BAP1 Immunohistochemistry and p16 FISH to SeparateBenign From Malignant Mesothelial Proliferations

Brandon S. Sheffield, MD,* Harry C. Hwang, MD,w Anna F. Lee, MD,zKim Thompson, HT, ASCP, QIHC,w Stephanie Rodriguez, HT, MB, ASCP,wChristopher H. Tse, MBBS,w Allen M. Gown, MD,w and Andrew Churg, MD*

Abstract: A variety of immunohistochemical (IHC) stains have

been proposed to mark either benign or malignant mesothelial

proliferations. Loss of the p16 tumor suppressor (CDKN2A),

through homozygous deletions of 9p21, is a good marker of

mesotheliomas but lacks sensitivity. Recent reports indicate that

some mesotheliomas are associated with loss of BRCA-asso-

ciated protein 1 (BAP1) expression. Here we investigate BAP1

and p16 as potential markers of malignancy and compare test

characteristics with previously proposed markers using a well-

characterized tissue microarray. BAP1 protein expression was

interrogated by IHC. The p16 locus was examined by fluo-

rescence in situ hybridization (FISH) directed toward chromo-

some 9p21. Loss of BAP1 was identified in 7/26 mesotheliomas

and 0/49 benign proliferations. Loss of p16 was identified in 14/

27 mesotheliomas and 0/40 benign proliferations, yielding 100%

specificity and positive predictive value for each marker. To-

gether, BAP1 IHC and p16 FISH were 58% sensitive for de-

tecting malignancy. Various combinations of p53, EMA, IMP3,

and GLUT1 showed reasonably high specificity (96% to 98%)

but poor to extremely poor sensitivity. Combined BAP1 IHC/

p16 FISH testing is a highly specific method of diagnosing

malignant mesotheliomas when the question is whether a mes-

othelial proliferation is benign or malignant and is particularly

useful when tissue invasion by mesothelial cells cannot be

demonstrated. However, combined BAP1/p16 FISH testing is

not highly sensitive, and negative results do not rule out a

mesothelioma. The test characteristics of previously proposed

markers EMA, p53, GLUT1, IMP3 suggest that, even in

combination, these markers are not useful tools in this clinical

setting.

Key Words: BAP1, BRCA-associated protein 1, p16, CDKN2A,

9p21, mesothelioma

(Am J Surg Pathol 2015;00:000–000)

Immunohistochemistry (IHC) has been instrumental inallowing the pathologist to distinguish mesothelial

proliferations from other processes.1 Delineation betweenbenign reactive and malignant mesothelial proliferationscontinues to present a diagnostic challenge and is offundamental importance to patient treatment and prog-nosis.

Routine morphology, and particularly stromal in-vasion, remains the most useful way to separate benignfrom malignant mesothelial processes.2,3 As the histologicdiagnosis remains a challenge, numerous IHC markershave been previously investigated for this purpose.Membranous staining for epithelial membrane antigen(EMA), p53 nuclear positivity, GLUT1 and IMP3 pos-itivity,4,5 and lack of desmin expression have all beenproposed as markers of malignancy; however, each ofthese tests lacks the specificity required for clinical use.6–9

Homozygous deletion of the 9p21 region has beenpreviously proposed as a marker of malignancy.10 Thiscytogenetic aberration results in loss of the p16(CDKN2A) tumor suppressor, among other genes, and isdetectable by fluorescence in situ hybridization (FISH).Demonstration of p16 deletion is highly specific for ma-lignancy but only demonstrable in a portion of meso-theliomas.11 Deletions of p16 offer more than a diagnosticmarker of malignancy, as they also act as a biomarker,conferring poorer prognosis in mesotheliomas, and mayalso represent a potential target for systemic therapy.12

Recently, loss of BRCA-associated protein 1 (BAP1)has been reported in some mesotheliomas. Germlinemutations in BAP1 have been associated with increasedrisk for malignant mesothelioma,13 as well as uvealmelanoma and renal cell carcinoma.14 Somatic BAP1mutations have been implicated in nonhereditary meso-theliomas.15 BAP1 IHC has been previously utilized forits prognostic value in uveal melanoma and renal cellcarcinoma.16,17

In this study, we examine the utility of BAP1 as amarker of malignancy in mesothelial proliferations, using apreviously described tissue microarray (TMA)9 and inves-tigate whether combining BAP1 IHC with p16 FISH im-proves the separation of benign from malignant mesothelialproliferations. We also follow up on our previous data withthis TMA to determine whether combining multiple IHCmarkers provides high enough specificity for clinical use.

