and accurate disease assessment and staging of fibrotic changes are critical for defining prognosis and indications for therapy in individual patients with CHC. In this view, two stages of liver fibrosis significantly modify the clinical management of patients with CHC: significant fibrosis, defined as at least portal fibrosis with few septa in liver histology, which represents a definitive indication to start antiviral treatment; and cirrhosis, which requires a specific follow-up including screening for oesophageal varices and for HCC by endoscopy, ultrasound and monitoring of alpha-fetoprotein levels [5].
Table 1. Staging systems for liver fibrosis in chronic hepatitis C
Liver Biopsy for Staging Fibrosis: the Perspective of Pathologists, Clinicians and Guidelines
Liver fibrosis staging systems
Liver biopsy has long remained the gold standard for assessment of liver fibrosis. It is based on direct histological assessment of the severity of liver disease, and several histological classification systems have been proposed to stage fibrosis and to grade necroinflammation in CHC (Table 1). The Knodell score is a quantitative score composed of four individual scores representing periportal and/or bridging necrosis (scored 0–10), intralobular degeneration and focal necrosis (scored 0–4), portal inflammation (scored 0–4) and fibrosis (scored 0–4) [6]. Although useful for assessing inflammatory activity, the Knodell score includes only three fibrosis stages, which limits its accuracy. The Ishak score, a revised version of the Knodell score that is frequently applied to CHC, preserves the grading and staging function of its predecessor while including six fibrosis stages [7]. Another widely used CHC classification system is the semiquantitative METAVIR system, which consists of grading and staging components [8,9].
Table 1. Staging systems for liver fibrosis in chronic hepatitis C
| Description | METAVIR (F) | Knodell | Ishak (F) | Batts–Ludwig (stage) |
|---|---|---|---|---|
| No fibrosis | 0 | 0 | 0 | 0 |
| Portal fibrosis without septa | 1 | 1 | 1–2 | 1–2 |
| Portal fibrosis with few septa | 2 | 1 | 3 | 3 |
| Septal fibrosis without cirrhosis | 3 | 2 | 4 | 3 |
| Cirrhosis | 4 | 3 | 5–6 | 4 |
Liver biopsy and the pathologist: the problem of sampling error
The quality and the size of the specimen obtained through a percutaneous liver biopsy in CHC are also an important issue. The main drawbacks lie in sampling error and intraobserver/interobserver variability. In a study by Regev et al. [10], samples taken from both right and left lobes revealed a difference of fibrosis stage of at least one grade in 33.1% of patients, while fibrosis was underdiagnosed in 14.5% of them. Cirrhosis may be missed on a single, blind percutaneous liver biopsy in 10–30% of cases [11,12]. Additionally, there is a significant degree of subjectivity in the pathologic assessment of liver biopsy samples. In a detailed study, Colloredo et al. [13] carefully analysed the impact of sample size on the correct staging of liver fibrosis in patients with CHC. By progressively reducing the dimensions of the same liver biopsy, they reported that the smaller the sample analysed, the milder the stage of fibrosis diagnosed by the pathologist. Interestingly, Rousselet et al. [14] reported that, even more than biopsy dimensions, the level of experience of the pathologist impacts on the diagnostic interpretation of liver biopsies. Pathologists have tried to define the features of an adequate liver biopsy sample. Some authors suggest that it should contain more than five complete portal tracts and be at least 15 mm in length [15,16]. In a critical review, Guido and Rugge [17] suggested that a biopsy sample of 20 mm or more containing at least 11 complete portal tracts should be considered reliable for grading and staging. Other authors have recommended even larger samples, up to 25 mm in length [18], and Scheuer stated that ‘bigger is better’[19]. Very recently, the American Association for the Study of Liver Diseases has recommended a biopsy sample of at least 20–30 mm in length and containing at least 11 complete portal tracts. Moreover, in clinical practice, a simple (METAVIR) rather than complex (Ishak) scoring system has been recommended for liver fibrosis [20]. In a recent report, Mehta et al. [21] suggested that liver biopsy is the best available standard of reference, but not the gold standard. Indeed, the performance of any surrogate is generally evaluated by the area under the receiving operating characteristic curve (AUC), using liver biopsy as the reference. Because liver biopsy is not the gold standard but the best available standard, a perfect surrogate will never reach a maximal value (that is 1). Taking into account a range of accuracies of the biopsy and a range of prevalences of significant fibrosis (that influence the AUC), the authors demonstrated that in the most favourable scenario, an AUC > 0.90 cannot be achieved when assessing significant fibrosis, even for a perfect marker [21].
Liver biopsy and the clinician: the problem of invasiveness
Liver biopsy has clear advantages, because it gives direct information not only about fibrosis, but also about necroinflammatory activity, steatosis, hepatic iron deposits and possible comorbidities. Nevertheless, it also presents some drawbacks for the clinician, including invasiveness and cost (Table 2). Pain is the most common complication of percutaneous liver biopsy, occurring in up to 84% of patients [22]. When moderate-to-severe pain occurs, usually in a small percentage of cases, the possibility of a more serious complication should be considered. The most important complication of a liver biopsy is bleeding, which occurs in 1/2500 to 1/10 000 cases. Death of a patient as a consequence of liver biopsy is very rare, less than or equal to 1/10 000 biopsies [20]. Regarding the impact of operator experience on the rate of complications, contradictory evidence exists. One study showed that the rate of complications in percutaneous liver biopsy was 3.2% if the operator had performed <20 biopsies and only 1.1% if the operator had performed more than 100 biopsies [23]. On the contrary, Chevallier et al. [24] showed that operator experience did not influence the final histological diagnosis or the degree of pain suffered by the patient.
