Application of Immunohistochemistry in Thyroid Pathology

Application of Immunohistochemistry in Thyroid Pathology Haiyan Liu, MD; Fan Lin, MD, PhD Context. Accurate classification of follicular-patterned thyroid lesions is not always an easy task on routine surgical hematoxylin-eosin stained or cytologic fine-needle aspiration specimens. The diagnostic challenges are partially due to differential diagnostic criteria that are often subtle and subjective. In the past decades, tremendous advances have been made in molecular gene profiling of tumors and diagnostic immunohistochemistry, aiding in diagnostic accuracy and proper patient management. Objective. To evaluate the diagnostic utility of the most commonly studied immunomarkers in the field of thyroid pathology by review of the literature, using the database of indexed articles in PubMed (US National Library of Medicine) from 1976 2013. Data Sources. Literature review, authors research data, and personal practice experience. Conclusions. The appropriate use of immunohistochemistry by applying a panel of immunomarkers and using a standardized technical and interpretational method may complement the morphologic assessment and aid in the accurate classification of difficult thyroid lesions. (Arch Pathol Lab Med. 2015;139:67 82; doi: 10.5858/ arpa.2014-0056-ra) In 2013, the American Cancer Society estimated that chemical markers and molecular markers, using somatic 60 220 new cases of thyroid carcinoma would be mutation, gene expression, and microrna analyses. The diagnosed in the United States, and 1850 persons would immunohistochemical markers are the scope of this review. die of the disease. The overall incidence of thyroid The objective of this article is to assess the diagnostic utility carcinoma has increased more rapidly than that of any other malignancy in recent years, especially in women, from 1.3 per 100 000 in 1935 to 16.3 per 100 000 in 2008. 1,2 Most of the most commonly studied and published immunomarkers in the current literature in the field of thyroid pathology. thyroid carcinomas (95%) are derived from follicular epithelial cells and are mainly well differentiated, including papillary thyroid carcinoma (PTC) and follicular thyroid IMMUNOMARKERS USED IN THYROID-ORGAN SPECIFIC DIFFERENTIATION carcinoma (FTC), with a minor fraction of poorly differentiated carcinomas (PDCs) and undifferentiated carcinomas 8 (PAX8), and TTF2 (FOXE1) are crucial for thyroid Thyroid transcription factors (TTFs) TTF1, paired box gene (UDCs). 3 Only 3% of thyroid carcinomas are medullary organogenesis and differentiation. 8 14 The transcription carcinomas (MCs) derived from parafollicular C cells. factors control the expression of thyroglobulin (TGB), The diagnosis of thyroid carcinomas is straightforward in thyroperoxidase (TPO), thyroid-stimulating hormone receptor, and thyroid iodine transporter. 15,16 Immunohisto- most cases. However, pathologists are not infrequently confronted with lesions exhibiting equivocal features that chemically, they serve as organ-specific immunomarkers. make the distinction of benign from malignant difficult. Thyroglobulin is the primary synthetic product of the Although the diagnostic criteria for PTC and FTC are clearly thyroid and the macromolecular precursor of thyroid outlined in textbooks, the application of the criteria is hormones (T3 and T4), serving as a specific marker for subjective and circumstantial. Interobservor or intraobservor thyroid follicular cell origin. 17 19 disagreements in the diagnosis of follicular thyroid lesions are well documented, even among expert pathologists. 4 7 In Thyroid Transcription Factor 1 recent years, tremendous efforts in search of objective Thyroid transcription factor 1, also named NKX2 homeobox 1 (NKX2.1), is a nuclear protein, approximately 38 kda, measures that differentiate benign from malignant lesions have been undertaken, mainly involving immunohisto- composed of a single polypeptide of 371 amino acids belonging to the family of homeodomain transcription Accepted for publication March 10, 2014. factors. Thyroid transcription factor 1 plays a crucial role in From the Department of Laboratory Medicine, Geisinger Medical the organogenesis and differentiation of thyroid and Center, Danville, Pennsylvania. lung. 9 11 In addition, TTF1 transcripts were detected in The authors have no relevant financial interest in the products or developing diencephalon, restricted to the hypothalamic companies described in this article. Reprints: Haiyan Liu, MD, Department of Laboratory Medicine, area and to the infundibulum at the earliest stages of their MC 01 31, Geisinger Medical Center, 100 N Academy Ave, differentiation, forming the posterior lobe of the pituitary, Danville, PA 17822 (e-mail: hliu1@geisinger.edu). Arch Pathol Lab Med Vol 139, January 2015 namely, the neurohypophysis. 9,20 Immunohistochemistry in Thyroid Pathology Liu & Lin 67

Source, y Table 1. Thyroid transcription factor 1 expression by immunohistochemical analysis was initially exclusively identified in thyroid and lung epithelial tissues, including normal, benign, and malignant tissues. 18,21 30 In routine practice, TTF1 became one of the most commonly used immunomarkers to identify thyroid or lung primary tumor in the setting of metastasis and to differentiate adenocarcinoma from squamous cell carcinoma in poorly differentiated non small cell carcinomas of the lung in small biopsy or cytologic specimens. In recent years, TTF1-positive, nonpulmonary, nonthyroid carcinomas have been reported, including colorectal, ovarian, breast, endometrial, and endocervical adenocarcinomas. 31 42 Comparing the 2 commercially available TTF1 monoclonal antibodies, SPT24 and 8G7G3/1, the rate of aberrant TTF1 expression in nonpulmonary, nonthyroid carcinomas is higher for SPT24, greater than 10% in colorectal adenocarcinomas as reported by some investigators. However, if using the 8G7G3/1 clone, the aberrant expression rate is low, less than 2% in general. Normal thyroid follicular cells and parafollicular cells show diffuse expression of TTF1. In thyroid neoplasm, TTF1 expression was reported in nearly 100% of PTCs, FTCs, and follicular adenomas (FAs); in approximately 90% of PDCs and MCs; and in none to fewer than 25% of UDCs.* Paired Box Gene 8 Paired box gene 8 is a member of the paired box (PAX) family of transcription factors that is expressed during organogenesis of the thyroid gland, müllerian tract, and kidney. 