A Clinical and Histopathologic Perspective on Evolving Noninvasive and Invasive

[Guest guest]

A Clinical and Histopathologic Perspective on Evolving Noninvasive

and Invasive Alternatives for Liver Biopsy

Clinical Gastroenterology and Hepatology, May 2008

Dirk J. Van LeeuwenâŽ, Balabaud‡, M.

Crawford§∥¶, ette Bioulac†" Sage§∥#, Amar P. DhillonâŽâŽ

⎠Section of Gastroenterology and Hepatology, Dartmouth Medical

School, Hanover, New Hampshire

‡ CHU Bordeaux, Pôle Hépato-Gastoentérologie Hôpital Saint

André, Bordeaux, France

§ INSERM E362, Bordeaux, France

∥ Université Victor Segalen Bordeaux 2, Bordeaux, France

¶ Department of Pathology, Immunology and Laboratory Medicine,

University of Florida, Gainesville, Florida

# CHU Bordeaux, Service d'Anatomie Pathologique Hôpital Pellegrin,

Bordeaux, France

âŽâŽ Department of Histopathology, Royal Free and University

College Medical School, London, United Kingdom

Noninvasive or minimally invasive alternatives are proposed as

substitutes for liver biopsy and include clinical indices, cross-

sectional imaging, serum biomarkers, liver stiffness measurement, and

portal pressure measurement. Most alternatives to liver biopsy assess

one aspect of liver disease and translate this into a numeric score.

Overlap between categories may limit applications. Liver biopsy

provides information about numerous variables: tissue architectural

changes; necroinflammatory injury; fibrotic stage; alterations of

parenchyma and bile duct epithelium; accumulation of fat, copper, and

iron; and molecular and genetic changes. Liver biopsy may identify

multiple disease etiologies. A single numeric score cannot be a

substitute for complete histologic assessment. However, within

defined clinical contexts, noninvasive assessment is an attractive

alternative for many patients given the ease, avoidance of risk from

invasive procedures, and validated contribution to clinical

management. Serum biomarkers and liver stiffness assessment may

become indispensable in longitudinal studies and to document outcome

of treatments. The accuracy of the more reliable techniques is

typically around 80%. Neither liver biopsy nor any single alternative

option represents an absolute assessment of liver disease. Biopsy and

alternatives are not mutually exclusive options. Liver biopsy and the

noninvasive alternatives require a clear understanding of

significance and limitations of each investigation. This places a

responsibility on the clinician to consider fully the results of any

of the investigative options used within the diagnostic and

prognostic context of each individual patient, and to choose

critically the most appropriate investigations for the patient's

needs.

Article Outline

• Abstract

• Underlying Assumptions When Applying Less-Invasive Testing

• Challenges

• Categoric Designations

• Comorbidity

• Closing Comments

The indications for liver biopsy continue to change because many of

the clinical requirements for liver biopsy (eg, evaluation of disease

grade and stage) increasingly often can be met by other

investigations.1, 2, 3 Conversely, morphologic assessment with

molecular reagents and digital imaging allows an increasingly

rigorous and sophisticated histopathologic assessment of tissue

inflammation and fibrosis.4, 5, 6 Alternatives to liver biopsy have

been proposed and are deemed to be as good as biopsy and less

damaging to the patient and include predictive tests for assessment

of fibrosis, inflammation, and fat.7 Biopsy limitations include cost,

complications, and sampling error.8, 9, 10, 11, 12 Liver biopsy is

not necessarily the gold standard for staging and grading of disease,

and imaging, for example, under some circumstances can be

preferable.13 Also, therapeutic strategies are evolving and include

the use of viral load and kinetics (hepatitis B virus DNA, hepatitis

C virus [HCV] RNA) rather than histopathologic end points.14, 15, 16,

17 This places liver biopsy in a new perspective: if all that is

required for making a decision about clinical management is evidence

of (some) scar formation, and viral clearance is the aim, to what

degree does liver biopsy histopathology still matter? The prevalence

of hepatitis B and C infection (500†" 800 million people worldwide),

make it neither feasible nor realistic to perform a liver biopsy in

all patients. Treatment is costly and side effects can be very

disabling. Being able to target high-risk groups for treatment, while

postponing treatment for those at low risk in anticipation of

improved therapy, will allow for optimal use of resources†" including

liver biopsy.18, 19 Resources thus could be reallocated from

performance of liver biopsy to treatment of viral infection.

