- Email: [email protected]
Metaplastic carcinomas of the breast: V. Metaplastic carcinoma with osteoclastic giant cells
Metaplastic Carcinomas of the Breast: V. Metaplastic Carcinoma With Osteoclastic Giant Cells ERIC S. WARGOTZ, MD, AND HENRY J. NORRIS, MD The clinical and pathologic features of 29 examples of mammary metaplastic carcinoma with osteoclastic giant cells (OGC) in the stroma are reported. A bland spindle cell or sarcomatous component dominated these neoplasms, although infiltrating duct carcinoma was present in 23 cases, and intraductal carcinoma was present in six cases. In all 29 neoplasms, the carcinoma was admixed or contiguous with the stroma. Osteoclastic giant cells were admixed within the cellular stroma, and were intimately associated with prominent thin-walled vessels. Hemorrhage and hemosiderin deposition were common. Osteoclastic giant cells were immunoreactive for vimentin and, to a lesser extent, actin, and uniformly not immunoreactive for keratins, confirming their mesenchymal nature. The stromal component of 63% of neoplasms tested was immunoreactive for keratin, 33% was immunoreactive for epithelial membrane antigen, 547~ reacted for S100 protein, 84% reacted for actin, and 100% was immunoreactive for vimentin. Nineteen neoplasms had osteoid, bone, or cartilage, but these were a prominent component in only five neoplasms and OGC were not limited to these areas. The diseasespecific cumulative 5-year survival rate for patients with metaplastic carcinoma with OGC was 68%, similar to rates for patients with matrix-producing carcinoma (68%), spindle cell carcinoma (64%) , and squamous carcinoma of ductal origin (63%‘), but notably different from that of patients with carcinosarcoma (49%). Of 17 women with axillary node dissection, only two had metastases. Eleven women developed distant metastases, most commonly to the lungs. Metastasis present at or following initial surgery was an ominous sign, as all 11 women with metastases died from tumor. Size and microscopic circumscription were significant factors in predicting disease progression. HUM PATHOL 21: 1142-l 150. G 1990 by W.B. Saunders Company.
Osteoclastic giant cells (OK) are a rare and unusual component of carcinoma of the breast. Occasionally they are found in otherlvise ordinary carcinomas, and are a component of osteosarcoma and osteoclastic sarcoma (giant cell malignant fibrous histiocytoma) of the breast. When present in metaplastic carcinomas, OGC are an integral part of the blandappearing spindle cell or sarcomatous stroma accompanying an adenocarcinoma or squamous carcinoma of the breast. The significance of‘ an OCX component in mammary carcinoma, including metaplastic carcinomas, has not been adequatelv examined, due to the
rarity of these neoplasms. In this report we present the clinical and pathologic features of 29 metaplastic carcinomas with osteoclastic giant cells, present the immunohistochemical findings of 16, and compare this neoplasm with ordinary carcinoma with OGC and other forms of metaplastic carcinoma with which it might be confused. MATERIALS AND METHODS All lesions of the breast coded as metaplastic carcinoma. spindle cell carcinoma, carcinosarcoma, carcinoma with spindle cell metaplasia, lJse~cdosar.conlatous stroma, OI osseous or cartilaginous metaplasia in the files of the Armed Forces Institute of Pathology (AFIP. Washington, DC) accessioned befijre January 1983 were retrieved and reviewed. Twenty-nine metaplastic carcinomas with OGC and clinical information were iclentifiecl. The distribution of AFIP cases among the five forms of’ metaplastic carci~miia which we have defined, studied. and reported is presented in Table 1. Definitions are given in the appendix. The stromal component formed the bulk of the neoplasm in all cases. Intraductal or infiltrating ~iclenocarcinoriia or squanious carcinoma was present and contiguous, or admixed with stroma containing the 0(X. .4n average of six blocks of tumor were studied per case. 1’11~clmIcal data were tabulated and analvLed in cow junction with :t:oss and microscopic characteristics. All JJ~Itients were ehglble for 5-year follow-LIP. Lift tables were constructed and data analyzecl using statistical methods described in prior reports.‘-:’ All P values were derived from two-tailed tests. I’leomorphism and mitotic activity in the spindle cell stroma were evaluated and graded as bef6re,‘,” along with the intensity,of’the inflammatory infiltrate when present. ‘I‘he degree of architectural clifftl’entiation of the overt carcinoma component was also graded as described elsewhere.“,” Cellularity was inversely proportional to the amount of‘ stromal collagen and was graded 1 + (25% or fewer cells) to 4+ (75% or more cells) after microscopic examination of the tissue sections through a x 25 objective. lmmunohistochemistry
Krv wrd5: metaplastic carcinoma with osteoclartic giant cells. osteoclkma, carcinosarroma. breast carcitloma with pseudosarcomatws stroma. metaplastir c-arcinoma. Address correspondence and reprint requests to Eric 5. Wargotz, MI). Division of Pathology. Maryland Medical Laboratories at Doctor’s Hospital, XI 1X Goodluck Rd. Lanham. MLI 20706. CJ I900 by W.B. Saunders Company. oo-lti-8 177/90/2 I I I-0000$5.00/0
1142
Immunohistochen~ical studies were performed on l~ai-af‘fin-embedded tissue from 16 neoplasms: polyclonal cytokeratin in 16, AE l/3 in 1 I, Gmentin and S- 100 protein in 13, and epithelial membrane antigen (EMA) and actin in 12. The l~olvclonal antiserum to keratin proteins was prepared at the AFIP as described previously.’ The immunohistochemical techniques used and the monospecificities of the commercial polyclonal antibody to S- 100 protein alid commercial mono&ma1 antibodies to keratins (AE l/3), EMA. actin, :md vimentin were evaluated as before. .l’he pattern of cellular inimunoreactivit) was semiquantitated and designated as diffuse (> SO%), intermediate (10% to 30%~). f’ocal (a single focus or multiple cliscrete foci), or- rare (occasional isolated cells).
METAPlASTlC CARCINOMA WITH OGC (Worgotz & Norris)
TABLE 1. Classification and Distribution of Cases of Metaplastic Carcinomas of the Breast
One woman gave a history of trauma prior to the occurrence of the tumor. Most of’ the women were seen prior to the availability of routine receptor analvsis, as none was performed. Gross and Microscopic
RESULTS Clinical F-indings ‘l’he patiwts with
twiged in age f&m 28 to Xl year-s, of 54. All were a nleal~ of‘ X and ;t median
women.
