[CANCER RESEARCH 50. 7057-7061. November I. 1990]

Immunocytochemical Localization of Estrogen and Progesterone Receptor and

Prognosis in Human Primary Breast CancerAngelika Reiner,1 Birgitt Neumeister, JürgenSpona, Georg Reiner, Michael Schemper, and Raimund Jakesz

Institute of Clinical Pathology [A. K., B. N.J, First University Clinic of Surgery [G. R., M. S., R. J.J, First University Clinic ofGynecology [J. S.J, University of I'ienna,A-1090 Vienna, Austria

ABSTRACT

An immunocytochemical assay (ICA) for the measurement of estrogenreceptor (ER) and progesterone receptor (PgR) has been evaluated in426 human primary breast carcinomas. For estrogen receptor determination ER ICA was used. PgR ICA was performed using the monoclonalantibody KD 68. Assay results for progesterone receptor immunocyto-chemistry were in agreement (/' < 0.0001) with those of biochemical

determination in 74%. Progesterone receptor positivity determined with

a semiquantified approach based on intensity and heterogeneity of immunocytochemical staining correlated significantly with biochemicallydetermined progesterone receptor levels (/' = 0.0001). Survival data

showed a significantly better overall survival for patients with either ERICA- or PgR ICA-positive carcinomas (ER ICA, P < 0.00001; PgR ICA,

P = 0.004). Patients with both negative ER ICA and PgR ICA showed a

poorer prognosis than patients with only one negative receptor. In ERICA- and PgR ICA-positive carcinomas a trend could be found that

patients whose carcinomas contained high numbers of receptor-positive

tumor cells had a better survival. This study demonstrates that ER ICAand PgR ICA are strong prognostic indicators and that the proportion ofsteroid hormone receptor-positive tumor cells seems to be of clinical

importance.

INTRODUCTION

Steroid hormone receptor status is known to be an importantpredictor of prognosis and response to endocrine therapy inbreast cancer patients. In the past it was determined by biochemical assay. This assay is associated with certain methodological difficulties. The major limitations are that a certainamount of tissue is required and that the tissue needs to behomogenized for extraction of the receptor prior to the assay.Therefore, the biochemical assay cannot be performed in verysmall tumors and intratumoral variations of steroid hormonereceptor cannot be determined. These problems were overcomeby the development of monoclonal antibodies against steroidhormone receptors and immunohistochemical assay (1-3). Although there exist reports from different centers about correlations of immunocytochemical assay of ER2 with biochemical

techniques (4-9), and a few reports about the more recentlydeveloped immunocytochemical assay for PgR (10-12), onlylittle is known about the clinical significance of the immunocytochemical assays (10, 13, 14).

Therefore, it was the aim of our study to describe how therecently developed PgR ICA works in our hands and moreimportantly to examine ER ICA and PgR ICA as prognosticindicators.

Received 11/28/89; accepted 7/20/90.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1To whom requests for reprints should be addressed, at Institute of ClinicalPathology. University of Vienna, Spitalgasse 4, A-1090 Vienna, Austria.

2The abbreviations used are: ER. estrogen receptor; PgR, progesterone receptor; ICA, immunocytochemical assay; DCC, dextran-coated charcoal assay.

MATERIALS AND METHODS

Patients

Tumor samples were obtained from 426 consecutive patients withprimary breast cancer. Six % of the patients were younger than 40years, 21 % were between 40 and 50 years, and 73% were more than 50years. Two hundred, thirty-two patients (54%) had a modified radicalmastectomy and in 174 cases (41%) a breast-conserving procedure waspossible, mainly combined with an axillary dissection. In 15 patientswith high operative risk a simple mastectomy was performed. Fivepatients with large tumors (T3, T4) had just a tumor biopsy prior tosystemic therapy and underwent modified radical mastectomy thereafter. The distribution of receptor-positive and receptor-negative patients was equal within the different operations. All patients withconservative surgery received radiation therapy to the remaining ipsi-lateral breast and an electron boost or interstitial radiotherapy with192Irto the tumor site. The administration of adjuvant systemic treat

