Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Short report
Discrepancy rates in liver biopsy reporting Richard Colling,1 Clare Verrill,1 Eve Fryer,1 Lai Mun Wang,1 Kenneth Fleming2 1
Cellular Pathology, John Radcliffe Hospital, Oxford University Hospitals NHS Trust, Oxford, UK 2 Medical Sciences Division, Oxford University, Oxford, UK Correspondence to Dr Richard Colling, Cellular Pathology, John Radcliffe Hospital, Oxford, OX3 9DU, UK; [email protected] Received 26 February 2014 Revised 29 April 2014 Accepted 2 June 2014 Published Online First 26 June 2014
ABSTRACT Medical liver biopsy reporting is challenging, and maintaining competency with small case numbers is potentially difficult. This study evaluates the discrepancies identified in cases referred to a specialist centre between the specialist reports and those of the referring general departments. Fifty consecutive recently referred cases were selected, and original and final reports were compared. Discrepancies were classified as per the Royal College of Pathologists guidelines and scored for potential clinical impact. The overall rate of discrepancy was 38% with most of these due to differences in interpretation of morphology. Seventy per cent of these discrepancies were judged to have major clinical impact (26% of all referred cases). This study highlights the need for robust systems of quality control of liver biopsies in a general setting.
INTRODUCTION Histological examination of liver core biopsy specimens is a useful investigation in the diagnosis and staging of many complex liver diseases.1 Liver biopsies are currently reported by pathologists in district general hospital settings and in tertiary referral centres.2 Although histopathology training ensures all independent histopathologists are competent in hepatopathology,3 maintaining this particular competency in practice may be challenging in a complex field, particularly with only small case numbers. In view of this, the aim of this study was to examine the rates of diagnostic discrepancy between our reports and those of our six regional referring district general hospitals, and to characterise the discrepancies encountered.
METHODS
To cite: Colling R, Verrill C, Fryer E, et al. J Clin Pathol 2014;67:825–827.
Histology reports stored on the departmental ‘FileMaker Pro’ database were searched for ‘referral’ and ‘liver’ cases in the period June 2012 to June 2013. Inclusion criteria were liver core biopsies that were referred to the department, of any gender and in all age groups. Cases with incomplete referral or departmental histological report data were excluded. Each case was retrospectively reviewed and the microscopic description, staging and final diagnoses were compared between the original referral and authorised final reports. Discrepancies identified were then categorised as per the Royal College of Pathologists (RCPath) guidelines (see table 1).4 Discrepancy types A and E were not the focus of this study, and were therefore not included. Discrepancy type D was obviously not applicable. This left Discrepancies B and C, essentially discrepancies due to either difference in interpretation of morphology (B) or due to a failure to correlate the findings with the clinical
Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
information (C). Cases with more than one discrepancy were categorised based on the most serious difference. Discrepancies were further scored according to the potential clinical impact we felt these could have: (1) no impact, (2) minor potential impact, (3) major potential impact.