| Advantages |
| Direct assessment of liver fibrosis |
| Semiquantitative |
| Evaluation of coexisting disorders |
| Limitations |
| Sampling error |
| Intraobserver and interobserver variability |
| Hospitalization often required |
| Complications |
| Pain |
| Haemorrhage |
| Pneumothorax |
| Haemothorax |
| Perforated viscus |
| Infection |
| Availability |
| High cost |
| Trained physician required |
| Contraindications |
| Absolute: uncooperative patient, severe coagulopathy and extrahepatic biliary obstruction |
| Relative: ascites, morbid obesity, possible vascular lesions and amyloidosis |
There is no clear consensus among clinicians about the role of liver biopsy in assessing liver fibrosis. A French survey of 1177 general practitioners concluded that liver biopsy may be refused by up to 59% of patients with hepatitis C and that 22% of the physicians share the same concern for the invasiveness of the procedure [25]. On this topic, a survey assessing consensus among Italian hepatologists on when and how to perform a liver biopsy in CHC showed great divergence in the management of the same subgroup of patients [26]. Another survey performed in a US clinical centre revealed that among 112 clinicians, 29.5% did not perform liver biopsy for the following reasons: concern about risks (72.7%), low reimbursement (66.7%) and logistic issues with space and recovery time (45.4%). Interestingly, biopsy without ultrasound guidance was routine practice for 53.2% of physicians [27]. A nationwide survey of French hepatologists regarding assessment of liver fibrosis in CHC revealed that only 4% of respondents still systematically performed liver biopsies [28].
Liver biopsy in the International Guidelines
The role of liver biopsy as reflected in the main international guidelines has evolved in recent years (Table 3). According to the 2002 Guidelines released by the National Institute of Health (NIH), liver biopsy is useful for defining the baseline state of liver disease and for informing treatment decisions by patients and healthcare providers regarding antiviral therapy [29]. Noninvasive tests do not currently provide the same information obtainable through liver biopsy. Information from a liver biopsy allows affected individuals to make more informed choices about the timing of antiviral treatment, thus liver biopsy can be a useful part of the informed consent process. Because a favourable response to current antiviral therapy occurs in 80% of patients infected with genotype 2 or 3, it may not always be necessary to perform a liver biopsy in these patients to make a treatment decision.
| Treatment guidelines | Role of liver biopsy |
|---|---|
| |
| NIH (2002) (29) | Liver biopsy is useful in defining baseline abnormalities of liver disease and supporting decisions regarding antiviral therapy. |
| Noninvasive tests do not currently provide the information that can be obtained through liver biopsy. | |
| Liver biopsy is a useful part of the informed consent process. | |
| As a favourable response to current antiviral therapy occurs in 80 per cent of patients infected with genotype 2 or 3, it may not be necessary to perform liver biopsy in these patients to make a decision to treat. | |
| APASL (2007) (30) | Treatment is indicated in those patients with histological stage of F1 or above on liver biopsy. |
| A liver biopsy is not mandatory to initiate therapy, especially if the subject is infected with HCV genotype 2 or 3. | |
| A liver biopsy before commencing therapy may provide information on prognosis. | |
| AASLD (2009) (5) | A liver biopsy may be unnecessary in HCV genotype 2 and 3 infection. An ongoing debate exists for patients infected with HCV-1 because of lower rates of SVR. |
| A liver biopsy should be considered in HCV patients to have more information for prognostic or therapeutic purposes. | |
| Available noninvasive tests may be useful in defining presence or absence of advanced fibrosis but should not replace liver biopsy in routine clinical practice. | |
| AISF (2010) (31) | Patients with normal ALT. |
| Antiviral treatment might be offered without the need for liver biopsy in patients with a high likelihood of achieving an SVR. | |
| In patients aged 50–65 years, and in those with a reduced likelihood of achieving an SVR, a liver biopsy may be used to evaluate the need for therapy, with treatment being recommended only for patients with fibrosis ≥F2 and a favourable HCV genotype. | |
| Biopsy and therapy are not recommended in elderly (>70 years) patients. | |
| Noninvasive assessments of fibrosis can be used to detect changes over time and indicate the need for biopsy or treatment on an individual patient basis. | |
| EASL (2011) | Assessment of the severity of liver fibrosis is important in decision-making. |
| Liver biopsy is still regarded as the reference method to assess the grade of inflammation and the stage of fibrosis. | |
| Transient elastography (TE) can be used to assess liver fibrosis. | |
| Noninvasive serum makers can be recommended for the detection of significant fibrosis (METAVIR score F2–F4). | |
| The combination of blood tests or the combination of transient elastography and a blood test improve accuracy and reduce the necessity of using liver biopsy to resolve uncertainty. | |
In 2007, the Asian Pacific Association for the Study of the Liver released a consensus statement about the management of CHC [30]. Overall, treatment is indicated in those patients with histological stage of F1 or above on liver biopsy. Patients with HCV genotype 2 or 3 can be treated regardless of stage. A liver biopsy is not mandatory before initiating therapy, especially if the subject is infected with HCV genotype 2 or 3. However, a liver biopsy before commencing therapy might provide information on prognosis.
