Document n9JrL8N8OJOmjKX4eOJDr5Vba

Guidelines for Pathologic Diagnosis of Malignant Mesothelioma A Consensus Statement from the International Mesothelioma Interest Group Aliya N. Husain, MD; Thomas V Colby, MD; Nelson G. Ordonez, MD; Thomas Krausz, MD; Alain Borczuk, MD; Philip T. Cagle, MD; Lucian R. Chirieac, MD; Andrew Churg, MD; Francoise Galateau-Salle, MD; Allen R. Gibbs, MBChB, FRCPath; Allen M. Gown, MD; Samuel P Hammar, MD; Leslie A. Litzky, MD; Victor L. Roggli, MD; William D. Travis, MD; Mark R. Wick, MD Context.--Malignant mesothelioma (MM) is an uncom mon tumor that can be difficult to diagnose. Objective.--To develop practical guidelines for the path ologic diagnosis of MM. Data Sources.--A pathology panel was convened at the International Mesothelioma Interest Group biennial meet ing (October 2006). Pathologists with an interest in the field also contributed after the meeting. Conclusions.--There was consensus opinion regarding (1) distinguishing benign from malignant mesothelial prolifera tions (both epithelioid and spindle cell lesions), (2) cytologic diagnosis of MM, (3) key histologic features of pleural and peritoneal MM, (4) use of histochemical and immunohistochemical stains in the diagnosis and differential diagnosis of MM, (5) differentiating epithelioid MM from various carci nomas (lung, breast, ovarian, and colonic adenocarcinomas and squamous cell and renal cell carcinomas), (6) diagnosis of sarcomatoid mesothelioma, (7) use of molecular markers in the differential diagnosis of MM, (8) electron microscopy in the diagnosis of MM, and (9) some caveats and pitfalls in the diagnosis of MM. Immunohistochemical panels are inte gral to the diagnosis of MM, but the exact makeup of panels used is dependent on the differential diagnosis and on the antibodies available in a given laboratory. Immunohistochem ical panels should contain both positive and negative mark ers. The International Mesothelioma Interest Group recom mends that markers have either sensitivity or specificity greater than 80% for the lesions in question. Interpretation of positivity generally should take into account the localiza tion of the stain (eg, nuclear versus cytoplasmic) and the per centage of cells staining (>10% is suggested for cytoplasmic membranous markers). These guidelines are meant to be a practical reference for the pathologist. (Arch Pathol Lab Med. 2009;133:1317-1331) As part of the International Mesothelioma Interest Group biennial meeting held in Chicago (October ologic diagnosis of malignant mesothelioma (MM). The discussion focused on practical diagnostic guidelines 2006), there was a pathology half-day workshop that in meant to be a reference for the pathologist, rather than a cluded invited lecturers and an open forum on the path- mandate or review of the literature. This article is the re sult of that discussion with additional input from other Accepted for publication October 16, 2008. pathologists who could not attend the meeting. From the Department of Pathology, University of Chicago, Chicago, Illinois (Drs Husain and Krausz); the Department of Pathology, Mayo Clinic College of Medicine, Scottsdale, Arizona (Dr Colby); the Section GENERAL RECOMMENDATIONS The diagnosis of MM must always be based on the re of Immunocytochemistry, Department of Pathology, The University of sults obtained from an adequate biopsy in the context of Texas M. D. Anderson Cancer Center, Houston (Dr Ordonez); the De partment of Pathology, Columbia University Medical Center, New York, New York (Dr Borczuk); the Department of Pathology, The Methodist Hospital, Houston, Texas (Dr Cagle); the Department of Pathology, Brigham and Women's Hospital, Boston, Massachusetts (Dr Chirieac); the Department of Pathology, University of British Columbia, Vancou appropriate clinical, radiologic, and surgical findings. A history of asbestos exposure should not be taken into con sideration by the pathologist when diagnosing MM. Lo cation of the tumor (pleural vs peritoneal) as well as the sex of the patient will affect the differential diagnosis as ver, Canada (Dr Churg); Groupe Mesopath, Laboratoire d'Anatomie discussed later. Specific information on antibody clones* S Pathologique, Caen, France (Dr Galateau-Salle); the Department of His- topathology, Llandough Hospital, Penarth, South Glamorgan, United Kingdom (Dr Gibbs); the Department of Pathology, PhenoPath Labo York, New York (DrTravis); and the Department of Pathology, University ratories, Seattle, Washington, and the Department of Pathology, Uni of Virginia Medical Center, Charlottesville (Dr Wick). versity of British Columbia, Vancouver, Canada (Dr Gown); Diagnostic Dr Churg serves as a consultant to various law firms involved in Specialties Laboratory, Bremerton, Washington (Dr Hammar); Pathol asbestos litigation. The other authors have no relevant financial interest ogy & Laboratory Medicine, University of Pennsylvania Medical Cen in the products or companies described in this article. ter, Philadelphia (Dr Litzky); the Department of Pathology, Duke Uni Reprints: Aliya