RESULTS From 484 liver biopsy cases in the period, there were 52 referred cases, 50 of which met the inclusion criteria. These were authorised cases sent by local general pathologists either for a second opinion or due to transfer of clinical care. All the original slides had been reviewed when we first reported the cases, and no additional work (levels or stains) had been requested. We found that the overall discrepancy rate was 38% (19 cases). A breakdown of the discrepancies found is displayed in table 2. There were five cases (26% of discrepancies) of the B1 category, three cases (16% of discrepancies) of the B2 category and nine cases (47% of discrepancies) of the B3 category. Two cases (11% of discrepancies) were of category C. Thirteen cases (68% of discrepancies) were thought potentially to have major clinical impact; none were felt to have had no clinical significance. The most common problem encountered was fibrosis staging. In our centre we use the Ishak staging system, or the non-alcoholic fatty liver disease scoring (NAFLDS) system for fibrosis, depending on the context. We also give a global comment based on this as mild (Ishak 1/2 or NAFLDS 1), moderate (Ishak 3/4 or NAFLDS 2/3) or severe (Ishak 5/6 or NAFLDS 4), which the clinicians find useful. Our local generalist centres gave a mixture of Ishak fibrosis scores or a similar global
Table 1 Categorisation of discrepancies in histopathology reporting Category
Description
A
Inadequate dissection, sampling or macroscopic description A diagnosis which one is surprised to see from any pathologist (eg, an obvious cancer reported as benign) A diagnosis which is fairly clearly incorrect, but which one is not surprised to see a small percentage of pathologists suggesting (eg, a moderately difficult diagnosis) A diagnosis where interobserver variation is known to be large Discrepancy due to lack of clinical correlation Failure to seek a second opinion in an obviously difficult case Discrepancy in report typography
B1
B2
B3 C D E
825
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Short report Table 2
A breakdown of each discrepancy found and the categorisation of this into discrepancy type and potential clinical impact score
Biopsy
Original report
Referral report
Discrepancy type
Impact score
Three cores, 14 portal tracts, 9 central veins* Two cores, 14 portal tracts, 14 central veins*
Normal, no fibrosis Granulomatous inflammation ?TB/sarcoid, no fibrosis Chronic active inflammation
ASH/NASH, chronic hepatitis with mild fibrosis PBC added to differential, mild fibrosis
B1 B1
3 3
B1
3
B1
3
B1
3
B2
3
Two cores, 16 and 6 mm, 14 portal tracts, 3 central veins Single 7 mm core, 2 portal tracts, 7 central veins*
Normal, no fibrosis
Four cores, 12 portal tracts, 17 central veins*
AIH, no fibrosis
Two cores, 14 and 16 mm, 17 portal tracts, 17 central veins Single disrupted core, 7 portal tracts, 3 central veins Two cores, 20 and 21 mm, 9 portal tracts, 8 central veins* Two cores, 16 portal tracts, 12 central veins* Two cores, 10 portal tracts, 8 central veins* Multiple fragmented cores, 14 portal tracts, 9 central veins* Three cores, 22 portal tracts, 9 central veins One core, 9 portal tracts, 4 central veins* Three cores, 22 portal tracts, 17 central veins* Three cores One core, 13 mm, 6 portal tracts, 5 central veins*
Non-specific minor abnormalities
Steatohepatitis, and bile duct injury with CBP deposition ?biliary disease Ductopaenia, mild fibrosis and CBP deposition? biliary disease Acute hepatitis with cholestasis and mild fibrosis, no evidence of AIH NRH and VO obstruction
Cirrhosis PBC no fibrosis
Cirrhosis with features of PSC PBC moderate fibrosis
B2 B2
3 2
HBV mild fibrosis ?PSC minor fibrosis HBV fibrosis cannot be assessed
HBV moderate fibrosis ?PSC moderate fibrosis HBV mild fibrosis
B3 B3 B3
2 2 2
PBC no AIH Cirrhosis with resolving steatohepatitis ?PSC Chronic inflammation, mild fibrosis ?PSC with mild fibrosis
B3 B3 B3 B3 B3
3 3 2 3 3
One core, 12 mm, 9 portal tracts, 3 central veins One core, 15 portal tracts, 4 central veins
PBC and AIH overlap Cirrhosis AIH Lymphoma, no fibrosis Non-specific minor abnormalities, no fibrosis ALD ?AIH PSC with moderate fibrosis
B3 C
2 3
Two cores, 19 portal tracts
Cirrhosis ?cause
ALD possible biliary component Gallstone injury (high alk phos, gallbladder of same case showed chronic cholecystitis) with mild fibrosis ALD (clinical history of high alcohol intake and raised GTT)
C
3
*Cases reviewed due to transfer of clinical care. AIH, autoimmune hepatitis; ALD, alcoholic liver disease; ASH, alcoholic steatohepatitis; CBP, copper binding protein; HBV, hepatitis B virus; NASH, non-alcoholic steatohepatitis; NRH, nodular regenerative hyperplasia; PBC, primary biliary sclerosis; PSC, primary sclerosing cholangitis; TB, tuberculosis; VO, venous outflow obstruction.