From the *Division of Anatomical Pathology, Vancouver GeneralHospital; zDivision of Anatomical Pathology, Children’s andWomen’s Hospital of British Columbia, Vancouver, BC, Canada;and wPhenoPath Laboratories, Seattle, WA.

Conflicts of Interest and Source of Funding: The authors have disclosedthat they have no significant relationships with, or financial interestin, any commercial companies pertaining to this article.

Correspondence: Andrew Churg, MD, Department of Pathology, Van-couver General Hospital, 910 West 10th Ave., Vancouver, BC,Canada V5Z 1M9 (e-mail: achurg@mail.ubc.ca).

Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

ORIGINAL ARTICLE

Am J Surg Pathol � Volume 00, Number 00, ’’ 2015 www.ajsp.com | 1

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.

METHODSAppropriate research ethics and review board per-

missions were obtained from the University of BritishColumbia and affiliated institutions before the com-mencement of this work.

A previously described9 TMA consisting of a mix-ture of 52 benign mesothelial proliferations and 31 ma-lignant mesotheliomas was utilized.

FISH studies were performed at PhenoPath Labo-ratories on formalin-fixed paraffin-embedded (FFPE)blocks from each patient. Briefly, 4-mm-thick FFPE sec-tions were deparaffinized and pretreated using a VP2000processor (Abbott Laboratories). After deparaffinizationand rehydration, the slides were acid treated in 0.2M HClfor 20 minutes, washed in 2� SSC, placed in 8.1% so-dium thiocyanate, washed in 2� SSC, digested in 0.8%pepsin for 10 minutes, washed in 2� SSC, and postfixedin 10% neutral buffered formalin. Slides were then de-natured for 16 minutes at 791C and then hybridized witha CDKN2A/CC9 probe (Cymogen Diagnostics) that wasdiluted 1:10 in Denhyb probe hybridization buffer (In-situs) and then incubated overnight (14 to 16 h) at 371C.The Cymogen CDKN2A/CC9 probe comprises a 298 kbporange fluor-labeled probe to 9p21.3 and a green fluor-labeled probe to the centromeric region of chromosome 9(9p11.1-q11.1). After incubation, slides were washed for3 minutes in 0.4� SSC/0.3% NP40 at 791C, followed bywashing at 251C in 2� SSC for 3 minutes and quick rinsein dH2O. After air drying, slides were counterstainedusing DAPI/Fluorguard solution and then coverslipped.Slides were then scanned using a Metasystems slidescanner (MetaSystems, Altlussheim, Germany). A pos-itive homozygous deletion event was defined as a DAPI-stained nucleus with no orange signals and at least 1 greensignal. Using the Metasystems-captured images, at least50 to 100 nuclei were counted for each case, and thepercentage of positive nuclei was calculated. The thresh-old, for p16 homozygous deletion, to account for sectiontruncation artifact, in FFPE sections, was 12%.

IHC studies were also performed at PhenoPathlaboratories under the supervision of A.M.G. and H.C.H.on deparaffinized 4-mm-thick sections after pretreatmentusing a Dako AutostainerLink 48. For BAP-1, mousemonoclonal anti-human BAP-1 clone C-4 (Santa CruzBiotechnology Inc., Dallas, TX) was used at 1:400 onslides pretreated for 20 minutes in a steamer in pH 9 Tris-EDTA buffer. BAP-1 was detected using Ultravision de-tection system (Thermo Scientific, Waltham, MA). BAP1staining produces a nuclear signal in all non-neoplasticcells, creating a useful internal control. BAP1 loss wasidentified by a homogenous loss of nuclear staining intumor cells. Non-nuclear staining was disregarded. AllTMA cores were scored as either BAP1 lost or intact. Forp53, mouse monoclonal anti-human p53 clone DO-7(Dako, Carpinteria, CA) was used at 1:1500 on slidespretreated for 8 minutes in a pressure cooker in pH 6citrate buffer. P53 was detected using mouse envisionreagent (Dako). For EMA, mouse monoclonal anti-hu-man EMA clone E29 (Dako) was used at 1:200 on slides

pretreated for 8 minutes in a pressure cooker in pH 6citrate buffer. EMA was detected using mouse envisionreagent (Dako).

