- Email: [email protected]
The role of immunohistochemistry in distinguishing squamous cell carcinoma from mesothelioma and adenocarcinoma in pleural effusion
Seminars in Diagnostic Pathology (2006) 23, 15-19
The role of immunohistochemistry in distinguishing squamous cell carcinoma from mesothelioma and adenocarcinoma in pleural effusion Qing Li, MD, PhD,a Neil Bavikatty, MD,b Claire W. Michael, MDa a
From the Department of Pathology, The University of Michigan Hospitals, Ann Arbor, Michigan; and the Department of Pathology, Munson Medical Center, Traver City, Michigan.
b
KEYWORDS Immunocytochemistry; Pleural effusions; Malignant mesothelioma; Squamous cell carcinoma; Adenocarcinoma
BACKGROUND Distinguishing metastatic squamous cell carcinoma (SCC) from malignant mesothelioma (MM) and adenocarcinoma (ADC) in pleural effusions may be particularly challenging by routine cytologic stains. We explored the utility of using a panel of six antibodies to differentiate SCC from MM and ADC. DESIGN 33 cases of pleural cytologic preparations retrieved from our archives consisted of 9 cases of SCC, 12 cases of epithelial MM, and 12 cases of adenocarcinoma of lung. Cell blocks were prepared by the thrombin clot technique followed by formalin-fixation and paraffin-embedding. Tissue sections of 4 m were stained with hematoxylin and eosin and the immunoperoxidase method visualized by the biotin-streptavidin-peroxidase system. The antibodies used were cytokeratins (CAM 5.2, K903, and CK 5/6), cell membrane glycoproteins (CEA and Ber-EP4), and calretinin. In all cases, the reactivity pattern was graded on a sliding scale from 0 to 4⫹ according to the percentage of reactive cells. RESULTS SCC was positive for K903 (100%), CK 5/6 (89%), CAM 5.2 (78%), and CEA (22%), and negative for Ber-EP4 (100%) and calretinin (100%). MM was positive for calretinin (100%), CAM 5.2 (100%), K903 (92%), CK 5/6 (92%), and negative for CEA (100%) and Ber-EP4 (100%). ADC was positive for CAM 5.2 (100%), CEA (83%), and Ber-EP4 (83%), and negative for calretinin (100%), K903 (92%) and CK 5/6 (92%). CONCLUSIONS Our studies confirm the role of the above panel of antibodies in distinguishing among these malignancies. Positive staining for K903, CK 5/6, and CAM 5.2 separated SCC and MM from ADC. Positive staining for calretinin separated MM from SCC and ADC. Positive staining for glycoproteins and predominantly negative staining for CK 5/6, K903 and calretinin separated ADC from SCC and MM. © 2006 Elsevier Inc. All rights reserved.
The diagnostic difficulty of malignant mesothelioma (MM) and adenocarcinoma (ADC) in pleural effusions has been well recognized and studied.1–3 Numerous studies have reported the histological and immunohistochemical differences of these two entities.4 – 6 In contrast, metastatic squamous cell carcinoma (SCC) in effusions is a relatively
Address reprint requests and correspondence: Claire W. Michael, MD, Department of Pathology, 2G322, University of Michigan Hospitals, 1500 East Medical Center Drive, Ann Arbor, MI 48109. E-mail address: [email protected].
0740-2570/$ -see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1053/j.semdp.2006.06.007
rare entity and not well studied. Well-differentiated SCC is usually not a diagnostic difficulty due to the easily identified diagnostic features, such as evidence of keratinization. However, poorly differentiated SCC can be confused with MM or ADC. It is important to the cytopathologist to recognize this entity and to make the correct diagnosis, particularly when the primary source is unknown. Recognized diagnostic features of SCC include keratinized pleomorphic cells with dense cytoplasm and a distinct cell border, tadpole cells, polygonal cells of flat and angulated appearance, keratin pearls, anucleated keratinized
16
Seminars in Diagnostic Pathology, Vol 23, No 1
Figure 1 Squamous cell carcinoma in pleural effusion. A cellular sphere with cytoplasmic vacuolation and slightly scalloped borders (A). A cell showing endo- ectoplasmic separation (B). Papanicolaou stain (600). (Color version of figure is available online.)
squames, and so-called third type squamous cells that are round or oval in shape with dense cytoplasm and a welldefined cell border.7,8 However, cases of poorly differentiated SCC frequently do not contain the keratinized pleomorphic cells, nor anucleated keratinized squames. They predominantly contain moderate to poorly differentiated, nonkeratinized squamous cells with a more centrally located nucleus, prominent nucleoli, and occasional cytoplasmic vacuoles. These morphological features are often similar to those of MM and ADC in effusions (Figure 1). For example, their abundant, dense cytoplasm may have an amphophilic staining pattern and zones of endoplasmic and ectoplasmic demarcation simulating cells of MM. On the other hand, their cytoplasmic vacuoles and prominent nucleoli may be misinterpreted as ADC. Therefore, the recognition of SCC in pleural effusion and its separation from MM and ADC may be difficult by routine cytologic stains. Immunohistochemical stains have been routinely used in distinguishing MM from ADC.3– 6 In this study, we applied a panel of six antibody markers, including cytokeratins (CAM 5.2, K903, and CK 5/6), cell membrane glycoproteins (CEA, Ber-EP4), and calretinin-binding protein (calretinin), to evaluate their utility in the differential diagnosis of malignant effusions, particularly in separating SCC from MM and ADC.
