During recent decades, many studies have investigated

Differential Diagnosis of Benign and Malignant Mesothelial Proliferations on Pleural Biopsies Philip T. Cagle, MD; Andrew Churg, MD, PhD Context. Although much of the pathology literature focuses on differential diagnosis of diffuse malignant mesothelioma from other types of cancer, the primary diagnostic challenge facing the pathologist is often whether a mesothelial proliferation on a pleural biopsy represents a malignancy or a benign reactive hyperplasia. Design. Based on previous medical publications, extensive personal consultations, and experience on the United States Canadian Mesothelioma Reference Panel and the International Mesothelioma Panel, salient information was determined about interpretation of benign versus malignant mesothelial proliferations on pleural biopsies. Results. Differentiation of benign reactive mesothelial hyperplasia from diffuse malignant mesothelioma is often difficult. Benign reactive mesothelial hyperplasia may mimic many features ordinarily associated with malignancy, and diffuse malignant mesothelioma may be cytologically bland. Entrapment of benign reactive mesothelial cells within organizing pleuritis may mimic tissue invasion. Conclusions. Various histologic clues favor a benign over a malignant mesothelial proliferation and vice versa. Invasion is the most reliable criterion for determining that a mesothelial proliferation is malignant. When there is any doubt that a pleural biopsy represents a malignancy, we recommend a diagnosis of atypical mesothelial proliferation. (Arch Pathol Lab Med. 2005;129:1421 1427) During recent decades, many studies have investigated the differential diagnosis of diffuse malignant mesothelioma (DMM) versus other types of cancer, especially epithelial DMM versus adenocarcinomas metastatic to the pleura. However, the surgical pathologist not infrequently encounters a different problem with pleural biopsies, namely, whether a mesothelial proliferation represents a malignant neoplasm or a benign reactive hyperplasia. 1 13 Reactive pleural processes may be associated with pleural effusion and thickening of the pleura and in some cases raise the clinical possibility of malignancy. Reactive mesothelial cell hyperplasia and organizing pleuritis may demonstrate increased cellularity and reactive cytologic atypia of epithelial mesothelial cells and of spindle mesothelial cells, fibroblasts, and endothelial cells, respectively. It is not unusual for both epithelial and spindle proliferating cells to occur together in a reactive process. These benign reactive mesothelial proliferations may mimic malignancy because of high cellularity, cytologic atypia, and other histologic features, whereas some DMMs are cytologically bland and may be sampled in minimally invasive areas that disguise their malignant nature. Obviously, it is of considerable importance to patients to know that they have a benign pleural process rather than DMM or stage IV carcinoma. Accepted for publication July 14, 2005. From the Departments of Pathology, The Methodist Hospital, Houston, Tex (Dr Cagle); and University of British Columbia, Vancouver (Dr Churg). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Philip T. Cagle, MD, Department of Pathology, The Methodist Hospital, 6565 Fannin St, Main Bldg Room 227, Houston, TX 77030 (e-mail: pcagle@tmh.tmc.edu). Biopsies that include DMM in the differential diagnosis can be referred to the United States Canadian Mesothelioma Reference Panel for the opinions of a panel of experts. In 22% of 217 cases circulated to the entire United States Canadian Mesothelioma Reference Panel between 1994 and 1999, 1 or more panelists disagreed with the majority on whether the case represented a benign reactive proliferation or a malignancy. 