2083 Estrogen Receptor Status, Determined by Immunohistochemistry, as a Predictor of the Recurrence of Stage I Endometrial Carcinoma Paola A. Gehrig, M.D.1 Linda Van Le, M.D.1 Babatunde Olatidoye, Joseph Geradts, M.D.2 BACKGROUND. The aim of this study was to compare the concordance between 2 M.D. 1 Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of North Carolina, Chapel Hill, North Carolina. 2 Department of Pathology and Laboratory Medicine, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina. Presented in part at the Society of Gynecologic Investigators meeting, Atlanta, Georgia, March 11– 14, 1999. Dr. Olatidoye’s current address: Department of Pathology, Box 3712 Duke University Medical Center, Durham, NC 27710. Dr. Geradts’ current address: Nuffield Department of Pathology and Bacteriology, University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom. Address for reprints: Linda Van Le, M.D., Division of Gynecologic Oncology, CB 7570 MacNider Building, University of North Carolina, Chapel Hill, NC 27599. Received March 1, 1999; revision received June 28, 1999; accepted June 28, 1999. © 1999 American Cancer Society immunohistochemical (IHC) and biochemical (RIA) methods for determining hormone receptor status in patients with endometrial carcinoma and to determine whether IHC expression of estrogen and progesterone receptors (ER and PR) has prognostic significance. METHODS. Paraffin blocks were obtained from patients diagnosed with endometrial carcinoma between 1987 and 1991. IHC analysis for ER and PR expression was performed and scored based on staining intensity and the percentage of tumor cells with nuclear staining. Biochemical assays were performed on frozen tissues. Concordance between the two methods was evaluated and hormone receptor status was correlated with tumor grade, stage, recurrence and survival. RESULTS. ER and PR expression, determined by IHC, correlated well with RIA levels (Spearmans correlation coefficient, P 5 0.006 and 0.0005, respectively). Determination of ER and PR expression by both methods was correlated with tumor grade. Hazards ratios revealed that the absence of ER and PR expression, determined by both IHC and RIA, independently correlated with recurrence in early stage disease (P , 0.05). CONCLUSIONS. Historically, receptors have been determined by RIA. In this study, IHC and RIA were equally suitable for determination of ER and PR. This is significant clinically as IHC has several advantages over RIA, including easier processing, lower cost, greater speed, and applicability to fixed tissue samples. In addition, ER negative status was predictive of the recurrence of Stage I tumors independent of tumor grade. ER status may aid the clinician in planning treatment when adjuvant treatment is controversial. Cancer 1999;86:2083–9. © 1999 American Cancer Society. KEYWORDS: endometrial carcinoma, estrogen receptors, progesterone receptors, immunohistochemistry. E ndometrial carcinoma is the most common gynecologic malignancy, accounting for 6% of all new cancer cases, with approximately 36,000 new cases diagnosed in the United States in 1997. Although endometrial carcinoma accounts for only 2% of all cancerrelated deaths, the number of women dying of this disease (4900 per year) still is significant.1 Although endometrial carcinoma has relatively good cure rates, 86% in Caucasians and 55% in African-Americans, there is motivation to determine which prognostic factors correlate with patient outcome. Well-described prognostic factors include stage of disease at the time of diagnosis, histologic type, degree of tumor differentiation, depth of myometrial invasion, and lymphovascular space invasion. 