Materials and methods Cases Thirty-three cell blocks from malignant pleural effusion were retrieved from the files of the Department of Pathology, University of Michigan Hospitals. Nine of the cases were diagnosed as poorly differentiated squamous cell carcinoma. Of these, 6 cases originated from the lung, 1 from the larynx, 1 was from the vulva, and 1 from the tongue. Twelve of the cases were diagnosed as diffuse malignant mesothelioma of epithelial type. The remain-
ing 12 cases were diagnosed as metastatic adenocarcinoma of lung primary. The cell blocks were prepared by the thrombin clot technique, were fixed in 10% neutral buffered formalin, and embedded in paraffin. Four-micrometer tissue sections were prepared and stained with hematoxylin and eosin (H&E). Smears and H&E-stained slides were reviewed to confirm the cytologic diagnosis and the presence of malignant cells.
Immunohistochemistry The immunoperoxidase studies were performed on deparaffinized, unstained sections using commercially available 6 antibodies: CAM 5.2 (Becton Dickinson, San Jose, CA), K903, CEA (polyclonal) and Ber-EP4 (DAKO, Carpinteria, CA), CK 5/6 (Chemicon International, Temecula, CA), and calretinin (Zymed, San Francisco, CA). The primary antibody was incubated for 30 minutes followed by an incubation of the secondary antibody for 30 minutes. Then, the chromogen diaminobenzidine and streptavidin-biotin labeling method was used for visualization (Dako Corporation). All the immunohistochemistry stains were performed by Dako autostainer ((Dako Corporation). For each of the antibodies, the endogenous peroxidase activity was blocked with a 5-minute incubation of peroxidase inhibitor. Table 1 represents the summary of antibodies used in this study and antigen retrieval methods for each antibody used in this study to enhance the signal. A negative control consisting of omission of the primary antibody was included with each antibody tested. The pattern of reactivity was based on cytoplasmic membrane staining for CAM 5.2, K903, CK 5/6, CEA, and Ber-EP4, and both nuclear and cytoplasmic membrane staining for calretinin. The reactive intensity of each antibody was graded semiquantitatively using a sliding scale from 0 to 4 as follows: no cell staining ⫽ 0, ⬍25% ⫽ 1⫹, 25-50% ⫽ 2⫹, 50-75% ⫽ 3⫹, and ⬎75% ⫽ 4⫹. A grade of 1⫹ or above was considered positive.
Li et al Table 1 Antibody CAM 5.2
Distinguishing SCC from MM and ADC The source and dilution of antibody in the study Company
Becton Dickinson K903 DAKO CEA DAKO Ber-EP4 DAKO Calretinin Zymed CK 5/6
17
Catalog No.
Dilution Pretreatment
349205
None
Proteinase K
M0630 A00115 M0804 18-0211
1:100 1:3000 1:100 1:200
Proteinase K None Enzymel 10mM citrate Microwave TRIS
Chemicon Intl MAB 1620 1:100
Abbreviations: CEA, carcinoembryonic antigen; EMA, epithelial membrane antigen; Ber-EP4, epithelial antigen clone BerEP4.
Results
Figure 3 Squamous cell carcinoma with intense cytoplasmic staining (Cytokeratin 5/6, 1000⫻). (Color version of figure is available online.)
Discussion
In metastatic SCC, 9/9 (100%) cases were positive for K903, and 8/9 (89%) cases were positive for CK 5/6 (Figures 2 and 3). CAM 5.2 (7/9 cases, 78%) were positive. CEA (2/9 cases, 22%) cases were positive. All of the cases were negative for Ber-EP4 (100%) and calretinin (100%). Of the malignant mesotheliomas, all 12 cases were positive for calretinin (100%) and CAM 5.2 (100%) (Figure 4). Eleven cases were positive for K903 (92%) and CK 5/6 (92%) (Figures 5 and 6). All cases were negative for CEA (100%) and Ber-EP4 (100%). Among the 12 cases of metastatic adenocarcinoma, all cases were positive for CAM 5.2 (100%). Ten cases were positive for CEA (83%) and Ber-EP4 (83%) (Figure 7). All cases were negative for calretinin (100%). Eleven cases were negative for CK 5/6 (92%). The overall immunohistochemistry staining results of CAM 5.2, K903, CK 5/6, CEA, Ber-EP4, and calretinin in SCC, MM, and ADC are summarized in Table 2. The control staining was satisfactory each time and not shown here.
The difficulty in distinguishing poorly differentiated SCC from MM and ADC is a well-recognized problem. In malignant effusions, the diagnostic difficulty is compounded by the fact that SCC frequently manifests considerable overlap with either adenocarcinoma (cellular spheres, lack of keratinization, prominent nucleoli, and cytoplasmic vacuolization) or malignant mesothelioma (cellular spheres, prominent nucleoli, and ecto-endoplasmic demarcation). In our experience, frequently, a malignant effusion is worked up for the differential diagnosis of ADC versus MM. Generally, a limited panel of epithelial markers, such as CEA, B72.3, Ber-EP4, Leu-M1, and one of the mesothelial markers such as calretinin, HBME-1, or WT-1 are used. The outcome in cases of SCC could be confusing as the stains could all be negative resulting in either additional workup of the patient, assuming poor preservation of sample, or sometimes a false diagnosis based on cytologic features only. Immunohistochemistry can greatly aid in solving such diagnostic dilemmas; however, currently available markers have different sensitivities and specificities for cells of epithelial or mesothelial differentiation.9 Several studies
Figure 2 Squamous cell carcinoma with intense cytoplasmic staining (K903, 1000⫻). (Color version of figure is available online.)