6 Only difficult cases are circulated to the entire panel, but this rate of disagreement among experts highlights the diagnostic dilemma that can be posed by the question of benign reactive proliferation versus malignancy on pleural biopsy. APPROACH TO THE DIFFERENTIAL DIAGNOSIS As with biopsies from other sites, it is always best to interpret pleural biopsies within the clinical context, including clinical history and findings as well as radiologic findings. Pleural processes that may be associated with benign reactive mesothelial hyperplasia and/or organizing pleuritis on a pleural biopsy are listed in Table 1. A unilateral, bloody pleural effusion is characteristic of many pleural effusions caused by malignancy, but several benign processes listed in Table 1 may be unilateral and some (pulmonary infarct, pneumothorax, surgery, trauma) may produce bloody effusions. Occasionally, a patient has exuberant pleuritis overlying a subpleural lung cancer that may or may not be invasive into the pleura, and the radiology may be very helpful in understanding the full context of the biopsy. A history of asbestos exposure should not be given any inordinate weight in making a diagnosis of DMM for at least 3 reasons: (1) asbestos exposure can cause benign pleural effusions, benign pleural plaques, and benign mesothelial reactions, and these outnumber malignant me- Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg 1421

Figure 1. High-power view shows flat, inconspicuous normal mesothelial cells lining the visceral pleural surface (hematoxylin-eosin, original magnification 300). Figure 2. Simple mesothelial hyperplasia shows a more conspicuous layer of relatively bland cuboidal cells regularly spaced along the pleural surface (hematoxylin-eosin, original magnification 350). 1422 Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg

Table 1. Causes of Benign Reactive Mesothelial Hyperplasia and Organizing Pleuritis Infections Collagen vascular diseases Pulmonary infarcts Drug reactions Pneumothorax Subpleural lung carcinomas Surgery Trauma Nonspecific Table 2. Features of Benign Reactive Mesothelial Hyperplasia That May Mimic Malignancy High cellularity Cytologic atypia/mitoses Necrosis Papillary excrescences or lumens Entrapment (pseudoinvasion with benign mesothelial cells entrapped in organizing pleuritis) sotheliomas; (2) asbestos exposure is fairly common and asbestos-exposed individuals are susceptible to the same nonasbestos causes of pleural effusions and reactions as everyone else; and (3) not all DMMs are caused by asbestos exposure. The impression of the surgeon on thoracoscopic or thoracotomy examination is helpful information. Classically, malignancies involving the pleura appear as multiple tumor nodules or encasement of the pleura by tumor. However, gross appearances may occasionally be misleading and, if the biopsy is not conclusive for malignancy, it is recommended that the pathologist not yield to pressure to make a diagnosis of malignancy based solely on reported gross impressions. On the other hand, if the surgeon states that the process appears benign on thoracoscopic or thoracotomy examination, the pathologist should be circumspect about diagnosing malignancy. Histologic features of benign reactive mesothelial hyperplasia that may potentially mimic histologic features of malignancy are listed in Table 2. In subsequent sections, we discuss the histologic features of mesothelial proliferations that can be encountered on a pleural biopsy. Diagnosis of malignancy on pleural biopsy may be problematic not only because of potential benign reactive mesothelial hyperplasia, but also because of sampling and other issues that may limit the usefulness of any biopsy. We advocate a very guarded approach to the diagnosis of malignancy on pleural biopsies when a reactive mesothelial proliferation is in the differential diagnosis. Given the consequences to the patient of a diagnosis of malignant mesothelioma, when there is any doubt about malignancy on a pleural biopsy, we recommend the term atypical mesothelial proliferation or atypical mesothelial hyperplasia be used for the diagnosis and additional tissue obtained if clinically indicated. MESOTHELIAL PROLIFERATIONS LIMITED TO THE PLEURAL SURFACE If a pleural biopsy shows only mesothelial cells along the surface with no mesothelial cells observed in the underlying tissue, the proliferation may represent simple mesothelial cell hyperplasia or atypical mesothelial hyperplasia. It is likely that some instances of the latter represent mesothelioma in situ, but that diagnosis cannot be made with current histologic criteria unless invasive malignancy is also present. The normal mesothelial lining of the pleura is composed of flat cells that are inconspicuous on casual histologic examination (Figure 1). Irritation of the pleural surface leads to simple hyperplasia of the mesothelium. The mesothelial cells in simple hyperplasia form a more conspicuous layer of relatively bland cuboidal cells regularly spaced