+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
Clinics and Research in Hepatology and Gastroenterology (2014) xxx, xxx—xxx
Available online at
ScienceDirect www.sciencedirect.com
COMMENTARY
Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma Paulette Bioulac-Sage a,b,∗, Christine Sempoux c, Charles Balabaud a a
Inserm U1053, Université Bordeaux Segalen, 33076 Bordeaux cedex, France Service de Pathologie, Hôpital Pellegrin, CHU de Bordeaux, Bordeaux, France c Service d’Anatomie pathologique, Cliniques Universitaires Saint-Luc, 1200 Bruxelles, Belgium b
It is clinically important to differentiate focal nodular hyperplasia (FNH) from inflammatory hepatocellular adenoma (IHCA) because FNH does not need surgical resection. However, these two types of benign liver tumors may share some pathological similarities, at least in some areas. In a recent article published in Modern Pathology [1], authors from three different US academic centers (San Francisco, Baltimore and Seattle) studied the diagnostic utility and limitations of glutamine synthetase (GS) and serum amyloid-associated (SAA) protein immunohistochemistry in the distinction of FNH from IHCA. Standard histology and immunohistochemistry were analyzed in 54 IHCA, 40 FNH, and 3 indeterminate lesions. Their results are summarized in Table 1. Before going into the analysis of the results of this study, it is important to remind that the interpretation of immunohistochemistry (IHC) is based on molecular data [2].
∗ Corresponding author. Inserm U1053, Université Bordeaux Segalen, 33076 Bordeaux cedex, France. Tel.: +33 5 57 57 17 72; fax: +33 5 56 51 40 77. E-mail address: [email protected] (P. Bioulac-Sage).
FNH Gene expression studies have revealed molecular features supporting the theory of vascular abnormalities as a driving event in the formation of FNH. Indeed, FNH present abnormal expression of genes involved in angiogenesis with an increase in expression of the ANGPT1/ANGPT2 ratio [2]. Another major finding is the overexpression of Wnt/b-catenin target genes, including GLUL [3] coding for glutamine synthetase (GS). In a normal liver lobule, GS is expressed only in 2 to 3 hepatocyte layers around the central vein. In contrast, in FNH, there is an overexpression of GS that shows a specific location with a map-like pattern predominantly in the periphery of the nodules [4]. This heterogeneous distribution of GS is caused by b-catenin activation without b-catenin mutations, in accordance with the polyclonal origin of FNH [3]. Despite the fact that GS staining is a major marker to identify FNH, the map-like pattern may be missing in rare cases, or difficult to interpret (especially in biospies), the final conclusion relying then on additional markers such as CK7 as well as on the elimination of well-identified HCA subtypes [5].
2210-7401/$ – see front matter © 2014 Published by Elsevier Masson SAS. http://dx.doi.org/10.1016/j.clinre.2014.01.001
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
2
P. Bioulac-Sage et al. Table 1
Immunohistochemistry data (GS and SAA) in FNH and IHCA obtained in the article [1].
SAA
GS
Tumor (54 cases) IHCA Positive 50/54 (93%) Focal 6/54 Diffuse 44/54 Negative 4/54 (should be 4 and not 8)
Negative 2/53 (4%) Perivascular and/or patchy 39/53 (74%) Pseudo map-like 8/53 (15%) Classical map-like: 0/54 Diffuse 4/53 (8%); 3 b-cat positive
Non-tumoral liver (23 cases) Positive 14/23 (61%) but less intense than in tumor
Normal staining in perivenular area
Tumor (40 cases) FNH Positive 7/40: 2 patchy, 5 diffuse Negative 33/40 (83%) Non-tumoral liver (9 cases) Positive 5/9 (56%) 3 cases with no staining in the tumor 2 cases with the same level than in the tumor
Classical map-like 36/40 (90%) Pseudo map-like 4/40 (10%) all on biopsies
Normal staining in perivenular area
GS: glutamine synthetase; SAA: serum amyloid-associated; FNH: focal nodular hyperplasia; IHCA: inflammatory hepatocellular adenoma.