N. Husain, MD, University of Chicago, Rm S627, 5841 versity Medical Center, Durham, North Carolina (Dr Roggli); the De S Maryland Ave, MC6101, Chicago, IL 60631 (e-mail: ahusain@bsd. partment of Pathology, Memorial Sloan-Kettering Cancer Center, New uchicago.edu or aliya.husain@uchospitals.edu). Arch Pathol Lab Med--Vol 133, August 2009 Malignant Mesothelioma Diagnosis--Husain et al 1317 and their source should be obtained from the current lit erature because this an evolving area and is outside of the scope of this article. BENIGN VERSUS MALIGNANT MESOTHELIAL CELL PROLIFERATIONS Separating benign versus malignant mesothelial prolif erations presupposes first that the process has been rec ognized as mesothelial. The diagnostic approach used when distinguishing reactive mesothelial hyperplasia from epithelioid mesothelioma is different from that used when distinguishing fibrous pleuritis from desmoplastic mesothelioma.1 The major problem areas are discussed in the following. Reactive Mesothelial Hyperplasia Versus Epithelioid MM It is well known that reactive mesothelial proliferations may mimic mesothelioma (or metastatic carcinoma). Some of the causes of reactive mesothelial hyperplasia in the pleural space include infections, collagen vascular diseas es, pulmonary infarcts, drug reactions, pneumothorax, subpleural lung carcinomas, surgery, trauma, and nonspe cific inflammation. Exuberant mesothelial reactions also are encountered in the peritoneum and pericardium, and the latter may be particularly worrisome. The specific features of a reactive mesothelial prolifer ation that may mimic a neoplasm include high cellularity, the presence of numerous mitotic figures and cytologic atypia, the presence of necrosis, the formation of papillary groups, and entrapment of mesothelial cells within fibrous tissue mimicking invasion (Figure 1). Features distinguish ing reactive mesothelial hyperplasia from mesothelioma are summarized in Table 1. Demonstration of stromal or fat invasion is the key fea ture in the diagnosis of MM (Figure 2). Invasion may be into visceral or parietal pleura (or beyond) and this can be highlighted with immunostains such as pancytokeratin or calretinin. Invasion by mesothelioma is often subtle and may be into only a few layers of collagenous tissue below the mesothelial space. Invasive mesothelial cells may also be deceptively bland in appearance and completely lack a desmoplastic reaction. However, it is emphasized that if a solid piece of malignant tumor, with histologic features of MM, is identified, the presence of invasion is not required for the diagnosis. Reactive mesothelial proliferations tend to show a uni formity of growth and this may be highlighted with pancytokeratin staining, which show regular sheets and sweeping fascicles of cells that respect mesothelial bound aries in contrast to the disorganized growth and haphaz ardly intersecting proliferations seen in mesothelioma. Although certain immunohistochemical stains are more typically positive in benign proliferations and others in malignant proliferations, they should not be solely relied on for diagnosis in individual cases. The best known of these include epithelial membrane antigen, p53, and desmin. The results of Attanoos et al,2 who studied 60 me sotheliomas and 40 cases of reactive mesothelial hyper plasia, are included in Table 2. This topic was reviewed by King et al3 who concluded that desmin and epithelial membrane antigen (membranous staining) were the most useful but that the diagnostic sensitivity and specificity for both was less than 90%. They noted that novel markers of proliferation such as MCM2 and AgNOR may be useful adjuncts in this situation but they present technical diffi 1318 Arch Pathol Lab Med--Vol 133, August 2009 culties and are not widely used. Although telomerase transcriptase expression was suggested to discriminate between hyperplastic and neoplastic mesothelium, sub sequent studies showed limited usefulness.4 In a recent study by Kato et al,5 GLUT-1 reactivity was found in 40 of 40 mesotheliomas; whereas all 40 cases of reactive mesothelium were negative. A study done at the University of Chicago6 showed all 40 benign mesothelial tissues (20 normal, 20 reactive cases) to be negative for GLUT-1, whereas of 45 MM, 9 (20%) were negative, 34 (53%) were weakly positive, and 12 (27%) were strongly positive. Thus, when GLUT-1 is positive, it is a helpful marker for MM, both epithelioid and sarcomatoid (Figure 3, A through D), but is not helpful when negative. Fibrous Pleurisy Versus Desmoplastic Variant of Sarcomatoid Mesothelioma The identification of features of malignancy in a des moplastic mesothelioma requires adequate tissue and the amount of tissue in a closed pleural biopsy is often insuf ficient. Large surgical biopsies are generally needed. High-grade sarcomas presenting in the pleura generally do not enter into the differential diagnosis of fibrous pleu risy versus desmoplastic mesothelioma. Features to sepa rate the latter two are shown in Table 3. In the study by Mangano et al,7 the distinction