Pancreatic epithelial neoplasms can be simply classified

Transcription

1 Utility of Immunohistochemistry in the Pancreatobiliary Tract Fan Lin, MD, PhD; Zongming Eric Chen, MD, PhD; Hanlin L. Wang, MD, PhD Context. Immunohistochemistry has become a useful ancillary study in the identification and classification of pancreatic neoplasms. The diagnostic accuracy has been significantly improved because of the continuous discoveries of tumor-associated biomarkers and the development of effective immunohistochemical panels. Objectives. To identify and classify pancreatic neoplasms by immunohistochemistry. Data Sources. Literature review and authors research data and personal practice experience were used. Conclusions. To better guide therapeutic decisions and predict the prognostic outcome, it is crucial to make an accurate diagnosis of a pancreatic neoplasm. Application of appropriate immunohistochemical panels enables pathologists to differentiate pancreaticobiliary adenocarcinomas from reactive conditions and to identify rare types of pancreatic neoplasms. Knowing the utilities and pitfalls of each tumor-associated biomarker is essential to avoiding a potential diagnostic error because an absolutely cancerspecific biomarker does not exist. This article reviews frequently used tumor-associated biomarkers, provides lists of effective immunohistochemical panels, and recommends a diagnostic algorithm as a standard approach to pancreatic neoplasms. (Arch Pathol Lab Med. 2015;139:24 38; doi: / arpa ra) Pancreatic epithelial neoplasms can be simply classified into solid and cystic neoplasms based on their gross appearances, as illustrated in Figure 1. Pancreatic ductal adenocarcinoma (DADC) and its variants, such as colloid carcinoma and medullary carcinoma, account for approximately 85% of neoplasms, followed by intraductal papillary mucinous neoplasms (IPMNs) at 3% to 5%, and pancreatic neuroendocrine tumors (P-NETs) at 3% to 4%. 1,2 Mucinous cystic neoplasm (MCNs), acinar cell carcinomas (ACCs), solid pseudopapillary tumors (SPT), and serous cystadenomas (SCA) are uncommon neoplasms of the pancreas. 1,2 Accurate diagnosis for each entity is crucial for making therapeutic decisions and predicting a prognostic outcome. Numerous immunohistochemical (IHC) markers to improve diagnostic accuracy have been described in the literature The selected immunomarkers for differentiating these neoplasms are summarized in Table 1. The applications and pitfalls of important markers will be reviewed throughout this article. PANCREATIC DADC AND ITS VARIANTS Distinguishing pancreatic DADC from nonneoplastic pancreatic tissues/diseases, including chronic pancreatitis Accepted for publication March 25, From the Department of Laboratory Medicine, Geisinger Medical Center, Danville, Pennsylvania (Drs Lin and Chen); and the Department of Pathology, University of California, Los Angeles (Dr Wang). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Fan Lin, MD, PhD, Department of Laboratory Medicine, MC 01-31, Geisinger Medical Center, 100 N Academy Ave, Danville, PA ( Flin1@geisinger.edu). and autoimmune pancreatitis, can be challenging, especially in small biopsies and fine-needle aspiration biopsy (FNAB) specimens. Numerous immunomarkers have been reported to be useful in solving that problem. Those markers include von Hippel-Lindau tumor suppressor (pvhl), placental S100 (S100P), mammary serine protease inhibitor (maspin), insulin-like growth factor II messenger RNA binding protein-3 (IMP3), mesothelin, prostate stem cell antigen (PSCA), annexin A8, fascin, claudin 4, claudin 18, p53, SMAD family member 4 (DPC4/SMAD4), carcinoembryonic antigen (CEA), cytokeratin (CK) 17, CK19, mucin (MUC) 1, MUC2, MUC5AC, cancer antigen 19-9 (CA19-9), and EPCAM More recently, 3 additional biomarkers (annexin A10, plectin 1, and aldo-keto reductase family 1, member B10 [AKR1B10]) have been reported as useful markers to differentiate pancreatic DADCs from benign/ reactive pancreatic ducts Specifically, annexin A10, a calcium and phospholipid-binding protein expressed in normal gastric mucosa, was found to be positive in 78% (57 of 73) of primary pancreatic DADCs and 83% (19 of 23) of metastatic pancreatic DADCs, 46% (154 of 337) of primary gastric ADCs and 47% (36 of 77) of metastatic gastric adenocarcinomas (ADCs), and only 2% of metastatic ADCs from other organs. 50 Plectin 1 was reported to be positive in 100% of primary and metastatic pancreatic DADCs and 60% of pancreatic intraepithelial neoplasia (PanIN) 3 cases, but it was negative in all chronic pancreatitis and normal pancreatic tissues. 51 AKR1B10 was found to be expressed in 70% (35 of 50) of pancreatic DADCs and most PanINs. 52 We do not have experience with these 3 newly described immunomarkers. Additional studies are needed to further validate the clinical utilities of those 3 markers. In our previous study, 48 we investigated the utility of 26 IHC markers (CAM 5.2, CK7, CK20, CK17, CK19, MUC1, 24 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

2 Figure 1. Summary of the immunohistochemical profile for each pancreatic neoplasm. Abbreviations: ACC, acinar cell carcinoma; ADC, adenocarcinoma; BCL, B cell/cll lymphoma; CA, carcinoma; CD, cluster of differentiation; CDX2, caudal type homeobox 2; CK, cytokeratin; CLL, chronic lymphocytic leukemia; DPC4, SMAD family member 4; ER, estrogen receptor; IMP3, insulin-like growth factor II messenger RNA binding protein-3; IPMN, intraductal papillary mucinous neoplasm; maspin, mammary serine protease inhibitor; MCN, mucinous cystic neoplasm; MSI, microsatellite instability; MUC, mucin; N, nuclear; NSE, neuron-specific enolase; PAX8, paired box 8; PB, pancreatoblastoma; PDX1, pancreatic and duodenal homeobox 1; P-NET, pancreatic neuroendocrine tumor; pvhl, von Hippel-Lindau tumor suppressor; PR, progesterone receptor; S100P, placental S100; SCA, serous cystadenoma; SPT, solid pseudopapillary tumor. MUC2, MUC4, MUC5AC, MUC6, p53, DPC4/SMAD4, caudal type homeobox 2 [CDX2], pvhl, S100P, IMP3, maspin, mesothelin, claudin 4, claudin 18, annexin A8, fascin, PSCA, EPCAM, CEA, and CA19-9) in 60 cases of pancreatic DADC on tissue microarray and routine tissue sections. We also performed immunohistochemical stains for maspin, S100P, IMP3, and pvhl on cell blocks of 67 pancreatic FNAB specimens, including 44 cases of pancreatic DADC and 27 benign/reactive cases. The results demonstrated that (1) more than 90% of pancreatic DADCs were positive for maspin, S100P, and IMP3; (2) nearly all pancreatic DADCs were negative for pvhl, whereas Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 25