More recently, the AASLD guidelines state that a liver biopsy should be considered in patients with CHC if the patient and healthcare provider want information regarding fibrosis stage for prognostic purposes or to make a treatment decision [5]. There is agreement that the high rate of sustained virological response in persons with genotype 2 or 3 infections means that biopsy is not necessary in these cases. There is, however, an ongoing debate about whether a biopsy is warranted for persons infected with HCV genotype 1, because of lower rates of SVR. Even more uncertain is whether a biopsy is necessary in persons infected with other, less common HCV genotypes (4 through 6). The currently available noninvasive tests may be useful for detecting the presence or absence of advanced fibrosis in persons with CHC infection, but should not replace the liver biopsy in routine clinical practice.
The recent guidelines from the Italian Association for the Study of the Liver put particular emphasis on the role of liver biopsy in patients with normal ALT levels [31]. Antiviral treatment might be offered without the need for liver biopsy in patients with a high likelihood of achieving an SVR (e.g. patients aged <50 years with highly treatable HCV genotype and low viral loads), in the absence of any contraindication and co-factors of poor responsiveness. In patients aged 50–65 years, and in those with a reduced likelihood of achieving an SVR, a liver biopsy may be used to evaluate the need for therapy, with treatment being recommended only for patients with more severe fibrosis (>F2) and a higher probability of responding, depending on the HCV genotype. Biopsy and therapy are not recommended in elderly patients (>70 years). Instead, it is recommended that these patients adopt lifestyle changes and undergo periodic ALT determinations. Noninvasive assessments of fibrosis can be used to detect changes over time and consequently indicate the need for biopsy or treatment on an individual basis.
The most recent guidelines about the management of CHC come from the European Association for the Study of the Liver (EASL) [32]. According to these, assessment of the severity of liver fibrosis is important in decision-making in patients with CHC. Liver biopsy is still regarded as the reference method to assess the grade of inflammation and the stage of fibrosis. Transient elastography can be used to assess liver fibrosis in patients with CHC. Noninvasive serum makers can be recommended for the detection of significant fibrosis (METAVIR score F2–F4). Interestingly, the EASL guidelines state also that the combination of blood tests or the combination of transient elastography and a blood test improve accuracy and reduce the necessity of using liver biopsy to resolve uncertainty.
In summary, current guidelines state that staging of liver fibrosis may be unnecessary in highly treatable HCV genotypes, while it may still have a role in HCV genotype 1 because of the lower SVR rate. Interestingly, the role of noninvasive methods for liver fibrosis has progressively evolved across the guidelines, starting from the NIH, in which they were not considered adequate, and moving towards the most recent guidelines (from American, Italian and European Associations), which do not recommend that noninvasive methods replace liver biopsy, but suggest that they may be used on an individual basis, especially to define the presence or absence of significant liver fibrosis and to reduce the number of liver biopsies.
Noninvasive Assessment of Liver Fibrosis: Serum Biomarkers
Millions of people worldwide are affected by CHC, but only 20–40% of them will ever develop liver fibrosis during their lifetime. Indeed, it would be unrealistic and extremely costly to stage liver fibrosis in all affected individuals with liver biopsies. On the other hand, fibrosis stage is the most important prognostic factor and is decisive for starting antiviral therapy in most cases. Currently, liver biopsy should be considered a diagnostic bottleneck for large-scale screening of liver fibrosis in CHC. Therefore, new noninvasive tests are necessary to limit the use of biopsies. During the past two decades, scientific interest has been focused in this direction and numerous potential serum biomarkers for the assessment of liver fibrosis have been evaluated [16,33].
Serum biomarkers can be broadly divided into direct or indirect. Direct markers are fragments of liver matrix components, such as hyaluronan and products of collagen synthesis or degradation produced by hepatic stellate cells (HSCs) during the fibrotic process, and the molecules involved in regulating fibrosis. Consequently, this group of biomarkers reflects the metabolism of hepatic ECM and has a pathophysiologic rationale. However, the routine clinical use of direct markers of liver fibrosis may be limited by test availability in some hospital settings.
In contrast, the indirect markers are biochemical parameters measurable in the peripheral blood that are routinely performed in patients with CHC. They are an indirect expression of liver damage and have a statistical association with liver fibrosis stage. These include molecules synthesized, regulated or excreted by the liver, such as clotting factors, bilirubin, transaminases and albumin.
While direct markers of liver fibrosis reflect the process of fibrogenesis, indirect markers satisfy the request for a simple and easy-to-perform marker [16,33]. An overview of the most validated biomarkers in CHC and of their performance is presented in Tables 4 & 5.
The accuracy of noninvasive methods for liver fibrosis varies among studies and is not considered adequate to warrant substituting this method for liver biopsy or implementing it in clinical practice. Indeed, although the most recent guidelines for management of CHC mention the possible use of noninvasive methods for liver fibrosis, they do not recommend their widespread use [5,31,32]. A number of studies have suggested that the diagnostic performance of noninvasive methods, especially for the diagnosis of significant fibrosis, may be improved by combining single tests into diagnostic algorithms [62]. The goal of combination algorithms is to use noninvasive methods when they have adequate accuracy, while reserving liver biopsy only to those patients in whom such tests are sufficiently accurate. In clinical practice, combination algorithms can provide the following information (and response): (i) presence/absence of significant fibrosis (whether to administer antiviral therapy); (ii) presence/absence of liver cirrhosis (whether to commence screening for oesophageal varices and HCC); and (iii) unclassifiable (liver biopsy needed to stage hepatic fibrosis). The combination approach may avoid the diagnostic bottleneck represented by liver biopsy, and it may stimulate general practitioners and patients to perform the initial screening for liver fibrosis in CHC. With this approach, liver biopsy and noninvasive markers for liver fibrosis may have a synergistic effect towards the goal of correctly classifying fibrosis in most patients with CHC. Table 8 summarizes a number of proposed sequential or synchronous combination algorithms for noninvasive detection of liver fibrosis.