comment (mild/moderate/severe). We found 11 cases (58% of discrepancies) discrepantly staged; most were understaged. Other problems included recognising and interpreting bile duct disorders, seen in seven cases (37% of discrepancies), and 4 cases (21% of discrepancies) had misdiagnoses of autoimmune hepatitis (AIH).
DISCUSSION Liver biopsy remains an important tool in the investigation of liver disease and is often referred to as the gold standard.1 However, interpretation of liver morphology is known to be difficult with marked interobserver variation for key features such as staging of fibrosis.5 Indeed, interobserver variation between specialists and generalists is a well-known phenomenon in general histopathology, as well as between individual ‘experts’.5 6 In view of this, development of expertise undoubtedly requires considerable experience. Furthermore, maintenance of this expertise, like other similar particular skills, presumably requires a continuous exposure to a minimum number of cases each year. As a specialist centre, we see around 500 liver biopsies a year, and our routine practice is for our team of four hepatopathologists to review all liver biopsies together and to authorise a consensus report. We feel this standardises reporting from our centre and minimises discrepancies. The generalist setting will likely see between 10 and 40 cases per year, and 826
these tend to be reported in isolation by one pathologist with a special interest. Previous studies have shown a high rate of discrepancy in the interpretation of liver biopsies between general pathologists and ‘experts’.7 Given the crucial importance of biopsy findings in the management of liver diseases,1 this raises the question whether all liver biopsies should be reported by ‘experts’. In this context, although non-invasive techniques for investigating and diagnosing liver disease are improving (suggesting that liver biopsy will become less important), specimen numbers are not falling and cases are becoming increasingly complex.1 8 Accordingly, the need for accurate diagnosis is becoming all the more important. To gain insight into the local situation, we decided to measure the discrepancy rate of liver biopsies of cases referred into the department for a second opinion, between our diagnosis and the referring diagnosis. We found a discrepancy rate of 38% which, although not out of line with results of similar investigations—rates up to 65% have been reported7 and, indeed, in general histopathology reporting, rates of up to 43% have been reported, although mean frequencies are in the 5–7% range9—we feel is higher than what patients would deem acceptable. The most common cause of discrepancy—58% (11/19)— related to difficulties with fibrosis staging. Bile duct abnormalities were commonly missed—37% (7/19)—unsurprising, as this is a known area of difficulty.7 Perhaps more unexpected is the Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Short report difficulty around diagnosis of AIH (21%—4/19). This may relate to the relative infrequency of the disease. Some of these were cases which were sent for a second opinion because they were challenging for a generalist. As expected then, a significant number of the discrepancies we identified were B3 which, by definition, are those in which interobserver variation is known to be high and opinions differ. However, 10 (53%) of the discrepant cases were sent for review because of a transfer of clinical care. Furthermore, a significant number (42%, 8 cases) of the discrepancies we felt were of the B1/B2 category, which should not have been authorised by any qualified general histopathologist. In such cases, failure to follow-up and undertake further investigation would have had major clinical and therapeutic implications, highlighting the potential major clinical impact, such discrepancies—the discrepancy in 26% of the referred cases was judged to have potentially had major clinical implications. As mentioned above, general pathologist reporting of liver specimens has a high rate of discrepancy, and some have called for specialist reporting of all biopsies.7 10 This is contentious and has major practical implications, but this study does highlight the issue and the problem of maintaining skills in this area. The recently published RCPath tissue pathways for medical liver biopsy now recommends a number of new measures for
non-specialist centres, including a named lead for liver pathology (formally in a pathologist's job plan), external quality assessment (EQA) participation, a minimum annual exposure (20 per pathologist is suggested) and passing lower case numbers to specialist centres.11 We also suggest that the service should be subject to external annual audit of a sample of cases (∼10%), and clear referral criteria or guidelines should be in place in all centres, and the threshold for referral should be low. Contributors KF conceived the project. RC collected the data and drafted the manuscript. All participants were involved in evaluating the data and editing the final draft. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed. Data sharing statement All data collected is published in full.
