Percutaneous needle biopsy of the liver is an important

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2005;3: Importance of Specimen Size in Accurate Needle Liver Biopsy Evaluation of Patients With Chronic Hepatitis C THOMAS D. SCHIANO,* SAMIA AZEEM,* CAROL A. BODIAN, HENRY C. BODENHEIMER JR, SUKMA MERATI, SWAN N. THUNG, and PRODROMOS HYTIROGLOU *Division of Liver Diseases, Department of Biostatistics, and the Lillian and Henry M. Stratton-Hans Popper Department of Pathology, The Mount Sinai Medical Center, New York; and the Division of Gastroenterology, Beth Israel Medical Center, New York, New York Background & Aims: In patients with chronic hepatitis C (CHC), percutaneous needle liver biopsy examination establishes the severity of necroinflammatory activity and fibrosis, thus guiding treatment decisions. Optimal biopsy specimen size remains controversial. We sought to determine how varying lengths of biopsy specimens influence the grading and staging of CHC. Methods: We used 100 liver biopsy specimens from patients with CHC. The slides were evaluated blindly using the METAVIR scoring system, after being covered with paper, so that only specific specimen lengths (5 mm, 10 mm, 15 mm, and >20 mm) were visible. In each case, the scores obtained with biopsies 5 mm, 10 mm, or 15 mm long were compared with the scores at 20 mm or greater by weighted statistics ( of >.75 signified excellent agreement). A subset of specimens 20 mm or greater was selected for a blinded repeat scoring to assess intraobserver agreement. The statistics for the designated features and lengths were compared using analysis of variance. Results: In assessing the stage of fibrosis, the weighted statistics for agreement with the 20-mm or greater score at 5 mm, 10 mm, and 15 mm were.75,.85, and.92, respectively. In assessing the histologic activity score, the corresponding figures were.73,.81, and.77, respectively. Average statistic comparisons showed that intraobserver agreement was significantly better than agreement between the 20-mm or greater scores and those at shorter lengths; the 5-mm scores were significantly lower than the others; and there was no significant difference between the 10-mm and 15-mm scores. Conclusions: Liver biopsy specimens measuring at least 10 mm usually reflect the grade and stage of CHC reliably. Relatively little improvement in diagnostic accuracy is obtained with longer specimens. Percutaneous needle biopsy of the liver is an important diagnostic tool used routinely in patients with chronic hepatitis C (CHC). Liver biopsy examination establishes the severity of necroinflammatory activity (grade) and the degree of fibrosis (stage), providing useful data to estimate the rate of disease progression. 1 3 Liver fibrosis is a key histologic feature in CHC, useful for evaluation of the severity of disease, for treatment decisions, and for assessing drug efficacy. 4 Liver biopsy examination guides almost all treatment decisions in CHC patients, and also excludes the presence of comorbid disease processes such as steatohepatitis. Although the needle liver biopsy procedure obtains a relatively small tissue specimen for analysis, it traditionally has been felt to be representative of any diffuse pathologic process affecting the liver. 5 7 It is accepted, however, that sampling error does occur. Liver pathologists believe that a tissue cylinder of at least 15 mm generally is adequate 7 ; however, a recent study has recommended a minimum size of 20 mm. 8 It also has been reported that there is a tendency for the predictive value of a positive and negative test to improve with increasing specimen size. 9 This becomes crucial in patients with CHC because the amount of fibrosis noted on liver biopsy examination is predictive of the response to interferon treatment, and also is prognostic of the progression to cirrhosis. 10,11 Noninvasive biochemical markers for hepatic fibrosis are under development, but to date have shown only modest success in predicting the degree of fibrosis. 9,12,13 With the anticipated addition of antifibrotic therapies to the CHC treatment armamentarium, very accurate measurements of hepatic fibrosis become even more crucial because liver biopsy examination is currently the only modality to assess the response over time of liver fibrosis to therapy. Ascertaining what an optimal liver biopsy specimen length is in patients with CHC will have an impact on the management of patients with coagulopathy in whom multiple passes of the biopsy needle when necessary could be associated with increased morbidity, and in patients entering antifibrotic drug Abbreviation used in this paper: CHC, chronic hepatitis C by the American Gastroenterological Association /05/$30.00 PII: /S (05)

