Guidelines for Pathologic Diagnosis of Mesothelioma: 2023 Update of the Consensus Statement From the International Mesothelioma Interest Group

The 2021 World Health Organization (WHO) classification of mesothelioma⁴  retains the 3 major histologic subtypes—epithelioid, biphasic, and sarcomatoid—but incorporates several architectural patterns and cytologic and stromal features that are prognostically significant. The diagnostic term mesothelioma is recommended, rather than malignant mesothelioma. To avoid confusion, the lesion previously termed well-differentiated papillary mesothelioma has been renamed well-differentiated papillary mesothelial tumor. As before, the term multicystic mesothelioma is discouraged, in favor of (multiloculated) peritoneal inclusion cyst.⁵  In short, the term mesothelioma now applies only to malignant tumors.

Epithelioid mesothelioma accounts for 60% to 70% of pleural and 80% to 90% of peritoneal mesotheliomas.^6–9  These tumors comprise polygonal, oval, or cuboidal cells that often mimic nonneoplastic, reactive mesothelial cells. Epithelioid mesothelioma exhibits several generally familiar architectural patterns, including tubulopapillary (Figure 1), trabecular (Supplemental Figure 1, see the supplemental digital content file, containing 18 figures and 2 tables at https://meridian.allenpress.com/aplm in the November 2024 table of contents.), micropapillary (Figure 2), adenomatoid (Figure 3), and solid (Figure 4). The identified architectural patterns should be reported for each tumor in both biopsy and resection specimens, and in definitive resection specimens (ie, extended pleurectomy/decortication or extrapleural pneumonectomy); percentage representation (to the nearest 10%) of each pattern should also be reported.¹⁰ 

Tubulopapillary, trabecular, and adenomatoid patterns are associated with a more favorable prognosis. Conversely, any micropapillary component or at least 50% solid pattern is associated with worse prognosis.¹¹  Micropapillary pattern correlates with a higher incidence of lymphatic invasion. Necrosis is seen in 30% of epithelioid mesotheliomas and is associated with a worse prognosis.¹² 

Variant cytologic and stromal features are recognized for epithelioid mesothelioma, and familiarity facilitates proper diagnosis. Those features of known prognostic significance should be reported when present.

Rare epithelioid mesotheliomas comprise clusters of mildly atypical tumor cells in a matrix of loose myxoid stroma (Supplemental Figure 2). Epithelioid mesotheliomas with at least 50% myxoid morphology and less than 50% solid growth pattern have a favorable prognosis.^13,14 

Rhabdoid cytologic features are prognostically unfavorable, defined by a variable proportion (15%–75%) of tumor cells morphologically similar to those of rhabdomyoblastic tumors, containing cytoplasmic globules that express cytokeratins but are negative for myogenin (Supplemental Figure 3).¹⁵ 

Epithelioid mesotheliomas with pleomorphic cytomorphology—defined by nuclear enlargement, hyperchromasia, prominent nucleoli, and (often) multinucleation, forming at least 10% of the tumor—behave like sarcomatoid and biphasic mesotheliomas.^16,17  The 2021 WHO classification recommends that tumors with pleomorphic cytomorphology (Supplemental Figure 4) be classified as epithelioid, biphasic, or sarcomatoid, based on the remaining tumor cell morphology, although the presence of a pleomorphic component should be documented and its poor prognostic significance noted.^4,18 

Mesothelioma with lymphohistiocytoid features is defined by polygonal tumor cells morphologically similar to histiocytes, admixed with a marked lymphoid infiltrate (Supplemental Figure 5, A). Immunostains might be necessary to distinguish this entity from a nonneoplastic inflammatory process, lymphoepithelial carcinoma, or lymphoma. “Lymphohistiocytoid features” does not simply refer to markedly inflamed mesothelioma; the tumor cells must show histiocytoid morphology. Mesothelioma with lymphohistiocytoid features can be classified as epithelioid, biphasic, or sarcomatoid from the morphology of the non-lymphohistiocytoid component. Lymphohistiocytoid features in sarcomatoid mesothelioma impart a favorable prognosis, although prognostic significance is less clear for epithelioid tumors.¹⁹ 

These features are not prognostically significant (Supplemental Figure 5, B; Supplemental Figure 6). However, familiarity with these morphologies facilitates distinction, respectively, from clear cell carcinomas, sarcoma, and melanoma²⁰ ; florid deciduosis or deciduoid carcinomas²¹ ; signet ring cell adenocarcinomas of the lung and gastrointestinal tract²² ; small cell carcinomas, sarcomas, and lymphomas²³ ; and adenoid cystic carcinoma. Note that mesotheliomas with small cell–like morphology do not show true neuroendocrine differentiation, nor do they stain with neuroendocrine markers, and use of the term small cell mesothelioma is discouraged, to avoid confusion with small cell carcinoma.

Sarcomatoid mesotheliomas account for 5% to 15% of pleural and less than 5% of peritoneal mesotheliomas.^8,9,24  They are composed of infiltrating sheets of spindle cells with variable cytologic atypia (Supplemental Figure 7). Tumors can show necrosis, atypical mitoses, and/or heterologous (including rhabdomyosarcomatous, osteosarcomatous, and chondrosarcomatous) elements, which when extensive must be distinguished from true sarcomas.²⁵  Sarcomatoid mesothelioma has a significantly poorer prognosis than epithelioid mesothelioma.^12,24,26 

Desmoplastic mesothelioma is a pattern of sarcomatoid mesothelioma characterized by a hypocellular population of bland spindle cells, arranged in a haphazard (patternless) fashion between bands of dense collagenous stroma (Figure 5). While the stroma is similar to that seen in pleural plaque, the tumor cells are arranged in a haphazard fashion as opposed to a linear arrangement usually parallel to the surface in the latter.²⁷  Desmoplastic mesothelioma might not be suspected unless frankly sarcomatoid areas or areas of invasion into adipose tissue or lung are found. A diagnosis of “desmoplastic mesothelioma” requires desmoplastic features in at least 50% of a tumor in a definitive resection specimen. If desmoplastic morphology is seen in a smaller biopsy specimen, the diagnostic qualification “with desmoplastic features” is recommended.