REFERENCES 1
2 3 4 5
Take home messages
6
▸ Liver core biopsy evaluation is a difficult skill to maintain. ▸ Recognition of bile duct abnormalities and staging fibrosis are common difficulties. ▸ Discrepancy rates between generalist and specialist reporting are high. ▸ Many of the discrepancies are those which should not be authorised by any qualified histopathologist. ▸ New RCPath Guidelines for non-specialist centres recommend a named liver pathologist who participates in EQA and reports a minimum of 20 cases per year.
7
Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
8
9 10
11
D’Incao RB, Silva MC, Almeida PR, et al. Percutaneous liver biopsy–2 decades of experience in a public hospital in the South of Brazil. Ann Hepatol 2013;12: 876–80. Bateman AC. Patterns of histological change in liver disease: my approach to ‘medical’ liver biopsy reporting. Histopathology 2007;51:585–96. The Royal College of Pathologists. Curriculum for specialty training in histopathology. The Royal College of Pathologists, 2010. The Royal College of Pathologists. Review of the categorisation of discrepancies in histopathology. The Royal College of Pathologists, 2008. Verkooijen HM, Peterse JL, Schipper MEI, et al. Interobserver variability between general and expert pathologists during the histopathological assessment of large-core needle and open biopsies of non-palpable breast lesions. Eur J Cancer 2003;39:2187–91. Latour M, Amin MB, Billis A, et al. Grading of invasive cribriform carcinoma on prostate needle biopsy: an interobserver study among experts in genitourinary pathology. Am J Surg Pathol 2008;32:1532–9. Bejarano PA, Koehler A, Sherman KE. Second opinion pathology in liver biopsy interpretation. Am J Gastroenterol 2001;96:3158–64. Myers RP, Fong A, Shaheen AA. Utilization rates, complications and costs of percutaneous liver biopsy: a population-based study including 4275 biopsies. Liver Int 2008;28:705–12. Raab SS, Nakhleh RE, Ruby SG. Patient safety in anatomic pathology: measuring discrepancy frequencies and causes. Arch Pathol Lab Med 2005;129:459–66. Krishnan B, Stares M, Rajabally H, et al. PTU-094 should liver biopsies be reported by pathologists with a subspecialist interest in hepatology? Gut 2013;62(Suppl 1): A83–4. The Royal College of Pathologists. Tissue pathways for liver biopsies for the investigation of medical disease and for focal lesions. The Royal College of Pathologists, 2014.
827
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Discrepancy rates in liver biopsy reporting Richard Colling, Clare Verrill, Eve Fryer, Lai Mun Wang and Kenneth Fleming J Clin Pathol 2014 67: 825-827 originally published online June 26, 2014
doi: 10.1136/jclinpath-2014-202261 Updated information and services can be found at: http://jcp.bmj.com/content/67/9/825
These include:
References Email alerting service
Topic Collections
This article cites 8 articles, 1 of which you can access for free at: http://jcp.bmj.com/content/67/9/825#BIBL Receive free email alerts when new articles cite this article. Sign up in the box at the top right corner of the online article.
