Secondary Logo

Journal Logo

Special Article

The Separation of Benign and Malignant Mesothelial Proliferations

New Markers and How to Use Them

Churg, Andrew MD*,†; Naso, Julia R. MD, PhD*,†

Author Information
The American Journal of Surgical Pathology: November 2020 - Volume 44 - Issue 11 - p e100-e112
doi: 10.1097/PAS.0000000000001565
  • Free


The separation of benign from malignant mesothelial proliferations is an important but often difficult problem in surgical pathology and one that is crucial to patient care, since the prognosis of pleural mesotheliomas is poor and these tumors are incurable, while the prognosis of some cases of peritoneal mesothelioma is considerably better but it is nonetheless unclear whether any peritoneal mesothelioma patients are actually cured.

Hematoxylin and eosin morphology plays a fundamental and important role in this separation, and remains the first approach, but in some proportion of cases routine morphology cannot separate benign reactive from malignant mesothelial proliferations.1 A further complication is that there are now reliable diagnostic criteria for the entity of malignant mesothelioma in situ, a malignant process that morphologically mimics a benign reaction and requires a confirmatory technique for diagnosis (see the Malignant mesothelioma in situ section below).

Over the approximately last 25 years a number of adjunctive markers have been proposed to help differentiate benign reactions and mesotheliomas. We2 have previously reviewed many of the older immunohistochemical (IHC) markers including EMA, desmin, IMP-3, and GLUT-1. Some of these, for example EMA and desmin staining, are simply based on empiric observations with no known mechanistic basis and do not provide consistent separation of benign from malignant. Others, such as GLUT-1 staining, reflect a known metabolic process in malignant neoplasms (aerobic glycolysis) but are difficult to interpret and not 100% specific for malignancy even when positive. Our general conclusion was that some of these stains provide statistical differences for groups of cases, but there is too much overlap between malignant and reactive processes to recommend their use in individual patients.

In the last roughly 10 years a number of new markers based largely on documented genomic alterations in malignant mesotheliomas have been proposed. Although the established ones are reliable in terms of specificity, use of these new markers is not straightforward because there are considerable variations in sensitivity depending on body cavity (pleural vs. peritoneal) and tumor morphology (epithelial vs. sarcomatous), and it is crucial to know exactly which fluourescent in situ hybridization (FISH) or staining pattern, for example, nuclear versus cytoplasmic loss, is diagnostic for each marker. Furthermore, no single marker is 100% sensitive, so that combinations need to be employed. At this point it is clear from consultation material that many pathologists are confused about the application and interpretation of these markers.

This review will cover these markers and provide rules for when and how to apply them. We will divide the review into markers for which there are a sufficient number of publications to consider the marker as established, and those for which there are very few publications or the data are contradictory, so that they should be regarded as proposals rather than clearly reliable tests. A summary of applications is provided in Table 1.

TABLE 1 - Separation of Benign From Malignant Mesothelial Proliferations by Test, Site, and Morphology
Markers Utility for Epithelial or Sarcomatous Pleural Mesothelioma Utility in Peritoneal Epithelial Mesothelioma (Compared With Pleural Epithelial Mesothelioma) Utility in Effusion Cytology Specimens Diagnostic Interpretation of a Positive Test Utility for Distinguishing Mesothelioma From Other Malignancies Advantages and Disadvantages/Comment
BAP1 IHC Sensitivity greater for epithelial (~70%) than sarcomatous (<20%) Appears to be similar to pleural (little data) Works well in effusion samples, but alcohol fixatives leach BAP1 Loss of nuclear staining (in most cases 100%) with positive internal control Nuclear loss provides high specificity for a diagnosis of mesothelioma vs. high grade serous carcinoma or non–small cell lung carcinoma Internal positive control (inflammatory or stromal cells) required for interpretation Sensitivity improved by combination with other markers (see Table 3)
CDKN2A (p16) FISH Useful for both Less sensitive (<50% for peritoneal vs. ~60%-80% in pleura) Works well in effusion samples Can be used on smears Homozygous deletion above laboratory truncation background (truncation background varies from laboratory to laboratory). Above 30% always malignant Not useful Section truncation background values need to be established by each laboratory that does test Selecting exact areas to probe can be difficult
MTAP IHC Useful for both Likely less sensitive given that CDKN2A loss is less prevalent in peritoneal compared with pleural mesotheliomas Works well in effusion samples, but MTAP may be leached by alcohol fixatives Loss of cytoplasmic staining (in most cases 100%) with positive internal control Not useful Good correlation with CDKN2A FISH. Nuclear loss is not specific. Cytoplasmic loss can be difficult to interpret in spindle cell proliferations Internal positive control (inflammatory or stromal cells) required for interpretation Implication of patchy loss of staining is uncertain
NF2 FISH Potentially useful for both Potentially useful No data Hemizygous deletion or monosomy 22, but extant data are contradictory and truncation values uncertain Not useful Section truncation background values need to be established Published data based on hemizygous loss or monosomy 22. Literature is inconsistent
Cyclin D1 IHC Useful for both (using different thresholds) No data No data Diagnosis of malignancy supported by >50% nuclear staining for epithelioid mesothelioma or >75% nuclear staining for sarcomatous mesothelioma Not useful Spindle cell processes sometimes difficult to interpret and may be confounded by vascular staining. Endothelial staining serves as an internal positive control Only a single publication
NF2 (Merlin) IHC Not useful No data No data No data Not useful Fails to separate benign from malignant Only a single publication
LATS1/2 IHC Not useful No data No data No data Not useful Fails to separate benign from malignant Only a single publication
YAP/TAZ IHC Not useful No data No data No data Not useful Fails to separate benign from malignant Only a single publication
EZH2 IHC Useful for both No data No data >50% of nuclear staining supports a diagnosis of malignancy Not useful Only 2 publications which are not completely consistent in suggested method of interpretation
5-HMC IHC Useful for both No data No data >50% nuclear loss indicative of malignancy Not useful Double immunostain (5-HMC+CAM5.2) required for accurate interpretation of spindle cell proliferations Only a single publication
PD-L1 IHC Useful for sarcomatous mesotheliomas No data No data Strong diffuse membranous staining in tumor cells (or >50% staining) indicative of malignancy Not useful Weak or negative staining not diagnostic Only 2 publications—both use antibody 22C3. No information on other antibodies
CD47 IHC Useful for epithelial mesotheliomas No data Works on cytology specimens Works in effusion samples Not useful Only a single publication (on cytology specimens). Not recommended for sarcomatous tumors because spindle cell reactive proliferations stain