Figure 4 Malignant mesothelioma with nuclear and membranous staining (Calretinin, 1000⫻). (Color version of figure is available online.)
18
Figure 5 Malignant mesothelioma with cytoplasmic staining (K903, 1000⫻). (Color version of figure is available online.)
have evaluated the separation of these entities by using immunohistochemical staining mainly focusing on the differential diagnosis between MM and ADC.6 To our knowledge, there is no such study that has compared the immunohistochemical staining patterns of metastatic SCC with MM and ADC in effusion. In this study, we used a panel of antibody markers, including several different keratins, cell membrane glycoprotein CEA and Ber-EP4 and Calretinin, to evaluate the reactive patterns of these antibody markers among SCC, MM, and ADC. Keratin antibodies have been extensively studied in distinguishing epithelial malignancy from mesothelioma. In summary, it has been reported that antibodies against high molecular weight keratins will react with most of MM and relatively few ADC; antibodies against low molecular weight keratins will react with both of those tumor groups.9,10 Recently, Ordonez reported that pulmonary adenocarcinomas were nonreactive for CK 5/6, whereas MM were positive.11 Our results demonstrated that SCCs were positive for both high and low molecular weight keratins, including CK5/6, and were mostly negative for other markers. ADCs were characteristically positive for CAM 5.2, Ber-EP4 and CEA, but predominantly negative for K903 and CK5/6. Similar to SCC, MMs were also positive for CAM 5.2, K903 and CK 5/6, however, they were also positive for calretinin.
Seminars in Diagnostic Pathology, Vol 23, No 1
Figure 7 Adenocarcinoma with distinct membranous staining (Ber-EP4, 1000⫻). (Color version of figure is available online.)
Calretinin is a cytoplasmic calcium-binding protein and is expressed in MM. It has been widely accepted that calretinin is one of the most positive markers for MM,12,13 although there are conflicting reports regarding the specificity of calretinin to mesothelioma. Some studies have shown that calretinin is also expressed in a few ADC particularly those of ovarian origin.12 Our results demonstrated that calretinin was 100% positive in MM; it was negative in all SCC and pulmonary ADC cases. There is a potential pitfall in the interpretation of calretinin staining patterns, since background benign mesothelial cells are positive for calretinin. In our ADC cases, all the benign reactive mesothelial cells are highlighted by calretinin and all malignant tumor cells are negative for calretinin (Figure 8). When we interpreted the staining pattern of calretinin, we excluded the background staining of benign mesothelial cells. Thus, the positive reactivity of calretinin separates MM from SCC and ADC. CEA has been considered as one of the most reliable markers to distinguish ADC from MM,14 with most of the MM negative for CEA.15 CEA is an oncoprotein in the cells of endodermal origin and expressed in carcinomas of the lung and the gastrointestinal tract. Wang and coworkers were the first to report that mesotheliomas fail to express CEA as determined by the immunoperoxidase method.16 Corson and Pinkus observed that some epithelial mesothe-
Table 2 Immunohistochemistry reactive patterns in malignant effusions
Figure 6 Malignant mesothelioma with membranous staining (Cytokeratin 5/6, 1000⫻). (Color version of figure is available online.)
Antibody
SCC (n ⫽ 9)
MM (n ⫽ 12)
ADC (n ⫽ 12)
CAM 5.2 K903 CK 5/6 CEA Ber-EP4 Calretinin
7 (78%) 9 (100%) 8 (89%) 2 (22%) 0 (0%) 0 (0%)
12 (100%) 11 (92%) 11 (92%) 0 (0%) 0 (0%) 12 (100%)
12 (100%) 1 (8%) 1 (8%) 10 (83%) 10 (83%) 0 (0%)
Abbreviations: SCC, squamous cell carcinoma; MM, malignant mesothelioma; ADC, adenocarcinoma; CEA, carcinoembryonic antigen; EMA, epithelial membrane antigen; Ber-EP4, epithelial antigen clone BerEP4.
Li et al
Distinguishing SCC from MM and ADC
19
References
Figure 8 Adenocarcinoma staining negative while the benign reactive mesothelial cells are positive (Calretinin, 1000⫻). (Color version of figure is available online.)