Articles on similar topics can be found in the following collections JCP Education (131) Clinical diagnostic tests (756)
Notes
To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions To order reprints go to: http://journals.bmj.com/cgi/reprintform To subscribe to BMJ go to: http://group.bmj.com/subscribe/
Short report
Discrepancy rates in liver biopsy reporting Richard Colling,1 Clare Verrill,1 Eve Fryer,1 Lai Mun Wang,1 Kenneth Fleming2 1
Cellular Pathology, John Radcliffe Hospital, Oxford University Hospitals NHS Trust, Oxford, UK 2 Medical Sciences Division, Oxford University, Oxford, UK Correspondence to Dr Richard Colling, Cellular Pathology, John Radcliffe Hospital, Oxford, OX3 9DU, UK; [email protected] Received 26 February 2014 Revised 29 April 2014 Accepted 2 June 2014 Published Online First 26 June 2014
ABSTRACT Medical liver biopsy reporting is challenging, and maintaining competency with small case numbers is potentially difficult. This study evaluates the discrepancies identified in cases referred to a specialist centre between the specialist reports and those of the referring general departments. Fifty consecutive recently referred cases were selected, and original and final reports were compared. Discrepancies were classified as per the Royal College of Pathologists guidelines and scored for potential clinical impact. The overall rate of discrepancy was 38% with most of these due to differences in interpretation of morphology. Seventy per cent of these discrepancies were judged to have major clinical impact (26% of all referred cases). This study highlights the need for robust systems of quality control of liver biopsies in a general setting.
INTRODUCTION Histological examination of liver core biopsy specimens is a useful investigation in the diagnosis and staging of many complex liver diseases.1 Liver biopsies are currently reported by pathologists in district general hospital settings and in tertiary referral centres.2 Although histopathology training ensures all independent histopathologists are competent in hepatopathology,3 maintaining this particular competency in practice may be challenging in a complex field, particularly with only small case numbers. In view of this, the aim of this study was to examine the rates of diagnostic discrepancy between our reports and those of our six regional referring district general hospitals, and to characterise the discrepancies encountered.
METHODS
To cite: Colling R, Verrill C, Fryer E, et al. J Clin Pathol 2014;67:825–827.
Histology reports stored on the departmental ‘FileMaker Pro’ database were searched for ‘referral’ and ‘liver’ cases in the period June 2012 to June 2013. Inclusion criteria were liver core biopsies that were referred to the department, of any gender and in all age groups. Cases with incomplete referral or departmental histological report data were excluded. Each case was retrospectively reviewed and the microscopic description, staging and final diagnoses were compared between the original referral and authorised final reports. Discrepancies identified were then categorised as per the Royal College of Pathologists (RCPath) guidelines (see table 1).4 Discrepancy types A and E were not the focus of this study, and were therefore not included. Discrepancy type D was obviously not applicable. This left Discrepancies B and C, essentially discrepancies due to either difference in interpretation of morphology (B) or due to a failure to correlate the findings with the clinical
Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
information (C). Cases with more than one discrepancy were categorised based on the most serious difference. Discrepancies were further scored according to the potential clinical impact we felt these could have: (1) no impact, (2) minor potential impact, (3) major potential impact.