{Of the 24 with known racial extraction, two ;111tl the remaining 22 (92%) were CaucaGui. Tweritv-three of the 29 ti:omen were postmenopa”“;‘l. ,411 patietlts presented with a single palpable firm ~nass. .I‘he tumor was fixed to the deep tissue in two wome~i. Nipple retraction was present in two women, ;111d skin retraction was present in one woman. None of’ the women had nipple discharge or ulceration. Axillary node enlargement was absent in all patients. Sixteen r~oplasn~s occurred in the rjght breast, 12 in rhe left breast, and one was not designated. Of cases wirh the quadrant specified, 42’2 occurred in the up32% were in the upper inner I)er outt‘r quadrant. quadrant. and the remainder presented with near-l) rqual frequency in the lower inner and outer quadrants, or were described as being on the border of two cluadrants. Neoplasms had been present from 1 to 3 was reported in Iwo cases. weeks. Pain (11. tenderness were
Mack,
FIGURE 1. Breast biopsy specimen with a poorly circumscribed 3.0 cm centrally hemorrhagic and necrotic metaplastic carcinoma with 3GC
Findings
carcinomas with O<;C were All metaplastic within the mammary parenchyma. and ranged in sile from 1 .?I cm to I5 cm, with a mean diameter of 5.5 cm and a median of 5.0 cm. (irossly, 17 had a nodular contour, five had an irregular contour, and in seven the outline was not described. Typically, neoplasms were firm to hard, gritty. with a tan to pink cut surface. and often had a cvstic center containing hemorrhagic necrotic debris (Fig 1). Microscopically, only 10 of’ the 17 grossly nodular neoplasms were circumscribed, and none was encapsulated. Seventeen neoplasms had areas of infiltration into the surrounding mammary stroma. The margin of two neoplasms could not i>e adequatel! evaluated. Foreign or polarizable marerial was not identified in any of the neoplasms. None of the neoplasms had leaf-like processes or elongated c9eft-like spaces lined by epithelium as in cystosarc’oma. All 29 neoplasms iv&e similarly composed 01‘ a dominant stromal component containing OGC, and varying degrees of carcinoma (intraductal, infiltr..rting, or both) admixed or contiguous with the stroma (Figs 2 through 5). Hemorrhage and hemosiderin were presenr in all neoplasms. and prominant in most. A majority of the neoplasms also containt~d prominent curved. branching, thin-walled vessels. Osteoclastic giant cells were clustered around and within these vessels (Figs 6 and 7j as well as diffusely throughout more cellular areas of stroma. Osteocla& giant cells
HUMAN PATHOLOGY
Volume 21, No. 11 (November 1990)
FIGURE 2. Infiltrating carcinoma and OGC within a high-grade polymorpl 1ous anification X 200.) stroma. (Mac
contained numerous round nuclei, often concentrated near the center of the cell. Most cells had bland-appearing nuclei, but in some the nuclei were enlarged, irregular, and contained prominent nucleoli. Mononuclear and binuclear forms were also present. All OK had abundant finely granular, homogeneous, eosinophilic-staining cytoplasm. Mitotic figures were observed (Fig 8). The stroma of seven neoplasms was predominantly grade 1, composed of bland-appearing spindle
cells with uniform nuclei having finely granular cytoplasm, similar in appearance to the spindle cells of spindle cell metaplastic carcinoma.2 Four neoplasms had grade 2 stroma composed of interweaving and overlapping fascicles of atypical cells, and increased mitotic activity consistent with fibrosarcoma. Grade 3 stroma (polymorphous) was the dominant stroma in 18 cases, and resembled a pleomorphic malignant fibrous histiocvtoma in terms of cellular composition and mitotic activity: however. all 18 lacked a stori-
FIGURE 3. Carcinoma, OGC. and osteoid within a high-grade polymorphous stroma. (Magnification x 200.)
METAPLASTIC CARCINOMA WITH OGC (Wargotz & Norris]
FIGURE 4. Pleomorphic OGC, bone formation, and carcinoma within a highgrade stroma [Magnification )i 200.)
!i,rnt patterrl. (~raci~ 3 strcltna was markedI\ vzllulat bvith pleomotphic nuclei, prominent nucleoli, ;tncl OFif31 COrltiGIlell niultindeate turnOr giant (rlls. Of the 29 ~~tx~pl.;~s~iis,tlw strorn;t contained osteoid or hone in nine, #fight fi)rmecf osteoicl or bone along Lvith cart ilaqe, and two f ornled cartilage alone (Figs 3 through 5). ‘Twel,r y (of thv 29 did not producr a hetcrologous ~trornal conif~otirnt.
FIGURE 5. Carcinoma. OGC, and cartlage within a polymorphous stromo. (Magnification * 75.)
Inf‘iltt-aring duct carcinoma wah j)resent in 23 neoplasnis. Intraductal carcinonw alonr~ w2s fold in six cases. ‘l‘hirteen had both. acitnixetl with ;I spindle wll stronia. A niajoritv of the infiltt-atirig cat~cinonms were moclet-atel~‘dif~~retitialecl (2.3% to 75’1( of‘ thr cxrcinonla having lutnen f’ormation) or poorly tlifferentiated (Ia4 than 9:‘,‘% of. thv c;tr(~inotm h;tving- lu-
HUMAN PATHOLOGY
Volume 21, No. 11 (November 19901
FIGURE 6. Thin-walled branching vessels with associated OGC typical of metaplastic carcinoma with OGC. (Magnification x 150.1
mixed or contiguous with the osteoclastic stroma in all 29 neoplasms. Unlike other carcinomas of the breast with OGC, the overt carcinoma component of metaplastic carcinoma with OGC did not have admixed OGC. Transitional areas between the epithelium and osteoclastic stroma were subtle. Eleven neoplasms had an admixed inflammatory infiltrate composed predominantly of lymphocytes and plasma cells. The intensity of the infiltrate was
men formation). Only one was well-differentiated (greater than i5Lr( of the carcinoma having lumen formation) with prominent cribriform and papillary patterns, a feature common to many nonmetaplastic carcinomas with OX. Infiltrating apocrine carcinoma was present in two neoplasms, adenoid cystic and adenocarcinoma patterns were mixed in one, and squamous differentiation of the carcinoma was present in three neoplasms. The carcinoma was ad-
FIGURE 7. A full spectrum of osteoclastic cells, including mononuclear, binuclear, and multinuclear forms, is identified within this high-grade polymorphous stroma, and associated with a thin-walled branching vessel. (Magnification x 200.)
1146
METAPLASTIC CARCINOMA WITH OGC [Wargok & Noms)
FIGURE 8. Atypical OGC and one with nuclei in mitosis. Poorly differentiated carcinoma and thin-walled branching vessels are also present wlthin the polymorphous stroma. (Magnification i 150.)
in ti)ur neoplasms. intermediate in five, and marked in two neoplasms. Lymphoid follicles were 1lot present. Necrosis, usually of coagulati\,e type, was ;I common feature. (~ranulomatous inflammation was not ohser\,ed. sparse
Immunohistochemistry OstecJclaaric giant cells were utiiti)rmly negative for keratins. EMA. and S-100 protein. In all neoimmunoreactive for \imen~Jlasnis, sI:)nit’ 0(X: were t;Jr actin. tin. and a f’ew were also imrnunoreactive ‘I‘his pattern of inlrnunoreactivit~ is consistent with their hist ioc\ tic origin. ‘I-he str;)ma of 10 of I6 neoplasms tested (ti3%#) for keratin: four were diffusel) \\as iIllI~~llII0I~~it~Ii~e three were immunoreac tive. one was intermedia1.e. f’ocal. and hL’0 had rare positive cells. Ot‘ the 11 neoplasms studied Lvith both polyclonal and monoclonal (.ZE l/3) i~mrnunoreagenrs. there was concordance in onlb 82%‘. due to the absence of AE 113 immunoreacti;itv in two polyclonal immunocytokeratin-reactive neoplasms. ‘I‘he stromal component in four of 12 neoplasms wxs immunoreactive for EVA, varying from focal to diffuse. C’imentin was detected in the ;mcl OK of all 13 neoplasms stromal component trsted. usually in a diffuse pattern. but it was not detected in the epithelial component. The stromal component in aevtdn of I3 neoplasms (.M’;r,) was immunoreactive for S- 100 protein. with the reactivity varying from f&al to diffuse. A focal to diffuse pattern of itumunoreactivitv for actin was observed in the stroma of 10 of’ I:! neoplasms (83%). Some of these stlbstances were coexpressed by the same population of spindle cells. Stromal myofibroblasts and fibroblasts accounted for at least some of the actin and
vimentin
irIirnilnoreactivitv.