along the pleural surface (Figure 2). Distinct nucleoli may sometimes be present. The hyperplastic cells usually show regular spacing along the membrane. Atypical mesothelial hyperplasia is a more florid mesothelial proliferation along the pleural surface than simple hyperplasia. There is a spectrum of cytologic atypia and cellularity between cases of simple hyperplasia to cases of atypical mesothelial hyperplasia that are worrisome for noninvasive malignancy. Atypical mesothelial hyperplasia may consist of a single layer of cells with varying degrees of cytologic atypia along the surface (Figure 3) or may consist of piled-up accumulations of cells on the surface (Figure 4). The cells of atypical mesothelial hyperplasia are cuboidal to elongate, may have enlarged nuclei, and often have prominent nucleoli with significant cell-tocell variation (Figures 5 and 6). Simple mesothelial hyperplasia is usually not of diagnostic concern and clearly represents a benign reactive process. Some examples of atypical mesothelial hyperplasia represent the more cellular and more cytologically atypical spectrum of benign reactive processes, whereas other examples probably represent noninvasive malignancy, either at the margin of an invasive tumor or possibly a preinvasive stage of malignancy. Inflammatory changes and clinical history/findings of a benign process favor a benign reaction, and a lack of inflammatory changes or dense underlying fibrosis may increase concern for a malignancy, but neither situation is definitive. On the other hand, bulk tumor in nodules or masses on the pleural surface should be considered malignant (Figure 7). Progressive morphologic steps from normal epithelium through varying degrees of dysplasia to carcinoma in situ Figure 3. This mesothelium demonstrates minimal features of atypical mesothelial hyperplasia consisting of a single layer of cells along the pleural surface with minimal to mild cytologic atypia (hematoxylin-eosin, original magnification 350). Figure 4. This example of atypical mesothelial hyperplasia consists of piled-up cuboidal to elongate cells with nuclear atypia and prominent nucleoli in many cells. There are papillary-like tufts of cells projecting into the pleural space (hematoxylin-eosin, original magnification 300). Figure 5. Atypical mesothelial hyperplasia with proliferating mesothelial cells piled up above the pleural surface. Some of the cells have prominent nucleoli (hematoxylin-eosin, original magnification 350). Figure 6. Atypical mesothelial hyperplasia composed of cytologically atypical cells with prominent nucleoli and papillary tufts. This histology is suggestive of malignancy, but is not definitive and no invasion was identified in the biopsy (hematoxylin-eosin, original magnification 350). Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg 1423

Figure 7. Bulk tumor on the pleural surface, as shown here, should be considered malignant (hematoxylin-eosin, original magnification 150). Figure 8. Very early stage tubulopapillary diffuse malignant mesothelioma (DMM) on the visceral surface has microfoci of superficial invasion. These types of lesions and extrapolation from other types of cancers lead many to believe that there is a preinvasive stage of DMM. Currently, there are no reliable criteria to separate mesothelioma in situ from reactive mesothelial hyperplasia and, therefore, we recommend against making diagnoses of mesothelioma in situ at the present time (hematoxylin-eosin, original magnification 250). Figure 9. Bland necrosis of a sarcomatous diffuse malignant mesothelioma (DMM) is seen in the lower left half of the figure. Bland necrosis, consisting of ischemic-type necrosis without significant inflammatory cells or cellular debris, is characteristic of DMM, as opposed to benign reactive processes (hematoxylin-eosin, original magnification 150). Figure 10. Papillary excrescences of benign mesothelial cells are seen in this reactive pleuritis. A partially en face cut also gives the impression of very superficial invasion. The mesothelial cells are very regular and bland (hematoxylin-eosin, original magnification 200). 1424 Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg

Table 3. Histologic Features Favoring a Diagnosis of Malignancy Over Reactive Hyperplasia Unequivocal invasion of underlying tissues Cellular nodules with stromal expansion Cellularity of the atypical cells throughout the full thickness of the pleura (lack of zonation) Unequivocal malignant features of severe pleomorphism, atypical mitoses, etc Bland necrosis to minimally invasive malignancy are established in the cervix, lung, and other organs. Therefore, it is probable that there are preneoplastic and in situ phases in the development of DMM. 9 Some examples of atypical mesothelial hyperplasia could conceivably represent these preneoplastic and preinvasive phases of DMM. Whitaker et al 3 and Henderson et al 4 proposed potential criteria for a morphologic diagnosis of mesothelioma in situ, but noted that this diagnosis could only be confirmed when accompanied by adjacent or subsequent invasive tumor (Figure 8). Currently, there are no reliable morphologic criteria for the diagnosis of mesothelioma in situ in the absence of invasive tumor, and we suggest that the term mesothelioma in situ should not be used in diagnosis. Instead, a surface mesothelial proliferation of indeterminate malignancy should be diagnosed as atypical mesothelial hyperplasia or atypical mesothelial proliferation, with the understanding that some of these cases will eventually turn out to be benign and that others may turn out to be malignant. When in doubt, it is best to avoid a diagnosis that cannot be substantiated and instead ask for additional tissue if clinically indicated. DIAGNOSIS OF MALIGNANCY Pathologists recognize certain histologic features as being characteristic of malignancy, but are aware that some reactive processes may share degrees of these features. Mesothelial cells are particularly notorious for demonstrating greater degrees of cytologic atypia in benign reactive processes than they sometimes do when they are malignant. As is well known to cytologists, malignant mesothelial cells can be very cytologically bland. Indeed, although malignant mesothelial cells may occasionally be highly pleomorphic, the presence of highly pleomorphic cells is much more suggestive of a carcinoma than a DMM. Mitoses, which are often associated with malignancy, can be increased in a benign reactive mesothelial hyperplasia. Necrosis may occur in both DMM and benign reactive pleural processes. In DMM, tumor necrosis is typically bland, consisting of areas of ischemic necrosis without significant inflammation or cellular debris (Figure 9), whereas in inflammatory pleural conditions like tuberculosis, the necrosis is typically associated with inflammatory cells and cellular debris. On occasion, benign reactive mesothelial hyperplasias may exhibit architectural features like papillary excrescences and tubular lumens that mimic the architectural patterns of DMM or other malignancies. There are some architectural clues that favor a benign proliferation over a malignancy. Papillary excrescences of benign hyperplastic mesothelial cells tend to be tufts of cells without true fibrovascular cores (Figure 10). Tubule-like structures in benign mesothelial proliferations tend to be small, simple, nonbranching structures (Figure 11), but it should be noted that some DMMs and metastatic carcinomas are also composed of small, simple, nonbranching tubules or glands. Complex branching tubules or glands favor malignancy (Figure 12). Histologic features strongly favoring a diagnosis of DMM over a benign reactive mesothelial hyperplasia are listed in Table 3. Invasion of underlying tissues is the most reliable criterion for the diagnosis of DMM, whether epithelial or sarcomatous. When invasion is minimal or equivocal, keratin immunostaining can help highlight the keratin-immunopositive neoplastic mesothelial cells infiltrating into underlying fat, connective tissue, or lung tissue (Figure 13). However, the mere fact that proliferating mesothelial cells are immunopositive for keratin, calretinin, or other markers does not make a diagnosis of benign versus malignant; the immunostains are used to demonstrate invasion of the cells into underlying tissues. In addition to invasion, certain other criteria favor malignancy, both epithelial and sarcomatous types. Obvious cellular nodules of neoplastic cells may expand into the stroma, or neoplastic cells may have such conspicuous malignant cytologic features (severe pleomorphism, abnormal mitoses) that a diagnosis of malignancy is warranted. Benign reactive mesothelial proliferations tend to demonstrate zonation; that is, the proliferating cells are found toward the pleural surface