RESULTS From 484 liver biopsy cases in the period, there were 52 referred cases, 50 of which met the inclusion criteria. These were authorised cases sent by local general pathologists either for a second opinion or due to transfer of clinical care. All the original slides had been reviewed when we first reported the cases, and no additional work (levels or stains) had been requested. We found that the overall discrepancy rate was 38% (19 cases). A breakdown of the discrepancies found is displayed in table 2. There were five cases (26% of discrepancies) of the B1 category, three cases (16% of discrepancies) of the B2 category and nine cases (47% of discrepancies) of the B3 category. Two cases (11% of discrepancies) were of category C. Thirteen cases (68% of discrepancies) were thought potentially to have major clinical impact; none were felt to have had no clinical significance. The most common problem encountered was fibrosis staging. In our centre we use the Ishak staging system, or the non-alcoholic fatty liver disease scoring (NAFLDS) system for fibrosis, depending on the context. We also give a global comment based on this as mild (Ishak 1/2 or NAFLDS 1), moderate (Ishak 3/4 or NAFLDS 2/3) or severe (Ishak 5/6 or NAFLDS 4), which the clinicians find useful. Our local generalist centres gave a mixture of Ishak fibrosis scores or a similar global
Table 1 Categorisation of discrepancies in histopathology reporting Category
Description
A
Inadequate dissection, sampling or macroscopic description A diagnosis which one is surprised to see from any pathologist (eg, an obvious cancer reported as benign) A diagnosis which is fairly clearly incorrect, but which one is not surprised to see a small percentage of pathologists suggesting (eg, a moderately difficult diagnosis) A diagnosis where interobserver variation is known to be large Discrepancy due to lack of clinical correlation Failure to seek a second opinion in an obviously difficult case Discrepancy in report typography
B1
B2
B3 C D E
825
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Short report Table 2
A breakdown of each discrepancy found and the categorisation of this into discrepancy type and potential clinical impact score
Biopsy
Original report
Referral report
Discrepancy type
Impact score
Three cores, 14 portal tracts, 9 central veins* Two cores, 14 portal tracts, 14 central veins*
Normal, no fibrosis Granulomatous inflammation ?TB/sarcoid, no fibrosis Chronic active inflammation
ASH/NASH, chronic hepatitis with mild fibrosis PBC added to differential, mild fibrosis
B1 B1
3 3
B1
3
B1
3
B1
3
B2
3
Two cores, 16 and 6 mm, 14 portal tracts, 3 central veins Single 7 mm core, 2 portal tracts, 7 central veins*
Normal, no fibrosis
Four cores, 12 portal tracts, 17 central veins*
AIH, no fibrosis
Two cores, 14 and 16 mm, 17 portal tracts, 17 central veins Single disrupted core, 7 portal tracts, 3 central veins Two cores, 20 and 21 mm, 9 portal tracts, 8 central veins* Two cores, 16 portal tracts, 12 central veins* Two cores, 10 portal tracts, 8 central veins* Multiple fragmented cores, 14 portal tracts, 9 central veins* Three cores, 22 portal tracts, 9 central veins One core, 9 portal tracts, 4 central veins* Three cores, 22 portal tracts, 17 central veins* Three cores One core, 13 mm, 6 portal tracts, 5 central veins*
Non-specific minor abnormalities
Steatohepatitis, and bile duct injury with CBP deposition ?biliary disease Ductopaenia, mild fibrosis and CBP deposition? biliary disease Acute hepatitis with cholestasis and mild fibrosis, no evidence of AIH NRH and VO obstruction
Cirrhosis PBC no fibrosis
Cirrhosis with features of PSC PBC moderate fibrosis
B2 B2
3 2
HBV mild fibrosis ?PSC minor fibrosis HBV fibrosis cannot be assessed
HBV moderate fibrosis ?PSC moderate fibrosis HBV mild fibrosis
B3 B3 B3
2 2 2
PBC no AIH Cirrhosis with resolving steatohepatitis ?PSC Chronic inflammation, mild fibrosis ?PSC with mild fibrosis
B3 B3 B3 B3 B3
3 3 2 3 3
One core, 12 mm, 9 portal tracts, 3 central veins One core, 15 portal tracts, 4 central veins
PBC and AIH overlap Cirrhosis AIH Lymphoma, no fibrosis Non-specific minor abnormalities, no fibrosis ALD ?AIH PSC with moderate fibrosis
B3 C
2 3
Two cores, 19 portal tracts
Cirrhosis ?cause
ALD possible biliary component Gallstone injury (high alk phos, gallbladder of same case showed chronic cholecystitis) with mild fibrosis ALD (clinical history of high alcohol intake and raised GTT)
C
3
*Cases reviewed due to transfer of clinical care. AIH, autoimmune hepatitis; ALD, alcoholic liver disease; ASH, alcoholic steatohepatitis; CBP, copper binding protein; HBV, hepatitis B virus; NASH, non-alcoholic steatohepatitis; NRH, nodular regenerative hyperplasia; PBC, primary biliary sclerosis; PSC, primary sclerosing cholangitis; TB, tuberculosis; VO, venous outflow obstruction.
comment (mild/moderate/severe). We found 11 cases (58% of discrepancies) discrepantly staged; most were understaged. Other problems included recognising and interpreting bile duct disorders, seen in seven cases (37% of discrepancies), and 4 cases (21% of discrepancies) had misdiagnoses of autoimmune hepatitis (AIH).
