|Home | About | Journals | Submit | Contact Us | Français|
The NSABP (National Surgical Adjuvant Breast and Bowel Project) B-24 study demonstrated significant benefit with adjuvant tamoxifen in patients with ductal carcinoma in situ (DCIS) after lumpectomy and radiation. Patients were enrolled without knowledge of hormone receptor status. The current study retrospectively evaluated the relationship between receptors and response to tamoxifen.
Estrogen (ER) and progesterone receptors (PgR) were evaluated in 732 patients with DCIS (41% of original study population). An experienced central laboratory determined receptor status in all patient cases with available paraffin blocks (n = 449) by immunohistochemistry (IHC) using comprehensively validated assays. Results for additional patients (n = 283) determined by various methods (primarily IHC) were available from enrolling institutions. Combined results were evaluated for benefit of tamoxifen by receptor status at 10 years and overall follow-up (median, 14.5 years).
ER was positive in 76% of patients. Patients with ER-positive DCIS treated with tamoxifen (v placebo) showed significant decreases in subsequent breast cancer at 10 years (hazard ratio [HR], 0.49; P < .001) and overall follow-up (HR, 0.60; P = .003), which remained significant in multivariable analysis (overall HR, 0.64; P = .003). Results were similar, but less significant, when subsequent ipsilateral and contralateral, invasive and noninvasive, breast cancers were considered separately. No significant benefit was observed in ER-negative DCIS. PgR and either receptor were positive in 66% and 79% of patients, respectively, and in general, neither was more predictive than ER alone.
Patients in NSABP B-24 with ER-positive DCIS receiving adjuvant tamoxifen after standard therapy showed significant reductions in subsequent breast cancer. The use of adjuvant tamoxifen should be considered for patients with DCIS.
Ductal carcinoma in situ (DCIS) was gradually recognized as a disease distinct from invasive breast cancer (IBC) during the early 20th century.1,2 Between then and the introduction of screening mammography 70 to 80 years later, DCIS accounted for fewer than 5% of newly diagnosed breast cancers.3–6 Since the introduction and widespread use of screening mammography in the United States, the incidence of DCIS has increased dramatically, and today it accounts for 20% to 30% of all breast cancers.3,5,6
In the not too distant past, the standard treatment for DCIS was mastectomy and complete axillary dissection, which eradicated the disease, but at the high cost of disfiguring surgery.4,7,8 Local excision largely replaced this radical approach, and today lumpectomy is the standard of care for the majority of patients with DCIS.9–11 Early experience with lumpectomy alone was disappointing, with local recurrence rates of 30% or more, mainly because of the difficulty of obtaining clear surgical margins.4,12–16 Adjuvant radiotherapy was added to lumpectomy to sterilize occult residual tumor.17–19 The seminal NSABP (National Surgical Adjuvant Breast and Bowel Project) B-17 DCIS trial demonstrated that the addition of radiotherapy reduced the relative risk of local recurrence by 61%, although 13% of patients still eventually developed an ipsilateral recurrence, comparable to patients with IBC treated in a similar manner.20,21 As a result of its success in treating IBCs,22,23 the NSABP conducted its B-24 clinical trial to evaluate adjuvant tamoxifen in DCIS after lumpectomy and radiation. This trial demonstrated an additional 37% reduction in relative risk of local recurrence and a decrease in contralateral breast cancer of comparable magnitude.24 Results were similar in the United Kingdom-ANZ (United Kingdom, Australia, and New Zealand) trial, which also evaluated the efficacy of adjuvant tamoxifen in DCIS.25
We now know that the benefit of tamoxifen in treating IBC is essentially restricted to estrogen receptor (ER) –positive and/or progesterone receptor (PgR) –positive disease, consistent with the biologic mechanism of action of the drug.26 B-24 was initiated before this relationship was firmly established, and receptor status was not involved in patient enrollment. The purpose of the current study was to evaluate retrospectively ER and PgR and their relationships to response to adjuvant tamoxifen in the B-24 trial.