2084 CANCER November 15, 1999 / Volume 86 / Number 10 Although the importance of hormone receptors in the management of breast carcinoma patients is well established,2 the clinical significance of hormone receptor expression in endometrial carcinoma is less clear. Estrogen receptor (ER) status is believed to provide prognostic information independent of tumor stage and grade in women with endometrial carcinoma.3– 6 However, progesterone receptor (PR), and not ER, is suggested to be more predictive of disease free survival.6 –11 The methods by which hormone receptor status in patients with endometrial carcinoma have been evaluated are immunohistochemistry (IHC)4,5,10,12,13 and radioimmunoassay (RIA);6 –9,11,14 IHC has a benefit over RIA, in that hormone receptor status can be determined for archival tissues in contrast to RIA, which must be performed on tissues before histologic processing occurs. IHC has other advantages over RIA, including easier tissue handling, greater speed, lower cost, and wider availability. Several authors have shown that IHC can be as sensitive and specific as RIA;4,15–20 however, few have made the correlation between IHC methodology and disease free interval and survival.4,5,10 In addition, until recently, RIA was the preferred method for determining ER and PR status in gynecologic specimens. We therefore undertook this study to determine the concordance between IHC and RIA determination of hormone receptor status and the prognostic significance and clinical applicability of ER and PR status in patients with endometrial carcinoma as determined by IHC. MATERIALS AND METHODS Patients Of the surgically staged patients with endometrial carcinoma who were treated at the University of North Carolina Hospitals between 1987 and 1991, 72 patients were evaluated for the presence of estrogen and progesterone receptors. The patients were 72 consecutive cases in whom the attending physician requested receptor status by RIA. The 72 patient charts were reviewed for disease stage at the time of initial surgery, histologic tumor grade, and cell type in accordance with the International Federation of Gynecology and Obstetrics (FIGO) criteria. Those patients who were staged surgically before 1988 were reclassified according to the 1988 FIGO criteria. Charts also were reviewed for disease free interval, location of tumor recurrence, and survival. This study met the exemption criteria set forth by the Institutional Review Board at the University of North Carolina. RIA RIA analysis was performed by the SmithKline Beecham clinical laboratories. The method used was a dextran-coated, charcoal-scatchard analysis with tritium. Materials Monoclonal anti-ER antibody 1D5 was purchased from Dako Corporation (Carpinteria, CA). Monoclonal anti-PR antibody 1A6 was obtained from Novocastra via Vector Laboratories (Burlingame, CA). The secondary detection reagents, Elite ABC kit, were from Vector Laboratories. IHC The immunostains were performed on an automated stainer (Biotek 2000; Ventana Systems, Phoenix, AZ). The primary incubation time for both assays was 30 minutes after antigen retrieval in 0.01 M citrate buffer (20 minutes at 95–100°C). Anti-ER antibody 1D5 was used at a dilution of 1:20, and anti-PR antibody 1A6 was used at a dilution of 1:15. The detection reaction used the Vector Elite ABC kit (manufacturer-recommended protocol). Diaminobenzidine and hematoxylin counterstains were used for color development. Interpretation Although RIA was performed for 72 patients, only 67 tumors were evaluable for ER and PR by IHC. The two study pathologists (J. G. and B. A. O.), blindly and independently, evaluated hormone receptor expression by IHC. The average nuclear staining intensity was graded as 0 (absent), 1 (weak), 2 (moderate), and 3 (strong). The percentage of tumor cells with positive nuclear staining also was estimated. The immunohistochemical score was determined by multiplying the average staining intensity by the percentage of tumor cells stained.16 Thus, the theoretical range of scores was from 0 to 300. Significant differences in interpretation (for example, negative vs. positive or low positive vs. high positive scores) were resolved in conference. However, such differences in opinion only occurred in a very small subset of cases. In all other cases, the two observer scores were averaged. In preliminary analysis, the best agreement between IHC analysis and RIA analysis with regard to separating negative cases from positive cases was found when the cut-off values were set at an IHC score of 20 and an RIA value of 50 fmol/mg protein (ER, correlation coefficient [r] 5 0.63; PR, r 