Liver fibrosis staging represents a parameter of paramount importance for the management and prognosis of patients with CHC. Currently, it is inconceivable to stage fibrosis with liver biopsy in all patients because it is an invasive procedure with a number of drawbacks. Noninvasive methods are accurate for diagnosing both significant fibrosis and cirrhosis, the main clinical end-points. Combining noninvasive methods may optimize diagnostic performance. Algorithms for this purpose could drastically reduce the number of biopsies needed to correctly classify liver fibrosis in CHC. However, available data indicate that liver biopsy cannot be completely substituted with noninvasive assessment. The most rational approach is to screen patients first with algorithms that combine the most validated noninvasive methods for liver fibrosis and then perform a liver biopsy only when the accuracy of noninvasive methods is not reliable.
Giada Sebastiani and Alfredo Alberti contributed to the conception and first draft of the article. All authors approved the final version of the article. Alfredo Alberti is the guarantor. None of the authors of this manuscript has declared any conflict of interest within the last 3 years which may arise from being named as an author on the manuscript.
While direct markers of liver fibrosis reflect the process of fibrogenesis, indirect markers satisfy the request for a simple and easy-to-perform marker [16,33]. An overview of the most validated biomarkers in CHC and of their performance is presented in Tables 4 & 5.
Direct markers for liver fibrosis
The most investigated direct markers of liver fibrosis in CHC include hyaluronan, laminin, procollagen III, type IV collagen and YKL-40 (Tables 4 & 5). Hyaluronan is a glycosaminoglycan synthesized in HSCs and degraded by the liver sinusoidal cells [34] and has been extensively studied in CHC. In a study conducted in 326 patients, the AUC for significant fibrosis was 0.86 and the specificity 95%, while the AUC for cirrhosis was 0.92 and the specificity 89.4% when a cut-off level of 110 μg/L was used [34]. However, another cohort study with more than 400 cases has reported a lower AUC (0.73) for significant fibrosis [35]. In that study, cirrhosis could be excluded with a 100% negative predictive value (NPV) using a cut-off value of 50 μg/L; the AUC (0.97) was excellent as well. Similar results were reported in another study of 486 patients in which hyaluronan levels <60 μg/L excluded cirrhosis with a 99% NPV [36].
Laminin is a noncollagenous glycoprotein synthesized by HSCs. Its diagnostic value is inferior to that of hyaluronan and type IV collagen [37]. In a detailed study of 243 patients with chronic liver diseases, laminin was 77% accurate for detection of significant fibrosis in CHC [38]. Another study of 37 patients with CHC showed a slightly better performance (AUC = 0.82) [39].
YKL-40 is a glycoprotein that belongs to the chitinase family. It is strongly expressed in human cartilage and liver and preformed adequately at detecting significant fibrosis in 109 patients with CHC (AUC, 0.81; sensitivity, 78%; specificity, 81%) [40]. In the same study, however, performance in predicting cirrhosis was lower (AUC = 0.795).
Collagen molecules have also been used as markers. Type IV collagen performed well for detecting significant fibrosis (AUC = 0.83) [39]. Several studies evaluated the use of procollagen III in CHC. However, in comparative studies, procollagen III performed less well than type IV collagen and hyaluronan [34,40].
Panels of markers for liver fibrogenesis have also been proposed, as a strategy for increasing diagnostic performance. Fibrometer® (BioLiveScale, Angers, France) is a patented test that combines patient age, platelets, prothrombin index, AST, ![[alpha]](http://onlinelibrarystatic.wiley.com/undisplayable_characters/0003b1.gif)
2-macroglobulin, hyaluronan and urea and has an AUC in the range of 0.85–0.89 for significant fibrosis and 0.91 for cirrhosis in patient with CHC [41,42]. However, large-scale independent studies confirming this good performance are still lacking.
2-macroglobulin, hyaluronan and urea and has an AUC in the range of 0.85–0.89 for significant fibrosis and 0.91 for cirrhosis in patient with CHC [41,42]. However, large-scale independent studies confirming this good performance are still lacking.In the initial report on a panel of markers known as Fibrospect® (Prometheus Lab., San Diego, CA) [hyaluronan, tissue inhibitor of metalloproteinase-1 (TIMP-1) and 2-macroglobulin] indicated an AUC of 0.832 for significant fibrosis, with a positive predictive value (PPV) of 74.3% and a NPV of 75.8% [43]. Subsequent studies revealed an AUC ranging from 0.82 to 0.87 for diagnosis of significant fibrosis, with 71.8–93% sensitivity, 66–73.9% specificity and an overall test accuracy ranging from 73.1% to 80.2% [44–46]. The performance of hyaluronan, Fibrospect® and YKL-40 in diagnosing significant fibrosis in CHC were compared, and interestingly, the AUC of Fibrospect® was only 0.66, compared with hyaluronan, which had an AUC of 0.76 [47].