ment depended on eligibility to enter different adjuvant trials on operable breast cancer or on hormone receptor status and nodal status.Thus, 79 patients received adjuvant polychemotherapy, 133 patientshad adjuvant endocrine therapy (tamoxifen), 25 had a combination ofpolychemotherapy and hormonal therapy, and 189 had no adjuvantsystemic therapy. In summary, most of the patients with hormonereceptor-positive tumors were treated with tamoxifen alone or combined with polychemotherapy. Only 29 patients with ER ICA-positivetumors (10%) and 50 patients with ER ICA-negative tumors (37%) hadpolychemotherapy alone. Patients without systemic adjuvant treatment(n = 189) had ER ICA-negative tumors (40%) and ER ICA-positivetumors (60%). The clinical follow-up was conducted at 3-month intervals for 36 months and at 6-month intervals thereafter. The patientswere followed up to 62 months (median 36 months).

On pathological examination tumors were smaller than 2 cm in 62%and 2 cm or larger in 38% of the cases. Lymph node métastaseswerefound in 48% of patients. Histological tumor types were invasive ductalcarcinoma NOS in 71%, invasive lobular carcinoma in 13%, lobulo-ductal carcinoma in 4%, and mucoid, medullary, and ductal in situcarcinoma in 2% each of the cases. The remaining carcinomas consistedof rare tumor types as tubular, papillary, or comedo carcinoma. Histological tumor grade according to the World Health Organization wasclassified in 25% as highly differentiated, in 45% as intermediate, andin 30% as poorly differentiated.

Receptor Analysis

Immediately after surgery tumor specimens were trimmed of adjacentfat and connective tissue.

Biochemical Receptor Analysis. For biochemical analysis specimenswere quick-frozen in liquid nitrogen and ER and PgR measured usingthe DCC assay and Scatchard analysis. ER and PgR values >10 fmol/mg protein were considered positive, and values <10 fmol/mg proteinwere considered negative.

Immunocytochemical Assay. For immunocytochemistry tumor samples were either quick-frozen in liquid nitrogen and stored in anUltrafreeze unit at -70°C for as many as 4 years or were frozenimmediately in a Cryocut unit at -20°C using Tissue-Tek (Miles

Laboratories, Inc., Elkhart, IN) and processed thereafter. All tumorsamples were sectioned at 5 ^m thickness at -20°C, thaw mounted

onto glass slides, and fixed in 3.7% phosphate-buffered formaldehydefor 10 min. All further processing for ER ICA was performed accordingto the instructions in the Abbott ER ICA Monoclonal Kit (Abbott

7057

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ER ICA AND PgR ICA AND PROGNOSIS IN BREAST CANCER

Laboratories, Diagnostics Division, North Chicago, IL). For PgR ICAprediluted rat monoclonal antibody against human PgR (KD 68) wasprovided by Abbott Laboratories. The immunohistochemical methodused was the same as for ER ICA. In summary, it consisted of thesequential application of monoclonal anti-PgR antibody for 30 min,followed by incubations with bridging antibody and peroxidase-antiper-oxidase complex for 30 min each. All incubations were performed atroom temperature. The sites of immunoprecipitate formation wereidentified by diaminobenzidine. All reagents used for PgR ICA wereexactly the same as provided with the ER ICA kit. Controls consistedof ER- and PgR-positive cells as provided with the kit and treated witheither anti-ER or anti-PgR antibody and control antibody, respectively.In addition a series of tumor specimens was treated with the primaryand control antibodies. Specimens showing positive results with theprimary antibodies failed to show positive reaction with the controlantibody. Tumor specimens were classified as ER ICA or PgR ICApositive when positively stained cells could be identified. In cases ofpositive staining further parameters analyzed were intensity and heterogeneity of staining. The intensity of staining was subjectively evaluatedas 1, 2, or 3, with 1 representing the weakest and 3 the most intensestaining. When some variability in staining intensity within a tumorwas noticed, the average degree of intensity was considered representative. Heterogeneity represented the proportion of positively stainedtumor cells. It was evaluated as 1,2, 3, or 4 by estimation on screeningwide areas within each tissue section: 1, <10% stained cells; 2, <30%;3, <70%; and 4, >70% stained cells, compared with the total of tumorcells. Results of determination of intensity and heterogeneity werecombined and evaluated as follows: 2 and 3 points, low; 4 and 5 points,intermediate; and 6 and 7 points, high positively of ER ICA or PgRICA. All cases with negative staining were scored as 0.