of fibrous pleurisy from desmoplastic mesothelioma could be made by identifying one or more of the following features in a spindle cell proliferation of the pleura: invasive growth, bland necrosis, frankly sarcomatoid areas, and metastatic disease. Stromal invasion is often more difficult to recognize in spindle cell proliferations of the pleura than in epithelioid proliferations. The invasive malignant cells are often de ceptively bland, resembling fibroblasts, and pancytokeratin staining is invaluable in highlighting the presence of cytokeratin-positive malignant cells in regions where they should not normally be present: in the connective tissue, adipose tissue, or skeletal muscle deep to the parietal pleura or invading the visceral pleura and lung tissue (or other extrapleural structures present in the sample) (Fig ure 4, A and B). Although Mangano et al7 also found bland necrosis of paucicellular fibrous tissue to be a reli able criterion of malignancy, it may be subtle and one may be reluctant to base a diagnosis of malignancy solely on its presence. Fortunately, most cases that show bland ne crosis also show invasive growth. Similarly, the presence of ''frankly sarcomatoid foci'' is a distinctly subjective de termination and one would be reluctant to base a diag nosis of malignancy on its presence alone because reactive processes may show marked cytologic atypia, albeit typ ically at the surface of the process. Uniformity of growth and thickness of the process, sur face atypia with deep maturation, and perpendicular thinwalled vessels are typical of reactive fibrous pleuritis (Fig ure 5, A and B), in contrast to the disorganized growth pattern and variable thickness of desmoplastic mesotheli omas. A helpful clue in desmoplastic mesotheliomas is the presence of expansile nodules of varying sizes with abrupt changes in cellularity between nodules and their sur rounding tissue. CYTOLOGIC FEATURES OF MM There is limited role of cytology in the primary diag nosis of MM, but this is debated between cytopathologists Malignant Mesothelioma Diagnosis--Husain et al Figure 1. Reactive mesothelial hyperplasia within fibrous tissue mimicking invasion (hematoxylin-eosin, original magnification XI00). Figure 2. Epithelioid malignant mesothelioma invading fat (hematoxylin-eosin, original magnification X100). Figure 3. GLUT-1 staining by immunohistochemistry of mesothelial lesions. A, Malignant mesothelioma with strong immunoreactivity for GLUT-1 (original magnification X400). B, Mesothelial hyperplasia negative for GLUT-1. As expected, red blood cells are strongly positive (original magnification X200). C, Well-differentiated papillary mesothelioma is negative for GLUT-1 (original magnification X200). D, Sarcomatoid meso thelioma with positive staining (original magnification X400). Arch Pathol Lab Med--Vol 133, August 2009 Malignant Mesothelioma Diagnosis--Husain et al 1319 Figure 4. Desmoplastic mesothelioma. A, Proliferation of bland-appearing spindle cells with haphazard growth pattern (hematoxylin-eosin, orig inal magnification X200). B, Keratin immunohistochemical staining highlights infiltration into fat (original magnification X200). Figure 5. Reactive fibrous pleuritis. A, There is uniformity of growth and thickness of the reactive process (hematoxylin-eosin, original magnifi cation X40). B, Perpendicular uniformly paced thin-walled vessels with fibrin on one surface and progressive maturation of fibrous tissue into deeper part (hematoxylin-eosin, original magnification X100). MM includes sarcomas and other spindle cell neoplasms, and the differential diagnosis of mixed MM includes other mixed or biphasic tumors such as synovial sarcoma. Des moplastic mesotheliomas may mimic fibrous pleuritis. Be cause each broad histologic category has its own distinc tive differential diagnosis, the immunostains selected for further workup of a MM are dictated by the histologic category into which it falls.9-16 The most common histologic type of MM is epithelioid. The common secondary patterns of epithelioid MM are readily recognized by most pathologists: tubulopapillary pattern, acinar (glandular) pattern, adenomatoid pattern (also termed microglandular), and the solid mesothelialcell pattern. Some epithelioid MMs have a distinctive rec ognizable secondary pattern consisting of clusters of epi thelioid cells floating in pools of hyaluronic acid. The clear cell, deciduoid, adenoid cystic, signet ring, small cell, and rhabdoid patterns are less common secondary patterns of epithelioid MM that are more likely to be confused with other types of cancer. The tubulopapillary pattern, acinar (glandular) pattern, and adenomatoid (microglandular) pattern must be dif Arch Pathol Lab Med--Vol 133, August 2009 ferentiated from adenocarcinoma metastatic to the pleura. The tubulopapillary pattern consists of a mixture of pap illary structures lined by bland flat, cuboidal, or polygonal cells with fibrovascular cores and glandlike tubules. The acinar pattern consists of elongated or branching glandlike lumina lined by relatively