Study Population

The study cohort is 4,150 female patients with newly diagnosed, invasive breast cancer in British Columbia, whose tumor specimens were tested by a central (ER) laboratory at Vancouver Hospital between 1986 and 1992. The DCC protocol is as published^10-12 and reproduced on our web site (www.gpec.ubc.ca/index.php?content=papers/ER.php). Median follow-up was 12.4 years and age at diagnosis 60 years. All patients had been referred to the British Columbia Cancer Agency and have staging, pathology, treatment, and follow-up information.^13,14 During the study era, 75% of breast cancer patients in the province were referred; nonreferred patients were generally elderly or treated by mastectomy without indications for adjuvant therapy.¹⁵

Abstracted clinical information includes age; histology; grade; tumor size; number of involved axillary nodes; lymphatic or vascular invasion (LVI); ER status by the DCC method¹¹; type of local and initial adjuvant systemic therapy (AST); and dates of diagnosis, first local, regional, or distant recurrence, and death. A subset of these patients was included in a recent population-based study validating the prognostic model ADJUVANT!¹⁶ Table A3 (online only) summarizes cohort characteristics. The study was approved by the Clinical Research Ethics Board of the University of British Columbia and BC Cancer Agency.

Tissue Microarrays and IHC

The Vancouver Hospital ER laboratory retained single archival blocks from each patient. This material had been frozen before neutral buffered formalin fixation. Slides from these blocks stained with hematoxylin and eosin were reviewed by two pathologists to identify areas of invasive breast carcinoma. Tissue microarrays (TMAs) were constructed as described^17,18 (details reproduced at www.gpec.ubc.ca/index.php?content=papers/ER.php). Using one core per patient, 17 TMA blocks were required. Thin sections (4 μm) were immunostained using DakoCytomation EnVision and System-HRP (Dako Corporation, Carpinteria, CA) in a two-step technique. Slides were deparaffinized with xylene and rehydrated through three alcohol changes. Endogenous peroxidase activity was quenched by incubating 5 minutes with 0.03% hydrogen peroxide/sodium azide. Slides were then incubated with one of two primary anti-ER antibodies, 1D5 (1:100; Dako Corp, Carpinteria, CA) or SP1 (1:250; LabVision, Fremont, CA), followed by the peroxidase-labeled polymer, in a Tris-HCl buffer containing stabilizing protein and an antimicrobial agent, using sequential 30-minute incubations. Staining was completed by a 10-minute incubation with 3,3′-diaminobenzidine plus substrate-chromogen. Primary antibody was omitted in negative controls. External positive controls were slides from breast cancers with previously documented ER expression. A previous study using a duplicate-redundancy 431-case TMA demonstrated 96% agreement between duplicate cores, for both 1D5 and SP1⁹ (Tables A1 and A2).

Stained TMA slides were digitally scanned and linked to a relational database,¹⁹ and are available for review (https://www.gpecimage.ubc.ca/tma/web/viewer.php; username, ersp11D5; password, er4150).

ER Scoring System

TMAs were scored visually by two pathologists (D.T., A.M.G.) for percentage of tumor cell nuclear positivity, and scored as negative (< 1%); positive, 1+ (1% to 25%); positive, 2+ (25% to 75%); or positive, 3+ (> 75%). For most analyses, IHC scores are dichotomized at ≥ 1 = ER positive.²⁰ Pathologists were blinded to clinical outcomes.

Statistical Analysis

Statistical analysis was performed using SPSS software version 13.0 (SPSS Inc, Chicago, IL) and R 2.1.1 (www.r-project.org). In univariate analyses, Overall survival (OS), breast cancer–specific survival (BCSS), and relapse-free survival (RFS) were estimated using Kaplan-Meier²¹ curves, and survival differences determined by log-rank tests.^22,23 A trend test was used when three or more ordered groups were compared. For BCSS, survival time was censored at death if the cause was not breast cancer, or if the patient was still alive at the end of the study. Six patients with unknown cause of death were excluded from BCSS analysis. For RFS, survival time was also censored at death if the cause was not breast cancer or the patient was alive without relapse at the end of the study. For OS, survival time was censored if the patient was still alive at the end of the study.

Cox proportional hazards^22,24 models were used to calculate adjusted hazard ratios (HRs) accounting for covariates. Hypothesis testing was performed using Wald's statistic. Smoothed plots of weighted Schoenfeld residuals were used to test proportional hazard assumptions.²⁵ κ²⁶ statistics and Kendall's τ-b²⁷ tests were used to measure agreement between the two ER immunostains and DCC assay, and correlation of ER status to pathologic variables. Differences involving pathologic factors were compared using Pearson's χ²⁸ and Mann-Whitney U²⁹ for categoric and continuous variables, respectively. All statistical tests were two-sided and P < .05 was considered significant.