immunoreac-ti~it): is similar plastic carcinoma.zJi except rhe stromal cells for keratin fi)r spindle cell carcinoma.
rl‘tIis pattern of stromal to other ta)rms of metathat immunoreac.tivit~ of’ is less than that observed
Initial Therapy and Lymph Node Status hlastectomy was the initial therapy in 2 1 of the 29 women. Kadical niastectom\’ iv;is performed cm nine, on eight. and simple modified radical mastectomy mastectomy ~~2s dorle on fi)ur. Partial mastectonic OII three women. (wide local’excision) was performed and excisional biopsy bias the initial thc*rapy for the remaining five women. Axillar) dissections lvere performed on 17 two ( 12% ) had lymph node mewomen and only tastases. One woman with a 3.7-cm primat-! tumor had a single micrometastasis of‘ overt carcinoma to a small lymphatic channel in the capsule of one of 26 nodes examined. I‘he second woman had a IO-cm primary tumor. and six of seven nodes \Vere positi\,e. The metastatic component was poorly dif‘ferrntiated carcinoma lvithout OGC in five of the nodes, but the sixth node contained carcinoma and a pleomorphic spindle cell proliferation with 0(X and osteoid production. Two patients are known to have received prophylactic postoperative radiation and chemotherap> wlthin 3 months of initial surgery. One woman received total of X.000 rads to the breast, axilla, and supraclavicular region following excisional biopsy. She also received 18 cycles of melphalan. A second woman received postoperative radiation and chemotherapy following excisional biopsv. 1147
HUMAN PATHOLOGY
Volume 21. No. 11 (November 19901
Survival Follow-up information obtained on all 29 women ranged from 6 months for a patient who died from tumor to 15 years for a patient free of disease. The cumulative disease-specific j-year survival rate was 68%. Five patients died of other causes, without evidence of local or distant recurrence. The mean follow-up interval was 4.9 years for all patients. The mean follow-up interval for patients without local recurrence, metastasis, or death from tumor was 7.1 years. Two of the five women treated only by excisional biopsy developed local recurrence: one woman developed recurrences at 6 and 12 months, and the other developed recurrence at 32 months. Recurrence in both patients was subsequently treated by extended simple mastectomy. There were no metastases to axillary lymph nodes, and both women are free of tumor at 6.3 and 8.3 years, respectively, following initial therapy. The two patients who received prophylactic postoperative radiation and chemotherapy within 3 months of excisional biopsy did not develop local recurrence or metastasis, and both have been followed for at least 8.5 vears. Of the 1 1 women who developed metastases, ‘eight developed them from 3 months to 4.9 years (mean, 1.7 years; median, 11.5 months) following initial therapy. In three fi-orn initial therapy to local 01 women the interval distant recurrence is unknown. All 11 women with metastases died from tumor at 6 months to 5.8 years (mean of 2.3 years. median of 1.4 years) following initial therapy. Three of these patients had been treated by radical mastectomy, one also received bilateral oophorectomy. Four of the 1 1 women who died from tumor had been treated by modified radical mastectomy, two by simple mastectomy, one had
a partial mastectomy, and one had excisional biopsy only. Seven women developed lung metastases following initial therapy, including the two women with axillary node metastases. In addition to lung metastases, one woman had metastases to mediastinal lymph nodes, and one woman developed metastases in the liver. One patient developed multiple bone meAt least tastases without evidence of lung metastasis. two women with lung metastases received radiation and chemotherapy. ,4n additional patient was treated by pneumonectomy. Sections of lung metastases from two women were reviewed and, in both, metastases contained abundant OGC within a grade 3 sarcomatous stroma without an overt carcinoma component (Fig 9). Disease-related mortality rates fiJlowing each of’ the five surgical approaches were as follows: 33% for the nine women with radical mastectomy, 50c/c8for the eight treated by modified radical mastectomy, 50% for the four women having simple mastectomy, 33% for partial mastectomy (one of three patients). and ‘LO%,for the five patients treated by excisional biopsy, although an additional two patients treated initialIF by excisional biopsy developed local recurrence and were subsequentlv treated by mastectomy. Analysis of Recurrences The 11 women who developed metastases were not significantly older (mean age 57) than patients without them (mean age 56); however, their neoplasms were significantly larger (P = ,039, mean diof ameter of 7.2 cm versus 4.5 cm). The relationship size to metastatic potential and death from tumor is illustrated in Table 2. Four of five patients initially treated by excisional biopsy had neoplasms less than
FIGURE 9. Peribronchiolar metastatic metaplastic carcinoma with OGC. Osteoelastic giant cells within a polymorphous stroma without overt carcinoma. (Magnification x 100.)
1148
METAPLASTIC CARCINOMA WITH OGC (Wargotz & Norris]
TABLE 2. .~____
Relationship of Size to Metastatic Potential
35 un in clianieter. A fif’th patient treated by hiops> for ;I neoplasm 7.5 cm in diameter died from metaaratic tumor. A gross notlular contour was not ;I sig: nifiant f;rc.tor against metiistasis. as se\en of the 1 i grossI\ ciirc umscrifxd neoplasms metastasized. Microscwpicdly. mayins were important. as growth char;icterixd t)v ii c~lrcttniscrihecl or pushing margin h2s present in eight of’ 14 neoplasms (57’7) which did not recur 1oc;tllv or metastasix, whereas only tb.0 ot‘ 13 neopl;tsrr~s (1.5%) which subsequentlv r-e&l-red locall\ or metastasized (P = .O’LS) had &s f’eafuw. Other histolo#c features such as grade of‘the stroma (blantlappeanng spindle ccII versus sarconiatous stroma), tlegrre of’ cellularitv. mitotic activitv. or presence of’ bone or cartilage or an inflanimatc~r-y infiltrate, and tlyree oi‘ clif’ferentiatiori of the carcinoma did not diff’er sigriXic~;intly between neoplasms whit h metastasi/.ecl and thaw which did not (‘I‘al~le ?i).
DISCUSSION hlarnrnan cwcinom;ts with (XX:. f’irst described over HO vcxrs ;igo. ’ are rare neoplasms lvith approximatelv !
\Itct-ceropic, i 11.(.11tIlsct.ij)tioII Itttlartttnatotv ittfiltt~;ttr 13one ot. cat-tihgt
.N~mt-e(.LIt-t-ent
I :i ‘4
ii7
(I’ = .t11’:1j
Iti’% “3’7
:+I’( St;‘{
(I’ = ,071)
Stt~mrd: Ket-atin itrlrttua[c,tc.bc ti\tt\ (kllularit~ f I’le(lnlclr-phi\n1 \litotic tigttws pr 10 k1l’P ---
iti’,?