bland cuboidal cells. The ade nomatoid pattern consists of bland flat to cuboidal cells lining small glandlike structures. The solid pattern consists of nests, cords, or sheets of round, oval, or polygonal cells with abundant eosinophilic cytoplasm and round, vesicular nuclei with prominent nu cleoli. These cells resemble nonneoplastic, reactive mesothelial cells, and the differential diagnosis may include re active mesothelial hyperplasia, solid adenocarcinoma, and even squamous cell carcinoma due to the abundant pink cytoplasm. The solid poorly differentiated pattern consists of sheets and nests of relatively discohesive polygonal to round cells with uniform nuclei. Lymphomas and poorly differentiated carcinomas enter into the differential diag nosis of solid poorly differentiated MM.9-15 The clear cell pattern is composed of MM cells with clear cytoplasm that should be differentiated from clear Malignant Mesothelioma Diagnosis--Husain et al 1321 Figure 6. Cytologic features of malignant mesothelioma. A, Numerous large clumps of cells are present in effusion of malignant mesothelioma (Papanicolaou, original magnification X40). B, The clumps have a berrylike external contour (Papanicolaou, original magnification X200). C, Multiple binucleated cells are seen (Papanicolaou, original magnification X400). D, Cell block also shows frequent clumps and can be very useful in performing special stains (hematoxylin-eosin, original magnification X200). Figure 7. Calretinin immunohistochemical staining. A, Malignant mesothelioma has diffuse strong nuclear and cytoplasmic positivity (original magnification X200). B, Adenocarcinoma is usually negative but may show focal positivity as shown here (original magnification X400). 1322 Arch Pathol Lab Med--Vol 133, August 2009 Malignant Mesothelioma Diagnosis--Husain et al V V?'. Mr OB Figure 10. D2-40 immunohistochemical staining. A, Strong membra nous staining in malignant mesothelioma (original magnification X200). B, Focal staining in squamous cell carcinoma (original magni fication X400). Figure 9. WT-1 immunohistochemical staining. A, Strong nuclear staining in malignant mesothelioma invading fat. Note that endothelial cells show cytoplasmic staining only (original magnification X200). B, Strong granular cytoplasmic staining in large cell carcinoma (original magnification X400). C, Cytoplasmic staining in adenocarcinoma of lung (original magnification X400). keratin (CK) 5/6 and WT-1 in PSC and the infrequent staining for carcinoembryonic antigen in PSC limits the ability of those markers to discriminate between these en tities. Estrogen receptor may be helpful in difficult cases, as would a positive result for progesterone receptor. In 1326 Arch Pathol Lab Med--Vol 133, August 2009 male patients, WT-1 (nuclear staining) and D2-40 are use ful markers in addition to calretinin, and for nonserous adenocarcinoma, B72.3, MOC-31, BG8, and Ber-EP4 all have high sensitivity and specificity. In addition, carcinoembryonic antigen may also be useful in the setting when PSC is not in the differential diagnosis. Recently, hcaldesmon has been reported to be highly useful as a mesothelial marker.38 SARCOMATOID MESOTHELIOMA The criteria for distinguishing reactive fibrous pleurisy from sarcomatoid mesothelioma have been well charac terized and were summarized earlier in the article. An immunohistochemical panel that can be useful for the initial evaluation of a sarcomatoid tumor involving the pleura should include cytokeratins, calretinin, and D2-40. Multiple cytokeratin antibodies including AE1/3, CAM 5.2 (or CK18), and CK7 should be used as cytokeratin ex pression can be focal, weak, and/or variable.39,40 Other positive markers that are used in the evaluation of epithe lioid mesothelioma such as WT-1 and CK5/6, as well as adenocarcinoma markers such as Ber-EP4, carcinoembryonic antigen, and MOC-31, do not provide much added utility in sarcomatoid tumors. D2-40 and calretinin have Malignant Mesothelioma Diagnosis--Husain et al Figure 12. Thyroid transcription factor 1 immunohistochemical stain ing shows strong nuclear staining in lung adenocarcinoma (original magnification X400). Figure 13. BG8 immunohistochemical shows strong membranous staining in large cell carcinoma (original magnification X400). Figure 11. MOC-31 immunohistochemical staining. A, Large cell car cinoma with membranous staining (original magnification X400). B, Papillary adenocarcinoma of lung with strong staining (original mag nification X400). C, Focal staining in malignant mesothelioma (original magnification X200). been the 2 positive mesothelial markers most consistently expressed in sarcomatoid mesotheliomas in a variable per centage of cases.40-42 False positives can occur by the mis interpretation of positive reactivity within benign en trapped lymphatics or reactive mesothelial elements. A histologically malignant sarcomatoid tumor that is Arch Pathol Lab Med--Vol 133, August 2009 strongly and diffusely cytokeratin positive usually limits the differential diagnosis to sarcomatoid mesothelioma, sarcomatoid