CD34 and α smooth muscle actin distinguish verrucous ...

6
CD34 and a smooth muscle actin distinguish verrucous hyperplasia from verrucous carcinoma Kristen M. Paral, MD, a Jerome B. Taxy, MD, b and Mark W. Lingen, DDS, PhD, FRCPath c University of Chicago, Chicago, IL, USA; NorthShore University Health System, Evanston, IL, USA Objective. This study evaluated the use of stromal biomarkers CD34 and a smooth muscle actin (a-SMA) to distinguish verrucous carcinoma (VC) from verrucous hyperplasia (VH). Study Design. Thirteen VH, 15 VC, 20 squamous cell carcinoma (SCC), and 16 of uninvolved adjacent stroma specimens were analyzed for a-SMA and CD34 expression by immunohistochemistry. Results. Stromal a-SMA positivity was observed in 100% (20 of 20) of the SCC and in 93% (14 of 15) of the VC, whereas none of the VH (0 of 13) or adjacent uninvolved stroma (0 of 16) demonstrated a-SMA reactivity. Stromal CD34 positivity was observed in 100% (13 of 13) of VH and adjacent stroma (16 of 16), while 20% (3 of 15) of VC and 11% (2 of 18) of SCC stroma expressed CD34. The SCC and VC groups differed significantly from the VH and uninvolved stroma groups for both a-SMA and CD34 expression (P < .0001). Conclusions. Stromal CD34 and a-SMA protein expression patterns may aid in distinguishing between VC and VH in challenging cases. (Oral Surg Oral Med Oral Pathol Oral Radiol 2014;117:477-482) Verrucous carcinoma (VC) consists of deceptively bland, broad-based, inwardly undulating epithelium devoid of the usual cytologic criteria for malignancy. 1 Verrucous hyperplasia (VH) is described as a lesion that resembles VC but differs in that its base lies above the line connecting the bases of the uninvolved epithelium on both sides of the lesion. 2 This feature may be difcult or impossible to evaluate, owing to inadequate segments of surrounding normal tissue or technical difculties with embedding and cutting. A lesion exhibiting growth both above and below such a line creates further diagnostic confusion. Epithelial features do not clearly resolve the issue, as both lesions lack malignant cytology. The stroma, however, may provide an alternative means by which to distinguish these lesions. The stroma associated with invasive carcinoma is characterized by a loss of CD34 þ dendritic cells and a gain of a smooth muscle actinepositive (a-SMA þ ) myobroblasts; the opposite is seen in un- involved adjacent stroma. These principles hold true regardless of anatomic location, as previously described in squamous cell carcinoma (SCC) of the skin, 3 cer- vix, 4,5 esophagus, 6 and upper aerodigestive tract 7,8 ; in ductal carcinoma of the breast 9-11 ; and in adenocarci- noma of the pancreas 12 and colon. 13 Additionally, a gradation of these stromal alterations has been described in low- to high-risk intraepithelial lesions of the breast, 9 cervix, 4 and oral mucosa. 8 The previous reports in the skin and upper aero- digestive tract include one report of myobroblasts in oral VC, but VC and VH have not been specif- ically addressed. 3,7,8 This study was performed to compare the stromal reactions among conventional inltrating SCC, VC, VH, and uninvolved adjacent stroma with regard to CD34 and a-SMA protein expression. MATERIALS AND METHODS After institutional review board approval, 13 VH speci- mens, 15 VC specimens, and 20 conventional SCC specimens from mucocutaneous body sites were retrieved from the pathology archives. Using established criteria, 14 histopathologic interpretation of the formalin- xed, parafn-embedded, hematoxylin-eosinestained sections was performed by 3 pathologists (K.P., J.B.T., and M.W.L.) to conrm the diagnosis of VH, VC, or SCC. Among the cases analyzed, 16 contained tumor- free stroma for comparison (5 VC, 4 VH, and 7 SCC). The stroma associated with the lesional epithelium was dened as tumor stroma, whereas the uninvolved adja- cent stroma associated with the normal epithelium was designated as tumor-free stroma. a Resident, Department of Pathology, University of Chicago. b Professor, Department of Pathology, NorthShore University Health System. c Professor, Department of Pathology, University of Chicago. Received for publication Nov 6, 2013; returned for revision Dec 11, 2013; accepted for publication Dec 14, 2013. Ó 2014 Elsevier Inc. 2212-4403 http://dx.doi.org/10.1016/j.oooo.2013.12.401 Statement of Clinical Relevance The differentiation between VH and VC can be challenging in histopathologic diagnosis. This study proposes that stromal expression of CD34 and a smooth muscle actin proteins may aid in dis- tinguishing between these 2 entities. 477 Vol. 117 No. 4 April 2014 Open access under CC BY-NC-ND license.