2 September 2005 BIOPSY SPECIMEN SIZE IN CHRONIC HEPATITIS C 931 treatment trials in whom larger amounts of liver tissue presumably are needed to measure the degree of fibrosis reliably. 14,15 We thus sought to determine whether varying sizes of liver biopsy specimens influence the grading and staging of CHC. Patients and Methods By using a computerized database in the Division of Liver Diseases, we identified all patients with CHC undergoing a percutaneous needle liver biopsy procedure at the Mount Sinai Medical Center from September 1996 to November For all of these biopsy specimens, a reusable 17-gauge metal Menghini needle with a length of 70 mm was used. All transjugular and intraoperative liver biopsy specimens were excluded. Liver biopsy specimens of patients with concurrent hepatitis B, human immunodeficiency virus co-infection, alcoholic liver disease, and liver transplantation were excluded after review of clinical data. Only those patients with liver biopsy specimens that were at least 20 mm in size (cumulative size if the tissue was fragmented) were studied. The study was approved by the Mount Sinai institutional review board. A total of 464 patients and liver biopsy specimens met the strict criteria and were entered into a master list. One hundred biopsy specimens then were selected at random in batches of For each case, 2 slides, stained with H&E and with Masson s trichrome, were examined microscopically by an experienced blinded liver pathologist (P.H.) for 4 separate sessions. The sessions were spaced out over several months, and included examination of the specimens at lengths of 5 mm, 10 mm, and 15 mm and with the entire material available. Each slide was covered by opaque paper so that only the length assigned to that session was visible; lengths were measured using a scaled ruler by a separate blinded investigator (S.A.). A random number generator was used separately for each batch to determine the order in which the cases would be presented for each of the 4 reading sessions. A random permutation of the integers 1 4 then was generated for each case to determine the amount of specimen to be exposed in successive cycles. The longer specimens included the tissue visible in the corresponding shorter segments. For example, in one scenario, the first case selected was the 10th one reviewed in the first session with 10 mm of tissue exposed. It was then the 21st reviewed in the second session with 15 mm of tissue exposed, the 9th in the third session with 20 mm exposed, and the 17th in the fourth session with 5 mm exposed. Specimens were graded and staged according to the METAVIR scoring system (Figure 1). 16 Steatosis also was graded on a scale of 0 3. The METAVIR scoring system was used because of its excellent reproducibility in assessing fibrosis and inflammation. 17 At the conclusion of the study, 24 specimens were selected at random and repeat evaluation by the same blinded liver pathologist was conducted in a similar manner to assess the degree of intraobserver variability. Figure 1. The METAVIR scoring system. Statistical Analysis For each histologic feature from the METAVIR scoring system, agreement between the 5-mm, 10-mm, and 15-mm exposure lengths with the 20-mm or greater length (considered the gold standard) was estimated by using weighted statistics with the software Stat Xact-4 for Windows (Cytel Software Corporation, Cambridge, MA). Reproducibility of the score pairs of the sections that were 20 mm or greater was estimated in the same manner. Estimated statistics were compared using analysis of variance for a 2-factor factorial design, with Tukey s Studentized Range Tests for 1-tailed pair-wise comparisons, implemented with SAS (SAS Institute, Cary, NC). 18 The statistical evaluation was based on the prerequisite that the biopsy length of 20 mm or greater was the gold standard, and that smaller biopsy specimens were compared with that gold standard. The interpretation of the magnitude of the weighted score is that greater than.75 signifies excellent agreement, with the closer to 1.0 the greater the agreement; and less than.40 indicates poor agreement. 