Transitional features in mesothelioma are an uncommon and recently described pattern, defined by sheetlike growth and cytomorphology intermediate between sarcomatoid and epithelioid tumor cells, with elongated but plump and cohesive cells (Figure 6). These were likely variably classified historically, but they are now regarded as a pattern of sarcomatoid morphology, as their prognosis and transcriptomic profile is similar to sarcomatoid mesotheliomas.^28,29  Reticulin staining can be helpful in identifying the transitional pattern, as staining is present around individual cells as in sarcomatoid mesothelioma, in contrast to surrounding clusters of cells in the epithelioid subtype.²⁹ 

Biphasic mesothelioma accounts for 15% to 30% of pleural and 10% to 20% of peritoneal mesotheliomas.^6–9  These tumors contain malignant epithelioid and sarcomatoid components (Supplemental Figure 8), with transitional morphology regarded as sarcomatoid in this context (ie, if transitional features are seen in an otherwise epithelioid mesothelioma, a diagnosis of biphasic mesothelioma should be rendered). In a definitive resection specimen, at least 10% of each component must be present to render the diagnosis of biphasic mesothelioma. In biopsy specimens, biphasic mesothelioma should be diagnosed if malignant epithelioid and sarcomatoid components are present, regardless of the percentage of each component (ie, <10% epithelioid or sarcomatoid component does not preclude a diagnosis of biphasic mesothelioma in a biopsy—this change in the 2021 WHO classification⁴  reflects the typically higher percentage of sarcomatoid morphology seen in resection versus biopsy specimens).^7,11,30  In peritoneal tumors, some evidence suggests that even focal (<10%) sarcomatoid morphology imparts a poorer prognosis than is typical of epithelioid peritoneal mesothelioma,³¹  so it is prudent to note even focal (<10%) sarcomatoid growth in resected peritoneal mesotheliomas.

The prognosis of biphasic mesothelioma is intermediate between that of pure epithelioid and pure sarcomatoid mesothelioma.^8,32  While data are limited, there is some suggestion that biphasic mesotheliomas with predominant (>50%) sarcomatoid morphology have a poorer prognosis,^33,34  so the percentage contribution of the epithelioid component and the sarcomatoid component should be reported in biopsies and resections of biphasic mesothelioma.

Historically, expert pathologists showed moderate interobserver agreement in diagnosis of biphasic mesothelioma, though improved agreement can be attained through application of strict diagnostic criteria.^28,34,35  Epithelioid mesothelioma associated with a reactive spindled mesothelial population presents a challenging differential for biphasic mesothelioma, particularly when the spindled component lacks overtly atypical morphologic features. In this setting, immunohistochemical and molecular studies may help define the spindled mesothelial component as malignant (see Ancillary Studies in Diagnosis of Biphasic Mesothelioma below).

Although an exhaustive review of the molecular biology of mesothelioma is beyond the scope of these guidelines, a brief overview is provided as context for subsequent discussion of diagnostic molecular assays and immunostains. The genes most commonly affected by somatic mutation and copy number alteration in mesothelioma are BAP1, NF2, CDKN2A, TP53, LATS1/2, and SETD2, with mutation rates in particular genes varying by both tumor site and histotype.^36,37  Gene fusions are rare in mesothelioma, though a subset of peritoneal mesotheliomas in young patients harbor ALK rearrangements or EWSR1::ATF1 or EWSR1/FUS::CREB fusions, while rare cases of EWSR1::YY1 fusion are reported in peritoneal mesothelioma in middle-aged patients.^38–42 ALK rearrangements are very rarely detected in pleural mesothelioma.^43,44 

BRCA1-associated protein-1 (BAP1) is a tumor suppressor gene encoding a deubiquitylase with roles in cell cycle regulation, cellular differentiation, cell death, and DNA damage response.⁴⁵ BAP1 is inactivated in approximately 60% of pleural and 70% of peritoneal mesotheliomas by missense, truncating, and splice site mutations; truncating fusion events; and copy number loss via chr 3p21.1 deletion. BAP1 mutations appear to be an early event in mesothelioma pathogenesis, representing the most common molecular alteration in reported cases of mesothelioma in situ (see Mesothelioma In Situ below). BAP1 alterations are found more often in epithelioid than in sarcomatoid tumors.⁴⁶  Germline BAP1 mutations produce the BAP1 tumor predisposition syndrome (see Germline Predisposition to Mesothelioma below).

CDKN2A (p16) resides on the chr 9p21 locus alongside CDKN2B (p15INK5), p14RF, and MTAP. Homozygous deletion of CDKN2A is found in approximately 70% of pleural mesotheliomas (including 90%–100% of sarcomatoid mesotheliomas and 40%–70% of epithelioid and biphasic types) but just 10% to 15% of peritoneal mesotheliomas.^47–49 CDKN2A point mutations are exceptionally rare in mesothelioma.⁴⁸  Like BAP1 alterations, CDKN2A deletions are an early event in mesothelioma pathogenesis and are occasionally detected in mesothelioma in situ (see Mesothelioma In Situ below). Some evidence suggests that CDKN2A deletion plays a role in evolution from epithelioid to biphasic mesothelioma in some cases.⁴⁸  Approximately 75% to 90% of mesotheliomas with CDKN2A deletion show codeletion of the neighboring MTAP gene,^47,49  which can therefore be used as an immunohistochemical surrogate for CDKN2A deletion.

NF2, located in the chromosomal region 22q12, encodes Merlin, a tumor suppressor protein in the Hippo signaling pathway.⁵⁰ NF2 inactivation in mesothelioma is predominantly via truncating mutations or gene deletion, and is somewhat more common in biphasic and sarcomatoid (∼70%) than in epithelioid (∼40%) tumors.⁴⁸ NF2 inactivation appears to be a fairly late event in mesothelioma pathogenesis, and intratumoral heterogeneity for NF2 alterations is common.⁵¹ 

Reactive and malignant mesothelial proliferations can overlap morphologically with nonmesothelial lesions and malignancies. Establishing mesothelial lineage is a crucial early step in proper diagnosis of a serosal lesion. Although mesothelial cells show certain characteristic morphologic features, panels of “mesothelial” and “epithelial” immunostains are now almost universally applied in routine diagnosis to confirm the morphologic impression of mesothelial differentiation. Markers of specific epithelial lineages also play a role, depending on the clinical and morphologic differential diagnosis. Rarely, the differential for mesothelioma includes nonepithelial (eg, mesenchymal, hematolymphoid) tumors, to which the immunopanel must be tailored.

Before undertaking immunostains for evaluation of mesothelioma, the responsible laboratory should have performed a rigorous validation to determine ideal conditions for routine use in their hands. Immunostains should be interpreted with caution in minute biopsies, in those with crush artifact (which may induce false-positive or false-negative staining), and around the edges of biopsy specimens (which may show artifactual positive immunostaining). Careful attention should be paid to avoid misinterpretation of immunostaining in entrapped benign mesothelial or epithelial structures.

Immunohistochemical stains for broad-spectrum cytokeratin (eg, pancytokeratin, AE1/AE3, CAM 5.2, CK OSCAR) are highly sensitive for mesothelioma, including sarcomatoid mesothelioma. In one large study, 93% of sarcomatoid mesotheliomas exhibited immunoreactivity for at least 1 cytokeratin; that percentage may be even higher if a cytokeratin cocktail is used, there is adequate sampling of the tumor, and the tissue is well fixed.²⁴  For sarcomatoid neoplasms, cytokeratin positivity is additionally useful in excluding spindle cell sarcoma or melanoma,⁵²  although rare sarcomas and melanomas can be positive for cytokeratins, and areas of heterologous differentiation in mesothelioma are often cytokeratin negative. Note that reactive mesothelial stroma is also keratin positive; keratin stain does not differentiate benign from malignant mesothelial proliferations.