Statistical Analysis

Overall survival and disease-free survival was described by Kaplan-Meier (15) estimates, using the programs of the BMDP statistic package. Possible prognostic differences of groups were analyzed by the logrank test (16) and the generalized Wilcoxon test (17). The proportionalhazards model by Cox ( 18) was used to investigate the prognostic effectof receptor measurements while simultaneously adjusting for otherfactors of assumed prognostic importance. The prognostic strength ofa receptor is described by an estimate of the relative risk of dying withinshort time intervals for "negative" versus "positive" patients. The

probability that such or greater relative risk estimate differs from 1only by chance is shown by a two-sided P value.

For DCC dichotomized at 10 fmol/mg, association with PgR ICA(positive versus negative) was tested by standard x2 analysis. For DCC

continuous, monotone association with PgR ICA (O, +, ++, +++) wasconfined by the test for Kendall's T (19). The hypothesis of monotone

association of heterogeneity and intensity of staining with DCC wasexamined similarly.

RESULTS

Localization of Progesterone Receptor in Breast Cancer Specimens. Positive staining with PgR ICA was localized to thenuclei of breast cancer cells in all histológica! specimens (Fig.1). Heterogeneity of staining was observed frequently. No tumorspecimen presented with 100% stained cells. PgR ICA waspositive in 58% of the carcinomas.

Relationship between PgR ICA and Biochemically DeterminedPgR Status. A statistically significant correlation (P < 0.0001)was observed between PgR ICA and biochemically determinedPgR status. Results for both methods agreed in 74% of thecases. In only 4% of the cases PgR ICA was negative, comparedwith DCC. In 22% of the cases PgR ICA was positive, whereasPgR was negative in DCC (Table 1). The relationship betweenPgR status as determined by DCC and semiquantified PgRICA is shown in Fig. 2. There was a significant correlation of

l s

Fig. 1. Immunocylochemical localization of PgR in an infiltrating ductalcarcinoma. A, histológica! specimen treated with PgR ICA (x 220). B, negativecontrol treated with control antibody (x 220).

Table 1 Comparison between immunocytochemical and biochemical PgR statusn = 349, />< 0.0001.

DCC

PgR ICA Negative Positive

NegativePositive

132(38)°

76(22)15(4)

126(36)1Number (%) of tumors.

100

to

M

í TO

Z (0

SO

1 ,0

0 10 20 SO 100 200 500 1000 2000

PgR fmol/fflq

Fig. 2. Comparison of PgR values determined by DCC and scmiquantifiedPgR ICA.

PgR levels as determined by both methods (P = 0.0001 ). AmongPgR ICA high tumors 73% had PgR values >50 fmol/mgprotein. In PgR ICA intermediate and low carcinomas PgRvalues >50 fmol occurred less frequently (64 and 43%, respectively). However, in all groups of PgR ICA-positive carcinomasa few biochemically PgR-negative carcinomas occurred, but thefrequency decreased with increasing PgR positivity (48, 30, and22%). In PgR ICA-negative carcinomas there occurred a fewbiochemically PgR-positive tumors, but PgR values never exceeded 200 fmol/mg protein.

Relationship between ER ICA and PgR ICA. Eighty-one % ofthe carcinomas showed concordant positive or negative resultsin immunocytochemical ER and PgR status (P< 0.0001 ) (Table

7058

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ER ICA AND PgR ICA AND PROGNOSIS IN BREAST CANCER

2). In 3% carcinomas were observed which were positive in PgRICA only. The majority of these carcinomas (7 of 10) showedonly weak PgR ICA positivity and 8 of 10 carcinomas presentedwith <30% PgR ICA-positive tumor cells. However, two carcinomas were observed with intermediate and one carcinomawith high PgR ICA positivity.