3 Table 1. Summary of Useful Markers in Selected Pancreatic Epithelial Neoplasms Antibodies DADC ACC P-NET SPT PB CK7 þ or þ þ or þ or CK19 þ or þ or þ þ or Mesothelin þ þ or S100P þ þ or Maspin þ þ or Glypican 3 þ or ND b-catenin M M or N M N and M N and C, or Mþ E-cadherin þ þ þ or þ Chromogranin þ or þ CD10 þ þ or þ IMP3 þ þ or or þ Trypsin þ þ than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; ACC, acinar cell carcinoma; C, cytoplasmic staining; CD10, cluster of differentiation 10; CK, cytokeratin; DADC, ductal adenocarcinoma; IMP3, insulin-like growth factor II messenger RNA binding protein-3; M, membranous staining; maspin, mammary serine protease inhibitor; N, nuclear staining; ND, no data; PB, pancreatoblastoma; P-NET, pancreatic neuroendocrine tumor; S100P, placental S100; SPT, solid pseudopapillary tumor. Antibody order based on the application of the group of antibodies to a specific entity. nonneoplastic ducts and acini were positive for pvhl in all cases; (3) normal/reactive pancreatic ducts were frequently positive for CK7, CK19, MUC1, MUC6, CA19-9, EPCAM, PSCA, mesothelin, annexin A8, claudin 4, and claudin 18; (4) normal pancreatic ducts were usually negative for IMP3, maspin, S100P, CK17, MUC2, MUC4, and MUC5AC; (5) 60% of pancreatic DADCs were negative for DPC4/SMAD4, and 50% of cases were strongly positive for p53; (6) loss of pvhl expression and S100P overexpression were observed in all PanINs, regardless of grade; and (7) strong background staining was frequently seen with fascin, PSCA, and annexin A8. 48 Markers like tumor-associated glycoprotein 72 (TAG 72/B72.3), EPCAM, and epithelial cell adhesion molecule (Ber-EP4) were usually positive in pancreatic DADCs, but they were frequently positive or weakly positive in normal or reactive ducts as well. In contrast, monoclonal CEA (mcea) was strongly positive in most pancreatic DADCs, and the adjacent stroma also showed positivity because of the leaking of mcea into the adjacent tissue. Normal/reactive ducts tended to be negative or only weakly positive for mcea. 48 Based on that study and review of the literature, 48 we concluded that pvhl, maspin, S100P, IMP3, CK17, MUC5AC, and DPC4/SMAD4 were the best diagnostic panel of immunomarkers for confirming the diagnosis of pancreatic DADCs in both surgical and FNAB specimens. In an autoimmune pancreatitis, the infiltrating plasma cells were predominately positive for immunoglobulin G4 (IgG4). An immunostain for IgG4 may be helpful in diagnosing a difficult case The presence of abundant IgG4-positive plasma cells does not, however, preclude the diagnosis of pancreatic DADC because, in a small subset of pancreatic DADC cases, the cancer-adjacent tissue may show features of autoimmune pancreatitis. The useful immunomarkers for the distinction between a pancreatic DADC and reactive ducts are summarized in Table 2. Additional discussion of pvhl, S100P, maspin, IMP3, MUC5AC, and CK17 will follow in the section below. An example of pancreatic DADC showing expression of pvhl, maspin, pvhl/maspin double stain, S100P, IMP3, MU- Table 2. Pancreatic Ductal Adenocarcinoma (DADC) Versus Chronic Pancreatitis Antibodies Pancreatic DADC Pancreatitis pvhl þ Maspin þ S100P þ or C þ only IMP3 þ CK17 þ or Usually MUC5AC þ or p53 þ or or very weakly þ mcea þ Usually or focally þ DPC4/SMAD4 Loss (60%) þ Mesothelin þ or weakly þ than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; C, cytoplasmic staining; CK17, cytokeratin 17; DPC4, SMAD family member 4; IMP3, insulin-like growth factor II messenger RNA binding protein-3; mcea, monoclonal carcinoembryonic antigen; MUC5AC, mucin 5AC; pvhl, von Hippel-Lindau tumor suppressor; maspin, mammary serine protease inhibitor; S100P, placental S100. Antibody order based on the application of the group of antibodies to a specific entity. C5AC, and DPC4/SMAD4 is illustrated in Figure 2, A through H. VHL (von Hippel-Lindau) is a tumor suppressor gene. Inactivation of the VHL gene on band 3p25-26 by mutation, deletion, or hypermethylation is a frequent event in both hereditary and sporadic clear cell renal cell carcinomas (RCCs). 56,57 Genetic alteration of the VHL gene appears to be a crucial step in the initiation and progression of clear cell RCC. Our previous study demonstrated pvhl was expressed in most renal cell neoplasms, including clear cell RCCs, papillary RCCs, chromophobe RCCs, and oncocytomas 47 and in more than 90% of metastatic RCCs. In addition, 90% of clear cell carcinomas of the ovary and uterus were positive for pvhl. 47 In contrast, many tumors from various organs, including carcinomas of the lung, pancreas, gastrointestinal tract, adrenal gland, prostate, and bladder, were Figure 2. A through H, A representative case of pancreatic ductal adenocarcinoma on hematoxylin-eosin stained section (A), and loss of expression of von Hippel-Lindau tumor suppressor (pvhl) (B), positive for mammary serine protease inhibitor (maspin) (C), positive for maspin and negative for pvhl (double staining, brown-maspin and red-pvhl; D), positive for placental S100 (E), insulin-like growth factor II messenger RNA binding protein-3 (F), mucin 5AC (G), and loss of SMAD family member 4 (H) (original magnifications 3200 [A, D, and E], 3400 [B and C and F through H]). 26 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

4 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 27

5 negative for pvhl. 47 Interestingly, normal/reactive pancreatic ducts and acinar cells were positive for pvhl. Expression of pvhl has been also tested in ADCs of the gallbladder, extrahepatic bile ducts and intrahepatic bile ducts The results demonstrated (1) only 6% of gallbladder ADCs were positive for pvhl, 60 (2) 7.5% of extrahepatic bile duct ADCs were positive for pvhl, 58 and (3) in contrast, 71% of intrahepatic cholangiocarcinomas (ICCs) were positive for pvhl. 61 Therefore, pvhl may have a role in the distinction of pancreatic DADC, gallbladder adenocarcinoma (ADC), and extrahepatic ADC from ICC. S100P belongs to the family of S100 calcium binding proteins. It is a 95 amino acid protein first purified from the human placenta. The level of S100P expression has been found to increase during the progression from PanIN to invasive ADC ,46 Expression of S100P has been reported in more than 90% of pancreatic DADCs, gallbladder ADCs, and extrahepatic ADCs. 46,48,58 60 In contrast, only 27%% of ICCs were positive for S100P. 61 Similar observations have also been reported by Aishima et al, 62 who found positive nuclear S100P staining in 8 of 69 (12%) of the peripheraltype ICCs, in contrast to 28 of 41 (68%) of the perihilar cholangiocarcinomas. Tsai et al 63 reported a higher frequency of S100P expression in peripheral ICCs (53%), but in most of their positive cases, the staining was heterogeneous with mixed nuclear, cytoplasmic and extracellular patterns. It is unclear how many of those cases showed nuclear S100P expression. In addition, one-half of their positive cases exhibited only focal staining. 