Broad-spectrum cytokeratins are virtually 100% sensitive for epithelioid mesothelioma. If an epithelioid malignant neoplasm causing diffuse serosal thickening is negative for multiple broad-spectrum cytokeratins, other diagnoses should be considered, such as melanoma, epithelioid hemangioendothelioma or angiosarcoma, and lymphoma.

Occasional tumors do not stain with any marker. This often reflects artifact, such as overfixation in formalin, or alcohol fixation followed by antigen retrieval (commonly used for cytology specimens), so some knowledge about the fixative is important, as is proper laboratory validation for alcohol-fixed tissue.^53–55  Assessment of internal controls is helpful. If needed, vimentin can be used to assess baseline immunoreactivity of the tissue.

The most common morphologic differential diagnosis for mesothelioma is carcinoma, and panels of mesothelial and epithelial immunomarkers are routinely used to establish mesothelial lineage and distinguish benign or malignant mesothelial proliferations from epithelial mimics. The specific markers used will depend on the differential diagnosis, and as noted above, nonepithelial (eg, mesenchymal, melanocytic, or hematolymphoid) tumors may occasionally enter the differential, requiring appropriate immunopanel modifications.

The best-characterized and most common mesothelial markers include calretinin (Supplemental Figure 9), CK5 or CK5/6 (Supplemental Figure 10), WT1 (Wilms tumor-1; Supplemental Figure 11), and podoplanin (D2-40) (Supplemental Figure 12). Each of these markers shows greater than 80% sensitivity for epithelioid mesothelial proliferations, with lower sensitivity for sarcomatoid proliferations.^56,57  HEG1 is a promising mesothelial marker for epithelioid tumors, with similar sensitivity to other mesothelial markers, and possibly greater specificity in the differential with carcinoma of the lung.⁵⁸  Importantly, none of these markers is entirely specific for mesothelial origin, and all can be positive (usually focal, though sometimes diffuse) in a subset of carcinomas.⁵⁶  Further, different mesothelial immunostains show different patterns of positive staining. For calretinin, combined cytoplasmic and nuclear staining is typically present in mesothelial cells. For WT1, only nuclear staining is considered positive, and cytoplasmic-only staining should be disregarded. CK5/6 is cytoplasmic, and D2-40 and HEG1 show membranous staining. While there is no validated standard for the percentage tumor cell staining required to be called “positive,” using 10% seems like a reasonable minimum.

The most common and generally most reliable epithelial markers are claudin-4 (Supplemental Figure 13), MOC-31 (Supplemental Figure 14), and Ber-EP4, all of which show membranous staining. Of note, claudin-4 may occasionally show dotlike cytoplasmic reactivity, which should be interpreted as negative. A variety of older markers (including CEA [carcinoembryonic antigen], CD15 [LeuM1], BG-8, and B72.3) also remain reliable options. When properly validated, each of these markers is greater than 80% sensitive for epithelial lineage (albeit less sensitive for sarcomatoid carcinoma) and greater than 80% specific in distinction from mesothelial tissues.^56,57  Note that diffuse MOC-31 and Ber-EP4 expression are more specific for epithelial lineage than patchy staining, as ∼10% to 15% of mesotheliomas show patchy MOC-31 or Ber-EP4 staining.^56,59 

Since none of these markers are perfectly sensitive or specific, it is recommended that, in addition to broad-spectrum cytokeratin, 2 mesothelial and 2 epithelial markers be included in a first-line immunopanel to establish mesothelial lineage. If results are concordant, the diagnosis can be considered established. If discordant, the immunopanel can be expanded for a second round of staining, with additional antibodies selected according to the differential diagnosis. A different tissue block can also be stained, if available. Given the range of reliable options and the likelihood of interlaboratory variation, no specific first-line antibody panel is recommended. Instead, each laboratory should test staining conditions for the antibodies of choice, ideally verifying sensitivity and specificity of at least 80% with appropriate controls.

Emerging evidence suggests that the novel mesothelial marker HEG1 and the epithelial marker claudin-4 may be sufficiently sensitive and specific to be used as a 2-marker panel to distinguish epithelioid mesothelioma from non–small cell lung carcinoma.⁵⁸  In the authors’ experience, claudin-4 is sufficiently reliable to serve as the sole epithelial marker in most differentials⁵⁹  (though claudin-4 is not expressed in proximal renal tubules or hepatocytes and is therefore not highly sensitive for renal cell or hepatocellular carcinoma). Despite its high sensitivity, HEG1 is not yet in widespread clinical use, and its expression in 50% of serous ovarian cancers and 100% of thyroid cancers limits its application. At present, it seems prudent to continue using immunopanels to establish mesothelial lineage.^59–62 

In addition to the broad-spectrum epithelial markers discussed above, immunomarkers specific to particular types of carcinoma are useful in certain differential diagnoses.

Tables 1 and 2 list markers that are useful in distinguishing epithelioid pleural mesothelioma from adenocarcinoma and squamous cell carcinoma of the lung, respectively. TTF-1 (thyroid transcription factor-1; 8G73/1 DAKO clone) and Napsin A are highly specific for lung adenocarcinoma in the differential with epithelioid mesothelioma.^63,64  WT1 is expressed in just 2% of squamous cell carcinomas and is therefore highly specific for epithelioid mesothelioma in this differential,⁶⁵  while claudin-4 and p40 (and, to a lesser degree, p63) are specific for squamous cell carcinoma in this setting.^65–69  p40 also assists in distinguishing squamous cell carcinoma from adenocarcinoma.⁷⁰ 

Because most breast carcinomas express estrogen receptor, gross cystic disease fluid protein-15, and/or mammaglobin, these markers are often useful in distinguishing metastatic breast carcinoma and mesothelioma.⁷¹  Calretinin and CK5/6 can be positive in high-grade basal-type breast carcinomas, which may also be negative for estrogen and progesterone receptor.^72,73  SOX10 expression favors breast cancer in this scenario. One-third to one-half of epithelioid mesotheliomas are positive for GATA3, limiting its usefulness in this scenario.^71,74 

