Fibrosis in the liver role of histology – is it the gold standard? BSG Annual meeting post- graduate course 20 March 20

1. Fibrosis in the liverrole of histology – is it the gold standard?BSG Annual meeting post- graduate course20 March 2006 Amar Paul Dhillon Royal Free & University College Medical School

2. Fibrosis in the liverrole of histology– is it the gold standard?Outline • Introduction. • What is “the gold standard”. • Histopathological assessment of liver biopsy fibrosis. • Liver biopsy adequacy. • An adequate sample is essential for reliable assessment. • Histopathological description of liver disease stage. • What do we mean by “liver fibrosis”? • Histopathological stage “scoring”. • Stage scoring is a description, not a measurement. • Histopathological measurement of liver fibrosis. • Image analysis. • Validation of surrogate markers of liver fibrosis. • Conclusion. • Histology is the gold standard for liver fibrosis but it must be improved. • The limitations of the current histological methodology have to be understood much better.

3. “Gold standard” 1970Encycl. Brit. X. 547/2 Gold standard, a monetary system in which the standard unit is a fixed weight of gold or is kept at the value of a fixed weight of gold. Claassen J. BMJ 330, 1121; 2005. The gold standard: not a golden standard. • “Studies that evaluate a new diagnostic test, procedure, or method should do so by comparing it with a time honoured alternative that is considered to be the current standard in the field. In this context the meaning of the word standard is “authoritative or recognised exemplar of quality or correctness”. • a gold standard in its true meaning, derived from the monetary gold standard, merely denotes the best tool available at that time to compare different measures. • Even in its glory days, the monetary gold standard was never considered perfect. It was subject to endless debate, and in the end it was abandoned for a better system. • Similarly, today’s gold standard tests will be replaced by better ones. As was eloquently stated by Versi: “It is the absolute truth that is never reached; gold standards are constantly challenged and superseded when appropriate.”

4. Liver fibrosis: histology is the gold standard • The answer to the given question is quite straightforward because examination of liver tissue is the only direct way of assessing liver collagen. • If we wish to assess changes of liver collagen related to disease or treatment of patients with chronic liver disease, histopathology is the only realistic option because other routine diagnostic requirements can still be satisfied. • The alternative of biochemical assay destroys the tissue and is confounded by the large amount of normal (structural) collagen that is present in large portal tracts, blood vessels, and liver capsular tissue (which the histopathologist can ignore).

5. BSG guidelines on the use of liver biopsy in clinical practice Oct 2004 • Indications for liver biopsy: • 5.2 HCV viral infection • “biopsy at present, as with other indications, remains the only reliable method of assessing the degree of fibrosis”. • 5.11 Non-alcoholic liver disease • “Liver tests do not reliably confirm or refute the diagnosis or stage the extent of liver fibrosis”.

6. Liver fibrosis: is histology the gold standard? The answer is not difficult, but the question is tricky. • The unstated relevant aspects of the question are: • Is histological assessment of connective tissue and architectural changes in human liver biopsies the best way of assessing chronic hepatitis disease related changes and response to antifibrotic treatment? • The importance of the question, and most of the work on this subject is driven by the need to have a practical way of managing and monitoring liver disease stage in patients with chronic viral hepatitis. • This was a redundant question when “progressive” (advanced) stages of liver disease were considered to be irreversible: now it is not. • The answer is still “yes”, but the histological assessment of liver biopsies must be improved and done more critically. • Until the histological assessment of liver biopsy fibrosis is done better, we will not know if surrogate “non-invasive liver fibrosis” markers are valid (“sufficient”) or not.

7. Liver biopsy adequacy • “Most hepatopathologists are satisfied with a biopsy specimen containing at least six to eight portal tracts”. • Bravo AA et al NEJM 344, 495; 2001. • We should not be primarily concerned with the size of biopsy that “satisfies” the pathologist. • The correct question is: • What size of biopsy will provide a reliable assessment for the patient’s management?

8. Liver biopsy adequacy • Guido M and Rugge M. • Semin Liv Dis 24, 89; 2004. • In most diffuse liver diseases examination of 12-15 complete portal tracts is necessary. • ~20mm of a 1.4mm diameter (17 gauge) needle biopsy. • Progressively longer samples of thinner biopsies are needed.

9. Stage of disease affects adequacy of liver biopsy samples for fibrosis measurement Bedossa et al. Hepatology 38,1449;2003 Fig. 5. Percentage of correctly classified biopsies with the converted METAVIR score of fibrosis according to length of biopsy specimen. (A) Value for different stages

10. Liver biopsy adequacy • Sample adequacy depends on: • Disease aetiology. • Disease distribution. • Stage of disease. • Diameter of needle biopsy. • In most diffuse liver diseases examination of 12-15 complete portal tracts is necessary. • Inadequate samples will give misleading results: • Inadequate small samples tend to underestimate the degree of disease grade and stage.