The study samples were derived from patients enrolled onto NSABP Protocol B-24,24 in which 1,804 women with DCIS were randomly assigned to placebo (n = 902) or tamoxifen (n = 902) after lumpectomy and local radiotherapy (Fig 1). Briefly, local radiotherapy (50 Gy) was administered beginning no later than 8 weeks after surgery. Adjuvant placebo or tamoxifen therapy (10 mg twice daily) was initiated within 56 days of surgery and continued for 5 years. Follow-up included physical examination every 6 months and annual mammography. Local recurrences were confirmed histologically. Distant metastases were confirmed clinically and by radiology and/or histology. End point information was based on data through the closure of follow-up to the study on May 1, 2007, obtained and received by the NSABP Biostatistical Center by June 8, 2007. The median time on study for patients in the analysis cohort at this time was 14.5 years. Informed consent was required for all participants. The protocol and consent forms were approved by the National Institutes of Health and the institutional review boards of all participating institutions.
ER and PgR results for this study were available in a subset of patients (41%; n = 732) enrolled onto the original clinical trial. This subset was statistically similar to the remaining study population with regard to the distribution of treatment (placebo v tamoxifen) and other available factors including age, race, eligibility status, tumor size, and surgical margin status (data not shown).
Formalin-fixed paraffin-embedded tissue (FFPET) blocks of DCIS were available for 449 patents and were used to determine ER and PgR status by immunohistochemistry (IHC) in an experienced central laboratory. The IHC assays used mouse monoclonal antibodies 6F11 for ER (Novacastra, Burlingame, CA) and 1294 for PgR (Dako, Carpenteria, CA) and were comprehensively validated in several previous studies of IBCs.27–30 The Allred score was used to estimate the proportion and intensity of positive cells (range, 0 to 8), and scores of 3 or greater were defined as clinically positive.28,30 Representative examples of positive and negative IHC results for ER in DCIS are shown in Figure 2.
ER and PgR results from laboratories used by enrolling institutions were available for an additional 283 patients, although FFPET blocks were not available for central laboratory testing for those patients. These results were determined by a variety of methods reflecting daily practice at each institution (IHC, 62%; radio-labeled ligand binding assay, 24%; enzyme immunoassay, 5%; unknown, 9%). Local interpretations of results, which also varied methodologically, were relied on to define positive and negative receptor status. There were no significant differences in the proportions of patient cases evaluated in central versus outside laboratories regarding treatment (placebo v tamoxifen), age, race, tumor size, surgical margin status, or the presence of comedo necrosis (data not shown).
The primary end point of analyses in this study was time to the occurrence of any breast cancer as a first event subsequent to the original diagnosis of DCIS. Events included all local, regional, or distant recurrences and contralateral breast cancers. Other primary cancers or deaths without breast cancer were censored for the analysis. Secondary analyses included time to ipsilateral and contralateral breast cancers (IBC and/or DCIS) as first events. Events used in the analyses were medically reviewed and classified as being either invasive or noninvasive (ie, DCIS).
Cox proportional hazards models were employed to make formal inferences about group comparisons, and Kaplan-Meier curves were used to quantify the percentage of patients who were free of recurrence over time. Both unadjusted and adjusted Cox regression analyses were employed to quantify hazard ratios (HRs) comparing the two treatment groups.31 Adjustments were made for stratification of age at entry. Formal tests of treatment by ER status interactions were also conducted. To further explore the variability of treatment effects within ER subgroups, age-adjusted models were fitted. P values associated with HRs were based on large sample approximations, except in settings where expected events were quite rare (defined as < five in any treatment group). For the latter cases, Fisher's exact tests were employed by assuming that the numbers of failures were governed by Poisson processes.31 Analyses were also performed to determine if significant treatment by stratification variable interactions existed with respect to the end points.32 The incidences of site-specific failures were calculated using cumulative incidence curves.33–35 P values for treatment comparisons of cumulative incidence curves were obtained using cause-specific hazard rates that adjusted for the stratification variables.35
ER and PgR were positive in 76% and 66% of patients, respectively, based on results from central and institutional results combined (n = 732). In patient cases with both ER and PgR results (n = 714), 64% were ER positive/PgR positive, 13% were ER positive/PgR negative, 2% were ER negative/PgR positive, and 21% were ER negative/PgR negative. Patients were almost evenly distributed between the placebo and tamoxifen arms of the study (78% v 82%; P = .30). There were no significant differences in receptor status or other available and potentially confounding factors between the two treatment groups, including tumor size, margin status, comedo necrosis, age, race, and method of detection (Table 1).