Supplemental Table 1 lists markers that are considered useful in distinguishing mesothelioma and metastatic renal cell carcinoma. Because of their sensitivity and specificity, calretinin, D2-40, and CK5/6 are the best mesothelial markers in this context.⁷⁵  A panel of epithelial markers may be necessary, as claudin-4, MOC-31, and Ber-EP4 are expressed in 90%, 40%, and 40% of renal cell carcinomas, respectively. Carbonic anhydrase IX is expressed in virtually all epithelioid pleural mesotheliomas and therefore not useful to distinguish it from renal cell carcinomas.⁷⁶  PAX8 or PAX2 can be very useful, as these are expressed in most renal cell carcinomas but not in pleural mesotheliomas,^76–78  though PAX8 (using both polyclonal and monoclonal antibodies) is positive in ∼15% of peritoneal mesotheliomas.⁷⁹  The sensitivity and specificity of renal cell carcinoma (RCC) marker and CD15 for renal cell carcinoma are not high.⁷⁶  Most renal cell carcinomas express CD10, but half of epithelioid mesotheliomas are also positive.⁸⁰ 

Serous ovarian carcinomas are virtually always positive for WT1 and may be positive for HEG1 (50%), CK5/6 (∼30%), and D2-40 (∼20%).^56,81  Conversely, as noted above, ∼15% of peritoneal mesothelioma are PAX8 positive, more often in women (25%) than men (5%).^79,82  Estrogen and progesterone receptor expression is rare in peritoneal mesothelioma (7% and 2%, respectively) (Table 3).⁸²  Strong diffuse p53 (ie, >80% staining throughout tumor) does not exclude peritoneal mesothelioma, as 15% harbor TP53 mutation.⁸³ 

Adenocarcinomas of the gastrointestinal tract (Supplemental Table 2) and prostate can be distinguished from epithelioid mesotheliomas by the demonstration of CDX2⁸⁴  and prostate-specific antigen (and more recently NKX3.1),⁸⁵  respectively. SATB2 (a marker of colorectal adenocarcinoma and osteosarcoma) is typically negative in mesothelioma.⁸⁶ 

The role of broad-spectrum cytokeratin immunostains in the diagnosis of sarcomatoid mesothelioma is discussed above. In a cytokeratin-positive sarcomatoid malignancy, distinguishing sarcomatoid mesothelioma from sarcomatoid carcinoma requires a panel of immunomarkers, as sarcomatoid mesotheliomas are often negative for 1 or more mesothelial markers, and none of these markers is entirely specific. D2-40 and calretinin are each expressed in 50% to 60% of sarcomatoid mesotheliomas,^24,87–89  though staining may be focal, and specificity is limited, as D2-40 and calretinin each stain ∼20% to 25% of sarcomatoid lung carcinomas. D2-40 reactivity in entrapped lymphatics or reactive mesothelial elements is a potential pitfall. Claudin-4, MOC31, and BerEP4 also show low sensitivity for sarcomatoid areas of carcinomas—approximately 33%, 38%, and 23%, respectively, across studies.⁵⁹  Positive staining for TTF-1, Napsin A, or p40/p63 supports a diagnosis of sarcomatoid lung carcinoma. Diffuse GATA3 expression is seen in 70% of sarcomatoid mesotheliomas, but focal GATA3 expression can also be seen in sarcomatoid lung carcinoma,^90–92  while 29% of sarcomatoid urothelial carcinomas and 50% of sarcomatoid (metaplastic) breast carcinomas express GATA3. PAX8 is positive in 44% to 69% of sarcomatoid renal cell carcinomas but generally negative in sarcomatoid mesothelioma.⁹³  Conversely, sarcomatoid renal cell carcinoma is reportedly negative for CK5/6 and calretinin, though CK5/6 is particularly limited by its low sensitivity for sarcomatoid mesothelioma.⁷⁵ 

Synovial sarcoma is characteristically cytokeratin positive, and angiosarcoma and melanoma may also express cytokeratins. The diagnosis of synovial sarcoma can be confirmed with mutation-specific SS18-SSX and SSX-C-terminus immunostains, or with molecular confirmation of the distinctive X;18 translocation. Immunohistochemistry for TLE-1 can be used but is less specific.⁹⁴ 

After extensive workup and with appropriate clinical and radiologic features, 5% to 10% of sarcomatoid mesotheliomas are cytokeratin negative.^25,89,95,96  In the absence of convincing cytokeratin positivity, positive calretinin and/or D2-40 staining should not be interpreted as evidence of mesothelial differentiation, as these markers are variably positive in some sarcomas (including synovial sarcoma, malignant peripheral nerve sheath tumor, and angiosarcoma),^87,97  for which additional immunohistochemical markers would be warranted. The expanded differential might include epithelioid hemangioendothelioma, angiosarcoma, liposarcoma, myogenic or neurogenic sarcoma, undifferentiated pleomorphic sarcoma, malignant solitary fibrous tumor, melanoma, and histiocytic sarcomas. Morphology and clinical context should guide the differential and selection of appropriate immunohistochemical stains. Note that muscle-specific actin (HHF-35) and α-smooth muscle actin are often positive (occasionally diffusely) in sarcomatoid mesothelioma,⁹⁸  though desmin expression is quite rare in sarcomatoid mesothelioma.^98,99  Rare tumors cause diffuse pleural thickening, show extensive heterologous differentiation (eg, osteosarcomatous), and are negative for cytokeratin and mesothelial markers. After exclusion of a separate primary, these can be regarded as consistent with mesothelioma.

In current practice, histochemical staining is rarely used by experienced practitioners to distinguish mesothelioma from carcinoma (eg, in tumors expressing contradictory immunohistochemical markers). The cytoplasmic epithelial mucin in adenocarcinomas is positive by periodic acid–Schiff after diastase digestion (PAS-D) or by Alcian blue after hyaluronidase treatment. In contrast, cytoplasmic vacuoles in mesothelioma are generally negative by PAS-D, and the cytoplasmic hyaluronic acid in mesotheliomas stains positively with Alcian blue but is digestible by hyaluronidase.¹⁰⁰  PAS-D can be positive in hyaluronidase crystals. Mucicarmine can be positive in mesothelioma or adenocarcinoma and is not recommended for this distinction.

The ultrastructural features of mesothelioma are well described.¹⁰¹  However, the advent of novel immunomarkers has reduced the role of electron microscopy in routine diagnosis. Electron microscopy occasionally helps establish a diagnosis of mesothelioma when immunohistochemistry is equivocal, though tumors without diagnostic morphologic and immunophenotypic features of mesothelioma frequently lack specific ultrastructural findings, as well.^102,103  Formalin-fixed material retrieved from a paraffin block may be satisfactory, as microvilli and tonofilament bundles tend to be preserved. As routine diagnostic experience with electron microscopy wanes, such cases are likely best referred to subspecialists with expertise in this domain.