DISCUSSION Liver biopsy remains an important tool in the investigation of liver disease and is often referred to as the gold standard.1 However, interpretation of liver morphology is known to be difficult with marked interobserver variation for key features such as staging of fibrosis.5 Indeed, interobserver variation between specialists and generalists is a well-known phenomenon in general histopathology, as well as between individual ‘experts’.5 6 In view of this, development of expertise undoubtedly requires considerable experience. Furthermore, maintenance of this expertise, like other similar particular skills, presumably requires a continuous exposure to a minimum number of cases each year. As a specialist centre, we see around 500 liver biopsies a year, and our routine practice is for our team of four hepatopathologists to review all liver biopsies together and to authorise a consensus report. We feel this standardises reporting from our centre and minimises discrepancies. The generalist setting will likely see between 10 and 40 cases per year, and 826
these tend to be reported in isolation by one pathologist with a special interest. Previous studies have shown a high rate of discrepancy in the interpretation of liver biopsies between general pathologists and ‘experts’.7 Given the crucial importance of biopsy findings in the management of liver diseases,1 this raises the question whether all liver biopsies should be reported by ‘experts’. In this context, although non-invasive techniques for investigating and diagnosing liver disease are improving (suggesting that liver biopsy will become less important), specimen numbers are not falling and cases are becoming increasingly complex.1 8 Accordingly, the need for accurate diagnosis is becoming all the more important. To gain insight into the local situation, we decided to measure the discrepancy rate of liver biopsies of cases referred into the department for a second opinion, between our diagnosis and the referring diagnosis. We found a discrepancy rate of 38% which, although not out of line with results of similar investigations—rates up to 65% have been reported7 and, indeed, in general histopathology reporting, rates of up to 43% have been reported, although mean frequencies are in the 5–7% range9—we feel is higher than what patients would deem acceptable. The most common cause of discrepancy—58% (11/19)— related to difficulties with fibrosis staging. Bile duct abnormalities were commonly missed—37% (7/19)—unsurprising, as this is a known area of difficulty.7 Perhaps more unexpected is the Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Short report difficulty around diagnosis of AIH (21%—4/19). This may relate to the relative infrequency of the disease. Some of these were cases which were sent for a second opinion because they were challenging for a generalist. As expected then, a significant number of the discrepancies we identified were B3 which, by definition, are those in which interobserver variation is known to be high and opinions differ. However, 10 (53%) of the discrepant cases were sent for review because of a transfer of clinical care. Furthermore, a significant number (42%, 8 cases) of the discrepancies we felt were of the B1/B2 category, which should not have been authorised by any qualified general histopathologist. In such cases, failure to follow-up and undertake further investigation would have had major clinical and therapeutic implications, highlighting the potential major clinical impact, such discrepancies—the discrepancy in 26% of the referred cases was judged to have potentially had major clinical implications. As mentioned above, general pathologist reporting of liver specimens has a high rate of discrepancy, and some have called for specialist reporting of all biopsies.7 10 This is contentious and has major practical implications, but this study does highlight the issue and the problem of maintaining skills in this area. The recently published RCPath tissue pathways for medical liver biopsy now recommends a number of new measures for
non-specialist centres, including a named lead for liver pathology (formally in a pathologist's job plan), external quality assessment (EQA) participation, a minimum annual exposure (20 per pathologist is suggested) and passing lower case numbers to specialist centres.11 We also suggest that the service should be subject to external annual audit of a sample of cases (∼10%), and clear referral criteria or guidelines should be in place in all centres, and the threshold for referral should be low. Contributors KF conceived the project. RC collected the data and drafted the manuscript. All participants were involved in evaluating the data and editing the final draft. Competing interests None. Provenance and peer review Not commissioned; externally peer reviewed. Data sharing statement All data collected is published in full.