Transcript of CD34 and α smooth muscle actin distinguish verrucous ...

Page 1: CD34 and α smooth muscle actin distinguish verrucous ...

Vol. 117 No. 4 April 2014

CD34 and a smooth muscle actin distinguish verrucoushyperplasia from verrucous carcinomaKristen M. Paral, MD,a Jerome B. Taxy, MD,b and Mark W. Lingen, DDS, PhD, FRCPathc

University of Chicago, Chicago, IL, USA; NorthShore University Health System, Evanston, IL, USA

Objective. This study evaluated the use of stromal biomarkers CD34 and a smooth muscle actin (a-SMA) to distinguish

verrucous carcinoma (VC) from verrucous hyperplasia (VH).

Study Design. Thirteen VH, 15 VC, 20 squamous cell carcinoma (SCC), and 16 of uninvolved adjacent stroma specimens

were analyzed for a-SMA and CD34 expression by immunohistochemistry.

Results. Stromal a-SMA positivity was observed in 100% (20 of 20) of the SCC and in 93% (14 of 15) of the VC, whereas

none of the VH (0 of 13) or adjacent uninvolved stroma (0 of 16) demonstrated a-SMA reactivity. Stromal CD34 positivity

was observed in 100% (13 of 13) of VH and adjacent stroma (16 of 16), while 20% (3 of 15) of VC and 11% (2 of 18) of

SCC stroma expressed CD34. The SCC and VC groups differed significantly from the VH and uninvolved stroma groups

for both a-SMA and CD34 expression (P < .0001).

Conclusions. Stromal CD34 and a-SMA protein expression patterns may aid in distinguishing between VC and VH in

challenging cases. (Oral Surg Oral Med Oral Pathol Oral Radiol 2014;117:477-482)

Verrucous carcinoma (VC) consists of deceptivelybland, broad-based, inwardly undulating epitheliumdevoid of the usual cytologic criteria for malignancy.1

Verrucous hyperplasia (VH) is described as a lesionthat resembles VC but differs in that its base lies abovethe line connecting the bases of the uninvolvedepithelium on both sides of the lesion.2 This featuremay be difficult or impossible to evaluate, owing toinadequate segments of surrounding normal tissue ortechnical difficulties with embedding and cutting. Alesion exhibiting growth both above and below such aline creates further diagnostic confusion. Epithelialfeatures do not clearly resolve the issue, as both lesionslack malignant cytology. The stroma, however, mayprovide an alternative means by which to distinguishthese lesions. The stroma associated with invasivecarcinoma is characterized by a loss of CD34þ dendriticcells and a gain of a smooth muscle actinepositive(a-SMAþ) myofibroblasts; the opposite is seen in un-involved adjacent stroma. These principles hold trueregardless of anatomic location, as previously describedin squamous cell carcinoma (SCC) of the skin,3 cer-vix,4,5 esophagus,6 and upper aerodigestive tract7,8; inductal carcinoma of the breast9-11; and in adenocarci-noma of the pancreas12 and colon.13 Additionally, agradation of these stromal alterations has been

aResident, Department of Pathology, University of Chicago.bProfessor, Department of Pathology, NorthShore University HealthSystem.cProfessor, Department of Pathology, University of Chicago.Received for publication Nov 6, 2013; returned for revision Dec 11,2013; accepted for publication Dec 14, 2013.� 2014 Elsevier Inc.2212-4403http://dx.doi.org/10.1016/j.oooo.2013.12.401

Open access under CC BY-NC-ND license.

described in low- to high-risk intraepithelial lesions ofthe breast,9 cervix,4 and oral mucosa.8

The previous reports in the skin and upper aero-digestive tract include one report of myofibroblastsin oral VC, but VC and VH have not been specif-ically addressed.3,7,8 This study was performed tocompare the stromal reactions among conventionalinfiltrating SCC, VC, VH, and uninvolved adjacentstroma with regard to CD34 and a-SMA proteinexpression.