liomas stained weakly for CEA by using polyclonal antiCEA antibody, but the intensity was much less than that observed in adenocarcinoma.14 Monoclonal antibodies to CEA have been reported to give cleaner backgrounds, less nonspecific cross-reactivity, and higher specificity for adenocarcinoma, with no false positives in 28 cases of pleural mesothelioma.17 However, the sensitivity of this antibody was reduced with only 72% (36 of 50) of positivity in adenocarcinoma.18 In this study, we used polyclonal antiCEA antibodies. This may explain the positive staining of some SCC. When we interpreted the staining pattern of mesothelioma, we excluded the background staining, since inflammatory cells and areas of necrosis may stain positive with polyclonal anti-CEA antibody. Ber-EP4 is another epithelial marker and initially thought to be specific for ADC18,19; however, it is now known that 10% to 20% of MM has focal positivity for this marker.20,21 In our study, ADC was 83% positive for CEA and 83% positive for Ber-EP4, with all ADC staining at least with one of them. Both of CEA and Ber-EP4 were negative for MM. All cases of SCC were negative for Ber-EP4. Our results demonstrate a high positive reaction pattern toward CEA and Ber-EP4 in the cases of ADC, negative reaction pattern in the cases of MM, and predominantly negative pattern in SCC. In summary, the differential diagnosis of SCC from mesothelioma and adenocarcinoma may be challenging on routine cytologic stains. To support any subtle morphologic distinctions among these entities, a panel of antibody markers can be applied to a prepared cell block. Our studies demonstrate that negative staining for calretinin and glycoproteins separated SCC from MM and ADC. Positive staining for calretinin separated MM from SCC and most ADC. Positive staining for glycoproteins and predominantly negative staining for CK 5/6, K903 and calretinin separated ADC from SCC and MM. We also recommend that the diagnosis of MM should not be based solely on a positive CK 5/6 and negative epithelial markers to avoid missing a SCC.
1. Shield PW, Callan JJ, Devine PL: Markers for metastatic adenocarcinoma in serous effusion specimens. Diagn Cytopathol 11:237-245, 1994 2. Matsuo T, Ito H, Anami M, et al: Malignant peritoneal mesothelioma with squamous metaplasia. Cytopathology 4:373-378, 1993 3. Ordonez NG: The immunohistochemical diagnosis of epithelial mesothelioma. Hum Pathol 30:313-323, 1999 4. Motoyama T, Watanabe T, Okazaki E, et al: Immunohistochemical properties of malignant mesothelioma cells in histologic and cytologic specimens. Acta Cytol 39:164-170, 1995 5. Moran CA, Wick MR, Suster S: The role of immunohistochemistry in the diagnosis of malignant mesothelioma. Semin Diagn Pathol 17:178183, 2000 6. Brockstedt U, Gulyas M, Dobra K, et al: An optimized battery of eight antibodies that can distinguish most cases of epithelial mesothelioma from adenocarcinoma. Am J Clin Pathol 114:203-209, 2000 7. Smith-Purslow MJ, Kini SR, Naylor B: Cells of squamous cell carcinoma in pleural, peritoneal and pericardial fluids. Origin and morphology. Acta Cytol 33:245-253, 1989 8. Frost JK: The cell in health and disease. An evaluation of cellular morphologic expression of biologic behavior. Karger, George L. Wied, 1986, pp 165-186 9. Otis CN, Carter D, Cole S, et al: Immunohistochemical evaluation of pleural mesothelioma and pulmonary adenocarcinoma. Am J Surg Pathol 11:445-456, 1987 10. Chu PG, Weiss LM: Expression of cytokeratin 5/6 in epithelial neoplasms: an immunohistochemical study of 509 cases. Mod Pathol 15:6-10, 2002 11. Ordonez NG: Value of cytokeratin 5/6 immunostaining in distinguishing epithelial mesothelioma of the pleura from lung adenocarcinoma. Am J Surg Pathol 22:1215-1221, 1998 12. Doglioni C, Tos AP, Laurino L, et al: Calretinin: a novel immunocytochemical marker for mesothelioma. Am J Surg Pathol 20:1037-1046, 1996 13. Ordonez NG: Value of calretinin immunostaining in differentiating epithelial mesothelioma from lung adenocarcinoma. Mod Pathol 11: 929-933, 1998 14. Corson JM, Pinkus GS: Mesothelioma: profile of keratin proteins and carcinoembryonic antigen: an immunoperoxidase study of 20 cases and comparison with pulmonary adenocarcinomas. Am J Pathol 108: 80-87, 1982 15. Athanassiadou P, Athanassiades P, Lazaris D, et al: Immunocytochemical differentiation of reactive mesothelial cells and adenocarcinoma cells in serous effusions with the use of carcinoembryonic antigen and fibronectin. Acta Cytol 38:718-722, 1994 16. Wang NS, Huang SN, Gold P: Absence of carcinoembryonic antigenlike material in mesothelioma. Cancer 44:937-943, 1979 17. Sheibani K, Battifora H, Burke JS: Antigenic phenotype of malignant mesotheliomas and pulmonary adenocarcinomas: an immunohistologic analysis demonstrating the value of Leu M1 antigen. Am J Pathol 123:212-219, 1986 18. Bailey ME, Brown RW, Mody DR, et al: Ber-EP4 for differentiating adenocarcinoma from reactive and neoplastic mesothelial cells in serous effusions. Comparison with carcinoembryonic antigen, B72.3 and Leu-M1. Acta Cytol 40:1212-1216, 1996 19. Riera JR, Astengo-Osuna C, Longmate JA, et al: The immunohistochemical diagnostic panel for epithelial mesothelioma: a reevaluation after heat-induced epitope retrieval. Am J Surg Pathol 21:1409-1419, 1997 20. Maguire B, Whitaker D, Carrello S, et al: Monoclonal antibody BerEP4: its use in the differential diagnosis of malignant mesothelioma and carcinoma in cell blocks of malignant effusions and FNA specimens. Diagn Cytopathol 10:130-134, 1994 21. Moch H, Oberholzer M, Dalquen P, et al: Diagnostic tools for differentiating between pleural mesothelioma and lung adenocarcinoma in paraffin embedded tissue. Part I: Immunohistochemical findings. Virchows Arch A Pathol Anat Histopathol 423:19-27, 1993
The role of immunohistochemistry in distinguishing squamous cell carcinoma from mesothelioma and adenocarcinoma in pleural effusion Qing Li, MD, PhD,a Neil Bavikatty, MD,b Claire W. Michael, MDa a
From the Department of Pathology, The University of Michigan Hospitals, Ann Arbor, Michigan; and the Department of Pathology, Munson Medical Center, Traver City, Michigan.