REFERENCES 1
2 3 4 5
Take home messages
6
▸ Liver core biopsy evaluation is a difficult skill to maintain. ▸ Recognition of bile duct abnormalities and staging fibrosis are common difficulties. ▸ Discrepancy rates between generalist and specialist reporting are high. ▸ Many of the discrepancies are those which should not be authorised by any qualified histopathologist. ▸ New RCPath Guidelines for non-specialist centres recommend a named liver pathologist who participates in EQA and reports a minimum of 20 cases per year.
7
Colling R, et al. J Clin Pathol 2014;67:825–827. doi:10.1136/jclinpath-2014-202261
8
9 10
11
D’Incao RB, Silva MC, Almeida PR, et al. Percutaneous liver biopsy–2 decades of experience in a public hospital in the South of Brazil. Ann Hepatol 2013;12: 876–80. Bateman AC. Patterns of histological change in liver disease: my approach to ‘medical’ liver biopsy reporting. Histopathology 2007;51:585–96. The Royal College of Pathologists. Curriculum for specialty training in histopathology. The Royal College of Pathologists, 2010. The Royal College of Pathologists. Review of the categorisation of discrepancies in histopathology. The Royal College of Pathologists, 2008. Verkooijen HM, Peterse JL, Schipper MEI, et al. Interobserver variability between general and expert pathologists during the histopathological assessment of large-core needle and open biopsies of non-palpable breast lesions. Eur J Cancer 2003;39:2187–91. Latour M, Amin MB, Billis A, et al. Grading of invasive cribriform carcinoma on prostate needle biopsy: an interobserver study among experts in genitourinary pathology. Am J Surg Pathol 2008;32:1532–9. Bejarano PA, Koehler A, Sherman KE. Second opinion pathology in liver biopsy interpretation. Am J Gastroenterol 2001;96:3158–64. Myers RP, Fong A, Shaheen AA. Utilization rates, complications and costs of percutaneous liver biopsy: a population-based study including 4275 biopsies. Liver Int 2008;28:705–12. Raab SS, Nakhleh RE, Ruby SG. Patient safety in anatomic pathology: measuring discrepancy frequencies and causes. Arch Pathol Lab Med 2005;129:459–66. Krishnan B, Stares M, Rajabally H, et al. PTU-094 should liver biopsies be reported by pathologists with a subspecialist interest in hepatology? Gut 2013;62(Suppl 1): A83–4. The Royal College of Pathologists. Tissue pathways for liver biopsies for the investigation of medical disease and for focal lesions. The Royal College of Pathologists, 2014.
827
Downloaded from http://jcp.bmj.com/ on March 13, 2015 - Published by group.bmj.com
Discrepancy rates in liver biopsy reporting Richard Colling, Clare Verrill, Eve Fryer, Lai Mun Wang and Kenneth Fleming J Clin Pathol 2014 67: 825-827 originally published online June 26, 2014
doi: 10.1136/jclinpath-2014-202261 Updated information and services can be found at: http://jcp.bmj.com/content/67/9/825
These include:
References Email alerting service
Topic Collections
This article cites 8 articles, 1 of which you can access for free at: http://jcp.bmj.com/content/67/9/825#BIBL Receive free email alerts when new articles cite this article. Sign up in the box at the top right corner of the online article.
Articles on similar topics can be found in the following collections JCP Education (131) Clinical diagnostic tests (756)
Notes
To request permissions go to: http://group.bmj.com/group/rights-licensing/permissions To order reprints go to: http://journals.bmj.com/cgi/reprintform To subscribe to BMJ go to: http://group.bmj.com/subscribe/