7 1’;
3..i +
3.1+
Lpdr 20
2.2
$pdc
“.fi
2.t
* .%lcarr ‘.~‘lIW\. t Itrclt~df:~ 10~.(I tc‘ou urret~ce> ;tnd distal nwt;t*tdw~
(XX (‘I‘able 1). such as matrix-producing carcinoma (6X%),’ spindle cell carcinoma (64g ).” and pure squamous carcinoma of ductal origin (ES%,). I7 The .?-year survival ratrs of’ metaplastic carcinoma with C)(;(: (MC%) and carcinosarcoma (49%). however. are notably different.:’ Patients with metaplastic. carcinoma ivith 0(X: have a S-year sur-\,i\,al rate similar to or better than patients with ordinal-\ inf’iltrating duct (.arcinonia. Although metaplastic carcinoma with (XX: ma! have a stroma that varies from bland-appearing spindle cells to high-grade sarcoma, it is distinguished f’rom spindle cell carcinoma and carcinosarcoma b) the presence of 0C:C within cellular areas of stroma. frequelitly associated with osteoid or cartilage (59%. of neoplasms in this study). In contrast. othrr forms ot‘metapfastic carcinoma ‘with osteoid. twne, or cartilage. such as matrix-producing carcinoma, spindle cell carcmoma, and carcinosarcorria. Lick an associated OK component. ‘-:3.’’ In addition. matrix-producing carcinoma lacks a spindle cell component (see appendix). St I-omal hemorrha&e with “bloodfake” formation, hemosiderin deposItwn. and prominent thin-rvalled vessels forming cleft-like spaces are features of nietaplastic carcinoma kvith (XC:. Osteoelastic ,giant cells are commonlv found clustered around anti within these thin-walled vessels. ‘I‘he characteristic \,ascularitv as well as the 0(X are distinguishing features of’this form of metaplastic carcinoma. .IIle stromal hemorrhage. hemosiderin, and prominent \ asculature, however. have been described 21s f’wtures of‘ other breast c‘arc inomas with (XX:, but these carcinomas lack the pronlinent metaplastic stromal components that dominate metaplastic carcinoma. Also, in metaplastic carcinoma with 0(X, the osteoclastic giant cells arc not aclmi?c.c~tl tvith overt carcinoma, as the), are in breast carcinomas with OK that lack the metaplastic stroma.!‘,’ i I” ‘l‘he grade of‘ the stroma in metaplastic carcinoma wit11 0(X was not a significant prognostic indicator for thr likelihood of local recurrence or distant metastasis. Furthermore, although 0(X are more likely to occur with osteoid, bone, or cartilage f’ormatioll. their presence is not exclusively dependent on the pro~luction of‘ these stromal elements. Several theories regarding the histogenesis of’the 0(X: in breast carcinomas have been ad\,ancecl. Most investigators have recognized that this fi)rm of giant cell is not epithelial but of histiocyic, (,ri~in.7,‘.‘.“.‘i.“i We found 0(X; were irnmunoreactive i;,r \imentin. to a lesser extent for actin, and a11 were negative for keratin ancl EMA. ‘These results also are f urtlier evidence against epithelial origin. ‘I‘he presence of OK in carcinomas of‘thc breast and other sites represents a response I:() substances elaborated bv the tumor cells or to tum01‘ cell surface antigens. although the factor5 which stimulate the development of’ OCX are unclear. Several studies using murine and human breast tissue indicate that in some instances carcinomas of the breast stimulate angiogenesis. ’ H-Z”’I’he prolif’eration of‘ thin-walled vessels and the tendencv of OK to associate intimatelv with
1149
HUMAN PATHOLOGY
Volume 21, No. 11 (November 1990)
these vessels suggests there is perhaps a common stimulating factor responsible for both stromal angiogenesis and development of OGC. APPENDIX-GLOSSARY Definitions Overt carcinoma with diMnlris-producing carcimm2u. rect transition to a cartilaginous and/or osseous stromal matrix without an intervening spindle cell zone. .Spindlr crll curcinomcl. A dominant spindle cell proliferation composed of bland-appearing bipolar cells with plump nuclei, insignificant pleomorphism, and low mitotic activity growing as feathered. overlapping, or interweaving wavy fdscicles. often in a storiform pattern. Collagen produciion is characteristic, as are myxoid ancl angioid areas. There must be in situ carcinoma or ductal. lobular, squamous, or mixed infiltrating carcinoma, which is contiguous or subtly merges with the spindle cell component. If overt carcinoma is absent, the epithelial nature must be proven electron microscopy. 01 through immunohistochemistry, some other method. with at least half Cnr-cin0.m rco7rilln. Biphasic neoplasms of the tumor composed of a cellular malignant-appearing spindle cell component consisting of pleomorphic bipolar cells or a polymorphous cell population. with nuclear atypia and high mltotic activity, and an in situ or ductal, squamous, or mixed infiltratmg carcinoma, which is contiguous or admixed with the sarcomatous component. Pzo-r sq~~umous ~11 carcinoma. o_f ductul origin. Infiltrating carcinoma that is exclusively squamous without involvement of the skin, or intraductal carcinoma which is exclusively squamous. M&plastic cm-inomu ulith ostrockttic gircmt ~11.~. Intraductal or infiltrating carcinoma contiguous or admixed with a bland-appearing spindle cell stroma or sarcomatous stroma in which osteoclastic giant cells are immixed. REFERENCES I. Wargotz ES, Norris H,l: Metaplastic carcinomas of the breast. I. Matrix-producing carcinoma. HUM P.vrHoL ?O:f528-6%. 1989 2. Wargotz ES. l)eos PH. Norris H,J: Metaplastic carcinomas
of the breast. II. Spindle cell carcinoma. Ht1.w PATHOI. %):7:32-740. 19X9 3. Wargotz ES, Norris FiJ: Sletaplastic carcinomas of the breast. 111. Carcinosarcoma. Cancer 64:1490-1199. 1989 4. Lecene P: Les tumeurs mixtes du sein. Re\ (:hir (Chir] 33:434-439. 1906 .5. Leroux R: Reaction giganto cellulaire du stnmla dam un epithelioma mammaire. Hull Cancer ?0:6924i97. 193 1 6. I)uhoucher H. Montpellier J, Laffargue F: E:pithelioma mammaire avcc reaction de tk pe myeloplaxique. Ann Anat Pathol 70:787-79 I. 193:3 7. Nielsen BB, Kiaer HLV: Carcinoma of the breast with stroma multinucleated giant cells. Histopathology 9: 1X3- 193. 198.3 8. Kobayashi S. ‘Tobioka N. Samoto T. et al: Breast cancer with osteoclast like multinucleated giant cells. Acta l’athol ,]pn 34:147.5-14x1. 19x4 9. Holland R. Llrhain I(;. tan Haelst hl: hlammarp carcmoma with osteoclast like giant cells. Additional observations on six easer. Cancer .53:lW-1973, 1984 10. Huvos A(;. Lucas JC, Foote FW: Metaplastic breast carcinoma. NY State J Xled 73: 107X-10X2. 1973 1 I. Fisher ER, Palekar AS, (iregorio KM. et al: Mucoepidermoid and squamous cell carcinomas of breast with reference to squamous metaplasia and giant cell tumors. ArnJ Surg Pathol 7: I.5 “7, 1983 l?. Factor FM, Biempica 1.. Katner 1. et al: (:arcinoma of the bt-east with multinucleated reactive stromal giant cells. A light and electron microscopic study of two cases. Virchow’s Arch [A] 374: l12. 1!,77 13. Gras M, Zafiridou H, Karaiossifides (:: Unusual stl-omal response in a malignant breast ttlnlor. Presented at the \‘I Congress of the European Societb of Pathology. 1977
II. Oberman HA: Metaplastic carcinonla of the breast. A clinicopathologic studs of 29 patients. Am J Surg Pathol 1 I :918-929. 1987 1.i. Tavassoli FA. Norris H,] : Breast carcinoma with osteoclastlike giant cells. Arc-h Pathol Lab Med 1 lO:tX-639. 1986 16. Agnantis N?‘, Rosen PP: Mammary,carcinoma with ostt‘oclast-like giant cells. A study of eight cases wnh follow-up data. Am J Clin Pathol 72:3X3-3X9. 1979 17. Wargotz ES. Norris H,J: hletaplastic tarcinomas of the bl-east. IV. Squamous cell carcinoma. Cancel- 65:272-276. 1991) 1X. (;imbrone MA Jr. (;ullino PM: Neovasc~llar-ization induced hv intraocular xenografts of normal, preneoplastic and neoplastic mouse mammar> tissues. JN(:I 56:305-318. 1976 19. (Cullino PM: Natural history of breast cancer. Progression from hyperplasia to neoplasia as predicted by angiogenesis. (:ancer XJ:2697-“703 I . 1977 20. Brem SS. ]ensrn HM. (;ullino l’hl: Angiogenesis as a marker of preneoplastic lesions of the human breast. (;anc-er -1I :“39-“44 197X -.