b
KEYWORDS Immunocytochemistry; Pleural effusions; Malignant mesothelioma; Squamous cell carcinoma; Adenocarcinoma
BACKGROUND Distinguishing metastatic squamous cell carcinoma (SCC) from malignant mesothelioma (MM) and adenocarcinoma (ADC) in pleural effusions may be particularly challenging by routine cytologic stains. We explored the utility of using a panel of six antibodies to differentiate SCC from MM and ADC. DESIGN 33 cases of pleural cytologic preparations retrieved from our archives consisted of 9 cases of SCC, 12 cases of epithelial MM, and 12 cases of adenocarcinoma of lung. Cell blocks were prepared by the thrombin clot technique followed by formalin-fixation and paraffin-embedding. Tissue sections of 4 m were stained with hematoxylin and eosin and the immunoperoxidase method visualized by the biotin-streptavidin-peroxidase system. The antibodies used were cytokeratins (CAM 5.2, K903, and CK 5/6), cell membrane glycoproteins (CEA and Ber-EP4), and calretinin. In all cases, the reactivity pattern was graded on a sliding scale from 0 to 4⫹ according to the percentage of reactive cells. RESULTS SCC was positive for K903 (100%), CK 5/6 (89%), CAM 5.2 (78%), and CEA (22%), and negative for Ber-EP4 (100%) and calretinin (100%). MM was positive for calretinin (100%), CAM 5.2 (100%), K903 (92%), CK 5/6 (92%), and negative for CEA (100%) and Ber-EP4 (100%). ADC was positive for CAM 5.2 (100%), CEA (83%), and Ber-EP4 (83%), and negative for calretinin (100%), K903 (92%) and CK 5/6 (92%). CONCLUSIONS Our studies confirm the role of the above panel of antibodies in distinguishing among these malignancies. Positive staining for K903, CK 5/6, and CAM 5.2 separated SCC and MM from ADC. Positive staining for calretinin separated MM from SCC and ADC. Positive staining for glycoproteins and predominantly negative staining for CK 5/6, K903 and calretinin separated ADC from SCC and MM. © 2006 Elsevier Inc. All rights reserved.
The diagnostic difficulty of malignant mesothelioma (MM) and adenocarcinoma (ADC) in pleural effusions has been well recognized and studied.1–3 Numerous studies have reported the histological and immunohistochemical differences of these two entities.4 – 6 In contrast, metastatic squamous cell carcinoma (SCC) in effusions is a relatively
Address reprint requests and correspondence: Claire W. Michael, MD, Department of Pathology, 2G322, University of Michigan Hospitals, 1500 East Medical Center Drive, Ann Arbor, MI 48109. E-mail address: [email protected].
0740-2570/$ -see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1053/j.semdp.2006.06.007
rare entity and not well studied. Well-differentiated SCC is usually not a diagnostic difficulty due to the easily identified diagnostic features, such as evidence of keratinization. However, poorly differentiated SCC can be confused with MM or ADC. It is important to the cytopathologist to recognize this entity and to make the correct diagnosis, particularly when the primary source is unknown. Recognized diagnostic features of SCC include keratinized pleomorphic cells with dense cytoplasm and a distinct cell border, tadpole cells, polygonal cells of flat and angulated appearance, keratin pearls, anucleated keratinized
16
Seminars in Diagnostic Pathology, Vol 23, No 1
Figure 1 Squamous cell carcinoma in pleural effusion. A cellular sphere with cytoplasmic vacuolation and slightly scalloped borders (A). A cell showing endo- ectoplasmic separation (B). Papanicolaou stain (600). (Color version of figure is available online.)
squames, and so-called third type squamous cells that are round or oval in shape with dense cytoplasm and a welldefined cell border.7,8 However, cases of poorly differentiated SCC frequently do not contain the keratinized pleomorphic cells, nor anucleated keratinized squames. They predominantly contain moderate to poorly differentiated, nonkeratinized squamous cells with a more centrally located nucleus, prominent nucleoli, and occasional cytoplasmic vacuoles. These morphological features are often similar to those of MM and ADC in effusions (Figure 1). For example, their abundant, dense cytoplasm may have an amphophilic staining pattern and zones of endoplasmic and ectoplasmic demarcation simulating cells of MM. On the other hand, their cytoplasmic vacuoles and prominent nucleoli may be misinterpreted as ADC. Therefore, the recognition of SCC in pleural effusion and its separation from MM and ADC may be difficult by routine cytologic stains. Immunohistochemical stains have been routinely used in distinguishing MM from ADC.3– 6 In this study, we applied a panel of six antibody markers, including cytokeratins (CAM 5.2, K903, and CK 5/6), cell membrane glycoproteins (CEA, Ber-EP4), and calretinin-binding protein (calretinin), to evaluate their utility in the differential diagnosis of malignant effusions, particularly in separating SCC from MM and ADC.
