Appropriateness of breast-conserving treatment of breast carcinoma in women with germline mutations in BRCA1 or BRCA2

A clinic-based series

Authors


Abstract

BACKGROUND

Although BRCA1 and BRCA2 were identified in 1994 and 1995, respectively, to the authors' knowledge the optimal management of women with BRCA-associated breast carcinoma remains incompletely defined. The current study evaluates the appropriateness of breast-conserving therapy (BCT) in women with BRCA mutations.

METHODS

Between May 1992 and October 2003, 87 female participants in genetic testing protocols were identified who 1) were found to have deleterious germline BRCA mutations and 2) reported a history of invasive breast carcinoma that was treated with wide local excision and radiation therapy. Clinical records were reviewed and follow-up was updated.

RESULTS

The 87 subjects underwent BCT for 95 invasive breast tumors (8 women received BCT for metachronous bilateral tumors). In all 95 treated breasts, the 5-year and 10-year probabilities of metachronous ipsilateral breast carcinoma (MIBC) were 11.2% and 13.6%, respectively. Among the 87 subjects, the 5-year and 10-year probabilities of metachronous contralateral breast carcinoma (CBC) after treatment of the index tumor were 11.9% and 37.6%. No clinical factors were identified that were associated with either MIBC or CBC, including the use of tamoxifen or chemotherapy.

CONCLUSIONS

Women with BRCA-associated breast carcinoma who undergo BCT appear to have risks of MIBC that are similar to those reported for young women without known mutations. The indications for unilateral mastectomy in this group of women should be the same as those for women with nonhereditary carcinoma. However, significant risks of CBC and possibly late MIBC may prompt the serious consideration of bilateral mastectomy as a preventive measure. Cancer 2005. © 2004 American Cancer Society.

The discovery of BRCA1 and BRCA2 in 1994 and 1995, respectively,1, 2 raised a number of questions regarding the optimal management of women with hereditary breast carcinoma. The appropriateness of breast-conserving therapy (BCT) for women with hereditary breast carcinoma is one such area of ongoing discussion.3 A number of reports have described short-term (5-year) risks of ipsilateral tumor development that are similar to those experienced by women who do not have a hereditary predisposition but increased risks developing a new contralateral primary tumor.4–12 However, in other studies, an increased risk of metachronous ipsilateral disease has been observed.13–16 These conflicting data are a source of confusion for women with BRCA mutations and a recently diagnosed breast tumor who are choosing between the dramatically different alternatives of BCT (lumpectomy and radiation) and bilateral mastectomy. A prospective randomized trial would be optimal to determine the relative appropriateness of these strategies. However, such a study is unlikely to be acceptable to women with mutations. Therefore, in the current study we attempted to explore this issue further by describing the outcomes of a prospectively followed cohort of women with invasive breast carcinoma and BRCA mutations who were treated with BCT.

MATERIALS AND METHODS

Patients

The subjects of this report were women who were identified as carriers of deleterious BRCA1 or BRCA2 mutations by the Clinical Genetics Service at Memorial Sloan-Kettering Cancer Center between May 1992 and August 2004 and who underwent BCT for breast carcinoma. Women provided written informed consent for genetic testing and follow-up as part of studies approved by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center. Mutations were identified either by complete coding sequence analysis of DNA extracted from peripheral blood leukocytes or by targeted analysis of specific founder mutations. All mutations were confirmed by direct DNA sequencing. The subjects of this study had BRCA1 or BRCA2 mutations that were predicted to result in a truncated protein product or that have been generally accepted to be deleterious by virtue of the predicted effect on gene function. No individuals with genetic variants of uncertain significance were included in the current study.

Data Collection and Follow-Up

Baseline information regarding patient characteristics and family history were derived from baseline questionnaires completed as part of the specific protocol in which the subject was enrolled, from the genetic counseling record, and from the clinical record. Information regarding tumor characteristics was extracted from the pathology report, with the final stage of disease at the time of diagnosis being assigned based on the pathology report and the clinical record. Because the majority of women were diagnosed before the 2003 revision of the American Joint Committee on Cancer (AJCC) staging system for breast carcinoma,17 and sufficient information for restaging was not always available, stage assignment according to the 1988 AJCC system was retained for the purposes of this report. Information regarding treatment was extracted from the baseline information provided by the subject during study enrollment and from the medical record. Follow-up and vital status were determined through a review of the clinical and research records and, when necessary, through telephone contact with the patient, next-of-kin, or primary physician.