1150
Osteoclastic giant cells (OK) are a rare and unusual component of carcinoma of the breast. Occasionally they are found in otherlvise ordinary carcinomas, and are a component of osteosarcoma and osteoclastic sarcoma (giant cell malignant fibrous histiocytoma) of the breast. When present in metaplastic carcinomas, OGC are an integral part of the blandappearing spindle cell or sarcomatous stroma accompanying an adenocarcinoma or squamous carcinoma of the breast. The significance of‘ an OCX component in mammary carcinoma, including metaplastic carcinomas, has not been adequatelv examined, due to the
rarity of these neoplasms. In this report we present the clinical and pathologic features of 29 metaplastic carcinomas with osteoclastic giant cells, present the immunohistochemical findings of 16, and compare this neoplasm with ordinary carcinoma with OGC and other forms of metaplastic carcinoma with which it might be confused. MATERIALS AND METHODS All lesions of the breast coded as metaplastic carcinoma. spindle cell carcinoma, carcinosarcoma, carcinoma with spindle cell metaplasia, lJse~cdosar.conlatous stroma, OI osseous or cartilaginous metaplasia in the files of the Armed Forces Institute of Pathology (AFIP. Washington, DC) accessioned befijre January 1983 were retrieved and reviewed. Twenty-nine metaplastic carcinomas with OGC and clinical information were iclentifiecl. The distribution of AFIP cases among the five forms of’ metaplastic carci~miia which we have defined, studied. and reported is presented in Table 1. Definitions are given in the appendix. The stromal component formed the bulk of the neoplasm in all cases. Intraductal or infiltrating ~iclenocarcinoriia or squanious carcinoma was present and contiguous, or admixed with stroma containing the 0(X. .4n average of six blocks of tumor were studied per case. 1’11~clmIcal data were tabulated and analvLed in cow junction with :t:oss and microscopic characteristics. All JJ~Itients were ehglble for 5-year follow-LIP. Lift tables were constructed and data analyzecl using statistical methods described in prior reports.‘-:’ All P values were derived from two-tailed tests. I’leomorphism and mitotic activity in the spindle cell stroma were evaluated and graded as bef6re,‘,” along with the intensity,of’the inflammatory infiltrate when present. ‘I‘he degree of architectural clifftl’entiation of the overt carcinoma component was also graded as described elsewhere.“,” Cellularity was inversely proportional to the amount of‘ stromal collagen and was graded 1 + (25% or fewer cells) to 4+ (75% or more cells) after microscopic examination of the tissue sections through a x 25 objective. lmmunohistochemistry
Krv wrd5: metaplastic carcinoma with osteoclartic giant cells. osteoclkma, carcinosarroma. breast carcitloma with pseudosarcomatws stroma. metaplastir c-arcinoma. Address correspondence and reprint requests to Eric 5. Wargotz, MI). Division of Pathology. Maryland Medical Laboratories at Doctor’s Hospital, XI 1X Goodluck Rd. Lanham. MLI 20706. CJ I900 by W.B. Saunders Company. oo-lti-8 177/90/2 I I I-0000$5.00/0
1142
Immunohistochen~ical studies were performed on l~ai-af‘fin-embedded tissue from 16 neoplasms: polyclonal cytokeratin in 16, AE l/3 in 1 I, Gmentin and S- 100 protein in 13, and epithelial membrane antigen (EMA) and actin in 12. The l~olvclonal antiserum to keratin proteins was prepared at the AFIP as described previously.’ The immunohistochemical techniques used and the monospecificities of the commercial polyclonal antibody to S- 100 protein alid commercial mono&ma1 antibodies to keratins (AE l/3), EMA. actin, :md vimentin were evaluated as before. .l’he pattern of cellular inimunoreactivit) was semiquantitated and designated as diffuse (> SO%), intermediate (10% to 30%~). f’ocal (a single focus or multiple cliscrete foci), or- rare (occasional isolated cells).
METAPlASTlC CARCINOMA WITH OGC (Worgotz & Norris)
TABLE 1. Classification and Distribution of Cases of Metaplastic Carcinomas of the Breast
One woman gave a history of trauma prior to the occurrence of the tumor. Most of’ the women were seen prior to the availability of routine receptor analvsis, as none was performed. Gross and Microscopic
RESULTS Clinical F-indings ‘l’he patiwts with
twiged in age f&m 28 to Xl year-s, of 54. All were a nleal~ of‘ X and ;t median
women.
{Of the 24 with known racial extraction, two ;111tl the remaining 22 (92%) were CaucaGui. Tweritv-three of the 29 ti:omen were postmenopa”“;‘l. ,411 patietlts presented with a single palpable firm ~nass. .I‘he tumor was fixed to the deep tissue in two wome~i. Nipple retraction was present in two women, ;111d skin retraction was present in one woman. None of’ the women had nipple discharge or ulceration. Axillary node enlargement was absent in all patients. Sixteen r~oplasn~s occurred in the rjght breast, 12 in rhe left breast, and one was not designated. Of cases wirh the quadrant specified, 42’2 occurred in the up32% were in the upper inner I)er outt‘r quadrant. quadrant. and the remainder presented with near-l) rqual frequency in the lower inner and outer quadrants, or were described as being on the border of two cluadrants. Neoplasms had been present from 1 to 3 was reported in Iwo cases. weeks. Pain (11. tenderness were
Mack,
FIGURE 1. Breast biopsy specimen with a poorly circumscribed 3.0 cm centrally hemorrhagic and necrotic metaplastic carcinoma with 3GC
Findings
carcinomas with O<;C were All metaplastic within the mammary parenchyma. and ranged in sile from 1 .?I cm to I5 cm, with a mean diameter of 5.5 cm and a median of 5.0 cm. (irossly, 17 had a nodular contour, five had an irregular contour, and in seven the outline was not described. Typically, neoplasms were firm to hard, gritty. with a tan to pink cut surface. and often had a cvstic center containing hemorrhagic necrotic debris (Fig 1). Microscopically, only 10 of’ the 17 grossly nodular neoplasms were circumscribed, and none was encapsulated. Seventeen neoplasms had areas of infiltration into the surrounding mammary stroma. The margin of two neoplasms could not i>e adequatel! evaluated. Foreign or polarizable marerial was not identified in any of the neoplasms. None of the neoplasms had leaf-like processes or elongated c9eft-like spaces lined by epithelium as in cystosarc’oma. All 29 neoplasms iv&e similarly composed 01‘ a dominant stromal component containing OGC, and varying degrees of carcinoma (intraductal, infiltr..rting, or both) admixed or contiguous with the stroma (Figs 2 through 5). Hemorrhage and hemosiderin were presenr in all neoplasms. and prominant in most. A majority of the neoplasms also containt~d prominent curved. branching, thin-walled vessels. Osteoclastic giant cells were clustered around and within these vessels (Figs 6 and 7j as well as diffusely throughout more cellular areas of stroma. Osteocla& giant cells
HUMAN PATHOLOGY
Volume 21, No. 11 (November 1990)
FIGURE 2. Infiltrating carcinoma and OGC within a high-grade polymorpl 1ous anification X 200.) stroma. (Mac
contained numerous round nuclei, often concentrated near the center of the cell. Most cells had bland-appearing nuclei, but in some the nuclei were enlarged, irregular, and contained prominent nucleoli. Mononuclear and binuclear forms were also present. All OK had abundant finely granular, homogeneous, eosinophilic-staining cytoplasm. Mitotic figures were observed (Fig 8). The stroma of seven neoplasms was predominantly grade 1, composed of bland-appearing spindle
cells with uniform nuclei having finely granular cytoplasm, similar in appearance to the spindle cells of spindle cell metaplastic carcinoma.2 Four neoplasms had grade 2 stroma composed of interweaving and overlapping fascicles of atypical cells, and increased mitotic activity consistent with fibrosarcoma. Grade 3 stroma (polymorphous) was the dominant stroma in 18 cases, and resembled a pleomorphic malignant fibrous histiocvtoma in terms of cellular composition and mitotic activity: however. all 18 lacked a stori-
FIGURE 3. Carcinoma, OGC. and osteoid within a high-grade polymorphous stroma. (Magnification x 200.)