Materials and methods Cases Thirty-three cell blocks from malignant pleural effusion were retrieved from the files of the Department of Pathology, University of Michigan Hospitals. Nine of the cases were diagnosed as poorly differentiated squamous cell carcinoma. Of these, 6 cases originated from the lung, 1 from the larynx, 1 was from the vulva, and 1 from the tongue. Twelve of the cases were diagnosed as diffuse malignant mesothelioma of epithelial type. The remain-
ing 12 cases were diagnosed as metastatic adenocarcinoma of lung primary. The cell blocks were prepared by the thrombin clot technique, were fixed in 10% neutral buffered formalin, and embedded in paraffin. Four-micrometer tissue sections were prepared and stained with hematoxylin and eosin (H&E). Smears and H&E-stained slides were reviewed to confirm the cytologic diagnosis and the presence of malignant cells.
Immunohistochemistry The immunoperoxidase studies were performed on deparaffinized, unstained sections using commercially available 6 antibodies: CAM 5.2 (Becton Dickinson, San Jose, CA), K903, CEA (polyclonal) and Ber-EP4 (DAKO, Carpinteria, CA), CK 5/6 (Chemicon International, Temecula, CA), and calretinin (Zymed, San Francisco, CA). The primary antibody was incubated for 30 minutes followed by an incubation of the secondary antibody for 30 minutes. Then, the chromogen diaminobenzidine and streptavidin-biotin labeling method was used for visualization (Dako Corporation). All the immunohistochemistry stains were performed by Dako autostainer ((Dako Corporation). For each of the antibodies, the endogenous peroxidase activity was blocked with a 5-minute incubation of peroxidase inhibitor. Table 1 represents the summary of antibodies used in this study and antigen retrieval methods for each antibody used in this study to enhance the signal. A negative control consisting of omission of the primary antibody was included with each antibody tested. The pattern of reactivity was based on cytoplasmic membrane staining for CAM 5.2, K903, CK 5/6, CEA, and Ber-EP4, and both nuclear and cytoplasmic membrane staining for calretinin. The reactive intensity of each antibody was graded semiquantitatively using a sliding scale from 0 to 4 as follows: no cell staining ⫽ 0, ⬍25% ⫽ 1⫹, 25-50% ⫽ 2⫹, 50-75% ⫽ 3⫹, and ⬎75% ⫽ 4⫹. A grade of 1⫹ or above was considered positive.
Li et al Table 1 Antibody CAM 5.2
Distinguishing SCC from MM and ADC The source and dilution of antibody in the study Company
Becton Dickinson K903 DAKO CEA DAKO Ber-EP4 DAKO Calretinin Zymed CK 5/6
17
Catalog No.
Dilution Pretreatment
349205
None
Proteinase K
M0630 A00115 M0804 18-0211
1:100 1:3000 1:100 1:200
Proteinase K None Enzymel 10mM citrate Microwave TRIS
Chemicon Intl MAB 1620 1:100
Abbreviations: CEA, carcinoembryonic antigen; EMA, epithelial membrane antigen; Ber-EP4, epithelial antigen clone BerEP4.
Results
Figure 3 Squamous cell carcinoma with intense cytoplasmic staining (Cytokeratin 5/6, 1000⫻). (Color version of figure is available online.)
Discussion
In metastatic SCC, 9/9 (100%) cases were positive for K903, and 8/9 (89%) cases were positive for CK 5/6 (Figures 2 and 3). CAM 5.2 (7/9 cases, 78%) were positive. CEA (2/9 cases, 22%) cases were positive. All of the cases were negative for Ber-EP4 (100%) and calretinin (100%). Of the malignant mesotheliomas, all 12 cases were positive for calretinin (100%) and CAM 5.2 (100%) (Figure 4). Eleven cases were positive for K903 (92%) and CK 5/6 (92%) (Figures 5 and 6). All cases were negative for CEA (100%) and Ber-EP4 (100%). Among the 12 cases of metastatic adenocarcinoma, all cases were positive for CAM 5.2 (100%). Ten cases were positive for CEA (83%) and Ber-EP4 (83%) (Figure 7). All cases were negative for calretinin (100%). Eleven cases were negative for CK 5/6 (92%). The overall immunohistochemistry staining results of CAM 5.2, K903, CK 5/6, CEA, Ber-EP4, and calretinin in SCC, MM, and ADC are summarized in Table 2. The control staining was satisfactory each time and not shown here.
The difficulty in distinguishing poorly differentiated SCC from MM and ADC is a well-recognized problem. In malignant effusions, the diagnostic difficulty is compounded by the fact that SCC frequently manifests considerable overlap with either adenocarcinoma (cellular spheres, lack of keratinization, prominent nucleoli, and cytoplasmic vacuolization) or malignant mesothelioma (cellular spheres, prominent nucleoli, and ecto-endoplasmic demarcation). In our experience, frequently, a malignant effusion is worked up for the differential diagnosis of ADC versus MM. Generally, a limited panel of epithelial markers, such as CEA, B72.3, Ber-EP4, Leu-M1, and one of the mesothelial markers such as calretinin, HBME-1, or WT-1 are used. The outcome in cases of SCC could be confusing as the stains could all be negative resulting in either additional workup of the patient, assuming poor preservation of sample, or sometimes a false diagnosis based on cytologic features only. Immunohistochemistry can greatly aid in solving such diagnostic dilemmas; however, currently available markers have different sensitivities and specificities for cells of epithelial or mesothelial differentiation.9 Several studies
Figure 2 Squamous cell carcinoma with intense cytoplasmic staining (K903, 1000⫻). (Color version of figure is available online.)