carcinoma, and, on occasion, synovial sarco ma or metastatic sarcomatoid renal cell carcinoma. Al though synovial sarcomas of the pleura (or primary pul monary synovial sarcomas involving the pleura) usually present as localized solid tumors, they can present with diffuse pleural thickening that is similar to MM. The di agnosis of synovial sarcoma should be considered when there is a highly cellular neoplasm with very little cyto plasm in tumor cells resulting in overlapping nuclei, focal hemangiopericytoma-like blood vessels are present, and there is limited keratin expression. The diagnosis can be confirmed by molecular testing for its distinctive X;18 translocation in formalin-fixed, paraffin-embedded tissue. Unless there is convincing calretinin and D2-40 positivity, it is difficult to separate out the spindled cell component of a partially sampled sarcomatoid carcinoma from sar comatoid mesothelioma. Heterologous elements may be present in both tumors. A possible distinguishing feature is when the tumor has areas in which the malignant cells are infiltrating through densely collagenized fibrosis (as is characteristic of desmoplastic mesothelioma). This pattern Malignant Mesothelioma Diagnosis--Husain et al 1327 Table 6. Immunohistochemical Markers Used in the Differential Diagnosis Between Epithelioid Pleural Mesothelioma and Renal Cell Carcinoma Marker Current Value/Comments Epithelioid mesothelioma (positive mesothelioma markers) Keratin 5/6 Mesothelin Calretinin Podoplanin (D2-40) WT-1 protein Very useful. Seventy-five percent to 100% of mesotheliomas are positive. Renal cell carcinomas are negative. Very useful. One hundred percent of mesotheliomas are positive. Renal cell carcinomas are negative. Very useful. Virtually all mesotheliomas are positive and the staining is often strong and diffuse with nuclear and cytoplasmic staining. Four percent to 10% of renal cell carcinomas are focally positive. Very useful. Eighty-six percent to 100% of mesotheliomas show positivity along the cell membrane. Renal cell carcinomas are negative. Useful. Forty-three percent to 93% of mesotheliomas show nuclear positivity. Four percent of renal cell carci nomas are positive. Renal cell carcinoma (positive carcinoma markers) CD15 (Leu-M1) RCC Ma MOC-31 Ber-EP4 BG8 (LewisY) Useful. Sixty-three percent of renal cell carcinomas are positive. Mesotheliomas only rarely show focal positivi ty. Can stain any necrotic tissue. Useful. Fifty percent to 70% of renal cell carcinomas are positive. Eight percent of mesotheliomas are focally positive. Limited utility. Fifty percent of renal cell carcinomas are positive. Two percent to 10% of mesotheliomas show focal staining. Not useful. Forty-two percent of renal cell carcinomas are positive. Up to 20% of mesotheliomas are focally positive. Not useful. Four percent of renal cell carcinomas and 3%-7% of mesotheliomas are positive. is quite typical of MM and favors that diagnosis, although ultimately, in this instance, the diagnosis may have to in corporate other gross and clinical features. In some cases, especially with limited biopsy material, it may be difficult to distinguish metastatic sarcomatoid carcinoma from sar comatoid mesothelioma. Carcinoma markers such as carcinoembryonic antigen and thyroid transcription factor 1 (for lung) can be tried. Sarcomatoid renal cell carcinoma can metastasize to the pleura and grow like a mesotheli oma producing a pseudomesotheliomatous sarcomatoid- type pattern. Differential cytokeratin positivity profiles, other than CK5/6, have not been reported to date in the differential diagnosis of these 2 tumors. CK5/6 has been reported to be negative in sarcomatoid renal cell carci nomas, but the low sensitivity of CK5/6 as a marker in sarcomatoid mesothelioma greatly limits its utility.43 Cal retinin and D2-40 positivity have not been extensively studied in sarcomatoid renal cell carcinomas. One series reported calretinin negativity in all 4 sarcomatoid renal cell carcinomas tested, but at this point it would be pru- Table 7. Peritoneal Malignant Mesothelioma (PMM) Versus Papillary Serous Carcinoma (PSC) and Nongynecologic Adenocarcinoma (AdCa) Marker Current Value/Comments Positive mesothelioma markers Calretinin D2-40 CK5/6 WT-1 Useful. Positive in 85%-100% of PMM cases, but reactivity in 0%-38% PSC limits its use as a single marker. Potentially useful. Positive in 93%-96% of PMM cases, but wide spectrum of positivity in PSC from 13% to 65%; requires more data in this context. Not useful. Positive in 53%-100% of PMM cases, but positive in 22%-35% of PSC cases. Not useful. Positive in 43%-93% of PMM, but 89%-93% of PSC are positive. PSC markers MOC-31 BG8 Ber-EP4 B72.3 CEA Estrogen receptor Progesterone receptor Very useful. Positive in 5% of PMM and 98% of PSC. Very useful. Positive in 3%-9% of PMM and 73% of PSC. Useful. Positive in 9%--13% of PMM and 83%-100% of PSC. Limited utility. Positive in 0%-3% of PMM versus 65%-100% of PSC, but many trace/focal. Not useful. Zero percent PMM but 0%-45% of PSC (mean, 20%), sensitivity in PSC is too low