One important point to bear in mind is that these tests need to be applied to processes that are first documented to be mesothelial, because, with the arguable exception of BRCA-1 related protein-1 (BAP1), most of the genomic abnormalities to be discussed are not specific to mesotheliomas but occur in many different types of malignant neoplasms.3

Given the fact that staining results can change a diagnosis from benign to malignant, the importance of properly validating IHC stains must also be stressed, and particularly ensuring that positive internal control staining is present for those stains that normally have a positive internal control (eg, BAP1, methylthioadenosine phosphorylase [MTAP]). Ideally the on-slide control run with the test sample should include both normal and abnormal tissue


BAP1 IHC Staining

BAP1 is a deubiquitinase with a complex set of functions including DNA repair, cellular metabolism, and calcium-dependent apoptosis. Deletion or loss of function mutation of the BAP1 gene prevents cells undergoing apoptosis after DNA damage, allowing proliferation of cells with deleterious mutations.4BAP1 thus functions as a tumor suppressor gene. BAP1 somatic mutations/deletions are found in ∼60% of malignant mesotheliomas,5 and germline mutations in a small fraction,6 probably <1% of mesotheliomas.

Stain Interpretation

In normal cells BAP1 protein will be detected in cell nuclei by IHC, whereas in cells of malignant mesotheliomas that harbor BAP1 mutations/deletions either no protein expression is seen at all, or there is cytoplasmic accumulation of protein with no nuclear staining, the latter event reflecting loss of the nuclear localization sequence. Loss of nuclear staining is thus the diagnostic IHC result (Fig. 1). However, cytoplasmic staining is not reliable because it can be influenced by staining conditions, particularly high primary antibody concentrations. In our experience BAP1 protein is somewhat fixation sensitive, so sections stained for BAP1 must show positive staining of stromal and inflammatory cells which serve as an internal control (Fig. 1).

Core needle biopsy suspicious for mesothelioma on hematoxylin and eosin. Only a few groups of very bland mesothelial cells (inset) are present, but BAP1 stain shows loss of BAP1 in the mesothelial cells, confirming that this is malignant. Note positive stromal and inflammatory cells that serve as an internal control.

Most authors report 100% loss of nuclear staining in mesotheliomas with BAP1 mutations, but some studies find partial loss of staining to be common and employ cutoff values; for example, Hida et al7 use >19% as an indicator of malignancy. Occasional tumors with what appear to be BAP1 lost and BAP1 retained clones have been described.8 We have found that partial loss of staining in histologic sections is uncommon, and extensive loss with a small fraction of cells that have retained BAP1 should be regarded as lost.

Sensitivity, Specificity, and Differences by Histologic Pattern/Tumor Location

Loss of BAP1 staining in histologic sections of mesotheliomas as a diagnostic tool was first described by Sheffield et al9 and Cigognetti et al.8 Subsequently a large number of articles have been published with remarkably consistent results (reviewed in Churg and colleagues2,5,7) and the conclusion that loss of BAP1 nuclear staining is a reliable marker of malignancy.

Wang et al5 performed a meta-analysis of 12 published studies with 1824 tissue and cytology specimens. In tissue sections they determined that the specificity of nuclear loss for malignancy was 100%. Reports tracing outcomes in cases called atypical by morphology have come to the same conclusion8,10 (see the Malignant mesothelioma in situ section below). The Wang et al5 analysis showed that the sensitivity of BAP1 staining of tissue specimens for epithelial mesotheliomas was 74%, for mixed epithelial and sarcomatous mesotheliomas 50%, and for sarcomatous mesotheliomas 7%. Some individual studies report sensitivities of up to 37% in sarcomatous tumors,11 and some report sensitivities of 0.12 In our experience loss of staining in sarcomatous mesothelioma can occasionally be seen but is infrequent.13 When mixed epithelial and sarcomatous mesotheliomas show BAP1 loss it is most commonly in the epithelial compartment with retention in the sarcomatous component14 (Fig. 2).

A mixed epithelial and sarcomatous mesothelioma demonstrating loss of BAP1 in the epithelial component but not in the sarcomatous component.

Most studies on BAP1 have looked at pleural mesotheliomas. To our knowledge there are no studies directly comparing BAP1 staining in pleural versus peritoneal mesotheliomas, but Singhi et al15 found BAP1 loss in 49 of 86 (57%) and Tandon et al16 loss in 57 of 137 (42%) peritoneal mesotheliomas. In our experience BAP1 loss is common in peritoneal tumors, and is the defining feature of both pleural and peritoneal malignant mesothelioma in situ (see the Malignant mesothelioma in situ section).

Application to Effusion Cytology Specimens

BAP1 staining works well in effusion cytology samples (Fig. 3), but alcohol fixatives such as CytoLyt may leach antigen and formalin fixation is more reliable. In their meta-analysis, Wang et al5 concluded that the sensitivity for BAP1 staining in effusion specimens from 5 different studies was 58% and the specificity of nuclear loss for malignancy 100%. These were essentially all epithelial processes, since sarcomatous mesotheliomas generally do not shed cells into effusions.

Pleural effusion cytology specimen from a malignant mesothelioma. In this case both BAP1 and MTAP are lost. Note the positive internal control cells in both stains.

Effusion samples are often more difficult to interpret than tissue sections because of mixtures of malignant and reactive mesothelial cells, and in fact the most important use of BAP1 staining in effusion specimens is making a definitive diagnosis of mesothelioma when atypical but not overtly malignant mesothelioma cells are found. Andrici et al17 reported loss of BAP1 in 8 of 57 effusion samples with atypical mesothelial cells but no overt tumor; malignant mesothelioma was subsequently diagnosed in all 6 cases that had follow-up. Similarly, Cigognetti et al8 found that all 6 of 6 patients with atypical mesothelial cells and BAP1 loss eventually developed mesotheliomas.

BAP1 Cancer Predisposition Syndrome and Other Tumors That Can Show Loss of BAP1

Germline mutations of BAP1 were initially described in families with mesotheliomas, uveal melanomas, cutaneous melanomas, and clear cell renal cell carcinomas (6), and some forms of benign cutaneous nevi. This scenario is often labeled the BAP1 cancer predisposition syndrome (reviewed by Walpole et al18). More recent data suggest that such families may also develop meningiomas, basal cell carcinomas, breast carcinomas, neuroendocrine carcinomas, and cholangiocarcinomas.18,19 Some of the tumor types associated with the BAP1 cancer predisposition syndrome show BAP1 loss at a considerable frequency (eg, 84% of uveal melanomas and 10% of meningiomas with rhabdoid features),19 such that one must be cognizant of their presence in the differential diagnosis of a lesion with BAP1 loss.