2-macroglobulin] indicated an AUC of 0.832 for significant fibrosis, with a positive predictive value (PPV) of 74.3% and a NPV of 75.8% [43]. Subsequent studies revealed an AUC ranging from 0.82 to 0.87 for diagnosis of significant fibrosis, with 71.8–93% sensitivity, 66–73.9% specificity and an overall test accuracy ranging from 73.1% to 80.2% [44–46]. The performance of hyaluronan, Fibrospect® and YKL-40 in diagnosing significant fibrosis in CHC were compared, and interestingly, the AUC of Fibrospect® was only 0.66, compared with hyaluronan, which had an AUC of 0.76 [47].A panel known as Hepascore® (Pathwest, University of Western Australia, Australia) (bilirubin, ![[gamma]](http://onlinelibrarystatic.wiley.com/undisplayable_characters/0003b3.gif)
GT, hyaluronan, 2-macroglobulin, age and sex) performed very well in CHC, with AUC ranging from 0.79 to 0.85 for diagnosis of significant fibrosis and from 0.89 to 0.94 for cirrhosis [42,48]. Becker et al. [49] validated Hepascore® in almost 400 patients with CHC, reporting an AUC for significant fibrosis of 0.81–0.83.
GT, hyaluronan, 2-macroglobulin, age and sex) performed very well in CHC, with AUC ranging from 0.79 to 0.85 for diagnosis of significant fibrosis and from 0.89 to 0.94 for cirrhosis [42,48]. Becker et al. [49] validated Hepascore® in almost 400 patients with CHC, reporting an AUC for significant fibrosis of 0.81–0.83.The ELF study group has proposed a panel of direct noninvasive markers that includes age, hyaluronan, type III collagen and TIMP-1. In a cohort study of more than one thousand patients with chronic liver disease, this panel detected advanced fibrosis with AUC 0.77 in patients with CHC [50]. Unfortunately, large-scale independent studies are still lacking for most of these panels.
Indirect markers for liver fibrosis
Indirect noninvasive markers for liver fibrosis include a growing number of serum parameters and their combinations that are generally economical and routinely measured in patients with CHC. The AST-to-ALT ratio (AAR) was one of the first indirect markers for staging liver fibrosis in patients with CHC. An increase in AAR reflects progressive impairment of liver functional (normal value <0.8), while a ratio >1 is indicative of cirrhosis [51]. In CHC, the AAR has an accuracy ranging from 60% to 83.6%, sensitivity between 31.5% and 81.3% and specificity between 55.3% and 100% in distinguishing cirrhotic from noncirrhotic patients [51,52]. The overall performance of AAR was very variable across studies, with an AUC ranging from 0.51 to 0.83. The AAR does not identify significant fibrosis, and its value can be altered by the use of alcohol [38].
The AST-to-platelet ratio index (APRI) has been proposed as another simple score. It is calculated through AST and platelet count, thus it has virtually no cost [53]. APRI can be used to confirm or exclude both significant fibrosis (cut-off 1.5 and 0.5, respectively) and cirrhosis (cut-off 1 and 2, respectively), but in both cases, there is an intermediate range of values in which the performance is not satisfactory and 30–50% of cases cannot be classified. To date, APRI is one of the most investigated noninvasive markers for liver fibrosis [16]. In the initial study by Wai et al. [53], APRI was highly accurate for predicting significant fibrosis and cirrhosis, with an AUC of 0.88 and 0.94, respectively. More recent studies, however, show variable performance in CHC, with AUC ranging between 0.69 and 0.88 for significant fibrosis and between 0.61 and 0.94 for cirrhosis [16,54]. Interestingly, a recent systematic review of the diagnostic accuracy of APRI indicated that the major strength of this index is its good performance in excluding HCV-related fibrosis [55]. The authors concluded that future studies of novel markers need to demonstrate improved accuracy and cost-effectiveness compared with this economical and widely available index. A further evolution of APRI is the Lok index, which combines platelet count, International Normalised Ratio (INR) and AAR [56]. The Lok index uses two cut-off values for the diagnosis of cirrhosis: 0.2 to rule out cirrhosis and 0.5 to confirm cirrhosis. Lok index values between 0.2 and 0.5 are considered unclassified. In a cohort study of 1141 patients with CHC, Lok et al. (56) report an AUC of 0.78–0.81 for diagnosis of cirrhosis, with 86% NPV for 0.2 cut-off and 75% PPV for 0.5 cut-off. Subsequent studies reported similar diagnostic performance but no clear advantage compared with APRI [57]. The main limit of this index is that it does not provide information about significant fibrosis. Forns et al. [58] developed a simple panel for the prediction of fibrosis based on routine clinical variables: age, GT, cholesterol levels and platelet count. Forns’ index uses two cut-off values, 4.2 to exclude significant fibrosis and 6.9 to confirm significant fibrosis. Values between 4.2 and 6.9 are considered unclassified. The index has been extensively investigated in CHC. In a detailed study including 476 patients with CHC, Forns’ index had good diagnostic performance (AUC of 0.81–0.86) for diagnosis of significant fibrosis [58]. Notably, the lower cut-off value (4.2) had 96% NPV to exclude significant fibrosis. On the other hand, the upper cut-off value (6.9) had only 66% PPV to confirm significant fibrosis. Interestingly, subsequent studies reported a slightly lower performance, with an AUC ranging from 0.76 to 0.79 [42,54]. Its main limits are a lack of information regarding cirrhosis and a significant number of unclassified cases. Another combination of simple markers, named Fib-4, is based on AST, ALT, age and platelet count [59]. Fib-4 uses two cut-off values: 1.45 to exclude significant fibrosis and 3.25 to confirm significant fibrosis. In a detailed study of 592 HCV-infected patients, Fib-4 correctly identified patients with severe fibrosis with an AUC of 0.85 [59]. Similar results have been reported by other authors [60]. This index does not provide information about cirrhosis and leaves a considerable group of patients unclassified.