47 52 Expression of PAX8 was mainly reported in thyroid and renal neoplasms and infrequently in bladder Summary of Hector Battifora Mesothelia-1 (HBME-1) Studies a d PTC, No. (%) FVPTC, FTC, PDC, UDC, HCC, Barroeta et al, 114 2006 10/11 (91) 3/4 (75) 5/7 (71) 3/3 (100) 4/5 (80) 1/4 (25) Park et al, 102 2007 166/181 (92) 22/25 (88) Scognamiglio et al, 112 2006 43/49 (88) 25/29 (86) Nasr et al, 115 2006 49/51 (96) 9/10 (90) Prasad et al, 108 2005 57/67 (85) 3/6 (50) 0/4 (0) 1/8 (13) Cheung et al, 118 2001 76/138 (55) 38/84 (45) 2/4 (50) 4/6 (67) 1/2 (50) 2/7 (29) Liu et al, 109 2008 39/53 (74) 8/11 (73) 2/13 (15) Rossi et al, 107 2006 39/42 (93) 12/14 (86) Saggiorato et al, 106 2005 39/42 (93) 21/33 (64) Rossi et al, 120 2013 15/15 (100) 0/9 (0) de Matos et al, 110 2005 79/84 (94) 21/25 (84) 24/38 (63) 0/2 (0) Nechifor-Boila et al, 121 2013 94/98 (96) 73/90 (81) 4/9 (44) Torregrossa et al, 122 2007 190/200 (95) 128/138 (93) Volante et al, 104 2004 28/32 (88) e 26/70 (37) Mai et al, 116 2002 f 42/42 (100) 19/29 (66) 2/12 (17) Mase et al, 123 2003 35/36 (97) 33/39 (85) 2/2 (100) Papotti et al, 124 2005 14/14 (100) Ito et al, 125 2005 37/37 (100) 84/138 (61) Saleh et al, 95 2010 18/20 (90) 11/12 (92) 18/22 (82) Abbreviations: FA, follicular adenoma; FTC, follicular thyroid carcinoma; FVPTC, follicular variant papillary thyroid carcinoma; Graves, Graves disease; HA, Hürthle cell adenoma; HCC, Hürthle cell carcinoma; LTis, lymphocytic thyroiditis; MC, medullary thyroid carcinoma; Neg, negative; NG, nodular goiter; NL, normal thyroid tissue; PDC, poorly differentiated thyroid carcinoma; PTC, papillary thyroid carcinoma; UDC, undifferentiated thyroid carcinoma (anaplastic carcinoma). a Asa, 128 2005: HBME-1 was expressed in 40% of malignant thyroid tumors, including PTC and FTC. b Zhu et al, 126 2010: HBME-1 expression was identified in 95% (148 of 155) of PTCs. c Barut et al, 105 2010: HBME-1 expression was identified in 72% (47 of 65) of malignant thyroid neoplasms; rare focal reactivity was identified in benign lesions. d Wiseman et al, 127 2008: HBME-1 expression was identified in 54% (53 of 99) of malignant thyroid neoplasms; only 5% (5 of 100) of benign lesions. e Hürthle cell variant PTC. f Mai et al, 116 2002: In addition, HBME-1 expression in Hürthle cell PTC was 38% (5 of 13). * References 18, 25, 26, 28, 30, 31, 39, 43 46. tumors, as well as müllerian-origin malignancies (including ovarian malignant neoplasms). 26,46,53 65 Few studies 53,54 also reported its expression in thymic and parathyroid neoplasms. Recently, PAX8 expression was documented in pancreatic neuroendocrine tumors. 66 68 Tacha et al 54 reported that other cancers, including carcinomas of breast, lung, prostate, gastrointestinal tract, liver, mesothelioma, melanoma, and testicular tumors, lacked PAX8 expression; 14% (3 of 22) of pancreatic adenocarcinomas and 40% (2 of 5) of rhabdomyosarcomas were PAX8 positive. In normal tissues, strong nuclear staining for PAX8 was observed in follicular cells of the thyroid; müllerian epithelial cells including endometrium, endocervix, and secretory cells of the fallopian tube; renal tubular epithelium; epithelial lining of the vas deferens; as well as islet cells of the pancreas. 54 Diffuse weak to moderate nuclear staining was observed in some cases of parathyroid tissue and nonneoplastic thymic epithelial cells. Urothelium and squamous mucosa were reported to have weak and patchy staining. Several studies investigated PAX8 expression in thyroid neoplasms, reporting a positive rate of nearly 100% in PTCs and FAs, 43,46,53,54,69 91% to 100% in FTCs, 46,53,54 75% to 100% in PDCs, 46,53,58 and 50% to 80% in UDCs. 46,53,54,56,58 Medullary carcinomas were reported to be PAX8 nonreactive with rare exceptions: positive in 75% (6 of 8, most 1þ) 46 and 41% (13 of 32) 54 of cases. Thyroid Transcription Factor 2 Thyroid transcription factor 2, also named forkhead box E1 (FOXE1), was originally identified as a thyroid-specific, forkhead-domain containing nuclear protein capable of recognizing and binding to a DNA sequence present in the promoters of both TGB and TPO, 2 genes expressed exclusively in thyroid follicular cells. 15,70 Thyroid transcrip- 68 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin

MC FA, HA, Table 1. Extended NG, LTis, Graves, NL, 1/3 (33) 1/18 (6) 2/12 (17) 1/4 (25) 17/35 (49) 11/54 (20) 2/49 (4) 0/6 (0) 0/14 (0) 4/19 (21) 0/8 (0) 0/10 (0) 2/21 (10) 1/29 (3) 0/14 (0) 0/59 (0) 0/35 (0) 0/40 (0) 1/12 (9) 0/14 (0) 0/10 (0) 1/17 (6) 0/41 (0) 2/50 (4) 0/5 (0) 10/18 (56) 9/10 (90) 4/12 (33) 0/170 (0) 4/13 (31) 6/50 (12) 19/45 (42) 4/48 (8) 4/4 (100) 17/62 (27) 8/62 (13) 1/15 (7) 47/155 (30) 17/98 (17) Neg 26/46 (57) 9/52 (17) tion factor 2 expression is restricted to the thyroid; anterior pituitary; epithelia of the oropharynx, trachea, and esophagus; exocrine cells of the seminiferous tubules of the testis; and epidermis and hair follicles. 12,71 78 Thyroid transcription factor 2 is 1 of 3 TTFs that play critical roles in thyroid organogenesis and differentiation. However, TTF2 seems to play a role in the development of the negative controller of thyroid-specific gene expression. 79,80 Mutations in TTF2 (FOXE1) may be involved in human disorders characterized by congenital hypothyroidism, thyroid dysgenesis, and cleft palate. 81,82 There are only few studies in literature investigating TTF2 expression by immunohistochemical analysis in human tumors and normal tissues. Nonaka et al 46 reported that TTF2 was expressed exclusively in normal thyroid follicular cells and a few C cells in thyroid C-cell hyperplasia and thyroid neoplasms, including 100% of cases of PTC (17 of 17), FA (18 of 18), FTC (16 of 16), and PDC (7 of 7); 75% (6 of 8) of cases of MC in a focal staining pattern; and only 7% (2 of 28, focal 1þ, 1% 25% of nuclei) of cases of UDC. All other neoplastic and nonneoplastic tissues showed no reactivity, including lung, esophagus, stomach, colon, pancreas, kidney, breast, ovary, prostate, urinary bladder, skin, testis, lymph node, and soft tissue. Matoso et al 83 studied 6 cases of spindle cell foci in thyroid glands by immunohistochemical analysis and found that 100% (6 of 6) of cases showed nuclear staining for TTF2. Increased expression of TTF2 by in situ hybridization was reported in epidermis and basal cell carcinomas. Zhang et al 