By IHC these tumors and the benign nevi all show loss of nuclear BAP1. However, IHC will not sort out germline from somatic mutations. Both show up as lack of nuclear staining because the germline mutation only affects one allele, and a second mutation/deletion event affecting the other allele is required before nuclear staining is lost.

BAP1 somatic loss is a rare event in other tumors. This phenomenon is sometimes diagnostically useful for separating morphologically/IHC equivocal mesotheliomas from non–small cell lung cancers and high grade serous carcinomas, because both of these types of tumor lose BAP1 in <1% of cases.8,20–22 However, rare cases of somatic and germline BAP1 mutations in lung adenocarcinoma have been reported.23,24

Use of BAP1 Plus Another Marker of Mesothelial Malignancy

There are a number of articles discussing combinations of BAP1 and MTAP IHC, or BAP1 immunostains plus cyclin-dependent kinase inhibitor 2A (CDKN2A) FISH. These combinations often improve sensitivity (see the Use of combined markers to improve sensitivity section).

CDKN2A (p16) FISH and Methylthioadenosine Phosphorylase IHC Staining

CDKN2A and methylthioadenosine phosphorylase (MTAP) are genes located within 200 KB of each other on chromosome 9p21.3. CDKN2A is a tumor suppressor gene that codes for 2 proteins, p16INK4A (frequently just referred to as “p16”) and p14ARF. The former inhibits transition from G1 to S phase of the cell cycle by inhibiting cyclin-dependent kinases 4 and 6 and thus activating retinoblastoma protein; while the latter activates p53. CDKN2A is deleted or mutated in many types of malignancy,3 not only many mesotheliomas, hence blind application of CDKN2A FISH can be very misleading if the process is not first established as mesothelial.

MTAP protein is an enzyme involved in purine metabolism that plays a role in the recycling of adenosine and methionine. Like CDKN2A, MTAP is mutated or deleted in many different types of malignancies.25MTAP may function as a tumor suppressor gene in its own right,26,27 but is of diagnostic interest principally because it is frequently co-deleted with CDKN2, and loss of MTAP can be detected by IHC, thus providing a much simpler diagnostic test. Krasinkas et al,28 using FISH probes for each gene, showed exactly matching loss or retention of CDKN2A and MTAP in a series of peritoneal mesotheliomas. The immunochemical correspondence is not quite as good: Hida et al7 found that MTAP immunochemistry and CDKN2A FISH were correlated in 74% of cases, and this idea was confirmed by Berg et al29 who found the 2 markers concordant in 82% of cases, and by Chapel et al30 who reported that MTAP loss is 78% sensitive and 96% specific for homozygous deletion of CDKN2A as determined by FISH.

Interpretation and Technical Issues

CDKN2A FISH is used to detect deletion of the gene (Fig. 4). Two probes are used: a centromere 9 probe that serves as an internal/copy number control, and a CDKN2A probe. A normal cell should show 2 centromere 9 and 2 CDKN2A signals. Most laboratories only evaluate homozygous deletion; that is, loss of both CDKN2A signals. Because the size of a mesothelial cell nucleus is larger than the thickness of a 5 µm tissue section, there is always a background rate of spurious “loss” due to truncation artifact. For this reason each laboratory must determine the background truncation rate of normal mesothelial cells before running the test. The corollary is that an adequate number of mesothelial cells, typically 50 to 100, must be counted in a test sample to be sure that the results are above the truncation background. Although it is easy to count this number of cells when there is a solid nodule of potential tumor, it can be much harder to pick out the correct cells (under oil with DAPI fluorescent staining of the nuclei) if they are scattered, particularly when mixed with inflammatory or stromal cells. Thus CDKN2A FISH requires careful area selection by the pathologist and experience in using FISH.

A pleural biopsy from a patient with nodular pleural thickening on imaging, thought suspicious for mesothelioma. The biopsy contained only one small focus of bland mesothelial cells (circled), not sufficient by themselves for a diagnosis of malignancy. However, FISH (insert) showed >90% loss of CDKN2A signals, thus confirming a diagnosis of malignancy.

The truncation upper limit reported in the literature varies from laboratory to laboratory, but generally is in the range of 10% to 20%. Our own view is that test results a little above the truncation background should be interpreted cautiously and with careful reference to the truncation background in the laboratories performing the test, but homozygous deletion rates of 30% or greater are always diagnostic of malignancy (Fig. 4).

In benign cells MTAP IHC always stains the cytoplasm and often stains the nucleus (Figs. 3, 5), but examination of mesotheliomas shows that nuclear loss is inconsistent and only cytoplasmic loss is diagnostic of malignancy.31 MTAP does not appear to be particularly fixation sensitive, but as is true of BAP1 IHC, sections stained for MTAP must show a positive internal control of inflammatory and/or stromal cells to be interpretable (Figs. 3, 5).

A malignant mesothelioma in which BAP1 (A) is retained but MTAP (B) is lost. Note the positive internal control cells in both stains. For BAP1 loss of nuclear staining is diagnostic, but for MTAP only loss of cytoplasmic staining is diagnostic of malignancy and nuclear staining/loss is not informative.

In our experience, most mesotheliomas that demonstrate MTAP loss show cytoplasmic loss in every tumor cell, but a few show loss in only a fraction of tumor cells (see the Cytology section). Some authors use a numeric cutoff; for example, Hida et al7 proposed that loss of staining in >32% of cells was indicative of malignancy, but such numbers are lab-specific and at this point there are no clear rules for interpreting partial loss.

There are also sometimes discrepancies between MTAP IHC and CDKN2A FISH, such that MTAP may be retained while CDKN2A is lost, or vice versa. These discrepancies presumably reflect deletion/mutation/epigenetic events (hypermethylation) that specifically target one gene but not the other. For example, Savic et al32 found CDKN2A promoter hypermethylation in 4 of 5 mesotheliomas in which FISH did not show homozygous deletion.

Findings by Tumor Morphology and Tumor Location

There are no meta-analyses available for either CDKN2A FISH or MTAP IHC. We2 previously reviewed the literature for CDKN2A FISH and concluded that the sensitivity for epithelial mesotheliomas in the pleura was probably on the order of 60% to 70% but with a large spread of results from different papers. For peritoneal epithelial mesotheliomas the numbers were lower, typically <50%, and this has been confirmed in other studies.15,29 For pleural sarcomatous mesotheliomas reported numbers are often in the 80% to 100% range.11,13 Sarcomatous mesotheliomas of the peritoneum are rare and no data are available. There are no reports of homozygous deletions above the truncation background in reactive mesothelial processes. Thus far there are also no reports of MTAP loss in benign reactive mesothelial processes,7,29–31 although the number of cases examined is relatively small.