compared with other choices. Useful. Sixty percent to 93% in PSC, and negative or very low positive rate (0%-8%) in PMM. Limited utility. Lower sensitivity than estrogen receptor, but uniformly negative in PMM. May be valuable if positive. PMM versus nongynecologic AdCa (biliary, pancreatic, gastric) Calretinin WT-1 D2-40 CK5/6 MOC-31 BG8 CEA B72.3 Ber-EP4 Very useful. Positive in 85%-100% of PMM but also positive in 10% of pancreatic AdCa, so limited as a single marker. Very useful. Positive in 43%-93% of PMM, 3% of gastric AdCa, negative in pancreatic AdCa. Potentially useful. Positive in 93%-96% of PMM, negative in pancreatic and gastric AdCa (but limited data). Not useful. Positive in 53%-100% of PMM, but 38% pancreatic AdCa positive. Very useful. Positive in 5% of PMM and 87% of AdCa. Very useful. Positive in 3%-9% of PMM and 89% of AdCa. Very useful. Positive in 81% of AdCa, negative in PMM. Very useful. Positive in 84% of pancreas, 89% of bile duct, 98% of colon AdCa, 0%-3% of PMM. Useful. Positive in >98% of pancreatic and gastric AdCa, 9%-13% of PMM. Abbreviations: CEA, carcinoembryonic antigen; CK, cytokeratin. 1328 Arch Pathol Lab Med--Vol 133, August 2009 Malignant Mesothelioma Diagnosis--Husain et al dent to incorporate additional gross and clinical correla tion.43 A histologically malignant sarcomatoid tumor that is ei ther focally cytokeratin positive or cytokeratin negative should be cautiously interpreted, and a diagnosis of me sothelioma very carefully considered. Focal cytokeratin positivity has been reported in many different types of sarcomas. It is also possible that the focal cytokeratin pos itivity represents entrapment of benign pleural elements. If the initial round of cytokeratins proves to be negative or there is only focal cytokeratin positivity, additional blocks should be selected and stained, and cytokeratin an tigen retrieval techniques as well as antibody source and dilutions should be reviewed. A vimentin stain is useful in assessing the general antigenic integrity of the tissue. Particularly in the absence of convincing cytokeratin pos itivity, calretinin and/or D2-40 positivity alone should not be interpreted as evidence of mesothelial differentiation. These markers are variably positive in many different types of sarcomas and other immunohistochemical mark ers should be added at this point. The expanded differ ential diagnosis might include other sarcomas (epithelioid hemangioendothelioma/angiosarcoma, synovial sarcoma, liposarcoma, myogenic or neurogenic tumors), malignant solitary fibrous tumor, melanoma, and lymphoma. The marker panel should be expanded accordingly to include antibodies such as CD31, CD34, desmin, myoglobin, S100, and LCA. It should be noted that some muscle markers are often positive at least focally, and on occasion more diffusely, in sarcomatoid mesotheliomas.44 These markers include muscle-specific actin (HHF-35) and a smooth mus cle actin. Desmin positivity in pure sarcomatoid mesothe liomas is quite rare.44,45 After extensive workup and with appropriate clinical and radiologic features, cytokeratinnegative sarcomatoid mesotheliomas have been published as a diagnosis of exclusion.39,46,47 MOLECULAR MARKERS IN MM The road to mesothelioma development is known to be long and winding and studies suggest that mesothelioma development is related to complex cytogenetic and molec ular events attesting to the long latency period of more than 30 to 40 years and to the multistep process. There are no molecular markers as yet that are useful in con firming a diagnosis of mesothelioma. However, detecting t(X;18) is most useful in the differential diagnosis of sy novial sarcoma. Begueret et al48 have confirmed the pres ence of this translocation in 90% of purely sarcomatoid primary synovial sarcoma of the pleura, whereas this translocation has never been detected in MM.49 ELECTRON MICROSCOPY OF MM The electron microscopic features of MM are well described.12,50 There is no single ultrastructural feature that is diagnostic of mesothelioma but rather the combination of several features that may be diagnostically useful. For epithelioid mesotheliomas these include very long thin apical microvilli that do not have a glycocalyx, as opposed to the generally shorter microvilli of adenocarcinomas that usually do have a glycocalyx; perinuclear tonofilament bundles; the presence of basal lamina; and long desmosomes. Electron microscopy is most valuable for well to mod erately well differentiated epithelioid tumors because sar comatous mesotheliomas typically do not show mesothe- Arch Pathol Lab Med--Vol 133, August 2009 lial features. It may also be very useful in establishing the correct diagnosis when the immunohistochemical results are equivocal or further support of a diagnosis of either MM or serous carcinoma is needed.51 Formalin-fixed ma terial retrieved from a paraffin block can sometimes be useful because microvilli and tonofilament bundles tend to be preserved. However, the limitations of electron mi croscopy need to be recognized. For one thing there may be