Test characteristics were calculated for the in-dividual markers as well as both markers in combination.Sensitivity and specificity were calculated by computing(true positives)/(true positives+false negatives) and (truenegatives)/(false positives+true negatives), respectively,with associated 95% confidence intervals calculated bys±1.96�O(s� ((1�s)/n)), where s=either sensitivity orspecificity, and n= the total number of cases evaluated.In all calculations, the original histologic diagnosis wasused as a gold standard.

Previously published9 IHC data for GLUT1 andIMP3 on the same TMA was reanalyzed in combinationwith the new IHC data. Combination analysis of markerswas performed either by interpreting simultaneous pos-itivity of markers as positive (and), or by interpretingpositivity in any marker as being positive (or).

RESULTSEighty-three mesothelial proliferations were in-

cluded in the TMA, including 31 malignant meso-theliomas (27 pleural, 3 peritoneal, and 1 pericardial siteof origin) and 52 benign mesothelial proliferations (de-termination of benign vs. malignant was made in theoriginal clinical diagnosis, and indeterminate cases wereexcluded from the TMA). Multiple cores were included torepresent each proliferation.

Scorable results for BAP1 were obtained for 75/83(93%) cases, 8 cases were unscored because of technicalissues. None of the benign mesothelial proliferationsshowed loss of BAP1 expression (0/49 [0%]). Seven ma-lignant mesotheliomas showed loss of BAP1 expression(7/26 [27%]) (Fig. 1). The calculated sensitivity and spe-cificity (95% confidence interval) of BAP1 loss for diag-nosing malignant mesothelioma were 27% (17%-37%)and 100% (100%-100%), respectively (Table 1).

FISH results for p16 were obtained for 67/83 (81%)cases (16 cases were not scored because of technical issues).Zero of 40 (0%) benign proliferations demonstrated p16deletion, whereas 14/27 (52%) malignant mesotheliomasshowed homozygous loss at this locus (Fig. 2). The cal-culated sensitivity and specificity of p16 FISH for diag-nosing malignant mesothelioma were 52% (40% to 64%)and 100% (100% to 100%), respectively (Table 1).

When combining p16 FISH and BAP1 IHC as apanel, results from both markers were available in 61/83(73%) cases. Fourteen of 24 mesotheliomas showed eitherloss of BAP1 expression or homozygous p16 deletion (14/24 [58%]). Loss of BAP1 and homozygous p16 deletionwas not found in any of the 37 benign cases (0/37 [0%]).Three cases of mesothelioma showed both BAP1 loss andhomozygous p16 deletion (3/24 [12%]). The calculatedsensitivity and specificity (95% confidence interval) ofBAP1 IHC and p16 FISH used as a panel to diagnosemesothelioma were 58% (46%-71%) and 100% (100%-100%), respectively (Table 2).

Sheffield et al Am J Surg Pathol � Volume 00, Number 00, ’’ 2015

2 | www.ajsp.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.

For comparison, we performed IHC for p53 andEMA, 2 previously proposed IHC markers for malignantmesothelioma. Scorable results were available for 77/83(93%) and 81/83 (98%) cases for p53 and EMA, re-spectively. Alone, neither marker achieved clinicallymeaningful specificity (53% for p53, and 84% forEMA; Table 1). In combination, p53 nuclear positivitytogether with strong membranous EMA expressionshowed 88% specificity for diagnosing malignancy, with asensitivity of 25% (Table 2).

In addition, we utilized previously published IHCdata on the same TMA for additional comparisons. Thedata published by Lee and colleagues show that a panel of

A B

C

FIGURE 1. Examples of retained (A) and lost (B and C) BAP1 expression. A, TMA core of benign reactive mesothelial atypiashowing intact BAP1 protein expression (yellow arrows). Note nonmesothelial stromal and inflammatory cells showing positivenuclear BAP1 expression (red and green arrows); these cells serve as positive internal controls. B and C, TMA cores of malignantmesotheliomas showing loss of nuclear BAP1 expression (yellow arrows). Note nuclear positivity within stromal and inflammatorycells (red and green arrows). When assessing BAP1 status, only nuclear positivity represents a signal of interest; cytoplasmic,membranous and extracellular staining is disregarded.