36 documented the lack of expression of TTF2 by immunohistochemical evaluation in 212 cases of ovarian and endometrial malignancies to ascertain that the TTF1-positive tumors in their study were indeed not follicular cell in origin. Although there are only few studies evaluating TTF2 expression in neoplastic and nonneoplastic tissues by immunohistochemistry, the data are appealing. Further study in a large series of cases is deemed necessary to assess the diagnostic utility of TTF2 in routine practice. Thyroglobulin Thyroglobulin, a thyroid hormone precursor, is a glycoprotein synthesized by thyrocytes, transported to the apical surface and secreted into the follicles, constituting the major component of colloid. Immunohistochemically, TGB expression is diffuse in 100% of normal thyroid follicular epithelial cells and 83% to 100% of FAs. 18,43,84 However, the staining pattern in lesions with Hürthle cell morphology was reported as weak or focal. In primary thyroid carcinomas, the expression of TGB, both at the messenger RNA and protein level, showed a certain degree of correlation with tumor differentiation. Studies 18,43,84 88 have reported 100% expression of TGB in PTC, 75% to 96% in FTC, 57% to 92% in PDC, and near lack of expression in UDC and MC. Metastatic thyroid carcinomas were reported as showing staining patterns similar to those of the primary tumors. 18,89 In nonthyroid tissues and tumors, including parathyroid gland, lung, stomach, pancreas, ovary, kidney, colon, salivary glands, prostate, breast, and glomus body, TGB expression was lacking. 43,89 91 In our previous study, 92 TGB expression was demonstrated in 14 of 14 (100%) cases of normal thyroid tissue, 45 of 45 (100%) cases of PTC, 36 of 36 (100%) cases of FTC, and 0 of 10 (0%) cases of MC. IMMUNOMARKERS USED IN THE DIFFERENTIAL DIAGNOSIS OF THYROID NEOPLASMS The histomorphologic diagnosis of thyroid neoplasm remains the cornerstone in the classification of thyroid follicular lesions. However, for those tumors that are poorly differentiated or undifferentiated, not follicular derived, and exhibit equivocal histomorphologic features, the application of immunohistochemical biomarkers may play an active or complementary role in their accurate classification. 93 99 Among the variety of biomarkers reported in the literature, Hector Battifora mesothelial 1 (HBME-1), galectin-3 (GAL- 3), cytokeratin 19 (CK19), Cbp/p300-interacting transactivator with Glu/Asp-rich carboxy-terminal domain, 1 (CIT- ED1), and TPO are most promising. However, there is no single marker sensitive enough to provide a definitive malignant diagnosis. Therefore, different panels of combined immunomarkers were proposed by many investigators. 98,100 112 The combination of HBME-1, GAL-3, and CK19 was by far the most common panel evaluated by investigators, and their diffuse expression has not been reported in benign lesions. Other markers, such as Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin 69

Source, y Table 2. PTC, Summary of Cytokeratin 19 (CK19) Studies a h FVPTC, FTC, PDC, UDC, HCC, Barroeta et al, 114 2006 10/11 (91) 4/4 (100) 3/7 (43) 2/3 (67) 4/5 (80) 1/4 (25) Park et al, 102 2007 175/181 (97) 11/25 (44) Scognamiglio et al, 112 2006 49/49 (100) 26/29 (90) Nasr et al, 115 2006 51/51 (100) 10/10 (100) Prasad et al, 108 2005 48/67 (72) 3/6 (50) 1/4 (25) 4/8 (50) Erkilic et al, 131 2002 25/25 (100) Sahoo et al, 132 2001 15/15 (100) 10/10 (100) Cheung et al, 118 2001 91/138 (66) 48/84 (57) 0/4 (0) 3/6 (50) 0/2 (0) 2/7 (29) Liu et al, 109 2008 41/53 (78) 2/11 (22) 0/13 (0) Murphy et al, 130 2008 20/20 (100) 2/9 (18) 6/14 (43) Rossi et al, 107 2006 36/42 (86) 11/14 (79) Saggiorato et al, 106 2005 37/42 (88) 23/33 (70) Beesley and McLaren, 111 2002 26/26 (100) 5/12 (42) 1/1 (100) Song et al, 133 2011 425/441 (96) de Matos et al, 110 2005 61/84 (73) 13/25 (52) 8/38 (21) 0/2 (0) Baloch et al, 129 1999 39/39 (100) 26/26 (100) 0/2 (0) Bose et al, 134 2012 22/22 (100) Saleh et al, 95 2010 17/20 (85) 10/12 (83) 19/22 (86) Abbreviations: FA, follicular adenoma; FTC, follicular thyroid carcinoma; FVPTC, follicular variant papillary thyroid carcinoma; Graves, Graves disease; HA, Hürthle cell adenoma; HCC, Hürthle cell carcinoma; LTis, lymphocytic thyroiditis; MC, medullary thyroid carcinoma; Neg, negative; NG, nodular goiter; NL, normal thyroid tissue; PDC, poorly differentiated thyroid carcinoma; PTC, papillary thyroid carcinoma; UDC, undifferentiated thyroid carcinoma (anaplastic carcinoma). a Asa, 128 2005: CK19 expression identified in 60% of PTC cases; nonreactive or focal positivity in follicular neoplasm and NG, usually inflamed tissue. b Zhu et al, 126 2010: CK19 expression was identified in 87% (135 of 155) of PTC cases. c Nechifor-Boila et al, 121 2013: CK19 expressed in 46% (93 of 204) of PTC cases. d Cameron and Berean, 234 2003: Diffuse CK19 expression is identified in 100% of PTCs, including FVPTC; FTC, FA, and NG were negative or showed focal reactivity. e Casey et al, 135 2003: CK19 expression is identified in 100% (30 of 30) of malignant tumors, however with significant reactivity in nonneoplastic thyroid tissue. f Barut et al, 105 2010: CK19 expression identified in 77% (50 of 65) of malignant thyroid tumors; only focal reactivity noted in benign lesions. g Wiseman et al, 127 2008: CK19 is expressed in 97% (96 of 99) of malignant tumors. h Nakamura et al, 100 2006: A panel consisting of Hector Battifora mesothelial-1 (HBME-1), galectin-3 (GAL-3), and CK19 or HBME-1, Cbp/p300- interacting transactivator with Glu/Asp-rich carboxy-terminal domain, 1 (CITED-1), and GAL-3: useful in distinguishing FA from FVPTC. i One hundred fifty-one cases of NG and FA. antibodies against proteins involved in cell proliferation and in the regulation of cell cycle, as well as oncogene proteins, are also being studied and documented in the literature. Recently, we found a distinct membranous staining pattern for trophoblastic cell surface antigen 2 (TROP2), a 35-kDa type 1 transmembranous glycoprotein, in PTCs; in contrast, follicular neoplasms (FAs and FTCs) were nonreactive or showed only rare focal, weak cytoplasmic staining. We propose that TROP2 is a potential novel immunomarker for the identification of PTC that can be used in a panel to increase diagnostic accuracy when encountering a difficult follicular cell derived lesion. 