FNH-like nodules Classically they are large regenerative nodules (a different entity from FNH) described in cirrhotic patients, and presenting an immunohistochemical profile similar to that of other cirrhotic samples and different from typical FNH [3]. In particular, in FNH-like nodules, b-catenin-induced perivenous GS is significantly down regulated compared to non-tumoral liver and FNH [3]. Consistent with this result, very little parenchymal GS staining is observed in FNH-like nodules [3]; however, in rare cases, GS staining is occasionally mildly positive, resembling FNH [6]. Notably, in the final stage of cirrhosis, GS is barely detectable in nodules, whereas in the early stage of compensated cirrhosis, GS expression is quasi normal [7]. In addition, the expression of C reactive protein (CRP) in FNH-like nodules is either absent or weak, focally or diffusely [6]. CRP an inflammatory protein is used to identify inflammatory HCA (see below).
Inflammatory HCA (IHCA) The main feature of IHCA is the activation of the JAK/STAT pathway [2]. Five different molecular drivers, namely IL6ST (coding for gp130), STAT3, GNAS, Jak1 and FRK were discovered as oncogenes activated by mutations in IHCA [2,8]. Each mutation is exclusive from the others and the different mutated genes can explain more than 80% of the overall IHCAs. So, in approximately 20% of IHCAs, no gene defect that could explain the inflammatory phenotype has been identified so far. In general, the recent paper of Joseph et al. [1] who quantified IHC data in FNH and IHCA confirms previous data showing the progress made in the discrimination of FNH from IHCA using GS for the identification of FNH and inflammatory proteins (serum amyloid A [SAA] or CRP) for the identification of IHCA (Table 1). The authors, however, pinpoint the limits of the results obtained with the 2 markers (GS and
SAA) they used, and the pitfalls in the identification of the 2 types of nodules (see Table 5 of ref [1]). There are several explanations for these results: • as molecular biology cannot be used in routine practice, IHC is a valid and robust surrogate technique. However, the interpretation of IHC should correspond to strict welldefined criteria [5]. The criteria defined in this paper have not yet been validated; • IHC interpretation is based on the comparison between tumoral and non-tumoral livers. However, in this study, non-tumoral tissues were available in only 23/54 IHCA and 9/40 FNH, respectively. It is also necessary to avoid any misinterpretation related to rare technical problems; • it is easy to understand that IHC interpretation is more difficult in liver biopsy; • the lack of information related to background liver disease; • the partial lack of basic clinical (age and gender were available in 43/54 IHCA and 35/40 FNH, no information on liver disease, etc.), biological, radiological and pathological data particularly the size of the tumor, and/or the length and diameter of the biopsy, evidence of remodeling by bleeding, necrosis. In remodeled HCA, fibrotic bands with some degree of ductular reaction are not rare creating occasionally confusion with FNH. In our institution, interpretation of IHC is based on molecular data as much as possible [5] (Table 2) as described below.
SAA or CRP staining interpretation In HCA Positive staining is interpreted as a surrogate marker of IHCA [2,5]. In the paper, the authors scored positivity from
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
Immunohistochemical pitfalls in the diagnosis of benign hepatocellular tumors Table 2
3
Our personal data interpretation of immunohistochemistry (GS and CRP) in FNH and IHCA developed in a normal livera .
CRP Tumor IHCA Positivity diffuse (positivity is the hallmark of the diagnosis) [9] (variable intensity, but usually strong; in case of extremely weak CRP staining the diagnosis of IHCA cannot be made without searching for mutations [MB]) CRP cannot be interpreted if CRP is positive in NT [9] Atrophic hepatocytes (major sinusoidal dilatation) can be CRP negative Non-tumoral liver Negativeb Weak centrolular staining can be observed (no diagnosis implication) b-IHCA As above
Tumor FNH [4] Negative (negativity is one of the hallmark) CRP is often focally positive around Inflammatory cells (no diagnosis implication) CRP can be positive in case of a generalized inflammatory stateb Non-tumoral liver Negativeb Weak centrolobular staining can be observed (no diagnosis implication)
GS Negativity Hepatocytes surrounding veins [9] can be positive as well as the periphery of the tumor (unknown meaning)
Faint background staining (diffuse or focal) is difficult to interpret as possibly related to b-cat mutations (usually negative by IHC) need for MB Strong: see b-IHCA below
Normal perivenular staining
Diffusely homogeneously positive, b-catenin often + good correlation with b-catenin mutations (MB) Patchy [9] (from mild to strong, usually diffuse) correlation with b-catenin mutations is not yet well establish, except if b-catenin staining is positive
Positive (classical map-like [staining mild to strong and more or less spread within the nodule]) Negativity (particularly on a biopsy) does not eliminate the diagnosis (sampling problem)
Normal perivenular staining
GS: glutamine synthetase; CRP: C reactive protein; FNH: focal nodular hyperplasia; IHCA: inflammatory hepatocellular adenoma; MB: molecular biology; NT: non-tumoral liver. a Data in the cirrhotic liver may be different. b Can be positive in particular cases (portal embolization. . .).