Table 2 shows published data comparing MTAP IHC to CDKN2A FISH sensitivity for detecting malignant mesotheliomas. In general CDKN2A FISH appears to be slightly more sensitive than MTAP IHC. Thus there is something to be gained from running both tests if both are available, but the improvement in diagnostic sensitivity from adding CDKN2A FISH to MTAP IHC is small.

TABLE 2 - Comparison of CDKN2A FISH and MTAP IHC on the Same Cases
References Morphology MTAP IHC*, n/N (%) CDKN2A FISH†, n/N (%)
Hida et al7 Epithelial/mixed 24/51 (47) 31/51 (61)
Berg et al29 Epithelial 13/20 (65) 12/17 (71)
Kinoshita et al11 Sarcomatous 25/29 (86) 28/30 (93)
Chapel et al30 Not specified/includes cytology specimens 26/56 (46) 32/56 (57)
*Loss of cytoplasmic staining.
†Homozygous deletion.

Application to Effusion Cytology Specimens

CDKN2A FISH works well in effusion specimens31,33,34 and in fact the first description of this test was on cytology preparations.35 Both cell blocks and smears can be used, and smears offer the advantage that there are no truncation artifacts. Hida et al36 showed a good correlation between CDKN2A loss by FISH in effusion smears and tissue biopsies, and found that only 1.3% of benign mesothelial cells showed apparent homozygous deletion in smears.

MTAP IHC also works well in cytology samples (Fig. 3). It is our impression that alcohol-based fixatives leach MTAP but this has not been formally studied. Kinoshita et al34 found loss of MTAP cytoplasmic staining in 37 of 73 (51%) of cases and homozygous deletion of CDKN2A in 44 of 73 (60%). Berg et al37 found complete MTAP loss in effusion specimens of 7 of 21 confirmed mesothelioma cases, complete retention of staining in 11 of 21 cases, and partial loss in 3 of 21 cases. The effusion cases with complete loss or no loss correlated exactly with the pattern in the corresponding biopsy specimens, but the partial loss cases did not. Thus, as is true for tissue sections, cases with partial loss should be reported cautiously.

Role of p16 IHC

p16 IHC does not provide the same results as examination of CDKN2A homozygous deletion by FISH.7,29 For example, p16 IHC may show no staining in benign mesothelial cells when FISH does not show any loss. p16 IHC should be not be substituted for FISH testing in either tissue sections or cytology preparations.


Detection of Mutations/Deletions in the Hippo Pathway

The Hippo pathway is a signaling kinase cascade normally involved in control of apoptosis, stress responses, and cell proliferation, particularly determination of organ size. The general effect of Hippo signaling is to downregulate these processes by degrading the signaling molecules YAP/TAZ in the cytoplasm38; if the Hippo pathway becomes nonfunctional because of mutations/deletions of its various components, there is uncontrolled translocation of YAP/TAZ to the nucleus with subsequent activation of a variety of genes affecting cell proliferation.38,39

Hippo pathway genomic abnormalities are present in numerous types of malignancy (reviewed by Han39). In an analysis of genomic abnormalities in 74 pleural mesothelioma, Hippo somatic mutations/deletions were found in 51% of cases, and in 20% >1 Hippo component was affected, most commonly NF2, LATS2, and MOB3B.37,40 However, attempts to detect/exploit abnormalities in the Hippo pathway for separating benign from malignant mesothelial proliferations have been somewhat disappointing.


Singhi et al15 reported that heterozygous deletions of NF2 could be detected by FISH in 30 of 86 (35%) of a series of peritoneal mesotheliomas; no homozygous deletions were identified. However, the same laboratory did not find any instances of either homozygous or heterozygous NF2 deletion in a series of 20 pleural mesotheliomas that did harbor a variety of BAP1, MTAP, and CDKN2A deletions/mutations.29 But Kinoshita et al41 reported monosomy 22 or hemizygous NF2 deletion in 25 of 47 (53%) pleural mesotheliomas and 0 of 27 reactive mesothelial proliferations. The reasons for these discrepancies are unclear, although only 2 of the Kinoshita cases actually showed hemizygous deletion while 23 showed monosomy. At this point more data are needed before recommending NF2 FISH for general use.

Hippo Pathway IHC

The protein product of NF2 is named Merlin. Sheffield et al42 used IHC to look for loss of Merlin, loss of LATS2, and nuclear translocation of YAP/TAZ in a tissue microarray (TMA) series of 25 mesotheliomas and 46 reactive proliferations. Loss of Merlin was detected in only 1 mesothelioma, whereas loss of LATS2 and nuclear YAP/TAZ were found in both mesotheliomas and reactive processes, with no statistical differences.

Cyclin D1 IHC

One of the downstream effects of YAP/TAZ signaling is increased transcription of cyclin D1. Pors et al43 reported cyclin D1 staining in a TMA series of 103 reactive mesothelial proliferations, 54 epithelial mesotheliomas, and 22 sarcomatous mesotheliomas. Positive staining of endothelial cell nuclei served as a positive internal control.

Most reactive epithelial mesothelial proliferations showed no staining or only a few percent of cells staining, whereas reactive spindle cell proliferations tended to show more staining. The authors concluded that for epithelial proliferations, >50% of cells staining was only present in epithelial mesotheliomas, while for spindle cell proliferations, >75% of cells staining was only seen in sarcomatous mesotheliomas (Fig. 6). Using these cutoffs, the sensitivity of cyclin D1 staining for epithelial mesotheliomas was 57% and for sarcomatous mesotheliomas 45%.

A desmoplastic mesothelioma showing infiltration of chest wall fat by spindle cells. MTAP stain shows loss of MTAP in the spindle cells with positive internal control of inflammatory cells, while cyclin D1 stain shows strong diffuse positivity in almost all the tumor cells.

Conclusions Re Hippo Pathway

At this point the question of whether NF2 FISH is useful for separating benign from malignant processes is unresolved, and IHC to examine other Hippo components is disappointing. Cyclin D1 is not strictly part of the Hippo pathway, but may be useable, although the cutoff values need to be validated in other laboratories. This whole area needs further examination.