TABLE 1. TMA Demographics

n (%)

Benign mesothelium 52 (63)Pleural 27 (52)Peritoneal 25 (48)

Malignant mesothelioma 31 (37)Pleural 27 (87)Peritoneal 3 (10)Pericardial 1 (3)Epithelioid 15 (48)Sarcomatoid 11 (35)Biphasic 5 (16)

Total 83

Am J Surg Pathol � Volume 00, Number 00, ’’ 2015 BAP-1, p16 FISH, and Mesothelioma

Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ajsp.com | 3

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.

IMP3 and GLUT1 can be used to diagnose malignancywith a sensitivity of 43% and a specificity of 96%. Addingp53 or EMA to a panel of IMP3 and GLUT1 reduced thesensitivity of the panel from 43% to 36% and 20%,

respectively, and did not produce any clinically mean-ingful change in the specificity, which modestly improvedfrom 96% to 98% in the combined panel of EMA, IMP3,and GLUT1 (Table 3).

DISCUSSIONThe separation of benign from malignant meso-

thelial proliferations is crucial to patient treatment andprognosis. Pleural mesotheliomas are universally fatal,even with radical triple modality therapy. The prognosisfor peritoneal mesotheliomas is considerably better, buteven here extensive surgery and hyperthermic intra-peritoneal chemotherapy are required.18 However, mostbenign mesothelial processes require nothing more thanwatchful waiting and patient reassurance.

The importance of making this separation meansthat ancillary tests should ideally have a specificity andpositive predictive value of 100%, higher values than weordinarily expect from tests designed to separate onemalignancy from another. As noted in the introduction,there are a variety of IHC tests that have been proposedas markers of mesothelial malignancy or benignancy. Anumber of them work as statistical predictors in largeseries of cases.6–8 For example, we found that about 70%of patients with <10% of cells staining for p53 or EMAsurvived 5 years, compared with 30% of patients with>10% staining for these markers.3 Here, overall survivalis used as a surrogate gold standard for malignancy,showcasing the unreliability of these 2 markers. Using thesame TMA as the current study, we reported that IMP3was positive in 53% of mesotheliomas and 27% of benigncases, whereas GLUT1 was positive in 60% of meso-theliomas and 13% of benign cases. However, these dif-ferences are not good enough for use on an individualbasis.

We noted in Lee and colleagues’ study that com-bining IMP3 and GLUT1 data resulted in positivity in43% of malignant cases and only 4% of benign cases.This is certainly an improvement over the single markerdata and led us to suggest that combined use of multiplemarkers may enable a sufficiently specific panel for di-agnosing malignancy. We have explicitly tested that hy-pothesis in this paper by adding p53 or EMA to IMP3and GLUT1; however, the improvement in specificity onusing just IMP3 plus GLUT1 is small to nonexistent, andthe sensitivity is extremely low (Table 2).

Accordingly, we evaluated another approach usingp16 FISH or BAP1 IHC. Although the total number of

FIGURE 2. Examples of p16 retained (A) and p16 deleted (B)mesothelial proliferations. A, TMA core of a benign mesothelialproliferation showing dual nuclear signals, orange dots cor-respond to the p16 (CDKN2A) locus, green dots correspond tothe centromere region of chromosome 9. The majority of cellsshow both signals in an approximately equal ratio indicatingthat no loss of p16 has occurred. B, TMA core of mesotheliomashowing homozygous loss of p16. Majority of cells show greensignal(s) without corresponding orange signal.