Association of ER ICA and PgR ICA with Clinical Follow-up.Women with either ER ICA-positive or PgR ICA-positivecarcinomas had a more favorable overall survival than womenwith steroid hormone receptor-negative carcinomas (Fig. 3).Regarding women with steroid hormone receptor-positive carcinomas there seemed to be no clear difference in survivaldepending on the grade of positivity, although patients with ERICA high and PgR ICA high carcinomas had the best overallsurvival (97% for each receptor). Overall, patients with ERICA-positive tumors are alive in 93% and those with PgR ICA-positive tumors in 91% of the cases. Patients with ER ICA-negative or PgR ICA-negative carcinomas had a significantlyworse prognosis (75% for each receptor). Survival data for bothsteroid hormone receptors in immunohistochemistry are demonstrated in Fig 4. Patients with both negative ER ICA andPgR ICA show a poorer prognosis (only 60% are alive) thanpatients with only one negative receptor. In contrast patientswith both ER ICA- and PgR ICA-positive carcinomas are alivein 90% of the cases. There was no additive effect in survival inpatients when both receptors were positive. In Fig. 5 overallsurvival of patients with receptor-positive carcinomas in rela-

1001

801

60% p < 0.00001n = 266

ER»PgR+ER* PgR-

ER - PgR »

ER- PgR-

" 6 12 18 21 30 36 42 18 months

Fig. 4. Kaplan-Meier estimated survival by steroid hormone receptors in ICA.

100X —I

801<30J

p = 0.08 , n.s.n = 291

12 18 21 30 36 12 48 months

Table 2 Comparison ofER-ICA and PgR-ICA

359, P < 0.0001.

PgR-ICA

ER-ICA Negative Positive

NegativePositive

93 (26)"

58(16)10(3)

198(55)1Number (%) of tumors.

1001

80S

60*

- ER-ICA <•++

ER-ICA +*

ER-ICA +

p<0.00001n - 126

ER-ICA -

42 48 months

1001

80S

•0.004' 286

-P9R-ICA'PgR-ICA

PgR-ICA -

48 months

BFig. 3. Kaplan-Meier estimated survival by ICA. A, ER ICA; B, PgR ICA.

1001

801

60»

>70»S70J

«301

<10X

p •=0.2 , n.s.n = 158

18 42 48 months

BFig. 5. Kaplan-Meier estimated survival in patients with steroid hormone

receptor-positive carcinomas in relation to the proportion of receptor-positivetumor cells. Percentages are the proportion of positive tumor cells. A, ER ICA;B, PgR ICA.

tion to the proportion of receptor-positive tumor cells is demonstrated. A trend could be found that patients whose carcinomas contained >30% receptor-positive tumor cells had a bettersurvival.

In Fig. 6 the relationship between survival of patients withreceptor-positive carcinomas and staining intensity is demonstrated. Although patients with carcinomas with high stainingintensity have the best prognosis, staining intensity alone doesnot discriminate satisfactory.

Various results on the prognostic effect of receptors obtainedfrom Cox analyses are summarized in Table 3. The results forER ICA and biochemically determined ER status are based on304 patients, and those for PgR ICA and biochemically determined PgR status are based on 212 patients. For comparabilitythe regression models and corresponding univariate analysiswere based on the same patients. The results from the regressionmodel are all adjusted for the following factors: histológica!tumor grade, axillary lymph node status, tumor size, conserva-

7059

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ER ICA AND PgR ICA AND PROGNOSIS IN BREAST CANCER

IMIn

801

6OT

nignintermedióte

0.5291

2«

1001

80ÃŽ

60t

-nign

•Intermedióte

0.3158

«8montns

BFig. 6. Kaplan-Meier estimated survival in patients with steroid hormone

receptor-positive carcinomas in relation to staining intensity. +, weak; ++.intermediate; +++, high staining intensity. A, ER ICA; fi, PgR ICA.

Table 3 Comparison of prognostic effect of receptors in immunocytochemical andbiochemical assays by Cox analysis

UnivariateanalysisReceptorER

ICAER-DCCPgRICA

PgR-DCCRelative

risk3.032.272.81

1.56P0.0001

0.00020.001

0.003Regression

modelRelative

risk4.08

2.812.01

2.90P0.001

0.010.047

0.28

live surgery or radical mastectomy, and adjuvant therapy beingeither none, chemoendocrine, or combined chemotherapy andhormone therapy. However, the receptors are not mutuallyadjusted in accordance with the aims of this project. As it isdemonstrated ER ICA and PgR ICA are prognosticates independent of other prognostic factors. Moreover, the lvalues forER ICA and PgR ICA are of higher statistical significance thanthose for the biochemical assay.