Affordable and simple diagnostic alternatives would serve broad

populations of patients. Currently proposed alternatives to liver

biopsy include imaging, blood tests (biomarkers), clinical indices,

and liver stiffness measurement,20, 21 and a number of these are

summarized in Table 1. Some have advocated cessation of liver biopsy

now that quality biomarkers are available.22 Others favor

validation,23, 24 stepwise combination algorithms of noninvasive

markers to diagnose significant fibrosis,25, 26 or monitoring of long-

term disease evolution.27

We are members of an international liver pathology study group and we

provide our views of the merits and pitfalls on the use of the

increasing number of alternatives that have been proposed as

replacements for liver biopsy.

Underlying Assumptions When Applying Less-Invasive TestingÂ

Proposing alternatives to liver biopsy prompts the question of how

they compare with liver biopsy? What are the underlying assumptions

when applying such tests?

A liver biopsy specimen is an estimated 1/50,000th of the total

liver. It is remarkable how often it is asserted that liver biopsy is

the reference, the gold standard, against which all other techniques

should be measured. This happens despite abundant evidence that both

in acute and chronic disease the risk of biopsy sampling error is

considerable because of uneven distribution of pathologic features.

Sampling errors are associated particularly with underestimation of

liver disease stage.8, 9, 11, 28, 29 Additional problems include

interobserver and intraobserver variations.

Alternative modalities have some features in common: they have been

developed to reflect disease processes more adequately in the liver

as a whole. Biomarkers exploit serologic correlates of liver disease,

using single assays or combinations of measurements. Cross-sectional

imaging identifies changes associated with liver disease of organ

size, shape, and vascularity (nodular liver outline, flow,

collaterals, spleen size), and more sophisticated applications such

as dedicated magnetic resonance imaging may combine shape, flow, and

collagen/elasticity assessment. Pressure recordings are achieved by

intravascular wedging in the liver and trying to obtain a

representation of hemodynamic changes in larger segments of the liver

draining into the wedged hepatic vein branch. Some tests such as

serum pro-collagen (III)-amino peptide are based on measurement of

components of collagen metabolism.

One of the newest tests that has gained popularity rapidly in Western

Europe and may soon gain approval in the US clinical practice is

liver stiffness measurement. This is performed with a Fibroscan

(EchoSens, Paris, France), a medical device based on properties of

elastometry.20, 21 The device uses a probe that contains an

ultrasonic transducer mounted on the axis of a vibrator. Vibration is

transmitted to induce an elastic shear wave that propagates through

the tissue of interest. Pulse-echo ultrasound acquisitions are

performed synchronously to assess the velocity of propagation of the

shear wave, which is related directly to tissue stiffness (or elastic

modulus). The harder the tissue, the faster the shear wave

propagates. The Fibroscan measures liver stiffness in a cylinder of

space with an approximate diameter of 1 cm and an axial length of 4

cm. This is 100 times the volume of a typical needle biopsy specimen

and therefore is said to be more representative of the entire hepatic

parenchyma than liver biopsy.

ChallengesÂ

The challenge of all tests is at least to reflect relevant pathologic

changes occurring in liver disease accurately, and these may include

hepatocellular and cholangiocellular damage, inflammation, and the

storage of fat, iron, and other substances. Desmet and Roskams29, 30

reminded us of the issues pathologists face when diagnosing fibrosis

and cirrhosis. Relevant pathophysiologic considerations include

structurally abnormal nodules and intrahepatic vascular shunts

between afferent (portal vein and hepatic artery) and efferent

(hepatic vein) vessels of the liver linking portal tracts and central

veins. They discuss details of the dynamic process that leads to

cirrhotic transformation and the potential of its reversibility:

there is more to the diagnosis of cirrhosis than merely fibrosis or a

stage score. The proposed alternatives to liver biopsy tend also to

have a shared characteristic in that they simplify and represent (an

element of) liver histopathology (fibrosis) by a single number, be it

a score, a pressure, or other measurement. Most noninvasive tests

focus on predicting the presence of fibrosis or cirrhosis.