MATERIALS AND METHODSAfter institutional review board approval, 13 VH speci-mens, 15 VC specimens, and 20 conventional SCCspecimens from mucocutaneous body sites wereretrieved from the pathology archives. Using establishedcriteria,14 histopathologic interpretation of the formalin-fixed, paraffin-embedded, hematoxylin-eosinestainedsections was performed by 3 pathologists (K.P., J.B.T.,and M.W.L.) to confirm the diagnosis of VH, VC, orSCC. Among the cases analyzed, 16 contained tumor-free stroma for comparison (5 VC, 4 VH, and 7 SCC).The stroma associated with the lesional epithelium wasdefined as tumor stroma, whereas the uninvolved adja-cent stroma associated with the normal epithelium wasdesignated as tumor-free stroma.

Statement of Clinical Relevance

The differentiation between VH and VC can bechallenging in histopathologic diagnosis. This studyproposes that stromal expression of CD34 and asmooth muscle actin proteins may aid in dis-tinguishing between these 2 entities.

477

Page 2: CD34 and α smooth muscle actin distinguish verrucous ...

Table I. Stromal patterns of a-SMAþ myofibroblasticcells

Pattern � þ þþTumor-free 16 0 0Verrucous hyperplasia 13 0 0Verrucous carcinoma 1 7 7Squamous cell carcinoma 0 11 9

a-SMA, a smooth muscle actin.

ORAL AND MAXILLOFACIAL PATHOLOGY OOOO

478 Paral, Taxy and Lingen April 2014

ImmunohistochemistrySections were deparaffinized and rehydrated withxylene and serial dilutions of ethyl alcohol to distilledwater. Tissue sections were incubated in 1 � sodiumcitrate buffer at a pH of 6 and heated in a steamerfor 20 minutes. Sections were pretreated withDAKO target retrieval solution S1699 (DAKO NorthAmerica Inc, Carpinteria, CA, USA) for a-SMA.Anti-a-SMA antibody (DAKO North America;M0851, mouse immunoglobulin G [IgG], dilution1:100) and anti-CD34 antibody (Leica MicrosystemsInc, Buffalo Grove, IL, USA; NCL-END, mouse IgG,dilution 1:25) were applied on tissue sections for a1-hour incubation at room temperature in a humiditychamber. The antigen-antibody binding was detectedwith labeled antimouse polymerehorseradish perox-idase Envision þ system (K4001; DAKO NorthAmerica) and DAB þ chromogen system (K3468;DAKO North America). Tissue sections were brieflyimmersed in hematoxylin for counterstaining. In allcases, staining of blood vessels served as positiveinternal controls.

Semiquantitative assessment and statistical analysisThe percentage of immunoreactive tumor-free andtumor-associated stromal cells excluding vessels wasrecorded as follows: �, no positive cells; þ, focal(<50% positive cells); and þþ, strong (>50% positivecells), as described by Barth et al.7 The data wereanalyzed statistically using STATA 12.1 (StataCorpLP) in the University of Chicago Biostatistics Lab-oratory. Differences in ordinal immunohistochemicallevels among groups were analyzed by the Kruskal-Wallis test followed by the Wilcoxon rank sum test forpairwise comparisons. In addition, because uninvolvedadjacent tumor-free and tumor-associated stroma tis-sues were frequently evaluated in the same patient,proportional odds models were fit with an adjustedvariance estimate to allow for the within-patient cor-relation. Sensitivity and specificity were calculatedseparately.

RESULTSOne hundred percent (20 of 20) of the tumor-associatedstroma from the SCC specimens and 93% (14 of 15)from the VC cases demonstrated a-SMA positivity(Table I; Figures 1 and 2). Strong a-SMA staining wasobserved in 45% (9 of 20) of SCC and 46% (7 of 15) ofVC, whereas focal staining was seen in 55% (9 of 20)of SCC and 46% (7 of 15) of VC. Conversely, none ofthe VH (0 of 13) or areas of adjacent tumor-free stroma(0 of 16) demonstrated a-SMA reactivity (Figure 3). AllVH (13 of 13) as well as the adjacent tumor-free stroma(16 of 16) exhibited CD34 positivity, with 69% (9 of

13) of the VH and 81% (13 of 16) of the adjacentstroma showing strong positivity (Table II). How-ever, only 20% (3 of 15) of VC expressed CD34, withall positive cases being focal. Of note, CD34 wasfocally lost in association with inflammatory infiltrates(Figure 4). The SCC and VC groups differed signifi-cantly from the VH and tumor-free stroma groups forboth a-SMA and CD34 (P < .0001). There was nosignificant difference between SCC and VC (P ¼ .91and P ¼ .41 for a-SMA and CD34, respectively) orbetween VH and tumor-free stroma (P ¼ 1.0 andP ¼ .73 for a-SMA and CD34, respectively). Resultsfor within-patient correlation were similar to thoseof the nonparametric tests (data not shown). Stronga-SMA positivity combined with complete CD34negativity was 100% specific for carcinoma-associatedstroma, whereas diffuse CD34 positivity combined withcomplete a-SMA negativity was 100% specific forbenign-associated stroma. a-SMA was 93% sensitivefor VC-associated stroma, and CD34 was 100% sensi-tive for adjacent stroma.

DISCUSSIONThe present series supports a role for combined CD34and a-SMA analysis in distinguishing VH from VC.Loss of CD34þ dendritic cells, together with a gain ofa-SMAþ myofibroblasts, supports a diagnosis of VC,whereas the opposite reaction supports VH. DiffuseCD34 positivity is strongly indicative of benignity,and its loss is a sensitive marker for malignant stromalalterations. CD34 loss is not specific for malignancy,however, as inflammatory infiltrates were noted todisrupt the dendritic meshwork, a finding that hasbeen previously described.7 Nonspecific CD34 losshas also been seen in association with scars andbiopsy sites in skin, in breast tissue, and in the upperaerodigestive tract.7,11,15-17 In the same body sites,a-SMA positivity has been noted in the stroma ofgranulation tissue.7,11,15-17 In the present series, pos-itivity for a-SMA was seen only in malignantstroma; however, not all carcinoma cases were posi-tive for a-SMA. Thus, the discriminatory utility ofimmunohistochemistry for VH and VC hinges on thecombined patterns of the 2 markers along with

Page 3: CD34 and α smooth muscle actin distinguish verrucous ...

Fig. 1. A-C, Hematoxylin-eosinestained sections of SCC demonstrating infiltrative islands of squamous epithelium showingcytologic features of malignancy (original magnification � 10, � 50, and � 200, respectively). D-F, Immunohistochemicalstaining for CD34, demonstrating the absence of a meshwork from the stroma surrounding the tumor cells (originalmagnification � 10, � 50, and � 200, respectively). G-I, Immunohistochemical staining for a-SMA demonstrating myofi-broblasts enveloping the SCC nests (original magnification � 10, � 50, and � 200, respectively). (SCC, squamous cell car-cinoma; a-SMA, a smooth muscle actin.)

OOOO ORIGINAL ARTICLE

Volume 117, Number 4 Paral, Taxy and Lingen 479

contextual prudence in cases with prior surgicalmanipulation.

In addition to offering a diagnostic adjunct to thecurrent histologic criteria, these results stimulatediscourse with respect to the biologic relationshipof VH to VC. Whether VH is the same as VC, aprecursor to VC, or a separate entity altogether hasbeen a subject of debate. Since it was first describedin 1980 by Shear and Pindborg,2 VH has been seenin association with leukoplakia (53%), VC (29%),epithelial dysplasia (66%), and conventional SCC(10%),18 but its placement on a spectrum withthese lesions has not been established. Many authorshave suggested that VH and VC are the same en-tity.18-22 Supporting this perspective is the sugges-tion that anatomic location influences whether alesion resembles VH or VC.18 Alternatively, VHmay represent a precursor to VC or may be anentirely separate entity. The results herein suggestthat VH and VC are not the same lesion, as theyevoke opposite stromal reactions. Alterations in the

stromal phenotype not only provide helpful diag-nostic clues in epithelial lesions but also highlightthe importance of the stromal-epithelial relationship.In fact, recent evidence suggests that alterations ofthe stromal compartment alone are sufficient toinduce epithelial malignancy.23 During the transitionfrom benign to malignant, resident CD34þ dendriticcells in normal stroma acquire a cancer-activatedphenotype characterized by a-SMA expression.24