Enhancer of Zeste Homolog 2 IHC Staining

Enhancer of zeste homolog 2 (EZH2) is part of the polycomb repressive complex 2, which mediates gene silencing through regulation of chromatin structure. EZH2 is involved in cell proliferation, cell cycle progression, cell differentiation, and apoptosis, and plays a crucial role in DNA damage repair. EZH2 can also activate oncogenic pathways independent of polycomb repressive complex 2.44 EZH2 mutations or aberrant function, typically overexpression, are found in a variety of malignancies.

Shinozaki-Ushiku et al45 examined EZH2 expression in 32 mesotheliomas (mostly epithelial or mixed epithelial/sarcomatous) and 18 reactive mesothelial proliferations, dividing expression into “low” (1 to 49) and “high” (>49) percent of cells showing nuclear staining. High EZH2 scores were found in 21 mesotheliomas and low scores in 11. In the reactive mesothelial proliferations low scores were found in all 18 and no high scores were observed. The authors calculated a sensitivity of 66% and specificity of 100% for high staining. However, many reactive proliferations scored in the 10% to 49% range and many mesotheliomas in the 50% to 90% range, so the reliability of a 50% separation is uncertain. No correlation was found between EZH2 overexpression and BAP1 IHC.

Yoshimura et al46 performed a similar study on 38 mesotheliomas (mostly epithelial) and 29 reactive proliferations and also counted proportion of cells staining. For reactive proliferations, staining proportions ranged from 0% to around 40%, with most cases <20%. In the mesotheliomas staining ranged from 0% to 94%, with roughly half the tumors above 50% (Fig. 7). Their calculated sensitivity was 45% and specificity 100% at a 50% staining cutoff; however, some of their reactive proliferations showed 40% staining, suggesting that a 50% cutoff may be too low. No details were provided on staining of histologic subtypes. No correlations were found between EZH2 overexpression and other markers (BAP1 IHC, MTAP IHC, and CDKN2A FISH).

A mixed epithelial and sarcomatous malignant mesothelioma showing strong diffuse nuclear positivity for EZH2 (courtesy Dr Kazuki Nabeshima, Fukuoka University). This intensity and diffuseness of strong staining supports a diagnosis of malignancy.

EZH2 staining might thus be usable for separating benign from malignant proliferations, but on the existing data would require the finding of staining in a very high proportion of cells (probably>75%) to reliably call a lesion malignant. This approach needs to be confirmed with further studies before EZH2 IHC can be recommended for general use.

5-Hydroxymethyl Cytosine IHC Staining

Epigenetic silencing of genes through methylation of cytosine residues and formation of 5-methyl cytosine (5-MC) was initially thought to be an irreversible event, but was later discovered to be a reversible process. Demethylation proceeds through the action of TET methylcytosine dioxygenases which oxidize 5-MC to 5-hydroxymethyl cytosine (5-HMC). 5-HMC is further oxidized, and the resulting base excised and replaced with intact cytosine.47 Upstream mutations in IDH1, IDH2, or the TET enzymes may lead to reduced levels of 5-HMC, and reduced levels have been found in a variety of hematologic and solid cancers.47

Chapel et al48 used an antibody against 5-HMC in a series of 49 mesotheliomas and 14 reactive mesothelial proliferations and counted the number of cells with loss of nuclear 5-HMC. Loss rates were greatest in sarcomatous mesotheliomas (mostly>80% of cells), intermediate in mixed mesotheliomas, and more widely distributed in epithelial mesotheliomas, whereas the reactive proliferations did not exhibit >20% loss. Chapel and colleagues proposed that a >50% loss rate was indicative of malignancy. Using this cutoff the overall sensitivity of 5-HMC staining for mesotheliomas was 92% and sensitivity 100%.

One of the problems these authors encountered was that the distinction of spindled mesothelial cells from spindled fibroblasts was difficult, and for that reason they recommended use of a double 5-HMC/CAM5.2 stain. 5-HMC staining appears to be a potentially useful approach to separating benign from malignant mesothelial processes, but these data need to be replicated and the requirement for double staining may limit the technique.

Use of Combined Markers to Improve Sensitivity

None of the individual markers discussed above is 100% sensitive for a diagnosis of mesothelioma versus a reactive mesothelial proliferation and for this reason combinations of markers have frequently been employed. For the combination of BAP1 IHC and either MTAP IHC or CDKN2A FISH, sensitivities of ~80% and often greater for detecting mesotheliomas are consistently reported in both tissue and cytology specimens (Table 3).

TABLE 3 - Sensitivity of Combined Tests Using BAP1 IHC Plus Another Marker for Separating Benign From Malignant Proliferations
Marker(s) References Specimen Type Mesothelioma Morphology Sensitivity of Marker Alone, n/N (%) Sensitivity of BAP1, n/N (%) Sensitivity of Marker Combined With BAP1, n/N (%)
CDKN2A (p16) FISH Hwang et al33 Cytology Epithelial 8/11 (73) 10/15 (67) 11/11 (100)
Kinoshita et al34 Cytology Epithelial 28/45 (62) 27/45 (60) 38/45 (84)
Hwang et al33 Surgical Epithelial 12/15 (80) 10/15 (67) 15/15 (100)
Berg et al29 Surgical Epithelial 12/17 (71) 12/20 (60) 15/17 (88)
Hwang et al13 Surgical Sarcomatous 16/20 (80) 3/20 (15) 17/20 (85)
Kinoshia et al11 Surgical Sarcomatous 28/30 (93) 11/28 (39) 30/30 (100)
Singhi et al15 Surgical 75 epithelial, 11 nonepithelial (all peritoneal) 25/86 (29) 49/86 (57) 56/86 (65)
Yoshimura et al46 Surgical 27 epithelial, 6 biphasic, and 5 sarcomatous 25/38 (66) 18/38 (47) 31/38 (82)
Hida et al7 Surgical 44 epithelial, 7 biphasic 31/51 (61) 31/51 (61) 43/51 (84)
MTAP IHC Kinoshita et al34 Cytology Epithelial 27/45 (60) 19/45 (42) 35/45 (78)
Berg et al37 Cytology Epithelial 7/21 (33) 12/19 (63) 15/21 (71)
Berg et al37 Surgical Epithelial 6/14 (43) 10/13 (77) 12/14 (86)
Berg et al29 Surgical Epithelial 13/20 (65) 12/20 (60) 19/20 (95)
Hida et al7 Surgical 44 epithelial, 7 biphasic 23/51 (45) 31/51 (61) 39/51 (76)
Kinoshita et al11 Surgical Sarcomatous 25/29 (86) 11/28 (39) 27/30 (90)
NF2 FISH Singhi et al15 Surgical 75 epithelial, 11 nonepithelial 30/86 (35) 49/86 (57) 61/86 (71)
Kinoshita et al41 Surgical Not specified 25/47 (53) 27/47 (57) 37/47 (79)
EZH2 IHC Yoshimura et al46 Surgical 27 epithelial, 6 biphasic, and 5 sarcomatous 17/38 (45) 18/38 (47) 28/38 (74)
Shinozaki-Ushiku et al45 Surgical/cytology 23 epithelial, 7 biphasic, and 2 sarcomatous 21/32 (66) 17/32 (53) 28/32 (87.5)
5-HMC/CAM5.2 IHC double stain Chapel et al48 Surgical 17 epithelial, 22 biphasic, 10 sarcomatous 45/49 (92) 25/49 (51) 48/49 (98)