METAPLASTIC CARCINOMA WITH OGC (Wargotz & Norris]
FIGURE 4. Pleomorphic OGC, bone formation, and carcinoma within a highgrade stroma [Magnification )i 200.)
!i,rnt patterrl. (~raci~ 3 strcltna was markedI\ vzllulat bvith pleomotphic nuclei, prominent nucleoli, ;tncl OFif31 COrltiGIlell niultindeate turnOr giant (rlls. Of the 29 ~~tx~pl.;~s~iis,tlw strorn;t contained osteoid or hone in nine, #fight fi)rmecf osteoicl or bone along Lvith cart ilaqe, and two f ornled cartilage alone (Figs 3 through 5). ‘Twel,r y (of thv 29 did not producr a hetcrologous ~trornal conif~otirnt.
FIGURE 5. Carcinoma. OGC, and cartlage within a polymorphous stromo. (Magnification * 75.)
Inf‘iltt-aring duct carcinoma wah j)resent in 23 neoplasnis. Intraductal carcinonw alonr~ w2s fold in six cases. ‘l‘hirteen had both. acitnixetl with ;I spindle wll stronia. A niajoritv of the infiltt-atirig cat~cinonms were moclet-atel~‘dif~~retitialecl (2.3% to 75’1( of‘ thr cxrcinonla having lutnen f’ormation) or poorly tlifferentiated (Ia4 than 9:‘,‘% of. thv c;tr(~inotm h;tving- lu-
HUMAN PATHOLOGY
Volume 21, No. 11 (November 19901
FIGURE 6. Thin-walled branching vessels with associated OGC typical of metaplastic carcinoma with OGC. (Magnification x 150.1
mixed or contiguous with the osteoclastic stroma in all 29 neoplasms. Unlike other carcinomas of the breast with OGC, the overt carcinoma component of metaplastic carcinoma with OGC did not have admixed OGC. Transitional areas between the epithelium and osteoclastic stroma were subtle. Eleven neoplasms had an admixed inflammatory infiltrate composed predominantly of lymphocytes and plasma cells. The intensity of the infiltrate was
men formation). Only one was well-differentiated (greater than i5Lr( of the carcinoma having lumen formation) with prominent cribriform and papillary patterns, a feature common to many nonmetaplastic carcinomas with OX. Infiltrating apocrine carcinoma was present in two neoplasms, adenoid cystic and adenocarcinoma patterns were mixed in one, and squamous differentiation of the carcinoma was present in three neoplasms. The carcinoma was ad-
FIGURE 7. A full spectrum of osteoclastic cells, including mononuclear, binuclear, and multinuclear forms, is identified within this high-grade polymorphous stroma, and associated with a thin-walled branching vessel. (Magnification x 200.)
1146
METAPLASTIC CARCINOMA WITH OGC [Wargok & Noms)
FIGURE 8. Atypical OGC and one with nuclei in mitosis. Poorly differentiated carcinoma and thin-walled branching vessels are also present wlthin the polymorphous stroma. (Magnification i 150.)
in ti)ur neoplasms. intermediate in five, and marked in two neoplasms. Lymphoid follicles were 1lot present. Necrosis, usually of coagulati\,e type, was ;I common feature. (~ranulomatous inflammation was not ohser\,ed. sparse
Immunohistochemistry OstecJclaaric giant cells were utiiti)rmly negative for keratins. EMA. and S-100 protein. In all neoimmunoreactive for \imen~Jlasnis, sI:)nit’ 0(X: were t;Jr actin. tin. and a f’ew were also imrnunoreactive ‘I‘his pattern of inlrnunoreactivit~ is consistent with their hist ioc\ tic origin. ‘I-he str;)ma of 10 of I6 neoplasms tested (ti3%#) for keratin: four were diffusel) \\as iIllI~~llII0I~~it~Ii~e three were immunoreac tive. one was intermedia1.e. f’ocal. and hL’0 had rare positive cells. Ot‘ the 11 neoplasms studied Lvith both polyclonal and monoclonal (.ZE l/3) i~mrnunoreagenrs. there was concordance in onlb 82%‘. due to the absence of AE 113 immunoreacti;itv in two polyclonal immunocytokeratin-reactive neoplasms. ‘I‘he stromal component in four of 12 neoplasms wxs immunoreactive for EVA, varying from focal to diffuse. C’imentin was detected in the ;mcl OK of all 13 neoplasms stromal component trsted. usually in a diffuse pattern. but it was not detected in the epithelial component. The stromal component in aevtdn of I3 neoplasms (.M’;r,) was immunoreactive for S- 100 protein. with the reactivity varying from f&al to diffuse. A focal to diffuse pattern of itumunoreactivitv for actin was observed in the stroma of 10 of’ I:! neoplasms (83%). Some of these stlbstances were coexpressed by the same population of spindle cells. Stromal myofibroblasts and fibroblasts accounted for at least some of the actin and
vimentin
irIirnilnoreactivitv.
immunoreac-ti~it): is similar plastic carcinoma.zJi except rhe stromal cells for keratin fi)r spindle cell carcinoma.