Figure 4 Malignant mesothelioma with nuclear and membranous staining (Calretinin, 1000⫻). (Color version of figure is available online.)
18
Figure 5 Malignant mesothelioma with cytoplasmic staining (K903, 1000⫻). (Color version of figure is available online.)
have evaluated the separation of these entities by using immunohistochemical staining mainly focusing on the differential diagnosis between MM and ADC.6 To our knowledge, there is no such study that has compared the immunohistochemical staining patterns of metastatic SCC with MM and ADC in effusion. In this study, we used a panel of antibody markers, including several different keratins, cell membrane glycoprotein CEA and Ber-EP4 and Calretinin, to evaluate the reactive patterns of these antibody markers among SCC, MM, and ADC. Keratin antibodies have been extensively studied in distinguishing epithelial malignancy from mesothelioma. In summary, it has been reported that antibodies against high molecular weight keratins will react with most of MM and relatively few ADC; antibodies against low molecular weight keratins will react with both of those tumor groups.9,10 Recently, Ordonez reported that pulmonary adenocarcinomas were nonreactive for CK 5/6, whereas MM were positive.11 Our results demonstrated that SCCs were positive for both high and low molecular weight keratins, including CK5/6, and were mostly negative for other markers. ADCs were characteristically positive for CAM 5.2, Ber-EP4 and CEA, but predominantly negative for K903 and CK5/6. Similar to SCC, MMs were also positive for CAM 5.2, K903 and CK 5/6, however, they were also positive for calretinin.
Seminars in Diagnostic Pathology, Vol 23, No 1
Figure 7 Adenocarcinoma with distinct membranous staining (Ber-EP4, 1000⫻). (Color version of figure is available online.)
Calretinin is a cytoplasmic calcium-binding protein and is expressed in MM. It has been widely accepted that calretinin is one of the most positive markers for MM,12,13 although there are conflicting reports regarding the specificity of calretinin to mesothelioma. Some studies have shown that calretinin is also expressed in a few ADC particularly those of ovarian origin.12 Our results demonstrated that calretinin was 100% positive in MM; it was negative in all SCC and pulmonary ADC cases. There is a potential pitfall in the interpretation of calretinin staining patterns, since background benign mesothelial cells are positive for calretinin. In our ADC cases, all the benign reactive mesothelial cells are highlighted by calretinin and all malignant tumor cells are negative for calretinin (Figure 8). When we interpreted the staining pattern of calretinin, we excluded the background staining of benign mesothelial cells. Thus, the positive reactivity of calretinin separates MM from SCC and ADC. CEA has been considered as one of the most reliable markers to distinguish ADC from MM,14 with most of the MM negative for CEA.15 CEA is an oncoprotein in the cells of endodermal origin and expressed in carcinomas of the lung and the gastrointestinal tract. Wang and coworkers were the first to report that mesotheliomas fail to express CEA as determined by the immunoperoxidase method.16 Corson and Pinkus observed that some epithelial mesothe-
Table 2 Immunohistochemistry reactive patterns in malignant effusions
Figure 6 Malignant mesothelioma with membranous staining (Cytokeratin 5/6, 1000⫻). (Color version of figure is available online.)
Antibody
SCC (n ⫽ 9)
MM (n ⫽ 12)
ADC (n ⫽ 12)
CAM 5.2 K903 CK 5/6 CEA Ber-EP4 Calretinin
7 (78%) 9 (100%) 8 (89%) 2 (22%) 0 (0%) 0 (0%)
12 (100%) 11 (92%) 11 (92%) 0 (0%) 0 (0%) 12 (100%)
12 (100%) 1 (8%) 1 (8%) 10 (83%) 10 (83%) 0 (0%)
Abbreviations: SCC, squamous cell carcinoma; MM, malignant mesothelioma; ADC, adenocarcinoma; CEA, carcinoembryonic antigen; EMA, epithelial membrane antigen; Ber-EP4, epithelial antigen clone BerEP4.
Li et al
Distinguishing SCC from MM and ADC
19
References
Figure 8 Adenocarcinoma staining negative while the benign reactive mesothelial cells are positive (Calretinin, 1000⫻). (Color version of figure is available online.)
liomas stained weakly for CEA by using polyclonal antiCEA antibody, but the intensity was much less than that observed in adenocarcinoma.14 Monoclonal antibodies to CEA have been reported to give cleaner backgrounds, less nonspecific cross-reactivity, and higher specificity for adenocarcinoma, with no false positives in 28 cases of pleural mesothelioma.17 However, the sensitivity of this antibody was reduced with only 72% (36 of 50) of positivity in adenocarcinoma.18 In this study, we used polyclonal antiCEA antibodies. This may explain the positive staining of some SCC. When we interpreted the staining pattern of mesothelioma, we excluded the background staining, since inflammatory cells and areas of necrosis may stain positive with polyclonal anti-CEA antibody. Ber-EP4 is another epithelial marker and initially thought to be specific for ADC18,19; however, it is now known that 10% to 20% of MM has focal positivity for this marker.20,21 In our study, ADC was 83% positive for CEA and 83% positive for Ber-EP4, with all ADC staining at least with one of them. Both of CEA and Ber-EP4 were negative for MM. All cases of SCC were negative for Ber-EP4. Our results demonstrate a high positive reaction pattern toward CEA and Ber-EP4 in the cases of ADC, negative reaction pattern in the cases of MM, and predominantly negative pattern in SCC. In summary, the differential diagnosis of SCC from mesothelioma and adenocarcinoma may be challenging on routine cytologic stains. To support any subtle morphologic distinctions among these entities, a panel of antibody markers can be applied to a prepared cell block. Our studies demonstrate that negative staining for calretinin and glycoproteins separated SCC from MM and ADC. Positive staining for calretinin separated MM from SCC and most ADC. Positive staining for glycoproteins and predominantly negative staining for CK 5/6, K903 and calretinin separated ADC from SCC and MM. We also recommend that the diagnosis of MM should not be based solely on a positive CK 5/6 and negative epithelial markers to avoid missing a SCC.