0 to 3 (0 absent, 3 strong) and classified SAA positivity as focal < 50% and diffuse > 50%. In our hands, the staining is positive (from mild to strong) or negative; diffuse positivity means close to 100% positivity and focal positivity means expression around inflammatory cells, a frequent finding in FNH and HCA which is not considered significant for the diagnosis of either nodules. CRP has always been more reliable than SAA in our experience. In addition, so far, CRP negative nodule has never been proven to be IHCA by molecular biology.
positive staining of CRP or SAA in the non-tumoral liver may occur and this excludes totally the interpretation of their positivity in the tumor (FNH or HCA). Indeed a positive CRP staining is observed in the non-tumoral liver after portal vein embolization, or bleeding and, probably, in all circumstances in which there is an inflammatory syndrome [9]; therefore, the positivity in the non-tumoral liver has not the same physiopathological meaning than in the tumor.
In FNH
In HCA, GS is interpreted as a surrogate marker of b-catenin mutation [2,5]. The distinction between perivascular on one hand, and patchy and perivascular on the other hand, as described in this paper, either in FNH or IHCA [1] seems confusing. Indeed
A diffuse positive immunohistochemical expression of CRP or SAA in FNH is puzzling if the non-tumoral liver is negative, and we have never seen such a case. By contrast,
GS staining interpretation
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
4
P. Bioulac-Sage et al.
we do not know if it makes sense to interpret GS in IHCA as in the normal liver. In IHCA, GS staining goes from absent, to few hepatocytes in one single row to several rows of hepatocytes, always around veins. It can also be strong and diffuse indicating therefore a beta-catenin activated IHCA (b-IHCA). ‘‘Perivascular and/or patchy staining (predominantly perivascular or perivascular staining accompanied by staining of variable intensity in clusters of hepatocytes beyond the perivascular region)’’ as defined by Joseph et al., has to our knowledge no specific meaning sustained by molecular data. The term patchy [9] has been previously used when the staining was irregularly dispersed in the whole tumor (not only perivascular), with a variable intensity, and with the understanding that it could be linked to b-catenin mutations, in spite of the fact that b-catenin nuclear staining was often absent. Molecular studies are still ongoing to understand this finding. The real problem concerns GS ‘‘pseudo map-like staining’’. For the authors it means a staining in clusters of hepatocytes beyond the perivascular region showing interconnected areas. It differs from the classical map-like staining in three aspects: • the positive interconnected bands of hepatocytes are narrow and inconspicuous; • staining intensity is variable, typically mild to moderate, and fades in intensity away from perivascular areas; • the staining is focal and confined to periphery. Anything focal and localized at the periphery, in our experience, should be interpreted cautiously and should not lead to a specific category of lesions until proven to be related to b-catenin mutations by molecular biology. Furthermore it is necessary to take into account the possibility of peritumoral hyperplasia as suggested by the work of Wanless et al. [10]. From what has been mentioned above, it is obvious that the interpretation of GS and CRP/SAA staining can already be difficult in surgical specimen, so even more difficult on a biopsy [11] and all kinds of misinterpretation (summarized in Table 5 of ref [1]) can be considered if the interpretation does not take into account: • the pathomolecular relationships; • the geographical distribution of the markers; • the staining of the background liver. Not surprisingly there are still indeterminate lesions that require molecular analysis. In this article, three cases could not be classified as FNH or IHCA and were labeled as indeterminate. In these 3 cases, SAA staining was positive and GS presented a pseudo map-like staining. To our opinion, these nodules require further investigations (including CRP immunostaining) before to conclude to indeterminate lesions. It is important in clinical practice to proceed step-by-step with IHC, as previously described [12]. In addition, we prefer to perform CRP staining, alone or in association with SAA, particularly if SAA staining is not straightforward. To extend the discussion, the interpretation of GS and CRP staining to differentiate FNH from IHCA is not a major issue in our experience. The main issue remains the differential diagnosis
between FNH without typical map-like pattern of GS staining and unclassified HCA (particularly when they are remodeled) with some pathological features observed in FNH. In conclusion: • the final diagnosis of FNH or HCA must be made in light of the clinical, radiological, and morphological features, using appropriately IHC markers such as GS and CRP whose interpretation requires knowledge of underlying molecular pathways; • there is an urgent need to discuss the difficulties in IHC interpretation among experienced liver pathologists to assess the range of varieties in the immunostaining and to avoid confusion in the diagnosis of benign hepatocellular tumors.