63 Overexpression of S100P was also frequently observed in IPMNs and MCNs but not in ACCs, P-NETs, and SCAs. 64,65 Reactive ducts may show cytoplasmic staining for S100P. As a rule, only nuclear positivity or both nuclear and cytoplasmic positivity was regarded as positive staining for S100P. Maspin, also known as serpin B5, is a tumor suppressor gene expressed in some human epithelial cells. Maspin overexpression can be seen in carcinomas from various organs including most ADCs of the pancreas and gallbladder and in ICCs;* however, maspin overexpression was also reported in nontumorous gallbladder epithelium in patients with cholelithiasis and intestinal metaplasia. 66,67 Therefore, maspin alone cannot be used as a reliable diagnostic marker to differentiate benign from malignant gallbladder glandular lesions. Normal and reactive pancreatic ducts are usually negative for maspin. 48 IMP3 is an oncofetal RNA-binding protein, also known as K homology domain containing protein overexpressed in cancer, IGF2BP3, and L523S. Immunohistochemical expression of IMP3 has been reported in various cancers, including ADCs of the pancreas, endocervix, and neuroendocrine carcinomas of the lung ,48,58 60 Expression of IMP3 was demonstrated in 37 of 38 (97%), of 112 (88.4%), 22 and 54 of 60 (90%) of the ADCs of the pancreas. 48 Similar findings of IMP3 expression in FNABs of the pancreas were reported. 23 Expression of IMP3 was reported in 81%, 64%, and 90% of ADCs of the gallbladder, extrahepatic bile ducts, and intrahepatic bile ducts, respectively MUC5AC is a high molecular-weight glycoprotein of the mucin family, which has been shown to serve an important role in pancreatic tumorigenesis. 68,69 Overexpression of MUC5AC has been reported in 60% to 100% of pancreatic * References 24, 25, 44, 48, 60, 61, 66, 67. DADCs in several studies. 10,48,70,71 Similarly, a high frequency of MUC5AC expression has been reported in cholangiocarcinoma 10,72 75 and in gallbladder ADC. 60,72,76 In the studies in which ICCs and extrahepatic cholangiocarcinomas were separately evaluated, it was apparent that extrahepatic cholangiocarcinomas more frequently expressed MUC5AC than ICCs. 70,72 Another interesting observation was that peripheral-type ICCs showed MU- C5AC expression less commonly than hilar cholangiocarcinomas. In the study by Guedj et al, of 52 hilar cholangiocarcinomas (62%) showed positivity for MUC5AC, in contrast to 13 of 59 (22%) peripheral ICCs. Similarly, Aishima et al 78 reported positive MUC5AC staining in 69% to 72% of hilar tumors but only 25% to 27% of peripheral ICCs (2 different anti-muc5ac antibodies were used in that study). The study from our group showed a similar finding of a higher frequency of MUC5AC expression in pancreatic DADCs (67%) than in ICCs (12%). 61 CK17 is a low molecular-weight keratin that is normally expressed in myoepithelial and basal cells and subsets of hair shaft epithelia. 10 A few studies have demonstrated it to be of value in the distinction between pancreaticobiliary and nonpancreaticobiliary ADCs. 7,10,79,80 Our data showed CK17 expression in 60% of pancreatic DADCs but only 12% of ICCs. 48,61 Our findings of CK17 expression in ICCs seem to be different from those reported by Chu et al, 10 who detected CK17 expression in 17 of 24 ICCs (71%). However, it is unclear whether hilar or perihilar cholangiocarcinomas were included in that study. To support our findings, Sarbia et al 79 showed positive CK17 staining in 10 of 17 ADCs (59%) of the extrahepatic bile ducts, a frequency similar to ours. 48,61 Two variants of pancreatic DADC should be mentioned here. One is colloid carcinoma, and the other is medullary carcinoma. Colloid carcinoma is usually positive for CDX2 and MUC2, but negative for MUC1, which is the opposite to the IHC profile of a conventional pancreatic DADC. 1,2,81 Identification of colloid carcinoma is important because it carries a much better prognosis than does a conventional pancreatic DADC. Medullary carcinoma of the pancreas is rare and frequently demonstrates loss of DNA mismatch repair proteins, especially MutL homolog 1 (MLH1) and postmeiotic segregation increased 2 (PMS2). 1,2,82 84 Our unpublished data demonstrated that both colloid carcinoma and medullary carcinoma show a similar IHC profile for the expression of pvhl, S100P, maspin, and IMP3 when compared with pancreatic DADCs. DIFFERENTIAL DIAGNOSIS OF P-NET, ACC, SPT, AND PANCREATOBLASTOMA In addition to the distinction between pancreatic DADCs and chronic pancreatitis/reactive ducts, another frequently encountered diagnostic problem is to differentiate P-NET from its mimics, such as ACC, SPT and rarely pancreatoblastoma (PB). The frequently used immunomarkers in making that differential diagnosis are summarized in Table 3. The applications and pitfalls for some of the markers are discussed in detail below. Acinar Cell Carcinoma Acinar cell carcinoma of the pancreas is a very rare, nonductal carcinoma, which accounts for about 1% of pancreatic neoplasms in adults and approximately 15% in 28 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

6 Table 3. Summary of Useful Markers for Acinar Cell Carcinoma (ACC), Pancreatic Neuroendocrine Tumor (P-NET), Solid Pseudopapillary Tumor (SPT), and Pancreatoblastoma (PB) Antibodies ACC P-NET SPT PB b-catenin M or rarely N þ C M N þ C N þ CorMþ E-cadherin þ þ or þ Chromogranin þ or þ CD10 þ þ or þ Trypsin þ þ Glypican 3 þ or ND BCL10 þ ND than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; BCL10, B cell CLL/lymphoma 10; C, cytoplasmic staining; CD10, cluster of differentiation 10; CLL, chronic lymphocytic leukemia; M, membranous staining; N, nuclear staining; ND, no data. Antibody order based on the application of the group of antibodies to a specific entity. pediatric patients. 1,2 The differential diagnosis from pancreatic DADC is relatively straightforward based on its morphologic features and the IHC profile. The diagnostic challenge is to differentiate ACC from a neuroendocrine neoplasm and an SPT. Perhaps the most difficult diagnostic problem is how to distinguish ACC from PB; sometimes, it is nearly impossible because of the significant overlapping of both morphologic and IHC features, especially in a smallcore biopsy or an FNAB specimen. The classic IHC profile of an ACC is outlined in Table 4. Our experience and that of others showed CK7 is frequently negative or only focally positive in ACCs, which provides a useful clue for differentiating an ACC from a pancreatic DADC, which is nearly always diffusely positive for CK7. A histochemical stain of periodic acid Schiff diastase is usually positive in ACC. Trypsin is usually positive but may give strong background staining. Up to 25% of ACCs may show both nuclear and cytoplasmic positivity for b-catenin, 85,86 which can create a problem in the distinction of an ACC from an SPT or a PB. Chromogranin was usually negative or showed only scattered positivity in endocrine cells. When more than 25% of tumor cells are positive for endocrine markers, the tumor is regarded as a mixed acinar and endocrine carcinoma. Based on our experience, other markers, including mesothelin, pvhl, S100P, and maspin, are usually negative in ACCs. Several recently described immunomarkers have been reported to be useful in confirming a diagnosis of ACC. The antibody against COOH-terminal BCL10 was shown to be positive in normal acinar cells and in 82% (14 of 17) of ACCs by Hosoda et al 87 and 85% (40 of 47) of ACCs by La Rosa et al. 88 BCL10 expression was not seen in P-NETs, pancreatic DADCs, SPTs, SCAs, MCNs, or IPMNs. 