|

Received: 28 November 2017    Accepted: 29 November 2017

DOI: 10.1111/liv.13653

REVIEW ARTICLE

Diagnosis of non-alcoholic fatty liver disease/non-alcoholic

steatohepatitis: Why liver biopsy is essential

Pierre Bedossa

Service d’Anatomie Pathologique, Hôpital

Beaujon, Assistance Publique-Hôpitaux de
Paris, INSERM UMR 1149-CRI, Université
Denis Diderot Paris-7, Clichy, France
*Correspondence
Pierre Bedossa, M.D., Ph.D., Service
d’Anatomie Pathologique, Hôpital Beaujon,
Assistance Publique-Hôpitaux de Paris,
INSERM UMR 1149-CRI, Université Denis
Diderot Paris-7, Clichy, France.
Email: pierre.bedossa@inserm.fr
Funding information
GENFIT, Allergan, Intercept, Inventiva
Handling Editor: Francesco Negro
Abstract
The pattern of non-alcoholic fatty liver disease is complex with an association of several
lesions, each of them related to different pathophysiological mechanisms. Despite the
progress in non-invasive tools, liver biopsy remains the only diagnostic procedure that
can reliably assess these various patterns, their related severity and associations. The
most important difficulties of liver biopsy can be avoided if this procedure is performed
by an experienced hepatologist and read by a liver pathologist. However, for obvious
reasons, biopsy should be restricted to selected patients, especially in the context of clini-
cal trials. Indeed, liver biopsy is considered mandatory by regulatory authorities as a
­surrogate to assess drug efficacy in Phase 3 clinical trials. In addition to the clinical diag-
nosis, liver biopsy can be used to score the various histological patterns of disease (NASH-­
CRN, SAF scores) and accurately assess the extent of fibrosis, both of which are useful
when follow-­up biopsies are performed. When treatment becomes available in the near
future, liver biopsy could remain useful for choosing the most suitable therapeutic option
based on the main predominant histological features (activity, steatosis, fibrosis).

KEYWORDS
liver biopsy, non-alcoholic fatty liver disease, personalized treatment

1 | INTRODUCTION 2 | LIVER BIOPSY: A RELIABLE

Non-alcoholic fatty liver disease (NAFLD) covers a spectrum of histo-
logical lesions ranging from steatosis to a complex pattern that associ-
ates hepatocyte injury, inflammation and fibrosis. Although progress
has been made in the development of NIT (non-invasive tools), liver
biopsy is the only diagnostic procedure that can reliably assess these
various patterns and their association. These features are essential to
determine the treatment strategy and stratify the prognostic risk.1,2
However, as mentioned by L Castera, this is an invasive procedure
with a low, but real risk of morbidity and mortality. Thus, consider-
ing the large number of potential patients with NAFLD, liver biopsy
should be reserved for selected patients.
Abbreviations: CPA, collagen proportional area; FLIP, fally liver inhibition of progression;
NAFLD, non-alcoholic fatty liver disease; NASH, non-alcoholic steatohepatitis; NIT, non-­
invasive tool; SAF, steatosis-activity-fibrosis.
PROCEDURE
Most of the previously mentioned limitations of liver biopsy may be
avoided. It is often stated that since the volume of a needle biopsy sam-
ple represents only a very minor fraction of the entire liver, sampling var-
iations are relevant with a risk that is inversely proportional to the length
of the biopsy. Although a 25 mm biopsy is considered to be optimal for
assessing and quantifying detailed lesions, a 15 mm biopsy usually pro-
vides robust information for a global evaluation.3 Existing recommen-
dations on the size of the specimen were established for patients with
chronic viral hepatitis and may be different for NAFLD. Indeed, unlike
chronic hepatitis, NAFLD lesions tend to have a robust and characteristic
lobular systematization that mainly affects the centrilobular zone.4 Since
the size of a liver lobule is 0.5-­1 mm, the threshold for the minimally
required length may be lower. However, this must be confirmed.
Furthermore, the quality of the biopsy is highly dependent upon
who performs it and how. Liver tissue specimens should be collected