Disclosure of interest The authors declare that they have no conflicts of interest concerning this article.
References [1] Joseph NM, Ferrell LD, Jain D, Torbenson MS, Wu TT, Yeh MM, et al. Diagnostic utility and limitations of glutamine synthetase and serum amyloid-associated protein immunohistochemistry in the distinction of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Mod Pathol 2013;27: 62—72. [2] Nault JC, Bioulac-Sage P, Zucman-Rossi J. Hepatocellular benign tumors-from molecular classification to personalized clinical care. Gastroenterology 2013;144:888—902. [3] Rebouissou S, Couchy G, Libbrecht L, Balabaud C, Imbeaud S, Auffray C, et al. The beta-catenin pathway is activated in focal nodular hyperplasia but not in cirrhotic FNH-like nodules. J Hepatol 2008;49:61—71. [4] Bioulac-Sage P, Laumonier H, Rullier A, Cubel G, Laurent C, Zucman-Rossi J, et al. Overexpression of glutamine synthetase in focal nodular hyperplasia: a novel easy diagnostic tool in surgical pathology. Liver Int 2009;29:459—65. [5] Bioulac-Sage P, Rebouissou S, Thomas C, Blanc JF, Saric J, Sa Cunha A, et al. Hepatocellular adenoma subtype classification using molecular markers and immunohistochemistry. Hepatology 2007;46:740—8. [6] Bioulac-Sage P, Sempoux C, Paradis V, Ferrell L, Scoazec JY, Leteutre E, et al. Characterization of liver nodules in patients with Budd-Chiari syndrome (BCS) and portal vein agenesis (PVA). Hepatology 2012;52:607 [Abstract]. [7] Fleming KE, Wanless IR. Glutamine synthetase expression in activated hepatocyte progenitor cells and loss of hepatocellular expression in congestion and cirrhosis. Liver Int 2013;33:525—34. [8] Pilati C, Nault JC, Latouzé E, Imbeaud S, Mallet M, Boulai A, et al. Integrative genomic profiling of hepatocellular adenoma identify mutational processes involved in malignant transformation. Hepatology 2013;58:55 [Abstract]. [9] Bioulac-Sage P, Cubel G, Balabaud C, Zucman-Rossi J. Revisiting the pathology of resected benign hepatocellular nodules using new immunohistochemical markers. Semin Liver Dis 2011;31:91—103. [10] Arnason T, Fleming KE, Wanless IR. Peritumoral hyperplasia of the liver: a response to portal vein invasion by hypervascular neoplasms. Histopathology 2013;62:458—64.
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
Immunohistochemical pitfalls in the diagnosis of benign hepatocellular tumors [11] Bioulac-Sage P, Cubel G, Taouji S, Scoazec JY, Leteurtre E, Paradis V, et al. Immunohistochemical markers on needle biopsies are helpful for the diagnosis of focal nodular hyperplasia and hepatocellular adenoma subtypes. Am J Surg Pathol 2012;36:1691—9.
5
[12] Balabaud C, Al-Rabih WR, Chen PJ, Evason K, Ferrell L, Hernandez-Prera JC, et al. Focal nodular hyperplasia and hepatocellular adenoma around the world viewed through the scope of the immunopathological classification. Int J Hepatol 2013;2013:268625.