87 Two of 4 adenosquamous carcinomas (50%) were positive for BCL Glypican-3, an established immunomarker for hepatocellular carcinoma and yolk sac tumor, was positive in 58% (7 of 12) of ACCs but negative in pancreatic DADCs and P-NETs. 89 Carboxyl ester lipase has been found to be another sensitive marker positive in 91% (29 of 32) of ACCs. 88 Pancreatic Neuroendocrine Tumor Pancreatic neuroendocrine tumors are uncommon, accounting for approximately 3% of all pancreatic neoplasms. 1,2 Diagnosis of P-NET can be challenging because of the limited experience a pathologist may have with it and the variety of morphologic variants, including spindle cell, oncocytic cell, rhabdoid cell, pleomorphic cell, lipid-rich, and clear cell. The current World Health Organization classification of P-NETs was based on counting mitoses or mindbomb homolog 1 (MIB1/Ki-67) proliferative index. 1 Pancreatic neuroendocrine tumors are divided into (1) P- NET grade 1 (0 1 mitosis/10 high-power fields or,2% MIB1 index), (2) P-NET grade 2 (2 20 mitoses/10 highpower fields or 3% 20% MIB1 index), and (3) pancreatic neuroendocrine carcinoma (large cell neuroendocrine carcinoma or small cell carcinoma;.20 mitoses/10 high-power fields or.20% MIB1 index). Neuroendocrine features can be confirmed by immunoreactivity to synaptophysin, chromogranin, and CD56. Our experience and that of others showed nearly 100% of P-NETs were positive for both synaptophysin and chromogranin. In contrast, our experience showed CD56 was expressed in only 44% (7 of 16) of P-NETs. Pancreatic neuroendocrine tumors were frequently negative for both CK7 and CK20. Other investigators 90 have reported approximately 10% of P-NETs were positive for CK7 or CK20. CK19 positivity in P-NETs may be associated with a more-aggressive clinical behavior. 91,92 Several relatively P-NET-specific immunomarkers have been reported, including progesterone receptor (PR), paired box gene 8 (PAX8), pancreatic and duodenal homeobox 1 (PDX1) and islet-1. 90, Progesterone receptor has been reported in 55% to 58.5% of P-NETs, 93 which was similar to our finding of PR immunoreactivity in 56% (9 of 16) of P- NETs. Progesterone receptor immunoreactivity showed a significant correlation with the absence of metastasis and Table 4. Immunomarkers for Acinar Cell Carcinoma Antibodies Literature Trypsin þ BCL10 þ Carboxyl ester lipase þ Glypican-3 þ or b-catenin M or rarely N þ C pvhl Mesothelin S100P CK7 or focal þ Chromogranin or rare cell positive Vimentin þ or than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; BCL10, B cell CLL/lymphoma 10; C, cytoplasmic staining; CK7, cytokeratin 7; CLL, chronic lymphocytic leukemia; M, membranous staining; N, nuclear staining; pvhl, von Hippel-Lindau tumor suppressor; S100P, placental S100. Antibody order based on the application of the group of antibodies to a specific entity. Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 29

7 Figure 3. A through D, A pancreatic neuroendocrine neoplasm on hematoxylin-eosin stained section (A), and positive for progesterone receptor (B), paired box gene 8 (C), and islet-1 (D) (original magnification 3400 [A through D]). lack of invasion into the adjacent organs and large vessels. 93,94 Progesterone receptor expression in other neuroendocrine tumors (NETs) was very rare, with the exception of rare cases reported in lung carcinoid and duodenal NETs. 93,94 Estrogen receptor expression in P-NETs was generally absent. PAX8 is a member of the paired box (PAX) family of transcription factors involved in the development of thyroid follicular cells and the expression of thyroid-specific genes and, together with PAX2, is involved in the regulation of the organogenesis of the kidney and the Müllerian system. PAX8 has been demonstrated to be a highly sensitive and relatively specific marker for thyroid follicular cell tumors, RCCs, ovarian carcinomas, endometrial ADCs, and thymic tumors. Immunoreactivity to PAX8 was seen in 50% to 74% of P-NETs One study 95 showed PAX8 expression was seen in a significant percentage of duodenal and rectal NETs and 27% of SPTs of the pancreas. The immunoreactivity to PAX8 has been proven a consistent staining artifact by the cross-reaction between the anti-pax8 antibody and PAX6 antigen. 98 The human insulin gene enhancer-binding protein islet-1 is a transcription factor involved in the differentiation of pancreatic endocrine cells. Graham et al 99 showed islet-1 expression in 90% of P-NETs, 89% of duodenal NETs, 100% of rectal NETS, 38% of colonic NETs, and a small percentage of NETs from other organs. Koo and coworkers 100 reported that islet-1 was positive in 27 of 33 (82%) and in 19 of 28 (68%) of primary and metastatic P-NETs, respectively. Additionally, Agaimy et al 101 demonstrated that islet-1 expression was seen in nearly all extrapancreatic, poorly differentiated neuroendocrine carcinomas, such as pulmonary small cell carcinoma, Merkel cell carcinoma, medullary thyroid carcinoma, and adrenal neuroblastoma, whereas pancreatic poorly differentiated neuroendocrine carcinomas usually lacked islet-1 expression. These findings argue against the diagnostic value of using islet-1 to identify a pancreatic origin when dealing with a poorly differentiated neuroendocrine carcinoma. PDX1 is a Hox-type transcription factor that regulates both exocrine and endocrine pancreatic differentiation and maintains b-cell function. Chan et al 90 reported expression of PDX1 in 72% (18 of 25) of P-NETs, 10% (3 of 29) of pulmonary NETs, and 4% (1 of 26) of gastrointestinal NETs. The expression of PDX1 was also seen in 5 of 5 metastatic P- NETs (100%) in the liver and 2 of 2 metastatic duodenal NETs (100%) in the liver. 90 They suggested PDX1 was a highly sensitive and specific marker for P-NETs. 90 Another 30 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