TABLE 2. Test Characteristics of p16 FISH, BAP1, EMA, p53, IMP3, and GLUT1 IHC

Marker BAP1 IHC p16 FISH EMA p53 IMP3* GLUT1*

n 75 67 81 77 78 78Benign 0/49 0/40 8/50 23/49 13/48 6/48Malignant 7/26 14/27 10/31 16/28 16/30 18/30Sensitivity (95% CI) (%) 27 (17-37) 52 (40-64) 32 (22-42) 57 (46-68) 53 (42-64) 60 (49-71)Specificity (95% CI) (%) 100 (100-100) 100 (100-100) 84 (76-92) 53 (42-64) 73 (63-83) 88 (80-95)

*Data from Lee et al.9

Sheffield et al Am J Surg Pathol � Volume 00, Number 00, ’’ 2015

4 | www.ajsp.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.

reported cases is fairly small, homozygous deletion of p16by FISH has only been found in mesotheliomas and neverfound in benign proliferations.11,19,20 We reach that sameconclusion in the present study: homozygous loss of p16was 100% specific in our TMA series. The sensitivity forp16 FISH alone was 52%. This is not dissimilar to thereported values in the literature.21–23

BAP1 IHC has been used as a screen for familialcancer syndromes24 and to identify prognostic subgroupsin uveal melanoma and renal cell carcinoma.17,25,26 BAP1loss of expression by IHC correlates well with BAP1mutations and deletions at the DNA level.16 To ourknowledge, the use of BAP1 IHC for separating benignfrom malignant mesothelial proliferations has not beenreported in the literature. We show here that BAP1 loss is100% specific for mesothelioma, but by itself the markeris quite insensitive, although the value of 27% in Table 2is similar to what has been reported in the screeningstudies of mesothelioma.27 Of more interest, combiningBAP1 and p16 FISH produced a sensitivity of 58% with aspecificity of 100%, suggesting that this combination is agood, although not sensitive, diagnostic test. Whencomparing the various panels presented in Table 2 it isclear that BAP1 IHC combined with p16 FISH offers thislevel of diagnostic certainty, albeit only in about 60% ofcases.

The test characteristics presented in this report arelimited by the fact that they are calculated on the basis ofTMA sections. For FISH assays, this limits the number ofnuclei available for assessment. For both IHC and FISH,TMA methodology limits assessment for intratumoralheterogeneity. Additional studies are needed to determinewhether focal BAP1 loss may affect interpretation, as thisphenomenon has been previously reported in other ma-lignancies.28

Unlike previously proposed IHC markers, loss ofthe 9p21 region plays a role in mesothelioma develop-ment. Homozygous deletion of this region results in theloss of the CDKN2A tumor-suppressor gene (p16), as wellas the ARF tumor-suppressor gene (p14). Homozygousp16 deletion has been shown to identify a poorer prog-nostic subgroup. Closely linked, and commonly code-leted, is the MTAP gene, which may also be relevant inthe pathogenesis and possibly treatment of meso-theliomas. Thus FISH for p16 not only provides diag-

nostic utility but also prognostic and potentiallypredictive information.

Similarly, BAP1 may be more than just a marker ofmalignancy. Previous studies have shown that BAP1mutations do not confer any overall survival differencefrom BAP1 wild-type tumors.29 Nonetheless, BAP1 is atumor-suppressor gene that has been proposed as a po-tential target for systemic therapy,29 and its causative rolein mesothelioma is demonstrated by the heritable riskseen in germline BAP1 mutation carriers. Loss of BAP1expression by IHC may therefore be predictive of un-derlying germline BAP1 mutation. Hence, IHC may beused to aid in diagnosis and secondarily used to guideBAP1 germline testing in cases with supportive clinicalfeatures or family history. Additional studies are neededto determine the overall utility of BAP1 IHC as part ofthe diagnostic algorithm for hereditary mesothelioma (acomponent of the larger inherited BAP1 tumor predis-position syndrome).30

A panel of BAP1 IHC and p16 FISH will allow fora diagnosis of malignant mesothelioma in about 60%, butthe sensitivity of these tests, in combination, is insufficientto exclude a mesothelioma in the setting of a negativepanel result. Additional studies, and likely novel markers,are needed before the ideal malignant mesothelialscreening panel is developed for clinical use. Making themost of what is currently available, these 2 tests can beuseful to pathologists and clinicians when they are ap-plied correctly.