Some of the other combinations in Table 3 look extremely good, but these are often based on single papers. Until there are more confirmatory studies and more data on test interpretation (in particular, reproducible cutoffs), these results should be interpreted as guidelines to areas needed further investigation.

Immune Checkpoint Inhibitor IHC Staining

The idea that blockade of immune checkpoint inhibitors may allow inflammatory cells to attack tumors is well established. What is generally not been appreciated is that, apart from inflammatory cells, most benign tissues do not express immune checkpoint inhibitor proteins or express them at a very low level, although there are exceptions to this rule.

Derakhshan et al49 examined the use of programmed death-ligand 1 (PD-L1) staining to separate mesotheliomas from reactive processes using antibody 22C3 on TMAs containing 62 mesotheliomas and 88 reactive proliferations. They found that the vast majority of epithelial or spindle reactive mesothelial proliferations showed no staining and a small number showed weak staining. Most epithelial mesotheliomas demonstrated a similar pattern. However, the majority of sarcomatous mesotheliomas tested showed strong and diffuse staining (Fig. 8) and this was confirmed using whole sections. Similarly, Chung et al50 reported that benign mesothelial reactions were always negative for PD-L1 with antibody 22C3, whereas >50% of cells staining was seen in 8% of epithelial mesotheliomas and 46% of sarcomatous tumors. These results are in a sense not surprising because most PD-L1 positive mesotheliomas reported in the literature are sarcomatous.

Strong diffuse staining for PD-L1 (clone 22C3) in a sarcomatous mesothelioma (tissue microarray core). This intensity and diffuseness of strong staining supports a diagnosis of malignancy.

The potential advantage of using PD-L1 staining is that many laboratories already run this test. The downside is that one needs to use some type of cutoff to define a positive result and there are no clear rules as yet on what that cutoff should be. For the time being we recommend that strong diffuse staining such as shown in Figure 8 be viewed as positive, but weak or very focal staining as nondiagnostic; confirmatory studies would be useful. Presumably the same results will be obtained with PD-L1 antibodies other than 22C3, but this has not been reported.

CD47, also known as Integrin Associated Protein-1 is a somewhat what different type of immune checkpoint inhibitor that functions primarily as a “don’t eat me” signal to macrophages.51 CD47 is expressed at low levels in most normal tissues, levels that may or may not be detectable by IHC. Recently Schurch et al52 suggested that immunostaining for CD47 in effusion cytology specimens could detect malignant mesotheliomas because benign mesothelial cells either did not express detectable CD47 or showed very weak expression. They reported a sensitivity of 63% and specificity of 100%

We investigated the use of CD47 in a small TMA and found that reactive epithelial mesothelioma proliferations generally did not show detectable expression, but reactive spindle cell proliferations did, so that separation from sarcomatous mesotheliomas was not feasible. However, roughly half of the epithelial mesotheliomas tested showed moderate to strong staining.

Both the PD-L1 and CD47 results warrant further investigation; the potential problem with both is creating accurate cutoffs between benign and malignant.

Malignant Mesothelioma In Situ

Malignant mesothelioma in situ (MIS) has been talked about in the pathology literature for many years, but without any agreed upon rules for diagnosis (and considerable skepticism that the diagnosis could be made at all), because reactive surface mesothelial cells are often quite atypical and malignant mesothelial cells often innocuous-appearing.

We53 recently described 10 cases of MIS with diagnosis based on the finding of a single layer of generally quite bland slightly reactive-appearing mesothelial cells that showed loss of BAP1 (Fig. 9), no evidence of tumor on imaging or direct inspection of the serosal surfaces, and no evidence of invasive mesothelioma for at least 1 year. Seven of these cases developed invasive mesothelioma with a median time from biopsy showing MIS to invasive tumor of 60 months.

Malignant mesothelioma in situ. The hematoxylin and eosin stain shows a single layer of bland cuboidal mesothelial cells on the pleural surface. By itself this appearance is not diagnostic of malignancy and could well be benign, but there is diffuse loss of BAP1 in these cells, indicating that the correct diagnosis is mesothelioma in situ.

Almost all of these patients had recurrent pleural or peritoneal effusions of unknown etiology, and this lead us to suggest that, given a clinical scenario of repeated effusions with no obvious etiology and no evidence of invasive tumor by imaging or visual inspection, staining of what appears to be an innocuous surface layer of mesothelial cells in a pleural/peritoneal biopsy for BAP1 is crucial, since mesothelioma in situ is potentially curable.

Given the extensive data on BAP1 discussed above, loss of BAP1 on IHC is always an indicator of malignancy. However, only ~70% of epithelial mesotheliomas show mutation/deletion of BAP1 detectable as loss of nuclear staining,5 so that MIS might instead take the form of a single layer of mesothelial cells that have lost MTAP (or CDKN2A by FISH); thus these tests should be carried out if BAP1 is retained. There is in fact one case report54 of MIS in which MTAP showed loss of staining and CDKN2A FISH demonstrated homozygous deletion, but BAP1 was retained. Presumably more such cases will surface over time.


At this point there are 3 well established adjunctive markers for separating benign from malignant mesothelial proliferations, BAP1 IHC, MTAP IHC, and CDKN2A FISH, and we suggest that these should be the first approach to the problem, with consideration given to the morphology (epithelial vs. spindle) and location (pleura vs. peritoneum). Table 4 shows our recommended tests and order of testing.

