Immunoreactivity ofProstate-Specific Antigen in MaleBreast CarcinomasTwo Examples of a Diagnostic Pitfall inDiscriminating a Primary Breast Cancerfrom Metastatic Prostate CarcinomaRaj K. Gupta, M.D., F.I.A.C.*

Prostatic-specific antigen (PSA) is regarded as a specific markersecreted by normal and neoplastic acinar epithelial cells of theprostate gland; its detection by immunocytochemistry has beenaccepted as an indication of metastatic prostate cancer. This isascribed to the commonly held belief that PSA is not found inextraprostatic tissues. However, this concept has recently beenchallenged, based on the observations that certain nonprostatictissues and their neoplasms can also secrete PSA. Such a question-able belief could result in a diagnostic pitfall when using immuno-staining for PSA on fine-needle aspiration (FNAC) cytologysamples to differentiate metastatic prostate cancer from a primarycarcinoma of an extraprostatic organ. In this communication, twocases of primary carcinomas of the male breast are reported inwhich PSA immunopositivity on FNAC led to the suggestion of adiagnosis of metastatic carcinoma of the prostate.Diagn. Cytopa-thol. 1999;21:167–169. r 1999 Wiley-Liss, Inc.

Key Words:fine-needle aspiration cytology; breast; carcinoma; prostate-specific antigen; metastatic prostate cancer; immunostaining

Prostatic-specific antigen (PSA) has been regarded as aspecific marker of both normal and neoplastic prostaticepithelium; its detection by immunoperoxidase has beenwidely accepted as a confirmation of the prostatic origin ofmetastatic prostate cancer.1–3 This belief is substantiated bythe commonly held arguments that PSA immunostainingshould be done as a standard investigation, especially inmale patients suspected of a metastatic carcinoma of un-known origin and with a known cancer of the prostatepresenting with suspected metastases at unusual sites.2

Recently, the commonly held belief that immunostainingof PSA is specific for prostate epithelium has been ques-

tioned, due to its demonstration in nonprostatic tissues4,5 andnonprostatic neoplasms.5–12 Therefore, it seems reasonableto assume that expression of PSA is not restricted to the prostategland, and its demonstration in other neoplasms cannot beconsidered as an unequivocal evidence of prostatic origin.

In this communication, two cases of primary male breastductal cancers are reported, which were diagnosed byfine-needle aspiration cytology (FNAC) and which wereconfounded by a positive staining for PSA, thus creating adiagnostic pitfall of discriminating a primary breast carci-noma from a metastatic prostate cancer.

Materials and MethodsThe aspirate was obtained using a disposable 10-ml syringeand 22-gauge disposable needle using multiple passes in themass, maintaining negative pressure. The aspirated materialwas collected as needle and syringe washings in 30% ethylalcohol in physiologic saline. From half of this material,filter preparations were made on 3-µm size Gelman filtersand stained by a modified Papanicolaou method. Theremainder of needle and syringe washings were centrifuged,and cell blocks were made and fixed, embedded, cut at 5 µm,and stained with hematoxylin-eosin. No smears were madefrom aspirated material.

The excised breast mass and the mastectomy specimen,including the lymph nodes, were fixed, sectioned, processed,cut at 5 µm, and stained with hematoxylin-eosin. Sections ofcell block of the aspirate and excised tumor tissue werefurther utilized for immunostaining for polyclonal PSA,PSAP (prostatic-specific acid phosphatase), B72.3, andcarcinoembryonic antigen (CEA) by the immunoperoxidasemethod, using commercially purchased kits. Known positiveand negative controls were always used during the immuno-staining procedure.

Cytology Unit, Wellington Hospital and School of Medicine, Wellington,and Valley Diagnostic Laboratories, Ltd., Lower Hutt, New Zealand

*Correspondence to: Raj K. Gupta, M.D., F.I.A.C., Cytology Unit,Wellington Hospital, Wellington, New Zealand.

Received 1 December 1998; Accepted 6 April 1999

r 1999 WILEY-LISS, INC. Diagnostic Cytopathology, Vol 21, No 3 167

The clinical information in both cases was obtained byreviewing the charts and asking clinicians for pertinentdetails and outcomes of laboratory tests and other investiga-tions. These are summarized in Table I.