The article in a recent issue of Clinical Gastroenterology and

Hepatology by Patel et al31 may serve as an example of the complexity

and limitations in interpreting data from alternative testing. Their

research report is one of the many dealing with attempts to integrate

noninvasive testing into clinical practice. In a clinico-industrial

collaborative effort they assessed the value of the biomarker

FIBROSpect versus histopathology and morphometric quantification of

fibrosis using a collagen stain (Sirius red) and image analysis. The

biomarker incorporates hyaluronic acid, serum tissue inhibitor of

metalloproteinase-1, and α2-macroglobulin, and in a retrospective

study the test was shown to have a 75% accuracy in differentiating

mild-stage chronic hepatitis C (METAVIR score F0†" F1) from moderate

to severe (METAVIR F2†" 4) disease. Patel et al31 conclude that

multiple factors limit the current use of biomarkers and they discuss

some of the many problematic issues including difficulties with

fibrosis morphometry.

Their findings31 underscore the inherent challenges to pathologic

diagnosis or even the limited assessment of fibrosis or cirrhosis

using a surrogate biochemical or indirect physical measurements.

Although one can appreciate that the information obtained from

alternatives to biopsy might as good as, or more meaningful

clinically, than liver biopsy in some situations in part because of

increased virtual sample volume, all have limitations.

First, no alternative alone can or should be translated literally

into the more complex designation of cirrhosis, and neither should

the term fibrosis be used synonymously with cirrhosis. Liver biopsy

and the alternatives are assessments of a different and complementary

nature. There is an important and fundamental difference between a

clinicopathologic diagnosis of cirrhosis and a score of fibrosis

generated either by histologic scoring systems, stiffness, or

whatever measurement. Second, the box-plots of the results of

alternative testing show substantive overlap between individual

values for each liver disease stage. Third, similar to the liver

biopsy, even a larger volume sample is not immune to the issue of

heterogeneity in the organ distribution of fibrosis. A test such as

liver stiffness measurement at least allows multiple samples to be

obtained easily in one session, but continues to have limitations.

Fourth, discordance in reading, even by experts, is common.

In accepting surrogate markers to assess histopathologic disease and

fibrosis, further considerations arise and downstream-associated

assumptions must be treated cautiously. For example, cirrhosis is

considered to be a precancerous condition, but it would be incorrect

to link uncritically substitute markers with the risk of developing

hepatocellular carcinoma. Put differently, fibrosis and

carcinogenesis are not necessary linked; it is the pathobiology of

underlying chronic liver disease that is linked to carcinogenesis.

Hence, the increase of a surrogate marker of fibrosis by itself does

not necessarily predict an increased cancer risk.

Categoric DesignationsÂ

Disease categories are to some extent designated arbitrarily, based

on assumed relevance to disease management. Boundaries (overlap)

between the categoric assignments may be more or less well defined,

and the distinction between the categories will vary according to the

diagnostic tests used (imaging, biomarkers, stiffness, biopsy, portal

pressure). The importance of the overlap (ie, blurring of the

categoric boundaries) will depend on the clinical relevance of such

categoric distinctions. The article by Patel et al31 re-emphasizes an

observation found commonly when applying surrogate disease markers.

Tests more easily and reliably differentiate between minimal or no

disease (Ishak stage 0†" 2; Metavir stage F0†" F1) and more advanced

disease (Ishak stage 3†" 6; Metavir stage F2†" F4). For group

comparisons, this can be helpful and adequate. In contrast, in making

decisions regarding individual patients, this may not be so: a

patient facing an arduous year of HCV treatment with agents with

potentially major side effects and a 50% to 60% chance for cure, may

in case of limited (F2) disease choose to wait 5 to 10 years and then

receive possibly improved therapy, whereas a patient with clinically

well-compensated, histopathologically established (stage F4)

cirrhosis disease might be better advised to be treated promptly

before decompensation occurs. A recent study highlighted that

differentiating between very extensive fibrosis and minimal disease

was possible only in about one third of patients when combining

tests.32 Whatever sample error of liver biopsy may present, if

significant fibrosis is present, this is usually very helpful

information.