TABLE 4 - Recommended IHC Staining or FISH Approach by Morphology and Location
Pleura-epithelial: BAP1 IHC then MTAP IHC, or BAP1 IHC then CDKN2A FISH
Pleural-spindle cell: MTAP IHC or CDKN2A FISH, then BAP1 IHC (low yield test)
Peritoneal-epithelial: BAP1 IHC then MTAP IHC (low yield test in the peritoneum), or BAP1 IHC then CDKN2A FISH (low yield test in the peritoneum)
Peritoneal-spindle cell: no information (sarcomatous mesotheliomas extremely uncommon in the peritoneum)

BAP1 and MTAP IHC in general offer a yes or no (complete loss of staining or complete retention of staining) dichotomy, although occasional cases show partial loss and in that setting the test may be equivocal. The other markers described in this review either have inconsistent published results (NF2 FISH) or simply too few publications to recommend them. Most of these also suffer from the problem that, as thus far reported, staining or loss of staining needs to exceed a threshold, and at this point such thresholds are really laboratory-specific, thus limiting generalizability of the results.


1. Churg A, Galateau-Salle F. The separation of benign and malignant mesothelial proliferations: update 2012. Arch Pathol Lab Med. 2012;136:1217–1226.
2. Churg A, Sheffield BS, Galateau-Salle F. New markers for separating benign from malignant mesothelial proliferations: are we there yet? Arch Pathol Lab Med. 2016;140:318–321.
3. ICGC/TCGA Pan-Cancer Analysis of Whole Genomes Consortium. Pan-cancer analysis of whole genomes. Nature. 2020;578:82–93.
4. Bononi A, Giorgi C, Patergnani S, et al. BAP1 regulates IP3R3-mediated Ca(2+) flux to mitochondria suppressing cell transformation. Nature. 2017;546:549–553.
5. Wang LM, Shi ZW, Wang JL, et al. Diagnostic accuracy of BRCA1-associated protein 1 in malignant mesothelioma: a meta-analysis. Oncotarget. 2017;8:68863–68872.
6. Testa JR, Cheung M, Pei J, et al. Germline BAP1 mutations predispose to malignant mesothelioma. Nat Genet. 2011;43:1022–1025.
7. Hida T, Hamasaki M, Matsumoto S, et al. Immunohistochemical detection of MTAP and BAP1 protein loss for mesothelioma diagnosis: comparison with 9p21 FISH and BAP1 immunohistochemistry. Lung Cancer. 2017;104:98–105.
8. Cigognetti M, Lonardi S, Fisogni S, et al. BAP1 (BRCA1-associated protein 1) is a highly specific marker for differentiating mesothelioma from reactive mesothelial proliferations. Mod Pathol. 2015;28:1043–1057.
9. Sheffield BS, Hwang HC, Lee AF, et al. BAP1 immunohistochemistry and p16 FISH to separate benign from malignant mesothelial proliferations. Am J Surg Pathol. 2015;39:977–982.
10. Pillappa R, Maleszewski JJ, Sukov WR, et al. Loss of BAP1 expression in atypical mesothelial proliferations helps to predict malignant mesothelioma. Am J Surg Pathol. 2018;42:256–263.
11. Kinoshita Y, Hamasaki M, Yoshimura M, et al. A combination of MTAP and BAP1 immunohistochemistry is effective for distinguishing sarcomatoid mesothelioma from fibrous pleuritis. Lung Cancer. 2018;125:198–204.
12. McGregor SM, Dunning R, Hyjek E, et al. BAP1 facilitates diagnostic objectivity, classification, and prognostication in malignant pleural mesothelioma. Hum Pathol. 2015;46:1670–1678.
13. Hwang HC, Pyott S, Rodriguez S, et al. BAP1 immunohistochemistry and p16 FISH in the diagnosis of sarcomatous and desmoplastic mesotheliomas. Am J Surg Pathol. 2016;40:714–718.
14. Nicholson AG, Sauter JL, Nowak AK, et al. EURACAN/IASLC proposals for updating the histologic classification of pleural mesothelioma: towards a more multidisciplinary approach. J Thorac Oncol. 2020;15:29–49.
15. Singhi AD, Krasinskas AM, Choudry HA, et al. The prognostic significance of BAP1, NF2, and CDKN2A in malignant peritoneal mesothelioma. Mod Pathol. 2016;29:14–24.
16. Tandon RT, Jimenez-Cortez Y, Taub R, et al. Immunohistochemistry in peritoneal mesothelioma: a single-center experience of 244 cases. Arch Pathol Lab Med. 2018;142:236–242.
17. Andrici J, Sheen A, Sioson L, et al. Loss of expression of BAP1 is a useful adjunct, which strongly supports the diagnosis of mesothelioma in effusion cytology. Mod Pathol. 2015;28:1360–1368.
18. Walpole S, Pritchard AL, Cebulla CM, et al. Comprehensive study of the clinical phenotype of germline BAP1 variant-carrying families worldwide. J Natl Cancer Inst. 2018;110:1328–1341.
19. Kobrinski DA, Yang H, Kittaneh M. BAP1: role in carcinogenesis and clinical implications. Transl Lung Cancer Res. 2020;9(suppl 1):S60–S66.
20. Andrici J, Jung J, Sheen A, et al. Loss of BAP1 expression is very rare in peritoneal and gynecologic serous adenocarcinomas and can be useful in the differential diagnosis with abdominal mesothelioma. Hum Pathol. 2016;51:9–15.
21. Owen D, Sheffield BS, Ionescu D, et al. Loss of BRCA1-associated protein 1 (BAP1) expression is rare in non-small cell lung cancer. Hum Pathol. 2017;60:82–85.
22. Carbone M, Shimizu D, Napolitano A, et al. Positive nuclear BAP1 immunostaining helps differentiate non-small cell lung carcinomas from malignant mesothelioma. Oncotarget. 2016;7:59314–59321.
23. Abdel-Rahman MH, Pilarski R, Cebulla CM, et al. Germline BAP1 mutation predisposes to uveal melanoma, lung adenocarcinoma, meningioma, and other cancers. J Med Genet. 2011;48:856–859.
24. Jensen DE, Rauscher FJ III. BAP1, a candidate tumor suppressor protein that interacts with BRCA1. Ann N Y Acad Sci. 1999;886:191–194.
25. Shao X, Lv N, Liao J, et al. Copy number variation is highly correlated with differential gene expression: a pan-cancer study. BMC Med Genet. 2019;20:175.
26. Huang HY, Li SH, Yu SC, et al. Homozygous deletion of MTAP gene as a poor prognosticator in gastrointestinal stromal tumors. Clin Cancer Res. 2009;15:6963–6972.