5 0.67). Statistical Analysis The correlations between tumor grade, disease stage, and ER and PR status, determined by IHC and RIA, ER Predicts Endometrial Carcinoma Recurrence/Gehrig et al. 2085 FIGURE 2. Concordant and discordant expression of hormone receptors by FIGURE 1. The American Joint Committee on Cancer (AJCC) Staging for Endometrial Carcinoma. were calculated using both the Kruskal–Wallis test and the Fisher exact test. Because the correlations were not linear, the degree of correlation between the IHC and RIA measures for both ER and PR was evaluated using Spearmans correlation coefficient. Finally, Cox proportional hazards regression models were used to evaluate the ability of ER and PR (with both IHC and RIA) to predict the time to disease recurrence and the time to death. Three models were used for the Cox proportional hazards regression: The first model included only ER and PR as variables, the second model included both ER and PR variables as well as a variable for grade, and the third model included a variable for disease stage and tumor grade in addition to the ER and PR variables. RESULTS Although biochemical receptor data were available for 72 patients, complete clinical data, including follow-up and IHC score, were available for only 55 patients. Because one of the main objectives was to compare the prognostic value of RIA hormone receptor levels with IHC ER and PR status, analysis was restricted to the 55 patients with complete data. Among the 17 patients who were excluded, 10 had Stage I disease, 4 had Stage II disease, and 3 had Stage III or IV disease, as determined by the American Joint Committee on Cancer staging system (Fig. 1). Of the 17 excluded patients, 2 are alive with disease, and 2 others are dead. Among the 55 patients who were evaluated, 36 patients had surgical Stage I disease (65.5%), 8 had Stage II disease (14.5%), and 11 had Stage III–IV disease (20%). When the patients were divided according to histologic tumor grade, 13 had radioimmunoassay (RIA) and immunohistochemistry (IHC). Concordant expression: (A) Strongly positive progesterone receptor (PR) staining of Grade 1 carcinoma (IHC score, 292; RIA value, 2827 fmol/mg). (B) Negative PR staining of Grade 3 carcinoma (IHC score, ,20; RIA value, 13 fmol/mg). Note the distinct staining of the nonneoplastic epithelium. Discordant expression: (C) Strongly positive estrogen receptor (ER) stain of Grade 1 carcinoma (IHC score, 285; RIA value, 0 fmol/mg). (D) Negative PR stain of Grade 3 carcinoma (IHC score, ,20; RIA value, 120 fmol/mg). Note the distinct staining in the admixed stroma. Original magnification, 3200 in A and C, 3400 in B and D. Grade 1 tumors (23.6%), 24 had Grade 2 tumors (43.6%), and 18 had Grade 3 tumors (32.8%). The IHC score was determined by multiplying the average nuclear staining intensity for ER and PR with the percentage of tumor cells stained16 (Fig. 2). Of the 55 tumors evaluated by IHC, 48 tumors (87.3%) were ER positive, 42 tumors (76.4%) were PR positive, 42 tumors (76.4%) were ER positive and PR positive, 6 tumors (10.9%) were ER positive and PR negative, 0 tumors were ER negative and PR positive, and 7 tumors (12.7%) were ER negative and PR negative. Of the 55 tumors evaluated by RIA, 32 tumors (58.2%) were ER positive, 42 tumors (76.4%) were PR positive, 30 tumors (54.6%) were ER positive and PR positive, 2 tumors (3.6%) were ER positive and PR negative, 12 tumors (21.8%) were ER negative and PR positive, and 11 tumors (20%) were ER negative and PR negative. Examples of concordance between IHC and RIA expression of hormone receptors are shown in Figure 2A,B, and discordant cases are shown in Figure 2C,D. Intratumor receptor heterogeneity, in which expression was found in some areas of malignant epithelium but not in others, also was encountered and is exemplified in Figure 3. It is noteworthy that the RIA PR level of this tumor was elevated markedly at 201 fmol/mg. To determine whether hormone receptor expression by RIA correlated with expression by IHC, kappa statistics for both ER and PR were calculated using RIA 2086 CANCER November 15, 1999 / Volume 86 / Number 10 TABLE 2 Agreement of