Statistical Analysis

All survival data were calculated using the Kaplan–Meier method, without correction for competing causes of death. Freedom from metachronous ipsilateral breast carcinoma (MIBC) was defined as the time from the initial diagnosis of breast carcinoma to pathologic confirmation of a second malignancy within the treated breast, bilateral prophylactic mastectomy, or last follow-up. It was not possible in all cases to reliably distinguish between a true recurrence (near the site of original tumor) and a second primary malignancy within the same breast; therefore, all metachronous ipsilateral tumors were considered identically.

Distant disease-free survival (DDFS) was defined as the time from the first breast carcinoma diagnosis to the development of distant metastases or last follow-up. Survival free from contralateral breast carcinoma (CBC) and ovarian carcinoma-free survival were defined as the time from the first diagnosis of breast carcinoma to the time of pathologic diagnosis of the relevant second malignancy, prophylactic removal of the at-risk organs, or last follow-up. Breast carcinoma-specific survival (BCSS) was defined as the time from diagnosis to last follow-up or death from breast carcinoma, and overall survival (OS) was defined as the time from diagnosis to death from any cause.

Univariate analyses evaluating the effect of patient characteristics and clinical variables on time to the endpoint of interest (MIBC or CBC) were performed utilizing two-sided log-rank tests. For continuous variables such as age and time from diagnosis to genetic testing, comparisons between groups were performed using nonparametric tests. All P values are two-tailed, with α = 0.05. Statistical analyses were performed using SPSS for Windows, version 10.0 (SPSS Inc., Chicago, IL).

RESULTS

Patient and Tumor Characteristics

The subjects of the current study were 87 women with germline BRCA mutations who underwent BCT for invasive breast carcinoma. These women underwent BCT for 95 invasive tumors, with 8 women undergoing wide local excision and radiation therapy for a second, metachronous contralateral cancer. The characteristics of the study subjects and their tumors are presented in Table 1. The median age at the time of the first diagnosis was 43 years, with no significant difference noted between women with BRCA1 or BRCA2 mutations. The median interval from the time of diagnosis to genetic testing was 34 months (range, -60–225 months), with 37 women (42.5%) being tested within 2 years of their first diagnosis of breast carcinoma. The median follow-up from the time of first breast carcinoma diagnosis for the 87 subjects was 76 months (range, 12.1–251.8 months).

Table 1. Patient and Tumor Characteristics
FeaturePatients (n = 87)Tumors (n = 95)
  1. NOS: not otherwise specified; ER: estrogen receptor; PR: progesterone receptor; Gy: grays.

Mutation  
 BRCA162 (71.3%)68 (71.6%)
 BRCA225 (28.7%)27 (28.4%)
Median age at initial diagnosis (yrs) (range)43 (27–82) 
Median interval from diagnosis to testing (mos) (range)34 (−60–225) 
Family history of breast carcinoma  
 First-degree64 (73.6%) 
 Second-degree only14 (16.1%) 
 None9 (10.3%) 
Histology  
 Infiltrating ductal, NOS 84 (88.4%)
 Medullary/atypical medullary 7 (7.4%)
 Tubular 1 (1.1%)
 Infiltrating lobular 3 (3.2%)
Median tumor size (cm) (range) 1.4 (0.1–5.0)
T classification  
 T1 71 (74.7%)
 T2 22 (23.2%)
 Unknown 1 (1.1%)
Histologic grade  
 1 2 (2.1%)
 2 7 (7.4%)
 3 70 (73.7%)
 Unknown 16 (16.8%)
Hormone receptor (ER or PR)  
 Positive 34 (35.8%)
 Negative 45 (47.4%)
 Unknown 16 (16.8%)
Axillary lymph nodes  
 Positive 28 (29.5%)
 Negative 63 (66.3%)
 Unknown 4 (4.2%)
Median radiation dose (Gy) (range)  
 Whole breast 46.8 (40–54)
 Boost 14 (0–22.8)
Treatment  
 Chemotherapy (± tamoxifen) 70 (73.7%)
 Tamoxifen (± chemotherapy) 28 (29.5%)

The characteristics of the breast tumors diagnosed in the study subjects are also presented in Table 1. The predominant histology was infiltrating ductal carcinoma, not otherwise specified (88.4%), with a minority of cases being described as medullary or atypical medullary carcinoma (7.4%). In general, the tumors were in a relatively early stage at the time of diagnosis, with 75% being ≤ 2 cm in maximum dimension and only 28 tumors (29.5%) being associated with axillary lymph node metastases. Of the 28 tumors associated with axillary lymph node metastases, 1–3 lymph nodes were involved in 25 cases (89.3%), 4–9 lymph nodes were involved in 2 cases (7.1%), and ≥ 10 lymph nodes were involved in 1 case (3.6%). Despite these favorable tumor characteristics, the subjects received adjuvant chemotherapy after diagnosis in 73.7% of cases. Adjuvant tamoxifen was given after 28 diagnoses (29.5%). The median dose of radiation delivered to the whole breast was 46.8 grays (Gy) and the median boost dose was 14 Gy. No patient received brachytherapy.