27. Illei PB, Rusch VW, Zakowski MF, et al. Homozygous deletion of CDKN2A and codeletion of the methylthioadenosine phosphorylase gene in the majority of pleural mesotheliomas. Clin Cancer Res. 2003;9:2108–2113.
28. Krasinskas AM, Bartlett DL, Cieply K, et al. CDKN2A and MTAP deletions in peritoneal mesotheliomas are correlated with loss of p16 protein expression and poor survival. Mod Pathol. 2010;23:531–538.
29. Berg KB, Dacic S, Miller C, et al. Utility of methylthioadenosine phosphorylase compared with BAP1 immunohistochemistry, and CDKN2A and NF2 fluorescence in situ hybridization in separating reactive mesothelial proliferations from epithelioid malignant mesotheliomas. Arch Pathol Lab Med. 2018;142:1549–1553.
30. Chapel DB, Schulte JJ, Berg K, et al. MTAP immunohistochemistry is an accurate and reproducible surrogate for CDKN2A fluorescence in situ hybridization in diagnosis of malignant pleural mesothelioma. Mod Pathol. 2020;33:245–254.
31. Hamasaki M, Kinoshita Y, Yoshimura M, et al. Cytoplasmic MTAP expression loss detected by immunohistochemistry correlates with 9p21 homozygous deletion detected by FISH in pleural effusion cytology of mesothelioma. Histopathology. 2019;75:153–155.
32. Savic S, Franco N, Grilli B, et al. Fluorescence in situ hybridization in the definitive diagnosis of malignant mesothelioma in effusion cytology. Chest. 2010;138:137–144.
33. Hwang HC, Sheffield BS, Rodriguez S, et al. Utility of BAP1 immunohistochemistry and p16 (CDKN2A) FISH in the diagnosis of malignant mesothelioma in effusion cytology specimens. Am J Surg Pathol. 2016;40:120–126.
34. Kinoshita Y, Hida T, Hamasaki M, et al. A combination of MTAP and BAP1 immunohistochemistry in pleural effusion cytology for the diagnosis of mesothelioma. Cancer Cytopathol. 2018;126:54–63.
35. Illei PB, Ladanyi M, Rusch VW, et al. The use of CDKN2A deletion as a diagnostic marker for malignant mesothelioma in body cavity effusions. Cancer. 2003;99:51–56.
36. Hida T, Matsumoto S, Hamasaki M, et al. Deletion status of p16 in effusion smear preparation correlates with that of underlying malignant pleural mesothelioma tissue. Cancer Sci. 2015;106:1635–1641.
37. Berg KB, Churg AM, Cheung S, et al. Usefulness of methylthioadenosine phosphorylase and BRCA-associated protein 1 immunohistochemistry in the diagnosis of malignant mesothelioma in effusion cytology specimens. Cancer Cytopathol. 2020;128:126–132.
38. Zygulska AL, Krzemieniecki K, Pierzchalski P. Hippo pathway—brief overview of its relevance in cancer. J Physiol Pharmacol. 2017;68:311–335.
39. Han Y. Analysis of the role of the Hippo pathway in cancer. J Transl Med. 2019;17:116.
40. Hmeljak J, Sanchez-Vega F, Hoadley KA, et al. Integrative molecular characterization of malignant pleural mesothelioma. Cancer Discov. 2018;8:1548–1565.
41. Kinoshita Y, Hamasaki M, Yoshimura M, et al. Hemizygous loss of NF2 detected by fluorescence in situ hybridization is useful for the diagnosis of malignant pleural mesothelioma. Mod Pathol. 2020;33:235–244.
42. Sheffield BS, Lorette J, Shen Y, et al. Immunohistochemistry for NF2, LATS1/2, and YAP/TAZ fails to separate benign from malignant mesothelial proliferations. Arch Pathol Lab Med. 2016;140:391.
43. Pors J, Naso J, Berg K, et al. Cyclin D1 immunohistochemical staining to separate benign from malignant mesothelial proliferations. Mod Pathol. 2020;33:312–318.
44. Christofides A, Karantanos T, Bardhan K, et al. Epigenetic regulation of cancer biology and anti-tumor immunity by EZH2. Oncotarget. 2016;7:85624–85640.
45. Shinozaki-Ushiku A, Ushiku T, Morita S, et al. Diagnostic utility of BAP1 and EZH2 expression in malignant mesothelioma. Histopathology. 2017;70:722–733.
46. Yoshimura M, Kinoshita Y, Hamasaki M, et al. Highly expressed EZH2 in combination with BAP1 and MTAP loss, as detected by immunohistochemistry, is useful for differentiating malignant pleural mesothelioma from reactive mesothelial hyperplasia. Lung Cancer. 2019;130:187–193.
47. Vasanthakumar A, Godley LA. 5-hydroxymethylcytosine in cancer: significance in diagnosis and therapy. Cancer Genet. 2015;208:167–177.
48. Chapel DB, Husain AN, Krausz T. Immunohistochemical evaluation of nuclear 5-hydroxymethylcytosine (5-HMC) accurately distinguishes malignant pleural mesothelioma from benign mesothelial proliferations. Mod Pathol. 2019;32:376–386.
49. Derakhshan F, Ionescu D, Cheung S, et al. Use of programmed death ligand-1 (PD-L1) staining to separate sarcomatoid malignant mesotheliomas from benign mesothelial reactions. Arch Pathol Lab Med. 2020;144:185–188.
50. Chung YS, Kim M, Cha YJ, et al. Expression of V-set immunoregulatory receptor in malignant mesothelioma. Mod Pathol. 2020;33:263–270.
51. Russ A, Hua AB, Montfort WR, et al. Blocking “don’t eat me” signal of CD47-SIRPalpha in hematological malignancies, an in-depth review. Blood Rev. 2018;32:480–489.
52. Schurch CM, Forster S, Bruhl F, et al. The “don’t eat me” signal CD47 is a novel diagnostic biomarker and potential therapeutic target for diffuse malignant mesothelioma. Oncoimmunology. 2017;7:e1373235.
53. Churg A, Galateau-Salle F, Roden AC, et al. Malignant mesothelioma in situ: morphologic features and clinical outcome. Mod Pathol. 2020;33:297–302.
54. Minami K, Jimbo N, Tanaka Y, et al. Malignant mesothelioma in situ diagnosed by methylthioadenosine phosphorylase loss and homozygous deletion of CDKN2A: a case report. Virchows Arch. 2020;476:469–473.

malignant mesothelioma; reactive mesothelial proliferation; BAP1; MTAP; CDKN2A; malignant mesothelioma in situ

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.