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
ARTICLE IN PRESS
Clinics and Research in Hepatology and Gastroenterology (2014) xxx, xxx—xxx
Available online at
ScienceDirect www.sciencedirect.com
COMMENTARY
Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma Paulette Bioulac-Sage a,b,∗, Christine Sempoux c, Charles Balabaud a a
Inserm U1053, Université Bordeaux Segalen, 33076 Bordeaux cedex, France Service de Pathologie, Hôpital Pellegrin, CHU de Bordeaux, Bordeaux, France c Service d’Anatomie pathologique, Cliniques Universitaires Saint-Luc, 1200 Bruxelles, Belgium b
It is clinically important to differentiate focal nodular hyperplasia (FNH) from inflammatory hepatocellular adenoma (IHCA) because FNH does not need surgical resection. However, these two types of benign liver tumors may share some pathological similarities, at least in some areas. In a recent article published in Modern Pathology [1], authors from three different US academic centers (San Francisco, Baltimore and Seattle) studied the diagnostic utility and limitations of glutamine synthetase (GS) and serum amyloid-associated (SAA) protein immunohistochemistry in the distinction of FNH from IHCA. Standard histology and immunohistochemistry were analyzed in 54 IHCA, 40 FNH, and 3 indeterminate lesions. Their results are summarized in Table 1. Before going into the analysis of the results of this study, it is important to remind that the interpretation of immunohistochemistry (IHC) is based on molecular data [2].
∗ Corresponding author. Inserm U1053, Université Bordeaux Segalen, 33076 Bordeaux cedex, France. Tel.: +33 5 57 57 17 72; fax: +33 5 56 51 40 77. E-mail address: [email protected] (P. Bioulac-Sage).
FNH Gene expression studies have revealed molecular features supporting the theory of vascular abnormalities as a driving event in the formation of FNH. Indeed, FNH present abnormal expression of genes involved in angiogenesis with an increase in expression of the ANGPT1/ANGPT2 ratio [2]. Another major finding is the overexpression of Wnt/b-catenin target genes, including GLUL [3] coding for glutamine synthetase (GS). In a normal liver lobule, GS is expressed only in 2 to 3 hepatocyte layers around the central vein. In contrast, in FNH, there is an overexpression of GS that shows a specific location with a map-like pattern predominantly in the periphery of the nodules [4]. This heterogeneous distribution of GS is caused by b-catenin activation without b-catenin mutations, in accordance with the polyclonal origin of FNH [3]. Despite the fact that GS staining is a major marker to identify FNH, the map-like pattern may be missing in rare cases, or difficult to interpret (especially in biospies), the final conclusion relying then on additional markers such as CK7 as well as on the elimination of well-identified HCA subtypes [5].
2210-7401/$ – see front matter © 2014 Published by Elsevier Masson SAS. http://dx.doi.org/10.1016/j.clinre.2014.01.001
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
2
P. Bioulac-Sage et al. Table 1
Immunohistochemistry data (GS and SAA) in FNH and IHCA obtained in the article [1].
SAA
GS
Tumor (54 cases) IHCA Positive 50/54 (93%) Focal 6/54 Diffuse 44/54 Negative 4/54 (should be 4 and not 8)
Negative 2/53 (4%) Perivascular and/or patchy 39/53 (74%) Pseudo map-like 8/53 (15%) Classical map-like: 0/54 Diffuse 4/53 (8%); 3 b-cat positive
Non-tumoral liver (23 cases) Positive 14/23 (61%) but less intense than in tumor
Normal staining in perivenular area
Tumor (40 cases) FNH Positive 7/40: 2 patchy, 5 diffuse Negative 33/40 (83%) Non-tumoral liver (9 cases) Positive 5/9 (56%) 3 cases with no staining in the tumor 2 cases with the same level than in the tumor
Classical map-like 36/40 (90%) Pseudo map-like 4/40 (10%) all on biopsies
Normal staining in perivenular area
GS: glutamine synthetase; SAA: serum amyloid-associated; FNH: focal nodular hyperplasia; IHCA: inflammatory hepatocellular adenoma.