The acquisition of smooth muscle actin is a featureof myofibroblasts, which are prominent in the des-moplastic reaction to infiltrating solid tumors, and itmay be reasonable to attribute the immunostain-ing pattern for VC to a type of desmoplasia that isbelow the level of appreciation by routine micro-scopy. Supplementing routine histologic assessmentwith these 2 simple and widely available immunos-tains may contribute diagnostic information. Anotherdiagnostic tool, nuclear cytometry on Feulgen-stained histologic sections, has been reported asuseful for differentiating VC from benign lesions,25

Page 4: CD34 and α smooth muscle actin distinguish verrucous ...

Fig. 2. A-C, Hematoxylin-eosinestained sections demonstrating VC with its broad-based invasive pegs (originalmagnification � 10, � 50, and � 200, respectively). D-F, Immunohistochemical staining for CD34 demonstrates ameshwork of reactive cells in association with the adjacent uninvolved stroma (original magnification � 10, � 50,and � 200, respectively). Conversely, CD34 reactivity is limited to the vasculature in the tumor-associated stroma.G-I, Immunohistochemical staining for a-SMA demonstrating a network of slender, bipolar myofibroblasts beneath theinvading rete pegs (original magnification � 10, � 50, and � 200, respectively). (VC, verrucous carcinoma; a-SMA, asmooth muscle actin).

ORAL AND MAXILLOFACIAL PATHOLOGY OOOO

480 Paral, Taxy and Lingen April 2014

but to our knowledge, this method is not readilyavailable to most pathologists.

Since its initial description in 1985,26 proliferativeverrucous leukoplakia (PVL) has posed both diag-nostic and clinical management challenges. PVL is acondition of unknown etiology that is most oftenobserved in middle-aged women. Initially, the clin-ical presentation may be of a singular leukoplakiclesion that becomes multifocal over time. Impor-tantly, PVL has been reported to have a high rateof malignant transformation.26-29 Although severalcriteria for PVL have been proposed,30 the diagnosisof PVL is often made in a retrospective fashion manyyears later, after the lesion has progressed to SCC.Recently, both DNA aneuploidy and the expressionof several candidate protein biomarkers for the pre-diction of malignant changes in PVL have beeninvestigated for their utility in the diagnosis ofPVL.31-33 It would be interesting to test the hypoth-esis that the stromal CD34 and a-SMA expressionpatterns might aid in both distinguishing between VH

and PVL in challenging cases and predicting whichPVL cases are at most risk for undergoing malignanttransformation.

In conclusion, given that the mainstay of therapyfor either of these lesions is complete surgicalexcision, the distinction between the 2 entities maybe more of a histopathologic endeavor rather than aclinical one. For the pathologist, the significance ofdistinguishing these 2 entities rests on the findingthat 20% of VC cases have foci of conventionalSCC, for which the term hybrid tumor is used,34,35

because these hybrid tumors are treated moreaggressively.35 The biologic relationship betweenVC and VH has been, and continues to be, a sourceof a diagnostic dilemma. This study presents a po-tential additional step in discriminating betweenthem.

The authors thank Theodore Karrison, PhD, for his statisticalexpertise and Thomas Krausz, MD, FRCPath, for hisinspiration.

Page 5: CD34 and α smooth muscle actin distinguish verrucous ...

Fig. 3. A-C, Hematoxylin-eosinestained sections demonstrating VH that is characterized by hyperplastic, undulating epitheliumconfined to the level of the adjacent uninvolved epithelium (original magnification � 10, � 50, and � 200, respectively). D-F, Immu-nohistochemical staining for CD34 demonstrates a dendritic meshwork beneath the lesion that is continuous with the uninvolved stromaand consists of triangular cell bodies with delicate, interdigitating cytoplasmic extensions (original magnification� 10,� 50, and� 200,respectively). G-I, Immunohistochemical staining for a-SMA demonstrating negative stromal staining with positive staining of thevasculature (original magnification� 10,� 50, and� 200, respectively). (VH, verrucous hyperplasia; a-SMA, a smooth muscle actin.)