GT, cholesterol levels and platelet count. Forns’ index uses two cut-off values, 4.2 to exclude significant fibrosis and 6.9 to confirm significant fibrosis. Values between 4.2 and 6.9 are considered unclassified. The index has been extensively investigated in CHC. In a detailed study including 476 patients with CHC, Forns’ index had good diagnostic performance (AUC of 0.81–0.86) for diagnosis of significant fibrosis [58]. Notably, the lower cut-off value (4.2) had 96% NPV to exclude significant fibrosis. On the other hand, the upper cut-off value (6.9) had only 66% PPV to confirm significant fibrosis. Interestingly, subsequent studies reported a slightly lower performance, with an AUC ranging from 0.76 to 0.79 [42,54]. Its main limits are a lack of information regarding cirrhosis and a significant number of unclassified cases. Another combination of simple markers, named Fib-4, is based on AST, ALT, age and platelet count [59]. Fib-4 uses two cut-off values: 1.45 to exclude significant fibrosis and 3.25 to confirm significant fibrosis. In a detailed study of 592 HCV-infected patients, Fib-4 correctly identified patients with severe fibrosis with an AUC of 0.85 [59]. Similar results have been reported by other authors [60]. This index does not provide information about cirrhosis and leaves a considerable group of patients unclassified.The most validated noninvasive serum test in CHC: Fibrotest®
Fibrotest® (Biopredictive, Paris, France) (GT, total bilirubin, haptoglobin, apolipoprotein A1 and 2-macroglobulin – adjusted for gender and age) is the most validated liver fibrosis panel in CHC, with more than 50 studies conducted [16,61,62]. Although many of these were conducted by the group that patented the test, the total number of patients included in independent studies approaches 5,000 [41,42,54,59,63–67]. Factors causing error for Fibrotest® include conditions that alter its single components, including Gilbert’s syndrome, haemolysis and extrahepatic cholestasis. The first report suggested that Fibrotest® values (from 0 to 1) correlate with liver fibrosis stages according to METAVIR classification [68]. The AUC of Fibrotest® ranges from 0.74 to 0.87 for significant fibrosis and from 0.71 to 0.87 for cirrhosis [16,54,65,68]. One of the first independent, prospective studies to compare the performance of Fibrotest® to other noninvasive markers was from our Unit [54]. In this comparative study, the performance of Fibrotest®, APRI and Forns’ index was tested in 190 patients with CHC. Fibrotest® was the most accurate: AUC 0.81 for significant fibrosis and 0.71 for cirrhosis. A recent systematic review by independent investigators included nine studies, with a total of 1679 cases of CHC [69]. The authors found that Fibrotest® has excellent diagnostic accuracy for identifying of HCV-related cirrhosis, but is less useful for earlier stages of fibrosis. They concluded that Fibrotest® and other noninvasive tests for liver fibrosis are not ready to replace liver biopsy. Interestingly, the use of Fibrotest® has been recommended recently in France by the Haute Autorité de Santé for first-line assessment of liver fibrosis in patients with CHC, because validation in several studies was considered adequate.
Table 7. Limitations and risk factors for error with Fibroscan®
GT, total bilirubin, haptoglobin, apolipoprotein A1 and 2-macroglobulin – adjusted for gender and age) is the most validated liver fibrosis panel in CHC, with more than 50 studies conducted [16,61,62]. Although many of these were conducted by the group that patented the test, the total number of patients included in independent studies approaches 5,000 [41,42,54,59,63–67]. Factors causing error for Fibrotest® include conditions that alter its single components, including Gilbert’s syndrome, haemolysis and extrahepatic cholestasis. The first report suggested that Fibrotest® values (from 0 to 1) correlate with liver fibrosis stages according to METAVIR classification [68]. The AUC of Fibrotest® ranges from 0.74 to 0.87 for significant fibrosis and from 0.71 to 0.87 for cirrhosis [16,54,65,68]. One of the first independent, prospective studies to compare the performance of Fibrotest® to other noninvasive markers was from our Unit [54]. In this comparative study, the performance of Fibrotest®, APRI and Forns’ index was tested in 190 patients with CHC. Fibrotest® was the most accurate: AUC 0.81 for significant fibrosis and 0.71 for cirrhosis. A recent systematic review by independent investigators included nine studies, with a total of 1679 cases of CHC [69]. The authors found that Fibrotest® has excellent diagnostic accuracy for identifying of HCV-related cirrhosis, but is less useful for earlier stages of fibrosis. They concluded that Fibrotest® and other noninvasive tests for liver fibrosis are not ready to replace liver biopsy. Interestingly, the use of Fibrotest® has been recommended recently in France by the Haute Autorité de Santé for first-line assessment of liver fibrosis in patients with CHC, because validation in several studies was considered adequate.Noninvasive Assessment of Liver Fibrosis: Transient Elastography (Fibroscan®)