Separating benign from malignant mesothelial proliferations requires certainty that the process is mesothelial (see Establishing Mesothelial Lineage above). The diagnostic approach used when distinguishing reactive mesothelial hyperplasia from epithelioid mesothelioma differs from that used when distinguishing fibrous pleuritis from desmoplastic mesothelioma.¹⁰⁴  While morphology is paramount, a supportive immunophenotype is necessary for a definitive diagnosis of mesothelioma. The role of molecular studies in routine diagnosis is also evolving.

Reactive mesothelial proliferations may mimic mesothelioma, as they can show high cellularity, numerous mitoses, cytologic atypia, necrosis, papillary formations, and mesothelial entrapment within fibrosis, mimicking invasion (Figure 7).¹⁰⁴  Morphologic features that help distinguish reactive mesothelial hyperplasia from mesothelioma are summarized in Table 4.

Demonstration of tissue invasion is a key feature in diagnosis of mesothelioma (Supplemental Figure 15). Invasion by mesothelioma is often subtle, involving only a few layers of collagenous tissue subjacent to the mesothelial space and eliciting no obvious desmoplastic reaction. Invasion may be highlighted with immunostains, such as pancytokeratin or calretinin. However, inflammatory pleural processes can entrap mesothelial cells in granulation tissue deep to the pleura, typically arranged parallel to the pleural surface. Tubular collections of reactive mesothelial cells may also be seen, again parallel to the pleural surface. These patterns do not connote malignancy. Diagnosis is best performed in a well-oriented specimen (ie, cut perpendicularly to the pleural surface, including the full thickness of the pleura with adjacent adipose tissue, skeletal muscle, and/or lung parenchyma), as a tangential section (ie, taken parallel to the pleural surface) can give a false impression of a full-thickness mesothelial proliferation.

When a substantial amount of solid, malignant tumor (ie, tumor nodule[s]) with histologic features of mesothelioma is identified, the presence of invasion is not required for diagnosis. Tumor necrosis is also a feature of malignancy.

Identification of malignant features in desmoplastic mesothelioma requires adequate tissue, and large surgical biopsy samples are generally (but not always) needed. Features separating fibrous pleuritis from desmoplastic mesothelioma are shown in Table 5.

Fibrous pleuritis tends to show a uniformity of growth, with regular sheets and sweeping parallel fascicles of bland spindle cells that respect mesothelial boundaries. In contrast, sarcomatoid or desmoplastic mesothelioma shows disorganized growth, haphazardly intersecting fascicles, and expansile nodules of varying sizes with abrupt demarcation and changes in cellularity between nodules and their surrounding tissue (Figure 8, A). These different patterns can be highlighted by a broad-spectrum cytokeratin immunostain. Note that reactive mesothelial stroma is also keratin positive; keratin stain does not differentiate benign from malignant mesothelial proliferations; however, it can show the pattern of growth.

Stromal invasion is often more difficult to recognize in spindle cell than in epithelioid proliferations, as the invasive malignant cells in the former are often deceptively bland, resembling fibroblasts. Broad-spectrum cytokeratin staining is invaluable in highlighting bland, cytokeratin-positive malignant cells in areas where they would not normally be present (eg, adipose tissue, skeletal muscle, or lung or other visceral tissue in the pleura or peritoneum) (Figure 8, B). Although identification of invasion is often straightforward with the aid of broad-spectrum cytokeratin staining, fatlike spaces (termed fake fat) can be encountered in organizing pleuritis, probably reflecting artifactual changes in dense, fibrous connective tissue (Figure 9, A).¹⁰⁵  In such cases, horizontally oriented, cytokeratin-positive cells may be seen around the fatlike spaces (Figure 9, B). In addition, S100 protein, laminin, and collagen IV are usually positive in true adipose tissue and can help distinguish it from “fake fat,” which is negative for all 3 markers (Figure 9, C).

In small biopsy specimens, morphology alone may be inadequate for a definitive diagnosis of malignancy. Many immunostains previously purported to distinguish benign from malignant mesothelial proliferations—including GLUT1, IMP3, desmin, and epithelial membrane antigen (EMA)—are of little diagnostic value in individual cases. However, loss of nuclear BAP1 by immunohistochemistry, loss of cytoplasmic methylthioadenosine phosphorylase (MTAP) by immunohistochemistry, and homozygous deletion of CDKN2A by FISH, though not present in all cases of mesothelioma, are by definition never found in benign mesothelium (Supplemental Figures 16 and 17).^48,106–113  Both BAP1 and MTAP immunostains must be interpreted in the presence of a positive internal control, typically intratumoral inflammatory or stromal cells. These 3 techniques are very useful and can be applied in an algorithmic fashion in both tissue sections and cell block preparations (Figure 10).

Sensitivities of BAP1 immunohistochemistry, MTAP immunohistochemistry, and CDKN2A FISH depend on both tumor histologic subtype and primary site. Loss of nuclear BAP1 staining is seen in 60% to 70% of epithelioid pleural mesotheliomas but just 20% of sarcomatoid tumors.^{48,107,112,114}  Conversely, CDKN2A deletion and corresponding loss of cytoplasmic MTAP staining are seen in 60% and 40% of epithelioid pleural mesotheliomas, respectively, compared to 93% and 75%, respectively, of sarcomatoid pleural mesotheliomas.^107,115,116  Recent data indicate that MTAP immunohistochemistry may be significantly less sensitive than CDKN2A FISH in desmoplastic mesothelioma, given its scant cytoplasm, highlighting the continued role of CDKN2A FISH in clinical diagnosis.¹¹⁶ 

By primary site, BAP1 loss is seen in 50% to 60% of pleural and 60% to 70% of peritoneal mesotheliomas. Conversely, MTAP loss is seen in 50% of pleural compared to just 5% to 10% of peritoneal mesotheliomas, reflecting the different proportions of histologic subtypes and relative rates of underlying CDKN2A deletion at these sites.^48,83 

BAP1 loss can also support a diagnosis of mesothelioma in the differential with certain carcinomas. In particular, BAP1 loss is seen in less than 1% of lung cancers and serous ovarian carcinomas.^117,118  However, BAP1 loss is not specific for mesothelioma in isolation or in all contexts, as it may be seen in ∼15% of clear cell renal cell carcinomas and in subsets of intrahepatic cholangiocarcinoma, thymic carcinoma, and melanoma. MTAP loss does not help distinguish mesothelioma from other malignancies in any reported context.

p53 immunostaining has a controversial history in diagnosis of mesothelioma, largely because of prior misunderstanding about wild-type versus aberrant staining, but recent data indicate that strong diffuse p53 immunostaining (ie, in ≥80% of tumor cells) can be seen in mesothelioma but not in reactive mesothelial proliferations. Sensitivity is limited, with strong diffuse p53 staining seen in only ∼10% to 15% of pleural mesotheliomas, including rare cases in which this might be the only immunohistochemical evidence for malignancy.^48,119  There is conflicting data on the specificity of null-pattern (ie, complete absence of) p53 staining for underlying TP53 mutation.