Table II. Stromal patterns of CD34þ dendritic cells

Pattern � þ þþTumor-free 0 3 13Verrucous hyperplasia 0 4 9Verrucous carcinoma 12 3 0Squamous cell carcinoma 18 2 0

Fig. 4. Immunohistochemical staining for the CD34þ den-dritic meshwork demonstrating focal disruption by inflam-matory cell infiltrates (asterisk; original magnification � 200).

OOOO ORIGINAL ARTICLE

Volume 117, Number 4 Paral, Taxy and Lingen 481

REFERENCES1. Batsakis JG, Hybels R, Crissman JD, Rice DH. The pathology of

head and neck tumors: verrucous carcinoma. Head Neck Surg.1982;5:29-38.

2. Shear M, Pindborg JJ. Verrucous hyperplasia of the oral mucosa.Cancer. 1980;46:1855-1862.

3. Kacar A, Arikok AT, Kokenek Unal TD, et al. Stromal expressionof CD34, a-smooth muscle actin and CD26/DPPIV in squamouscell carcinoma of the skin: a comparative immunohistochemicalstudy. Pathol Oncol Res. 2012;18:25-31.

4. Barth PJ, Ramaswamy A, Moll R. CD34(þ) fibrocytes in normalcervical stroma, cervical intraepithelial neoplasia III, and invasivesquamous cell carcinoma of the cervix uteri. Virchows Arch.2002;441:564-568.

5. Lindenmayer AE, Miettinen M. Immunophenotypic features ofuterine stromal cells. CD34 expression in endocervical stroma.Virchows Arch. 1995;426:457-460.

6. Tsuzuki S, Ota H, Hayama M, Sugiyama A, Akamatsu T,Kawasaki S. Proliferation of a-smooth muscle actin-containingstromal cells (myofibroblasts) in the lamina propria subjacent tointraepithelial carcinoma of the esophagus. Scand J Gastro-enterol. 2001;36:86-91.

7. Barth PJ, Schenck zu Schweinsberg T, Ramaswamy A, Moll R.CD34þ fibrocytes, alpha-smooth muscle antigen-positive myofi-broblasts, and CD117 expression in the stroma of invasivesquamous cell carcinomas of the oral cavity, pharynx, and larynx.Virchows Arch. 2004;444:231-234.

Page 6: CD34 and α smooth muscle actin distinguish verrucous ...

ORAL AND MAXILLOFACIAL PATHOLOGY OOOO

482 Paral, Taxy and Lingen April 2014

8. Chaudhary M, Gadbail AR, Vidhale G, et al. Comparison ofmyofibroblasts expression in oral squamous cell carcinoma, ver-rucous carcinoma, high risk epithelial dysplasia, low risk epithe-lial dysplasia and normal oral mucosa. Head Neck Pathol. 2012;6:305-313.

9. Chauhan H, Abraham A, Philips JR, Pringle JH, Walker RA,Jones JL. There is more than one kind of myofibroblast: analysisof CD34 expression in benign, in situ, and invasive breast lesions.J Clin Pathol. 2003;56:271-276.

10. Barth PJ, Ebrahimsade S, Ramaswamy A, Moll R. CD34þ

fibrocytes in invasive ductal carcinoma, ductal carcinoma in situ,and benign lesions. Virchows Arch. 2002;440:298-303.

11. Ramaswamy A, Moll R, Barth PJ. CD34þ fibrocytes in tubularcarcinomas and radical scars of the breast. Virchows Arch.2003;443:536-540.

12. Barth PJ, Ebrahimsade S, Hellinger A, Moll R, Ramaswamy A.CD34þ fibrocytes in neoplastic and inflammatory pancreatic le-sions. Virchows Arch. 2002;440:128-133.

13. Nakayama H, Enzan H, Miyazaki E, Kuroda N, Naruse K,Hiroi M. Differential expression of CD34 in normal colorectaltissue, peritumoral inflammatory tissue, and tumour stroma.J Clin Pathol. 2000;53:626-629.

14. Gnepp DR, ed. Diagnostic Surgical Pathology of the Head andNeck. 2nd ed. Philadelphia, PA: Saunders; 2009:52-54.

15. Erdag G, Qureshi HS, Patterson JW, Wick MR. CD34-positivedendritic cells disappear from scars but are increased in peri-cicatricial tissue. J Cutan Pathol. 2008;35:752-756.