The measurement of liver stiffness by transient elastography is a validated method for the noninvasive assessment of liver fibrosis. Liver stiffness is measured through a device called Fibroscan® (Echosens, Paris, France), which consists of an ultrasound transducer probe mounted on the axis of a vibrator. Vibrations of mild amplitude and low frequency are transmitted by the transducer, inducing an elastic shear wave that propagates through the underlying tissues: the stiffer the tissue, the faster the shear wave propagates [70]. Fibroscan® examination is painless and rapid (<5 min). It is performed with the patient in the supine position, with the right arm tucked behind the head. The probe transducer is placed on the skin, between the rib bones at the level of the right lobe of the liver where a biopsy would be performed. The operator performs 10 valid acquisitions, and then Fibroscan® software calculates the median value. The software itself determines whether each measurement is successful or not. Results are expressed in kiloPascals (kPa). Liver stiffness values range from 2.5 to 75 kPa. According to the manufacturer, the validity of a Fibroscan® examination should be based on two parameters: the interquartile range (IQR), which reflects the variability of the validated measures, should not exceed 30% of the median value; and the success rate, which is the percentage of valid measurement, should be at least 60%[64,70]. Fibroscan® results are interpreted on the basis of cut-off values expressed in kPa: according to various studies of patients with CHC, significant fibrosis is defined by a cut-off value ranging from 5.2 to 8.9 and cirrhosis is diagnosed by a cut-off value ranging from 10.1 to 17.6 [64,70–75]. Table 6 summarizes some of the studies assessing the performance of Fibroscan® in CHC. In a number of studies, the accuracy of Fibroscan® was similar to that of noninvasive serum markers for the diagnosis of significant fibrosis, sometimes with inadequate figures (AUC < 0.80). On the other hand, for the diagnosis of cirrhosis, all studies revealed an excellent AUC (>0.90). Indeed, a recent meta-analysis investigated the performance of Fibroscan® [76]. The authors concluded that for the diagnosis of significant fibrosis, Fibroscan® is not sufficient for use in clinical practice. Inclusion of Fibroscan® in an algorithm with a combination of noninvasive serum markers may be considered. On the other hand, Fibroscan® can be used in clinical practice as an excellent tool for the confirmation of cirrhosis when other parameters are inconclusive. Like Fibrotest®, the French health authorities also consider Fibroscan® to be validated for first-line assessment of liver fibrosis in patients with CHC.
Table 6. Performance of Fibroscan® in chronic hepatitis C in various studies
Table 6. Performance of Fibroscan® in chronic hepatitis C in various studies
| Reference | Cut-off for ≥F2 (kPa) | Cut-off for F4 (kPa) | AUC for ≥F2 | AUC for F4 | Number of patients included |
|---|---|---|---|---|---|
| 70 | 7.6 | 14.4 | 0.88 | 0.99 | 106 |
| 64 | 7.1 | 12.5 | 0.83 | 0.95 | 183 |
| 74 | 8.7 | 14.5 | 0.79 | 0.97 | 327 |
| 75 | 6.8 | 17.6 | 0.79 | 0.91 | 935 |
| 73 | 7.8 | 14.8 | 0.91 | 0.98 | 150 |
| 72 | 8.9 | 10.1 | 0.89 | 0.97 | 187 |
| 71 | 5.2 | 12.9 | 0.75 | 0.90 | 913 |
Limitations and risk factors for error with Fibroscan®
Considering the widespread use of the Fibroscan® technique, a number of studies have analysed risk factors for error and the limitations of Fibroscan® (Table 7). Fibroscan® can be difficult to perform in obese patients or in those with narrow intercostal space and is impossible in patients with ascites [70]. Failure rates range between 2.4% and 9.4% among studies [64,70,77]. The reproducibility of Fibroscan® examination has been investigated in a detailed study of 200 patients by Fraquelli et al. [77]. Considering intraobserver and interobserver agreement, the authors concluded that Fibroscan® is highly reproducible. Factors associated with inter- and intraobserver variability were BMI > 25, high-grade hepatic steatosis and mild fibrosis (F0–F1 by METAVIR). Although Fibroscan® has excellent reproducibility, its applicability may not be as wide as that of biomarkers. In a very recent study, liver stiffness measurements were uninterpretable in nearly one in five cases (failure to obtain any measurement in 4% and unreliable results not meeting the manufacturer’s recommendations in 17%) [78]. The principal reasons were obesity and limited operator experience. A report suggested that acute viral hepatitis increases liver stiffness measured by Fibroscan®, thus the authors recommend that the extent of necroinflammatory activity should be carefully considered in future studies, particularly in patients with absent or low-stage liver fibrosis [79]. Recently, Millonig et al. [80] reported that liver stiffness correlates significantly with bilirubin levels in patients with extrahepatic cholestasis, which causes false-positive Fibroscan® results. Accordingly, Fibroscan® results tend to normalize following successful biliary drainage. Another cause of false positivity has been suggested by a case report in which vascular hepatic congestion led to elevated Fibroscan® results to a level unambiguously diagnostic for liver cirrhosis [81]. This increase in elastometry was entirely reversible upon correction of cardiovascular dysfunction. A detailed study of 254 patients with CHC showed that the ratio of IQR to ‘median value of liver stiffness’ was associated with discordance between Fibroscan® measurement and liver biopsy, indicating overestimation of liver fibrosis [82].