Progesterone Receptor Positivity and Negativity Determined by Immunohistochemistry and Radioimmunoassay IHCa RIAb Negative Positive Negative Positive 11 2 2 40 IHC: immunohistochemistry; RIA: radioimmunoassay. a Negative IHC, score ,20; positive IHC, score .20. b Negative RIA, ,50 fmol/mg protein; positive RIA, .50 fmol/mg protein. c P 5 0.0005. FIGURE 3. Intratumor receptor heterogeneity shown in PR staining of a Grade 3 adenocarcinoma: The RIA level for this tumor was 201 fmol/mg. Note the positive staining in the area of the tumor on the right (arrow) and the complete negativity in the tumor on the left. In addition, nonneoplastic glands also stained strongly (arrowhead). TABLE 1 Agreement of Estrogen Receptor Positivity and Negativity by Immunohistochemistry and Radioimmunoassay IHCa RIAb Negative Positive Negative Positive 6 1 17 31 IHC: immunohistochemistry; RIA: radioimmunoassay. a Negative IHC, score ,20; positive IHC, score .20. b Negative RIA, ,50 fmol/mg protein; positive RIA, .50 fmol/mg protein. c P 5 0.006. and IHC dichotomies, in which tumors were categorized as either positive or negative based on previously determined cut-off points. The concordance between IHC and RIA analysis for both ER (Table 1) and PR (Table 2) was statistically significant using this method. The actual RIA and IHC values for ER and PR also were correlated significantly (data not shown) with P 5 0.0005. However, as evidenced by Spearmans correlation coefficients, the PR correlation (0.75) between methods was considerably stronger than the ER correlation (0.52). Correlations between hormone receptors, tumor grade and disease stage were determined using the Kruskal–Wallis test; the medians for both ER measures and PR measures were related to disease stage and tumor grade. The median values of both ER measures and PR measures show a linear correlation with tumor grade when patients were separated into groups ac- cording to low-grade and high-grade disease. In all patients, the higher the grade, the lower the ER and PR values (Fig. 4A,B). Although the trend for ER was not statistically significant, there was a steady decrease in median values with increasing tumor grade. The median values for ER and PR did not show a consistent correlation with disease stage (Fig. 4C,D). Only the ER measure (IHC) showed a consistent, although not statistically significant, trend in relation to disease stage. The median follow-up was 59 months for both recurrences and death. There were 12 patients who had disease recurrences (21.8%), with a median time to recurrence of 11 months, and 4 patients died (7%), with a median time to death of 18 months. It is noteworthy that 7 of the recurrences and 3 of the deaths occurred among the patients with Stage I disease. The Cox proportional hazards regression model was used to evaluate the ability of ER and PR, determined by either IHC or RIA, to predict the time to recurrence (Tables 3, 4) and the time to death. The hazards ratios for disease recurrence (Table 3) for ER among all patients all were .1.0, with the “null” hazard set at 1.0. However, when adjusted for advanced stage disease, there was a significant risk if the ER status by IHC was negative (HR 5 5.2; P 5 0.03). When all patients were considered, the hazards correlating PR and tumor grade were elevated, although none reached significance. To further evaluate deaths occurring in patients with Stage I disease, the regression models were applied to these patients in a separate analysis (Table 4). When considering ER status, all hazards ratios were elevated significantly, with the ratios for IHC (P 5 0.01) slightly higher than the ratios for RIA (P 5 0.04). In Stage I patients, the hazards ratios of recurrence for PR, when performed by RIA, were elevated significantly for RIA alone and for RIA adjusted for tumor grade (P 5 0.01 and P 5 0.02, respectively). Ratios were elevated, but not significantly, for PR by IHC in Stage I patients. Therefore, when controlling for tumor ER Predicts Endometrial Carcinoma Recurrence/Gehrig et al. 2087 FIGURE 4. Median ER and PR values, as determined by RIA (fmol/mg protein) and IHC score by tumor grade (A,B) and