rl‘tIis pattern of stromal to other ta)rms of metathat immunoreac.tivit~ of’ is less than that observed
Initial Therapy and Lymph Node Status hlastectomy was the initial therapy in 2 1 of the 29 women. Kadical niastectom\’ iv;is performed cm nine, on eight. and simple modified radical mastectomy mastectomy ~~2s dorle on fi)ur. Partial mastectonic OII three women. (wide local’excision) was performed and excisional biopsy bias the initial thc*rapy for the remaining five women. Axillar) dissections lvere performed on 17 two ( 12% ) had lymph node mewomen and only tastases. One woman with a 3.7-cm primat-! tumor had a single micrometastasis of‘ overt carcinoma to a small lymphatic channel in the capsule of one of 26 nodes examined. I‘he second woman had a IO-cm primary tumor. and six of seven nodes \Vere positi\,e. The metastatic component was poorly dif‘ferrntiated carcinoma lvithout OGC in five of the nodes, but the sixth node contained carcinoma and a pleomorphic spindle cell proliferation with 0(X and osteoid production. Two patients are known to have received prophylactic postoperative radiation and chemotherap> wlthin 3 months of initial surgery. One woman received total of X.000 rads to the breast, axilla, and supraclavicular region following excisional biopsy. She also received 18 cycles of melphalan. A second woman received postoperative radiation and chemotherapy following excisional biopsv. 1147
HUMAN PATHOLOGY
Volume 21. No. 11 (November 19901
Survival Follow-up information obtained on all 29 women ranged from 6 months for a patient who died from tumor to 15 years for a patient free of disease. The cumulative disease-specific j-year survival rate was 68%. Five patients died of other causes, without evidence of local or distant recurrence. The mean follow-up interval was 4.9 years for all patients. The mean follow-up interval for patients without local recurrence, metastasis, or death from tumor was 7.1 years. Two of the five women treated only by excisional biopsy developed local recurrence: one woman developed recurrences at 6 and 12 months, and the other developed recurrence at 32 months. Recurrence in both patients was subsequently treated by extended simple mastectomy. There were no metastases to axillary lymph nodes, and both women are free of tumor at 6.3 and 8.3 years, respectively, following initial therapy. The two patients who received prophylactic postoperative radiation and chemotherapy within 3 months of excisional biopsy did not develop local recurrence or metastasis, and both have been followed for at least 8.5 vears. Of the 1 1 women who developed metastases, ‘eight developed them from 3 months to 4.9 years (mean, 1.7 years; median, 11.5 months) following initial therapy. In three fi-orn initial therapy to local 01 women the interval distant recurrence is unknown. All 11 women with metastases died from tumor at 6 months to 5.8 years (mean of 2.3 years. median of 1.4 years) following initial therapy. Three of these patients had been treated by radical mastectomy, one also received bilateral oophorectomy. Four of the 1 1 women who died from tumor had been treated by modified radical mastectomy, two by simple mastectomy, one had
a partial mastectomy, and one had excisional biopsy only. Seven women developed lung metastases following initial therapy, including the two women with axillary node metastases. In addition to lung metastases, one woman had metastases to mediastinal lymph nodes, and one woman developed metastases in the liver. One patient developed multiple bone meAt least tastases without evidence of lung metastasis. two women with lung metastases received radiation and chemotherapy. ,4n additional patient was treated by pneumonectomy. Sections of lung metastases from two women were reviewed and, in both, metastases contained abundant OGC within a grade 3 sarcomatous stroma without an overt carcinoma component (Fig 9). Disease-related mortality rates fiJlowing each of’ the five surgical approaches were as follows: 33% for the nine women with radical mastectomy, 50c/c8for the eight treated by modified radical mastectomy, 50% for the four women having simple mastectomy, 33% for partial mastectomy (one of three patients). and ‘LO%,for the five patients treated by excisional biopsy, although an additional two patients treated initialIF by excisional biopsy developed local recurrence and were subsequentlv treated by mastectomy. Analysis of Recurrences The 11 women who developed metastases were not significantly older (mean age 57) than patients without them (mean age 56); however, their neoplasms were significantly larger (P = ,039, mean diof ameter of 7.2 cm versus 4.5 cm). The relationship size to metastatic potential and death from tumor is illustrated in Table 2. Four of five patients initially treated by excisional biopsy had neoplasms less than
FIGURE 9. Peribronchiolar metastatic metaplastic carcinoma with OGC. Osteoelastic giant cells within a polymorphous stroma without overt carcinoma. (Magnification x 100.)
1148
METAPLASTIC CARCINOMA WITH OGC (Wargotz & Norris]
TABLE 2. .~____
Relationship of Size to Metastatic Potential
35 un in clianieter. A fif’th patient treated by hiops> for ;I neoplasm 7.5 cm in diameter died from metaaratic tumor. A gross notlular contour was not ;I sig: nifiant f;rc.tor against metiistasis. as se\en of the 1 i grossI\ ciirc umscrifxd neoplasms metastasized. Microscwpicdly. mayins were important. as growth char;icterixd t)v ii c~lrcttniscrihecl or pushing margin h2s present in eight of’ 14 neoplasms (57’7) which did not recur 1oc;tllv or metastasix, whereas only tb.0 ot‘ 13 neopl;tsrr~s (1.5%) which subsequentlv r-e&l-red locall\ or metastasized (P = .O’LS) had &s f’eafuw. Other histolo#c features such as grade of‘the stroma (blantlappeanng spindle ccII versus sarconiatous stroma), tlegrre of’ cellularitv. mitotic activitv. or presence of’ bone or cartilage or an inflanimatc~r-y infiltrate, and tlyree oi‘ clif’ferentiatiori of the carcinoma did not diff’er sigriXic~;intly between neoplasms whit h metastasi/.ecl and thaw which did not (‘I‘al~le ?i).
DISCUSSION hlarnrnan cwcinom;ts with (XX:. f’irst described over HO vcxrs ;igo. ’ are rare neoplasms lvith approximatelv !
\Itct-ceropic, i 11.(.11tIlsct.ij)tioII Itttlartttnatotv ittfiltt~;ttr 13one ot. cat-tihgt
.N~mt-e(.LIt-t-ent
I :i ‘4
ii7
(I’ = .t11’:1j
Iti’% “3’7
:+I’( St;‘{
(I’ = ,071)
Stt~mrd: Ket-atin itrlrttua[c,tc.bc ti\tt\ (kllularit~ f I’le(lnlclr-phi\n1 \litotic tigttws pr 10 k1l’P ---
iti’,?