DISCUSSION

Monoclonal antibodies to ER and PgR were used to localizesteroid hormone receptors in histological breast cancer samples.PgR localization was always limited to the cell nucleus. This isin agreement with other reports on PgR immunocytochemistry(3,10, 20) and with the concept of intranuclear steroid hormonereceptor localization. Intranuclear localization of ER was already reported in many previous studies (1,2, 4-9). Similar toER ICA we could demonstrate a significant correlation betweenPgR ICA and biochemically determined PgR status. Overallagreement of PgR positivity or negativity by both methods was74%. This result is comparable to the data of others (10).However, there was some disagreement mainly in cases which

were PgR ICA positive but biochemically PgR negative. Thisfinding is also similar to ER ICA (9). Partly, this may be dueto the fact that in our series of carcinomas only 40% werebiochemically PgR positive. Although this percentage of PgRpositivity is within the range of different reports (21-23), thereexist reports with considerably higher frequencies of PgR-positive carcinomas (10, 24). Other reasons for discrepantresults between immunohistochemistry and biochemical receptor determination may be prolonged specimen storage, alterations of the epitope during tissue fixation, and differencesbetween the biochemical and immunocytochemical steroid hormone receptor assays themselves.

As is known for biochemically determined steroid hormonereceptors there exists an association between ER and PgRstatus. Only 19% of the carcinomas were positive for onereceptor only. The majority was positive for ER alone and in3% only PgR positivity was found. This percentage is lowerthan for biochemical assays in which approximately 10% of thecarcinomas are ER negative but PgR positive. The majority ofthese carcinomas showed only weak PgR ICA positivity. However, there was one case with high PgR ICA positivity. Anexplanation for this result could be that in rare cases a dissociation of the function of ER and PgR exists in tumor cells. Thiswas suggested for mammary cancer cell lines (25). It is alsosupported by simultaneous immunocytochemical demonstration of ER and PgR within the same section of breast cancer,where cells exist which coexpress both receptors or only eitherone receptor (26).

For semiquantification of steroid hormone receptor immunocytochemistry we used a simple score based on heterogeneityand staining intensity. As in a former report for ER ICA (9) weare now able to demonstrate that this score seems to work quitewell also for PgR ICA. This score is simple and easy to performand therefore can be applied by any competent pathologist.

Immunocytochemical analysis of both ER and PgR was foundto predict overall survival. This was true for univariate and alsofor multivariate analysis when receptor immunocytochemistryprognostic power was kept independently of all other importantprognostic parameters. A similar result for ER ICA on a limitednumber of patients is found in the literature (13). In this studyER ICA was related independently to prognosis of lymph nodestatus and treatment. Moreover, in our study the comparisonof the prognostic significance of immunocytochemically andbiochemically determined steroid hormone receptor statusshowed that the P values were of higher significance for immunocytochemistry than for biochemical analysis. This suggests that the immunocytochemical assay is of higher discriminating power. This is in agreement with the result in anotherstudy (14). In that study, ER ICA predicted survival and disease-free survival for the whole observation time of 10 years, whilebiochemically determined ER status lost its predictive valueafter 5 years. Moreover, we found differences in survival forboth ER ICA and PgR ICA, depending on the grade of positivity. The differences between the different grades of positivitywere not statistically significant. But the observation time forsteroid hormone receptor-positive carcinomas in our study wasrather short and therefore it seems possible that differences canbe demonstrated more clearly at a later date. Therefore, weassume that the semiquantitative evaluation of steroid hormonereceptor immunocytochemistry is of clinical importance. Thisagrees with some reports in the literature in which a correlationbetween biochemical steroid hormone receptor values and survival was described (27).

7060

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ER ICA AND PgR ICA AND PROGNOSIS IN BREAST CANCER

Immunocytochemistry also gives information about the heterogeneity of steroid hormone receptor distribution. As oursurvival data suggest this seems to be of clinical importance,since patients whose carcinomas contain high levels of steroidhormone receptor-positive cells (>30%) show a better overallsurvival. This agrees with the results that high rates of recurrences occur in patients whose tumors contain high proportionsof ER-negative tumor cells (28). On the basis of this it istempting to speculate that in the future patients with only afew steroid hormone receptor-positive, but a large proportionof negative tumor cells may be selected for adjuvant combinedendocrine and chemotherapy.