Comparison of ER Expression by SP1 and 1D5 With DCC and Clinicopathologic Characteristics

Table 1 summarizes the IHC staining results for SP1 and 1D5 on TMAs. Figures 1A to 1C are from the same tumor sample that was negative for ER by 1D5 but strongly positive by SP1. This sample had an ER concentration of 48 fmol/mg by DCC. Figures 1D to 1F are from another tumor sample, which was moderately positive by 1D5 but strongly positive by SP1 (ER concentration, 174 fmol/mg).

The number of interpretable samples for the two immunostains is slightly different due to occasional core dropout from TMAs during sectioning and staining. Overall, 69.5% of samples were positive by SP1 versus 63.1% positive by 1D5. In absolute terms, the SP1 antibody had 11% more strong 3+ stains than 1D5, and 1D5 had 6.4% more negative stains than SP1. There is a strong positive correlation between the two antibodies (Kendall's τ-b, 0.790; P < 1.0 × 10⁻¹⁷).

Among clinical tumor samples, 3,884 samples were originally tested by DCC assay and categorized into four groups^10,11: negative (≤ 1 fmol/mg), low (2 to 9 fmol/mg), moderate (10 to 159 fmol/mg), and high (≥ 160 fmol/mg). Frequencies were 2.9% (111 of 3,884), 17.5% (678 of 3,884), 41.1% (1,596 of 3,884), and 38.6% (1,499 of 3,884), respectively. Values of DCC ≥ 10.0 fmol/mg were considered positive (the clinical cut point).^11,12 ER status by SP1 (κ, 0.654; sensitivity, 86%; specificity, 92%; P < 1.0 × 10⁻¹⁷) agreed better with DCC than did 1D5 (κ, 0.536; sensitivity, 78%; specificity, 92%; P < 1.0 × 10⁻¹⁷; Table 2).

A total of 4,105 samples were interpretable for both antibodies. SP1 stained 8% more positive samples that otherwise were negative by 1D5 (Table 3). For these 337 discrepant samples, 92% were DCC positive (median concentration, 67.0 fmol/mg). Among the 77 discrepant cases identified as negative by SP1 but positive by 1D5, 65% were DCC positive (median concentration, 34.0 fmol/mg). The SP1-positive/1D5-negative discordant samples had significantly higher ER concentrations by DCC (P = .021) and significantly lower tumor grades (P = .022) than the SP1-negative/1D5-positive discordant samples; there was no significant difference in tumor size (P = .110). Among the SP1-negative/1D5-negative group, 64% were DCC negative; among the SP1-positive/1D5-positive group, 98% were DCC positive.

SP1-positive, 1D5-positive, and DCC-positive status all correlate with older age at diagnosis (Kendall's τ-b, 0.149 [P < 1.0 × 10⁻¹⁷], 0.173 [P < 1.0 × 10⁻¹⁷], and 0.165 [P < 1.0 × 10⁻¹⁷], respectively), but correlate inversely with grade (Kendall's τ-b, −0.241 [P = 9.96 × 10⁻⁶⁰], −0.196 [P = 1.88 × 10⁻³⁸], and −0.282 [P = 3.25 × 10⁻⁸⁶]) and tumor size (Kendall's τ-b, −0.107 [P = 4.56 × 10⁻¹²], −0.083 [P = 7.42 × 10⁻⁸], and −0.108 [P = 1.2 × 10⁻¹¹]).

Prognostic Values of SP1-, 1D5-, and DCC-Positive Expression

To assess the prognostic value of ER in a general population, we examined SP1, 1D5, and DCC in the entire cohort. Both immunostains and DCC identified ER-positive patients as having an improved BCSS. DCC assay demonstrates the most significant and strongest linear trend for expression: the higher expression, the better BCSS (Fig 2A [SP1], P = 4.78 × 10⁻¹³; Fig 2B [1D5], P = 1.65 × 10⁻⁷; and Fig 2C [DCC], P = 6.08 × 10⁻¹⁶). A similarly significant trend is observed for RFS (SP1, P = 3.98 × 10⁻⁸; 1D5, P = 3.71 × 10⁻⁶; and DCC, P = 9.25 × 10⁻¹⁴) but not OS (SP1, P = .132; 1D5, P = .511; and DCC, P = .154). The 5- and 10-year survival probabilities of binarized ER status from Kaplan and Meier analyses are listed in Table A4 (online only). The increased survival for ER-positive patients starts to attenuate after 5 years. The same phenomenon is seen with all three detection methods.

AST for this cohort was prescribed according to guidelines based on age, tumor size, LVI, nodal status, and DCC-determined ER level.¹⁶ High risk was defined as node positive, or if node negative, the presence of LVI or tumor more than 2 cm and ER negative (DCC < 10 fmol/mg). Patients deemed low risk at the time of diagnosis were not given any AST; they had a better outcome (10-year BCSS, 82%; 95% CI, 80% to 84%) than patients receiving AST such as the tamoxifen group (10-year BCSS 69%; 95% CI, 66% to 71%; P = 8.4 × 10⁻²¹). Therefore, survival analyses based on ER status were done separately in these two subgroups.