Homozygous or heterozygous NF2 deletion (detected by FISH or molecular sequencing) is specific for mesothelioma in the differential with reactive mesothelial proliferation.¹²⁰  In recent reports, immunohistochemical loss of the NF2 protein product, Merlin, correlates strongly with underlying NF2 mutation, and Merlin loss has been detected in 40% to 50% of pleural mesotheliomas.^48,121  Preliminary results are promising, but full endorsement of Merlin immunohistochemistry for this application awaits additional data.

Morphology alone cannot always reliably distinguish biphasic mesothelioma from epithelioid mesothelioma associated with a reactive spindled mesothelial population. Immunohistochemical and molecular studies can attempt to establish a clonal relationship between the epithelioid and spindled populations. Although studies using this approach have shown variable results,^34,122–124  recent data show strong concordance for BAP1 staining pattern (ie, lost versus retained) in the epithelioid and sarcomatoid components of biphasic mesothelioma. In contrast, MTAP staining is more often discordant, with retained expression in the epithelioid component and loss in the sarcomatoid component.⁴⁸  At present, it is recommended that BAP1 or MTAP loss (or CDKN2A deletion by FISH) in the spindle cell component of a biphasic mesothelial proliferation be regarded as evidence of malignancy (ie, supporting diagnosis of biphasic mesothelioma). Conversely, if BAP1 or MTAP loss is confined to the epithelioid component (ie, retained expression in the spindled component), a diagnosis of biphasic mesothelioma should be made only if the spindled population shows unequivocal morphologic features of malignancy.

Tumor molecular profiling with large next-generation sequencing panels has the potential to provide diagnostic, prognostic, and therapeutic information in a single assay. In one study of resection specimens, a 447-gene next-generation sequencing panel showed 95% sensitivity for diagnosis of mesothelioma.⁴⁸  At present, routine genomic sequencing of mesotheliomas is performed only in select referral or academic centers, and it is not currently recommended for routine clinical use. Immunohistochemical studies remain the ancillary assay of choice, with targeted molecular studies (eg, CDKN2A FISH, ALK FISH, sequencing) in select cases.

The International Collaboration on Cancer Reporting has recently published a 3rd edition to their guidelines for reporting mesothelioma, which provide a valuable resource for routine practice.¹²⁵ 

The Union for International Cancer Control and American Joint Committee on Cancer (AJCC) Cancer Staging Manual, 8th edition,¹²⁶  represents the most widely applied TNM system and should be reported for all pleural mesotheliomas resected via extended pleurectomy/decortication or extrapleural pneumonectomy (now rarely performed). The TNM staging system for pleural mesothelioma evaluates resectability but is generally not a good predictor of prognosis. Importantly, the AJCC 8th edition does not include mesothelioma in situ. There is no consensus TNM staging for peritoneal, pericardial, or paratesticular mesothelioma.

The 2021 WHO classification⁴  also recognizes a variety of important pathologic factors beyond tumor histologic subtype (ie, favorable prognosis for epithelioid, intermediate prognosis for biphasic, and poor prognosis for sarcomatoid tumors).^10,127  As noted above, certain architectural, cytologic, and stromal features are linked to prognosis, and pathologists should routinely report this information.¹¹  In brief, prognostically favorable morphologic findings include tubulopapillary, trabecular, and adenomatoid architecture; low nuclear grade; high tumor-associated immune microenvironment (eg, lymphohistiocytoid cytologic features); and myxoid-rich stromal matrix.^12,13,128  Conversely, adverse prognosis is associated with any micropapillary or greater than 50% solid architecture; high nuclear grade; rhabdoid, pleomorphic, transitional, or desmoplastic morphology; and necrosis.^15–17,29 

Several studies have validated a 3-tiered grading system for epithelioid mesothelioma of the pleura and peritoneum, based on mitotic activity and nuclear atypia.^8,12,128  After clinician input at a multidisciplinary meeting of mesothelioma experts, necrosis was added to mitotic activity and nuclear atypia to create a 2-tiered grading system (ie, low versus high grade), which better facilitates clinical decision-making (Table 6). The prognostic significance of the 2-tiered system has been validated in a large series of pleural mesothelioma,^11,26  though its applicability to peritoneal mesothelioma remains unclear.⁸  When assigning nuclear grade, tumor foci with the highest-grade features should be used. The current 2021 WHO classification recommends routine reporting according to this 2-tiered grading system for biopsies and resections of epithelioid diffuse pleural mesothelioma.⁴ 

Homozygous deletion of CDKN2A and MTAP loss by immunohistochemistry both portend poor prognosis (shorter overall and disease-free interval) among mesothelioma cases.^36,129,130  In contrast, loss of nuclear BAP1 by immunohistochemistry is a favorable prognostic marker, at least partly reflecting the significantly improved prognosis and treatment responsiveness in patients with germline BAP1 mutation.^131,132  Rates of CDKN2A deletion, MTAP loss, and BAP1 correlate with tumor histologic subtype and primary site (see Distinguishing Benign Versus Malignant Mesothelial Proliferations above).

Germline testing should now be considered for all patients with mesothelioma, as it affords improved response to platinum-based chemotherapy and relatively favorable prognosis (despite advanced stage at presentation), potential access to novel therapies, and genetic counseling for the patient and family, which is relevant to surveillance for other tumors.^132–134 

Genomic profiling indicates that at least 12% of mesotheliomas arise in carriers of pathogenic germline mutations.^{45,132,133,135} BAP1 germline mutations (which cannot be distinguished from somatic mutations by BAP1 immunohistochemistry) account for approximately half of such cases. The prevalence of germline predisposition is higher in younger patients (ie, germline mutations identified in >50% of patients with mesothelioma and younger than 50 years¹³⁴ ), in those with peritoneal disease, in tumors with low-grade epithelioid morphology and a high tumor immune response, in patients with longer overall survival (median survival >5 years¹³⁵ ), and in those with a personal or family history of multiple cancers (especially melanoma, clear cell renal cell carcinoma, and breast cancer).¹³² 

Pathogenic germline mutations in patients with mesothelioma most commonly affect DNA damage repair pathways, which serve as a potential therapeutic target for poly (ADP-ribose) polymerase (PARP) inhibitors.¹³⁶  Use of PARP inhibitors alone or combined with platinum-based regimens in mesotheliomas with germline homologous recombination defects is under evaluation.

US Food and Drug Administration approval for combined nivolumab and ipilimumab in untreated unresectable diffuse pleural mesothelioma followed publication of the CheckMate 743 trial.¹³⁷  The benefits are primarily observed among patients with nonepithelioid disease, which is typically most refractory to conventional chemotherapy. By immunohistochemistry, PD-L1 is positive (>1% tumor cell staining) in 10% to 49% of epithelioid, 9% to 67% of biphasic, and 22% to 100% of sarcomatoid mesotheliomas,^138–140  with some variability between PD-L1 antibody clones.^139,140  Routine PD-L1 immunostaining is not currently indicated for mesothelioma.