Table 2 shows the risk of subsequent breast cancer in all patients with DCIS stratified by treatment arm and ER status at overall follow-up. Patients with ER-positive DCIS who received adjuvant tamoxifen versus placebo showed significant reductions in any breast cancer event (HR, 0.58; P = .001), any IBC (HR, 0.53; P = .005), and any contralateral breast cancer (HR, 0.50; P = .02). Reductions were also observed for any DCIS, any ipsilateral cancer, ipsilateral IBC, ipsilateral DCIS, contralateral IBC, and contralateral DCIS, but none were statistically significant. Results were similar in analyses of patients stratified by PgR and receptor (ER and/or PgR) status (Appendix Table A1, online only) although overall, they were not more predictive than when ER status was considered alone. Analyses at 10 years were similar to those of overall follow-up (Appendix Table A1). No significant reductions were associated with ER-negative DCIS in any setting.
Figure 3 shows Kaplan-Meier curves comparing treatment groups for ER-positive and -negative patients, demonstrating that the reduction of subsequent breast cancer was restricted to patients with receptor-positive DCIS treated with tamoxifen. In multivariable analysis (Table 3) of patients with available ER results, the only independently significant predictors of subsequent breast cancer were treatment status (tamoxifen v placebo; HR, 0.64; P = .003) and age at entry (≤ 49 v ≥ 50 years; HR, 0.61; P < .001). Results were similar for all patients in B-24.
The NSABP B-24 clinical trial demonstrated that adjuvant tamoxifen reduces subsequent ipsilateral and contralateral breast cancers in women with DCIS after lumpectomy and radiation.24 United Kingdom-ANZ, the only other major randomized trial evaluating tamoxifen in DCIS, failed to show a benefit in the lumpectomy plus radiation arm of the study (comparable to B-24), although significant benefit was observed in the lumpectomy plus tamoxifen alone arm.25 Tamoxifen has now been shown to significantly reduce subsequent breast cancer in randomized clinical trials of patients with IBC (eg, NSABP B-14),23,36 patients with DCIS (eg, NSABP B-24 and United Kingdom-ANZ),24,25,37 and women who are at high risk for developing breast cancer (eg, NSABP P-1).38
Numerous studies and clinical trials over the past 25 years have shown that the benefit of adjuvant tamoxifen in IBC is restricted to patients with receptor-positive disease.36 The B-24 trial was initiated before hormone receptor status was routinely evaluated in DCIS, and patients were enrolled without this information. The current study was undertaken to evaluate, retrospectively, the relationship between adjuvant tamoxifen and receptor status in DCIS, with the expectation that the results would be similar to those in IBC. Indeed, with prolonged follow-up, adjuvant tamoxifen was shown to significantly reduce subsequent ipsilateral breast cancer only in patients with ER-positive DCIS. Tamoxifen reduced contralateral breast cancer in patients with ER-positive and -negative DCIS, as expected (ie, prevention). No ipsilateral benefit was observed in ER-negative disease, emphasizing that tamoxifen must bind to functional ER to exert its beneficial effect on pre-existing residual tumor cells.26,39