Outcomes

MIBCs were diagnosed after the treatment of 12 invasive breast tumors (12.6%) at a median of 51.8 months after the initial diagnosis was made (range, 16.2–182.7 months). The crude average annual incidence of MIBC was 18.2 per 1000 woman-years. The 5-year, 10-year, and 15-year probabilities of remaining free of MIBC were 88.8%, 86.4%, and 76.6%, respectively (Fig. 1). Of the 12 cases of MIBC, 8 (66.7%) were ductal carcinoma in situ (DCIS) and 4 cases (33.3%) were invasive carcinomas. In 7 cases (58.3%) the MIBCs were diagnosed in BRCA1 mutation carriers and in 5 cases (41.7%) the diagnosis was made in BRCA2 carriers. Patient age at diagnosis, gene mutation (BRCA1 vs. BRCA2), tumor size, lymph node status, hormone receptor status, histologic grade, and treatment (chemotherapy and/or tamoxifen) were all examined as possible predictors of MIBC in BRCA mutation carriers with invasive disease who received radiation therapy. None of these variables were found to be statistically significantly associated with MIBC. The mean interval from the time of the initial diagnosis to genetic testing was longer in women with MIBC than in those without (74.9 months vs. 43.1 months; P = 0.04).

Figure 1.

Probability of remaining free of metachronous ipsilateral breast carcinoma.

Of the 87 women in the current study, 20 (23.0%) developed a CBC at a median of 67.4 months (range, 18.6–214.4 months) after their first diagnosis. The crude average annual incidence of CBC was 39.3 per 1000 woman-years. The 5-year, 10-year, and 15-year probabilities of remaining free of CBC were 88.1%, 62.4%, and 46.8%, respectively (Fig. 2). In 15 cases (75.0%) the contralateral tumors were diagnosed in BRCA1 mutation carriers, and 5 cases (25.0%) were diagnosed in BRCA2 mutation carriers. Patient age at the time of the first diagnosis of breast carcinoma, gene mutation (BRCA1 vs. BRCA2), hormone receptor status, histologic grade, and treatment (chemotherapy and/or tamoxifen) were all examined as possible predictors of CBC in BRCA mutation carriers. None were found to be statistically significantly associated with CBC. The mean interval from the time of the initial diagnosis to genetic testing was longer in women with CBC than in those without (81.1 months vs. 39.4 months; P = 0.001), but there was no difference noted with regard to the probability of being free of CBC at 5 years in women whose initial diagnosis was made < 2 years before testing compared with women whose index diagnosis was made ≥ 2 years before testing (85.8% vs. 89.4%; P = 0.60 by the log-rank test).

Figure 2.

Probability of remaining free of metachronous contralateral breast carcinoma (all patients).

Among the 87 subjects in the current study, 8 cases of ovarian carcinoma were diagnosed during 482 woman-years of follow-up with ovarian tissue at risk (crude average annual incidence of 16.6 per 1000 women). The 5-year and 10-year probabilities of remaining free of ovarian carcinoma were 95.1% and 84.9%, respectively. All ovarian tumors were diagnosed in BRCA1 mutation carriers.

The DDFS, event-free survival (EFS), BCSS, and OS rates were 92.9%, 70.7%, 96.9%, and 95.6%, respectively, at 5 years after the initial diagnosis of breast carcinoma. At 10 years after the initial diagnosis, the DDFS, EFS, BCSS, and OS rates were reported to be 87.3%, 42.7%, 90.1%, and 89.4%, respectively.