1. Shield PW, Callan JJ, Devine PL: Markers for metastatic adenocarcinoma in serous effusion specimens. Diagn Cytopathol 11:237-245, 1994 2. Matsuo T, Ito H, Anami M, et al: Malignant peritoneal mesothelioma with squamous metaplasia. Cytopathology 4:373-378, 1993 3. Ordonez NG: The immunohistochemical diagnosis of epithelial mesothelioma. Hum Pathol 30:313-323, 1999 4. Motoyama T, Watanabe T, Okazaki E, et al: Immunohistochemical properties of malignant mesothelioma cells in histologic and cytologic specimens. Acta Cytol 39:164-170, 1995 5. Moran CA, Wick MR, Suster S: The role of immunohistochemistry in the diagnosis of malignant mesothelioma. Semin Diagn Pathol 17:178183, 2000 6. Brockstedt U, Gulyas M, Dobra K, et al: An optimized battery of eight antibodies that can distinguish most cases of epithelial mesothelioma from adenocarcinoma. Am J Clin Pathol 114:203-209, 2000 7. Smith-Purslow MJ, Kini SR, Naylor B: Cells of squamous cell carcinoma in pleural, peritoneal and pericardial fluids. Origin and morphology. Acta Cytol 33:245-253, 1989 8. Frost JK: The cell in health and disease. An evaluation of cellular morphologic expression of biologic behavior. Karger, George L. Wied, 1986, pp 165-186 9. Otis CN, Carter D, Cole S, et al: Immunohistochemical evaluation of pleural mesothelioma and pulmonary adenocarcinoma. Am J Surg Pathol 11:445-456, 1987 10. Chu PG, Weiss LM: Expression of cytokeratin 5/6 in epithelial neoplasms: an immunohistochemical study of 509 cases. Mod Pathol 15:6-10, 2002 11. Ordonez NG: Value of cytokeratin 5/6 immunostaining in distinguishing epithelial mesothelioma of the pleura from lung adenocarcinoma. Am J Surg Pathol 22:1215-1221, 1998 12. Doglioni C, Tos AP, Laurino L, et al: Calretinin: a novel immunocytochemical marker for mesothelioma. Am J Surg Pathol 20:1037-1046, 1996 13. Ordonez NG: Value of calretinin immunostaining in differentiating epithelial mesothelioma from lung adenocarcinoma. Mod Pathol 11: 929-933, 1998 14. Corson JM, Pinkus GS: Mesothelioma: profile of keratin proteins and carcinoembryonic antigen: an immunoperoxidase study of 20 cases and comparison with pulmonary adenocarcinomas. Am J Pathol 108: 80-87, 1982 15. Athanassiadou P, Athanassiades P, Lazaris D, et al: Immunocytochemical differentiation of reactive mesothelial cells and adenocarcinoma cells in serous effusions with the use of carcinoembryonic antigen and fibronectin. Acta Cytol 38:718-722, 1994 16. Wang NS, Huang SN, Gold P: Absence of carcinoembryonic antigenlike material in mesothelioma. Cancer 44:937-943, 1979 17. Sheibani K, Battifora H, Burke JS: Antigenic phenotype of malignant mesotheliomas and pulmonary adenocarcinomas: an immunohistologic analysis demonstrating the value of Leu M1 antigen. Am J Pathol 123:212-219, 1986 18. Bailey ME, Brown RW, Mody DR, et al: Ber-EP4 for differentiating adenocarcinoma from reactive and neoplastic mesothelial cells in serous effusions. Comparison with carcinoembryonic antigen, B72.3 and Leu-M1. Acta Cytol 40:1212-1216, 1996 19. Riera JR, Astengo-Osuna C, Longmate JA, et al: The immunohistochemical diagnostic panel for epithelial mesothelioma: a reevaluation after heat-induced epitope retrieval. Am J Surg Pathol 21:1409-1419, 1997 20. Maguire B, Whitaker D, Carrello S, et al: Monoclonal antibody BerEP4: its use in the differential diagnosis of malignant mesothelioma and carcinoma in cell blocks of malignant effusions and FNA specimens. Diagn Cytopathol 10:130-134, 1994 21. Moch H, Oberholzer M, Dalquen P, et al: Diagnostic tools for differentiating between pleural mesothelioma and lung adenocarcinoma in paraffin embedded tissue. Part I: Immunohistochemical findings. Virchows Arch A Pathol Anat Histopathol 423:19-27, 1993