and the deeper tissues are more fibrotic (Figure 14). Finding mesothelial cells deep in the pleural connective tissue or through the full thickness of the pleura favors malignancy. Bland necrosis is characteristic of DMM, particularly the desmoplastic variety, which may otherwise be difficult to differentiate from fibrous pleuritis or pleural plaque. A soft criterion that favors a benign diagnosis is capillaries in a biopsy that are parallel to each other and that are perpendicular to the pleural surface (Figure 15). Capillaries in malignant mesotheliomas, particularly desmoplastic mesotheliomas, tend to be inconspicuous and arranged in a haphazard fashion. MIMICS OF INVASION Entrapment refers to benign reactive mesothelial cells that have remained on a previous pleural surface that is now covered by fibrinous, organizing, or fibrotic pleuritis, giving an impression of invasion. As layers of fibrin and/ or fibrous connective tissue spread over the pleural surface in an inflammatory process, the entrapped mesothelial cells tend to form linear arrays marking the site of the previous surface (Figures 16 and 17). In some cases, including when there are adhesions of the visceral and parietal surfaces, the reactive mesothelial cells may line elongated clefts that may sometimes resemble glands (Figure 18). Histologic features that favor entrapment of nonneoplastic mesothelial cells over invasion are listed in Table 4. Figure 11. Cytologically bland mesothelial cells line round to ovoid, simple, nonbranching tubules in this reactive proliferation (hematoxylineosin, original magnification 350). Figure 12. Complex branching tubules favor a diagnosis of diffuse malignant mesothelioma (hematoxylin-eosin, original magnification 150). Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg 1425

Figure 13. Keratin immunostaining highlights microinvasion of keratin-positive mesothelioma cells into adipose tissue, confirming the malignant nature of the mesothelial proliferation (keratin, original magnification 350). Figure 14. Zonation characteristic of a benign process demonstrates proliferation of mesothelial cells, some forming simple nonbranching tubules, toward the pleural surface and absent in deeper fibrotic tissues (hematoxylin-eosin, original magnification 150). 1426 Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg

Table 4. Histologic Features Favoring Nonneoplastic Entrapment Over Invasion Active fibrin deposition with active inflammation Linear arrays of individual cells and small glands Favors nonneoplastic entrapment Simple nonbranching glands Proliferating mesothelial cells separated by large amounts of stroma En face cuts of unoriented or poorly oriented pleural biopsies may also give a false impression of invasion. Sections cut parallel to the free serosal surface rather than at right angles to the surface may give the impression that benign surface reactive mesothelial cells are surrounded by connective tissue. This may be particularly difficult to recognize on small biopsies. As always, we urge caution when the possibility of an en face section cut is encountered. The presence of mesothelial cells on 2 surfaces is a clue that a specimen has been cut en face. IMMUNOHISTOCHEMICAL STUDIES Immunohistochemical examination is generally of little value in differentiating a benign reactive mesothelial proliferation from a DMM. As previously mentioned, broadspectrum keratin immunostains can be used to confirm invasion into underlying tissues such as adipose tissue when the hematoxylin-eosin stain is equivocal, but virtually all active benign mesothelial reactions are also keratin positive; what keratin stains really do is to show the distribution of mesothelial cells within the tissue. While there are a number of markers such as calretinin that often indicate a mesothelial origin, both reactive and malignant mesothelial proliferations stain positively. Although a number of markers such as p53 and epithelial membrane antigen have been proposed to provide secondary support of a diagnosis of malignancy on pleural biopsy, none of these markers is absolutely specific or sensitive and, therefore, cannot be relied on to make a diagnosis. 14 19 References 1. McCaughey WTE, Al-Jabi M. Differentiation of serosal hyperplasia and neoplasia in biopsies. Pathol Annu. 1986;21(pt 1):271 293. 