8 Table 5. Markers for Pancreatic Neuroendocrine Tumor a Antibodies Literature GML Data, No. (%) Synaptophysin, n ¼ 16 þ 16 (100) Chromogranin, n ¼ 16 þ 16 (100) CD56, n ¼ 16 þ 7 (44) PR, n ¼ 16 þ or 9 (56) PAX8, n ¼ 32 þ or 15 (47) PDX1 þ ND Islet-1 þ ND b-catenin, n ¼ 16 0 (0) CAM 5.2, n ¼ 16 þ 16 (100) CK7, n ¼ 16 or þ 0 (0) CK20, n ¼ 16 1 (6) Vimentin, n ¼ 16 6 (38) CDX2, n ¼ 16 V 1 (6) Insulin, n ¼ 16 V 2 (12) CK19, n ¼ 16 þ or 4 (25) Abbreviations:þ, usually more than 75% of cases are positive;,lessthan 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; CD56, cluster of differentiation 56; CDX2, caudal type homeobox 2; CK, cytokeratin; GML, Geisinger Medical Laboratories; ND, no data; PAX8, paired box gene 8; PDX1, pancreatic and duodenal homeobox 1; PR, progesterone receptor; V, variable. Antibody order based on the application of the group of antibodies to a specific entity. a Reprinted from 3 Lin F, Prichard JW, Liu H, Wilkerson M, Schuerch C, eds. Handbook of Practical Immunohistochemistry: Frequently Asked Questions. New York, NY: Springer; 2011, with permission from Springer Science þ Business Media. Approximately 50% of SPTs expressed CD117 (c-kit); however, CD117 expression was not associated with underlying mutations of the KIT gene. 111 A representative SPT case is shown in Figure 4, A through F. Pancreatoblastoma Pancreatoblastoma is a very rare pancreatic neoplasm in adults, accounting for less than 1% of all pancreatic tumors. However, it is the most common neoplasm in childhood, accounting for about 25% of all pancreatic neoplasms seen before age 10 years. 1,112 Most PBs consist of both acinar and squamous components; some may also contain endocrine and ductal components. 1,112 The immunostaining results are largely dependent on the components in the tumor. Nuclear staining of b-catenin has been reported in a significant percentage of cases, which is similar to the findings in SPTs and ACCs. 85,113 The squamous component usually lacks the typical squamous immunophenotype, that is, being positive for CK5/6, CK14, and CK17. Instead, that component is usually positive for epithelial membrane antigen, CK8, CK18, and CK19 but negative for CK7. Loss of DPC4 expression has been reported in up to 22% of cases, and p53 expression was usually not seen. 85 a-fetoprotein may be positive in some cases, which is in keeping with the primitive nature of this neoplasm. The useful immunomarkers for PB are summarized in Table 7. study demonstrated that PDX1 expression was frequently associated with the type of hormone secreted by a NET, regardless of whether it was of pancreatic or duodenal origin. 102 Most insulin- and gastrin-secreted NETs were positive for PDX1, whereas glucagon-, somatostatin-, or serotonin-positive NETs usually lacked PDX1 expression. 102 Based on those data, PDX1 appears to be a specific marker for identifying insulin-positive or gastrin-positive NETs that originate from either the pancreas or duodenum. An example of a pancreatic neuroendocrine neoplasm with positive staining for PR, PAX8, and islet-1 is shown in Figure 3, A through D. The useful immunomarkers for P- NETs, including Geisinger Medical Center (Danville, Pennsylvania) data, are summarized in Table 5. A panel of immunomarkers that includes CK7, CK20, thyroid transcription factor 1 (TTF1), CDX2, cadherin-17 (CDH17), special AT-rich sequence-binding protein 2 (SATB2), PR, PDX1, and PAX8 is useful in differentiating a P-NET from a NET of other organs, which was described in the Immunohistochemistry in Undifferentiated Neoplasm/Tumor of Uncertain Origin article by Lin and Liu. 126 SEROUS CYSTADENOMA Serous cystadenoma is an uncommon neoplasm of the pancreas, accounting for 1% to 2% of all pancreatic neoplasms. 1,2 It more often occurs in a middle-aged or older woman in the body or tail of the pancreas. 1,2 A malignant counterpart, serous cystadenocarcinoma, has been reported, which has a similar IHC profile as a benign Table 6. Antibodies Markers for Solid-Pseudopapillary Neoplasm of the Pancreas Literature b-catenin N þ C positive E-cadherin Chromogranin CD10 þ Vimentin þ AE1/AE3 or focally þ CK7 or focally þ Trypsin a-1 antitrypsin þ CD56 þ Solid Pseudopapillary Tumor of the Pancreas NSE þ or Synaptophysin or þ Solid pseudopapillary tumor is rare and only accounts for 1% to 2% of pancreatic exocrine neoplasms and 5% of cystic Claudin 5 Claudin 7 M þ or focally C þ neoplasms. 1,2 It is defined as a low-grade malignant neoplasm Progesterone receptors (PR) þ or PAX8 or þ that occurs predominately in adolescent girls or young Estrogen receptors (ER) women. Numerous immunomarkers for SPT have been CD99 þ (C dot stain) reported in the literature, as listed in Table b-catenin, E-cadherin, CD10, vimentin, and chromogranin are the most Galectin 3 CD117 (c-kit) þ þ or effective IHC panel for confirming the diagnosis of SPT More than 90% of SPTs showed both nuclear and cytoplasmic than 5% of cases are positive; þ or, usually more than 50% of cases staining for b-catenin and loss of E-cadherin expression. 104,105 are positive; or þ, less than 50% of cases are positive; C, cytoplasmic staining; CD, cluster of differentiation; CK, cytokeratin; M, membranous staining; N, nuclear staining; NSE, neuron-specific enolase; PAX8, Chromogranin was consistently negative in SPT, which was a useful feature in differentiating it from a P-NET. Both CD10 paired box gene 8. Antibody order based on the application of the and vimentin were positive in nearly all reported SPTs. group of antibodies to a specific entity. Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 31

9 Figure 4. A through E, A representative case of solid-pseudopapillary tumor on hematoxylin-eosin stained section (A), and positive for b-catenin (B), cluster of differentiation 10 (C), vimentin (D), progesterone receptor (E), and loss of E-cadherin expression (F) (original magnification 3400 [A through E]). SCA. Periodic acid Schiff for glycogen is usually positive, and mucicarmine for mucin is usually negative. Both pvhl andmuc6tendtoshowdiffuseandstrongcytoplasmic and membranous staining 64 ; in contrast, in our experience both neuron-specific enolase (NSE) and inhibin-a more frequently show focal and weak staining. The reports of the expression of neuroendocrine markers, such as CD56, synaptophysin, and chromogranin A, were inconsistent. 114,115 Our experience and studies by others showed that SCA was usually negative for those 3 markers. 32 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