DISCUSSION

After nearly a decade of investigation, the optimal local therapy for women with BRCA-associated breast carcinoma remains controversial. In particular, there are conflicting data regarding the degree to which a genetic predisposition toward breast carcinoma results in an increased risk of metachronous ipsilateral tumor development after BCT. A number of studies published to date have examined the relation between a family history of breast carcinoma and the risk of ipsilateral disease after wide local excision and adjuvant radiation therapy.18–23 To our knowledge, none of these reports has reported family history to be an independent predictor of disease recurrence. Studies of documented BRCA mutation carriers have also generally been reassuring, with reported ipsilateral recurrence risks similar to those experienced by other women without mutations (Table 2). Previous reports from the Memorial Sloan-Kettering Cancer Center have noted a negative influence of germline BRCA status on BCSS, without a statistically significant influence noted on MIBC independent of young age at the time of diagnosis.7, 11, 12 However, to our knowledge, most studies to date have described relatively small numbers of mutation carriers, with limited follow-up, and two reports from one group have sounded a more cautionary note. In one study by Turner et al.,13 52 patients who developed an ipsilateral tumor recurrence at a median of 4.75 years after BCT were matched to controls who had not developed a disease recurrence. BRCA mutations were significantly more common among the 15 index (recurrence) patients who were initially diagnosed before age 40 years compared with their matched controls (6 of 15 index patients [40%] vs. 1 of 15 matched controls [6.7%]; P = 0.03), with clinical characteristics suggesting that the majority of the second tumors in the mutation carriers were second primary malignancies. In a later report from the same group, 127 women with conservatively treated early-onset breast carcinoma underwent genetic testing at a median of 12.7 years after their diagnosis.14 Among the 22 mutation carriers identified in this study, the actuarial risk of developing an ipsilateral tumor of the breast at 12 years was 49%, although only 6 breasts were at risk at this time point, which may limit the robustness of the estimate. Nonetheless, both of these studies13, 14 suggest that women with BRCA mutations may be at a significant risk for the development of ipsilateral second primary malignancies after BCT. The magnitude of the risk may be exaggerated by the designs of these studies, because women underwent testing after surviving for a prolonged period of time after their initial diagnosis, and thus after being at risk for an extended period. Unfortunately, the actuarial risk to a recently diagnosed mutation carrier may be significantly lower than reported in these studies because there is a risk that she will not survive her breast carcinoma long enough to experience a delayed risk of developing a ipsilateral second primary tumor.

Table 2. Studies of BCT in BRCA Carriers
SeriesAscertainmentGenesnMIBC (no.)Risk of MIBC (%)
  1. BCT: breast-conserving treatment; MIBC: metachronous ipsilateral breast carcinoma; AJ: Ashkenazi Jewish; NS: not stated.

Foulkes et al., 19974Unselected AJBRCA111220 (5-yr)
Robson et al., 199827Early-onset AJBRCA1/291NS
Verhoog et al., 19986Clinic-basedBRCA118NS14 (5-yr)
Gaffney et al., 19985Clinic-basedBRCA1/271NS
Robson et al., 19997Unselected AJBRCA1/2355  15 (5-yr)22 (10-yr)
Haffty et al., 200214Early-onsetBRCA1/22211  22 (5-yr) 41 (10-yr)
Pierce et al., 20039Clinic-basedBRCA1/2170 12.5 (10-yr)
Bremer et al., 200315Hospital-basedBRCA1/29 (bilateral)329 (5-yr)
Delaloge et al., 200310Clinic-basedBRCA1/237 BRCA116 BRCA2NS   9 (10-yr)37 (10-yr)
Seynaeve et al., 200416Clinic-basedBRCA1/2264NS for BRCA
Robson et al., 200411Unselected AJBRCA1/261NS12 (10-yr)
Metcalfe et al., 200412Clinic-basedBRCA1/2188NS11.5 (10-yr)
Current studyClinic-basedBRCA1/2951211.2 (5-yr)13.6 (10-yr)

The current report, which describes outcomes in a clinic-based series of BRCA mutation carriers undergoing BCT for breast carcinoma, provides some reassurance to women with BRCA-associated breast carcinoma. In this series, the 5-year and 10-year risks of MIBC after treatment of invasive breast carcinoma were 11.2% and 13.6%, respectively. No clinical variables were found to be predictive of the occurrence of an MIBC, and risk did not appear to be significantly reduced by tamoxifen or chemotherapy. It is possible that the clinic-based design of this study may have underestimated the risk of MIBC if, for some reason, women with mutations experiencing an MIBC were less likely to present for genetic testing, perhaps because of reduced survival. With regard to this possibility, it is reassuring that the risk of MIBC reported in the current study is similar to that experienced by mutation carriers in previous anonymized, hospital-based series of unselected women of Ashkenazi descent,7, 11 and in 2 large, clinic-based collaborative ascertainments,9, 12 albeit substantially lower than the 41% risk at 10 years described in the report of Haffty et al.14 However, given the small number of breasts at risk at longer follow-up times in all studies, it is entirely possible that the 95% confidence intervals of the risk estimates would overlap.