One can conclude that steroid hormone receptor immuno-

cytochemistry adds important information to the currently usedbiochemical steroid hormone receptor assay. The most important findings are that both ER ICA and PgR ICA are strongprognosticators for overall survival and that the proportion ofsteroid hormone receptor-positive tumor cells seems to be ofclinical importance. Although our results are encouraging, onehas to keep in mind that immunocytochemistry detects thereceptors but does not assure that the receptors are functional.

REFERENCES

1. Greene, G. L., Sobel, N. B., King, W. J., and Jensen, E. V. Immunochemicalstudies of estrogen receptors. J. Steroid Biochem., 20: 51-56, 1984.

2. King, W. J., and Greene, G. L. Monoclonal antibodies localize oestrogenreceptor in the nuclei of target cells. Nature (London), 307: 745-747, 1984.

3. Press, M. F., Udove, J. A., and Greene, G. L. Progesterone receptor distribution in the human endometrium. Analysis using monoclonal antibodies tothe human progesterone receptor. Am. J. Pathol., 131: 112-124, 1988.

4. McCarty, K. S., Jr., Miller, L. S., Cox, E. B., Konrath, J.. and McCarty. K.S., Sr. Estrogen receptor analysis. Correlation of biochemical and immuno-histochemical methods using monoclonal antireceptor antibodies. Arch. Pathol. Lab. Med., 709: 716-721, 1985.

5. King, W. J.. De Sombre, E. R., Jensen, E. V., and Greene, G. L. Comparisonof immunocytochemical and steroid-binding assays for estrogen receptor inhuman breast tumors. Cancer Res., 45: 293-300. 1985.

6. Hawkins, R. A., Sangster, K., and Krajewski, A. Histochemical detection ofoestrogen receptors in breast carcinoma: a successful technique. Br. J. Cancer,55:407-410. 1986.

7. Pertschuk. L. P.. Eisenberg. K. B., Carter, A. C., and Feldman, J. G.Immunohistologic localization of estrogen receptors in breast cancer withmonoclonal antibodies: correlation with biochemistry and clinical endocrineresponse. Cancer (Philadelphia), 55: 1513-1518, 1985.

8. Di Fronzo, G., Clemente. C., Capelletti, V.. Miodini. P.. Coradini, D..Ronchi, E., Andreola, S., and Rilke, F. Relationship between ER ICA andconventional steroid receptor assays in human breast cancer. Breast CancerRes. Treat., S: 35-43, 1986.

9. Reiner, A., Spona, J., Reiner, G., Schemper, M., Kolb, R.. Kwasny, W..Függer,R.. Jakesz. R., and Holzner, J. H. Estrogen receptor analysis onbiopsies and fine-needle aspirates from human breast carcinoma. Correlationof biochemical and immunohistochemical methods using monoclonal anti-

receptor antibodies. Am. J. Pathol., 125: 443-449, 1986.10. Pertschuk, L. P., Feldman. J. G., Eisenberg, K. B., Carter, A. C.. Thelmo,

W. L., Cruz, W. P., Thorpe, S. M., Christensen, J., Rasmussen. B. B., Rose,C., and Greene, G. L. Immunocytochemical detection of progesterone receptor in breast cancer with monoclonal antibody. Relation to biochemical assay,disease-free survival, and clinical endocrine response. Cancer (Philadelphia),«2:342-349, 1988.

11. Elashry-Stowers, D., Zava, D. T., Speers, W. C., and Edwards, D. P.Immunocytochemical localization of progesterone receptors in breast cancerwith anti-human receptor monoclonal antibodies. Cancer Res., 48: 6462-6474, 1988.

12. Charpin, C., Jacquemier. J., Andrac, L., Vacherei, H., Habib, M. C., Devic-tor, B., Lavaut, M. N., and Toga, M. Multiparametric analysis (SAMBA200) of the progesterone receptor immunocytochemical assay in nonmalig-nant and malignant breast disorders. Am. J. Pathol., 132: 199-211, 1988.