ALK rearrangements can be identified by immunohistochemistry or molecular testing in a small number of mesotheliomas, with a predilection for children and young adults, women, and peritoneal tumors.^40,141,142 ALK rearrangement appears mutually exclusive with other genetic events commonly observed in mesothelioma. ALK-fusion–positive peritoneal mesotheliomas afford a small group of patients access to novel targeted treatment with tyrosine kinase inhibitors, with reportedly dramatic treatment response.¹⁴³  There is no formal guideline on screening for ALK rearrangement in mesothelioma, but given its clinical implications, it appears reasonable to perform ALK immunohistochemistry in young patients and patients with peritoneal tumors, particularly if other molecular alterations (eg, BAP1 loss, MTAP loss, Merlin loss) are not detected.

Mesothelioma in situ (MIS) as the noninvasive precursor to diffuse mesothelioma was initially proposed in 1992.¹⁴⁴  The morphology of MIS is now recognized as variable and includes flat or cuboidal cells with or without cytologic atypia, small or complex papillary proliferations, or small surface nodules with moderate to severe cytologic atypia. Invasion is absent by definition, and there must be no clinically or radiographically identifiable mass lesion or diffuse process. MIS cannot be diagnosed by morphology alone, and immunohistochemical loss of nuclear BAP1 and/or demonstration of CDKN2A homozygous deletion (by FISH or by MTAP immunohistochemistry) must also be demonstrated on a rigorously validated assay with appropriate controls (Figure 11, A through D).^{47,111,112,115,145–147} 

There are no published criteria on minimum acceptable sample size, but caution is warranted when biopsy samples are very small or crushed. The WHO recommends thoracoscopic evaluation with large biopsy specimens (ideally 100–200 mm²) from different areas of the pleura in patients with nonresolving effusions.⁴  Similarly, no criteria are published for which samples should be tested for BAP1, MTAP, and CDKN2A alterations, but a low threshold is suggested for patients with unexplained recurrent effusions, history of occupational exposures, genetic predisposition, history of chest radiation, or atypical histologic features.

The differential diagnosis of MIS includes reactive mesothelial atypia and well-differentiated papillary mesothelial tumor (WDPMT), depending on lesional architecture.^148,149  BAP1 and MTAP immunostains should be performed in WDPMT-like lesions identified on investigation of an effusion or related symptoms, with aberrant results supporting MIS with WDPMT-like morphology, in the correct clinical context. Absence of BAP1, MTAP, and CDKN2A alterations does not exclude MIS, and an expert opinion is advisable in difficult cases, given the potential for malignant behavior.

The WHO classification⁴  currently only describes MIS in the pleural space, but peritoneal, pericardial, and paratesticular presentations are described, and the same diagnostic criteria can be applied.^150–152  At present, MIS has only been established for epithelioid mesothelioma.

The 2021 WHO classification⁴  emphasizes that MIS is a multidisciplinary diagnosis. Communication with the clinical team is especially important because time to progression may range from 1 year^146,149  to 15 years.¹⁵³  No treatment guidelines exist currently.

Up to 90% of patients with diffuse pleural mesothelioma present with a pleural effusion. Consequently, cytology fluid specimens are often the first (and in patients who cannot tolerate additional procedures, the only) sample available for diagnosis.

The typical cytologic features of epithelioid mesothelioma were described more than 50 years ago and have been refined in numerous subsequent publications.^154–156  Mesothelioma can manifest cytologically as a highly cellular effusion with obvious nuclear atypia, numerous large tissue fragments and cell clusters, or as cellular fluid with single and clustered mesothelial cells exhibiting only subtle atypia, overlapping with reactive mesothelium. Cell-in-cell arrangements, a hump at the cell periphery, multinucleated cells, papillary groups with basement membrane cores, and orangeophilic cells are features concerning for mesothelioma (Figure 12, A and B). Sarcomatoid mesotheliomas generally do not shed malignant cells into effusions.

As in tissue specimens, the diagnosis of mesothelioma in cytology specimens is a 2-step process in which both mesothelial lineage and malignancy must be established.^157–159  Immunocytochemistry plays an invaluable role in both steps and can be performed on smears or cell blocks, although use of cell blocks is preferable, as they undergo processing similar to formalin-fixed, paraffin-embedded tissue specimens, thereby enabling preparation of serial sections for immunostaining and/or molecular studies. Given that cell blocks can be prepared in various ways, it is important to recognize that differences in fixation can affect immunostaining results and to use appropriate controls.^53,54  When cell blocks are not prepared, the cell-transfer technique (in which a Papanicolaou-stained sample is divided into several pieces and transferred to multiple slides) can facilitate multiple immunocytochemical stains on limited materials.¹⁶⁰ 

The same antibodies used to distinguish mesothelial from epithelial lineage in tissue samples (see Establishing Mesothelial Lineage above) can be applied to cytology specimens, provided they have been validated for this application in the responsible laboratory. As in tissue specimens, it is currently recommended to use 2 epithelial and 2 mesothelial markers, though claudin-4 immunocytochemistry can be used as the sole epithelial marker if well validated. When a cytology specimen comprises single cells or loose aggregates, the differential diagnosis may also include melanoma, lymphoma, sarcoma, germ cell, and other tumors, and the immunopanel should be tailored to the differential diagnosis.

BAP1 and MTAP immunocytochemistry, and CDKN2A FISH have been validated for use in cell block preparations (Supplemental Figure 18).^{108,110,154,158,159,161,162}  In distinction of mesothelioma from reactive mesothelial proliferation in pleural effusion cytology specimens, a recent meta-analysis of 65 studies found 100% specificity for CDKN2A homozygous deletion by FISH, and 99% specificity for both BAP1 loss and MTAP loss by immunocytochemistry. The same study found a diagnostic sensitivity of 83% for the combination of BAP1 immunocytochemistry and CDKN2A FISH.¹⁶³  FISH for NF2 deletion is also reportedly specific for mesothelioma in effusion cytology, though not yet in widespread use.^120,164  Although previously viewed as potentially useful markers, it is now clear that positive staining for EMA, IMP-3, CD146, or GLUT1 alone is insufficient to diagnose a cytology specimen as malignant.