considerable overlap of ultrastructural features be tween adenocarcinomas and epithelioid mesotheliomas. As well, the characteristic ultrastructural features of me sothelioma tend to be seen in tumors for which the di agnosis is obvious by light microscopy; in tumors that are poorly differentiated, electron microscopy usually does not solve the problem.52,53 FEATURES NOT USEFUL IN MAKING THE DIAGNOSIS OF MM History of Asbestos Exposure Because there is an association of asbestos exposure and the development of MM, many pathologists may adopt the position that a history of asbestos exposure makes a tumor more likely to be a mesothelioma, and, conversely, in the absence of such a history, they are reluctant to diagnose mesothelioma. However, the history of exposure to asbes tos or the absence of such a history is not useful to the pathologist in making a diagnosis of mesothelioma. The situation is analogous to that of lung cancer: Although most lung cancers occur in cigarette smokers, no one would hesitate to diagnose a lung cancer if told that the patient was a nonsmoker. For mesothelioma a similar sce nario applies: The diagnosis is based on clinical, radiolog ic, and, ultimately, pathologic features, and the issue of asbestos exposure is irrelevant. Presence of Psammoma Bodies Like many tumors with a papillary architecture, meso theliomas may occasionally contain psammoma bodies. Although the presence of psammoma bodies might sug gest certain carcinomas (eg, papillary carcinoma in the thyroid and serous papillary tumors of the female genital tract and peritoneal surface), their presence is not a di agnostic aid in either confirming or excluding a diagnosis of mesothelioma. Mucicarmine Positivity The evaluation of mucin stains in putative mesothelio mas is not straightforward and positivity in a tumor with some mucin stains should not be used as a criterion in confirming or excluding the diagnosis of mesothelioma. Although the presence of mucin has been classically used to support a diagnosis of adenocarcinoma rather than mesothelioma, mucin positivity may be encountered in mesotheliomas. With Alcian blue and colloidal iron staining mesotheliomas may contain hyaluronate droplets that show a positive reaction and this material is removed with hyaluronidase pretreatment. The spaces containing the positive material tend to be quite large and not typical of the intracellular mucin droplets seen in adenocarcino ma. Because hyaluronidase-sensitive material may be en countered in mesotheliomas, digested periodic acid-Schiff stains have been recommended for neutral (''epithelial'') mucins because this finding is almost entirely restricted to adenocarcinoma. Rarely otherwise typical mesothelio mas may contain mucicarmine-positive, digested periodic Malignant Mesothelioma Diagnosis--Husain et al 1329 acid-Schiff-positive, and Alcian blue-positive droplets that are resistant to hyaluronidase. Ultrastructural evalu ation in the small series reported by Hammar et al35 sug gested that this positivity was related to crystallized pro teoglycans. A rim of faint mucicarmine positivity is common in the vacuoles of mesotheliomas and should not be considered a feature diagnostic of adenocarcinoma. SV40 Virus Exposure The SV40 virus has been proposed by some as an etiologic agent in some mesotheliomas. Although this pro posal remains controversial, it is clear that the presence or absence of exposure to SV40 virus is not a criterion in confirming or excluding the diagnosis of mesothelioma.12 PITFALLS IN THE DIAGNOSIS OF MM The first ''port of call'' for the histologic diagnosis of MM is the hematoxylin-eosin stain. Immunohistochemical panels are adjuncts to diagnosis and should not be per formed automatically or blindly without considering sev eral factors. As stated previously the major determinants of which panel to use are (1) the location of the tumor-- it will vary as to whether it is pleural, peritoneal, or an other serosal surface; (2) the phenotypic problem--benign versus malignant, epithelioid, spindle, biphasic, small cell, or pleomorphic; and (3) the experience of the laboratory. A laboratory using immunohistochemical stains should perform them frequently, have well-established protocols, and should have an appreciation of their sensitivities and specificities with respect to the various morphologic prob lems. There is no single utopian immunohistochemical panel to cover all diagnostic ''mesothelial'' problems. One of the problems in comparing the results of partic ular antibodies from different studies is a lack of stan dardization of immunohistochemical procedures. This can result in conflicting results for sensitivity and specificity for various antibodies. The study of King et al3 tabulates the data for antibody clone, manufacturer, dilution, and antigen retrieval methods for 5 antibodies used in sepa rating benign and malignant mesothelial proliferations in 13 studies. It illustrates the wide variability used between the various studies. Prior to using an antibody for diag nosis a laboratory should have carried out