10 Table 7. Markers for Pancreatoblastoma Antibodies Acinar Endocrine Ductal CK7 þ þ CK19 þ þ CAM 5.2 þ þ þ Trypsin þ NSE þ Synaptophysin þ Chromogranin þ CEA þ TAG 72 (B72.3) þ than 5% of cases are positive; CEA, carcinoembryonic antigen; CK, cytokeratin; NSE, neuron-specific enolase; TAG72 (B72.3), tumorassociated glycoprotein 72. Antibody order based on the application of the group of antibodies to a specific entity. However, a recent study 115 of 12 cases of SCA demonstrated that CD56, synaptophysin, and chromogranin A were expressed in 75%, 92%, and 0% of the cases, respectively. Those data suggested that only chromogranin A was a useful marker in distinguishing SCA from P-NET. Many cases may be positive for EPCAM and CA19-9, which are also positive in a high percentage of pancreatic mucin producing neoplasms and DADCs. The useful immunomarkers from the literature and from our data are summarized in Table 8. An example of a solid variant of SCA with diffuse and strong positivity for pvhl, MUC6, and inhibin-a isshowninfigure5,athroughd. MUCINOUS CYSTIC NEOPLASM Mucinous cystic neoplasms of the pancreas are relatively uncommon neoplasms, accounting for approximately 8% of surgically resected pancreatic cystic lesions. 1,2 They occur almost exclusively in women, with a female to male ratio of 20:1. 1,2 Mucinous cystic neoplasms consist of 2 components, an epithelial lining, and a subepithelial ovarian-type stroma. The ovarian-type stroma is the key feature to differentiate MCN from IPMN, 1,2 which is usually positive for PR, ER, smooth muscle actin, CD10, and inhibin-a. 116 Expression of different types of mucin, such as MUC5AC, is not very useful in differentiating an MCN from an IPMN. MUC2 is frequently expressed in goblet cells of MCN. MUC1 tends to be negative in a noninvasive MCN. DPC4/SMAD4 expression is retained. Mucinous cystic neoplasms with an invasive component may show a loss of DPC4/SMAD4 expression and express MUC1. 1 Loss of expression of pvhl and overexpression of S100P were the frequent findings in MCNs, regardless of the grade of epithelial dysplasia. 64 The review of the literature and our Geisinger Medical Center data are summarized in Table 9. INTRADUCTAL PAPILLARY MUCINOUS NEOPLASM Intraductal papillary mucinous neoplasms are relatively common cystic neoplasms of the pancreas, accounting for approximately 20% of all cystic pancreatic lesions. 1,2 Intestinal-type IPMN is usually positive for MUC2, CDX2, and CK20. Gastric foveolar-type IPMN is usually positive for MUC5AC and negative for both MUC1 and MUC2. Pancreatobiliary-type IPMN is usually positive for MUC1 and MUC5AC and negative for MUC2 and CDX2. Oncocytic type IPMN is usually positive for MUC5AC and MUC6. Table 8. Markers for Serous Cystadenoma a Antibodies Literature GML Data, No. (%) pvhl, n ¼ 13 þ 13 (100) MUC6, n ¼ 13 þ or 12 (92) Inhibin-a, n¼ 13 þ 12 (92) NSE, n ¼ 13 þ 7 (54) CK7, n ¼ 13 þ 13 (100) CK20, n ¼ 13 0 (0) S100P, n ¼ 13 0 (0) Synaptophysin, n ¼ 13 or þ 0 (0) Chromogranin, n ¼ 13 0 (0) TAG 72 (B72.3), n ¼ 13 0 (0) CEA, n ¼ 13 0 (0) CA19-9, n ¼ 13 or þ 4 (31) EPCAM, n ¼ 13 9 (69) Mucicarmine, n ¼ 13 0 (0) than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; CA19-9, cancer antigen 19.9; CEA, carcinoembryonic antigen; CK, cytokeratin; GML, Geisinger Medical Laboratories; MUC6, mucin 6; NSE, neuron-specific enolase; pvhl, von Hippel-Lindau tumor suppressor; S100P, placental S100; TAG72 (B72.3), tumor-associated glycoprotein 72. Antibody order based on the application of the group of antibodies to a specific entity. a Reprinted from 3 Lin F, Prichard JW, Liu H, Wilkerson M, Schuerch C, eds. Handbook of Practical Immunohistochemistry: Frequently Asked Questions. New York, NY: Springer; 2011, with permission from Springer Science þ Business Media. Studies showed S100P expression and loss of expression of pvhl in all types of IPMN, regardless of the grade of epithelial dysplasia. 64,65 Expression of DPC4/SMAD4 was present in all tested cases in our study. Expression of MUC1, loss of expression of DPC4/SMAD4, and overexpression of p53 in IPMNs were frequently associated with invasive Table 9. Markers for Mucinous Cystic Neoplasm a Antibodies Literature GML Data, No. (%) CK7, n ¼ 12 þ 12 (100) S100P, n ¼ 12 þ 8 (67) pvhl, n ¼ 12 4 (33) CD10, n ¼ 12 þ 4 (33) Estrogen receptor, n ¼ 12 þ 3 (25) Inhibin-a, n¼ 12 þ or 8 (67) Progesterone receptor, n ¼ 12 þ or 6 (50) CK20, n ¼ 12 or þ 4 (33) CEA, n ¼ 12 þ 12 (100) CA19-9, n ¼ 12 þ 11 (92) CDX2, n ¼ 12 3 (25) MUC1, n ¼ 12 2 (17) MUC2, n ¼ 12 or þ 0 (0) MUC5AC, n ¼ 12 þ 8 (67) MUC6, n ¼ 12 6 (50) DPC4/SMAD4, n ¼ 12 þ 12 (100) than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; CA19-9, cancer antigen; CD, cluster of differentiation; CDX2, caudal type homeobox 2; CEA, carcinoembryonic antigen; CK, cytokeratin 19-9; DPC4/SMAD4, SMAD family member 4; GML, Geisinger Medical Laboratories; MUC, mucin; pvhl, von Hippel-Lindau tumor suppressor; S100P, placental S100. Antibody order based on the application of the group of antibodies to a specific entity. a Reprinted from 3 Lin F, Prichard JW, Liu H, Wilkerson M, Schuerch C, eds. Handbook of Practical Immunohistochemistry: Frequently Asked Questions. New York, NY: Springer; 2011, with permission from Springer Science þ Business Media. Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 33

11 Figure 5. A through D, A representative case of serous cystadenoma on hematoxylin-eosin stained section (A), and positive for von Hippel-Lindau tumor suppressor (B), mucin 6 (C), and inhibin-a (D) (original magnification 3200 [A through D]). carcinoma. 117 MUC6 tended to be expressed in the basal layer of epithelial cells; the papillary structures projecting into the cystic space were frequently negative for MUC6. Maspin expression increased with increasing grade of dysplasia in noninvasive lesions from intraductal papillary mucinous adenomas to intraductal papillary mucinous borderline neoplasm to noninvasive intraductal papillary mucinous carcinoma but decreased significantly in invasive intraductal papillary mucinous carcinomas. 118 The literature and our Geisinger Medical Center data are summarized in Table 10. COMMON METASTASES IN THE PANCREAS Metastatic neoplasms of the pancreas are rare, accounting for approximately 2% of all pancreatic malignancies. 1,2 The most common malignancies metastasizing to the pancreas are RCC, melanoma, non small cell carcinoma of the lung, colorectal ADC, and breast carcinoma. 119 The common metastases to the pancreas and useful immunomarkers for identifying them are summarized in Table 11. For a detailed discussion of how to work up a tumor of uncertain origin, please refer to the review article Immunohistochemistry in Undifferentiated Neoplasm/Tumor of Uncertain Origin by Lin and Liu. 126 Clear cell RCC is the most common metastasis; it is usually positive for PAX8, pvhl, renal cell carcinoma marker, CD10, and vimentin and is negative for CK7 and CK20. Pulmonary ADC is another frequent metastasis. Approximately 75% of cases are positive for TTF1 and napsin A. Both TTF1 and napsin A are negative in pancreatic DADCs; however, mucinous ADCs of the lung are frequently positive for CDX2 and CK20 and negative for TTF1 and napsin A. S100 is a highly sensitive (98%), but not specific, marker for screening of melanoma. If a desmoplastic melanoma is suspected, sex-determining region Y box 10 (SOX10) should also be included. Other markers, including melanomaassociated antigen recognized by T cells 1 (MART1), human melanoma black 45 (HMB45) and microphthalmia-associated transcription factor (MiTF), are helpful. Caution should be taken if the sample is fixed in alcohol because the S100 antigen is not preserved well after alcohol fixation; therefore, a false-negative result may occur. Some metastatic small cell carcinomas of the lung can be negative for both synaptophysin and chromogranin and even for broad-spectrum cytokeratin, but they are very infrequently negative for CD56. The MIB-1 (Ki-67) proliferative index tends to be very high (.50%). It would be 34 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al

