Current Concepts in Liver Disease: An Update, Part 2
Moderator: Dr. Linda D. Ferrell
Section 3 -
Update in Grading and Staging in Chronic Hepatitis: is Liver Biopsy Still Useful?
H˘pital Beaujon, Clichy
1. The changing pattern of liver biopsy reports: from a detailed analytical report to classification and scoring systems
Liver biopsy is a useful tool. In addition to provide help for the diagnosis or the etiology of a
chronic liver disease, liver biopsy provides valuable information on the severity of parenchymal
destruction and architectural disturbance. These are of major importance for long term prognosis in
chronic liver diseases. All of these data are usually put together in a comprehensive detailed biopsy
report and some of them should also be mentioned in an easily accessible and more understandable manner
for the hepatologist, i.e. scoring system.
Several scoring systems have been proposed. They respond all satisfactorily to this main objective
and differ only from one another on minor points.
The METAVIR scoring system assesses separately and linearly fibrosis and necroinflammation. A
detailed algorithm that allows providing a global score of activity according to two elementary features
(interface hepatitis and lobular inflammation)
has been defined . These scores have shown a good
reproducibility (excellent for score of fibrosis and acceptable for activity) and are largely diffused
among pathologists and hepatologist. The simplicity and reproducibility of the METAVIR score makes it
useful both for routine practice as well for clinical trials. The Ishak scoring system is also commonly
used . It is a modification of the Knodell system. Both grade and stage are more detailed and
provide more accurate information but to the disadvantage of reproducibility.
2. Why score liver lesions? Advantages of scoring liver biopsy
Scoring system provides major advantages both for cohort studies and also for individual patients:
- Scores are very useful for cohort studies. By categorizing patients
into homogeneous groups with the same extent and severity of a disease, scores add credit to large
multicentric epidemiological studies or clinical trials.
- In clinical practice and at the level of individual patients, scoring
systems are very useful to support medical decisions. The best example is the simplification in
treatment decision provided by scores of chronic viral hepatitis. In addition, scoring liver damages
allows the follow-up of the disease progression in a given patient spontaneously or under treatment. The
semi-quantitation provided by scores render the comparison of two biopsies performed in a same patient
easier than to compare detailed reports.
- In addition, a score makes a liver biopsy report more accessible to
non-specialized physicians and also to non-specialized pathologists. A score is easier to understand
than a formal report.
3. Limitations of scoring systems in liver biopsies
Although very useful, scoring has some drawbacks. Some are related to the liver biopsy procedure
itself while others are related directly to scoring systems.
- Statistical misuse
Scores are semi-quantitative values. Each scale integrates association of several different features,
all of them being descriptive lesions. At the end, scores are categories, not numbers. The consequence
is that scoring should not be handled statistically as linear values but as categories with adapted
statistical tests. Unfortunately, many studies in the literature, including clinical trials, did not
take this consideration into account, which limits the value of their scientific conclusions.
Although very easy to handle, a major drawback of a histological score is a reductive approach to very
few liver lesions. Fibrosis is a complex biological process with various development patterns. The
feature is integrated in other systems that might influence its development, rate of progression and
evolution of cirrhosis. Inflammation, necrosis, steatosis and iron overload are associated features of
significant importance which are not evaluated within a score of fibrosis. These associated features
should be carefully evaluated in order to develop tailored therapies adapted to each patient.
- Scores and sampling variability.
Even for chronic diseases supposed to be diffuse within the liver, sampling variability remains a
major drawback and sampling error strongly influences performance of a scoring system. Sampling
variability concerns the possible heterogeneous distribution of a histopathologic feature within the
liver. This variability depends on the type of pathological feature. While inflammation and steatosis
are relatively homogeneously distributed within the liver, fibrosis is much more heterogeneous. To limit
the extent of sampling variation, it is recommended to have a biopsy long and large enough, in
association with a scoring system with degree of precision that is not too high. It has been shown that
the longer the biopsy, then the lower the risk of sampling variability. The actual recommended minimal
size is 15 mm length with 11 portal tracts and the optimal size should be 25 mm. Even with these
thresholds, and taking the whole liver as the reference, 25 to 30% of biopsies are incorrectly scored for
fibrosis. Fortunately, most of discrepancies are only 1 stage in difference according to the referential
value. It is noteworthy that the complexity of a scoring system increases the risk of sampling error. A
scoring system that includes histopathologic features too rarely represented or too accurately evaluated
will not accommodate with the sampling limitation of a liver biopsy core. On the other hand, a too
simplistic score, although less sensitive to sampling error, will not provide enough information to the
clinician. Therefore an efficient score will be the one that has reached a delicate equilibrium between
the amount of information of the scoring system and the sampling variability.
- Scores and observer variation.