19 Results The 100 needle liver biopsy specimens that were assessed microscopically included 64 that were exactly 20 mm in length, 25 that were mm in length, 10 that were mm in length, and 1 that was more than 30 mm in length. Only 23% of the 464 liver biopsy specimens on the master list were greater than 20 mm in length. On microscopic examination, the various necroinflammatory changes and the stage of fibrosis in the material evaluated encompassed the entire spectrum of histopathology usually seen in CHC (Table 1). Twentynine patients either had severe fibrosis (stage F3), or established cirrhosis (stage F4). Only 1 patient had severe piecemeal and bridging necrosis. In assessing the stage of fibrosis, the weighted score at 5-mm, 10-mm, and 15-mm lengths was.75,.85, and.92, respectively. The results are shown in Table 2. Comparing the reading from the specimens 20 mm or greater with the 5-mm reading, there was perfect agreement in 56 cases, and a 1-stage difference in 31 cases

3 932 SCHIANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 Table 1. Histopathologic Features Assessed in the 20 mm or Longer Specimens Histologic feature Grade/stage Number of patients (total, n 100) Piecemeal necrosis Focal lobular necrosis Portal inflammation Bridging necrosis Yes 1 No 99 Histologic activity score A0 17 A1 55 A2 19 A3 9 Fibrosis score F0 16 F1 18 F2 37 F3 19 F4 10 Steatosis (87% perfect agreement or difference by 1 stage). Similarly, comparing the reading from the specimens 20 mm or greater with the 10-mm reading, there was perfect agreement in 72 cases, and a 1-stage difference in 22 cases (94% perfect agreement or difference by 1 stage). Between the readings from the specimens 20 mm or greater and the 15-mm specimens, there was perfect agreement in 80 cases, and a 1-stage difference in 18 cases (98% perfect agreement or difference by 1 stage). In assessing the overall grade of necroinflammatory activity, the weighted score at 5-mm, 10-mm, and 15-mm lengths was.73,.81, and.77, respectively. The level of agreement between the readings of the different lengths of biopsy specimens is shown in Table 3. Note the high percentage of exact agreement or difference by only 1 grade throughout practically all comparisons. For piecemeal necrosis, the weighted scores for increasing Table 3. Histologic Activity Score: Comparison of the Readings From Specimens 20 mm or Longer With Those of the Shorter Lengths Histologic activity score 15 mm 10 mm 5 mm Perfect agreement N 72 N 74 N 67 1-grade difference N 27 N 26 N 32 2-grade difference N 1 N 0 N 1 order of length were.75,.81, and.84, respectively (Table 4), whereas for focal lobular necrosis the scores were.58,.69, and.61, respectively (data not shown). A summary of the weighted statistics comparing different length readings with the gold standard is shown in Table 5. Statistical analysis using all 100 samples together was performed to assess whether there was agreement with the readings from the specimens 20 mm or greater (averaged over the 6 pathologic features) across all lengths. Intraobserver agreement was very good to excellent for all of the factors, with statistics ranging from.71 for focal lobular necrosis to 1.0 for piecemeal necrosis, and an average across all the features of.89. Note the.98 intraobserver agreement for fibrosis. Although the 10-mm and 15-mm statistics generally were lower than the statistics for the 20-mm or greater specimens, their average values were close to.80, still exceptionally similar in agreement. Overall, the average extent of agreement (values of the statistics) varied significantly among histologic features and among lengths (P.001 for each, by 2-way factorial analysis of variance). Pairwise comparisons were performed to explore the source of the differences among lengths. The comparisons showed that in averaging all the features, the scores of the 20-mm or greater specimens were significantly higher than each of the shorter specimens, and the scores of the 5-mm specimens were significantly lower than each of the longer specimens. The 10-mm and 15-mm scores did not differ significantly from each other. The analysis was repeated with features limited to fibrosis, histologic activity score, and piecemeal necrosis. The extent of agreement did not differ significantly among these 3 features, but the effect of specimen length Table 2. Assessment of Fibrosis: Comparison of the Readings From Specimens 20 mm or Longer With Those of the Shorter Lengths 15 mm 10 mm 5 mm Perfect agreement N 80 N 72 N 56 1-stage difference N 18 N 22 N 31 2-stage difference N 2 N 6 N 13 Table 4. Piecemeal Necrosis: Comparison of the Readings From Specimens 20 mm or Longer With Those of the Shorter Lengths Piecemeal necrosis 15 mm 10 mm 5 mm Perfect agreement N 84 N 81 N 75 1-grade difference N 16 N 19 N 25 2-grade difference N 0 N 0 N 0