disease stage (C,D). (A) ER (RIA), P 5 0.03; PR (RIA), P 5 0.0009. (B) ER (IHC), P 5 0.10; PR (IHC), P 5 0.02. (C) ER (RIA), P 5 0.29; PR (RIA), P 5 0.62. (D) ER (IHC), P 5 0.09; PR (IHC), P 5 0.15. TABLE 3 Disease Recurrence Hazard Ratios and 95% Confidence Intervals for Estrogen Receptor and Progesterone Receptor, as Measured by Radioimmunoassay and Immunohistochemistry, Alone and Adjusted for Tumor Grade and Disease Stage TABLE 4 Disease Recurrence Hazard Ratios and 95% Confidence Intervals for Estrogen Receptor and Progesterone Receptor, as Measured by Radioimmunoassay and Immunohistochemistry, for Stage I Patients Alone and Adjusted for Tumor Grade All stages (95% CI) Adjust for ER Nothing Grade 3 Stages III–IV PR Nothing Grade 3 Stages III–IV All stages (95% CI) RIA (<50) IHC (<20) 2.2 (0.7–6.9) 1.8 (0.6–5.9) 1.6 (0.5–5.3) 2.8 (0.8–10.3) 1.9 (0.5–7.7) 5.2 (1.2–2.3)a 2.8 (0.9–8.8) 1.7 (0.4–6.8) 2.4 (0.7–7.6) 2.0 (0.6–6.7) 1.1 (0.3–4.4) 1.7 (0.5–5.7) 95% CI: 95% confidence interval; RIA: radioimmunoassay; IHC: immunohistochemistry; ER: estrogen receptor; PR: progesterone receptor. a Significant hazard ratios. grade, negative ER status by RIA and IHC was found to correlate with increased disease recurrence in Stage I patients. In contrast, only negative PR status by RIA, and not by IHC, was found to correlate with disease recurrence in Stage I patients. The correlation between the 4 deaths and hormone receptor status also was evaluated; however, the hazards ratios were unstable, with very large 95% confidence intervals for this small number of patients (data not shown). Overall, patients with low ER and PR values by both RIA and IHC were at an increased risk Adjust for ER Nothing Grade 3 PR Nothing Grade 3 RIA (<50) IHC (<20) 5.7 (1.1–30)a 5.5 (1.1–29)a 8.1 (1.8–37)a 7.6 (1.6–36)a 7.7 (1.5–32)a 7.6 (1.5–40)a 4.2 (0.9–19) 4.1 (0.7–22) 95% CI: 95% confidence interval; RIA: radioimmunoassay; IHC: immunohistochemistry; ER: estrogen receptor; PR: progesterone receptor. a Significant hazard ratios. of death. Because of the small number of deaths in this series, statistical significance could not be reached. DISCUSSION Several studies have evaluated ER and PR status in patients with endometrial carcinoma.3–11,14 –19 The majority of these studies were performed using biochemical methods, such as RIA on frozen tissues.6 –9,11,14 A recent trend has been to evaluate ER and PR status using IHC methods on either frozen or paraffin embedded tissues; however, only a few studies have evaluated IHC and RIA simultaneously from pathologic specimens.15–16 Previous studies have shown reason- 2088 CANCER November 15, 1999 / Volume 86 / Number 10 able concordance between the two methods; however, in cases in which a correlation did not exist, the authors have proposed several explanations. One hypothesis is that, in cases in which a tumor is RIA positive but IHC negative, the fixation of the tissues may alter the antigenic site of the protein receptor, thus leading to the discordant result.4 In addition, another hypothesis suggests that delay in tissue handling may lead to partial proteolysis of the tissues, thus resulting in negative IHC results, whereas RIA hormone receptor levels are found to be elevated. In this study, there was excellent concordance between the IHC results and the RIA results using Spearmans correlation coefficient; however, there still were cases in which there was a disparity between hormone receptor expression as determined by IHC and RIA. These differences in hormone receptor expression may be explained by the technical limitations of the biochemical assay. Unlike breast carcinoma, which is surrounded by stromal tissues that contain no or relatively low levels of receptor, endometrial carcinoma is involved intimately with benign stroma that may contain significant amounts of receptor, thus elevating the hormone receptor levels falsely when measured by RIA.16 This may lead to erroneous results when measuring hormone receptors in whole-tissue homogenates. Several other studies also have suggested that IHC should be used for determination of hormone receptor levels because of the inaccuracies found with use of RIA.10,15,17,18 However, we acknowledge that any quantitation of immunohistochemical stains is an inexact science and that a certain interobserver variability exists. Fortunately, this is an issue only in the very small subset of patients who produce borderline negative or positive results, and it may be less significant than the often arbitrary choice of cutoff values used in biochemical assays. In the current study, there were 12 patients who were ER negative and PR positive as determined by RIA but not by IHC analysis. Biologically, the expression of PR is dependent on ER for up-regulation; therefore, there should not be any PR positive tumors without the influence of estrogen. Initially, as seen in Figure 1D, we concluded that PR levels determined by RIA were elevated falsely, because the benign stroma adjacent to the carcinoma strongly expressed PR. In addition, there were 12 patients with a negative ER status determined by RIA who, by IHC, were found to express ER in epithelial cells. Therefore, in addition to concluding that RIA does not allow one to control for the stromal component, we propose that freeze-thaw artifact prior to performing RIA may interfere with ligand binding, leading to false negative ER status. Because there were so few deaths in our study, we chose to evaluate disease recurrence as the endpoint. In this study, when evaluating the correlation between hormone receptor status and tumor recurrence, ER status determined by IHC was the most predictive of recurrent disease. This is in contrast to other studies in which it was found that PR status, and not ER status, was the more significant prognostic factor.7–11 Because the majority of recurrences occurred among Stage I patients, the regression models were applied to these patients in a separate analysis. Even after controlling for tumor grade, ER status still was found to be more significant than PR status as a prognostic factor. Other studies have evaluated hormone receptor status as a prognostic factor for death, and the results are variable. In a study evaluating 175 patients, survival was superior in patients with PR positive tumors, whereas ER status did not correlate with survival.8 Similarly, other investigators, using biochemical and immunohistochemical methods, have observed that PR status, and not ER status, was correlated with death from disease.9,10 However, only one study using RIA and another study using IHC analysis found that ER status was more predictive of patient survival.4,6 Whereas normal breast and endometrial epithelial cells express ERs, loss of ER suggests a more advanced molecular pathology with the deregulation of signaling pathways. Tumors that no longer express ER, therefore, are removed further from the normal cellular phenotype and may be considered more aggressive and refractory to therapy. Thus, the poor patient outcome associated with the absence of ER expression may reflect a more malignant tumor course than would be expected based only on the histologic diagnosis. Clearly, there is significant disagreement in the literature regarding the prognostic significance of ER and PR status in patients with endometrial adenocarcinoma. To complicate the situation further, in clinical practice, two very different methods have been used to evaluate hormone receptor expression. This study, in which both ER and PR were found to be related to disease recurrence, reports that ER status, as determined by IHC, can be used as an independent predictor for disease recurrence. Clinically, this is useful information, because IHC methodology is less expensive, more accurate, and easily performed on paraffin embedded tissues. The numbers in this study were too small to make a definitive statement on the importance of hormone receptor status in patients with endometrial carcinoma; however, there should be some consideration of using IHC to determine the presence of ER as a prognostic factor in the clinical management of these patients. ER Predicts Endometrial Carcinoma Recurrence/Gehrig et al. ACKNOWLEDGMENTS The authors thank Elizabeth Allred for her assistance with the statistical analysis. REFERENCES 1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics 1998. CA Cancer J Clin 1998;48:6. 