11. Inadequate samples give variable resultsSiddique et al. Scand J Gastro 38,427;2003 • 29 pairs of HCV liver biopsies with 4-5 portal tracts in each. • 2 separate R lobe samples, via the same puncture site. • Knodell scoring: ≥2 point difference in: • 20 cases (69%) total necroinflammatory score. • 6 cases (21%) fibrosis score.

12. Adequate samples give reproducible resultsPersico et al. Am J Gastro 97,491;2002 52 consecutive patients with HCV, paired liver biopsies. Separate, synchronous (18G needle) samples of R and L lobes of liver. Sample mean length R lobe = 2.8+/-1.1cm; L lobe = 2.5+/-0.9 cm. (samples <15mm in 2 patients who were excluded from the analysis) Ishak scoring system: no significant difference in stage or grade between R and L lobe samples.

13. Histopathological description of liver disease stage

14. What do we mean by “liver fibrosis”? Differences between morphological appearance, description, stage scoring and liver fibrosis measurement From: Standish R et al. An appraisal of the histopathological assessment of liver fibrosis. Gut in press.

15. What do we mean by “liver fibrosis”? Histological fibrosis/stage scores of liver are a mixture of descriptions of fibrotic and architectural histopathological changesKnodell RG et al. Hepatology 1,431;1981Ishak KG et al. J Hepatol 22,696;1995

16. Histological scoring systems for liver biopsy assessment of fibrosis • Up to now histological staging systems have been misunderstood by many, and there has been improper use of the histological fibrosis/stage “scores”. • Histological stage scores describe features that depend not only on the degree of fibrosis but also on architectural changes. • The “number” assigned in the fibrosis score is not a measurement. • The “number” is a shorthand label for a morphological description ie it is a categorical assignment. • Appropriate statistical methods for ordered categorical assignments must be used to analyse the results of trials that use fibrosis “scores”. • eg Q-Q analysis • Lagging LM et al. Liver 22, 136; 2002.

17. Histopathological measurement of liver fibrosis (collagen) • Stage scoring is different from collagen measurement. • To assess fibrosis (collagen) properly in a liver biopsy, it is necessary to stain for it specifically and measure that staining accurately. • Currently, reticulin staining is usually used (in the UK) to assess fibrosis and architecture. • Specific collagen staining is not used routinely. • Histopathological measurement of liver collagen requires: • Sirius red staining. • Image analysis.

18. Measurement of liver biopsy fibrosis • If liver collagen measurement is required, some sort of quantitative approach is unavoidable. • Sirius red stains most hepatic collagens, (including types I & III, which are the main types involved in liver fibrosis). • Sirius red staining correlates with chemical hydroxyproline assays for collagen under standardised conditions. • Specific staining of biopsies for collagen with interactive image analysis provides specific, precise (sensitive) numerical information about liver fibrosis.

19. Development of intelligent interactive image analysis protocols is required for the reliable measurement of (chronic hepatitis) disease related liver fibrosis Hoofring A et al. J Hepatol 39, 738–741; 2003. Three-dimensional reconstruction of hepatic bridging fibrosis in chronic hepatitis C viral infection. The affected portal tracts are up to 0.4 mm diameter.

20. The graph illustrates: • the relative stability of the cumulative average collagen proportionate area (CPA) • with a smaller measured area (compared to the unedited CPA). • The dominant proportion of collagen that resides in normal structures • in normal liver, and the relatively small amount of fibrous tissue that is • normally present in small (<0.5mm diameter) portal tracts. From: Standish R et al. An appraisal of the histopathological assessment of liver fibrosis.Gut in press.

21. Surrogate markers for the assessment of hepatic fibrosis in chronic liver disease. • Surrogate markers and non-invasive fibrosis tests have not been validated properly yet. • The “gold standard” liver histopathology assessments with which they have been compared to date have been inadequate.

22. A systematic review of the quality of liver biopsy specimens Cholangitas E et al. Am J Clin Pathol. In press • There were 32 percutaneous liver biopsy studies in which biopsy length and complete portal tract number (CP) was evaluated: • mean length 17.7±5.8 mm and CP number 7.5±3.4. • There were 15 transjugular liver biopsy studies: • mean length 13.5±4.5 mm and CP number 6.8±2.3. • Only 11 (5%) of 207 therapeutic studies for chronic hepatitis B and C documented length and/or CP number. • Of 12 studies evaluating non-invasive fibrosis tests, only 8 documented length or CP, and only 1 both length and portal tract number. • The performance of these tests has been evaluated using sub-optimal liver biopsies (<20-25 mm length and/or containing <11 CP) or the quality of the biopsy was not mentioned.

23. Fibrosis in the liverrole of histology– is it the gold standard?Conclusion • Histology is the gold standard for liver fibrosis but it must be improved. • Clinical studies of chronic liver disease have not measured biopsy tissue fibrosis critically and precisely in most studies. • Histopathological disease stage is not the same as liver collagen content. • Progress in this area can only be made when: • A better, more rational, approach is used to assess fibrosis in liver biopsies. • To assess fibrosis properly in a liver biopsy, it is necessary to: • examine a sample that is adequately representative of the disease in question. • stain for fibrosis specifically. • measure the staining accurately.