Table 7. Limitations and risk factors for error with Fibroscan®
| Limitations | Risk factors for error |
|---|---|
| Failure in 5% of cases (especially obese) | Acute viral hepatitis |
| Unreliable results in 15% of cases (obesity, ascites and limited operator experience) | Obesity |
| Interobserver and intraobserver variability influenced by liver steatosis, overweight and mild fibrosis stages | Extrahepatic cholestasis |
| Lower performance for diagnosis of significant fibrosis | Vascular hepatic congestion |
| Unable to discriminate between intermediate stages of fibrosis | Ratio interquartile range/median value |
Stepwise and Synchronous Combination Algorithms of Noninvasive Methods for Liver Fibrosis
Stepwise combination algorithm: SAFE biopsy and Bourliere’s algorithm
We proposed an approach called Sequential Algorithms for Fibrosis Evaluation (SAFE) biopsy that combines APRI and Fibrotest® in series, with the aim of increasing diagnostic accuracy and reducing the number of liver biopsies necessary to correctly stage liver fibrosis, while minimizing misclassified cases [54]. These two methods were chosen because they are the most validated biomarkers in the literature and they are widely available. The stepwise modelling of the algorithms for significant fibrosis and cirrhosis aimed to achieve >90%. In the model, APRI is used as a first-line test because it is simple and economical, while Fibrotest® is used as a second-line test because it is more accurate but costly. Liver biopsy is reserved as a third-line test in cases where noninvasive markers do not show adequate accuracy and/or in unclassified cases (only for APRI). In the initial study on patients with CHC, SAFE biopsy algorithms had excellent performance for both significant fibrosis and cirrhosis. More recently, SAFE biopsy has been validated in a multicenter, international study that enrolled more than 2035 patients with CHC for whom APRI and Fibrotest® were available; liver histology was used as a reference standard [83]. To date, this is the largest independent study on noninvasive biomarkers for liver fibrosis. As shown in Table 8, diagnostic performance was excellent and SAFE biopsy was able to avoid about 50% of liver biopsies for significant fibrosis and 80% for cirrhosis.
A stepwise algorithm combining Hepascore®, a patented test, and APRI was recently proposed [48,67]. The authors reported high diagnostic accuracy (91%), while avoiding 45% of liver biopsies for diagnosing significant fibrosis. However, the main limitation of this algorithm is that Hepascore® is not as validated in the literature as APRI, Fibrotest® and Forns’ index. Moreover, validation studies are needed to confirm the very good accuracy reported in this initial study.
Synchronous combination algorithms: Bordeaux algorithm, Leroy algorithm, Fibropaca algorithm and Angers’ algorithm
Castera et al. [64] have proposed an algorithm based on the concordance between Fibrotest® and Fibroscan® in CHC. This algorithm results in increased accuracy, especially for the diagnosis of significant fibrosis, with respect to the single methods. A recent collaborative study compared the algorithm combining Fibroscan® and Fibrotest® (the Bordeaux algorithm) and SAFE biopsy in 302 patients with CHC [84]. For the diagnosis of significant fibrosis, the Bordeaux algorithm avoided more liver biopsies, but SAFE biopsy had a significantly higher performance. For the diagnosis of cirrhosis, the Bordeaux algorithm showed a significantly higher accuracy. On the other hand, the Bordeaux algorithm uses Fibrotest® and Fibroscan® in all patients, while SAFE biopsy uses Fibrotest®, which has virtually no cost, in the subgroup of patients that are not well classified by APRI.
A synchronous combination approach has also been proposed for clinical use. Bourliere et al. [85] proposed a combination algorithm of APRI, Fibrotest® and Forns’ index, which performed very well for both significant fibrosis and cirrhosis, avoiding about 50% of liver biopsies for significant fibrosis and about 80% for diagnosis of cirrhosis. Another synchronous algorithm proposed by Leroy et al. [42], based on the concordance of Fibrotest® and APRI, had excellent performance for the diagnosis of significant fibrosis; however, relatively few liver biopsies were avoided, compared with other combination algorithms. Cales and colleagues reported that a combination of Fibrometer® and Fibrotest®, named Angers’ algorithm, may avoid 44.8% of liver biopsies with an overall accuracy of 95.3% [86]. The authors suggest that synchronous combination algorithms may be more effective than sequential algorithms such as SAFE biopsy. However, their synchronous algorithm combining Fibrometer® and Fibrotest® has some disadvantages when compared to SAFE biopsy. First, it uses two patented tests, with an average cost of €100 each, in all patients, while SAFE biopsy uses only one (Fibrotest®) and only in about half of the cases. Another important issue is that the SAFE biopsy, Fibropaca and Bordeaux algorithms all employ widely available and validated tests. Although Fibrometer® performed well in studies by the patenting group, its external validation is poor [41,66].
Combination algorithms of noninvasive methods in clinical practice: a synergistic approach aimed at reducing rather than eliminating liver biopsy
A less invasive approach to staging liver fibrosis in CHC undoubtedly implies many advantages in clinical practice, including broader first level screening, higher patient compliance and lower screening costs. Some special subgroups may benefit even more from less invasive screening, such as elderly HCV carriers and those with normal ALT levels. Indeed, a number of eminent hepatologists have recently agreed on the strategy of combining noninvasive methods to improve diagnostic accuracy for liver fibrosis.
In a recent review, Pinzani et al. [87] have proposed applying two unrelated noninvasive methods in CHC and performing liver biopsy only in a subgroup of patients. Similarly, Manning and Afdhal [33] have proposed annual measurements of biomarkers and Fibroscan® in patients with CHC. The choice of algorithm for use in clinical practice can be based on local availability of component tests, test reliability and validation, and patient comorbidities. Importantly, the APASL has recently produced the first consensus recommendations on liver fibrosis, which also recommend the stepwise algorithm approach. The main conclusions of these guidelines are as follows: (i) noninvasive tests are useful for identifying only those patients with no fibrosis or extreme levels of fibrosis; (ii) staging of liver fibrosis in the intermediate range cannot be satisfactorily predicted by any of the available tests; and (iii) a stepwise algorithm incorporating noninvasive markers of fibrosis may reduce the number of liver biopsies by about 30% [88].
0 comments: