spindle cell breast paid articl

7/28/2019 Spindle Cell Breast Paid Articl

1/7

An Immunohistochemical Study ofMetaplastic Spindle Cell Carcinoma,Phyllodes Tumor and Fibromatosisof the Breast

BARBARA DUNNE, MRCPATH, ANDREW H. S. LEE, MRCP, MRCPATH,SARAH E. PINDER, FRCPATH, JANE A. BELL, FIBMS,AND IAN O. ELLIS, FRCPATH

The diagnosis of metaplastic (sarcomatoid) carcinoma (MSC) ofbreast often requires immunohistochemistry with a cytokeratin (CK)panel to distinguish them from phyllodes tumors (PT), primary sar-comas, and fibromatoses. CK staining may be heterogeneous in meta-plastic carcinomas. The aim of the study was to investigate the theorythat MSCs show evidence of myoepithelial differentiation and toevaluate immunohistochemical markers that may be helpful in distin-guishing MSCs from PT and fibromatosis. We reviewed histology and

performed immunohistochemistry for AE1/AE3, 34E12, CK5 andCK14, Cam5.2, CK7 and CK19, epithelial membrane antigen (EMA)(B55), smooth muscle actin (SMA), S100, desmin, vimentin, CD31,CD34, and bcl-2 on paraffin-embedded tissue from 18 MSCs, 26 PTs,and 8 fibromatoses. We assessed staining by using a semiquantitativemethod. Sarcomatous areas in MSCs were positive for 34E12 in 11cases; for SMA in 10; for CK5 in 7; for CK14 in 6; for Cam5.2,

AE1/AE3, and S100 in 5; and for CK7 and CK19 in 3. No CKexpression was seen in stromal areas in PT or in fibromatoses. CD34and bcl-2 were more frequently expressed in spindle cell areas in PTs(18 and 12 of 26, respectively) than in MSCs (0 and 2 of 18, respec-tively). MSCs show strong evidence of myoepithelial differentiation.CD34 and, to a lesser extent, bcl-2 positivity in PTs may be helpful indifferentiating these two lesions from MSCs, particularly in smallbiopsies, because CK staining in MSCs may be heterogeneous. In our

hands, 34E12 was the CK most frequently expressed in sarcomatoidareas in MSCs. HUM PATHOL 34:1009-1015. 2003 Elsevier Inc. Allrights reserved.

Key words: breast, metaplastic carcinoma, myoepithelial pheno-type, phyllodes tumor, CD34 positivity.

Abbreviations: MSC, metaplastic sarcomatoid carcinoma; PT,phyllodes tumor; CK, cytokeratin.

Breast tumors with a sarcomatous growth patternare relatively rare and can present a diagnostic chal-lenge to the histopathologist. Monophasic spindle celltumors include fibromatosis, spindle cell myoepitheli-oma, myofibroblastoma, and inflammatory pseudotu-mor, which are benign, and metaplastic (sarcomatoid)

carcinoma (MSC) and the very rare primary breastsarcoma at the malignant end of the spectrum. Tumorswith a biphasic pattern composed of spindle and epi-thelial areas are also a heterogeneous group and en-compass fibroadenomas, PTs, and biphasic MSCs.MSCs, by definition, contain malignant sarcomatoid,usually spindle cell, areas thatshow, at least in part, anepithelial immunophenotype.1-4

MSCs can be biphasic and contain discrete epithe-lial areas in addition to sarcomatoid areas, or less com-monly, they can be composed of sarcomatoid areasalone. The sarcomatoid areas, in turn, can be purespindle cell or malignant fibrous histiocytoma (MFH)

like or show heterologous mesenchymal differentia-tion, most commonly cartilaginous or osseous in type.5

The theory that MSCs are derived from myoepithelialcells rather than epithelial cells or, indeed, from a

common stem cell that is capable of dual differentia-tion has been debated in the literature, the earliestsuch proposal coming from Hamperl.6

Myoepithelial neoplasms of the breast are ex-tremely uncommon, and their description has beenlimited to case reports. Adenomyoepitheliomas have

been divided into 3 subtypes based on their growthpattern:7,8 spindle cell, which resembles leiomyoma butcontains occasional glandular spaces, and tubular andlobular variants. They have been shown to behave aslow-grade malignant neoplasms,7-10 having the poten-tial to recur locally if not completely excised and, veryrarely, to metastasize.10 The entity of malignant myoep-ithelioma (myoepithelial carcinoma) in its pure form isreportedly composed of spindle cells that show evi-dence of myoepithelial differentiation by immunohis-tochemistry or electron microscopy.11,12 It is recog-nized that there is an overlap between this entity andmonophasic MSC, as both tumors may express S100,

actin, and cytokeratins (CKs).1,3,13-15

PTs, in contrast,are biphasic tumors that do not show an epithelial or amyoepithelial phenotype in the stroma. They are char-acterized by their leaflike architecture, cleftlike spaceslined by epithelium, and hypercellular stroma.16-18 Fi-bromatosis of breast has a similar histological appear-ance to the same lesion at other sites and is composedof a variably cellular proliferation of uniform, blandspindle cells arranged in fascicles with intervening hya-linized areas.19,20 It is important as a differential diag-nosis for monophasic MSC, but absence of an epithelialcomponent after careful sampling and CK negativityare useful differentiating features.

From the Department of Pathology, City Hospital, Nottingham,United Kingdom. Accepted for publication May 1, 2003.

Address correspondence and reprint requests to Barbara Dunne,MRCPath, Department of Pathology, Glasgow Royal Infirmary, Glas-gow G4 OSF, United Kingdom.

2003 Elsevier Inc. All rights reserved.0046-8177/03/3410-0007$30.00/0doi:10.1053/S0046-8177(03)00414-3

1009


2/7

The aim of this study was to perform a detailedimmunohistochemical study on spindle cell lesions ofbreast common to our practice to accomplish the fol-lowing: (1) investigate the theory that MSCs show phe-notypic evidence of myoepithelial differentiation and(2) show markers that may be helpful in differentiatingthese tumors, one from another. We used a range of

CKs, including broad-spectrum CKs (AE1/AE3) andCKs known specifically to be expressed in both basal(CK5, CK14, 34E12)21 and luminal (CK7,22 CK19,23

Cam5.2) epithelia. We included S100 and smooth mus-cle actin (SMA), which are also expressed in myoepi-thelial cells, and CD34 and bcl-2, which have beenshown to be positive in the stroma of phyllodes tumors(PT).24-26 We completed the stromal panel withdesmin, vimentin, and CD31.

MATERIALS AND METHODS

We collected 18 MSCs (16 biphasic, 2 uniphasic), 26 PTs

(18 benign, 5 borderline, 3 malignant), and 8 fibromatosesfrom the files of Nottingham City Hospital. All tumors hadbeen resected at this hospital between 1987 and 1999. Wereviewed the hematoxylin and eosin (H&E) slides. The PTs

were graded according to the criteria of Moffat et al.27

We performed immunohistochemistry on one represen-tative block from each case for Cam5.2 (Becton Dickinson,Franklin Lakes, NJ, 1:2), AE1/AE3 (Dako, Glostrup, Den-mark, 1:50), CK5 (The Binding Site, Birmingham, UK,1:100), CK14 (Novocastra, Newcastle Upon Tyne, UK, 1:100),34E12 (Dako, 1:80), CK7 (Dako, 1:50), CK19 (Dako, 1:100),epithelial membrane antigen (EMA [B55 antibody]; Notting-ham University, Nottingham, UK, 1:4000), SMA (Sigma, St.Louis, MO, 1:200), S100 (Dako, 1:1000), desmin (Dako,

1:50), vimentin (Dako, 1:250), CD31 (Dako, 1:100), CD34(Novocastra, 1:50), and bcl-2 (Dako, 1:100). An automaticimmunohistochemical stainer (Dako TechMate with Chem-Mate Detection Kit) was used, with microwave retrieval for allantibodies except S100, EMA, SMA, and desmin. We exam-ined the entire lesional area for each stain and assessedstaining semiquantitatively, where 1 staining represented10% positive cells; 2, 10% and 50% positive cells; 3,50% and 90% positive cells; and 4, 90% positive cells.

We considered 2 staining and above to represent positivity.Fishers exact test (SPSS for Windows, version 11) was used totest the significance of differences in staining between MSCsand PTs. This project involved the use of routinely stored,archival, paraffin-embedded tissue. All patients consented toremoval of their breast tumors for treatment purposes alone.The type and extent of surgery were not modified in any wayfor the benefit of subsequent research projects. Patient clin-ical notes were not consulted.

RESULTS

Age and Sex of Patients

All patients were female. The age range for theMSC patients was 27-82 years, and the mean, 57 years(standard deviation [SD], 16.3); for the PT patients,it was 20-92 years, and the mean, 52 years ( SD 12.9);and for the fibromatoses, it was 20-64 years, and themean, 44 years ( SD 19.4).

Macroscopy and Microscopic Review

Gross pathology was detailed from the pathologyreport. The macroscopic size for the MSCs ranged from10 to 80 mm in maximum extent. Where circumscrip-tion was documented, 9 were described as well defined,3 as moderately well defined, and 2 as poorly defined.

Two tumors had friable/necrotic areas. The size rangefor the PTs was between 17 and 180 mm in maximumextent. Most of the tumors were described as well-defined, lobulated, firm, cream and white lesions. Inthe case of the fibromatoses, 10-35 mm was the sizerange, and the lesions were typically ill-defined andfibrous in nature.

Table 1 summarizes the histological features of theMSCs. See also Figs 1 through 4. The fibromatoses had

TABLE 1. Histological Features of MetaplasticCarcinomas

Caseno Metaplastic components

Conventionalcarcinoma

Invasive In situ

1 Spindle

2 Spindle 3 Spindle small squamoid foci

4Spindle (arising in sclerosing lesion)

small cohesive foci 5 Spindle DCIS6 Spindle small cohesive foci 7 Spindle small cohesive foci 8 Spindle small squamoid foci NST DCIS9 Spindle small cohesive foci DCIS

10 Spindle small squamoid foci NST 11 Cartilage Lobular 12 Cartilage NST 13 Cartilage NST 14 Cartilage NST

15 SpindleNST

papillary DCIS16 Spindle cartilage osteoid NST DCIS17 Spindle cartilage NST DCIS18 Spindle cartilage NST DCIS

Abbreviations: DCIS, ductal carcinoma in situ; NST, no specialtype.

FIGURE 1. Biphasic metaplastic spindle cell carcinoma withsquamous cell differentiation (Case 3, see Table 1; H&E, mag-nification 200).

HUMAN PATHOLOGY Volume 34, No. 10 (October 2003)

1010
http://-/?-http://-/?-


3/7

typical appearances and showed broad fascicles of spin-dle cells with variable cellularity, sometimes interveninghyalinized areas and infiltrative edges.

Results of Immunohistochemical Staining

MSC: Spindle and Heterologous Staining

The differential staining pattern of CKs, EMA,S100, and SMA in spindle cell and heterologous areasof the MSCs is summarized in Table 2. Overall, agreater number of cases stained positively for the myo-epithelial markers 34E12 (11 cases), SMA (10 cases),CK5 (7 cases), and CK14 (6 cases) (see Fig 5) than for

the broad-spectrum CK AE1/AE3 (5 cases) and theluminal CKs cam5.2 (5 cases), CK7 (3 cases), and CK19(3 cases). Five cases stained positively for S100. In themajority of cases, the spindle areas stained positively forboth myoepithelial and luminal CKs, and in some casesthere appeared to be positive staining in the same cellsfor the 2 different subtypes of keratin. The staining waspatchy in most cases. The most diffuse staining was seenwith AE1/AE3, for which all positive cases showed pos-itive staining in 50% of cells. In the 7 cases in whichchondrosarcoma-like areas were present, SMA was pos-itive in 2 cases and S100 in 5, with diffuse4 staining in2 of these. Of the CKs, 34E12 was positive in cartilag-

inous areas in 5 cases (see Fig 6) and was positive forCK5, CK14, AE1/AE3, and cam5.2 in 1 case each; stain-ing for CK7 and CK19 was negative. There was nopositivity when using any of the antibodies in the os-teoid present in 1 case. EMA was negative in all cases.

A comparison of mesenchymal staining in MSCs,PTs, and fibromatoses is summarized in Table 3. Noneof the MSCs showed CD34 expression in the spindlecell component. There was positive staining for bcl-2 in2 MSCs; 1 of these was in a chondroid area, and 1 wasin a spindle cell area. CD31 was positive in pleomorphicspindle cells in 1 MSC. There was desmin positivity in 3MSCs, 2 in cartilage and 1 in spindle cells. Vimentin was

strongly positive in the sarcomatous areas in all but 1MSC.

MSC: Epithelial/Myoepithelial Staining

Epithelial and myoepithelial marker expression innonsarcomatous areas of MSCs is summarized in Table4. The epithelial areas in the MSCs showed mixedexpression for both basal and luminal CKs and EMA.Fourteen cases expressed 34E12; 13 expressed AE1/AE3, Cam5.2, and CK7; 12 expressed EMA; 11 ex-pressed CK5 and CK19; 10 expressed CK14; 7 expressedS100; and 4 showed SMA positivity. There was positivestaining for bcl-2 in the epithelial component of 7MSCs; two of these showed 4 staining, but there wasno positivity for CD34, CD31, desmin, and vimentin.

PTs: Stromal Staining

See Table 3 and Table 5. There was CD34 expres-sion in the stroma of 18 of 26 PTs. Positive staining wasseen in 1 (2 staining) of 3 malignant, in 3 (4 stain-ing in 1, 2 staining in 2) of 5 borderline, and in 14benign tumors (4 in 3, 3 in 2, and 2 in 9; see Fig

FIGURE 3. Uniphasic spindle cell carcinoma invading fat(Case 2, see Table 1; H&E, magnification 200).

FIGURE 4. Biphasic metaplastic carcinoma showing heterol-ogous cartilaginous differentiation (Case 17, see Table 1; H&E,magnification 200).

FIGURE 2. Biphasic metaplastic spindle cell carcinoma withwell-demarcated spindle and epithelial components (Case15, see Table 1; H&E, magnification 200).

IMMUNOHISTOCHEMICAL DIFFERENTIATION IN BREAST (Dunne et al)

1011


4/7

7). Bcl-2 was positive in the stroma of 12 PTs, 2 of 3malignant, 3 of 5 borderline, and 7 of 18 benign tumors(see Fig 8). There was 3 staining in the malignanttumor and in 2 of the benign lesions, but the remain-der of the staining was in 50% of cells. Seventeen PTsshowed stromal positivity for SMA, 2 of 3 malignant, 3of 5 borderline, and 11 of 18 benign lesions. Three PTs(1 malignant, 1 borderline, and 1 benign) showed stro-mal positivity for S100. CD31 staining was negative.

There was desmin positivity in 3 tumors (1 malignant, 1borderline, and 1 benign). Vimentin was diffusely pos-itive in the stromal component of all lesions and CK,and EMA staining was negative.

PTs: Epithelial Staining

All cases showed4 positivity for AE1/AE3, Cam5.2,CK7, and EMA, where in the case of each antibody, thepositive cells lined the luminal surfaces of the clefts orglandular spaces. 34E12 stained positively in all but onecase, and S100, positively in all cases, with a mixed basaland luminal pattern of positivity. CK5, CK14, and SMAshowed strong positivity in the epithelial areas of all cases,but this time, the pattern of staining was predominantly

basal, where the positive cells were those lying adjacent tothe basement membrane and there were often overlyingnegative luminal cells. There was positive staining in theepithelial areas for bcl-2 in 24 of 26 PTs; this was predom-

TABLE 2. Epithelial and Myoepithelial Marker Expression in Spindle and Heterologous Elements in MetaplasticSpindle Cell Carcinomas

Case number SMA S100 34E12 CK5 CK14 AE1/AE3 Cam5.2 CK7 CK19 EMA

1 S2 S0 S1 S1 S1 S1 S0 S1 S0 S02 S3 S1 S3 S4 S3 S3 S1 S0 S0 S03 S0 S0 S3 S2 S2 S4 S4 S2 S2 S04 S3 S0 S0 S1 S0 S0 S1 S1 S0 S0

5 S1 S0 S2 S4 S4 S3 S3 S0 S0 S06 S4 S0 S2 S4 S3 S3 S1 S2 S0 S07 S2 S0 S2 S4 S4 S0 S2 S0 S0 S08 S0 S0 S2 S3 S1 S0 S4 S2 S2 S09 S2 S1 S0 S0 S0 S0 S0 S0 S0 S0

10 S0 S0 S1 S0 S0 S0 S0 S1 S4 S011 C3 C4 C4 C0 C0 C0 C0 C0 C0 C012 C0 C4 C4 C2 C2 C3 C3 C1 C1 C013 C0 C2 C2 C0 C1 C1 C1 C0 C0 C014 C2 C1 C1 C0 C1 C1 C0 C0 C0 C015 S3 S1 S1 S0 S0 S0 S0 S1 S0 S016 S3, C1, B0 S2, C2, B0 S2, C2, B0 S0, C0, B0 S0, C0, B0 S0, C0, B0 S0, C0, B0 S0, C0, B0 S0, C0, B0 S0, C0, B017 S1, C0 S1, C0 S1, C0 S0, C0 S0, C0 S0, C0 S0, C0 S0, C0 S0, C0 S0, C018 S1, C0 S2, C2 S2, C2 S0, C0 S0, C0 S0, C0 S0, C0 S0, C0 S0, C0 S0, C0

Total positive 2 10 5 11 7 6 5 5 3 3 0

1 staining 10% positive cells, 2 10% and 50% positive cells, 3 50% and 90% positive cells, 4 90% positive cells.Abbreviations: S, spindle cell areas; C, cartilaginous areas; B, osteoid areas.

FIGURE 5. Uniphasic spindle cell carcinoma showing strongexpression for CK14 by immunohistochemistry (Case 2, seeTable 1; H&E, magnification 100).

FIGURE 6. Immunohistochemical staining for 34E12 in heter-ologous cartilage in a metaplastic carcinoma (Case 17, seeTable 1; H&E, magnification 400).


1012


5/7

inantly luminal in distribution. There was no positivity inthe epithelial areas of PTs for CD34, CD31, desmin, orvimentin.

Fibromatosis

See Table 3. Five of 8 fibromatoses showed positiv-

ity for SMA, 1 showed 4 positivity, and 4 showed 2positivity. Vimentin was diffusely positive in 7 of 8 cases.There was negative staining in all cases for S100, bcl-2,CD31, all CKs, and EMA. Staining for CD34 was alsointerpreted as negative, although some of the cases hadpositive staining at the interface between the lesion andnormal breast tissue. One case showed 2 positivity fordesmin.

DISCUSSION

We reviewed all of the spindle cell tumors of breastthat were recorded in our practice over a 12-year pe-

riod. We carried out a detailed immunohistochemical

analysis of the tumors by using antibodies against arange of CKs and mesenchymal markers. Our aim wasto try to better understand the histogenesis of thesetumors and to identify potential antibodies that mightaid the histopathologist in his or her diagnosis.

The mesenchymal elements of the MSCs showedheterogeneous immunohistochemical staining for bothepithelial and myoepithelial markers, and there wascoexpression of both luminal and basal CKs in some ofthe cases. Overall, the pattern of immunoreactivity fa-vored a myoepithelial phenotype, with 34E12 and CKs5 and 14 staining positively in 11 (61%), 7 (39%), and6 (33%) of cases, respectively, and SMA and S100 stain-ing positively in 10 (55%) and 5 (28%) of cases. AE1/AE3 (5 cases, 28%) and luminal CKs Cam5.2 (5 cases,28%), CK7 (3 cases, 16%), and CK19 (3 cases, 16%)were also positive in sarcomalike areas in some of thecases. Thus, in our hands, 34E12 is positive in mesen-chymal areas in the greatest number of MSCs. Thisfinding concurs with other reports in which this anti-body has been the most sensitive of all the CKs inmetaplastic carcinomas.5,28 Overall, however, the stain-ing for CKs was variable in distribution, which high-lights the need to use a panel of CK antibodies in theworkup of a potential MSC. Epithelial areas in thebiphasic MSCs showed strong diffuse staining for all ofthe CKs, both basal and luminal types and for S100.SMA was positive in epithelial areas in 4 (25%) cases.

Other investigators have also reported evidence ofa myoepithelial phenotype in MSCs by using immuno-histochemical methods, either alone1,3 or in combina-tion with ultrastructural examination.2,29,30 Contrastingreports state that the presence of similar tumors inother organs in which myoepithelial cells are absent

makes this an unlikely cell of origin.13 Although one

TABLE 3. Comparison of Staining in Spindle CellAreas of Metaplastic Carcinomas, Phyllodes Tumors,

and Fibromatoses

Staining

Metaplasticcarcinoma(n 18)

Phyllodestumour

(n 26)

P value(Fishers exact

test)Fibromatosis

(n 8)

CD34 0 18 0.0001 0

Bcl-2 2 12 0.02 0SMA 10 17 0.54 5S100 10 3 0.002 0CD31 1 0 0.41 0Desmin 3 3 0.66 1Vimentin 17 26 0.41 7

TABLE 4. Epithelial and Myoepithelial Marker Expression in Epithelial Areasin Metaplastic Spindle Cell Carcinomas

Case number SMA S100 34E12 CK5 CK14 AE1/AE3 Cam5.2 CK7 CK19 EMA

1 2 3 0 0 2 4 3 4 4 2 2 04 0 4 4 2 4 4 4 4 4 45 4 3 3 4 4 4 4 4 3 46 0 2 4 4 4 4 4 3 0 27 4 0 2 4 4 4 4 4 1 48 0 2 4 4 4 4 2 2 3 39 0 1 4 4 4 4 4 4 4 4

10 0 0 2 1 1 4 4 4 4 411 4 0 3 4 4 0 1 2 0 212 0 2 4 3 4 4 4 4 2 413 2 1 1 0 1 1 1 0 0 014 0 2 0 0 1 1 0 1 0 015 0 0 2 0 0 4 4 0 4 116 0 2 2 1 1 4 4 4 4 417 0 1 4 2 1 4 4 4 4 418 0 1 4 2 4 4 4 4 4 4

Total positive 2 4 7 14 11 10 13 13 13 11 12

NOTE.1 staining10% positive cells,210% and50% positive cells,350% and90% positive cells,490% positivecells.


1013


6/7

recent report has shown positive staining in MSCs forthe novel putative myoepithelial markers maspin, p-cadherin, and p63,31 immunohistochemical evidenceof myoepithelial differentiation previously has beenlimited to showing actin, S100, and nonspecific CK insarcomatoid areas. In this series, as well as actin andS100, we have shown frequent expression of CKs 5 and14, which are specifically expressed in myoepithelial

cells in both sarcomatoid and epithelial elements ofbiphasic MSCs, lending further and stronger evidencein favor of myoepithelial differentiation. It is most likelythat the sarcomatous areas of these tumors developalong a particular pathway of myoepithelial differenti-ation rather than that they are truly derived histogenet-ically from myoepithelium. This theory is in keepingwith our finding of strong diffuse staining for bothbasal and luminal CKs in epithelial areas of MSCs,coupled with differential expression of more basal-spe-cific proteins in mesenchymal areas in many of thetumors. Kaufman et al32 have suggested that the myo-epithelial phenotype may represent a transition be-tween the transformation from epithelial to sarcoma-tous lines of differentiation. Another study, usingmicrosatellite analysis, showed a single clonal origin forcarcinomatous and sarcomatoid elements in a case ofbiphasic MSC by showing similar loss of heterozygositypatterns in both areas.33

Because of the rarity of MSCs and thus the paucityof large series with long-term follow-up and because ofthe lack of uniformity of grading between series, it hasbeen difficult to fully elucidate the prognosis of theselesions. Some reports have suggested an unfavorableoutcome for the group as a whole,34 whereas othershave shown evidence of a variable prognosis dependingon the histological pattern.2,30,35 Clearly, much moreneeds to be learned about the biology of these tumorsif we are to predict survival and response to treatment.

The second importantfinding in this study was thesignificantly greater frequency of expression of CD34(P 0.0001) and bcl-2 (P 0.02) in mesenchymalareas of PTs compared with in MSCs. Previous reports

have similarly described expression of CD3424,25 andbcl-224,26 in the stroma of fibroadenomas and PTs.Combining the results of the current series with thoseof a previous series24 and with those from other recentcases in our department, we have seen CD34 expressionin 58 of 66 PTs. Of these, 38 of 40 benign tumors werepositive, 8 of 9 borderline tumors showed positivity, and12 of 15 malignant PTs were positive. S100 was found tobe expressed in a significantly higher number of MSCsthan PTs (P 0.002), and SMA was expressed in aroughly similar proportion of PTs and MSCs (P 0.54).

The differential diagnosis between biphasic MSC

and PT is clear-cut in some cases. On H&E staining, aleaflike architecture and bland epithelium lining cleft-like spaces points to the latter diagnosis, whereas ma-lignant epithelial elements and heterologous differen-tiation, if present, are more likely to be present in theformer. In the preoperative setting, in which only asmall quantity of tissue in a core biopsy may be availablefor assessment, positive stromal CK staining (or its ab-sence) can often be the deciding factor in differentiat-ing these two lesions. Because axillary surgery is moreoften undertaken in MSC than in PT, accurate preop-erative diagnosis is important. As we have shown, CKstaining is rarely diffuse in MSC, and heterologous

TABLE 5. Staining in Stromal Areas of PhyllodesTumors in Relation to Grade of Tumor

StainingBenign

(n 18)Borderline

(n 5)Malignant

(n 3)

CD34 14 3 1Bcl-2 7 3 2SMA 11 3 2

S100 1 1 1Vimentin 18 5 3Desmin 1 1 1CD31 0 0 0

FIGURE 7. Strong positive immunohistochemical staining forCD34 in a benign phyllodes tumor (H&E, magnification 100).

FIGURE 8. Positive immunohistochemical staining for bcl-2 ina benign phyllodes tumor (H&E, magnification 200).


1014


7/7

staining can lead to false negatives. We propose that theaddition of CD34, bcl-2, and S100 antibodies to a CKpanel would be of benefit in the workup of spindle celllesions of breast, CD34 and bcl-2 positivity would favorPT, and S100 positivity would favor MSC.

The fibromatoses showed positivity for vimentin in7 cases, for SMA in 5, and for desmin in 1 case. Stainingwith S100 and bcl-2 was negative. Unsurprisingly, no

immunoreactivity was seen with CKs or EMA, a featurethat differentiates these lesions from bland uniphasicMSCs. We saw no true lesional positivity for CD34 infibromatoses, although some of the cases showed posi-tivity at the edge of the lesion in the stromal interfacebetween the lesion and adjacent normal breast.

In summary, in our hands, MSCs of breast showedstrong immunohistochemical evidence of myoepithe-lial differentiation, often expressing basal-specific CKs.They also showed that they can coexpress specific lu-minal CKs, suggesting an origin from a stem cell that iscapable of dual differentiation. An immunohistochem-ical panel composed of S100, SMA, and at least 4 CKs,

including basal-specific 34E12 and CK14 or CK5, isvaluable in the assessment of a suspected MSC. CD34and, to a lesser extent, bcl-2 positivity in the stroma ofPTs may be useful in differentiating them from MSC ofbreast, particularly in the preoperative setting. Ourfindings are useful to diagnostic histopathologists intheir attempt to accurately categorize these lesions.

REFERENCES

1. Meis JM, Nelson Gordonez NG, Gallagher HS: Sarcomatoidcarcinoma of the breast: An immunohistochemical study of six cases.Virchows Arch A 410:415-421, 1987

2. Wargotz ES, Deos OH, Norris HJ: Metaplastic carcinomas of

the breast. II. Spindle cell carcinoma. Hum Pathol 20:732-740, 19893. Eusebi V, Cattani MG, Ceccarelli C, et al: Sarcomatoid carci-

nomas of the breast: An immunohistochemical study of 14 cases. ProgSurg Pathol 10:83-89, 1989

4. Ellis IO, Bell J, Ronan JE, et al: Immunocytochemical inves-tigation of intermediate filament proteins and epithelial membraneantigen in spindle cell tumours of the breast. J Pathol 154:157-165,1988

5. Rosen PP: Rosens Breast Pathology, 2nd ed. Philadelphia,PA, Lippincott Williams &Wilkins, 2001

6. Hamperl H: The myothelia (myoepithelial cells). Normalstate; regressive changes; hyperplasia; tumors. Curr Top Pathol 53;161-220, 1970

7. Tavassoli FA: Myoepithelial lesions of the breast: Myoepithe-liosis, adenomyoepithelioma and myoepithelial carcinoma. Am JSurg Pathol 15:554-568, 1991

8. Rosen PP: Adenomyoepithelioma of the breast. Hum Pathol18:1232-1237, 19879. Kaier H, Nielsen B, Paulsen S, et al: Adenomyoepithelial

adenosis and low-grade malignant adenomyoepithelioma of thebreast. Virchows Arch A Pathol Anat 405:55-67, 1984

10. Loose JH, Patchefsky AS, Hoolandes IJ, et al: Adenomyoepi-thelioma of the breast. A spectrum of biologic behaviour. Am J SurgPathol 16:868-876, 1992

11. Schurch W, Potvin C, Seemayer TA: Malignant myoepitheli-

oma (myoepithelial carcinoma) of the breast: An ultrastructural andimmunohistochemical study. Ultrastruct Pathol 8:1-11, 1985

12. Thorner PS, Kahn HJ, Baumal R, et al: Malignant myoepi-thelioma of the breast. An immunohistochemical study by light andelectron microscopy. Cancer 57:745-750, 1986

13. Foschini MP, Dina RE, Eusebi V: Sarcomatoid neoplasms ofthe breast: Proposed definitions for biphasic and monophasic sarco-matoid mammary carcinomas. Semin Diagn Pathol 10:128-136, 1993

14. Lakhani SR, OHare MJ, Monaghan P, et al: Malignant myo-

epithelioma (myoepithelial carcinoma) of the breast: A detailed cy-tokeratin study. J Clin Pathol 48:164-167, 199515. Foschino MP, Eusebi V: Carcinomas of the breast showing

myoepithelial cell differentiation. A review of the literature. VirchowsArch 432:303-310, 1998

16. Treves N: A study of cystosarcom phyllodes. Ann N Y AcadSci 114:922-936, 1964

17. Pietruszka M, Barnes L: Cystosarcoma phyllodes. A clinico-pathological analysis of 42 cases. Cancer 41:-1983, 1978 1974

18. Ward RM, Evans HL: Cystosarcoma phyllodes. A clinicopath-ologic study of 26 cases. Cancer 58:2282-2289, 1986

19. Wargotz ES, Norris HJ, Austin RM, et al: Fibromatosis of thebreast: A clinical and pathologic study of 28 cases. Am J Surg Pathol11:38-47, 1987

20. Bogomoletz WV, Boulenger E, Simatos A: Infiltrating fibro-matosis of the breast. J Clin Pathol 34:30-34, 1981

21. Purkis PE, Steel JB, Mackenzie IC, et al: Antibody markers ofbasal cells in complex epithelia. J Cell Sci 97:39-50s, 199022. Van Niekerk CC, Jap PHK, Ramaekers FCS, et al: Immuno-

histochemical demonstration of keratin 7 in routinely fixed paraffinembedded human tissues. J Pathol 165:145-152, 1991

23. Bartek J, Durban EM, Hallowes RC, et al: A subclass ofluminal epithelial cells in the human mammary gland, defined byantibodies to cytokeratins. J Cell Sci 75:17-33, 1985

24. Moore T, Lee AHS: Expression of CD34 and bcl-2 in phyl-lodes tumours, fibroadenomas and spindle cell lesions of the breast.Histopathology 38:62-67, 2001

25. Silverman JS, Tamsen A: Mammaryfibroadenoma and somephyllodes tumour stroma are composed of CD34 fibroblasts andfactor XIIIa dendrophages. Histopathology 29:411-419, 1996

26. Keunen-Boumeester V, Henzen-Logmans SC, TimmermansMM, et al: Altered expression of p53 and its regulated proteins in

phyllodes tumours of the breast. J Pathol 189:169-175, 199927. Moffat CJC, Pinder SE, Dixon AR, et al: Phyllodes tumours ofthe breast: A clinicopathologic review of thirty-two cases. Histopathol-ogy 27:205-218, 1995

28. Chieng C, Cranor M, Lesser ME, et al: Metaplastic carci-noma of the breast with osteocartilagenous heterologous elements.Am J Surg Pathol 22:188-194, 1998

29. Wargotz ES, Norris HJ: Metaplastic carcinomas of the breast.I. Matrix-producing carcinoma. Hum Pathol 20:628-635s, 1989

30. Wargotz ES, Norris HJ: Metaplastic carcinoma of the breast.III. Carcinosarcoma. Cancer 64:1490-1499, 1989

31. Reis-Filho JS, Milanezi F, Paredes J, et al: Novel and classicmyoepithelial/stem cell markers in metaplastic carcinomas of thebreast. Appl Immunohistochem Mol Morphol 11:1-8, 2003

32. Kaufman MW, Marti JR, Gallager HS, et al. Carcinoma of thebreast with pseudosarcomatous metaplasia. Cancer 53:-1917, 19841908

33. Kung FYL, Tse GMK, Lo KW, et al: Metachronous bilateralmammary metaplastic and infiltrating duct carcinomas: A molecularstudy for clonality. Hum Pathol 33:677-679, 2002

34. Oberman HA: Metaplastic carcinoma of the breast. A clini-copathologic study of 29 patients. Am J Surg Pathol 11:918-929, 1987

35. Sneige N, Yaziji H, Mandavilli SR, et al: Low-grade (fibroma-tosis-like) spindle cell carcinoma of the breast. Am J Surg Pathol25:1009-1016, 2001


1015

spindle cell breast paid articl

Documents