16. Barth PJ, Moll R, Ramaswamy A. Stromal remodeling andSPARC (secreted protein acid rich in cysteine) expression ininvasive ductal carcinomas of the breast. Virchows Arch.2005;446:532-536.

17. Betz P, Nerlich A, Wilske J, Tübel J, Penning R, Eisenmenger W.Time-dependent appearance of myofibroblasts in granulationtissue of human skin wounds. Int J Leg Med. 1992;105:99-103.

18. Suarez P, Batsakis JG, el-Naggar AK. Leukoplakia: still a galli-maufry or is progress being made? A review. Adv Anat Pathol.1998;5:137-155.

19. Luna MA, Tortoledo ME. Verrucous Carcinoma. In: Gnepp DR,ed. Pathology of the Head and Neck. New York, NY: ChurchillLivingstone; 1988:497-515.

20. Slootweg PJ, Müller H. Verrucous hyperplasia or verrucouscarcinoma: an analysis of 27 patients. J Oral Maxillofac Surg.1983;11:13-19.

21. Arendorf TM, Aldred MJ. Verrucous carcinoma and verrucoushyperplasia. J Dent Assoc South Africa. 1982;37:529-532.

22. Murrah VA, Batsakis JG. Proliferative verrucous leukoplakiaand verrucous hyperplasia. Ann Otol Rhinol Laryngol. 1994;103:660-663.

23. Hu B, Castillo E, Harewood L, et al. Multifocal epithelial tumorsand field cancerization from loss of mesenchymal CSL signaling.Cell. 2012;149:1207-1220.

24. Kalluri R, Zeisberg M. Fibroblasts in cancer. Nat Rev Cancer.2006:392-401.

25. Ferlito A, Antonutto G, Silvestri F. Histological appearancesand nuclear DNA content of verrucous squamous cell carci-noma of the larynx. J Otorhinolaryngol Relat Spec. 1976;38:65-85.

26. Hansen LS, Olson JA, Silverman S. Proliferative verrucous leu-koplakia: a long-term study of thirty patients. Oral Surg Oral MedOral Pathol. 1985;60:285-298.

27. Silverman S, Gorsky M. Proliferative verrucous leukoplakia: afollow-up study of 54 cases. Oral Surg Oral Med Oral PatholOral Radiol Endod. 1997;84:154-157.

28. Fettig A, Pogrel MA, Silverman S, Bramanti TE, Da Costa M,Regezi JA. Proliferative verrucous leukoplakia of the gingiva.Oral Med Oral Pathol Oral Radiol Endod. 2000;90:723-730.

29. Bagan JV, Jimenez-Soriano Y, Diaz-Fernandez JM, et al. Ma-lignant transformation of proliferative verrucous leukoplakia tooral squamous cell carcinoma: a series of 55 cases. Oral Oncol.2011;47:732-735.

30. Cerero-Lapiedra R, Balade-Martinez D, Moreno-Lopez LA,Esparza-Gomez G, Bagan JV. Proliferative verrucous leukopla-kia: a proposal for diagnostic criteria. Med Oral Patol Oral CirBuccal. 2010;15:e839-e845.

31. Klanrit P, Sperandio M, Brown AL, et al. DNA ploidy inproliferative verrucous leukoplakia. Oral Oncol. 2007;43:310-316.

32. Gouvea AF, Vargus PA, Coletta RD, Jorge J, Lopes MA. Clini-copathologic features and immunohistochemical expression ofp53, Ki-67, and Mcm-5 in proliferative verrucous leukoplakia.J Oral Path Med. 2010;39:447-452.

33. Gouvea AF, Santos Silva AR, Speight PM, et al. High incidenceof DNA ploidy abnormalities and increased Mcm2 expressionmay predict malignant change in oral proliferative verrucousleukoplakia. Histopathology. 2013;62:551-562.

34. Medina JE, Dichtel W, Luna MA. Verrucous-squamous carci-nomas of the oral cavity: a clinicopathologic study of 104 cases.Arch Otolaryngol. 1984;110:437-440.

35. Orvidas LJ, Olsen KD, Lewis JE, Suman VJ. Verrucous carci-noma of the larynx. Head Neck. 1998;20:197-203.

Reprint requests:

Mark W. LingenDepartment of PathologyUniversity of Chicago5841 S Maryland AveMC 6101Chicago, IL [email protected]