For the pure prognostic subset of patients receiving no AST, SP1-positive status was associated with 14% absolute increased BCSS (P = 5.0 × 10⁻⁸) at 10 years, 1D5-positive status was associated with 9% (P = 2.2 × 10⁻⁴), and DCC-positive status was associated with 19% (P = 2.3 × 10⁻¹²; Table A4). For RFS, SP1-positive and DCC-positive status were increased significantly by 6% (P = .002) and 13% (P = 5.1 × 10−7) at 10 years. 1D5-positive status showed 3% increased 10-year RFS (P = .06).

Among 1,377 patients receiving tamoxifen as their only AST, only 84 were ER negative by DCC. SP1, 1D5, and DCC all identified ER-positive patients as having better outcomes than ER-negative patients. The absolute increased 10-year BCSS for SP1-, 1D5-, and DCC-positive status are 14% (P = 4.7 × 10⁻⁶), 10% (P = 1.5 × 10⁻⁴), and 24% (P = 3.3 × 10⁻⁷); the increased 10-year RFS for SP1-, 1D5-, and DCC-positive status are 10% (P = 1.7 × 10⁻⁴), 7% (P = .001), and 18% (P = 3.9 × 10⁻⁵), respectively (Table A4).

Concordant and Discordant Cases Between SP1 and 1D5

As expected, patients with SP1/1D5 double-positive status have better BCSS (HR, 0.667; 95% CI, 0.592 to 0.753) compared with patients with SP1/1D5 double-negative status (P = 5.00 × 10⁻¹¹; Fig 3). The SP1-positive/1D5-negative group has a HR of 0.647 (95% CI, 0.423 to 0.989; P = .044) compared with the SP1-negative/1D5-positive group. Importantly, there are no significant survival differences between the SP1-positive/1D5-positive and SP1-positive/1D5-negative patients (P = .408). The SP1-negative/1D5-positive group does not have significantly different survival from the SP1-negative/1D5-negative group (HR, 0.93; 95% CI, 0.63 to 1.36; P = .698) but seems inferior to the SP1-positive group (HR, 1.45; 95% CI, 0.99 to 2.11; P = .055). In relapse-free survival analysis, the SP1-negative/1D5-positive patients have a similar hazard compared with the SP1-negative/1D5-negative group, with a HR of 0.972 (95% CI, 0.683 to 1.38), whereas SP1-positive/1D5-negative patients have relapse-free hazard similar to that of SP1-positive/1D5-positive patients (HR, 1.07; 95% CI, 0.885 to 1.28; (Fig A1, online only). The same pattern in BCSS and RFS is seen in both the subsets of patients not treated with AST or those treated only with tamoxifen (data not shown). These results also support that SP1 antibody is a better prognostic marker than the commonly used ER antibody 1D5.

Multivariate Analysis

Smoothed, rescaled Schoenfeld residuals plots were used to test proportional hazards assumptions. All covariates followed proportional hazards, except ER status, which varied slightly during the long period of follow-up. The hazard rate of breast cancer death and relapse is different in the first 5 years, consistent with data reported by Hess et al.³⁰ Table A5 summarizes the adjusted HRs of ER-positive status detected by the two immunostains and DCC; the reported values were determined from Cox regression models including age, tumor size, grade, LVI, and nodal status as covariates for BCSS and RFS (Table A5). The results show that SP1, 1D5, and DCC all work efficiently as independent prognostic factors for BCSS and RFS in a general population after adjusting for the listed clinicopathologic prognosticators. However, SP1, 1D5, or DCC did not remain significant among the low-risk patients receiving no AST.

For patients receiving tamoxifen as their only AST, SP1, 1D5, and DCC each retained significance when added individually to a model including the same clinical parameters listed above as covariates. The BCSS HRs for SP1-positive, 1D5-positive, and DCC-positive status were 0.636 (95% CI, 0.499 to 0.811; P = 2.55 × 10⁻⁴), 0.697 (95% CI, 0.559 to 0.870; P = 1.37 × 10⁻³), and 0.649 (95% CI, 0.450 to 0.935; P = 2.03 × 10⁻²), respectively. The RFS HRs for SP1-positive, 1D5-positive, and DCC-positive status were 0.683 (95% CI, 0.540 to 0.864; P = 1.45 × 10⁻³), 0.744 (95% CI, 0.601 to 0.921; P = 6.59 × 10⁻³), and 0.709 (95% CI, 0.491 to 1.03; P = .068), respectively. To test which detection method was more significantly associated with prognosis, a Cox regression model with age, tumor size, grade, LVI, and nodal status was fitted with each of SP1, 1D5, and DCC. SP1 was the most significant prognostic factor among the three ER detection methods, both in BCSS (HR, 0.664; 95% CI, 0.517 to 0.852; P = 1.27 × 10⁻³) and RFS (HR, 0.699; 95% CI, 0.550 to 0.888; P = 3.40 × 10⁻³) in this subgroup. The HRs of the clinicopathologic covariates are listed in Tables 4 and 5.