2. Tuder RM. Malignant disease of the pleura: a histopathological study with special emphasis on diagnostic criteria and differentiation from reactive mesothelium. Histopathology. 1986;10:851 865. 3. Whitaker D, Henderson DW, Shilkin KB. The concept of mesothelioma in situ: implications for diagnosis and histogenesis. Semin Diagn Pathol. 1992;9: 151 161. 4. Henderson DW, Shilin KB, Whitaker D. Reactive mesothelial hyperplasia vs mesothelioma, including mesothelioma in situ. Am J Clin Pathol. 1998;110:397 404. 5. 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. 1998;110:191 199. 6. Churg A, Colby TV, Cagle P, et al. The separation of benign and malignant mesothelial proliferations. Am J Surg Pathol. 2000;24:1183 1200. 7. Galateau-Sallé F, Cagle PT. Non-malignant versus malignant proliferations on pleural biopsy. In: Cagle PT, ed. Diagnostic Pulmonary Pathology. New York, NY: Marcel Dekker; 2000:555 567. 8. Cagle PT. Pleural histology. In: Light RW, Lee YCG, eds. Pleural Disease: An International Textbook. London, England: Arnold Publishers; 2003:249 255. 9. Cagle PT. Is there a preneoplastic mesothelial lesion? In: Vizjak A, Ferluga D, Bussolati G, eds. Update in Pathology. Ljubljana, Slovenia: University of Ljubljana; 2003:345 346. 10. Churg A, Roggli V, Galateau-Salle F, et al. Tumours of the pleura: mesothelial tumours. In: Travis WD, Brambilla E, Harris CC, Muller-Hermelink HK, eds. Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: IARC Press; 2004. World Health Organization Classification of Tumours; vol 10. 11. Laga AC, Allen T, Bedrossian C, Laucirica R, Cagle PT. Reactive mesothelial hyperplasia. In: Cagle PT, editor-in-chief. The Color Atlas and Text of Pulmonary Pathology. New York, NY: Lippincott Williams & Wilkins; 2005:587 588. 12. Churg A, Cagle PT, Roggli VL. Tumors of the Serosal Membranes. Washington, DC: Armed Forces Institute of Pathology. In press. Atlas of Tumor Pathology; 4th series, fascicle 10. 13. Galateau-Salle F, ed-in-chief; Brambilla E, Cagle PT, Churg A, et al, panel members. Pathology of Malignant Mesothelioma: An Update Prepared by the International Mesothelioma Panel. New York, NY: Springer Verlag. In press. 14. Mayall FG, Goddard H, Gibbs AR. p53 immunostaining in the distinction between benign and malignant mesothelial proliferations using formalin-fixed paraffin sections. J Pathol. 1992;168:377 381. 15. Cagle PT, Brown RW, Lebovitz RM. p53 immunostaining in the differentiation of reactive processes from malignancy in pleural biopsy specimens. Hum Pathol. 1994;25:443 448. 16. Hurlimann J. Desmin and neural marker expression in mesothelial cells and mesotheliomas. Hum Pathol. 1994;25:753 757. 17. Wolanski KD, Whitaker D, Shilkin KB, Henderson DW. The use of epithelial membrane antigen and silver-stained nucleolar organizer regions testing in the differential diagnosis of mesothelioma from benign reactive mesothelioses. Cancer. 1998;82:593 590. 18. Davidson B, Nielsen S, Christensen J, et al. The role of desmin and N- cadherin in effusion cytology: a comparative study using established markers of mesothelial and epithelial cells. Am J Surg Pathol. 2001;25:1405 1412. 19. Kumaki F, Kawai T, Churg A, et al. Expression of telomerase reverse transcriptase (TERT) in malignant mesotheliomas. Am J Surg Pathol. 2002;26:365 370. Figure 15. In this reactive granulation tissue, capillaries and fibroblasts are parallel to one another and perpendicular to the pleural surface (hematoxylin-eosin, original magnification 350). Figure 16. Entrapment of benign reactive mesothelial cells occurs when organizing fibrin and blood (upper half of figure) layer over the pleural surface lined by reactive mesothelial cells (lower half of figure) (hematoxylin-eosin, original magnification 250). Figure 17. With organization of the overlying pleural exudates, proliferating mesothelial cells are entrapped between the pleura (lower portion of figure) and newly formed fibrous tissue (upper portion of figure). In this example, the entrapped reactive mesothelial cells are hyperplastic and form small nonbranching tubules (hematoxylin-eosin, original magnification 250). Figure 18. As the fibrous tissue matures, a linear array of entrapped mesothelial cells is present between the pleura and overlying fibrous tissue. Especially with adhesions, elongate lumens may be formed (hematoxylin-eosin, original magnification 150). Arch Pathol Lab Med Vol 129, November 2005 Mesothelial Proliferations on Pleural Biopsies Cagle & Churg 1427