In the case of liver biopsy interpretation, the judgment of the

interpreting pathologist with input from the clinical team33 can be

critical. In an evidence-based approach to treating patients, the

pathologist plays a major management role on the basis of

histopathologic categorization of disease status. Conversely, with

other technical measurements of liver disease, there is little

influence of the interpretive skill of the pathologist. In this way

attention should be paid to the difference between the consultative

role of the pathologist versus the provision of quantitative data

variables from surrogate investigations. The value of any test will

be judged by the ability of that test to guide management of any

given individual patient at any given point in the course of their

treatment. The use of test combinations also must be evaluated

carefully because some of the different tests may merely duplicate

existing data, and may not increase diagnostic accuracy significantly.

Of note, the accuracy of noninvasive testing also may be affected by

other factors such as normal transaminase levels, strongly increased

transaminase levels, or the age of the patient.34, 35, 36

34. 34Kurosaki M, Izumi N. External validation of FIB-4: diagnostic

accuracy is limited in elderly populations. Hepatology.

2008;47:352. CrossRef

35. 35Sebastiani G, Vario A, Guido M, et al.. Performance of

noninvasive markers for liver fibrosis is reduced in chronic

hepatitis C with normal transaminases. J Viral Hepat.

2008;15:212†" 218.

36. 36Coco B, Oliveri F, Maina AM, et al.. Transient elastography: a

new surrogate marker of liver fibrosis influenced by major changes of

transaminases. J Viral Hepat. 2007;14:360†" 369. MEDLINE | CrossRefÂ

ComorbidityÂ

We have discussed as yet mainly alternatives to liver biopsy applied

to presumed single-etiology disease entities. However, liver biopsy

also may identify comorbid and clinically unsuspected conditions that

may be relevant for immediate management and assessment of long-term

outcome. A typical example is the human immunodeficiency virus†" HCV

co-infected patient who is not only at risk for HCV progression, but

also at risk for the hepatotoxic effects of highly active anti-

retroviral therapy and antibiotics to treat infections. Some of these

patients may be treated in relative isolation, that is, not receiving

the benefit from a multidisciplinary team that is focused on early

detection of the unusual and sometimes subtle presentations of

patients with multiple disease etiologies. Such patients may be

rechallenged with the wrong medication if the disease is ascribed to

HCV only (unpublished observations, Suriawinata 2008). Another

example is the patient treated for autoimmune hepatitis with steroids

and whose disease flare a year later under adequate immunosuppression

only was understood after liver biopsy: her considerable weight gain

was accompanied by florid nonalcoholic steatohepatitis, which had

replaced autoimmune hepatitis. This type of pathology will not be

identified easily by any test other than a liver biopsy, initiated by

an experienced hepatologist. Tissue examination can be performed with

increasingly sophisticated immunologic and molecular technology.

Advocates of noninvasive testing easily recognize many remaining

indications for liver biopsy.25

Comorbid pathology (fat, iron, and so forth) may be present in

addition to other histopathologic features and significantly impact

long-term outcome of disease or modify treatment options.37

Histopathologic assessment of disease and the assumed benefits of

therapy will continue to guide current and future management, but

liver biopsy will not be needed necessarily for all patients and

probably will be indicated in increasingly limited groups of patients

as surrogate, noninvasive alternatives establish their respective

positions in the clinical diagnostic armamentarium.

Closing CommentsÂ

Alternatives and surrogates for liver biopsy increasingly will play a

key role in the management of patients with liver disease. They are

often a reasonable alternative for the initial assessment of disease,

and under defined circumstances compare well with liver biopsy for

specific clinical purposes. They may increasingly also play a role in

longitudinal monitoring of patients. Alternatives to liver biopsy

provide information that is different from and complementary to the

information that can be obtained from liver biopsy. In some

circumstances the information that is obtainable by noninvasive

methods may make the liver biopsy unnecessary. Histopathology will

remain important in understanding liver pathology and the impact of

therapy, and the liver biopsy has its own advantages and limitations.

The context-specific relative advantages and disadvantages of all of

the diagnostic approaches for the assessment of patients with liver

disease including liver biopsy and its surrogates must be understood

by those who use these procedures to determine patient care. One

choice of diagnostic method will not suit all circumstances. As

further options become available, to some extent they increase the

burden on the hepatologist to understand the choices properly, and to

be certain that clinical and noninvasive assessment of disease is

indeed appropriate and sufficient for the management of their

individual patient. Many individual physicians and institutions will

face considerable challenges in deciding how to incorporate in a cost-

effective and meaningful way the abundance of information that has

emerged in recent years. A consensus conference and a clinical

guideline would be most timely in this respect