The inability to assess stromal invasion and sarcomatous elements, coupled with the grim prognosis and expensive and toxic therapy associated with the diagnosis, has contributed to a general reluctance to render a primary definitive diagnosis of mesothelioma based solely on effusion cytology. The reported sensitivity of cytologic diagnosis for mesothelioma is 30% to 75%, though specificity is 99% to 100%.^157,165  Sensitivity is almost certainly lower for sarcomatoid mesothelioma, which does not typically shed cells in effusions. Although the epithelioid component of biphasic mesothelioma can shed mesothelioma cells into effusions, biphasic mesothelioma cannot be differentiated from epithelioid mesothelioma on cytology effusion. Additionally, cytology alone cannot distinguish invasive mesothelioma from MIS,¹⁶⁶  and correlation with clinical and radiographic findings is recommended in all cases. Architectural subtyping and grading of epithelioid mesothelioma also cannot be performed on cytology specimens.

The above discussion pertains principally to effusion cytology, which is by far the most common setting for a cytologic diagnosis of mesothelioma. Similar considerations apply to other types of cytology samples. Mesothelioma may rarely be diagnosed in sputum or bronchial washing, lavage, or brushing, and a few studies from the 1980s report use of transthoracic or endoscopic bronchial ultrasound-directed fine-needle aspiration biopsy (FNAB) for diagnosis of mesothelioma. Though rarely reported, definitive primary diagnosis of mesothelioma using these techniques is possible in the appropriate clinical and radiologic context. The cytomorphology in such cases resembles that of mesothelioma in effusions, with high cellularity and papillary clusters, though other features characteristic of mesothelioma in effusions (eg, cell-in-cell arrangements, multinucleated cells) are less commonly observed. Because FNAB directly samples a clinical mass lesion, sarcomatoid mesothelioma is more likely to be diagnosed in an FNAB specimen than in effusion cytology.

The WHO Classification of Tumors of the female genital tract (which includes peritoneal mesothelial tumors) encourages the diagnostic term peritoneal inclusion cyst and discourages use of multicystic mesothelioma (and similar terms) to avoid confusion with (malignant) mesothelioma.⁵  Peritoneal inclusion cyst(s) may comprise 1 or multiple cysts lined by bland mesothelial cells without significant stratification, papillary formations, or infiltration of soft tissues. Lesions can be unifocal or multifocal within the pelvis and abdomen. Local recurrence rates as high as 50% have been reported in studies of florid multifocal lesions,¹⁶⁷  though a recent study including a more representative population of peritoneal inclusion cysts, as currently defined in the WHO, found a local recurrence rate of just 3%.¹⁶⁸  It remains unclear whether peritoneal inclusion cysts represent reactive or neoplastic lesions, though BAP1 and MTAP are universally retained.¹⁶⁹ 

Well-differentiated papillary mesothelial tumor (WDPMT) occurs principally in the peritoneum and is often an incidental surgical finding. Tumors are generally unifocal and small (<2 cm), though larger and/or multifocal examples are reported with otherwise classic morphology.¹⁷⁰  WDPMT is composed of slender papillae with hyalinized to myxoid cores lined by a single layer of bland mesothelial cells (Figure 13).¹⁷¹  Mitoses are rare to absent. Infiltration of underlying soft tissue is absent by definition, though occasional WDPMTs show so-called invasive foci, characterized by confluent papillary growth and/or percolating mesothelial nests/cords within papillary stroma (ie, confined to the WDPMT).¹⁷²  Recurrent mutations in TRAF7, CDC42, EHD1, ATM, FBXO10, and SH2D2A have been reported.^173,174 BAP1 and CDKN2A alterations are absent, with BAP1 and MTAP retained by immunohistochemistry.¹⁶⁹  Most (60%–95%) are positive for PAX8.^171,175 

“Invasive foci” and multifocality are associated with increased recurrence risk and should be reported when present. Malignant transformation is rare and can occur years or decades after original diagnosis. It is uncertain whether such cases represent true WDPMTs, versus MIS with WDPMT-like morphology and/or early epithelioid mesotheliomas with WDPMT-like foci.¹⁷⁶  BAP1 immunohistochemistry (and, if necessary, MTAP immunohistochemistry or CDKN2A FISH) should be performed on WDPMT-like lesions discovered during clinical workup for effusion or associated with multifocal or diffuse serosal involvement. BAP1 or MTAP loss supports a diagnosis of WDPMT-like mesothelioma or MIS, but retained staining does not exclude mesothelioma and requires multidisciplinary correlation.¹⁴⁸ 

Adenomatoid tumors are small, circumscribed, nodular lesions, most often involving the uterus and fallopian tube, and rarely other peritoneal sites.^177,178  An association with immunosuppression has been noted.¹⁷⁷  Tumors comprise acini, cords, and nests of plump to flattened, bland mesothelial cells. Single cells may be noted, often with a signet ring appearance. Stringlike bridges characteristically span tumor lumina, which may contain hyaluronic acid–rich myxoid material. BAP1 and MTAP are retained.¹⁷⁹ TRAF7 mutations have been reported in adenomatoid tumors at various sites.¹⁷⁷ 

Rare peritoneal mesothelial lesions show mixed features of peritoneal inclusion cyst, WDPMT, and adenomatoid tumor. The pathogenesis and prognosis of these hybrid lesions remain unclear.¹⁷¹  Occasional noninvasive, morphologically bland, and unifocal or oligofocal nodular mesothelial proliferations defy easy classification. A provisional entity of “solid papillary mesothelial tumor” attempts to encompass some of these lesions, which appear to show indolent behavior.¹⁸⁰  In such cases, a descriptive diagnosis with recommendation for clinical and radiographic correlation with close follow-up is generally appropriate.

This article provides broad guidelines for diagnosis of mesothelioma, which, though uncommon, carries a grave prognosis and frequently has medicolegal implications. We emphasize that morphology remains the cornerstone for classification of mesothelial proliferations in both biopsy and resection specimens. Numerous prognostically and clinically significant morphologic features are discussed, and these should be routinely reported in the diagnostic report, whenever possible. Immunohistochemistry and molecular studies play a growing and evolving role in diagnosis and management of mesothelioma. Immunohistochemical panels are routinely applied to establish mesothelial lineage, distinguish mesothelioma from malignant mimics, and distinguish mesothelioma from reactive mesothelial proliferations. Specific immunopanels should be tailored to the clinical and morphologic differential, and immunostains should be carefully validated for optimal performance. Molecular studies, including FISH and targeted sequencing panels, are useful in challenging cases with nondiagnostic morphologic and immunophenotypic findings. These same principles apply to the cytologic diagnosis of mesothelioma, with recognition that cytology does not permit evaluation of stromal invasion, that tissue fixation protocols may impact immunostain performance in cytology preparations, and that sarcomatoid mesothelioma does not typically shed in effusions. Importantly, the pathologist must always correlate morphology and ancillary study results with clinical, radiographic, and operative findings. Despite (or perhaps in part because of) rapid evolution in the field of mesothelioma diagnosis, this remains a challenging and evolving area of surgical and cytopathology, and expert opinion should be sought in difficult cases, when needed.

This article has been endorsed by the Board of the International Mesothelioma Interest Group.