12 Table 10. Markers for Intraductal Papillary Mucinous Neoplasm a Antibodies Literature GML Data, No. (%) CK7, n ¼ 18 þ 18 (100) S100P, n ¼ 18 þ 18 (100) pvhl, n ¼ 18 0 (0) CK19, n ¼ 16 þ 12 (75) CK20, n ¼ 16 or þ 10 (62) CDX2, n ¼ 16 þ or 6 (38) CEA, n ¼ 18 þ 18 (100) CA19-9, n ¼ 16 þ 10 (62) MUC1, n ¼ 18 V 9 (50) MUC2, n ¼ 18 V 8 (44) MUC5AC, n ¼ 18 þ 18 (100) MUC6, n ¼ 18 ND 14 (78) DPC4/SMAD4, n ¼ 18 þ 18 (100) than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; ND, no data; V, variable; CK, cytokeratin; S100P, placental S100; pvhl, von Hippel- Lindau tumor suppressor; CDX2, caudal type homeobox 2; CEA, carcinoembryonic antigen; CA19-9, cancer antigen 19-9; GML, Geisinger Medical Laboratories; MUC, mucin; DPC4/SMAD4, SMAD family member 4. Antibody order based on the application of the group of antibodies to a specific entity. a Reprinted from 3 Lin F, Prichard JW, Liu H, Wilkerson M, Schuerch C, eds. Handbook of Practical Immunohistochemistry: Frequently Asked Questions. New York, NY: Springer; 2011, with permission of Springer Science þ Business Media. extremely unusual to have a small cell carcinoma with only a moderately increased MIB-1 (Ki-67) proliferative index. Nearly 100% of metastatic colonic ADCs are positive for SATB2, CDH17, and CDX2; however, medullary carcinoma of the large intestine frequently demonstrates loss of expression of both CDX2 and CK20. In this instance, the tumor cells would likely demonstrate loss of expression of either MLH1/PMS2 or muts homolog 2 (MSH2)/MSH6, and are frequently positive for SATB2, CDH17, trefoil factor 3 (TFF3), MUC4, and calretinin. 120 Entirely tissue-specific markers for the distinction between pancreatic ADC and upper gastrointestinal ADC are not available now. However, loss of DPC4/SMAD4 expression, negative CDH17 expression, and expression of CK17 are suggestive of a pancreatic primary. Table 12. Gallbladder and Extrahepatic Bile Duct Adenocarcinoma (ADC) Versus Reactive Condition Marker Gallbladder ADC Extrahepatic Bile Duct ADC Reactive Condition pvhl þ Maspin þ ND or þ S100P þ þ IMP3 þ mcea þ þ than 5% of cases are positive; or þ, less than 50% of cases are positive; IMP3, insulin-like growth factor II messenger RNA binding protein-3; maspin, mammary serine protease inhibitor; mcea, monoclonal carcinoembryonic antigen; ND, no data; pvhl, von Hippel- Lindau tumor suppressor; S100P, placental S100. GALLBLADDER, EXTRAHEPATIC BILE DUCTS, AND INTRAHEPATIC BILE DUCTS Following an extensive study and review of literature, we believe the 4-marker panel of pvhl, maspin, IMP3, and S100P is the most sensitive and specific diagnostic panel to differentiate ADC of the gallbladder and extrahepatic bile ducts from reactive gallbladder/bile duct mucosa ; findings are summarized in Table 12. Other immunomarkers from recent publications are briefly reviewed here. Glucose transporter 1 (GLUT1) has been reported 121 to be expressed in 51.8% (N ¼ 115) of ADCs of the gallbladder and not expressed in normal/benign gallbladder mucosa. Riener and coworkers 122 studied the expression of P- cadherin, CD24, and p53 in 39 cases of gallbladder ADC and demonstrated their expression in 64%, 42%, and 45% of cases, respectively; in contrast, all 52 cases of normal/ reactive gallbladder mucosa were negative for all 3 proteins. Tamura et al 123 reported expression of regenerating isletderived family, member 4 (REG4) was seen in 56% of gallbladder carcinomas (N ¼ 34) and was negative in all normal/reactive cases (N ¼ 28). Li et al 124 reported the expression frequencies for CDX2 and hepatocyte paraffin-1 were 45.4% and 41.7% in gallbladder carcinomas (N ¼ 108) and 14.3% and 5.7% in chronic cholecystitis (N ¼ 35), respectively; p16 was shown to be expressed in 45% of highgrade dysplasias and in 27.6% of gallbladder ADCs (N ¼ 20) but was negative in 20 normal/reactive cases. 125 Table 11. Metastases in the Pancreas Markers PDADC Kidney Lung-A Melanoma Upper GI Lung-S Colon Breast CK7 þ þ þ þ or þ CK20 þ or þ S100 þ or þ or þ GATA3 þ ER þ TTF1 þ þ Napsin A or þ þ SATB2 or þ þ pvhl þ PAX8 þ Synap þ DPC4 or þ þ þ þ þ þ þ þ CK17 þ or than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; CK, cytokeratin; DPC4, SMAD family member 4; ER, estrogen receptors; GATA3, GATA binding protein 3; GI, gastrointestinal; lung-a, lung adenocarcinoma; lung-s, lung small cell carcinoma; PAX8, paired box gene 8; PDADC, pancreatic ductal adenocarcinoma; pvhl, von Hippel-Lindau tumor suppressor; SATB2, special AT-rich sequence-binding protein 2; synap, synaptophysin; TTF1, thyroid transcription factor 1. Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al 35

13 Table 13. Intrahepatic Cholangiocarcinomas (ICCs) Versus Pancreatic Ductal Adenocarcinomas (DADC) Marker Pancreatic DADC ICC pvhl þ S100P þ or þ CK17 þ or MUC5AC þ or than 5% of cases are positive; þ or, usually more than 50% of cases are positive; or þ, less than 50% of cases are positive; CK17, cytokeratin 17; MUC5AC, mucin 5AC; pvhl, von Hippel-Lindau tumor suppressor; S100P, placental S100. In contrast to ADCs of the pancreas, gallbladder, and extrahepatic bile ducts, ICC has a unique IHC profile, as summarized in Table 13, in which, 71% of ICCs were positive for pvhl, and the expression of S100P, MUC5AC, and CK17 was seen in only 27%, 12%, and 12% of cases, respectively. 61 In summary, application of IHC in the diagnosis and classification of a pancreatic solid or cystic tumor has become a useful ancillary study and plays a crucial role in reaching a definitive diagnosis. An accurate diagnosis of a pancreatic tumor is essential in understanding tumor pathobiology, in making an appropriate therapeutic decision, and in predicting a prognostic outcome. This review article (1) recommends a simple working algorithm for workup of a solid or cystic pancreatic tumor, (2) discusses the emerging biomarkers for common pancreatic tumors with a focus on their application and pitfalls, and (3) identifies the most effective IHC panels for solving frequently encountered diagnostic challenges and minimizing overuse or underuse of the currently available and rapidly growing list of biomarkers. With advances in molecular methodologies, additional tissue-specific biomarkers for ADCs of the pancreas, gallbladder, and extrahepatic and intrahepatic bile ducts will be discovered in the future. References 1. Hruban RH, Boffetta P, Hiraoka N, et al. Ductal adenocarcinoma of the pancreas. In: Bosman FT, Carneiro F, Hruban RH, and Theise ND, eds. WHO Classification of Tumours of the Digestive System. 