The limitations of this study must be emphasized.First, the approach of using p16 FISH and BAP1 IHConly applies to proliferations that have been establishedas mesothelial; they cannot be used to separate meso-theliomas from other tumors (and other types of malig-nancies can show loss of these markers). Second, whereasthe panel appears to be highly specific when 1 or bothmarkers is deleted, the converse does not apply: positivestaining for BAP1 by IHC or a positive p16 signal byFISH does not mean that a mesothelial proliferation isbenign, because 40% of mesotheliomas will not show lossof these markers. Third, although the combination of p16FISH and BAP1 IHC appears to be highly specificwhen either marker is lost, we caution against applyingthese tests in a blind manner. Tissue invasion is thebest marker of malignancy with mesothelial processes,

TABLE 3. Test Characteristics of BAP1 IHC or p16 FISH Compared With Other Proposed Panels*

Panel

BAP1 IHC or

p16 FISH

EMA or

p53 IHC

EMA and

p53 IHC

IMP3 and

GLUT1 IHCwp53 and IMP3 and

GLUT1 IHCz

EMA and IMP3

and GLUT1 IHCz

n 61 76 76 78 73 76Benign 0/37 24/48 6/48 2/48 2/45 1/46Malignant 14/24 18/28 7/28 13/30 10/28 6/30Sensitivity (95% CI) (%) 58 (46-71) 64 (54-75) 25 (15-35) 43 (32-54) 36 (25-47) 20 (11-29)Specificity (95% CI) (%) 100 (100-100) 50 (39-61) 88 (80-95) 96 (91-100) 96 (91-100) 98 (95-100)

*Combinations designated as OR, required positivity of either marker. Combinations designated as AND, required positivity for both markers.wData from Lee et al.9

zData represents a synthesis from Lee et al9 and this study.

Am J Surg Pathol � Volume 00, Number 00, ’’ 2015 BAP-1, p16 FISH, and Mesothelioma

Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved. www.ajsp.com | 5

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.

and p16 FISH/BAP1 IHC testing should be reserved forcases in which invasion cannot be demonstrated or isequivocal.

REFERENCES1. Husain AN, Colby T, Odonez N, et al. Guidelines for pathologic

diagnosis of malignant mesothelioma: 2012 update of the consensusstatement from the International Mesothelioma Interest Group.Arch Pathol Lab Med. 2013;137:647–667.

2. Henderson DW, Reid G, Kao SC, et al. Challenges andcontroversies in the diagnosis of mesothelioma: Part 1. Cytology-only diagnosis, biopsies, immunohistochemistry, discriminationbetween mesothelioma and reactive mesothelial hyperplasia, andbiomarkers. J Clin Pathol. 2013;66:847–853.

3. Churg A, Galateau-Salle F. The separation of benign and malignantmesothelial proliferations. Arch Pathol Lab Med. 2012;136:1217–1226.

4. Kato Y, Tsuta K, Seki K, et al. Immunohistochemical detection ofGLUT-1 can discriminate between reactive mesothelium andmalignant mesothelioma. Mod Pathol. 2007;20:215–220.

5. Shi M, Fraire AE, Chu P, et al. Oncofetal protein IMP3, a newdiagnostic biomarker to distinguish malignant mesothelioma fromreactive mesothelial proliferation. Am J Surg Pathol. 2011;35:878–882.

6. Cagle PT, Brown RW, Lebovitz RM. P53 immunostaining in thedifferentiation of reactive processes from malignancy in pleuralbiopsy specimens. Hum Pathol. 1994;25:443–448.

7. Attanoos RL, Griffin A, Gibbs AR. The use of immunohistochem-istry in distinguishing reactive from neoplastic mesothelium. A noveluse for desmin and comparative evaluation with epithelial mem-brane antigen, p53, platelet-derived growth factor-receptor,P-glycoprotein and Bcl-2. Histopathology. 2003;43:231–238.

8. King J, Thatcher N, Pickering C, et al. Sensitivity and specificity ofimmunohistochemical antibodies used to distinguish between benignand malignant pleural disease: a systematic review of publishedreports. Histopathology. 2006;49:561–568.

9. Lee AF, Gown AM, Churg A. IMP3 and GLUT-1 immunohis-tochemistry for distinguishing benign from malignant mesothelialproliferations. Am J Surg Pathol. 2013;37:421–426.

10. Illei PB, Ladanyi M, Rusch VW, et al. The use of CDKN2A deletionas a diagnostic marker for malignant mesothelioma in body cavityeffusions. Cancer. 2003;99:51–56.

11. Monaco SE, Shuai Y, Bansal M, et al. The diagnostic utility of p16FISH and GLUT-1 immunohistochemical analysis in mesothelialproliferations. Am J Clin Pathol. 2011;135:619–627.