2. Pichon MF, Broet P, Magdelenat H, Delarue JC, Spyratos F, Basuyau JP, et al. Prognostic value of steroid receptors after long-term follow-up of 2257 operable breast cancers. Br J Cancer 1996;73:1545–51. 3. Creasman WT. Prognostic significance of hormone receptors in endometrial cancer. Cancer 1993;71:1467–70. 4. Chambers JT, Carcangiu ML, Voynick IM, Schwartz PE. Immunohistochemical evaluation of estrogen and progesterone receptor content in 183 patients with endometrial carcinoma. Am J Clin Pathol 1990;94:255– 60. 5. Pertschuk LP, Masood S, Simone J, Feldman JG, Fruchter RG, Axiotis CA, et al. Estrogen receptor immunocytochemistry in endometrial carcinoma: a prognostic marker for survival. Gynecol Oncol 1996;63:28 –33. 6. Chambers JT, MacLusky N, Eisenfield A, Kohorn EI, Lawrence R, Schwartz PE. Estrogen and progestin receptor levels as prognosticators for survival in endometrial cancer. Gynecol Oncol 1988;31:65–77. 7. Ingram SS, Rosenman J, Heath R, Morgan TM, Moore D, Varia M. The predictive value of progesterone receptor levels in endometrial cancer. Int J Radiat Oncol Biol Phys 1989;17:21–7. 8. Ehrlich CE, Young PCM, Stehman FB, Sutton GP, Alford WM. Steroid receptors and clinical outcome in patients with adenocarcinoma of the endometrium. Am J Obstet Gynecol 1988;158:796 – 807. 9. Kleine W, Maier T, Geyer H, Pfleiderer A. Estrogen and progesterone receptors in endometrial cancer and their prognostic relevance. Gynecol Oncol 1990;38:59 – 65. 10. Fukuda K, Mori M, Uchiyama M, Iwai K, Iwasaka T, Sugimori H. Prognostic significance of progesterone receptor immunohistochemistry in endometrial carcinoma. Gynecol Oncol 1998;69:220 –5. 11. Kadar N, Malfetano JH, Homesley HD. Steroid receptor concentrations in endometrial carcinoma: effect on survival in surgically staged patients. Gynecol Oncol 1993;50:281– 6. 2089 12. Battifora H. Immunocytochemistry of hormone receptors in routinely processed tissues: the new gold standard. Appl Immunol 1994;2:143–5. 13. Taylor CR. Paraffin section immunocytochemistry for estrogen receptor: the time has come. Cancer 1996;77:2419 –22. 14. Palmer DC, Muir IM, Alexander AI, Cauchi M, Bennerr RC, Quinn MA. The prognostic importance of steroid receptors in endometrial carcinoma. Obstet Gynecol 1988;72:388 –93. 15. Pertschuk LP, Beddoe AM, Gorelic LS, Shain SA. Immunocytochemical assay of estrogen receptors in endometrial carcinoma with monoclonal antibodies: comparison with biochemical assay. Cancer 1986;57:1000 – 4. 16. Soper JT, Segreti EM, Novotny DB, Mutch D, Creasman WT, McCarty KS. Estrogen and progesterone receptor content of endometrial carcinomas: comparison of total tissue versus cancer component analysis. Gynecol Oncol 1990;36:363– 8. 17. Mutch DG, Soper JT, Novotny DA, Cox EB, Creasman WT, McCarty KS, et al. Endometrial adenocarcinoma estrogen receptor content: association of clinicopathologic features with immunohistochemical analysis compared with standard biochemical methods. Am J Obstet Gynecol 1987;157: 924 –32. 18. Nyholm HCJ, Nielsen AL, Lyndrup J, Norup P, Thorpe SM. Biochemical and immunohistochemical estrogen and progesterone receptors in adenomatous hyperplasia and endometrial carcinoma: correlations with stage and other clinicopathologic features. Am J Obstet Gynecol 1992;167:1334 – 42. 19. Budwit-Novotony DA, McCarty KS, Cox EB, Soper JT, Mutch DG, Creasman WT. Immunohistochemical analyses of estrogen receptor in endometrial adenocarcinoma using a monoclonal antibody. Cancer Res 1986;46:5419 –25. 20. Pertschuk LP, Feldman JG, Kim Y, Braithwaite L, Schneider F, Braverman AS, et al. Estrogen receptor immunocytochemistry in paraffin embedded tissues with ER1D5 predicts breast cancer endocrine response more accurately than H222Spg in frozen sections or cytosol-based ligand-binding assays. Cancer 1996;77:2514 –9. 21. Scambia G, Benedetti-Panici P, Ferrandina G, Distefano M, Salerno G, Romanini ME, et al. Epidermal growth factor, oestrogen and progesterone receptor expression in primary ovarian cancer: correlation with clinical outcome and response to chemotherapy. Br J Cancer 1995;72:361– 6.