4th ed. Lyon, France: IARC Press; 2010: World Health Organization Classification of Tumours; vol Hruban RH, Pitman MB, Klimstra DS. Tumors of the Pancreas. Washington, DC: American Registry of Pathology and Armed Forces Institute of Pathology; 2007: AFIP Atlas of Tumor Pathology; 4th series, fascicle Lin F, Wang HL. Pancreas and ampulla. In: Lin F, Prichard JW, Liu H, Wilkerson M, Schuerch C, eds. Handbook of Practical Immunohistochemistry: Frequently Asked Questions. New York, NY: Springer; 2011: Chu PG, Weiss LM. Tumors of the digestive system. In: Modern Immunohistochemistry. New York, NY: Cambridge University Press; 2009: Basturk O, Farris, AB III, Adsay NV. Immunohistology of the pancreas, biliary tract, and liver. In: Dabbs DJ, ed. Diagnostic Immunohistochemistry: Theranostic and Genomic Applications. 3rd ed. Philadelphia, PA: Saunders Elsevier; 2010: Geller SA, Dhall D, Alsabeh R. Application of immunohistochemistry to liver and gastrointestinal neoplasms: liver, stomach, colon, and pancreas. Arch Pathol Lab Med. 2008;132(3): Goldstein NS, Bassi D. Cytokeratins 7, 17, and 20 reactivity in pancreatic and ampulla of Vater adenocarcinomas. Percentage of positivity and distribution is affected by the cut-point threshold. Am J Clin Pathol. 2001;115(5): Hornick JL, Lauwers GY, Odze RD. Immunohistochemistry can help distinguish metastatic pancreatic adenocarcinomas from bile duct adenomas and hamartomas of the liver. Am J Surg Pathol. 2005;29(3): Chu P, Wu E, Weiss LM. Cytokeratin 7 and cytokeratin 20 expression in epithelial neoplasms: a survey of 435 cases. Mod Pathol. 2000;13(9): Chu PG, Schwarz RE, Lau SK, Yen Y, Weiss LM. Immunohistochemical staining in the diagnosis of pancreatobiliary and ampulla of Vater adenocarcinoma: application of CDX2, CK17, MUC1, and MUC2. Am J Surg Pathol. 2005; 29(3): Lau SK, Prakash S, Geller SA, Alsabeh R. Comparative immunohistochemical profile of hepatocellular carcinoma, cholangiocarcinoma, and metastatic adenocarcinoma. Hum Pathol. 2002;33(12): Bhardwaj A, Marsh WL Jr, Nash JW, Barbacioru CC, Jones S, Frankel WL. Double immunohistochemical staining with MUC4/p53 is useful in the distinction of pancreatic adenocarcinoma from chronic pancreatitis: a tissue microarray-based study. Arch Pathol Lab Med. 2007;131(4): Coppola D, Lu L, Fruehauf JP, et al. Analysis of p53, p21waf1, and TGFb1 in human ductal adenocarcinoma of the pancreas: TGF-b1 protein expression predicts longer survival. Am J Clin Pathol. 1998;110(1): Apple SK, Hecht JR, Lewin DN, Jahromi SA, Grody WW, Nieberg RK. Immunohistochemical evaluation of K-ras, p53, and HER-2/neu expression in hyperplastic, dysplastic, and carcinomatous lesions of the pancreas: evidence for multistep carcinogenesis. Hum Pathol. 1999;30(2): DiGiuseppe JA, Hruban RH, Goodman SN, et al. Overexpression of p53 protein in adenocarcinoma of the pancreas. Am J Clin Pathol. 1994;101(6): Werling RW, Yaziji H, Bacchi CE, Gown AM. CDX2, a highly sensitive and specific marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476 primary and metastatic carcinomas. Am J Surg Pathol. 2003;27(3): Moskaluk CA, Zhang H, Powell SM, Cerilli LA, Hampton GM, Frierson HF Jr. Cdx2 protein expression in normal and malignant human tissues: an immunohistochemical survey using tissue microarrays. Mod Pathol. 2003;16(9): De Lott LB, Morrison C, Suster S, Cohn DE, Frankel WL. CDX2 is a useful marker of intestinal-type differentiation: a tissue microarray-based study of 629 tumors from various sites. Arch Pathol Lab Med. 2005;129(9): Ordóñez NG. Value of claudin-4 immunostaining in the diagnosis of mesothelioma. Am J Clin Pathol. 2013;139(5): Yantiss RK, Woda BA, Fanger GR, et al. KOC (K homology domain containing protein overexpressed in cancer): a novel molecular marker that distinguishes between benign and malignant lesions of the pancreas. Am J Surg Pathol. 2005;29(2): Findeis-Hosey JJ, Xu H. The use of insulin like-growth factor II messenger RNA binding protein-3 in diagnostic pathology. Hum Pathol. 201;42(3): Wachter DL, Schlabrakowski A, Hoegel J, Kristiansen G, Hartmann A, Riener MO. Diagnostic value of immunohistochemical IMP3 expression in core needle biopsies of pancreatic ductal adenocarcinoma. Am J Surg Pathol. 2011; 35(6): Zhao H, Mandich D, Cartun RW, Ligato S. Expression of K homology domain containing protein overexpressed in cancer in pancreatic FNA for diagnosing adenocarcinoma of pancreas. Diagn Cytopathol. 2007;35(11): Ohike N, Maass N, Mundhenke C, et al. Clinicopathological significance and molecular regulation of maspin expression in ductal adenocarcinoma of the pancreas. Cancer Lett. 2003;199(2): Cao D, Zhang Q, Wu LS, et al. Prognostic significance of maspin in pancreatic ductal adenocarcinoma: tissue microarray analysis of 223 surgically resected cases. Mod Pathol. 2007;20(5): Wente MN, Jain A, Kono E, et al. Prostate stem cell antigen is a putative target for immunotherapy in pancreatic cancer. Pancreas. 2005;31(2): Argani P, Rosty C, Reiter RE, et al. Discovery of new markers of cancer through serial analysis of gene expression: prostate stem cell antigen is overexpressed in pancreatic adenocarcinoma. Cancer Res. 2001;61(11): McCarthy DM, Maitra A, Argani P, et al. Novel markers of pancreatic adenocarcinoma in fine-needle aspiration: mesothelin and prostate stem cell antigen labeling increases accuracy in cytologically borderline cases. Appl Immunohistochem Mol Morphol. 2003;11(3): Ordonez NG. Application of mesothelin immunostaining in tumor diagnosis. Am J Surg Pathol. 2003;27(11): Hassan R, Laszik ZG, Lerner M, Raffeld M, Postier R, Brackett D. Mesothelin is overexpressed in pancreaticobiliary adenocarcinomas but not in normal pancreas and chronic pancreatitis. Am J Clin Pathol. 2005;124(6): Frierson HF Jr, Moskaluk CA, Powell SM, et al. Large-scale molecular and tissue microarray analysis of mesothelin expression in common human carcinomas. Hum Pathol. 2003;34(6): Swierczynski SL, Maitra A, Abraham SC, et al. Analysis of novel tumor markers in pancreatic and biliary carcinomas using tissue microarrays. Hum Pathol. 2004;35(3): Baruch AC, Wang H, Staerkel GA, Evans DB, Hwang RF, Krishnamurthy S. Immunocytochemical study of the expression of mesothelin in fine-needle aspiration biopsy specimens of pancreatic adenocarcinoma. Diagn Cytopathol. 2007;35(3): Cao D, Maitra A, Saavedra JA, Klimstra DS, Adsay NV, Hruban RH. Expression of novel markers of pancreatic ductal adenocarcinoma in pancreatic nonductal neoplasms: additional evidence of different genetic pathways. Mod Pathol. 2005;18(6): Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in the Pancreatobiliary Tract Lin et al