13. DeSombre, E. R., Thorpe, S. M., Rose, C., Blough, R. R., Andersen, K. W.,Rasmussen, B. B., and King, W. J. Prognostic usefulness of estrogen receptorimmunocytochemical assay for human breast cancer. Cancer Res., 46(Suppl.): 4256s-4264s, 1986.

14. Kinsel, L. B., Szabo, E., Greene, G. L., Konrath, J., Leight, G. S., andMcCarty, K. S., Jr. Immunocytochemical analysis of estrogen receptors as apredictor of prognosis in breast cancer patients: comparison with quantitativebiochemical methods. Cancer Res., 49: 1052-1056, 1989.

15. Kaplan, E. L., and Meier, P. Nonparametric estimation from incompleteobservations. J. Am. Statist. Assoc.. 53: 457-481, 1958.

16. Mantel, N. Evaluation of survival data and two new rank order statisticsarising in its consideration. Cancer Chemother. Rep., 50: 163-170. 1966.

17. Breslow, N. A generalized Kruskal-Wallis test for comparing K samplessubject to unequal patterns of censorship. Biometrika, 57: 579-594, 1970.

18. Cox. D. R. Regression models and life tables (with discussion). J. R. Statist.Soc. B, 34: 187-220, 1972.

19. Kendall, M. G. Rank Correlation Methods, Ed. 3. London, Griffin, 1982.20. Lessey, B. A., Killam, A. P., Metzger. D. A., Haney, A. F., Greene, G. L.,

and McCarty, K. S., Jr. Immunohistochemical analysis of human uterineestrogen and progesterone receptors throughout the menstrual cycle. J. Clin.Endocrinol. Metab., 67: 334-339, 1988.

21. Saez, S., Cheix, F., and Asselain, B. Prognostic value of estrogen andprogesterone receptors in primary breast cancer. Breast Cancer Res. Treat.,5:345-354,1983.

22. Clark, G. M., and McGuire, W. L. Progesterone receptors and human breastcancer. Breast Cancer Res. Treat., 3: 157-163, 1983.

23. Martin, P. M., Rolland, P. H.. Jacquemier. J., Rolland, A. M., and Toga,M. Multiple steroid receptors in human breast cancer II. Estrogen andprogestin receptors in 672 primary tumors. Cancer Chemother. Pharmacol.,2: 107-113, 1979.

24. Thorpe. S. M., and Rose, C. Oestrogen and progesterone receptor determinations in breast cancer. Technology and biology. Cancer Surv., 5: 505-525,1986.

25. Reiner, G. C. A., and Katzenellenbogen, B. S. Characterization of estrogenand progesterone receptors and the dissociated regulation of growth andprogesterone receptor stimulation by estrogen in MDA-MB-134 humanbreast cancer cells. Cancer Res., 46: 1124-1131, 1986.

26. Bonkhoff, H., and Wernert, N. Immunohistochemical analysis of the topographical relationship between the estrogen- and progesterone receptor infive human breast cancers. A simultaneous demonstration of both nuclearreceptors in the same section by using a computer-assisted image processing.Virchows Arch. A Pathol. Anal., 414: 339-344, 1989.

27. Clark, G. M., McGuire, W. L., Hubay. C. A.. Pearson. O. H., and Marshall,J. S. Progesterone receptors as a prognostic factor in stage II breast cancer.N. Engl. J. Med.. 509: 1343-1347, 1983.

28. Walker, K. J., Bouzubar, N., Robertson, J., Ellis, I. O., Elston, C. W.,Blarney, R. W., Wilson. D. W.. Griffiths, K.. and Nicholson, R. I. Immunocytochemical localization of estrogen receptor in human breast tissue. CancerRes.. 48: 6517-6522. 1988.

7061

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1990;50:7057-7061. Cancer Res   Angelika Reiner, Birgitt Neumeister, Jürgen Spona, et al.   Cancer

BreastProgesterone Receptor and Prognosis in Human Primary Immunocytochemical Localization of Estrogen and

  Updated version

  http://cancerres.aacrjournals.org/content/50/21/7057

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/50/21/7057To request permission to re-use all or part of this article, use this link

on June 11, 2018. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from