113 Hector Battifora Mesothelial 1 Hector Battifora mesothelial 1 is an unelucidated membrane antigen found in the microvilli of mesothelial cells, normal tracheal epithelium, and adenocarcinoma of the lung, pancreas, and breast. In the past decade, several indicator of malignancy, especially true for PTC. The overall sensitivity of HBME-1 was 78.8% for thyroid malignancy, 87.3% for PTC, and 65.2% for FTC. The specificity was 82.1%. Expression of HBME-1 was also noted in benign thyroid lesions such as FA, nodular goiter, and lymphocytic thyroiditis (LTis), usually in a focal staining fashion, with a reported overall positive rate of 26%, 12%, and 19%, respectively. Hyaline trabecular tumors were reported to lack HBME-1 expression, suggesting its discriminating role in the distinction between hyaline trabecular tumor and PTC; however, Lenggenhager et al 119 recently reported that 37.5% (3 of 8) of cases of hyaline trabecular tumor showed patchy but strong membranous and/or cytoplasmic reactivity to HBME-1. Many investigators proposed using a panel of immunomarkers, most commonly a combination of HBME-1, CK19, and GAL-3, to increase the discriminating power between benign and malignant neoplasms when a histologically equivocal lesion is encountered. studies have investigated HBME-1 expression in benign and Cytokeratin 19 malignant thyroid tissues; the published studies are reviewed and summarized in Table 1. In normal thyroid Cytokeratin 19 is a low-molecular-weight cytokeratin tissue, there was virtually no expression of HBME- found in a variety of simple or glandular epithelia, both 1. 104,108,114 116 Overexpression of HBME-1 was demonstrated normal and their neoplastic counterparts. In the thyroid in malignant thyroid neoplasms, especially PTCs, with the gland, normal follicular epithelium usually has shown no exception of Hürthle cell carcinomas. Several investigators detectable CK19 expression 110,115,129 ; however, few resion 104,108,114 119 have reported a reduced or lack of expres- for HBME-1 in Hürthle cell carcinomas or thyroid References 95, 104 110, 112, 114 116,118,120 128. neoplasms with Hürthle cell features. Studies suggested that References 95, 104 110, 112, 114 116, 118, 119, 121, 123 125, overexpression of HBME-1 in a thyroid nodule was an 127. 70 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin

MC, FA, HA, Table 2. Extended NG, LTis, Graves, NL, 0/3 (0) 0/3 (0%) 3/18 (17) 8/12 (67) 6/13 (46) 1/4 (25) 10/35 (29) 5/54 (9) 7/49 (14) 5/6 (83) 6/14 (43) 16/19 (84) 2/8 (25) 10/10 (100) 1/21 (5) 5/25 (20) 20/20 (100) 7/35 (20) 8/40 (20) 0/12 (0) 0/14 (0) 1/10 (10) 4/15 (27) 0/11 (0%) 8/19 (42) 1/17 (6) 1/41 (2) 5/50 (10) 2/2 (100%) 5/20 (25) 2/8 (25) 39/151 (26%) i 1/5 (20) 6/18 (33) 6/10 (60) 2/12 (17) 0/170 (0) 0/4 (0) 0/5 (0) Neg 6/8 (75) 4/8 (50) 23/46 (50) 8/52 (15) ports 114,128,130 noted CK19 expression in normal thyroid tissue in a focal staining pattern, especially in inflamed tissue. Many studies reported a strong and diffuse staining pattern of CK19 in PTC. However, its expression in follicular cells of LTis 114,115 and follicular neoplasms (FA or FTC) was also demonstrated ; therefore, positive CK19 staining lacks specificity for PTC or malignancy. Table 2 summarizes the data from studies of CK19 expression in various thyroid lesions. The overall sensitivity of CK19 was 79.3% for malignancy, 82.2% for PTC, and 44.3% for FTC. The specificity was 63.1%. Overexpression of CK19 is a good indicator for PTC; however, the sensitivity for follicular carcinoma is low. There are high rates of reactivity in benign thyroid lesions, especially in LTis. 114,115 Compared with the diffuse, strong reactivity in PTC, most studies indicated focal reactivity in benign lesions. Cytokeratin 19 may have added value as part of a panel of immunomarkers in the diagnosis of PTC. Galectin-3 Galectin-3 is a member of a family of b-galactoside binding animal lectins shown to be involved in tumor progression and metastasis. Both nuclear and cytoplasmic expression of GAL-3 has been demonstrated in a variety of tissues and cells. Study of mouse 3T3 fibroblasts revealed the presence of both phosphorylated and nonphosphorylated GAL-3; the former resides in both nucleus and cytoplasm; the latter, exclusively in the nucleus. Cell proliferation is associated with an increased level of both forms, while alterations in nuclear versus cytoplasmic GAL- 3 localization have been shown to be associated with neoplastic progression. 136,137 Galectin-3 plays an important role in cell-cell/cell-matrix adhesion, cell growth, neoplastic transformation/spread, cell cycle regulation/apoptosis, and cell repair processes. In recent years, overexpression of GAL-3 has been reported in various human carcinomas, most noticeably in well-differentiated follicular-derived thyroid carcinomas. Many studies investigated GAL-3 expression in cytologic and histologic specimens of thyroid References 95, 104 108,110 112,114, 115,118,130 135. References 95, 103 112, 114, 117, 120, 122, 124, 126 128, 130, 133, 138 146. nodular lesions; these are summarized in Table 3. Galectin-3 has been found to be useful in differentiating malignant thyroid lesions (such as PTC and the follicular variant of PTC) from benign lesions. The overall sensitivity of GAL-3 was 84.6% for malignancy, 87.5% for PTC, and 72.6% for FTC. The specificity was 83.6%. In general, GAL-3 expression in benign lesions is often focal, and in contrast, is diffuse in malignant lesions. Trophoblastic Cell Surface Antigen 2 Trophoblastic cell surface antigen 2, also known as tumorassociated calcium signal transducer 2 (TAC-STD2), is a transmembrane glycoprotein associated with tumor development and progression in a variety of epithelial carcinomas including ovarian, colorectal, pancreatic, gastric, pulmonary, endometrioid endometrial, and oral cavity squamous cell carcinomas. 