ResultsCytohistologic and Immunoperoxidase FindingsPapanicolaou-stained filter preparations and hematoxylin-eosin-stained sections of cell block and breast mass showedbland groups of ductal cells with fine chromatin, smallnucleoli, minimal atypism, and little variation in size andshape (Figs. 1, 2). Sections of the mastectomy specimenshowed no further tumor in the breast tissue or lymph nodes.Immunostaining for PSA was positive in cell blocks of theaspirate in both cases (Fig. 3), while in tumor tissue it wasfound in traces. The PSAP was negative, while B72.3 wasfound in traces and CEA was positive (Table I).

In both cases the laboratory investigations were withinnormal range, except for a mildly elevated serum PSA of 0.4ng/ml (case 1) and 0.3 ng/ml (case 2) (normal range,,0.2ng/ml). Also, repeated chest X-ray, abdominal CT, and boneimaging studies were found to be normal. The serum PSAlevel remained stable following the mastectomy, and both

patients are alive and well without any residual or recurrenttumor.

DiscussionIn practice, immunostaining for PSA in specimens of tumorhas traditionally been regarded as useful in identification ofprostatic carcinomas.1,2 Also, this belief is generally ac-

Table I. Summary of Findings in the Two Cases of Breast Carcinomaa

CaseAge (yr)and sex Clinical findings Clinical diagnosis FNA diagnosis

Histologicconfirmation

Immunostainingresults

1 80, M 23 1.5 cm slightly firm mobilemass, of a few months’ dura-tion in left upper portion of thebreast. On and off urinaryincontinence for a few months.

Suspicious for carcinoma (pri-mary?) (metastatic?)

Carcinoma (metastatic prostate?) C and B PSA1 (C)PSA T (B)PASP2 (C and B)B72.3 T (C and B)CEA1 (C and B)

2 76, M 33 2.5 cm gritty mass aboveright areola, of 3 weeks’ dura-tion. On and off urinary incon-tinence for several months.

Carcinoma (primary?) (meta-static?)

Carcinoma (metastatic prostate?) C and B PSA1 (C)PSA T (B)PSAP2 (C and B)B72.3 T (C and B)CEA1 (C and B)

aC, cell block of aspirate; B, breast mass; T, trace positive;2, negative;1, positive.

Fig. 1. Papanicolaou-stained filter preparation, showing ductal groups ofmalignant cells of breast carcinoma (3250). Fig. 2. The findings in Figure 1 are shown histologically in a hematoxylin-

eosin stained section of the cell block from the aspirate (3250).

Fig. 3. Section of cell block from aspirate, showing immunopositivity forPSA (3400).

GUPTA

168 Diagnostic Cytopathology, Vol 21, No 3

cepted in male patients with metastatic carcinoma of un-known origin, and those with a known prostate cancer whodevelop metastases at unusual sites. Until recently, it wasbelieved that PSA is not present in other tissues, except foroccasional cells in teratomas13 and in renal tubules andrenal-cell carcinomas.14 However, some studies have chal-lenged this concept, since it has been shown that PSApositivity in some tumors of nonprostatic origin and ex-traprostatic tissue can be nonspecific and may pose adiagnostic pitfall.4,5,11,12Some examples of such nonspecificstaining have been seen in apocrine sweat gland tumors ofthe skin and breast, oncocytoma of the parotid gland,carcinomas of the pancreas, urinary bladder, ovary, andparaurethral glands, pleomorphic adenomas, Warthin’s tu-mors, mucoepidermoid carcinomas, and adenoid cysticcarcinomas of the salivary gland.5–12

Certainly, the two male patients who presented in thisstudy with malignant breast tumor and PSA positivity didcreate such a dilemma. Also, since in both cases, apart frombreast mass, symptoms of on-and-off urinary incontinencewere present. However, in neither of them was a significantelevation of PSA found biochemically nor did any otherinvestigations point toward a significant pathology in theprostate gland. To our knowledge, this is the first report ofprimary male breast ductal carcinomas in which PSApositivity suggested a possible metastasis from the prostate,although in none of the patients was a primary found in theprostate. In a recent study,5 the few cases of primarycarcinoma in the breast in which PSA positivity was shownwere of the apocrine type. In the same study5 it wassuggested that the detection of PSA in apocrine cells andsweat gland and breast apocrine tumors may be ascribed tosimilarities or sequential homologies with unknown proteinsin the apocrine cells.