64  |  wileyonlinelibrary.com/journal/liv
© 2018 John Wiley & Sons A/S. Liver International. 2018;38(Suppl. 1):64–66.
Published by John Wiley & Sons Ltd
BEDOSSA
using a 16-­
gauge (or larger) cutting needle (eg Bard, Microvasive,
TruCut) whenever possible. Use of suction needles (eg Menghini,
Jamshedi, Klatskin) should be avoided as they often cause fragmenta-
tion of fibrotic specimens and prevent effective evaluation of fibrosis.
Then, a well-­trained hepatologist (or radiologist) with sufficient expe-
rience and motivation is essential for a successful procedure and like
in other technical procedures, experience is essential.
The expertise of the pathologist who performs the histological
evaluation is also important. Indeed, the FLIP pathology consortium
showed that agreement was higher among pathologists when biopsies
were interpreted by a group of specialized academic liver pathologists
than by pathologists from a more general practice.5 However, the
study shows that training with adequate histological guidelines mark-
edly increases the accuracy of interpretation whatever the level of the
pathologist and his/her academic training. Therefore, a pathologist is
reliable as long as the hepatologist (or radiologist) can provide an ad-
equate sample.
3 | LIVER BIO PSY: A USEFUL PROCEDURE
Performing a liver biopsy in a patient with clinical signs of NAFLD
allows an accurate evaluation of the different possible histological
components of the disease and determining their relationship, since
this can result in a significantly different prognosis. Indeed, biopsy
provides more essential information than the presence or absence of
NASH. The histopathological spectrum of NASH includes a mixture of
various lesions that should be categorized into four main groups: stea-
tosis, hepatocellular injury, inflammation and fibrosis. The list of the
elementary features associated with these processes is provided in
Table 1. A detailed description is beyond the scope of this article and
the reader should consult recent histopathological reviews.2,6 A full
assessment of these lesions is crucial to characterize the severity of
changes to obtain a prognosis and differentiate the processes that are
considered to be non-progressive and not at risk of increasing mortal-
ity from liver disease (simple steatosis) from features linked to a poor
long-­term prognosis and/or progression of liver injury (steatohepatitis
NASH). The diagnosis of steatohepatitis is based on the association of
T A B L E   1   Main histological patterns in NAFLD
Steatosis Type: macro-­, medio-­, microvacuole
Amount: usually in %
Location: zone 3, periportal, azonal, diffuse
Hepatocellular Ballooning and Clarification of cytoplasm
injury Apoptotic body
Mallory-­Denk body
Inflammation Location: portal, periportal, lobular
Inflammatory cell type
Extent
Fibrosis Location: perisinusoidal, perivenular, portal
Extent: focal, bridging fibrosis, annular fibrosis
Architectural modification
Other Vacuolated nuclei
Megamitochondria
|
      65
Key points
• So far, none of the non-invasive tools can replace liver
biopsy for the evaluation of the various histological pat-
terns of disease, their severity and their association in
NAFLD.
• Except for the risk of morbidity, most of the limitations of
liver biopsy can be avoided if adequate precautions are
taken.
• Liver biopsy may still be indicated in the future to choose
the most relevant personalized treatment based on the
dominant histological features of disease.
liver fat (macrovacuolar or mediovesicular steatosis of >5%), hepato-
cyte ballooning and lobular inflammation.4 Perisinusoidal fibrosis is a
useful and frequent diagnostic feature but is not formally included in
the diagnostic criteria of steatohepatitis. In the early stages, the pat-
tern of injury follows a centrilobular accentuation, although at later
stages, the lobular architecture is mutilated and the zonal distribution
is no longer visible. Other histological features can be found in steato-
hepatitis but are not necessary for the diagnosis of NASH. Portal in-
flammation is frequent in paediatric NASH but can be found in adults
and may be associated with more severe disease.
Large cohort studies based on histological definitions by a cen-
tral pathologist and with long-­term follow-­up of clinical events have
shown that the stage of fibrosis (bridging fibrosis or cirrhosis) is the
main predictor of liver-­related mortality.7,8 As steatohepatitis most
probably drives fibrogenesis, it may be difficult to statistically deter-
mine the independent effect of steatohepatitis from that of fibrosis
because of colinearity between the two variables. Staging of fibrosis
is based on the NASH CRN fibrosis stage.9 Unfortunately, this scoring
system underestimates perisinusoidal (pericellular) fibrosis within the
lobule which is a common pattern, especially in patients with diabe-
tes. Furthermore, it does not provide a distinction between biopsies
with rare or short septa from those with numerous septa (unlike the
distinction between F2 and F3 with the METAVIR score for chronic
hepatitis). This limits the results of biopsy since there is no clear-­cut
border to differentiate significant fibrosis from those with advanced
fibrosis. Morphometry, which quantitatively assesses the amount of
fibrous tissue (or Collagen Proportional Area, CPA) and is performed in
most clinical trials, can be useful in these cases.