147 157 In normal tissues, TROP2 expression was reported as low or lacking. Many studies 158 163 have demonstrated that TROP2 overexpression is associated with tumor aggressiveness and poor prognosis. Immunohistochemical evaluation of TROP2 expression in thyroid tumors has not been well studied. Recently, we 113 evaluated TROP2 expression in 136 cases of thyroid neoplasms (48 cases of PTC, 37 cases of FTC, 51 cases of FA) and normal thyroid tissues (n ¼ 15) on tissue microarray (TMA) sections, as well as in routine sections of 61 atypical follicular lesions and 20 benign thyroid lesions (10 cases each of chronic LTis and nodular goiter) by immunohistochemical analysis with the Ventana BenchMark Ultra (Ventana Medical Systems Inc, Tucson, Arizona). In the TMAs of thyroid neoplasms, 90% (43 of 48) of PTCs demonstrated a strong membranous staining pattern, with most being diffuse (3þ or 4þ), as illustrated in Figure 1, A and B. In contrast, 96% (49 of 51) of FAs and 89% (33 of 37) of FTCs lacked TROP2 expression; only 2 of 51 FA cases and 4 of 37 FTC cases showed focal (1þ) strong cytoplasmic staining without membranous condensation, as illustrated in Figure 1, C and D. The staining results are summarized in Table 4. For the 61 routine sections of atypical follicular lesion, 70% (23 of 33) of PTCs were positive for TROP2, with a diffuse staining pattern (3þ or 4þ) in 83% (19 of 23) of the positive cases. All 11 cases of atypical follicular neoplasm and 17 cases of adenomatoid nodule with focal nuclear Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin 71

Source, y Table 3. PTC, Summary of Galectin-3 (GAL-3) Studies a e FVPTC, FTC, PDC, UDC, HCC, Barroeta et al, 114 2006 9/11 (82) 3/4 (75) 4/7 (57) 1/3 (33) 4/5 (80) 3/4 (75) Park et al, 102 2007 179/181 (99) 16/25 (64) Scognamiglio et al, 112 2006 47/49 (96) 26/29 (90) Prasad et al, 108 2005 63/67 (94) 4/6 (67) 4/4 (100) 7/8 (88) Liu et al, 109 2008 49/53 (92) 4/11 (36) 4/13 (31) Murphy et al, 130 2008 20/20 (100) 5/9 (55) 3/14 (21) Rossi et al, 107 2006 37/42 (88) 10/14 (71) Saggiorato et al, 106 2005 42/42 (100) 29/33 (88) Beesley and McLaren, 111 2002 22/26 (85) 12/12 (100) 0/1 (0) Song et al, 133 2011 427/441 (97) Rossi et al, 120 2013 15/15 (100) 0/9 (0) de Matos et al, 110 2005 61/84 (73) 13/25 (52) 8/38 (21) 0/2 (0) Weber et al, 138 2004 22/24 (92) 4/9 (44) Saggiorato et al, 139 2004 26/26 (100) 39/39 (100) Orlandi et al, 140 1998 18/18 (100) 17/17 (100) Saggiorato et al, 141 2001 17/17 (100) Savin et al, 103 2008 126/147 (86) 10/18 (56) Bartolazzi et al, 142 2008 67/86 (78) 11/15 (73) 5/7 (71) 18/22 (82) Torregrossa et al, 122 2007 152/200 (76) 90/138 (65) Volante et al, 104 2004 31/32 (97) g 66/70 (94) Volante et al, 117 2004 46/49 (94) Gaffney et al, 143 2003 55/60 (92) 14/21 (67) Coli et al, 144 2002 28/28 (100) 10/10 (100) 3/5 (60) 1/1 (100) 1/1 (100) Nascimento et al, 145 2001 9/11 (82) 11/14 (79) 10/17 (59) Papotti et al, 124 2005 13/14 (93) Saleh et al, 95 2010 18/20 (90) 10/12 (83) 18/22 (82) Bartolazzi et al, 146 2001 195/201 (97) 54/57 (95) 13/20 (65) 18/20 (90) 13/13 (100) Abbreviations: FA, follicular adenoma; FTC, follicular thyroid carcinoma; FVPTC, follicular variant papillary thyroid carcinoma; Graves, Graves disease; HA, Hürthle cell adenoma; HCC, Hürthle cell carcinoma; LTis, lymphocytic thyroiditis; MC, medullary thyroid carcinoma; NG, nodular goiter; NL, normal thyroid tissue; PDC, poorly differentiated thyroid carcinoma; PTC, papillary thyroid carcinoma; UDC, undifferentiated thyroid carcinoma (anaplastic carcinoma). a Zhu et al, 126 2010: GAL-3 expressed in 92% (142 of 155) of PTCs. b Barut et al, 105 2010: GAL-3 expressed in 72% (47 of 65) of malignant thyroid tumors. c Wiseman et al, 127 2008: GAL-3 expressed in 90% (89 of 99) of malignant thyroid tumors; 29% (29 of 100) of benign thyroid lesions. d Asa, 128 2005: CK19 is a marker of malignancy; inflamed NL and NG also showed reactivity. e Saggiorato et al, 139 2004; Orlandi et al, 140 1998; and Saggiorato et al, 141 2001: Studies were conducted by using fine-needle aspiration samples, with follow-up histologic tissue confirmation. Data provided in this table are study results on tissue. f Included NG and FA. g Hürthle cell variant PTC. atypia showed lack of TROP2 expression. The staining results are summarized in Table 5. The normal thyroid tissue on TMA revealed weak cytoplasmic staining; no membranous staining pattern was identified. The 10 surgical cases each of nodular hyperplasia and chronic LTis showed no TROP2 expression, except rare weak to moderate membranous staining in the lining cells of a degenerative cyst in 1 of the 10 cases of chronic LTis, as illustrated in Figure 1, E and F. Our data suggest that TROP2 is a potential novel immunomarker for identification of both classic and follicular variants of PTC. Based on our study, TROP2 appears to be a more specific marker than the 3 traditional markers (CK19, HBME-1, and GAL-3); however, additional studies in a large number of difficult cases from multiple institutions are warranted to substantiate the current findings. Thyroperoxidase Thyroperoxidase is a thyroid-specific enzyme reflecting normal thyroid function. Thyroperoxidase expression is demonstrated in normal thyroid follicular epithelial cells, usually in diffuse fashion. 103,164 During thyroid cell dedifferentiation, the expression of TPO is lost. Therefore, lack of TPO expression is regarded as a marker of malignancy. Several studies 103,120,164 169 investigated the diagnostic utility of TPO in thyroid lesions, yielding a sensitivity of 90% for PTC and 76% for FTC and a specificity of 88%. Table 6 summarizes the data from reported studies of TPO expression in thyroid lesions. The earlier studies produced more promising results than recent ones. Thyroperoxidase appears highly sensitive for PTC but borderline to poorly sensitive for FTC. The staining patterns of FTC and FA showed a certain degree of overlapping without meaningful discriminatory value. A couple of recent studies 103,129 used a combination of TPO and GAL-3 to improve diagnostic accuracy. They yielded a sensitivity of 96% to 98% for PTC and 44% to 67% for FTC, proposing that the application of TPO immunostain in combination with other biomarkers is useful, and with added value in the diagnosis of PTC. OTHER CELL ADHESION MOLECULES, CELL CYCLE REGULATORY PROTEINS, AND ONCOGENES E-cadherin, a calcium-dependent transmembrane cell adhesion molecule, is required for normal epithelial function. Down-regulation of E-cadherin expression has been observed in various carcinomas and is usually associated with an advanced stage and progression. In normal and benign thyroid lesions, high expression of E- cadherin was demonstrated and restricted to thyrocytes. 170 172 In thyroid carcinomas, E-cadherin expression was reduced in PDC, lost in UDC, and preserved in most 72 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin

MC, FA, HA, Table 3. NG, Extended LTis, Graves, NL, 1/3 (33) 1/18 (6) 6/13 (46) 1/35 (3) 3/54 (6) 9/49 (18) 2/21 (10) 16/29 (55) 1/14 (7) 0/59 (0) 0/12 (0) 0/14 (0) 1/10 (10) 5/15 (33) 7/11 (64) 8/19 (42) 0/17 (0) 0/41 (0) 3/50 (6) 1/2 (50) 2/20 (10) 3/8 (38) 77/151 (51) f 0/5 (0) 2/18 (11) 7/10 (70) 1/12 (8) 0/170 (0) 4/13 (31) 2/105 (2) 3/29 (10) 4/52 (8) 4/28 (14) 0/142 (0) 19/176 (11) 3/126 (2) 6/50 (12) 6/45 (13) 0/13 (0) 14/15 (93) 3/14 (21) 17/27 (63) 7/25 (28) 1/9 (11) 1/14 (7) 0/30 (0) 0/11 (0) 0/18 (0) 3/15 (2) 19/46 (41) 8/52 (15) 3/7 (43) 4/125 (3) 1/7 (14) 0/50 (0) 2/29 (7) 0/1 (0) 0/75 (0) minimally invasive carcinomas. 170 175 These studies suggest that loss of E-cadherin expression is the crucial event in dedifferentiation, progression, and metastatic spread of thyroid carcinomas. Fibronectins (FNs), extracellular matrix proteins produced by fibroblasts, are involved in cell adhesion, migration, and tumor progression. Oncofetal FNs, isoforms of FN, are highly expressed in fetal and neoplastic tissues, especially PTCs. 176 180 Prasad et al 108 investigated the expression of FN-1, CITED-1, HBME-1, and CK19 in various thyroid tissues by immunohistochemical analysis and revealed the expression of FN-1 in 91% (61 of 67) of PTCs, 100% (4 of 4) of UDCs, 50% (3 of 6) of FTCs, and 75% (6 of 8) of Hürthle cell carcinomas; in contrast, none of the normal (n ¼ 59) or benign thyroid lesions (n ¼ 43) expressed FN-1, with the exception of 2 of 29 cases of nodular goiter, which showed cytoplasmic and membranous expression of FN-1 in thyrocytes in association with fresh hemorrhage and fibrin deposition. Liu et al 109 evaluated a panel of immunomarkers on 100 benign thyroid tissues and 77 thyroid carcinomas on TMAs, revealing overexpression of FN-1 in 96% of PTCs and 86% of FTCs, suggesting that a panel consisting of GAL-3, FN-1, and intracellular sodium/iodide symporter demonstrates 98% accuracy in differentiating FA from malignant thyroid follicular lesions. Cluster of differentiation (CD) 44v6 is the isoform of CD44, a cell surface membrane glycoprotein that plays a role in the regulation of cell-cell and cell-matrix interactions as well as cell migration. 181 Expression of CD44v6 has been reported in several carcinomas. 182 185 Few studies investigating CD44v6 expression in thyroid lesions found overexpression of CD44v6 in well-differentiated PTCs and FTCs, suggesting the potential utility of CD44v6 in combination with GAL-3 in distinguishing benign from malignant thyroid neoplasms. 186 191 However, CD44v6 expression was also demonstrated in benign lesions and poorly differentiated or undifferentiated thyroid carcinomas at a lower rate of expression. Cyclin-dependent kinase inhibitor 1B (p27 Kip1 ), a nuclear cyclin-dependent kinase inhibitor, plays a major role in controlling progression from the G1 to the S phase of the cell cycle. The loss of regulatory control of the cell cycle, leading to unrestrained cell proliferation, is the hallmark of cancers. Down-regulation of p27 Kip1 has been shown to correlate with histologic loss of differentiation or high-grade morphology in various human carcinomas 192 196 and is emerging as an important prognostic factor. Studies 197,198 revealed strong nuclear reactivity in normal thyrocytes. In thyroid neoplasms, p27 Kip1 expression was significantly higher in benign lesions than malignant thyroid neoplasms, suggesting the potential utility of p27 Kip1 in the differentiation of benign from malignant follicular lesions. 197 201 Cyclin D1, a member of the family of cyclins, is a 36-kDa nuclear protein that functions as the regulatory subunit of cyclin-dependent kinases (CDK4 and CDK6) and controls the progression of the G1/S phases of the cell cycle. 202 Deregulation of cyclin expression results in the loss of control of normal cell growth and oncogenesis. Overexpression of cyclin D1 has been demonstrated in various human cancers, including thyroid carcinomas. 201,203 209 Normal thyrocytes are immunohistochemically negative for cyclin D1. Studies have been conducted to investigate the expression of cyclin D1 in thyroid neoplasms. Melck et al 210 evaluated the expression of cell cycle regulators, including cyclin D1, in TMAs of 100 benign and 105 malignant thyroid lesions. They demonstrated overexpression of cyclin D1 in 87.1% of malignant and 45.7% of benign thyroid lesions; similar findings were reported by Seybt et al 211 and Park et al. 102 Wang et al 212 investigated cyclin D1 expression in 34 conventional PTCs, 10 minimally invasive Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin 73

Figure 1. A through F, Trophoblastic cell surface antigen 2 (TROP2) staining pattern in thyroid neoplasm and lesions. A, Papillary thyroid carcinoma (PTC), follicular variant. B, Papillary thyroid carcinoma, follicular variant, diffuse (4þ) TROP2 staining, membranous pattern. Ninety percent of PTCs on tissue microarray sections express TROP2 in a membranous staining pattern. C, Follicular thyroid carcinoma (FTC). D, Follicular thyroid carcinoma, focal (1þ) strong cytoplasmic staining for TROP2. Follicular neoplasms (FTC and follicular adenoma [FA]) showed no TROP2 expression. Only 2 of 51 FAs and 4 of 37 FTCs showed focal (1þ) cytoplasmic staining without membranous pattern. E, Focal cystic degeneration in a case of lymphocytic thyroiditis (LTis). F, Focal membranous staining for TROP2 in the lining cells of the cyst. All 20 cases of benign thyroid lesions (10 Ltis,10 nodular goiter) did not show TROP2 staining (hematoxylin-eosin, original magnifications 3400 [A, C] and 3200 [E]; original magnifications 3400 [B, D] and 3200 [F]). 74 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin

Table 4. Trophoblastic Cell Surface Antigen (TROP2) Expression in 136 Cases of Thyroid Neoplasms Diagnosis 1þ 2þ 3þ 4þ Total Positive Cases (%) PTC (n ¼ 48) 5 10 18 10 43 (90) FA (n ¼ 51) 0 0 0 0 0 (0) FTC (n ¼ 37) 0 0 0 0 0 (0) Abbreviations: FA, follicular adenoma; FTC, follicular thyroid carcinoma; PTC, papillary thyroid carcinoma. Table 5. Trophoblastic Cell Surface Antigen (TROP2) Expression in 61 Cases of Atypical Follicular Lesions Diagnosis (n) TROP2, CK19, HBME-1, GAL-3, PTC (33) 23 (70) 22 (67) 33 (100) 33 (100) AFN (11) 0 (0) 2 (18) 9 (82) 8 (73) ANFNA (17) 0 (0) 3 (18) 7 (41) 6 (35) Abbreviations: AFN, atypical follicular neoplasm; ANFNA, adenomatoid nodule with focal nuclear atypia; CK19, cytokeratin 19; GAL-3, galectin-3; HBME-1, Hector Battifora mesothelial-1; n, number of cases; PTC, papillary thyroid carcinoma. Source, y Table 6. Summary of Studies With Thyroperoxidase (TPO), Monoclonal Antibody 47 Cutoff, a % PTC, FTC, UDC, No. (%) MC, No. (%) FA, HA, No. (%) NG, No. (%) Graves, Weber et al, 138 2004 5 12/24 (50) b 8/9 (89) c 13/13 (100) Savin et al, 103 2008 50 13/147 (9) 10/18 (56) 23/28 (82) 94/114 (83) Yousaf et al, 165 2008 80 0/17 (0) 3/6 (50) d 0/3 (0) 0/1 (0) 30/31 (97) 67/67 (100) Christensen et al, 166 2000 80 0/27 (0) e 25/26 (96) 71/71 (100) f De Micco et al, 167 1994 a 80 0/9 (0) 50/60 (83) 8/17 (47) De Micco et al, 164 1991 g 80 4/43 (9) 7/22 (32) 49/50 (98) 10/10 (100) 10/10 (100) de Micco et al, 168 2008 i 80 0/59 (0) 0/32 (0) 39/54 (72) 51/55 (93) Savin et al, 169 2006 80 11/52 (21) 24/40 (60) j 10/10 (100) Abbreviations: FA, follicular adenoma; FTC, follicular thyroid carcinoma; Graves, Graves disease; HA, Hürthle cell adenoma; MC, medullary thyroid carcinoma; NG, nodular goiter; NL, normal thyroid tissue; PTC, papillary thyroid carcinoma; UDC, undifferentiated thyroid carcinoma (anaplastic carcinoma). a Cutoff: The percentage of cells showing immunoreactivity to define positive. b Seven of 12 TPO-positive cases showed 3þ, diffuse staining; 5 of 12 were 3þ with fewer than 50% cells being positive. If 80% cutoff used, which is used by all other studies, the positive rate would be 7 of 24 (29%). c Seven of 8 TPO-positive cases showed 3þ, diffuse staining; 1 of 8 was 1þ with fewer than 50% cells being positive. If 80% cutoff used, the positive rate would be 7 of 9 (78%). d The 3 TPO-positive cases are minimally invasive follicular carcinomas with Hürthle cell morphology. e Twenty-seven follicular malignant cases; not specified whether PTC or follicular carcinoma. f Seventy-one benign thyroid tissues, not specified. g De Micco et al, 164 1991: Fine-needle aspiration specimens with confirmed diagnoses by surgical follow-up histology. Study was performed on half of the slides. h Number of cases was not specified. i de Micco et al, 168 2008: Study conducted on direct smears. j The 40 cases of follicular adenomas include cases with Hürthle cell morphology. NL, No. (%) 100 h FTCs, and 32 more aggressive thyroid carcinomas. Their study demonstrated overexpression of cyclin D1 in most aggressive thyroid carcinomas. Erickson et al 213 evaluated the expression of Ki-67 and cyclin D1 in Hürthle cell neoplasms (59 Hürthle cell adenomas, 55 Hürthle cell carcinomas, and 14 Hürthle cell neoplasms of uncertain malignant potential). This revealed overexpression of cyclin D1 in 18% of Hürthle cell carcinomas, compared with only 1.7% of the adenomas and none of the uncertain malignant potential tumors. It has been well demonstrated that cyclin D1 is overexpressed in malignant thyroid lesions compared to benign lesions; however, the staining patterns showed heterogeneity, which limits its utility as a diagnostic marker. b-catenin, a 92-kDa multifunctional protein, plays an important role in cell adhesion and signal transduction and serves as a downstream effector in the Wnt signaling pathway. 214 In normal resting cells, b-catenin forms cytoplasmic/membranous-bound complexes with E-cadherin. Upon activation, b-catenin translocates to the nucleus, promoting tumor growth through activation of the Wnt signaling pathway. 93,215 217 Normal thyroid follicular cells display a strong membranous immunoreactivity for b- catenin. 218,219 Studies were conducted to investigate the expression and the pattern of expression of b-catenin in various thyroid tumors. Garcia-Rostan et al 218 reported nuclear expression of b-catenin in 32.1% (9 of 28) of PDCs, 79.3% (23 of 29) of UDCs, and 0% (0 of 58) of the welldifferentiated carcinomas. Rossi et al 120 reported b-catenin expression in 80% (12 of 15) of PDCs but 0% (0 of 9) of UDCs. Ishigaki et al 220 observed cytoplasmic immunoreac- Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin 75

Table 7. Correlation of BRAF V600E Mutation in Papillary Thyroid Carcinoma by Immunohistochemistry (IHC) and Molecular Testing Source, y Antibody IHC Result Molecular Testing Comment Capper et al, 233 2011 VE1 9 cases positive by IHC 9 cases positive by PCR amplification and direct sequencing Koperek et al, 232 2012 VE1 39 cases positive by IHC 39 cases positive by gene sequencing Bullock et al, 231 2012 VE1 68 cases positive by IHC 59 cases positive for BRAF by Sanger sequencing Zagzag et al, 230 2013 VE1 25 of 28 BRAF-mutated cases positive by IHC 28 cases positive for BRAF by direct sequencing McKelvie et al, 228 2013 VE1 50 cases positive by IHC 41 cases positive for BRAF by C-PCR Routhier et al, 227 2013 VE1 13 cases positive by IHC 13 cases positive for BRAF by SNaPshot genotyping assay Routhier et al, 227 2013 Anti B-Raf mutation mouse antibody 16 cases positive by IHC including 3 cases that were negative by SNaPshot. 13 cases positive for BRAF by SNaPshot genotyping assay Abbreviations: BRAF, B-isoform of RAF kinase; C-PCR, competitive PCR; PCR, polymerase chain reaction. 100% sensitivity and specificity for IHC; 3 additional IHC-positive cases; molecular testing failed. 100% sensitivity and specificity; 1 strong IHC-positive cases was negative by molecular testing. In 11 discordant cases, further testing favored the original IHC results in 8 cases, suggesting that IHC was a more sensitive method than Sanger sequencing. Sensitivity 89% and specificity 100% by IHC. 8 of 9 IHC-positive and C-PCR negative cases were positive for BRAF by SNaPshot a method, suggesting that IHC was more sensitive than C-PCR method. 100% sensitivity and specificity by IHC. 100% sensitivity and 70% specificity by IHC. tivity for b-catenin in 67% (52 of 78) of PTCs, 25% (5 of 20) of FTCs, and only 9% (3 of 34) of FAs. The association of aberrant nuclear b-catenin expression and poor prognosis was observed by some investigators. 218 p53 is a tumor suppressor gene product that plays an important role in normal cell growth. Mutations of the TP53 gene lead to accumulation of p53, which can be detected by immunohistochemistry. p53 expression has been reported in various tumors, mainly in UDCs and PDCs and rarely in well-differentiated carcinomas as well as MCs in thyroid neoplasms. 220 226 BRAF Mutation-Specific Antibody The BRAF (the B-isoform of RAF kinase) oncogene is mutated in several types of tumors, such as colorectal adenocarcinoma, PTC, glioma, gastrointestinal adenocarcinoma, melanoma, and pulmonary adenocarcinoma. 