The current study was based on results from 732 patients with DCIS and known receptor status. Receptors in a majority (61%) were evaluated by IHC in an experienced central laboratory using comprehensively validated assays for ER and PgR.27–30,40 The remainder (39%) were evaluated by various methods (the majority by IHC) in institutional laboratories (FFPET blocks were not available for central confirmation). All results were combined to increase the statistical power of the study overall. However, a subset of patient cases (n = 102) was evaluated for ER by IHC in both central and institutional laboratories, and agreement (positive v negative) was only 74.5%. There was also a significant difference in the incidence of ER-negative patient cases when the results from central and outside laboratories were compared (20% v 30%; P = .002). Assuming the central results are correct, this suggests a substantial rate of false negatives from institutions. To investigate this issue, response to tamoxifen was compared in patients with ER-negative DCIS determined in central versus institutional laboratories (Table 4). Although it was not statistically significant, there was an apparent benefit in ER-negative patient cases determined by institutional laboratories (eg, HR, 0.58; P = .28 for any subsequent cancer), whereas no benefit was observed in patients with central results (HR, 1.09; P = .85 for any cancer). These results are consistent with a false-negative rate of 30% to 40% from institutional laboratories, which may be responsible for the small apparent benefit observed in ER-negative patient cases overall. Similar errors in receptor testing by IHC in IBCs have been observed in laboratories around the world.41–44 In response, the National Cancer Comprehensive Network, College of American Pathologists, and American Society of Clinical Oncologists published guidelines to help improve the quality IHC testing of hormone receptors in breast cancer.40,45
In summary, retrospective analyses of hormone receptors of patients enrolled onto the NSABP B-24 clinical trial showed a significant benefit for adjuvant tamoxifen in patients with ER-positive DCIS after standard therapy. This offers an additional therapeutic option for patients and physicians to consider.
We thank Edwin R. Fisher, MD (deceased), for his contributions to this study. D.C.A. also thanks Barbara C. Good, PhD, for editorial assistance.
|Event by Receptor Status||10 Years||Overall|
|Any breast cancer|
|ER positive||69||26||38||14||0.49||< .001||0.60||.003|
|PgR positive||58||27||33||14||0.47||< .001||0.55||.001|
|HR positive||71||26||40||14||0.51||< .001||0.61||.003|
|Invasive breast cancer only|
|Noninvasive breast cancer (DCIS) only|
Abbreviations: ER, estrogen receptor; DCIS, ductal carcinoma in situ; HR, hormone receptor; PgR, progesterone receptor.
See accompanying editorial on page 1249 and article on page 1384; listen to the podcast by Dr Baum at www.jco.org/podcasts
Supported by Public Health Service Grants No. U10-CA-37377, U10-CA-69974, U10-CA-12027, and U10-CA-69651 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, and by the Breast Cancer Research Foundation and AstraZeneca.
Presented in part at the 26th Annual San Antonio Breast Cancer Symposium, December 3-6, 2003, San Antonio, TX.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical trial information can be found for the following: PDQ: NSABP-B-24.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: D. Lawrence Wickerham, AstraZeneca Research Funding: Joseph P. Costantino, National Cancer Institute Expert Testimony: None Other Remuneration: None
Conception and design: D. Craig Allred, Stewart J. Anderson, Soonmyung Paik, Sandra M. Swain, Eleftherios P. Mamounas, Joseph P. Costantino, Norman Wolmark
Administrative support: Soonmyung Paik, Charles E. Geyer Jr, Joseph P. Costantino, Norman Wolmark
Provision of study materials or patients: Soonmyung Paik
Collection and assembly of data: Stewart J. Anderson, Soonmyung Paik, D. Lawrence Wickerham, Eleftherios P. Mamounas, Joseph P. Costantino, Stephanie R. Land
Data analysis and interpretation: Stewart J. Anderson, Soonmyung Paik, Iris D. Nagtegaal, Eleftherios P. Mamounas, Thomas B. Julian, Charles E. Geyer Jr, Joseph P. Costantino, Stephanie R. Land
Manuscript writing: All authors
Final approval of manuscript: All authors