Dense fibrosis in chronic pleuritis can be difficult to distinguish from the desmoplastic type of sarcomatoid mesothelioma, especially in small biopsies. Imaging and the operative note can provide important clues, as chronic fibrinous pleuritis typically involve first the visceral pleural surface, whereas mesotheliomas the parietal pleural surface. Grossly and histologically, fibrous pleuritis is typical of relatively uniform thickness, and the reactive spindled mesothelial cells are generally parallel to the surface (figure 8). In contrast, mesothelioma forms cartwheels and storiform structures, and grow into the chest wall at right angles. The pitfall of the “fake fat” artifact is well known, (2) and acellular scarring in itself (without cytokeratin positive spindle cells) can extend around real fat (Figure 9) Artifactual spaces can be confirmed to be fake fat by immunostaining for S-100 or calretinin but is rarely necessary if other histologic features are considered.

The pleomorphic and transitional patterns of epithelioid mesothelioma may be difficult to distinguish from sarcomatoid mesothelioma. Transitional pattern of epithelioid mesothelioma has been described as “sheet-like growth of cohesive, plump, elongated epithelioid cells with well-defined cell borders and a tendency to transition into spindle cells” (9) Another feature of epithelioid mesothelioma that is often difficult to distinguish from sarcomatoid component is the presence reactive fibrous stroma. In contrast to reactive fibroblasts, sarcomatoid neoplastic cells have frequent mitotic figures, atypical mitotic figures, and cytokeratin expression. BAP-1 loss by immunohistochemistry occurs in about one-half of biphasic mesotheliomas, of which a further one half (25% of total) show loss in the sarcomatoid component, which is also helpful in the distinction from reactive fibroblasts. (10,11)

Biphasic mesotheliomas are often treated like sarcomatoid mesotheliomas, depending on the percentage of the sarcomatoid component. Therefore, identifying a malignant spindled area of an epithelioid mesothelioma is important, its presence could prevent a patient from receiving surgical treatment.

Pleomorphic carcinoma of the lung is not always distinguishable histologically from sarcomatoid mesothelioma, but based on imaging and location of the tumor, this distinction is easily made in most cases. If there are areas of pleural plaques indicative of asbestos exposure, then desmoplastic mesothelioma is far more likely than pleomorphic carcinoma. As noted above, probably the only immunohistochemical markers that is useful is GATA3, which if diffusely positive points strongly towards mesothelioma. MTAP nuclear loss (see below), and deletion of CDKN2A are not useful in differentiating carcinoma from mesothelioma, as both can occur carcinomas of the lung and other organ sites.

These are usually straightforward to diagnose and are rarer even than mesothelioma. Malignant solitary fibrous tumor is usually a discrete, well-rounded mass, and projects into the lung parenchyma, and does not spread along pleural surfaces. In distinction to localized sarcomatoid mesothelioma, some areas should have histologic features and immunohistochemical staining patterns that would confirm a diagnosis. These include positivity for stat6, and negativity for pancytokeratin. Synovial sarcoma is the most common high-grade pleural-based sarcoma, and is readily diagnosed by molecular testing for X;16 translocation. In general, synovial sarcomas do not have a collagenous background, and have a monomorphous, cellular “dark blue” appearance that is quite different from sarcomatoid mesothelioma. Angiosarcomas may occur in the pleura, but are histologic quite distinct from mesothelioma, both in histologic appearance and immunoprofiles. Vascular markers have been shown to be uniformly negative for vascular markers including CD31. (12) An exception to this generalization are rare epithelioid hemangioendotheliomas that may mimic mesothelioma and be associated with asbestos exposure. (12) Undifferentiated sarcomas of the pleura are extremely rare, and if a tumor has the gross growth pattern of mesothelioma and is composed of pleomorphic spindle cells lacking any consistent markers by immunohistochemistry, other than patchy cytokeratin positivity, the tumor should probably be considered a sarcomatoid mesothelioma over a primary pleomorphic sarcoma of the pleura, especially if there is occupational history. Undifferentiated sarcoma of the pleura should be diagnosed only if there is no evidence of epithelial differentiation with negative markers for pancytokeratins.

BAP-1 (BRCA associated protein-1} mutations are detected by loss of nuclear staining by immunohistochemistry. Granular cytoplasmic staining is seen in about one-fourth of tumors that lose nuclear expression. There is abundant information on BAP-1 loss in mesotheliomas of various sites. A series of a large number of pleural mesotheliomas in showed that 22% of sarcomatoid mesothelioma showed loss of BAP-1, (11) although rates as low as 7% have been reported. (13) It is generally accepted that 60% of mesotheliomas overall show loss of BAP-1, with 1–2% of patients demonstrating germline mutations and a predisposition to breast and ovarian cancers. In a young woman with mesothelioma, germline mutations of BRCA1 should be suspected. BAP-1 loss does not have prognostic implications. (3)

Allelic loss of CDKN2A (p16^INK4A; 9p21.3) is generally detected by fluorescent in situ hybridization. There are various thresholds for heterozygous loss, homozygous loss, and wild type. In one study, a cut-off of 20% was used as a mean of all nuclei counted, with both alleles absent in >20% of nuclei for homozygous loss, or one absent in >20% of nuclei for heterozygous loss. (14) In another study, cutoff values for homo- and heterozygous loss were set at 10% and 47%, respectively, based on studies on reactive mesothelial proliferations in their laboratory. (15) FISH interpretation reports generally state these mean percentages and provide the thresholds used in the analysis. (In the case reported here, the value was 38% single allele loss, and 15% double allele loss).

In the largest series, 81% of sarcomatoid mesotheliomas of the pleura had homozygous loss of CKDN2A. In a study of biphasic tumors, 95% showed homozygous loss (in both morphologies in all cases). (16) Up to 15% of malignant mesotheliomas can show heterozygous loss, the actual value depending on the thresholding and length of the probe used. (15) In general, it is stated that p16 deletion (generally indicating homozygous) occurs in 90–100% of sarcomatoid types, and approximately 70% of epithelioid and biphasic type (3,17).

There is some evidence that prognosis of pleural mesothelioma is worse with homozygous loss at the CDKN2A locus compared to wild type, with heterologous loss imparting an intermediate prognosis. (15,18)

Immunohistochemical staining for MTAP has a high correlation with FISH results, and is a useful adjunct for diagnosis and, with BAP1, an essential immunostain for pathologists who see reasonable numbers of mesotheliomas. A recent large series showed, for all morphologies, a 78% sensitivity and 96% specificity for CDKN2A homozygous deletion for all mesotheliomas, without separation of sarcomatoid tumors. (19) There is little data on MTAP expression loss in sarcomatoid mesothelioma. It would be expected to be similar to FISH results for CDKN2a deletions (i.e., 80–90%). In Kinoshita et al’s series, 24 of 30 sarcomatoid mesotheliomas of the pleura showed nuclear loss. (20) In our experience, evaluation of nuclear staining loss is difficult in sarcomatoid and biphasic mesotheliomas, in contrast to BAP-1. (Figure 10)