227 The most common mutation in BRAF is due to a T to A switch at position 1796, which results in an alteration from valine to glutamate at the V600E. The BRAF V600E point mutation has been reported in approximately 50% of PTCs, with a higher frequency (approximately 70%) in tall cell variant and oncocytic variant of PTC and a much lower frequency (approximately 20%) in follicular variant of PTC. 228 The BRAF V600E mutation is generally negative in benign follicular lesions, normal thyroid tissue, medullary carcinoma, and follicular carcinoma. A recent meta-analysis of 5655 patients suggested PTC with the BRAF mutation is associated with a higher risk of recurrent, persistent disease, lymph node metastasis, and extrathyroidal extension. 229 Many molecular techniques have been used to detect the BRAF V600E point mutation, including single-strand conformation polymorphism, mutation-specific polymerase chain reaction, direct gene sequencing, and colorimetric mutation analysis. These methods tend to be expensive, time-consuming, labor-intensive, and difficult to validate and implement in some clinical settings. Recently, 2 mutation-specific antibodies against BRAF V600E have become commercially available; one is VE1 clone (Spring Bioscience, Pleasanton, California) and the other is anti B-Raf mouse monoclonal antibody (New East Bioscience, Malvern, Pennsylvania). Most studies used VE1 clone, and only rare studies used anti B-Raf mouse monoclonal antibody. 227,228,230 233 In general, BRAF mutation specific antibody has been shown to be useful in detection of the BRAF V600 mutation, with a sensitivity and specificity of more than 95% when compared to other molecular methods. 227,228,230 233 In fact, some studies 228,231 suggested that anti BRAF mutation-specific antibody may be more sensitive than molecular testing in detecting the BRAF mutation. Data from selected studies of BRAF mutation specific antibodies to detect the BRAF V600E mutation in PTCs, with the correlating molecular testing results, are summarized in Table 7. An example of the BRAF mutation in a papillary thyroid microcarcinoma, detected by immunohistochemistry using the VE1 clone, is shown in Figure 2, A through D. CONCLUSIONS Cytomorphology or histomorphology remains the cornerstone for the diagnosis of thyroid lesions. However, not infrequently, we encountered cases with equivocal morphology that posed a great challenge in reaching an accurate diagnosis. Ancillary studies, including somatic mutation testing, messenger RNA gene expression platforms, protein immunohistochemistry, and microrna panels, are becoming increasingly important. Immunohistochemistry by far is the most commonly used method to complement morphologic assessment. By reviewing the published data in the current literature, we have determined that there is no individual biomarker having sufficient sensitivity or speci- 76 Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin

Figure 2. A through D, Immunohistochemical detection of BRAF expression in a papillary thyroid microcarcinoma and its metastasis in a lymph node, using anti-ve1 antibody. A, Papillary thyroid microcarcinoma. B, Diffuse (4þ) BRAF cytoplasmic staining. C, Metastatic papillary thyroid microcarcinoma in a lymph node. D, Diffuse BRAF cytoplasmic staining in the metastatic thyroid carcinoma (hematoxylin-eosin, original magnification 3200 [A and C]; original magnification 3400 [B and D]). ficity to distinguish benign from malignant lesions. However, HBME-1 is often strongly and diffusely expressed in PTC; it can be used as a marker (in a panel) to aid in the PTC diagnosis. Galectin-3 (in a panel) can be useful in differentiating malignant from benign thyroid lesions; CK19 has low sensitivity and specificity. A panel of immunomarkers, as proposed by many investigators, is recommended when working on challenging thyroid follicular-derived lesions to improve diagnostic accuracy. The most commonly proposed panel is GAL-3, HBME-1, and CK19. We also recommend including TROP2 in the diagnostic panel. Other panels, such as the combination of 2 of the 3 markers or the addition of FN-1 or CITED-1, are also proposed. The application of a panel of immunomarkers improves the differential power of individual markers and aids in the accurate classification of challenging thyroid follicular-derived lesions. The authors would like to thank Melissa Erb, AAS, for her outstanding secretarial support, Tina Brosious, HT(ASCP), and Erin Powell, HT(ASCP), for construction of TMA blocks and cutting TMA sections, Jianhui Shi, MD, PhD, and Angie Bitting, HT(ASCP), QIHC, for their assistance with immunostains, and Kathy Fenstermacher, BA, for editing this manuscript. References 1. National Cancer Institute. SEER Stat Fact Sheets: thyroid cancer. http://seer. cancer.gov/statfacts/html/thyro.html. Accessed May 21, 2013. 2. Morris LG, Sikora AG, Tosteson TD, Davies L. The increasing incidence of thyroid cancer: the influence of access to care. Thyroid. 2013;23(7):885 891. 3. DeLellis RA, Lloyd RV, Heitz PU, et al, eds. Pathology and Genetics of Tumours of Endocrine Organs. Lyon, France: IARC Press; 2004. World Health Organization Classification of Tumours; vol 8. 4. Hirokawa M, Carney JA, Goellner JR, et al. Observer variation of encapsulated follicular lesions of the thyroid gland. Am J Surg Pathol. 2002; 26(11):1508 1514. 5. Franc B, de la Salmonière P, Lange F, et al. Interobserver and intraobserver reproducibility in the histopathology of follicular thyroid carcinoma. Hum Pathol. 2003;34(11):1092 1100. 6. Lloyd RV, Erickson LA, Casey MB, et al. Observer variation in the diagnosis of follicular variant of papillary thyroid carcinoma. Am J Surg Pathol. 2004; 28(10):1336 1340. 7. Elsheikh TM, Asa SL, Chan JK, et al. Interobserver and intraobserver variation among experts in the diagnosis of thyroid follicular lesions with borderline nuclear features of papillary carcinoma. Am J Clin Pathol. 2008; 130(5):736 744. 8. Mitchell PJ, Tijan R. Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins. Science. 1989;245(4916):371 378. 9. Kimura S, Hara Y, Pineau T, et al. The T/ebp null mouse: thyroid-specific enhancer-binding protein is essential for the organogenesis of the thyroid, lung, ventral forebrain, and pituitary. Genes Dev. 1996;10(1):60 69. Arch Pathol Lab Med Vol 139, January 2015 Immunohistochemistry in Thyroid Pathology Liu & Lin 77