Observer variation relates to differences in scoring between pathologists. Interobserver variation is
a major limitation in semi-quantitative histopathology and is highly dependent of the scoring system.
Observer variation must be tested for any new classification or scoring system and only scores with low
or acceptable interobserver variation should be retained. For observer variation to be low, a scoring
system must be related to unambiguously, well-defined pathological items with accurate definitions of
each grade of the scale. Observer variation increases with the complexity of the scoring system. In
order to rule out observer variation, scoring by two observers in a common setting or by two observers
independently with resolution of discordances in a common setting are the methods of choice especially in
the context of clinical trials. When several biopsies have been performed in a same patient, scoring of
the more recent one should be performed in conjunction with re-reading of the older one. Finally, it is
recommended that in the context of clinical trials, all biopsies should be graded by the same
pathologist. A recent study from France has shown that observer variation is not only related to scoring
system but also to pathologist experience. Observer variation is lower between senior pathologists
specialized in liver pathology and working in academic environment than a junior or general pathologist.
4. Surrogates to liver biopsy
Due to the several drawbacks of the liver biopsy and also to the risk of complications, major efforts
have been developed to produce non-invasive alternative methods . They are based either on imaging or
serum parameters measurement. MRI and ultrasonography have acceptable sensitivity and specificity to
diagnose cirrhosis but they are clearly inefficient for the diagnosis of earlier stages of fibrosis or
other associated lesions. Therefore they are not designed to be used for chronic hepatitis monitoring
unless the purpose is to screen for cirrhosis.
More promising is elastometry (Fibroscan©)
. This method assesses velocity of propagation of a
shear wave in the upper part of the liver as measured by ultrasonography. This technique provides a
continuous numerical value related to the hardness of the liver and cut-offs have been defined according
to the different stages of fibrosis. Although inefficient to detect early stages of fibrosis,
elastometry is performed to assess precirrhotic stages (F3) and cirrhosis. Moreover, elastometry allows
the discrimination of range of severity among cirrhosis that should be parallel to risk of decompensation
and complications. The major drawback of this approach is the cost of the equipment, obesity that can
render the data difficult to interpret, and major steatosis that should reduce hardness of fibrotic
Serum markers of fibrosis have also been developed . Combination of serum parameters, mostly
unrelated to fibrosis but to liver cell function and inflammation have been widely developed to
indirectly assess fibrosis. After several years of utilization, it clearly appears that these serum
markers are useful to discriminate among mild fibrosis and significant fibrosis/cirrhosis, with close
performance one to each other. At the lower values of the scale they show very high negative predictive
values to exclude significant fibrosis and for the higher values, these markers have excellent positive
predictive values to diagnosis severe fibrosis but again, they are inaccurate to assess stage of
Non invasive test or liver biopsy in chronic hepatitis ?
Pathologists must be aware of non-invasive tests, their advantages and limits. These surrogates now
belong to the armament that clinicians have access to. They are easy to use and well-accepted by
patients and physicians. Obviously, they are not so accurate as the biopsy to score fibrosis.
Furthermore, non-invasive tests assess fibrosis out of its pathologic context, ignoring associated
lesions that should have major impact in treatment efficacy. However, when discussing the respective
roles of non-invasive tests and liver biopsy, patient should always be at the center of the discussion.
It is therefore clear that place of liver biopsy has changed. Serologic, virologic and molecular tests
make the biopsy most often unnecessary to find etiology of chronic hepatitis. When no benefit is
anticipated from information that biopsy could provide, biopsy should not be performed because of the
risk of complications. Patients with chronic hepatitis due to virus C of genotype 3, that will be
systematically treated or patients whose clinical condition prevent antiviral treatment should not be
biopsied. In contrast, and in regards to the limited efficacy and side effects of antiviral treatments,
biopsy should be performed in all other cases. Since fibrosis stages dictate treatment indications,
biopsy is the only procedure that will appreciate fibrosis within its pathological context that might
modulate treatment and follow-up.
- METAVIR cooperative study group. An algorithm for the grading of activity in chronic hepatitis. Hepatology 1996;24:289-93
- Ishak K, Baptista A, Bianchi L et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995;22:696-9.
- Rockey DC, Bissell DM. Noninvasive measures of liver fibrosis. Hepatology. 2006 Feb;43(2 Suppl 1):S113-20.
- Castera L, Vergniol J, Foucher J, Le Bail B, Chanteloup E, Haaser M, Darriet M, Couzigou P, De Ledinghen V.Prospective comparison of transient elastography, Fibrotest, APRI, and liver biopsy for the assessment of fibrosis in chronic hepatitis C. Gastroenterology. 2005 Feb;128(2):343-50.
- Oh S, Afdhal NH. Hepatic fibrosis: are any of the serum markers useful? Curr Gastroenterol Rep. 2001 Feb;3(1):12-8. Review