7 1’;
3..i +
3.1+
Lpdr 20
2.2
$pdc
“.fi
2.t
* .%lcarr ‘.~‘lIW\. t Itrclt~df:~ 10~.(I tc‘ou urret~ce> ;tnd distal nwt;t*tdw~
(XX (‘I‘able 1). such as matrix-producing carcinoma (6X%),’ spindle cell carcinoma (64g ).” and pure squamous carcinoma of ductal origin (ES%,). I7 The .?-year survival ratrs of’ metaplastic carcinoma with C)(;(: (MC%) and carcinosarcoma (49%). however. are notably different.:’ Patients with metaplastic. carcinoma ivith 0(X: have a S-year sur-\,i\,al rate similar to or better than patients with ordinal-\ inf’iltrating duct (.arcinonia. Although metaplastic carcinoma with (XX: ma! have a stroma that varies from bland-appearing spindle cells to high-grade sarcoma, it is distinguished f’rom spindle cell carcinoma and carcinosarcoma b) the presence of 0C:C within cellular areas of stroma. frequelitly associated with osteoid or cartilage (59%. of neoplasms in this study). In contrast. othrr forms ot‘metapfastic carcinoma ‘with osteoid. twne, or cartilage. such as matrix-producing carcinoma, spindle cell carcmoma, and carcinosarcorria. Lick an associated OK component. ‘-:3.’’ In addition. matrix-producing carcinoma lacks a spindle cell component (see appendix). St I-omal hemorrha&e with “bloodfake” formation, hemosiderin deposItwn. and prominent thin-rvalled vessels forming cleft-like spaces are features of nietaplastic carcinoma kvith (XC:. Osteoelastic ,giant cells are commonlv found clustered around anti within these thin-walled vessels. ‘I‘he characteristic \,ascularitv as well as the 0(X are distinguishing features of’this form of metaplastic carcinoma. .IIle stromal hemorrhage. hemosiderin, and prominent \ asculature, however. have been described 21s f’wtures of‘ other breast c‘arc inomas with (XX:, but these carcinomas lack the pronlinent metaplastic stromal components that dominate metaplastic carcinoma. Also, in metaplastic carcinoma with 0(X, the osteoclastic giant cells arc not aclmi?c.c~tl tvith overt carcinoma, as the), are in breast carcinomas with OK that lack the metaplastic stroma.!‘,’ i I” ‘l‘he grade of‘ the stroma in metaplastic carcinoma wit11 0(X was not a significant prognostic indicator for thr likelihood of local recurrence or distant metastasis. Furthermore, although 0(X are more likely to occur with osteoid, bone, or cartilage f’ormatioll. their presence is not exclusively dependent on the pro~luction of‘ these stromal elements. Several theories regarding the histogenesis of’the 0(X: in breast carcinomas have been ad\,ancecl. Most investigators have recognized that this fi)rm of giant cell is not epithelial but of histiocyic, (,ri~in.7,‘.‘.“.‘i.“i We found 0(X; were irnmunoreactive i;,r \imentin. to a lesser extent for actin, and a11 were negative for keratin ancl EMA. ‘These results also are f urtlier evidence against epithelial origin. ‘I‘he presence of OK in carcinomas of‘thc breast and other sites represents a response I:() substances elaborated bv the tumor cells or to tum01‘ cell surface antigens. although the factor5 which stimulate the development of’ OCX are unclear. Several studies using murine and human breast tissue indicate that in some instances carcinomas of the breast stimulate angiogenesis. ’ H-Z”’I’he prolif’eration of‘ thin-walled vessels and the tendencv of OK to associate intimatelv with
1149
HUMAN PATHOLOGY
Volume 21, No. 11 (November 1990)
these vessels suggests there is perhaps a common stimulating factor responsible for both stromal angiogenesis and development of OGC. APPENDIX-GLOSSARY Definitions Overt carcinoma with diMnlris-producing carcimm2u. rect transition to a cartilaginous and/or osseous stromal matrix without an intervening spindle cell zone. .Spindlr crll curcinomcl. A dominant spindle cell proliferation composed of bland-appearing bipolar cells with plump nuclei, insignificant pleomorphism, and low mitotic activity growing as feathered. overlapping, or interweaving wavy fdscicles. often in a storiform pattern. Collagen produciion is characteristic, as are myxoid ancl angioid areas. There must be in situ carcinoma or ductal. lobular, squamous, or mixed infiltrating carcinoma, which is contiguous or subtly merges with the spindle cell component. If overt carcinoma is absent, the epithelial nature must be proven electron microscopy. 01 through immunohistochemistry, some other method. with at least half Cnr-cin0.m rco7rilln. Biphasic neoplasms of the tumor composed of a cellular malignant-appearing spindle cell component consisting of pleomorphic bipolar cells or a polymorphous cell population. with nuclear atypia and high mltotic activity, and an in situ or ductal, squamous, or mixed infiltratmg carcinoma, which is contiguous or admixed with the sarcomatous component. Pzo-r sq~~umous ~11 carcinoma. o_f ductul origin. Infiltrating carcinoma that is exclusively squamous without involvement of the skin, or intraductal carcinoma which is exclusively squamous. M&plastic cm-inomu ulith ostrockttic gircmt ~11.~. Intraductal or infiltrating carcinoma contiguous or admixed with a bland-appearing spindle cell stroma or sarcomatous stroma in which osteoclastic giant cells are immixed. REFERENCES I. Wargotz ES, Norris H,l: Metaplastic carcinomas of the breast. I. Matrix-producing carcinoma. HUM P.vrHoL ?O:f528-6%. 1989 2. Wargotz ES. l)eos PH. Norris H,J: Metaplastic carcinomas
of the breast. II. Spindle cell carcinoma. Ht1.w PATHOI. %):7:32-740. 19X9 3. Wargotz ES, Norris FiJ: Sletaplastic carcinomas of the breast. 111. Carcinosarcoma. Cancer 64:1490-1199. 1989 4. Lecene P: Les tumeurs mixtes du sein. Re\ (:hir (Chir] 33:434-439. 1906 .5. Leroux R: Reaction giganto cellulaire du stnmla dam un epithelioma mammaire. Hull Cancer ?0:6924i97. 193 1 6. I)uhoucher H. Montpellier J, Laffargue F: E:pithelioma mammaire avcc reaction de tk pe myeloplaxique. Ann Anat Pathol 70:787-79 I. 193:3 7. Nielsen BB, Kiaer HLV: Carcinoma of the breast with stroma multinucleated giant cells. Histopathology 9: 1X3- 193. 198.3 8. Kobayashi S. ‘Tobioka N. Samoto T. et al: Breast cancer with osteoclast like multinucleated giant cells. Acta l’athol ,]pn 34:147.5-14x1. 19x4 9. Holland R. Llrhain I(;. tan Haelst hl: hlammarp carcmoma with osteoclast like giant cells. Additional observations on six easer. Cancer .53:lW-1973, 1984 10. Huvos A(;. Lucas JC, Foote FW: Metaplastic breast carcinoma. NY State J Xled 73: 107X-10X2. 1973 1 I. Fisher ER, Palekar AS, (iregorio KM. et al: Mucoepidermoid and squamous cell carcinomas of breast with reference to squamous metaplasia and giant cell tumors. ArnJ Surg Pathol 7: I.5 “7, 1983 l?. Factor FM, Biempica 1.. Katner 1. et al: (:arcinoma of the bt-east with multinucleated reactive stromal giant cells. A light and electron microscopic study of two cases. Virchow’s Arch [A] 374: l12. 1!,77 13. Gras M, Zafiridou H, Karaiossifides (:: Unusual stl-omal response in a malignant breast ttlnlor. Presented at the \‘I Congress of the European Societb of Pathology. 1977
II. Oberman HA: Metaplastic carcinonla of the breast. A clinicopathologic studs of 29 patients. Am J Surg Pathol 1 I :918-929. 1987 1.i. Tavassoli FA. Norris H,] : Breast carcinoma with osteoclastlike giant cells. Arc-h Pathol Lab Med 1 lO:tX-639. 1986 16. Agnantis N?‘, Rosen PP: Mammary,carcinoma with ostt‘oclast-like giant cells. A study of eight cases wnh follow-up data. Am J Clin Pathol 72:3X3-3X9. 1979 17. Wargotz ES. Norris H,J: hletaplastic tarcinomas of the bl-east. IV. Squamous cell carcinoma. Cancel- 65:272-276. 1991) 1X. (;imbrone MA Jr. (;ullino PM: Neovasc~llar-ization induced hv intraocular xenografts of normal, preneoplastic and neoplastic mouse mammar> tissues. JN(:I 56:305-318. 1976 19. (Cullino PM: Natural history of breast cancer. Progression from hyperplasia to neoplasia as predicted by angiogenesis. (:ancer XJ:2697-“703 I . 1977 20. Brem SS. ]ensrn HM. (;ullino l’hl: Angiogenesis as a marker of preneoplastic lesions of the human breast. (;anc-er -1I :“39-“44 197X -.
1150