12. Ladanyi M. Implications of P16/CDKN2A deletion in pleuralmesotheliomas. Lung Cancer. 2005;49(S1):S95–S98.

13. Testa JR, CheungM, Pei J, et al. Germline BAP1 mutations predisposeto malignant mesothelioma. Nat Genet. 2011;43:1022–1025.

14. Murali R, Wiesner T, Scolyer RA. Tumours associated with BAP1mutations. Pathology. 2013;45:116–126.

15. Bott M, Brevet M, Taylor BS, et al. The nuclear deubiquitinaseBAP1 is commonly inactivated by somatic mutations and 3p21.1

losses in malignant pleural mesothelioma. Nat Genet. 2011;43:668–672.

16. Koopmans AE, Verdijk RM, Brouwer RW, et al. Clinicalsignificance of immunohistochemistry for deletion of BAP1mutations in uveal melanoma. Mod Pathol. 2014;27:1321–1330.

17. Joseph RW, Kapur P, Serie DJ, et al. Loss of BAP1 proteinexpression is an independent marker of poor prognosis in patientswith low-risk clear cell renal cell carcinoma. Cancer. 2014;120:1059–1067.

18. Cao C, Yan TD, Deraco M, et al. Importance of gender in diffusemalignant peritoneal mesothelioma. Ann Oncol. 2012;23:1494–1498.

19. Chung CT, Santos Gda C, Hwang DM, et al. FISH assaydevelopment for the detection of p16/CDKN2A deletion inmalignant pleural mesothelioma. J Clin Pathol. 2010;63:630–634.

20. Chiosea S, Krasinskas A, Cagle PT, et al. Diagnostic importance of9p21 homozygous deletion in malignant mesotheliomas. ModPathol. 2008;21:742–747.

21. Hwang H, Tse C, Rodriguez S, et al. p16 FISH deletion in surfaceepithelial mesothelial proliferations is predictive of underlyinginvasive mesothelioma. Am J Surg Pathol. 2014;38:681–688.

22. Wu D, Hiroshima K, Matsumoto S, et al. Diagnostic usefulnessof p16/CDKN2A FISH in distinguishing between sarcomatoidmesothelioma and fibrous pleuritis. Am J Clin Pathol. 2013;139:39–46.

23. Tochigi N, Attanoos R, Chirieac LR, et al. p16 Deletion insarcomatoid tumors of the lung and pleura. Arch Pathol Lab Med.2013;137:632–636.

24. Betti M, Casalone E, Ferrante D, et al. Inference on germline BAP1mutations and asbestos exposure from the analysis of familial andsporadic mesothelioma in a high-risk area. Genes ChromosomesCancer. 2015;54:51–62.

25. Kalirai H, Dodson A, Fagir S, et al. Lack of BAP1 proteinexpression in uveal melanoma is associated with increased meta-static risk and has utility in routine prognostic testing. Br J Cancer.2014;111:1373–1380.

26. Shah AA, Bourne TD, Murali R. BAP1 protein loss by immuno-histochemistry: a potentially useful tool for prognostic prediction inpatients with uveal melanoma Pathology. 2013;45:651–656.

27. Zauderer MG, Bott M, McMillan R, et al. Clinical characteristics ofpatients with malignant pleural mesothelioma harboring somaticBAP1 mutations. J Thorac Oncol. 2013;8:1430–1433.

28. De la Fouchardiere A, Cabaret O, Savin L, et al. Germline BAP1mutations predispose also to basal cell carcinomas. Clin Genet. 2014.[Epub ahead of print].

29. Ladanyi M, Zauderer MG, Krug LM, et al. New strategies inpleural mesothelioma: BAP1 and NF2 as novel targets fortherapeutic development and risk assessment. Clin Cancer Res.2012;18:4485–4490.

30. Battaglia A. The importance of multidisciplinary approach inearly detection of BAP1 tumor predisposition syndrome: clinicalmanagement and risk assessment. Clin Med Insights Oncol. 2014;8:37–47.

Sheffield et al Am J Surg Pathol � Volume 00, Number 00, ’’ 2015

6 | www.ajsp.com Copyright r 2015 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of the article is prohibited.