In conclusion, the results presented here suggest a cau-tious approach to PSA positivity as a differential marker fora primary vs. a metastatic carcinoma of the prostate, sincePSA positivity need not always indicate a prostatic origin.Also, it is felt that it is not necessary to routinely performPSA on every male malignant neoplasm to rule out ametastatic prostate carcinoma, since in most cases such ametastatic tumor is often associated with elevated serumPSA levels and will usually metastasize to the periaorticlymph nodes before involving supradiaphragmatic organs.

AcknowledgmentsThe author gratefully acknowledges the excellent secretarialassistance of Sharda Lallu in the preparation of this manu-script. The technical and photographic assistance of SarlaNaran and Robert Fauck is also acknowledged. The coopera-tion of various clinicians in providing clinical data isacknowledged.

References1. Katz RL, Raval P, Brooks TE, Ordonez NG. Role of immunocytochem-

istry in diagnosis of prostatic neoplasia by fine-needle aspirationbiopsy. Diagn Cytopathol 1985;1:28–32.

2. Tell DT, Khoury JM, Taylor HG, Veasey SP. Atypical metastasis fromprostate cancer: clinical utility of the immunoperoxidase technique forprostate-specific antigen. JAMA 1985;253:3574–3575.

3. Epstein JI, Cho KR. Metastatic prostatic carcinoma to supradiaphrag-matic lymph nodes: a clinicopathologic and immunohistochemicalstudy. Am J Surg Pathol 1987;11:457–463.

4. Elgamal AA, Ectors NL, Sunardhi-Widyaputra S, Van Poppel HP, VanDamme BJ, Baert LV. Detection of prostate specific antigen in pancreasand salivary glands: a potential impact on prostate cancer overestima-tion. J Urol 1996;156:464–468.

5. Papotti M, Paties C, Peveri V, Moscuzza L, Bussolati G. Immunocyto-chemical detection of prostate-specific antigen (PSA) in skin adnexaland breast tissues and tumors. Basic Appl Histochem 1989;33:25–29.

6. Yu H, Diamandis EP, Levesque M, Asa SL, Monne M, Croce CM.Expression of the prostate-specific antigen gene by a primary ovariancarcinoma. Cancer Res 1995;55:1603–1606.

7. Grignon DJ, Ro JY, Ayala AG, Johnson DE, Ordonez NG. Primaryadenocarcinoma of the urinary bladder: a clinicopathologic analysis of72 cases. Cancer 1991;67:2165–2172.

8. Zaviacic M, Sidlo J, Borovsky M. Prostate specific antigen and prostatespecific acid phosphatase in adenocarcinoma of Skene’s paraurethralglands and ducts. Virchows Arch [A] 1993;423:503–505.

9. Kuopio T, Ekfors TO, Nikkanen V, Nevalainen TJ. Acinar cellcarcinoma of the pancreas: a report of three cases. Acta PatholMicrobiol Immunol Scand 1995;103:69–78.

10. James GK, Pudek M, Berean KW, Diamandis EP, Archibald BL.Salivary duct carcinoma secreting prostate-specific antigen. Am J ClinPathol 1995;106:242–247.

11. Van Krieken JHJM. Prostate marker immunoreactivity in salivarygland neoplasms. Am J Surg Pathol 1993;17:410–414.

12. Holmes GF, Eisele DW, Rosenthal D, Westra WH. PSA immunoreactiv-ity in a parotid oncocytoma: a diagnostic pitfall in discriminatingprimary parotid neoplasms from metastatic prostate cancer. DiagnCytopathol 1998;19:221–225.

13. Taylor CR. Prostate, bladder, kidney, cervix, uterus. In: Taylor CR,editor. Immunomicroscopy: a diagnostic tool for surgical pathologist.Philadelphia: W.B. Saunders; 1986. p 260–282.

14. Franket AR, Rouse RV, Wang MC, Herzeuberg LA. Monoclonalantibodies to a human prostate antigen. Cancer Res 1982;42:3714–3718.

PSA STAINING IN BREAST CARCINOMAS

Diagnostic Cytopathology, Vol 21, No 3 169