an extensive workup to find the ideal conditions for routine use. The type of pathologic sample may affect results. For example, tiny needle biopsies may show crush artifact and false-positive immunostaining with various antibodies. Also the edges of biopsies may show artifactual positive immunostaining. There may also be variation in interpre tation of what is positive illustrated by the fact that some laboratories will only consider calretinin to be positive when there is nuclear staining, whereas a minority will consider cytoplasmic staining to be positive. This can sig nificantly affect sensitivity and specificity. Another problem associated with immunohistochemistry may be putting too much emphasis on focal immunopositivity. We would suggest that weak or focal staining of less than 10% of the cells should be considered as being negative when interpreting a panel of stains. Also one can observe positive immunostaining with mesothelial mark ers in reactive proliferations of submesothelial fibroblasts in the vicinity of nonmesothelial tumors and inflamma tory pleural diseases--it is important not to diagnose these as mesotheliomas. In contrast, mesotheliomas may 1330 Arch Pathol Lab Med--Vol 133, August 2009 invade the underlying lung and entrapped epithelial cells may show positive immunostaining with epithelial mark ers. Careful correlation with the hematoxylin-eosin sec tions is necessary to avoid misinterpretation. MESOTHELIOMA REVIEW PANELS Pathology review panels for putative mesothelioma cas es have been functioning since the 1960s. These panels have served as a referral source for pathologists facing diagnostic problems and, more recently, to confirm diag noses for large oncology treatment trials. The diagnosis of mesothelioma has been considered to be difficult although the increasing frequency of this malignancy and applica tion of immunohistochemical staining has made the di agnosis more reliable. In a 1991 report from the US and Canadian Mesothelioma Panel, only 70.5% of 200 cases had a three-fourths majority of the panel. It should be emphasized that these represented referral cases and by their nature were already difficult. In a more recent report from the same panel, agreement among panel members as to benign versus malignant was 78% for a group of 217 cases. SUMMARY This article gives broad guidelines for making the di agnosis of MM, which, although a rare tumor, has a grave prognosis and invariably has medicolegal implications. The salient recommendations are (1) histologic features and immunohistochemical panel are used in distinguish ing benign from malignant mesothelial proliferations; (2) there is limited usefulness of cytology, histochemical stains, electron microscopy, and molecular markers; pan els of antibodies are given, which need to be used accord ing to the differential diagnosis in each case; (3) and in the typical case in which all features are concordant, 2 mesothelioma markers and 2 carcinoma markers may be adequate to make a diagnosis; however, when there are discordant findings, additional markers should be per formed. The pathologist should always take the clinical, radiologic, and pathologic features into consideration and get expert second opinion in difficult cases, as necessary. We thank Hedy Kindler, MD, and Sam Armato, PhD, for or ganizing the 2006 International Mesothelioma Interest Group meeting in Chicago and providing the forum for the pathology panel discussion that led to this consensus article. References 1. Churg A, Colby TV, Cagle P, et al. The separation of benign and malignant mesothelial proliferations. Am J Surg Pathol. 2000;24:1 183-1200. 2. Attanoos RL, Griffin A, Gibbs AR. The use of immunohistochemistry in dis tinguishing reactive from neoplastic mesothelium: a novel use for desmin and comparative evaluation with epithelial membrane antigen, p53, platelet-derived growth factor-receptor, P-glycoprotein and Bcl-2. Histopathology. 2003;43:231238. 3. King J, Thatcher N, Pickering C, Hasleton P. Sensitivity and specificity of immunohistochemical antibodies used to distinguish between benign and malig nant pleural disease: a systematic review of published reports. Histopathology. 2006;49:561-568. 4. Krissman M, Muller K, Jaworska M, Johnen G. Pathological anatomy and molecular pathology. Lung Cancer. 2004;45S:S29-S33. 5. Kato Y, Tsuta K, Seki K, et al. Immunohistochemical detection of GLUT-1 can discriminate between reactive mesothelium and malignant mesothelioma. Mod Pathol. 2007;20:215-220. 6. Acurio A, Arif Q, Gattuso P, et al. Value of immunohistochemical markers in differentiating benign from malignant mesothelial lesions: United States and Canadian Academy of Pathology annual meeting [abstract 1527]. Mod Pathol. 2008;21:334A. 7. Mangano WE, Cagle PT, Churg A, Vollmer RT, Roggli VL. The diagnosis of desmoplastic malignant mesothelioma and its distinction from fibrous pleurisy: a histologic and immunohistochemical analysis of 31 cases including p53 immunostaining. Am J Clin Pathol. 1 998;110:191-199. Malignant Mesothelioma Diagnosis--Husain et al