4 September 2005 BIOPSY SPECIMEN SIZE IN CHRONIC HEPATITIS C 933 Table 5. Statistics for Agreement Between the Readings of the Specimens 20 mm or Longer and Those of the Corresponding Shorter Lengths Histologic feature 5 mm 10 mm 15 mm 20 mm Piecemeal necrosis Focal lobular necrosis Portal inflammation Histologic activity score Fibrosis score Steatosis Average overall histologic aspects NOTE. Intraobserver agreement is shown at the 20 mm column. remained the same: agreement on repeat evaluation of the scores from the 20-mm or greater specimens was significantly higher than agreement of any of the shorter specimen scores with those of the corresponding scores from the 20-mm or greater specimens; the 5-mm specimen scores were significantly lower than each of the other scores, whereas the 10-mm and 15-mm specimen scores were comparable. Discussion Percutaneous liver biopsy examination is an invaluable tool for the diagnosis and management of CHC. As new therapies for CHC emerge, there is an increasing need for accurate liver biopsy interpretation, including grading of necroinflammatory activity and staging of the fibrotic process. 15,17,20 22 In the present study, we evaluated the accuracy of CHC grading and staging in relation to specimen size in 100 percutaneous needle liver biopsy specimens. All interpretations were made by the same blinded pathologist to avoid interobserver variation. The specimens were examined 4 times after being assigned randomly to batches, as if each one represented 4 specimens of different lengths. Several weeks were allowed to pass between the evaluations of successive batches to minimize any recall effect from previous sessions and to ensure that the specimens were analyzed in as unbiased a manner as possible. The data concerning piecemeal necrosis, histologic activity score, fibrosis score, and steatosis obtained for either the 10-mm or the 15-mm specimens were in excellent agreement with those of the gold standard (ie, the 20-mm specimens, weighted scores.77). Agreement between the data of the 5-mm specimens and the gold standard was not as good; however, the corresponding weighted statistics were still in the range of In terms of the other parameters evaluated, agreement between the 10-mm and 15-mm specimens and the gold standard was good for portal inflammation ( statistics,.76 and.74, respectively), but not as good for focal lobular necrosis ( statistics,.69 and.61, respectively). The agreement of these 2 parameters between the 5-mm specimens and the gold standard was suboptimal ( statistics,.58 for both parameters). These findings indicate that the histologic changes of CHC involve the hepatic parenchyma in a relatively uniform way, with most 10-mm specimens reflecting accurately the pathology of corresponding longer specimens. However, the degree of uniformity is not the same for all parameters assessed. Our data suggest that piecemeal necrosis has a more regular distribution in the parenchyma than focal lobular necrosis. This is probably why the statistics showed better agreement in terms of focal lobular necrosis between the 10-mm specimens and the gold standard, as compared with between the 15-mm specimens and the gold standard. In this study, bridging necrosis was observed in only 1 specimen, confirming previous observations that this is an unusual histologic feature in CHC. 17 The importance of excellent agreement in fibrosis scores ( statistics for 10-mm and 15-mm specimens,.85 and.92, respectively) cannot be overemphasized because this score represents the measure of disease progression over time. Because semiquantitative assessment of histologic findings is a subjective process, at the end of this study a group of randomly selected specimens were re-evaluated to test for intraobserver variation. The excellent agreement between the 2 separate reviews of the same material that were performed several months apart attest to the reliability of our findings. The basic principle of the approach that we used in this project was introduced by 2 studies of liver biopsy specimens conducted by a group from Denmark over 20 years ago. 23,24 However, the goal of those studies was to establish the correct diagnosis rather than to score specific histologic changes, which was the focus of our study. After a review of 50 biopsy specimens with chronic aggressive hepatitis, those investigators concluded that