FNH-like nodules Classically they are large regenerative nodules (a different entity from FNH) described in cirrhotic patients, and presenting an immunohistochemical profile similar to that of other cirrhotic samples and different from typical FNH [3]. In particular, in FNH-like nodules, b-catenin-induced perivenous GS is significantly down regulated compared to non-tumoral liver and FNH [3]. Consistent with this result, very little parenchymal GS staining is observed in FNH-like nodules [3]; however, in rare cases, GS staining is occasionally mildly positive, resembling FNH [6]. Notably, in the final stage of cirrhosis, GS is barely detectable in nodules, whereas in the early stage of compensated cirrhosis, GS expression is quasi normal [7]. In addition, the expression of C reactive protein (CRP) in FNH-like nodules is either absent or weak, focally or diffusely [6]. CRP an inflammatory protein is used to identify inflammatory HCA (see below).
Inflammatory HCA (IHCA) The main feature of IHCA is the activation of the JAK/STAT pathway [2]. Five different molecular drivers, namely IL6ST (coding for gp130), STAT3, GNAS, Jak1 and FRK were discovered as oncogenes activated by mutations in IHCA [2,8]. Each mutation is exclusive from the others and the different mutated genes can explain more than 80% of the overall IHCAs. So, in approximately 20% of IHCAs, no gene defect that could explain the inflammatory phenotype has been identified so far. In general, the recent paper of Joseph et al. [1] who quantified IHC data in FNH and IHCA confirms previous data showing the progress made in the discrimination of FNH from IHCA using GS for the identification of FNH and inflammatory proteins (serum amyloid A [SAA] or CRP) for the identification of IHCA (Table 1). The authors, however, pinpoint the limits of the results obtained with the 2 markers (GS and
SAA) they used, and the pitfalls in the identification of the 2 types of nodules (see Table 5 of ref [1]). There are several explanations for these results: • as molecular biology cannot be used in routine practice, IHC is a valid and robust surrogate technique. However, the interpretation of IHC should correspond to strict welldefined criteria [5]. The criteria defined in this paper have not yet been validated; • IHC interpretation is based on the comparison between tumoral and non-tumoral livers. However, in this study, non-tumoral tissues were available in only 23/54 IHCA and 9/40 FNH, respectively. It is also necessary to avoid any misinterpretation related to rare technical problems; • it is easy to understand that IHC interpretation is more difficult in liver biopsy; • the lack of information related to background liver disease; • the partial lack of basic clinical (age and gender were available in 43/54 IHCA and 35/40 FNH, no information on liver disease, etc.), biological, radiological and pathological data particularly the size of the tumor, and/or the length and diameter of the biopsy, evidence of remodeling by bleeding, necrosis. In remodeled HCA, fibrotic bands with some degree of ductular reaction are not rare creating occasionally confusion with FNH. In our institution, interpretation of IHC is based on molecular data as much as possible [5] (Table 2) as described below.
SAA or CRP staining interpretation In HCA Positive staining is interpreted as a surrogate marker of IHCA [2,5]. In the paper, the authors scored positivity from
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
Immunohistochemical pitfalls in the diagnosis of benign hepatocellular tumors Table 2
3
Our personal data interpretation of immunohistochemistry (GS and CRP) in FNH and IHCA developed in a normal livera .
CRP Tumor IHCA Positivity diffuse (positivity is the hallmark of the diagnosis) [9] (variable intensity, but usually strong; in case of extremely weak CRP staining the diagnosis of IHCA cannot be made without searching for mutations [MB]) CRP cannot be interpreted if CRP is positive in NT [9] Atrophic hepatocytes (major sinusoidal dilatation) can be CRP negative Non-tumoral liver Negativeb Weak centrolular staining can be observed (no diagnosis implication) b-IHCA As above
Tumor FNH [4] Negative (negativity is one of the hallmark) CRP is often focally positive around Inflammatory cells (no diagnosis implication) CRP can be positive in case of a generalized inflammatory stateb Non-tumoral liver Negativeb Weak centrolobular staining can be observed (no diagnosis implication)
GS Negativity Hepatocytes surrounding veins [9] can be positive as well as the periphery of the tumor (unknown meaning)
Faint background staining (diffuse or focal) is difficult to interpret as possibly related to b-cat mutations (usually negative by IHC) need for MB Strong: see b-IHCA below
Normal perivenular staining
Diffusely homogeneously positive, b-catenin often + good correlation with b-catenin mutations (MB) Patchy [9] (from mild to strong, usually diffuse) correlation with b-catenin mutations is not yet well establish, except if b-catenin staining is positive
Positive (classical map-like [staining mild to strong and more or less spread within the nodule]) Negativity (particularly on a biopsy) does not eliminate the diagnosis (sampling problem)
Normal perivenular staining
GS: glutamine synthetase; CRP: C reactive protein; FNH: focal nodular hyperplasia; IHCA: inflammatory hepatocellular adenoma; MB: molecular biology; NT: non-tumoral liver. a Data in the cirrhotic liver may be different. b Can be positive in particular cases (portal embolization. . .).