As in other reports, the risk of CBC was significant in women with BRCA mutations, reaching 37.6% at 10 years. Again, no clinical variables were found to be predictive of the development of CBC, although the sample size was small and the power to identify such factors limited. Later patient age at the time of the initial diagnosis was not associated with a reduced risk of CBC, in contradistinction to the experience reported at another center.24 Although adjuvant treatment with either chemotherapy or tamoxifen did not appear to modify risk, the number of women receiving tamoxifen was too small to exclude a benefit.

It is interesting to note that the majority of MIBCs were DCIS, whereas nearly all CBCs were invasive. This raises the possibility that adjuvant radiation therapy may retard the development of invasive disease in the treated breast by eliminating clinically occult foci of early malignant disease. New primary tumors still may develop as a result of the genetic predisposition, but they presumably do so by progressing while the patient is under careful surveillance for the treated breast, increasing the likelihood that the newly developing malignancy will be identified at a preinvasive stage. However, on the contralateral, untreated side, clinically occult foci would not be retarded by radiation therapy, and may progress to invasive malignancy relatively quickly. This model would explain the observed differences noted in the rates of ipsilateral and contralateral tumors and the differences reported with regard to the ratio of DCIS to invasive carcinoma between MIBC and CBC. It also is consistent with the observation by other authors of a significant risk of developing a late second primary tumor within the treated breast.

The assessment of outcomes in prevalent cohorts of affected women with germline mutations may be distorted by an ascertainment bias introduced by the fact that women undergoing testing have, by definition, survived some period of time since their breast carcinoma diagnosis.25, 26 Such a bias largely influences the assessment of survival endpoints, but a similar bias may affect the evaluation of other outcomes. For example, women who have been diagnosed with bilateral breast carcinoma or both breast and ovarian carcinoma may be more likely to undergoing genetic testing and therefore be identified as mutation carriers, which may increase the apparent frequency of these events in a prevalent ascertainment. One also must be open to the possibility that the experience of an ipsilateral recurrence also may prompt women to seek risk assessment. In the current study, the mean time from diagnosis to genetic testing was longer in women experiencing MIBC or CBC compared with those not experiencing these outcomes, a finding that likely reflects the fact that women with a longer survival are more likely to experience these events because they are alive to do so. Nonetheless, it is also possible that a bias was introduced by the study of breast carcinoma survivors who were more likely to present for testing because of MIBC or CBC. Such a bias would lead to an overestimate, rather than an underestimate, of the risk of MIBC and CBC. Therefore, the figures derived from this experience likely represent the upper bounds of risks for these events.

Although the current study is focused on issues of MIBC and CBC, it is worth highlighting the significant risk for other events experienced by the women in this study. As mentioned, the apparent incidence of ovarian carcinoma is likely exaggerated by ascertainment bias, and the OS and BCSS rates may be overestimated for the same reason. Nevertheless, the women who were the subjects of the current study had a very high probability of developing some malignancy-related event after their initial diagnosis, with only 42.7% reported to remain event free 10 years after their index diagnosis, despite the relatively favorable prognostic features exhibited by their breast tumors. This substantial risk must be borne in mind by physicians caring for women with BRCA mutations, even if the prognosis of the index breast carcinoma appears to be favorable.

The current study describes a large series of women undergoing BCT for BRCA-associated breast carcinoma. As a prevalent ascertainment, the experience may be subject to biases that would tend to overestimate the risks of ipsilateral breast carcinoma and CBC and ovarian carcinoma, and to underestimate breast carcinoma-specific mortality in these women. However, the relatively low observed risk of MIBC over the short to intermediate term (5–10 years) suggests that BCT is a reasonable option for these women, and that the indications for unilateral mastectomy should be the same for both hereditary and nonhereditary breast carcinoma patients. Nevertheless, discussion of bilateral mastectomy is warranted because of the significant risk of CBC, particularly in light of other studies that have suggested the possibility of a substantial risk of MIBC because of the development of ipsilateral second primary malignancies in long-term survivors. Although the number of women being treated with tamoxifen was too small to exclude an effect on MIBC and CBC, no such benefit was evident, and additional research is necessary to identify nonsurgical prevention alternatives for these women at risk.

Ancillary