5 934 SCHIANO ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 9 acceptable diagnostic accuracy required a specimen length of at least 15 mm. 24 Since that time, significant changes have occurred in the way liver biopsy specimens are interpreted and new therapies for chronic hepatitis have been developed. Therefore, it is important to establish guidelines for specimen adequacy that reflect our current needs to grade and stage CHC reliably. It is widely believed that an adequate liver biopsy specimen should contain a minimum number of portal tracts. Most pathologists are satisfied with specimens containing 6 8 portal tracts, 7 although some studies now suggest that a greater number of portal tracts may be required for adequacy. 3,8,11 However, the spacing of portal tracts within the hepatic parenchyma varies significantly from specimen to specimen; therefore, any requirement regarding their number is of no practical value to the physician performing the biopsy procedure. This is why in the present study we chose to assess specimen adequacy on the basis of specimen length alone rather than on the number of portal tracts. Although all specimens were obtained using a Menghini needle, there is no reason to believe that these results would not apply to specimens obtained with other types of liver biopsy needles. In a recent study, Colloredo et al 8 evaluated specimen adequacy in a series of 161 liver biopsy specimens from patients with chronic hepatitis B or C. Their approach was similar to ours; however, there were a number of differences, such as: (1) a specimen length of 30 mm or greater was considered the gold standard; (2) the Ishak scoring system 25 was used for grading and staging; (3) all biopsy specimens of the same length were examined in 1 session; (4) statistics (the preferred method in this type of study 26 ) were not used; (5) both hepatitis B and C were evaluated; and (6) intraobserver variation was not assessed. These investigators 8 concluded that a specimen length of at least 20 mm should be recommended as a reliable size for grading and staging chronic hepatitis B and C. The earlier-mentioned differences in methodology do not allow a direct comparison between the findings of their study and the findings from our study. However, from a practical point of view, it should be taken into account that the majority of percutaneous needle biopsy specimens are not larger than the gold standard of 20 mm that we used. In fact, only 23% of the biopsy specimens included in our master list, with all procedures performed by experienced hepatologists, were greater than 20 mm in length. Because our assessment shows that there is minimal, if any, difference between the findings at the 10- and 20-mm specimen size, our results should be applicable to biopsy specimens with smaller sizes than the ones included in the present study. A limitation of our findings may apply to specimens with cirrhosis because only 10 of our patients had stage 4 fibrosis. Furthermore, biopsy specimens from cirrhotic livers, especially those with macronodular cirrhosis, are known to be prone to understaging because of sampling error. 9 In a recent study, Bedossa et al 27 addressed the reliability of fibrosis assessment in 17 surgical specimens with CHC. Evaluation of necroinflammatory activity was not part of their study. The investigators used image analysis; from the digitized image of each section, multiple virtual needle biopsy specimens of increasing length were produced. By using the METAVIR scoring system, 65% of the specimens measuring 15 mm were categorized correctly; this increased to 75% for the 25-mm specimens, without any substantial benefit for longer specimens. Therefore, the investigators suggested that a biopsy specimen length of at least 25 mm is necessary to evaluate fibrosis accurately. To obtain specimens consistently that are 25 mm in length or longer, however, is impractical to expect in the majority of needle biopsy procedures. The study by Bedossa et al 27 suggested that the needle biopsy procedure, irrespective of specimen length, has a significant sampling error as compared with the hepatic resection specimen. The findings of these investigators cannot be compared with ours, not only because of the differences in methodology, but also because of the different specimen types evaluated. As far as image analysis is concerned, it does provide an objective means of assessing fibrosis, but it is cumbersome, not routinely available, less standardized, and open to as much subjective interpretation as pathologist scoring. 