0 to 3 (0 absent, 3 strong) and classified SAA positivity as focal < 50% and diffuse > 50%. In our hands, the staining is positive (from mild to strong) or negative; diffuse positivity means close to 100% positivity and focal positivity means expression around inflammatory cells, a frequent finding in FNH and HCA which is not considered significant for the diagnosis of either nodules. CRP has always been more reliable than SAA in our experience. In addition, so far, CRP negative nodule has never been proven to be IHCA by molecular biology.
positive staining of CRP or SAA in the non-tumoral liver may occur and this excludes totally the interpretation of their positivity in the tumor (FNH or HCA). Indeed a positive CRP staining is observed in the non-tumoral liver after portal vein embolization, or bleeding and, probably, in all circumstances in which there is an inflammatory syndrome [9]; therefore, the positivity in the non-tumoral liver has not the same physiopathological meaning than in the tumor.
In FNH
In HCA, GS is interpreted as a surrogate marker of b-catenin mutation [2,5]. The distinction between perivascular on one hand, and patchy and perivascular on the other hand, as described in this paper, either in FNH or IHCA [1] seems confusing. Indeed
A diffuse positive immunohistochemical expression of CRP or SAA in FNH is puzzling if the non-tumoral liver is negative, and we have never seen such a case. By contrast,
GS staining interpretation
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
4
P. Bioulac-Sage et al.
we do not know if it makes sense to interpret GS in IHCA as in the normal liver. In IHCA, GS staining goes from absent, to few hepatocytes in one single row to several rows of hepatocytes, always around veins. It can also be strong and diffuse indicating therefore a beta-catenin activated IHCA (b-IHCA). ‘‘Perivascular and/or patchy staining (predominantly perivascular or perivascular staining accompanied by staining of variable intensity in clusters of hepatocytes beyond the perivascular region)’’ as defined by Joseph et al., has to our knowledge no specific meaning sustained by molecular data. The term patchy [9] has been previously used when the staining was irregularly dispersed in the whole tumor (not only perivascular), with a variable intensity, and with the understanding that it could be linked to b-catenin mutations, in spite of the fact that b-catenin nuclear staining was often absent. Molecular studies are still ongoing to understand this finding. The real problem concerns GS ‘‘pseudo map-like staining’’. For the authors it means a staining in clusters of hepatocytes beyond the perivascular region showing interconnected areas. It differs from the classical map-like staining in three aspects: • the positive interconnected bands of hepatocytes are narrow and inconspicuous; • staining intensity is variable, typically mild to moderate, and fades in intensity away from perivascular areas; • the staining is focal and confined to periphery. Anything focal and localized at the periphery, in our experience, should be interpreted cautiously and should not lead to a specific category of lesions until proven to be related to b-catenin mutations by molecular biology. Furthermore it is necessary to take into account the possibility of peritumoral hyperplasia as suggested by the work of Wanless et al. [10]. From what has been mentioned above, it is obvious that the interpretation of GS and CRP/SAA staining can already be difficult in surgical specimen, so even more difficult on a biopsy [11] and all kinds of misinterpretation (summarized in Table 5 of ref [1]) can be considered if the interpretation does not take into account: • the pathomolecular relationships; • the geographical distribution of the markers; • the staining of the background liver. Not surprisingly there are still indeterminate lesions that require molecular analysis. In this article, three cases could not be classified as FNH or IHCA and were labeled as indeterminate. In these 3 cases, SAA staining was positive and GS presented a pseudo map-like staining. To our opinion, these nodules require further investigations (including CRP immunostaining) before to conclude to indeterminate lesions. It is important in clinical practice to proceed step-by-step with IHC, as previously described [12]. In addition, we prefer to perform CRP staining, alone or in association with SAA, particularly if SAA staining is not straightforward. To extend the discussion, the interpretation of GS and CRP staining to differentiate FNH from IHCA is not a major issue in our experience. The main issue remains the differential diagnosis
between FNH without typical map-like pattern of GS staining and unclassified HCA (particularly when they are remodeled) with some pathological features observed in FNH. In conclusion: • the final diagnosis of FNH or HCA must be made in light of the clinical, radiological, and morphological features, using appropriately IHC markers such as GS and CRP whose interpretation requires knowledge of underlying molecular pathways; • there is an urgent need to discuss the difficulties in IHC interpretation among experienced liver pathologists to assess the range of varieties in the immunostaining and to avoid confusion in the diagnosis of benign hepatocellular tumors.