Liver biopsy has another significant advantage of providing an ac-
curate semiquantitative evaluation of the severity of injuries. Indeed,
although the dichotomized diagnostic approach (NAFL vs NASH) is
clinically useful, it is an oversimplification that does not reflect the his-
tological complexity of the disease. Like all other chronic liver diseases,
NAFLD can display a continuous spectrum of histological lesions
and dividing the disease into two categories is useful but artificial.
Therefore, a semiquantitative scoring system provides a better image
of the complex histological pattern of disease. Although these scoring
systems have limited value in common practice, for the moment, they
 |
66       BEDOSSA
are extremely useful for clinical trials. Both the NASH CRN from the
United States and the European FLIP consortium have helped develop
a more accurate histological evaluation of NAFLDs. The NAS (NAFLD
Activity Score) developed by the NASH CRN is the unweighted sum
of steatosis (0-­3), inflammation (0-­3) and ballooning (0-­2).9 It is not
designed to be a surrogate for the diagnosis of steatohepatitis, but a
crude evaluation of the severity of disease once the diagnosis of NASH
has been established by an overall pathological assessment. Although
the NAS is correlated with aminotransferase and HOMA values, the
prognostic value of this score has not been demonstrated. While most
patients with a NAS <3 and >4 can be confirmed as NAFL and NASH,
respectively, there is a grey zone (NAS = 3 or 4) that includes both
cases with NAFL and NASH. Thus, the clinical relevance of its use as a
histological outcome in therapeutic trials is questionable.
The SAF score, which was prospectively designed by the FLIP
Pathology consortium, increases agreement for the overall diagnosis
among observers when used in association with the FLIP algorithm.5
This score precisely defines the three main cardinal histological features
(steatosis, activity and fibrosis) of NAFLD in a didactic and easy-­to-­
understand form. Activity is a composite score of hepatocellular bal-
looning and lobular inflammation, two lesions that, in association, are
supposed to sustain the development of fibrosis (each graded from 0 to
2).10 Furthermore, this is the backbone for the diagnosis of NASH by the
FLIP algorithm. The SAF score has not been tested in therapeutic trials.
4 |  OTHER ADDED VALUES OF LIVER
BIOPSY
In addition, due to the high burden of the disease, comorbidities are
frequent and biopsy can help determine their relative contribution.
Moreover, during treatment, liver biopsy is the only recognized proce-
dure to assess the effect of drugs in controlled clinical trials. It should
also be mentioned that thus far, regulatory authorities consider liver
histology (but not NIT) to be the primary endpoint of Phase 3 clinical
trials in interim analysis.11
Finally, histology may become a primary goal in the treatment of
NAFLD in the near future as the use of personalized medicine is devel-
oped. Treatment of NAFLD with fibrosis and mild activity will require
different drugs than very active disease and limited fibrosis. Only liver
biopsy can evaluate this information and help determine the best ther-
apeutic option in these patients.
CO NFL I C TS O F I NT ER ES T
The authors do not have any disclosures to report.
O RC I D
Pierre Bedossa  http://orcid.org/0000-0002-8487-7322
REFERENCES
1. Ludwig J, Viggiano TR, McGill DB, Oh BJ. Nonalcoholic steatohepati-
tis: Mayo Clinic experiences with a hitherto unnamed disease. Mayo
Clin Proc. 1980;55:434‐438.
2. Kleiner DE, Brunt EM. Nonalcoholic fatty liver disease: pathologic
patterns and biopsy evaluation in clinical research. Semin Liver Dis.
2012;32:3‐13.
3. Bedossa P, Dargère D, Paradis V. Sampling variability of liver fibrosis in
chronic hepatitis C. Hepatology. 2003;38:1449‐1457.
4. Brunt EM. Nonalcoholic steatohepatitis: definition and pathology.
Semin Liver Dis. 2001;21:3‐16.
5. The Flip Pathology Consortium. Utility and appropriateness of the
FLIP algorithm and SAF score in the evaluation of biopsies of nonal-
coholic fatty liver disease. Hepatology. 2014;60:565‐575.
6. Bedossa P. Pathology of non-­alcoholic fatty liver disease. Liver Int.
2017;37(suppl 1):85‐89.
7. Ekstedt M, Hagström H, Nasr P, et al. Fibrosis stage is the strongest
predictor for disease-­specific mortality in NAFLD after up to 33 years
of follow-­up. Hepatology. 2015;61:1547‐1554.
8. Angulo P, Kleiner DE, Dam-Larsen S, et al. Liver fibrosis, but no
other histologic features, is associated with long-­ term outcomes
of patients with nonalcoholic fatty liver disease. Gastroenterology.
2015;149:389‐397.
9. Kleiner DE, Brunt EM, van Natta M, et  al. Design and validation of
a histological scoring system for nonalcoholic fatty liver disease.
Hepatology. 2005;41:1313‐1321.
10. Bedossa P, Poitou C, Veyrie N, et al. Histopathological algorithm and
scoring system for evaluation of liver lesions in morbidly obese pa-
tients. Hepatology. 2012;56:1751‐1759.
11. Shadab Siddiqui M, Harrison SA, Abdelmalek MF, et al. Case defini-
tions for inclusion and analysis of endpoints in clinical trials for NASH
through the lens of regulatory science. Hepatology. 2017; in press.
https://doi.org/10.1002/hep.29607.
How to cite this article: Bedossa P. Diagnosis of non-alcoholic
fatty liver disease/non-alcoholic steatohepatitis: Why liver
biopsy is essential. Liver Int. 2018;38(Suppl. 1):64–66.
https://doi.org/10.1111/liv.13653