26 It also should be kept in mind that the assessment of fibrosis in biopsy specimens by a pathologist is not limited to the overall amount of fibrous tissue, but also includes its distribution and features of the interface between the lobular parenchyma and the fibrous tissue. In conclusion, our findings indicate that liver biopsy specimens with a length of at least 10 mm usually reflect the grade of necroinflammatory activity and the stage of fibrosis reliably in patients with CHC. Relatively little improvement in diagnostic accuracy is afforded with longer specimens. This especially is true in the assessment of fibrosis; in our study, 94% of the 10-mm long specimens were assessed either identically to or within 1 stage difference of the gold standard of 20 mm or greater. This is reassuring because only a minority of the percutaneous liver biopsy procedures performed at our insti-

6 September 2005 BIOPSY SPECIMEN SIZE IN CHRONIC HEPATITIS C 935 tution were 20 mm or longer in length, which probably reflects what is seen at other centers. Further studies should be performed to ascertain whether these results can be extrapolated to other patient groups, such as in posttransplantation hepatitis C recurrence. References 1. Desmet VJ, Gerber MA, Hoofnagle JH, et al. Classification of chronic hepatitis: diagnosis, grading and staging. Hepatology 1994;19: Dienstag JL. The role of liver biopsy in chronic hepatitis C. Hepatology 2002;36:S152 S Zarski JP, McHutchison J, Bronowicki JP, et al. Rate of natural disease progression in patients with chronic hepatitis C. J Hepatol 2003;38: National Institute of Health. Consensus Development Conference Statement: Management of Hepatitis C: 2002 June 10 12, Hepatology 2002;36:S3 S Braunstein H. Needle biopsy of the liver in cirrhosis; diagnostic efficiency as determined by postmortem sampling. Arch Pathol 1956;62: Maharaj B, Maharaj RJ, Leary WP, et al. Sampling variability and its influence on the diagnostic yield of percutaneous needle biopsy of the liver. Lancet 1986;1: Bravo AA, Sheth SG, Chopra S. Current concepts: liver biopsy. N Engl J Med 2001;344: Colloredo G, Guido M, Sonzogni A, et al. Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease. J Hepatol 2003;39: Afdhal NN, Nunes D. Evaluation of liver fibrosis: a concise review. Am J Gastroenterol 2004;99: Yano M, Kumada H, Kage M, et al. The long-term pathological evolution of chronic hepatitis C. Hepatology 1996;23: Kaserer K, Fiedler R, Steindl P, et al. Liver biopsy is a useful predictor of response to interferon therapy in chronic hepatitis C. Histopathology 1998;32: Saadeh S, Cammell G, Carey WD, et al. The role of liver biopsy in chronic hepatitis C. Hepatology 2001;33: Fontana RJ, Lok ASF. Non-invasive monitoring of patients with chronic hepatitis C. Hepatology 2002;36:S57 S Poynard T, McHutchinson J, Manns M, et al. Biochemical surrogate markers of liver fibrosis and activity on a randomized trial of peginterferon alfa-2b and ribavirin. Hepatology 2003;38: Rosenberg WM, Voelker M, Thiel R, et al. Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology 2004;127: Bedossa P, Poynard T, for the French METAVIR Cooperative Study Group. An algorithm for grading activity in chronic hepatitis C. Hepatology 1996;24: The METAVIR Cooperative Group. Inter- and intra-observer variation in the assessment of liver biopsy of chronic hepatitis C. Hepatology 1994;20: SAS Institute Inc. SAS/STAT user s guide. Version 8. Cary, NC: SAS Institute Inc., Fleiss JL, Levin B, Paik MC. In: Statistical methods for rates and proportions. 3rd ed. New York: John Wiley & Sons, 2003: Ishak K, Baptista A, Bianchi L, et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995;22: Poynard T, McHutchison J, Manns M, et al. Impact of pegylated interferon alfa-2b and ribavirin on liver fibrosis in patients with chronic hepatitis C. Gastroenterology 2002;122: Scheuer PJ. Liver biopsy size matters in chronic hepatitis: bigger is better. Hepatology 2003;38: Holund B, Poulsen H, Schlichting P. Reproducibility of liver biopsy diagnosis in relation to the size of the specimen. Scand J Gastroenterol 1980;15: Schlichting P, Holund B, Poulsen H. Liver biopsy in chronic aggressive hepatitis. Diagnostic reproducibility in relation to size of specimen. Scand J Gastroenterol 1983;18: Ishak K, Baptista A, Bianchi L, et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995;22: Demetris AJ, Ruppert K. Pathologist s perspective on liver needle biopsy size? J Hepatol 2003;39: Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003;38: Address requests for reprints to: Thomas D. Schiano, MD, Division of Liver Diseases, Box 1104, The Mount Sinai Medical Center, One Gustave L. Levy Place, New York, New York thomas.schiano@msnyuhealth.org; fax: (212)