Disclosure of interest The authors declare that they have no conflicts of interest concerning this article.
References [1] Joseph NM, Ferrell LD, Jain D, Torbenson MS, Wu TT, Yeh MM, et al. Diagnostic utility and limitations of glutamine synthetase and serum amyloid-associated protein immunohistochemistry in the distinction of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Mod Pathol 2013;27: 62—72. [2] Nault JC, Bioulac-Sage P, Zucman-Rossi J. Hepatocellular benign tumors-from molecular classification to personalized clinical care. Gastroenterology 2013;144:888—902. [3] Rebouissou S, Couchy G, Libbrecht L, Balabaud C, Imbeaud S, Auffray C, et al. The beta-catenin pathway is activated in focal nodular hyperplasia but not in cirrhotic FNH-like nodules. J Hepatol 2008;49:61—71. [4] Bioulac-Sage P, Laumonier H, Rullier A, Cubel G, Laurent C, Zucman-Rossi J, et al. Overexpression of glutamine synthetase in focal nodular hyperplasia: a novel easy diagnostic tool in surgical pathology. Liver Int 2009;29:459—65. [5] Bioulac-Sage P, Rebouissou S, Thomas C, Blanc JF, Saric J, Sa Cunha A, et al. Hepatocellular adenoma subtype classification using molecular markers and immunohistochemistry. Hepatology 2007;46:740—8. [6] Bioulac-Sage P, Sempoux C, Paradis V, Ferrell L, Scoazec JY, Leteutre E, et al. Characterization of liver nodules in patients with Budd-Chiari syndrome (BCS) and portal vein agenesis (PVA). Hepatology 2012;52:607 [Abstract]. [7] Fleming KE, Wanless IR. Glutamine synthetase expression in activated hepatocyte progenitor cells and loss of hepatocellular expression in congestion and cirrhosis. Liver Int 2013;33:525—34. [8] Pilati C, Nault JC, Latouzé E, Imbeaud S, Mallet M, Boulai A, et al. Integrative genomic profiling of hepatocellular adenoma identify mutational processes involved in malignant transformation. Hepatology 2013;58:55 [Abstract]. [9] Bioulac-Sage P, Cubel G, Balabaud C, Zucman-Rossi J. Revisiting the pathology of resected benign hepatocellular nodules using new immunohistochemical markers. Semin Liver Dis 2011;31:91—103. [10] Arnason T, Fleming KE, Wanless IR. Peritumoral hyperplasia of the liver: a response to portal vein invasion by hypervascular neoplasms. Histopathology 2013;62:458—64.
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
+Model CLINRE-541; No. of Pages 5
ARTICLE IN PRESS
Immunohistochemical pitfalls in the diagnosis of benign hepatocellular tumors [11] Bioulac-Sage P, Cubel G, Taouji S, Scoazec JY, Leteurtre E, Paradis V, et al. Immunohistochemical markers on needle biopsies are helpful for the diagnosis of focal nodular hyperplasia and hepatocellular adenoma subtypes. Am J Surg Pathol 2012;36:1691—9.
5
[12] Balabaud C, Al-Rabih WR, Chen PJ, Evason K, Ferrell L, Hernandez-Prera JC, et al. Focal nodular hyperplasia and hepatocellular adenoma around the world viewed through the scope of the immunopathological classification. Int J Hepatol 2013;2013:268625.
Please cite this article in press as: Bioulac-Sage P, et al. Immunohistochemical pitfalls in the diagnosis of focal nodular hyperplasia and inflammatory hepatocellular adenoma. Clin Res Hepatol Gastroenterol (2014), http://dx.doi.org/10.1016/j.clinre.2014.01.001
